--- /dev/null
+<?php\r
+\r
+/**\r
+ * @file\r
+ * This file was auto-generated by generate-includes.php and includes all of\r
+ * the core files required by HTML Purifier. Use this if performance is a\r
+ * primary concern and you are using an opcode cache. PLEASE DO NOT EDIT THIS\r
+ * FILE, changes will be overwritten the next time the script is run.\r
+ *\r
+ * @version 4.11.0\r
+ *\r
+ * @warning\r
+ * You must *not* include any other HTML Purifier files before this file,\r
+ * because 'require' not 'require_once' is used.\r
+ *\r
+ * @warning\r
+ * This file requires that the include path contains the HTML Purifier\r
+ * library directory; this is not auto-set.\r
+ */\r
+\r
+\r
+\r
+/*! @mainpage\r
+ *\r
+ * HTML Purifier is an HTML filter that will take an arbitrary snippet of\r
+ * HTML and rigorously test, validate and filter it into a version that\r
+ * is safe for output onto webpages. It achieves this by:\r
+ *\r
+ * -# Lexing (parsing into tokens) the document,\r
+ * -# Executing various strategies on the tokens:\r
+ * -# Removing all elements not in the whitelist,\r
+ * -# Making the tokens well-formed,\r
+ * -# Fixing the nesting of the nodes, and\r
+ * -# Validating attributes of the nodes; and\r
+ * -# Generating HTML from the purified tokens.\r
+ *\r
+ * However, most users will only need to interface with the HTMLPurifier\r
+ * and HTMLPurifier_Config.\r
+ */\r
+\r
+/*\r
+ HTML Purifier 4.11.0 - Standards Compliant HTML Filtering\r
+ Copyright (C) 2006-2008 Edward Z. Yang\r
+\r
+ This library is free software; you can redistribute it and/or\r
+ modify it under the terms of the GNU Lesser General Public\r
+ License as published by the Free Software Foundation; either\r
+ version 2.1 of the License, or (at your option) any later version.\r
+\r
+ This library is distributed in the hope that it will be useful,\r
+ but WITHOUT ANY WARRANTY; without even the implied warranty of\r
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU\r
+ Lesser General Public License for more details.\r
+\r
+ You should have received a copy of the GNU Lesser General Public\r
+ License along with this library; if not, write to the Free Software\r
+ Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA\r
+ */\r
+\r
+/**\r
+ * Facade that coordinates HTML Purifier's subsystems in order to purify HTML.\r
+ *\r
+ * @note There are several points in which configuration can be specified\r
+ * for HTML Purifier. The precedence of these (from lowest to\r
+ * highest) is as follows:\r
+ * -# Instance: new HTMLPurifier($config)\r
+ * -# Invocation: purify($html, $config)\r
+ * These configurations are entirely independent of each other and\r
+ * are *not* merged (this behavior may change in the future).\r
+ *\r
+ * @todo We need an easier way to inject strategies using the configuration\r
+ * object.\r
+ */\r
+class HTMLPurifier\r
+{\r
+\r
+ /**\r
+ * Version of HTML Purifier.\r
+ * @type string\r
+ */\r
+ public $version = '4.11.0';\r
+\r
+ /**\r
+ * Constant with version of HTML Purifier.\r
+ */\r
+ const VERSION = '4.11.0';\r
+\r
+ /**\r
+ * Global configuration object.\r
+ * @type HTMLPurifier_Config\r
+ */\r
+ public $config;\r
+\r
+ /**\r
+ * Array of extra filter objects to run on HTML,\r
+ * for backwards compatibility.\r
+ * @type HTMLPurifier_Filter[]\r
+ */\r
+ private $filters = array();\r
+\r
+ /**\r
+ * Single instance of HTML Purifier.\r
+ * @type HTMLPurifier\r
+ */\r
+ private static $instance;\r
+\r
+ /**\r
+ * @type HTMLPurifier_Strategy_Core\r
+ */\r
+ protected $strategy;\r
+\r
+ /**\r
+ * @type HTMLPurifier_Generator\r
+ */\r
+ protected $generator;\r
+\r
+ /**\r
+ * Resultant context of last run purification.\r
+ * Is an array of contexts if the last called method was purifyArray().\r
+ * @type HTMLPurifier_Context\r
+ */\r
+ public $context;\r
+\r
+ /**\r
+ * Initializes the purifier.\r
+ *\r
+ * @param HTMLPurifier_Config|mixed $config Optional HTMLPurifier_Config object\r
+ * for all instances of the purifier, if omitted, a default\r
+ * configuration is supplied (which can be overridden on a\r
+ * per-use basis).\r
+ * The parameter can also be any type that\r
+ * HTMLPurifier_Config::create() supports.\r
+ */\r
+ public function __construct($config = null)\r
+ {\r
+ $this->config = HTMLPurifier_Config::create($config);\r
+ $this->strategy = new HTMLPurifier_Strategy_Core();\r
+ }\r
+\r
+ /**\r
+ * Adds a filter to process the output. First come first serve\r
+ *\r
+ * @param HTMLPurifier_Filter $filter HTMLPurifier_Filter object\r
+ */\r
+ public function addFilter($filter)\r
+ {\r
+ trigger_error(\r
+ 'HTMLPurifier->addFilter() is deprecated, use configuration directives' .\r
+ ' in the Filter namespace or Filter.Custom',\r
+ E_USER_WARNING\r
+ );\r
+ $this->filters[] = $filter;\r
+ }\r
+\r
+ /**\r
+ * Filters an HTML snippet/document to be XSS-free and standards-compliant.\r
+ *\r
+ * @param string $html String of HTML to purify\r
+ * @param HTMLPurifier_Config $config Config object for this operation,\r
+ * if omitted, defaults to the config object specified during this\r
+ * object's construction. The parameter can also be any type\r
+ * that HTMLPurifier_Config::create() supports.\r
+ *\r
+ * @return string Purified HTML\r
+ */\r
+ public function purify($html, $config = null)\r
+ {\r
+ // :TODO: make the config merge in, instead of replace\r
+ $config = $config ? HTMLPurifier_Config::create($config) : $this->config;\r
+\r
+ // implementation is partially environment dependant, partially\r
+ // configuration dependant\r
+ $lexer = HTMLPurifier_Lexer::create($config);\r
+\r
+ $context = new HTMLPurifier_Context();\r
+\r
+ // setup HTML generator\r
+ $this->generator = new HTMLPurifier_Generator($config, $context);\r
+ $context->register('Generator', $this->generator);\r
+\r
+ // set up global context variables\r
+ if ($config->get('Core.CollectErrors')) {\r
+ // may get moved out if other facilities use it\r
+ $language_factory = HTMLPurifier_LanguageFactory::instance();\r
+ $language = $language_factory->create($config, $context);\r
+ $context->register('Locale', $language);\r
+\r
+ $error_collector = new HTMLPurifier_ErrorCollector($context);\r
+ $context->register('ErrorCollector', $error_collector);\r
+ }\r
+\r
+ // setup id_accumulator context, necessary due to the fact that\r
+ // AttrValidator can be called from many places\r
+ $id_accumulator = HTMLPurifier_IDAccumulator::build($config, $context);\r
+ $context->register('IDAccumulator', $id_accumulator);\r
+\r
+ $html = HTMLPurifier_Encoder::convertToUTF8($html, $config, $context);\r
+\r
+ // setup filters\r
+ $filter_flags = $config->getBatch('Filter');\r
+ $custom_filters = $filter_flags['Custom'];\r
+ unset($filter_flags['Custom']);\r
+ $filters = array();\r
+ foreach ($filter_flags as $filter => $flag) {\r
+ if (!$flag) {\r
+ continue;\r
+ }\r
+ if (strpos($filter, '.') !== false) {\r
+ continue;\r
+ }\r
+ $class = "HTMLPurifier_Filter_$filter";\r
+ $filters[] = new $class;\r
+ }\r
+ foreach ($custom_filters as $filter) {\r
+ // maybe "HTMLPurifier_Filter_$filter", but be consistent with AutoFormat\r
+ $filters[] = $filter;\r
+ }\r
+ $filters = array_merge($filters, $this->filters);\r
+ // maybe prepare(), but later\r
+\r
+ for ($i = 0, $filter_size = count($filters); $i < $filter_size; $i++) {\r
+ $html = $filters[$i]->preFilter($html, $config, $context);\r
+ }\r
+\r
+ // purified HTML\r
+ $html =\r
+ $this->generator->generateFromTokens(\r
+ // list of tokens\r
+ $this->strategy->execute(\r
+ // list of un-purified tokens\r
+ $lexer->tokenizeHTML(\r
+ // un-purified HTML\r
+ $html,\r
+ $config,\r
+ $context\r
+ ),\r
+ $config,\r
+ $context\r
+ )\r
+ );\r
+\r
+ for ($i = $filter_size - 1; $i >= 0; $i--) {\r
+ $html = $filters[$i]->postFilter($html, $config, $context);\r
+ }\r
+\r
+ $html = HTMLPurifier_Encoder::convertFromUTF8($html, $config, $context);\r
+ $this->context =& $context;\r
+ return $html;\r
+ }\r
+\r
+ /**\r
+ * Filters an array of HTML snippets\r
+ *\r
+ * @param string[] $array_of_html Array of html snippets\r
+ * @param HTMLPurifier_Config $config Optional config object for this operation.\r
+ * See HTMLPurifier::purify() for more details.\r
+ *\r
+ * @return string[] Array of purified HTML\r
+ */\r
+ public function purifyArray($array_of_html, $config = null)\r
+ {\r
+ $context_array = array();\r
+ foreach($array_of_html as $key=>$value){\r
+ if (is_array($value)) {\r
+ $array[$key] = $this->purifyArray($value, $config);\r
+ } else {\r
+ $array[$key] = $this->purify($value, $config);\r
+ }\r
+ $context_array[$key] = $this->context;\r
+ }\r
+ $this->context = $context_array;\r
+ return $array;\r
+ }\r
+\r
+ /**\r
+ * Singleton for enforcing just one HTML Purifier in your system\r
+ *\r
+ * @param HTMLPurifier|HTMLPurifier_Config $prototype Optional prototype\r
+ * HTMLPurifier instance to overload singleton with,\r
+ * or HTMLPurifier_Config instance to configure the\r
+ * generated version with.\r
+ *\r
+ * @return HTMLPurifier\r
+ */\r
+ public static function instance($prototype = null)\r
+ {\r
+ if (!self::$instance || $prototype) {\r
+ if ($prototype instanceof HTMLPurifier) {\r
+ self::$instance = $prototype;\r
+ } elseif ($prototype) {\r
+ self::$instance = new HTMLPurifier($prototype);\r
+ } else {\r
+ self::$instance = new HTMLPurifier();\r
+ }\r
+ }\r
+ return self::$instance;\r
+ }\r
+\r
+ /**\r
+ * Singleton for enforcing just one HTML Purifier in your system\r
+ *\r
+ * @param HTMLPurifier|HTMLPurifier_Config $prototype Optional prototype\r
+ * HTMLPurifier instance to overload singleton with,\r
+ * or HTMLPurifier_Config instance to configure the\r
+ * generated version with.\r
+ *\r
+ * @return HTMLPurifier\r
+ * @note Backwards compatibility, see instance()\r
+ */\r
+ public static function getInstance($prototype = null)\r
+ {\r
+ return HTMLPurifier::instance($prototype);\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Converts a stream of HTMLPurifier_Token into an HTMLPurifier_Node,\r
+ * and back again.\r
+ *\r
+ * @note This transformation is not an equivalence. We mutate the input\r
+ * token stream to make it so; see all [MUT] markers in code.\r
+ */\r
+class HTMLPurifier_Arborize\r
+{\r
+ public static function arborize($tokens, $config, $context) {\r
+ $definition = $config->getHTMLDefinition();\r
+ $parent = new HTMLPurifier_Token_Start($definition->info_parent);\r
+ $stack = array($parent->toNode());\r
+ foreach ($tokens as $token) {\r
+ $token->skip = null; // [MUT]\r
+ $token->carryover = null; // [MUT]\r
+ if ($token instanceof HTMLPurifier_Token_End) {\r
+ $token->start = null; // [MUT]\r
+ $r = array_pop($stack);\r
+ //assert($r->name === $token->name);\r
+ //assert(empty($token->attr));\r
+ $r->endCol = $token->col;\r
+ $r->endLine = $token->line;\r
+ $r->endArmor = $token->armor;\r
+ continue;\r
+ }\r
+ $node = $token->toNode();\r
+ $stack[count($stack)-1]->children[] = $node;\r
+ if ($token instanceof HTMLPurifier_Token_Start) {\r
+ $stack[] = $node;\r
+ }\r
+ }\r
+ //assert(count($stack) == 1);\r
+ return $stack[0];\r
+ }\r
+\r
+ public static function flatten($node, $config, $context) {\r
+ $level = 0;\r
+ $nodes = array($level => new HTMLPurifier_Queue(array($node)));\r
+ $closingTokens = array();\r
+ $tokens = array();\r
+ do {\r
+ while (!$nodes[$level]->isEmpty()) {\r
+ $node = $nodes[$level]->shift(); // FIFO\r
+ list($start, $end) = $node->toTokenPair();\r
+ if ($level > 0) {\r
+ $tokens[] = $start;\r
+ }\r
+ if ($end !== NULL) {\r
+ $closingTokens[$level][] = $end;\r
+ }\r
+ if ($node instanceof HTMLPurifier_Node_Element) {\r
+ $level++;\r
+ $nodes[$level] = new HTMLPurifier_Queue();\r
+ foreach ($node->children as $childNode) {\r
+ $nodes[$level]->push($childNode);\r
+ }\r
+ }\r
+ }\r
+ $level--;\r
+ if ($level && isset($closingTokens[$level])) {\r
+ while ($token = array_pop($closingTokens[$level])) {\r
+ $tokens[] = $token;\r
+ }\r
+ }\r
+ } while ($level > 0);\r
+ return $tokens;\r
+ }\r
+}\r
+
+\r
+\r
+/**\r
+ * Defines common attribute collections that modules reference\r
+ */\r
+\r
+class HTMLPurifier_AttrCollections\r
+{\r
+\r
+ /**\r
+ * Associative array of attribute collections, indexed by name.\r
+ * @type array\r
+ */\r
+ public $info = array();\r
+\r
+ /**\r
+ * Performs all expansions on internal data for use by other inclusions\r
+ * It also collects all attribute collection extensions from\r
+ * modules\r
+ * @param HTMLPurifier_AttrTypes $attr_types HTMLPurifier_AttrTypes instance\r
+ * @param HTMLPurifier_HTMLModule[] $modules Hash array of HTMLPurifier_HTMLModule members\r
+ */\r
+ public function __construct($attr_types, $modules)\r
+ {\r
+ $this->doConstruct($attr_types, $modules);\r
+ }\r
+\r
+ public function doConstruct($attr_types, $modules)\r
+ {\r
+ // load extensions from the modules\r
+ foreach ($modules as $module) {\r
+ foreach ($module->attr_collections as $coll_i => $coll) {\r
+ if (!isset($this->info[$coll_i])) {\r
+ $this->info[$coll_i] = array();\r
+ }\r
+ foreach ($coll as $attr_i => $attr) {\r
+ if ($attr_i === 0 && isset($this->info[$coll_i][$attr_i])) {\r
+ // merge in includes\r
+ $this->info[$coll_i][$attr_i] = array_merge(\r
+ $this->info[$coll_i][$attr_i],\r
+ $attr\r
+ );\r
+ continue;\r
+ }\r
+ $this->info[$coll_i][$attr_i] = $attr;\r
+ }\r
+ }\r
+ }\r
+ // perform internal expansions and inclusions\r
+ foreach ($this->info as $name => $attr) {\r
+ // merge attribute collections that include others\r
+ $this->performInclusions($this->info[$name]);\r
+ // replace string identifiers with actual attribute objects\r
+ $this->expandIdentifiers($this->info[$name], $attr_types);\r
+ }\r
+ }\r
+\r
+ /**\r
+ * Takes a reference to an attribute associative array and performs\r
+ * all inclusions specified by the zero index.\r
+ * @param array &$attr Reference to attribute array\r
+ */\r
+ public function performInclusions(&$attr)\r
+ {\r
+ if (!isset($attr[0])) {\r
+ return;\r
+ }\r
+ $merge = $attr[0];\r
+ $seen = array(); // recursion guard\r
+ // loop through all the inclusions\r
+ for ($i = 0; isset($merge[$i]); $i++) {\r
+ if (isset($seen[$merge[$i]])) {\r
+ continue;\r
+ }\r
+ $seen[$merge[$i]] = true;\r
+ // foreach attribute of the inclusion, copy it over\r
+ if (!isset($this->info[$merge[$i]])) {\r
+ continue;\r
+ }\r
+ foreach ($this->info[$merge[$i]] as $key => $value) {\r
+ if (isset($attr[$key])) {\r
+ continue;\r
+ } // also catches more inclusions\r
+ $attr[$key] = $value;\r
+ }\r
+ if (isset($this->info[$merge[$i]][0])) {\r
+ // recursion\r
+ $merge = array_merge($merge, $this->info[$merge[$i]][0]);\r
+ }\r
+ }\r
+ unset($attr[0]);\r
+ }\r
+\r
+ /**\r
+ * Expands all string identifiers in an attribute array by replacing\r
+ * them with the appropriate values inside HTMLPurifier_AttrTypes\r
+ * @param array &$attr Reference to attribute array\r
+ * @param HTMLPurifier_AttrTypes $attr_types HTMLPurifier_AttrTypes instance\r
+ */\r
+ public function expandIdentifiers(&$attr, $attr_types)\r
+ {\r
+ // because foreach will process new elements we add, make sure we\r
+ // skip duplicates\r
+ $processed = array();\r
+\r
+ foreach ($attr as $def_i => $def) {\r
+ // skip inclusions\r
+ if ($def_i === 0) {\r
+ continue;\r
+ }\r
+\r
+ if (isset($processed[$def_i])) {\r
+ continue;\r
+ }\r
+\r
+ // determine whether or not attribute is required\r
+ if ($required = (strpos($def_i, '*') !== false)) {\r
+ // rename the definition\r
+ unset($attr[$def_i]);\r
+ $def_i = trim($def_i, '*');\r
+ $attr[$def_i] = $def;\r
+ }\r
+\r
+ $processed[$def_i] = true;\r
+\r
+ // if we've already got a literal object, move on\r
+ if (is_object($def)) {\r
+ // preserve previous required\r
+ $attr[$def_i]->required = ($required || $attr[$def_i]->required);\r
+ continue;\r
+ }\r
+\r
+ if ($def === false) {\r
+ unset($attr[$def_i]);\r
+ continue;\r
+ }\r
+\r
+ if ($t = $attr_types->get($def)) {\r
+ $attr[$def_i] = $t;\r
+ $attr[$def_i]->required = $required;\r
+ } else {\r
+ unset($attr[$def_i]);\r
+ }\r
+ }\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Base class for all validating attribute definitions.\r
+ *\r
+ * This family of classes forms the core for not only HTML attribute validation,\r
+ * but also any sort of string that needs to be validated or cleaned (which\r
+ * means CSS properties and composite definitions are defined here too).\r
+ * Besides defining (through code) what precisely makes the string valid,\r
+ * subclasses are also responsible for cleaning the code if possible.\r
+ */\r
+\r
+abstract class HTMLPurifier_AttrDef\r
+{\r
+\r
+ /**\r
+ * Tells us whether or not an HTML attribute is minimized.\r
+ * Has no meaning in other contexts.\r
+ * @type bool\r
+ */\r
+ public $minimized = false;\r
+\r
+ /**\r
+ * Tells us whether or not an HTML attribute is required.\r
+ * Has no meaning in other contexts\r
+ * @type bool\r
+ */\r
+ public $required = false;\r
+\r
+ /**\r
+ * Validates and cleans passed string according to a definition.\r
+ *\r
+ * @param string $string String to be validated and cleaned.\r
+ * @param HTMLPurifier_Config $config Mandatory HTMLPurifier_Config object.\r
+ * @param HTMLPurifier_Context $context Mandatory HTMLPurifier_Context object.\r
+ */\r
+ abstract public function validate($string, $config, $context);\r
+\r
+ /**\r
+ * Convenience method that parses a string as if it were CDATA.\r
+ *\r
+ * This method process a string in the manner specified at\r
+ * <http://www.w3.org/TR/html4/types.html#h-6.2> by removing\r
+ * leading and trailing whitespace, ignoring line feeds, and replacing\r
+ * carriage returns and tabs with spaces. While most useful for HTML\r
+ * attributes specified as CDATA, it can also be applied to most CSS\r
+ * values.\r
+ *\r
+ * @note This method is not entirely standards compliant, as trim() removes\r
+ * more types of whitespace than specified in the spec. In practice,\r
+ * this is rarely a problem, as those extra characters usually have\r
+ * already been removed by HTMLPurifier_Encoder.\r
+ *\r
+ * @warning This processing is inconsistent with XML's whitespace handling\r
+ * as specified by section 3.3.3 and referenced XHTML 1.0 section\r
+ * 4.7. However, note that we are NOT necessarily\r
+ * parsing XML, thus, this behavior may still be correct. We\r
+ * assume that newlines have been normalized.\r
+ */\r
+ public function parseCDATA($string)\r
+ {\r
+ $string = trim($string);\r
+ $string = str_replace(array("\n", "\t", "\r"), ' ', $string);\r
+ return $string;\r
+ }\r
+\r
+ /**\r
+ * Factory method for creating this class from a string.\r
+ * @param string $string String construction info\r
+ * @return HTMLPurifier_AttrDef Created AttrDef object corresponding to $string\r
+ */\r
+ public function make($string)\r
+ {\r
+ // default implementation, return a flyweight of this object.\r
+ // If $string has an effect on the returned object (i.e. you\r
+ // need to overload this method), it is best\r
+ // to clone or instantiate new copies. (Instantiation is safer.)\r
+ return $this;\r
+ }\r
+\r
+ /**\r
+ * Removes spaces from rgb(0, 0, 0) so that shorthand CSS properties work\r
+ * properly. THIS IS A HACK!\r
+ * @param string $string a CSS colour definition\r
+ * @return string\r
+ */\r
+ protected function mungeRgb($string)\r
+ {\r
+ $p = '\s*(\d+(\.\d+)?([%]?))\s*';\r
+\r
+ if (preg_match('/(rgba|hsla)\(/', $string)) {\r
+ return preg_replace('/(rgba|hsla)\('.$p.','.$p.','.$p.','.$p.'\)/', '\1(\2,\5,\8,\11)', $string);\r
+ }\r
+\r
+ return preg_replace('/(rgb|hsl)\('.$p.','.$p.','.$p.'\)/', '\1(\2,\5,\8)', $string);\r
+ }\r
+\r
+ /**\r
+ * Parses a possibly escaped CSS string and returns the "pure"\r
+ * version of it.\r
+ */\r
+ protected function expandCSSEscape($string)\r
+ {\r
+ // flexibly parse it\r
+ $ret = '';\r
+ for ($i = 0, $c = strlen($string); $i < $c; $i++) {\r
+ if ($string[$i] === '\\') {\r
+ $i++;\r
+ if ($i >= $c) {\r
+ $ret .= '\\';\r
+ break;\r
+ }\r
+ if (ctype_xdigit($string[$i])) {\r
+ $code = $string[$i];\r
+ for ($a = 1, $i++; $i < $c && $a < 6; $i++, $a++) {\r
+ if (!ctype_xdigit($string[$i])) {\r
+ break;\r
+ }\r
+ $code .= $string[$i];\r
+ }\r
+ // We have to be extremely careful when adding\r
+ // new characters, to make sure we're not breaking\r
+ // the encoding.\r
+ $char = HTMLPurifier_Encoder::unichr(hexdec($code));\r
+ if (HTMLPurifier_Encoder::cleanUTF8($char) === '') {\r
+ continue;\r
+ }\r
+ $ret .= $char;\r
+ if ($i < $c && trim($string[$i]) !== '') {\r
+ $i--;\r
+ }\r
+ continue;\r
+ }\r
+ if ($string[$i] === "\n") {\r
+ continue;\r
+ }\r
+ }\r
+ $ret .= $string[$i];\r
+ }\r
+ return $ret;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Processes an entire attribute array for corrections needing multiple values.\r
+ *\r
+ * Occasionally, a certain attribute will need to be removed and popped onto\r
+ * another value. Instead of creating a complex return syntax for\r
+ * HTMLPurifier_AttrDef, we just pass the whole attribute array to a\r
+ * specialized object and have that do the special work. That is the\r
+ * family of HTMLPurifier_AttrTransform.\r
+ *\r
+ * An attribute transformation can be assigned to run before or after\r
+ * HTMLPurifier_AttrDef validation. See HTMLPurifier_HTMLDefinition for\r
+ * more details.\r
+ */\r
+\r
+abstract class HTMLPurifier_AttrTransform\r
+{\r
+\r
+ /**\r
+ * Abstract: makes changes to the attributes dependent on multiple values.\r
+ *\r
+ * @param array $attr Assoc array of attributes, usually from\r
+ * HTMLPurifier_Token_Tag::$attr\r
+ * @param HTMLPurifier_Config $config Mandatory HTMLPurifier_Config object.\r
+ * @param HTMLPurifier_Context $context Mandatory HTMLPurifier_Context object\r
+ * @return array Processed attribute array.\r
+ */\r
+ abstract public function transform($attr, $config, $context);\r
+\r
+ /**\r
+ * Prepends CSS properties to the style attribute, creating the\r
+ * attribute if it doesn't exist.\r
+ * @param array &$attr Attribute array to process (passed by reference)\r
+ * @param string $css CSS to prepend\r
+ */\r
+ public function prependCSS(&$attr, $css)\r
+ {\r
+ $attr['style'] = isset($attr['style']) ? $attr['style'] : '';\r
+ $attr['style'] = $css . $attr['style'];\r
+ }\r
+\r
+ /**\r
+ * Retrieves and removes an attribute\r
+ * @param array &$attr Attribute array to process (passed by reference)\r
+ * @param mixed $key Key of attribute to confiscate\r
+ * @return mixed\r
+ */\r
+ public function confiscateAttr(&$attr, $key)\r
+ {\r
+ if (!isset($attr[$key])) {\r
+ return null;\r
+ }\r
+ $value = $attr[$key];\r
+ unset($attr[$key]);\r
+ return $value;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Provides lookup array of attribute types to HTMLPurifier_AttrDef objects\r
+ */\r
+class HTMLPurifier_AttrTypes\r
+{\r
+ /**\r
+ * Lookup array of attribute string identifiers to concrete implementations.\r
+ * @type HTMLPurifier_AttrDef[]\r
+ */\r
+ protected $info = array();\r
+\r
+ /**\r
+ * Constructs the info array, supplying default implementations for attribute\r
+ * types.\r
+ */\r
+ public function __construct()\r
+ {\r
+ // XXX This is kind of poor, since we don't actually /clone/\r
+ // instances; instead, we use the supplied make() attribute. So,\r
+ // the underlying class must know how to deal with arguments.\r
+ // With the old implementation of Enum, that ignored its\r
+ // arguments when handling a make dispatch, the IAlign\r
+ // definition wouldn't work.\r
+\r
+ // pseudo-types, must be instantiated via shorthand\r
+ $this->info['Enum'] = new HTMLPurifier_AttrDef_Enum();\r
+ $this->info['Bool'] = new HTMLPurifier_AttrDef_HTML_Bool();\r
+\r
+ $this->info['CDATA'] = new HTMLPurifier_AttrDef_Text();\r
+ $this->info['ID'] = new HTMLPurifier_AttrDef_HTML_ID();\r
+ $this->info['Length'] = new HTMLPurifier_AttrDef_HTML_Length();\r
+ $this->info['MultiLength'] = new HTMLPurifier_AttrDef_HTML_MultiLength();\r
+ $this->info['NMTOKENS'] = new HTMLPurifier_AttrDef_HTML_Nmtokens();\r
+ $this->info['Pixels'] = new HTMLPurifier_AttrDef_HTML_Pixels();\r
+ $this->info['Text'] = new HTMLPurifier_AttrDef_Text();\r
+ $this->info['URI'] = new HTMLPurifier_AttrDef_URI();\r
+ $this->info['LanguageCode'] = new HTMLPurifier_AttrDef_Lang();\r
+ $this->info['Color'] = new HTMLPurifier_AttrDef_HTML_Color();\r
+ $this->info['IAlign'] = self::makeEnum('top,middle,bottom,left,right');\r
+ $this->info['LAlign'] = self::makeEnum('top,bottom,left,right');\r
+ $this->info['FrameTarget'] = new HTMLPurifier_AttrDef_HTML_FrameTarget();\r
+\r
+ // unimplemented aliases\r
+ $this->info['ContentType'] = new HTMLPurifier_AttrDef_Text();\r
+ $this->info['ContentTypes'] = new HTMLPurifier_AttrDef_Text();\r
+ $this->info['Charsets'] = new HTMLPurifier_AttrDef_Text();\r
+ $this->info['Character'] = new HTMLPurifier_AttrDef_Text();\r
+\r
+ // "proprietary" types\r
+ $this->info['Class'] = new HTMLPurifier_AttrDef_HTML_Class();\r
+\r
+ // number is really a positive integer (one or more digits)\r
+ // FIXME: ^^ not always, see start and value of list items\r
+ $this->info['Number'] = new HTMLPurifier_AttrDef_Integer(false, false, true);\r
+ }\r
+\r
+ private static function makeEnum($in)\r
+ {\r
+ return new HTMLPurifier_AttrDef_Clone(new HTMLPurifier_AttrDef_Enum(explode(',', $in)));\r
+ }\r
+\r
+ /**\r
+ * Retrieves a type\r
+ * @param string $type String type name\r
+ * @return HTMLPurifier_AttrDef Object AttrDef for type\r
+ */\r
+ public function get($type)\r
+ {\r
+ // determine if there is any extra info tacked on\r
+ if (strpos($type, '#') !== false) {\r
+ list($type, $string) = explode('#', $type, 2);\r
+ } else {\r
+ $string = '';\r
+ }\r
+\r
+ if (!isset($this->info[$type])) {\r
+ trigger_error('Cannot retrieve undefined attribute type ' . $type, E_USER_ERROR);\r
+ return;\r
+ }\r
+ return $this->info[$type]->make($string);\r
+ }\r
+\r
+ /**\r
+ * Sets a new implementation for a type\r
+ * @param string $type String type name\r
+ * @param HTMLPurifier_AttrDef $impl Object AttrDef for type\r
+ */\r
+ public function set($type, $impl)\r
+ {\r
+ $this->info[$type] = $impl;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Validates the attributes of a token. Doesn't manage required attributes\r
+ * very well. The only reason we factored this out was because RemoveForeignElements\r
+ * also needed it besides ValidateAttributes.\r
+ */\r
+class HTMLPurifier_AttrValidator\r
+{\r
+\r
+ /**\r
+ * Validates the attributes of a token, mutating it as necessary.\r
+ * that has valid tokens\r
+ * @param HTMLPurifier_Token $token Token to validate.\r
+ * @param HTMLPurifier_Config $config Instance of HTMLPurifier_Config\r
+ * @param HTMLPurifier_Context $context Instance of HTMLPurifier_Context\r
+ */\r
+ public function validateToken($token, $config, $context)\r
+ {\r
+ $definition = $config->getHTMLDefinition();\r
+ $e =& $context->get('ErrorCollector', true);\r
+\r
+ // initialize IDAccumulator if necessary\r
+ $ok =& $context->get('IDAccumulator', true);\r
+ if (!$ok) {\r
+ $id_accumulator = HTMLPurifier_IDAccumulator::build($config, $context);\r
+ $context->register('IDAccumulator', $id_accumulator);\r
+ }\r
+\r
+ // initialize CurrentToken if necessary\r
+ $current_token =& $context->get('CurrentToken', true);\r
+ if (!$current_token) {\r
+ $context->register('CurrentToken', $token);\r
+ }\r
+\r
+ if (!$token instanceof HTMLPurifier_Token_Start &&\r
+ !$token instanceof HTMLPurifier_Token_Empty\r
+ ) {\r
+ return;\r
+ }\r
+\r
+ // create alias to global definition array, see also $defs\r
+ // DEFINITION CALL\r
+ $d_defs = $definition->info_global_attr;\r
+\r
+ // don't update token until the very end, to ensure an atomic update\r
+ $attr = $token->attr;\r
+\r
+ // do global transformations (pre)\r
+ // nothing currently utilizes this\r
+ foreach ($definition->info_attr_transform_pre as $transform) {\r
+ $attr = $transform->transform($o = $attr, $config, $context);\r
+ if ($e) {\r
+ if ($attr != $o) {\r
+ $e->send(E_NOTICE, 'AttrValidator: Attributes transformed', $o, $attr);\r
+ }\r
+ }\r
+ }\r
+\r
+ // do local transformations only applicable to this element (pre)\r
+ // ex. <p align="right"> to <p style="text-align:right;">\r
+ foreach ($definition->info[$token->name]->attr_transform_pre as $transform) {\r
+ $attr = $transform->transform($o = $attr, $config, $context);\r
+ if ($e) {\r
+ if ($attr != $o) {\r
+ $e->send(E_NOTICE, 'AttrValidator: Attributes transformed', $o, $attr);\r
+ }\r
+ }\r
+ }\r
+\r
+ // create alias to this element's attribute definition array, see\r
+ // also $d_defs (global attribute definition array)\r
+ // DEFINITION CALL\r
+ $defs = $definition->info[$token->name]->attr;\r
+\r
+ $attr_key = false;\r
+ $context->register('CurrentAttr', $attr_key);\r
+\r
+ // iterate through all the attribute keypairs\r
+ // Watch out for name collisions: $key has previously been used\r
+ foreach ($attr as $attr_key => $value) {\r
+\r
+ // call the definition\r
+ if (isset($defs[$attr_key])) {\r
+ // there is a local definition defined\r
+ if ($defs[$attr_key] === false) {\r
+ // We've explicitly been told not to allow this element.\r
+ // This is usually when there's a global definition\r
+ // that must be overridden.\r
+ // Theoretically speaking, we could have a\r
+ // AttrDef_DenyAll, but this is faster!\r
+ $result = false;\r
+ } else {\r
+ // validate according to the element's definition\r
+ $result = $defs[$attr_key]->validate(\r
+ $value,\r
+ $config,\r
+ $context\r
+ );\r
+ }\r
+ } elseif (isset($d_defs[$attr_key])) {\r
+ // there is a global definition defined, validate according\r
+ // to the global definition\r
+ $result = $d_defs[$attr_key]->validate(\r
+ $value,\r
+ $config,\r
+ $context\r
+ );\r
+ } else {\r
+ // system never heard of the attribute? DELETE!\r
+ $result = false;\r
+ }\r
+\r
+ // put the results into effect\r
+ if ($result === false || $result === null) {\r
+ // this is a generic error message that should replaced\r
+ // with more specific ones when possible\r
+ if ($e) {\r
+ $e->send(E_ERROR, 'AttrValidator: Attribute removed');\r
+ }\r
+\r
+ // remove the attribute\r
+ unset($attr[$attr_key]);\r
+ } elseif (is_string($result)) {\r
+ // generally, if a substitution is happening, there\r
+ // was some sort of implicit correction going on. We'll\r
+ // delegate it to the attribute classes to say exactly what.\r
+\r
+ // simple substitution\r
+ $attr[$attr_key] = $result;\r
+ } else {\r
+ // nothing happens\r
+ }\r
+\r
+ // we'd also want slightly more complicated substitution\r
+ // involving an array as the return value,\r
+ // although we're not sure how colliding attributes would\r
+ // resolve (certain ones would be completely overriden,\r
+ // others would prepend themselves).\r
+ }\r
+\r
+ $context->destroy('CurrentAttr');\r
+\r
+ // post transforms\r
+\r
+ // global (error reporting untested)\r
+ foreach ($definition->info_attr_transform_post as $transform) {\r
+ $attr = $transform->transform($o = $attr, $config, $context);\r
+ if ($e) {\r
+ if ($attr != $o) {\r
+ $e->send(E_NOTICE, 'AttrValidator: Attributes transformed', $o, $attr);\r
+ }\r
+ }\r
+ }\r
+\r
+ // local (error reporting untested)\r
+ foreach ($definition->info[$token->name]->attr_transform_post as $transform) {\r
+ $attr = $transform->transform($o = $attr, $config, $context);\r
+ if ($e) {\r
+ if ($attr != $o) {\r
+ $e->send(E_NOTICE, 'AttrValidator: Attributes transformed', $o, $attr);\r
+ }\r
+ }\r
+ }\r
+\r
+ $token->attr = $attr;\r
+\r
+ // destroy CurrentToken if we made it ourselves\r
+ if (!$current_token) {\r
+ $context->destroy('CurrentToken');\r
+ }\r
+\r
+ }\r
+\r
+\r
+}\r
+\r
+
+
+\r
+\r
+// constants are slow, so we use as few as possible\r
+if (!defined('HTMLPURIFIER_PREFIX')) {\r
+ define('HTMLPURIFIER_PREFIX', dirname(__FILE__) . '/standalone');\r
+ set_include_path(HTMLPURIFIER_PREFIX . PATH_SEPARATOR . get_include_path());\r
+}\r
+\r
+// accomodations for versions earlier than 5.0.2\r
+// borrowed from PHP_Compat, LGPL licensed, by Aidan Lister <aidan@php.net>\r
+if (!defined('PHP_EOL')) {\r
+ switch (strtoupper(substr(PHP_OS, 0, 3))) {\r
+ case 'WIN':\r
+ define('PHP_EOL', "\r\n");\r
+ break;\r
+ case 'DAR':\r
+ define('PHP_EOL', "\r");\r
+ break;\r
+ default:\r
+ define('PHP_EOL', "\n");\r
+ }\r
+}\r
+\r
+/**\r
+ * Bootstrap class that contains meta-functionality for HTML Purifier such as\r
+ * the autoload function.\r
+ *\r
+ * @note\r
+ * This class may be used without any other files from HTML Purifier.\r
+ */\r
+class HTMLPurifier_Bootstrap\r
+{\r
+\r
+ /**\r
+ * Autoload function for HTML Purifier\r
+ * @param string $class Class to load\r
+ * @return bool\r
+ */\r
+ public static function autoload($class)\r
+ {\r
+ $file = HTMLPurifier_Bootstrap::getPath($class);\r
+ if (!$file) {\r
+ return false;\r
+ }\r
+ // Technically speaking, it should be ok and more efficient to\r
+ // just do 'require', but Antonio Parraga reports that with\r
+ // Zend extensions such as Zend debugger and APC, this invariant\r
+ // may be broken. Since we have efficient alternatives, pay\r
+ // the cost here and avoid the bug.\r
+ require_once HTMLPURIFIER_PREFIX . '/' . $file;\r
+ return true;\r
+ }\r
+\r
+ /**\r
+ * Returns the path for a specific class.\r
+ * @param string $class Class path to get\r
+ * @return string\r
+ */\r
+ public static function getPath($class)\r
+ {\r
+ if (strncmp('HTMLPurifier', $class, 12) !== 0) {\r
+ return false;\r
+ }\r
+ // Custom implementations\r
+ if (strncmp('HTMLPurifier_Language_', $class, 22) === 0) {\r
+ $code = str_replace('_', '-', substr($class, 22));\r
+ $file = 'HTMLPurifier/Language/classes/' . $code . '.php';\r
+ } else {\r
+ $file = str_replace('_', '/', $class) . '.php';\r
+ }\r
+ if (!file_exists(HTMLPURIFIER_PREFIX . '/' . $file)) {\r
+ return false;\r
+ }\r
+ return $file;\r
+ }\r
+\r
+ /**\r
+ * "Pre-registers" our autoloader on the SPL stack.\r
+ */\r
+ public static function registerAutoload()\r
+ {\r
+ $autoload = array('HTMLPurifier_Bootstrap', 'autoload');\r
+ if (($funcs = spl_autoload_functions()) === false) {\r
+ spl_autoload_register($autoload);\r
+ } elseif (function_exists('spl_autoload_unregister')) {\r
+ if (version_compare(PHP_VERSION, '5.3.0', '>=')) {\r
+ // prepend flag exists, no need for shenanigans\r
+ spl_autoload_register($autoload, true, true);\r
+ } else {\r
+ $buggy = version_compare(PHP_VERSION, '5.2.11', '<');\r
+ $compat = version_compare(PHP_VERSION, '5.1.2', '<=') &&\r
+ version_compare(PHP_VERSION, '5.1.0', '>=');\r
+ foreach ($funcs as $func) {\r
+ if ($buggy && is_array($func)) {\r
+ // :TRICKY: There are some compatibility issues and some\r
+ // places where we need to error out\r
+ $reflector = new ReflectionMethod($func[0], $func[1]);\r
+ if (!$reflector->isStatic()) {\r
+ throw new Exception(\r
+ 'HTML Purifier autoloader registrar is not compatible\r
+ with non-static object methods due to PHP Bug #44144;\r
+ Please do not use HTMLPurifier.autoload.php (or any\r
+ file that includes this file); instead, place the code:\r
+ spl_autoload_register(array(\'HTMLPurifier_Bootstrap\', \'autoload\'))\r
+ after your own autoloaders.'\r
+ );\r
+ }\r
+ // Suprisingly, spl_autoload_register supports the\r
+ // Class::staticMethod callback format, although call_user_func doesn't\r
+ if ($compat) {\r
+ $func = implode('::', $func);\r
+ }\r
+ }\r
+ spl_autoload_unregister($func);\r
+ }\r
+ spl_autoload_register($autoload);\r
+ foreach ($funcs as $func) {\r
+ spl_autoload_register($func);\r
+ }\r
+ }\r
+ }\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Super-class for definition datatype objects, implements serialization\r
+ * functions for the class.\r
+ */\r
+abstract class HTMLPurifier_Definition\r
+{\r
+\r
+ /**\r
+ * Has setup() been called yet?\r
+ * @type bool\r
+ */\r
+ public $setup = false;\r
+\r
+ /**\r
+ * If true, write out the final definition object to the cache after\r
+ * setup. This will be true only if all invocations to get a raw\r
+ * definition object are also optimized. This does not cause file\r
+ * system thrashing because on subsequent calls the cached object\r
+ * is used and any writes to the raw definition object are short\r
+ * circuited. See enduser-customize.html for the high-level\r
+ * picture.\r
+ * @type bool\r
+ */\r
+ public $optimized = null;\r
+\r
+ /**\r
+ * What type of definition is it?\r
+ * @type string\r
+ */\r
+ public $type;\r
+\r
+ /**\r
+ * Sets up the definition object into the final form, something\r
+ * not done by the constructor\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ abstract protected function doSetup($config);\r
+\r
+ /**\r
+ * Setup function that aborts if already setup\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ public function setup($config)\r
+ {\r
+ if ($this->setup) {\r
+ return;\r
+ }\r
+ $this->setup = true;\r
+ $this->doSetup($config);\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Defines allowed CSS attributes and what their values are.\r
+ * @see HTMLPurifier_HTMLDefinition\r
+ */\r
+class HTMLPurifier_CSSDefinition extends HTMLPurifier_Definition\r
+{\r
+\r
+ public $type = 'CSS';\r
+\r
+ /**\r
+ * Assoc array of attribute name to definition object.\r
+ * @type HTMLPurifier_AttrDef[]\r
+ */\r
+ public $info = array();\r
+\r
+ /**\r
+ * Constructs the info array. The meat of this class.\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ protected function doSetup($config)\r
+ {\r
+ $this->info['text-align'] = new HTMLPurifier_AttrDef_Enum(\r
+ array('left', 'right', 'center', 'justify'),\r
+ false\r
+ );\r
+\r
+ $border_style =\r
+ $this->info['border-bottom-style'] =\r
+ $this->info['border-right-style'] =\r
+ $this->info['border-left-style'] =\r
+ $this->info['border-top-style'] = new HTMLPurifier_AttrDef_Enum(\r
+ array(\r
+ 'none',\r
+ 'hidden',\r
+ 'dotted',\r
+ 'dashed',\r
+ 'solid',\r
+ 'double',\r
+ 'groove',\r
+ 'ridge',\r
+ 'inset',\r
+ 'outset'\r
+ ),\r
+ false\r
+ );\r
+\r
+ $this->info['border-style'] = new HTMLPurifier_AttrDef_CSS_Multiple($border_style);\r
+\r
+ $this->info['clear'] = new HTMLPurifier_AttrDef_Enum(\r
+ array('none', 'left', 'right', 'both'),\r
+ false\r
+ );\r
+ $this->info['float'] = new HTMLPurifier_AttrDef_Enum(\r
+ array('none', 'left', 'right'),\r
+ false\r
+ );\r
+ $this->info['font-style'] = new HTMLPurifier_AttrDef_Enum(\r
+ array('normal', 'italic', 'oblique'),\r
+ false\r
+ );\r
+ $this->info['font-variant'] = new HTMLPurifier_AttrDef_Enum(\r
+ array('normal', 'small-caps'),\r
+ false\r
+ );\r
+\r
+ $uri_or_none = new HTMLPurifier_AttrDef_CSS_Composite(\r
+ array(\r
+ new HTMLPurifier_AttrDef_Enum(array('none')),\r
+ new HTMLPurifier_AttrDef_CSS_URI()\r
+ )\r
+ );\r
+\r
+ $this->info['list-style-position'] = new HTMLPurifier_AttrDef_Enum(\r
+ array('inside', 'outside'),\r
+ false\r
+ );\r
+ $this->info['list-style-type'] = new HTMLPurifier_AttrDef_Enum(\r
+ array(\r
+ 'disc',\r
+ 'circle',\r
+ 'square',\r
+ 'decimal',\r
+ 'lower-roman',\r
+ 'upper-roman',\r
+ 'lower-alpha',\r
+ 'upper-alpha',\r
+ 'none'\r
+ ),\r
+ false\r
+ );\r
+ $this->info['list-style-image'] = $uri_or_none;\r
+\r
+ $this->info['list-style'] = new HTMLPurifier_AttrDef_CSS_ListStyle($config);\r
+\r
+ $this->info['text-transform'] = new HTMLPurifier_AttrDef_Enum(\r
+ array('capitalize', 'uppercase', 'lowercase', 'none'),\r
+ false\r
+ );\r
+ $this->info['color'] = new HTMLPurifier_AttrDef_CSS_Color();\r
+\r
+ $this->info['background-image'] = $uri_or_none;\r
+ $this->info['background-repeat'] = new HTMLPurifier_AttrDef_Enum(\r
+ array('repeat', 'repeat-x', 'repeat-y', 'no-repeat')\r
+ );\r
+ $this->info['background-attachment'] = new HTMLPurifier_AttrDef_Enum(\r
+ array('scroll', 'fixed')\r
+ );\r
+ $this->info['background-position'] = new HTMLPurifier_AttrDef_CSS_BackgroundPosition();\r
+\r
+ $border_color =\r
+ $this->info['border-top-color'] =\r
+ $this->info['border-bottom-color'] =\r
+ $this->info['border-left-color'] =\r
+ $this->info['border-right-color'] =\r
+ $this->info['background-color'] = new HTMLPurifier_AttrDef_CSS_Composite(\r
+ array(\r
+ new HTMLPurifier_AttrDef_Enum(array('transparent')),\r
+ new HTMLPurifier_AttrDef_CSS_Color()\r
+ )\r
+ );\r
+\r
+ $this->info['background'] = new HTMLPurifier_AttrDef_CSS_Background($config);\r
+\r
+ $this->info['border-color'] = new HTMLPurifier_AttrDef_CSS_Multiple($border_color);\r
+\r
+ $border_width =\r
+ $this->info['border-top-width'] =\r
+ $this->info['border-bottom-width'] =\r
+ $this->info['border-left-width'] =\r
+ $this->info['border-right-width'] = new HTMLPurifier_AttrDef_CSS_Composite(\r
+ array(\r
+ new HTMLPurifier_AttrDef_Enum(array('thin', 'medium', 'thick')),\r
+ new HTMLPurifier_AttrDef_CSS_Length('0') //disallow negative\r
+ )\r
+ );\r
+\r
+ $this->info['border-width'] = new HTMLPurifier_AttrDef_CSS_Multiple($border_width);\r
+\r
+ $this->info['letter-spacing'] = new HTMLPurifier_AttrDef_CSS_Composite(\r
+ array(\r
+ new HTMLPurifier_AttrDef_Enum(array('normal')),\r
+ new HTMLPurifier_AttrDef_CSS_Length()\r
+ )\r
+ );\r
+\r
+ $this->info['word-spacing'] = new HTMLPurifier_AttrDef_CSS_Composite(\r
+ array(\r
+ new HTMLPurifier_AttrDef_Enum(array('normal')),\r
+ new HTMLPurifier_AttrDef_CSS_Length()\r
+ )\r
+ );\r
+\r
+ $this->info['font-size'] = new HTMLPurifier_AttrDef_CSS_Composite(\r
+ array(\r
+ new HTMLPurifier_AttrDef_Enum(\r
+ array(\r
+ 'xx-small',\r
+ 'x-small',\r
+ 'small',\r
+ 'medium',\r
+ 'large',\r
+ 'x-large',\r
+ 'xx-large',\r
+ 'larger',\r
+ 'smaller'\r
+ )\r
+ ),\r
+ new HTMLPurifier_AttrDef_CSS_Percentage(),\r
+ new HTMLPurifier_AttrDef_CSS_Length()\r
+ )\r
+ );\r
+\r
+ $this->info['line-height'] = new HTMLPurifier_AttrDef_CSS_Composite(\r
+ array(\r
+ new HTMLPurifier_AttrDef_Enum(array('normal')),\r
+ new HTMLPurifier_AttrDef_CSS_Number(true), // no negatives\r
+ new HTMLPurifier_AttrDef_CSS_Length('0'),\r
+ new HTMLPurifier_AttrDef_CSS_Percentage(true)\r
+ )\r
+ );\r
+\r
+ $margin =\r
+ $this->info['margin-top'] =\r
+ $this->info['margin-bottom'] =\r
+ $this->info['margin-left'] =\r
+ $this->info['margin-right'] = new HTMLPurifier_AttrDef_CSS_Composite(\r
+ array(\r
+ new HTMLPurifier_AttrDef_CSS_Length(),\r
+ new HTMLPurifier_AttrDef_CSS_Percentage(),\r
+ new HTMLPurifier_AttrDef_Enum(array('auto'))\r
+ )\r
+ );\r
+\r
+ $this->info['margin'] = new HTMLPurifier_AttrDef_CSS_Multiple($margin);\r
+\r
+ // non-negative\r
+ $padding =\r
+ $this->info['padding-top'] =\r
+ $this->info['padding-bottom'] =\r
+ $this->info['padding-left'] =\r
+ $this->info['padding-right'] = new HTMLPurifier_AttrDef_CSS_Composite(\r
+ array(\r
+ new HTMLPurifier_AttrDef_CSS_Length('0'),\r
+ new HTMLPurifier_AttrDef_CSS_Percentage(true)\r
+ )\r
+ );\r
+\r
+ $this->info['padding'] = new HTMLPurifier_AttrDef_CSS_Multiple($padding);\r
+\r
+ $this->info['text-indent'] = new HTMLPurifier_AttrDef_CSS_Composite(\r
+ array(\r
+ new HTMLPurifier_AttrDef_CSS_Length(),\r
+ new HTMLPurifier_AttrDef_CSS_Percentage()\r
+ )\r
+ );\r
+\r
+ $trusted_wh = new HTMLPurifier_AttrDef_CSS_Composite(\r
+ array(\r
+ new HTMLPurifier_AttrDef_CSS_Length('0'),\r
+ new HTMLPurifier_AttrDef_CSS_Percentage(true),\r
+ new HTMLPurifier_AttrDef_Enum(array('auto', 'initial', 'inherit'))\r
+ )\r
+ );\r
+ $trusted_min_wh = new HTMLPurifier_AttrDef_CSS_Composite(\r
+ array(\r
+ new HTMLPurifier_AttrDef_CSS_Length('0'),\r
+ new HTMLPurifier_AttrDef_CSS_Percentage(true),\r
+ new HTMLPurifier_AttrDef_Enum(array('initial', 'inherit'))\r
+ )\r
+ );\r
+ $trusted_max_wh = new HTMLPurifier_AttrDef_CSS_Composite(\r
+ array(\r
+ new HTMLPurifier_AttrDef_CSS_Length('0'),\r
+ new HTMLPurifier_AttrDef_CSS_Percentage(true),\r
+ new HTMLPurifier_AttrDef_Enum(array('none', 'initial', 'inherit'))\r
+ )\r
+ );\r
+ $max = $config->get('CSS.MaxImgLength');\r
+\r
+ $this->info['width'] =\r
+ $this->info['height'] =\r
+ $max === null ?\r
+ $trusted_wh :\r
+ new HTMLPurifier_AttrDef_Switch(\r
+ 'img',\r
+ // For img tags:\r
+ new HTMLPurifier_AttrDef_CSS_Composite(\r
+ array(\r
+ new HTMLPurifier_AttrDef_CSS_Length('0', $max),\r
+ new HTMLPurifier_AttrDef_Enum(array('auto'))\r
+ )\r
+ ),\r
+ // For everyone else:\r
+ $trusted_wh\r
+ );\r
+ $this->info['min-width'] =\r
+ $this->info['min-height'] =\r
+ $max === null ?\r
+ $trusted_min_wh :\r
+ new HTMLPurifier_AttrDef_Switch(\r
+ 'img',\r
+ // For img tags:\r
+ new HTMLPurifier_AttrDef_CSS_Composite(\r
+ array(\r
+ new HTMLPurifier_AttrDef_CSS_Length('0', $max),\r
+ new HTMLPurifier_AttrDef_Enum(array('initial', 'inherit'))\r
+ )\r
+ ),\r
+ // For everyone else:\r
+ $trusted_min_wh\r
+ );\r
+ $this->info['max-width'] =\r
+ $this->info['max-height'] =\r
+ $max === null ?\r
+ $trusted_max_wh :\r
+ new HTMLPurifier_AttrDef_Switch(\r
+ 'img',\r
+ // For img tags:\r
+ new HTMLPurifier_AttrDef_CSS_Composite(\r
+ array(\r
+ new HTMLPurifier_AttrDef_CSS_Length('0', $max),\r
+ new HTMLPurifier_AttrDef_Enum(array('none', 'initial', 'inherit'))\r
+ )\r
+ ),\r
+ // For everyone else:\r
+ $trusted_max_wh\r
+ );\r
+\r
+ $this->info['text-decoration'] = new HTMLPurifier_AttrDef_CSS_TextDecoration();\r
+\r
+ $this->info['font-family'] = new HTMLPurifier_AttrDef_CSS_FontFamily();\r
+\r
+ // this could use specialized code\r
+ $this->info['font-weight'] = new HTMLPurifier_AttrDef_Enum(\r
+ array(\r
+ 'normal',\r
+ 'bold',\r
+ 'bolder',\r
+ 'lighter',\r
+ '100',\r
+ '200',\r
+ '300',\r
+ '400',\r
+ '500',\r
+ '600',\r
+ '700',\r
+ '800',\r
+ '900'\r
+ ),\r
+ false\r
+ );\r
+\r
+ // MUST be called after other font properties, as it references\r
+ // a CSSDefinition object\r
+ $this->info['font'] = new HTMLPurifier_AttrDef_CSS_Font($config);\r
+\r
+ // same here\r
+ $this->info['border'] =\r
+ $this->info['border-bottom'] =\r
+ $this->info['border-top'] =\r
+ $this->info['border-left'] =\r
+ $this->info['border-right'] = new HTMLPurifier_AttrDef_CSS_Border($config);\r
+\r
+ $this->info['border-collapse'] = new HTMLPurifier_AttrDef_Enum(\r
+ array('collapse', 'separate')\r
+ );\r
+\r
+ $this->info['caption-side'] = new HTMLPurifier_AttrDef_Enum(\r
+ array('top', 'bottom')\r
+ );\r
+\r
+ $this->info['table-layout'] = new HTMLPurifier_AttrDef_Enum(\r
+ array('auto', 'fixed')\r
+ );\r
+\r
+ $this->info['vertical-align'] = new HTMLPurifier_AttrDef_CSS_Composite(\r
+ array(\r
+ new HTMLPurifier_AttrDef_Enum(\r
+ array(\r
+ 'baseline',\r
+ 'sub',\r
+ 'super',\r
+ 'top',\r
+ 'text-top',\r
+ 'middle',\r
+ 'bottom',\r
+ 'text-bottom'\r
+ )\r
+ ),\r
+ new HTMLPurifier_AttrDef_CSS_Length(),\r
+ new HTMLPurifier_AttrDef_CSS_Percentage()\r
+ )\r
+ );\r
+\r
+ $this->info['border-spacing'] = new HTMLPurifier_AttrDef_CSS_Multiple(new HTMLPurifier_AttrDef_CSS_Length(), 2);\r
+\r
+ // These CSS properties don't work on many browsers, but we live\r
+ // in THE FUTURE!\r
+ $this->info['white-space'] = new HTMLPurifier_AttrDef_Enum(\r
+ array('nowrap', 'normal', 'pre', 'pre-wrap', 'pre-line')\r
+ );\r
+\r
+ if ($config->get('CSS.Proprietary')) {\r
+ $this->doSetupProprietary($config);\r
+ }\r
+\r
+ if ($config->get('CSS.AllowTricky')) {\r
+ $this->doSetupTricky($config);\r
+ }\r
+\r
+ if ($config->get('CSS.Trusted')) {\r
+ $this->doSetupTrusted($config);\r
+ }\r
+\r
+ $allow_important = $config->get('CSS.AllowImportant');\r
+ // wrap all attr-defs with decorator that handles !important\r
+ foreach ($this->info as $k => $v) {\r
+ $this->info[$k] = new HTMLPurifier_AttrDef_CSS_ImportantDecorator($v, $allow_important);\r
+ }\r
+\r
+ $this->setupConfigStuff($config);\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ protected function doSetupProprietary($config)\r
+ {\r
+ // Internet Explorer only scrollbar colors\r
+ $this->info['scrollbar-arrow-color'] = new HTMLPurifier_AttrDef_CSS_Color();\r
+ $this->info['scrollbar-base-color'] = new HTMLPurifier_AttrDef_CSS_Color();\r
+ $this->info['scrollbar-darkshadow-color'] = new HTMLPurifier_AttrDef_CSS_Color();\r
+ $this->info['scrollbar-face-color'] = new HTMLPurifier_AttrDef_CSS_Color();\r
+ $this->info['scrollbar-highlight-color'] = new HTMLPurifier_AttrDef_CSS_Color();\r
+ $this->info['scrollbar-shadow-color'] = new HTMLPurifier_AttrDef_CSS_Color();\r
+\r
+ // vendor specific prefixes of opacity\r
+ $this->info['-moz-opacity'] = new HTMLPurifier_AttrDef_CSS_AlphaValue();\r
+ $this->info['-khtml-opacity'] = new HTMLPurifier_AttrDef_CSS_AlphaValue();\r
+\r
+ // only opacity, for now\r
+ $this->info['filter'] = new HTMLPurifier_AttrDef_CSS_Filter();\r
+\r
+ // more CSS3\r
+ $this->info['page-break-after'] =\r
+ $this->info['page-break-before'] = new HTMLPurifier_AttrDef_Enum(\r
+ array(\r
+ 'auto',\r
+ 'always',\r
+ 'avoid',\r
+ 'left',\r
+ 'right'\r
+ )\r
+ );\r
+ $this->info['page-break-inside'] = new HTMLPurifier_AttrDef_Enum(array('auto', 'avoid'));\r
+\r
+ $border_radius = new HTMLPurifier_AttrDef_CSS_Composite(\r
+ array(\r
+ new HTMLPurifier_AttrDef_CSS_Percentage(true), // disallow negative\r
+ new HTMLPurifier_AttrDef_CSS_Length('0') // disallow negative\r
+ ));\r
+\r
+ $this->info['border-top-left-radius'] =\r
+ $this->info['border-top-right-radius'] =\r
+ $this->info['border-bottom-right-radius'] =\r
+ $this->info['border-bottom-left-radius'] = new HTMLPurifier_AttrDef_CSS_Multiple($border_radius, 2);\r
+ // TODO: support SLASH syntax\r
+ $this->info['border-radius'] = new HTMLPurifier_AttrDef_CSS_Multiple($border_radius, 4);\r
+\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ protected function doSetupTricky($config)\r
+ {\r
+ $this->info['display'] = new HTMLPurifier_AttrDef_Enum(\r
+ array(\r
+ 'inline',\r
+ 'block',\r
+ 'list-item',\r
+ 'run-in',\r
+ 'compact',\r
+ 'marker',\r
+ 'table',\r
+ 'inline-block',\r
+ 'inline-table',\r
+ 'table-row-group',\r
+ 'table-header-group',\r
+ 'table-footer-group',\r
+ 'table-row',\r
+ 'table-column-group',\r
+ 'table-column',\r
+ 'table-cell',\r
+ 'table-caption',\r
+ 'none'\r
+ )\r
+ );\r
+ $this->info['visibility'] = new HTMLPurifier_AttrDef_Enum(\r
+ array('visible', 'hidden', 'collapse')\r
+ );\r
+ $this->info['overflow'] = new HTMLPurifier_AttrDef_Enum(array('visible', 'hidden', 'auto', 'scroll'));\r
+ $this->info['opacity'] = new HTMLPurifier_AttrDef_CSS_AlphaValue();\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ protected function doSetupTrusted($config)\r
+ {\r
+ $this->info['position'] = new HTMLPurifier_AttrDef_Enum(\r
+ array('static', 'relative', 'absolute', 'fixed')\r
+ );\r
+ $this->info['top'] =\r
+ $this->info['left'] =\r
+ $this->info['right'] =\r
+ $this->info['bottom'] = new HTMLPurifier_AttrDef_CSS_Composite(\r
+ array(\r
+ new HTMLPurifier_AttrDef_CSS_Length(),\r
+ new HTMLPurifier_AttrDef_CSS_Percentage(),\r
+ new HTMLPurifier_AttrDef_Enum(array('auto')),\r
+ )\r
+ );\r
+ $this->info['z-index'] = new HTMLPurifier_AttrDef_CSS_Composite(\r
+ array(\r
+ new HTMLPurifier_AttrDef_Integer(),\r
+ new HTMLPurifier_AttrDef_Enum(array('auto')),\r
+ )\r
+ );\r
+ }\r
+\r
+ /**\r
+ * Performs extra config-based processing. Based off of\r
+ * HTMLPurifier_HTMLDefinition.\r
+ * @param HTMLPurifier_Config $config\r
+ * @todo Refactor duplicate elements into common class (probably using\r
+ * composition, not inheritance).\r
+ */\r
+ protected function setupConfigStuff($config)\r
+ {\r
+ // setup allowed elements\r
+ $support = "(for information on implementing this, see the " .\r
+ "support forums) ";\r
+ $allowed_properties = $config->get('CSS.AllowedProperties');\r
+ if ($allowed_properties !== null) {\r
+ foreach ($this->info as $name => $d) {\r
+ if (!isset($allowed_properties[$name])) {\r
+ unset($this->info[$name]);\r
+ }\r
+ unset($allowed_properties[$name]);\r
+ }\r
+ // emit errors\r
+ foreach ($allowed_properties as $name => $d) {\r
+ // :TODO: Is this htmlspecialchars() call really necessary?\r
+ $name = htmlspecialchars($name);\r
+ trigger_error("Style attribute '$name' is not supported $support", E_USER_WARNING);\r
+ }\r
+ }\r
+\r
+ $forbidden_properties = $config->get('CSS.ForbiddenProperties');\r
+ if ($forbidden_properties !== null) {\r
+ foreach ($this->info as $name => $d) {\r
+ if (isset($forbidden_properties[$name])) {\r
+ unset($this->info[$name]);\r
+ }\r
+ }\r
+ }\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Defines allowed child nodes and validates nodes against it.\r
+ */\r
+abstract class HTMLPurifier_ChildDef\r
+{\r
+ /**\r
+ * Type of child definition, usually right-most part of class name lowercase.\r
+ * Used occasionally in terms of context.\r
+ * @type string\r
+ */\r
+ public $type;\r
+\r
+ /**\r
+ * Indicates whether or not an empty array of children is okay.\r
+ *\r
+ * This is necessary for redundant checking when changes affecting\r
+ * a child node may cause a parent node to now be disallowed.\r
+ * @type bool\r
+ */\r
+ public $allow_empty;\r
+\r
+ /**\r
+ * Lookup array of all elements that this definition could possibly allow.\r
+ * @type array\r
+ */\r
+ public $elements = array();\r
+\r
+ /**\r
+ * Get lookup of tag names that should not close this element automatically.\r
+ * All other elements will do so.\r
+ * @param HTMLPurifier_Config $config HTMLPurifier_Config object\r
+ * @return array\r
+ */\r
+ public function getAllowedElements($config)\r
+ {\r
+ return $this->elements;\r
+ }\r
+\r
+ /**\r
+ * Validates nodes according to definition and returns modification.\r
+ *\r
+ * @param HTMLPurifier_Node[] $children Array of HTMLPurifier_Node\r
+ * @param HTMLPurifier_Config $config HTMLPurifier_Config object\r
+ * @param HTMLPurifier_Context $context HTMLPurifier_Context object\r
+ * @return bool|array true to leave nodes as is, false to remove parent node, array of replacement children\r
+ */\r
+ abstract public function validateChildren($children, $config, $context);\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Configuration object that triggers customizable behavior.\r
+ *\r
+ * @warning This class is strongly defined: that means that the class\r
+ * will fail if an undefined directive is retrieved or set.\r
+ *\r
+ * @note Many classes that could (although many times don't) use the\r
+ * configuration object make it a mandatory parameter. This is\r
+ * because a configuration object should always be forwarded,\r
+ * otherwise, you run the risk of missing a parameter and then\r
+ * being stumped when a configuration directive doesn't work.\r
+ *\r
+ * @todo Reconsider some of the public member variables\r
+ */\r
+class HTMLPurifier_Config\r
+{\r
+\r
+ /**\r
+ * HTML Purifier's version\r
+ * @type string\r
+ */\r
+ public $version = '4.11.0';\r
+\r
+ /**\r
+ * Whether or not to automatically finalize\r
+ * the object if a read operation is done.\r
+ * @type bool\r
+ */\r
+ public $autoFinalize = true;\r
+\r
+ // protected member variables\r
+\r
+ /**\r
+ * Namespace indexed array of serials for specific namespaces.\r
+ * @see getSerial() for more info.\r
+ * @type string[]\r
+ */\r
+ protected $serials = array();\r
+\r
+ /**\r
+ * Serial for entire configuration object.\r
+ * @type string\r
+ */\r
+ protected $serial;\r
+\r
+ /**\r
+ * Parser for variables.\r
+ * @type HTMLPurifier_VarParser_Flexible\r
+ */\r
+ protected $parser = null;\r
+\r
+ /**\r
+ * Reference HTMLPurifier_ConfigSchema for value checking.\r
+ * @type HTMLPurifier_ConfigSchema\r
+ * @note This is public for introspective purposes. Please don't\r
+ * abuse!\r
+ */\r
+ public $def;\r
+\r
+ /**\r
+ * Indexed array of definitions.\r
+ * @type HTMLPurifier_Definition[]\r
+ */\r
+ protected $definitions;\r
+\r
+ /**\r
+ * Whether or not config is finalized.\r
+ * @type bool\r
+ */\r
+ protected $finalized = false;\r
+\r
+ /**\r
+ * Property list containing configuration directives.\r
+ * @type array\r
+ */\r
+ protected $plist;\r
+\r
+ /**\r
+ * Whether or not a set is taking place due to an alias lookup.\r
+ * @type bool\r
+ */\r
+ private $aliasMode;\r
+\r
+ /**\r
+ * Set to false if you do not want line and file numbers in errors.\r
+ * (useful when unit testing). This will also compress some errors\r
+ * and exceptions.\r
+ * @type bool\r
+ */\r
+ public $chatty = true;\r
+\r
+ /**\r
+ * Current lock; only gets to this namespace are allowed.\r
+ * @type string\r
+ */\r
+ private $lock;\r
+\r
+ /**\r
+ * Constructor\r
+ * @param HTMLPurifier_ConfigSchema $definition ConfigSchema that defines\r
+ * what directives are allowed.\r
+ * @param HTMLPurifier_PropertyList $parent\r
+ */\r
+ public function __construct($definition, $parent = null)\r
+ {\r
+ $parent = $parent ? $parent : $definition->defaultPlist;\r
+ $this->plist = new HTMLPurifier_PropertyList($parent);\r
+ $this->def = $definition; // keep a copy around for checking\r
+ $this->parser = new HTMLPurifier_VarParser_Flexible();\r
+ }\r
+\r
+ /**\r
+ * Convenience constructor that creates a config object based on a mixed var\r
+ * @param mixed $config Variable that defines the state of the config\r
+ * object. Can be: a HTMLPurifier_Config() object,\r
+ * an array of directives based on loadArray(),\r
+ * or a string filename of an ini file.\r
+ * @param HTMLPurifier_ConfigSchema $schema Schema object\r
+ * @return HTMLPurifier_Config Configured object\r
+ */\r
+ public static function create($config, $schema = null)\r
+ {\r
+ if ($config instanceof HTMLPurifier_Config) {\r
+ // pass-through\r
+ return $config;\r
+ }\r
+ if (!$schema) {\r
+ $ret = HTMLPurifier_Config::createDefault();\r
+ } else {\r
+ $ret = new HTMLPurifier_Config($schema);\r
+ }\r
+ if (is_string($config)) {\r
+ $ret->loadIni($config);\r
+ } elseif (is_array($config)) $ret->loadArray($config);\r
+ return $ret;\r
+ }\r
+\r
+ /**\r
+ * Creates a new config object that inherits from a previous one.\r
+ * @param HTMLPurifier_Config $config Configuration object to inherit from.\r
+ * @return HTMLPurifier_Config object with $config as its parent.\r
+ */\r
+ public static function inherit(HTMLPurifier_Config $config)\r
+ {\r
+ return new HTMLPurifier_Config($config->def, $config->plist);\r
+ }\r
+\r
+ /**\r
+ * Convenience constructor that creates a default configuration object.\r
+ * @return HTMLPurifier_Config default object.\r
+ */\r
+ public static function createDefault()\r
+ {\r
+ $definition = HTMLPurifier_ConfigSchema::instance();\r
+ $config = new HTMLPurifier_Config($definition);\r
+ return $config;\r
+ }\r
+\r
+ /**\r
+ * Retrieves a value from the configuration.\r
+ *\r
+ * @param string $key String key\r
+ * @param mixed $a\r
+ *\r
+ * @return mixed\r
+ */\r
+ public function get($key, $a = null)\r
+ {\r
+ if ($a !== null) {\r
+ $this->triggerError(\r
+ "Using deprecated API: use \$config->get('$key.$a') instead",\r
+ E_USER_WARNING\r
+ );\r
+ $key = "$key.$a";\r
+ }\r
+ if (!$this->finalized) {\r
+ $this->autoFinalize();\r
+ }\r
+ if (!isset($this->def->info[$key])) {\r
+ // can't add % due to SimpleTest bug\r
+ $this->triggerError(\r
+ 'Cannot retrieve value of undefined directive ' . htmlspecialchars($key),\r
+ E_USER_WARNING\r
+ );\r
+ return;\r
+ }\r
+ if (isset($this->def->info[$key]->isAlias)) {\r
+ $d = $this->def->info[$key];\r
+ $this->triggerError(\r
+ 'Cannot get value from aliased directive, use real name ' . $d->key,\r
+ E_USER_ERROR\r
+ );\r
+ return;\r
+ }\r
+ if ($this->lock) {\r
+ list($ns) = explode('.', $key);\r
+ if ($ns !== $this->lock) {\r
+ $this->triggerError(\r
+ 'Cannot get value of namespace ' . $ns . ' when lock for ' .\r
+ $this->lock .\r
+ ' is active, this probably indicates a Definition setup method ' .\r
+ 'is accessing directives that are not within its namespace',\r
+ E_USER_ERROR\r
+ );\r
+ return;\r
+ }\r
+ }\r
+ return $this->plist->get($key);\r
+ }\r
+\r
+ /**\r
+ * Retrieves an array of directives to values from a given namespace\r
+ *\r
+ * @param string $namespace String namespace\r
+ *\r
+ * @return array\r
+ */\r
+ public function getBatch($namespace)\r
+ {\r
+ if (!$this->finalized) {\r
+ $this->autoFinalize();\r
+ }\r
+ $full = $this->getAll();\r
+ if (!isset($full[$namespace])) {\r
+ $this->triggerError(\r
+ 'Cannot retrieve undefined namespace ' .\r
+ htmlspecialchars($namespace),\r
+ E_USER_WARNING\r
+ );\r
+ return;\r
+ }\r
+ return $full[$namespace];\r
+ }\r
+\r
+ /**\r
+ * Returns a SHA-1 signature of a segment of the configuration object\r
+ * that uniquely identifies that particular configuration\r
+ *\r
+ * @param string $namespace Namespace to get serial for\r
+ *\r
+ * @return string\r
+ * @note Revision is handled specially and is removed from the batch\r
+ * before processing!\r
+ */\r
+ public function getBatchSerial($namespace)\r
+ {\r
+ if (empty($this->serials[$namespace])) {\r
+ $batch = $this->getBatch($namespace);\r
+ unset($batch['DefinitionRev']);\r
+ $this->serials[$namespace] = sha1(serialize($batch));\r
+ }\r
+ return $this->serials[$namespace];\r
+ }\r
+\r
+ /**\r
+ * Returns a SHA-1 signature for the entire configuration object\r
+ * that uniquely identifies that particular configuration\r
+ *\r
+ * @return string\r
+ */\r
+ public function getSerial()\r
+ {\r
+ if (empty($this->serial)) {\r
+ $this->serial = sha1(serialize($this->getAll()));\r
+ }\r
+ return $this->serial;\r
+ }\r
+\r
+ /**\r
+ * Retrieves all directives, organized by namespace\r
+ *\r
+ * @warning This is a pretty inefficient function, avoid if you can\r
+ */\r
+ public function getAll()\r
+ {\r
+ if (!$this->finalized) {\r
+ $this->autoFinalize();\r
+ }\r
+ $ret = array();\r
+ foreach ($this->plist->squash() as $name => $value) {\r
+ list($ns, $key) = explode('.', $name, 2);\r
+ $ret[$ns][$key] = $value;\r
+ }\r
+ return $ret;\r
+ }\r
+\r
+ /**\r
+ * Sets a value to configuration.\r
+ *\r
+ * @param string $key key\r
+ * @param mixed $value value\r
+ * @param mixed $a\r
+ */\r
+ public function set($key, $value, $a = null)\r
+ {\r
+ if (strpos($key, '.') === false) {\r
+ $namespace = $key;\r
+ $directive = $value;\r
+ $value = $a;\r
+ $key = "$key.$directive";\r
+ $this->triggerError("Using deprecated API: use \$config->set('$key', ...) instead", E_USER_NOTICE);\r
+ } else {\r
+ list($namespace) = explode('.', $key);\r
+ }\r
+ if ($this->isFinalized('Cannot set directive after finalization')) {\r
+ return;\r
+ }\r
+ if (!isset($this->def->info[$key])) {\r
+ $this->triggerError(\r
+ 'Cannot set undefined directive ' . htmlspecialchars($key) . ' to value',\r
+ E_USER_WARNING\r
+ );\r
+ return;\r
+ }\r
+ $def = $this->def->info[$key];\r
+\r
+ if (isset($def->isAlias)) {\r
+ if ($this->aliasMode) {\r
+ $this->triggerError(\r
+ 'Double-aliases not allowed, please fix '.\r
+ 'ConfigSchema bug with' . $key,\r
+ E_USER_ERROR\r
+ );\r
+ return;\r
+ }\r
+ $this->aliasMode = true;\r
+ $this->set($def->key, $value);\r
+ $this->aliasMode = false;\r
+ $this->triggerError("$key is an alias, preferred directive name is {$def->key}", E_USER_NOTICE);\r
+ return;\r
+ }\r
+\r
+ // Raw type might be negative when using the fully optimized form\r
+ // of stdClass, which indicates allow_null == true\r
+ $rtype = is_int($def) ? $def : $def->type;\r
+ if ($rtype < 0) {\r
+ $type = -$rtype;\r
+ $allow_null = true;\r
+ } else {\r
+ $type = $rtype;\r
+ $allow_null = isset($def->allow_null);\r
+ }\r
+\r
+ try {\r
+ $value = $this->parser->parse($value, $type, $allow_null);\r
+ } catch (HTMLPurifier_VarParserException $e) {\r
+ $this->triggerError(\r
+ 'Value for ' . $key . ' is of invalid type, should be ' .\r
+ HTMLPurifier_VarParser::getTypeName($type),\r
+ E_USER_WARNING\r
+ );\r
+ return;\r
+ }\r
+ if (is_string($value) && is_object($def)) {\r
+ // resolve value alias if defined\r
+ if (isset($def->aliases[$value])) {\r
+ $value = $def->aliases[$value];\r
+ }\r
+ // check to see if the value is allowed\r
+ if (isset($def->allowed) && !isset($def->allowed[$value])) {\r
+ $this->triggerError(\r
+ 'Value not supported, valid values are: ' .\r
+ $this->_listify($def->allowed),\r
+ E_USER_WARNING\r
+ );\r
+ return;\r
+ }\r
+ }\r
+ $this->plist->set($key, $value);\r
+\r
+ // reset definitions if the directives they depend on changed\r
+ // this is a very costly process, so it's discouraged\r
+ // with finalization\r
+ if ($namespace == 'HTML' || $namespace == 'CSS' || $namespace == 'URI') {\r
+ $this->definitions[$namespace] = null;\r
+ }\r
+\r
+ $this->serials[$namespace] = false;\r
+ }\r
+\r
+ /**\r
+ * Convenience function for error reporting\r
+ *\r
+ * @param array $lookup\r
+ *\r
+ * @return string\r
+ */\r
+ private function _listify($lookup)\r
+ {\r
+ $list = array();\r
+ foreach ($lookup as $name => $b) {\r
+ $list[] = $name;\r
+ }\r
+ return implode(', ', $list);\r
+ }\r
+\r
+ /**\r
+ * Retrieves object reference to the HTML definition.\r
+ *\r
+ * @param bool $raw Return a copy that has not been setup yet. Must be\r
+ * called before it's been setup, otherwise won't work.\r
+ * @param bool $optimized If true, this method may return null, to\r
+ * indicate that a cached version of the modified\r
+ * definition object is available and no further edits\r
+ * are necessary. Consider using\r
+ * maybeGetRawHTMLDefinition, which is more explicitly\r
+ * named, instead.\r
+ *\r
+ * @return HTMLPurifier_HTMLDefinition\r
+ */\r
+ public function getHTMLDefinition($raw = false, $optimized = false)\r
+ {\r
+ return $this->getDefinition('HTML', $raw, $optimized);\r
+ }\r
+\r
+ /**\r
+ * Retrieves object reference to the CSS definition\r
+ *\r
+ * @param bool $raw Return a copy that has not been setup yet. Must be\r
+ * called before it's been setup, otherwise won't work.\r
+ * @param bool $optimized If true, this method may return null, to\r
+ * indicate that a cached version of the modified\r
+ * definition object is available and no further edits\r
+ * are necessary. Consider using\r
+ * maybeGetRawCSSDefinition, which is more explicitly\r
+ * named, instead.\r
+ *\r
+ * @return HTMLPurifier_CSSDefinition\r
+ */\r
+ public function getCSSDefinition($raw = false, $optimized = false)\r
+ {\r
+ return $this->getDefinition('CSS', $raw, $optimized);\r
+ }\r
+\r
+ /**\r
+ * Retrieves object reference to the URI definition\r
+ *\r
+ * @param bool $raw Return a copy that has not been setup yet. Must be\r
+ * called before it's been setup, otherwise won't work.\r
+ * @param bool $optimized If true, this method may return null, to\r
+ * indicate that a cached version of the modified\r
+ * definition object is available and no further edits\r
+ * are necessary. Consider using\r
+ * maybeGetRawURIDefinition, which is more explicitly\r
+ * named, instead.\r
+ *\r
+ * @return HTMLPurifier_URIDefinition\r
+ */\r
+ public function getURIDefinition($raw = false, $optimized = false)\r
+ {\r
+ return $this->getDefinition('URI', $raw, $optimized);\r
+ }\r
+\r
+ /**\r
+ * Retrieves a definition\r
+ *\r
+ * @param string $type Type of definition: HTML, CSS, etc\r
+ * @param bool $raw Whether or not definition should be returned raw\r
+ * @param bool $optimized Only has an effect when $raw is true. Whether\r
+ * or not to return null if the result is already present in\r
+ * the cache. This is off by default for backwards\r
+ * compatibility reasons, but you need to do things this\r
+ * way in order to ensure that caching is done properly.\r
+ * Check out enduser-customize.html for more details.\r
+ * We probably won't ever change this default, as much as the\r
+ * maybe semantics is the "right thing to do."\r
+ *\r
+ * @throws HTMLPurifier_Exception\r
+ * @return HTMLPurifier_Definition\r
+ */\r
+ public function getDefinition($type, $raw = false, $optimized = false)\r
+ {\r
+ if ($optimized && !$raw) {\r
+ throw new HTMLPurifier_Exception("Cannot set optimized = true when raw = false");\r
+ }\r
+ if (!$this->finalized) {\r
+ $this->autoFinalize();\r
+ }\r
+ // temporarily suspend locks, so we can handle recursive definition calls\r
+ $lock = $this->lock;\r
+ $this->lock = null;\r
+ $factory = HTMLPurifier_DefinitionCacheFactory::instance();\r
+ $cache = $factory->create($type, $this);\r
+ $this->lock = $lock;\r
+ if (!$raw) {\r
+ // full definition\r
+ // ---------------\r
+ // check if definition is in memory\r
+ if (!empty($this->definitions[$type])) {\r
+ $def = $this->definitions[$type];\r
+ // check if the definition is setup\r
+ if ($def->setup) {\r
+ return $def;\r
+ } else {\r
+ $def->setup($this);\r
+ if ($def->optimized) {\r
+ $cache->add($def, $this);\r
+ }\r
+ return $def;\r
+ }\r
+ }\r
+ // check if definition is in cache\r
+ $def = $cache->get($this);\r
+ if ($def) {\r
+ // definition in cache, save to memory and return it\r
+ $this->definitions[$type] = $def;\r
+ return $def;\r
+ }\r
+ // initialize it\r
+ $def = $this->initDefinition($type);\r
+ // set it up\r
+ $this->lock = $type;\r
+ $def->setup($this);\r
+ $this->lock = null;\r
+ // save in cache\r
+ $cache->add($def, $this);\r
+ // return it\r
+ return $def;\r
+ } else {\r
+ // raw definition\r
+ // --------------\r
+ // check preconditions\r
+ $def = null;\r
+ if ($optimized) {\r
+ if (is_null($this->get($type . '.DefinitionID'))) {\r
+ // fatally error out if definition ID not set\r
+ throw new HTMLPurifier_Exception(\r
+ "Cannot retrieve raw version without specifying %$type.DefinitionID"\r
+ );\r
+ }\r
+ }\r
+ if (!empty($this->definitions[$type])) {\r
+ $def = $this->definitions[$type];\r
+ if ($def->setup && !$optimized) {\r
+ $extra = $this->chatty ?\r
+ " (try moving this code block earlier in your initialization)" :\r
+ "";\r
+ throw new HTMLPurifier_Exception(\r
+ "Cannot retrieve raw definition after it has already been setup" .\r
+ $extra\r
+ );\r
+ }\r
+ if ($def->optimized === null) {\r
+ $extra = $this->chatty ? " (try flushing your cache)" : "";\r
+ throw new HTMLPurifier_Exception(\r
+ "Optimization status of definition is unknown" . $extra\r
+ );\r
+ }\r
+ if ($def->optimized !== $optimized) {\r
+ $msg = $optimized ? "optimized" : "unoptimized";\r
+ $extra = $this->chatty ?\r
+ " (this backtrace is for the first inconsistent call, which was for a $msg raw definition)"\r
+ : "";\r
+ throw new HTMLPurifier_Exception(\r
+ "Inconsistent use of optimized and unoptimized raw definition retrievals" . $extra\r
+ );\r
+ }\r
+ }\r
+ // check if definition was in memory\r
+ if ($def) {\r
+ if ($def->setup) {\r
+ // invariant: $optimized === true (checked above)\r
+ return null;\r
+ } else {\r
+ return $def;\r
+ }\r
+ }\r
+ // if optimized, check if definition was in cache\r
+ // (because we do the memory check first, this formulation\r
+ // is prone to cache slamming, but I think\r
+ // guaranteeing that either /all/ of the raw\r
+ // setup code or /none/ of it is run is more important.)\r
+ if ($optimized) {\r
+ // This code path only gets run once; once we put\r
+ // something in $definitions (which is guaranteed by the\r
+ // trailing code), we always short-circuit above.\r
+ $def = $cache->get($this);\r
+ if ($def) {\r
+ // save the full definition for later, but don't\r
+ // return it yet\r
+ $this->definitions[$type] = $def;\r
+ return null;\r
+ }\r
+ }\r
+ // check invariants for creation\r
+ if (!$optimized) {\r
+ if (!is_null($this->get($type . '.DefinitionID'))) {\r
+ if ($this->chatty) {\r
+ $this->triggerError(\r
+ 'Due to a documentation error in previous version of HTML Purifier, your ' .\r
+ 'definitions are not being cached. If this is OK, you can remove the ' .\r
+ '%$type.DefinitionRev and %$type.DefinitionID declaration. Otherwise, ' .\r
+ 'modify your code to use maybeGetRawDefinition, and test if the returned ' .\r
+ 'value is null before making any edits (if it is null, that means that a ' .\r
+ 'cached version is available, and no raw operations are necessary). See ' .\r
+ '<a href="http://htmlpurifier.org/docs/enduser-customize.html#optimized">' .\r
+ 'Customize</a> for more details',\r
+ E_USER_WARNING\r
+ );\r
+ } else {\r
+ $this->triggerError(\r
+ "Useless DefinitionID declaration",\r
+ E_USER_WARNING\r
+ );\r
+ }\r
+ }\r
+ }\r
+ // initialize it\r
+ $def = $this->initDefinition($type);\r
+ $def->optimized = $optimized;\r
+ return $def;\r
+ }\r
+ throw new HTMLPurifier_Exception("The impossible happened!");\r
+ }\r
+\r
+ /**\r
+ * Initialise definition\r
+ *\r
+ * @param string $type What type of definition to create\r
+ *\r
+ * @return HTMLPurifier_CSSDefinition|HTMLPurifier_HTMLDefinition|HTMLPurifier_URIDefinition\r
+ * @throws HTMLPurifier_Exception\r
+ */\r
+ private function initDefinition($type)\r
+ {\r
+ // quick checks failed, let's create the object\r
+ if ($type == 'HTML') {\r
+ $def = new HTMLPurifier_HTMLDefinition();\r
+ } elseif ($type == 'CSS') {\r
+ $def = new HTMLPurifier_CSSDefinition();\r
+ } elseif ($type == 'URI') {\r
+ $def = new HTMLPurifier_URIDefinition();\r
+ } else {\r
+ throw new HTMLPurifier_Exception(\r
+ "Definition of $type type not supported"\r
+ );\r
+ }\r
+ $this->definitions[$type] = $def;\r
+ return $def;\r
+ }\r
+\r
+ public function maybeGetRawDefinition($name)\r
+ {\r
+ return $this->getDefinition($name, true, true);\r
+ }\r
+\r
+ /**\r
+ * @return HTMLPurifier_HTMLDefinition\r
+ */\r
+ public function maybeGetRawHTMLDefinition()\r
+ {\r
+ return $this->getDefinition('HTML', true, true);\r
+ }\r
+ \r
+ /**\r
+ * @return HTMLPurifier_CSSDefinition\r
+ */\r
+ public function maybeGetRawCSSDefinition()\r
+ {\r
+ return $this->getDefinition('CSS', true, true);\r
+ }\r
+ \r
+ /**\r
+ * @return HTMLPurifier_URIDefinition\r
+ */\r
+ public function maybeGetRawURIDefinition()\r
+ {\r
+ return $this->getDefinition('URI', true, true);\r
+ }\r
+\r
+ /**\r
+ * Loads configuration values from an array with the following structure:\r
+ * Namespace.Directive => Value\r
+ *\r
+ * @param array $config_array Configuration associative array\r
+ */\r
+ public function loadArray($config_array)\r
+ {\r
+ if ($this->isFinalized('Cannot load directives after finalization')) {\r
+ return;\r
+ }\r
+ foreach ($config_array as $key => $value) {\r
+ $key = str_replace('_', '.', $key);\r
+ if (strpos($key, '.') !== false) {\r
+ $this->set($key, $value);\r
+ } else {\r
+ $namespace = $key;\r
+ $namespace_values = $value;\r
+ foreach ($namespace_values as $directive => $value2) {\r
+ $this->set($namespace .'.'. $directive, $value2);\r
+ }\r
+ }\r
+ }\r
+ }\r
+\r
+ /**\r
+ * Returns a list of array(namespace, directive) for all directives\r
+ * that are allowed in a web-form context as per an allowed\r
+ * namespaces/directives list.\r
+ *\r
+ * @param array $allowed List of allowed namespaces/directives\r
+ * @param HTMLPurifier_ConfigSchema $schema Schema to use, if not global copy\r
+ *\r
+ * @return array\r
+ */\r
+ public static function getAllowedDirectivesForForm($allowed, $schema = null)\r
+ {\r
+ if (!$schema) {\r
+ $schema = HTMLPurifier_ConfigSchema::instance();\r
+ }\r
+ if ($allowed !== true) {\r
+ if (is_string($allowed)) {\r
+ $allowed = array($allowed);\r
+ }\r
+ $allowed_ns = array();\r
+ $allowed_directives = array();\r
+ $blacklisted_directives = array();\r
+ foreach ($allowed as $ns_or_directive) {\r
+ if (strpos($ns_or_directive, '.') !== false) {\r
+ // directive\r
+ if ($ns_or_directive[0] == '-') {\r
+ $blacklisted_directives[substr($ns_or_directive, 1)] = true;\r
+ } else {\r
+ $allowed_directives[$ns_or_directive] = true;\r
+ }\r
+ } else {\r
+ // namespace\r
+ $allowed_ns[$ns_or_directive] = true;\r
+ }\r
+ }\r
+ }\r
+ $ret = array();\r
+ foreach ($schema->info as $key => $def) {\r
+ list($ns, $directive) = explode('.', $key, 2);\r
+ if ($allowed !== true) {\r
+ if (isset($blacklisted_directives["$ns.$directive"])) {\r
+ continue;\r
+ }\r
+ if (!isset($allowed_directives["$ns.$directive"]) && !isset($allowed_ns[$ns])) {\r
+ continue;\r
+ }\r
+ }\r
+ if (isset($def->isAlias)) {\r
+ continue;\r
+ }\r
+ if ($directive == 'DefinitionID' || $directive == 'DefinitionRev') {\r
+ continue;\r
+ }\r
+ $ret[] = array($ns, $directive);\r
+ }\r
+ return $ret;\r
+ }\r
+\r
+ /**\r
+ * Loads configuration values from $_GET/$_POST that were posted\r
+ * via ConfigForm\r
+ *\r
+ * @param array $array $_GET or $_POST array to import\r
+ * @param string|bool $index Index/name that the config variables are in\r
+ * @param array|bool $allowed List of allowed namespaces/directives\r
+ * @param bool $mq_fix Boolean whether or not to enable magic quotes fix\r
+ * @param HTMLPurifier_ConfigSchema $schema Schema to use, if not global copy\r
+ *\r
+ * @return mixed\r
+ */\r
+ public static function loadArrayFromForm($array, $index = false, $allowed = true, $mq_fix = true, $schema = null)\r
+ {\r
+ $ret = HTMLPurifier_Config::prepareArrayFromForm($array, $index, $allowed, $mq_fix, $schema);\r
+ $config = HTMLPurifier_Config::create($ret, $schema);\r
+ return $config;\r
+ }\r
+\r
+ /**\r
+ * Merges in configuration values from $_GET/$_POST to object. NOT STATIC.\r
+ *\r
+ * @param array $array $_GET or $_POST array to import\r
+ * @param string|bool $index Index/name that the config variables are in\r
+ * @param array|bool $allowed List of allowed namespaces/directives\r
+ * @param bool $mq_fix Boolean whether or not to enable magic quotes fix\r
+ */\r
+ public function mergeArrayFromForm($array, $index = false, $allowed = true, $mq_fix = true)\r
+ {\r
+ $ret = HTMLPurifier_Config::prepareArrayFromForm($array, $index, $allowed, $mq_fix, $this->def);\r
+ $this->loadArray($ret);\r
+ }\r
+\r
+ /**\r
+ * Prepares an array from a form into something usable for the more\r
+ * strict parts of HTMLPurifier_Config\r
+ *\r
+ * @param array $array $_GET or $_POST array to import\r
+ * @param string|bool $index Index/name that the config variables are in\r
+ * @param array|bool $allowed List of allowed namespaces/directives\r
+ * @param bool $mq_fix Boolean whether or not to enable magic quotes fix\r
+ * @param HTMLPurifier_ConfigSchema $schema Schema to use, if not global copy\r
+ *\r
+ * @return array\r
+ */\r
+ public static function prepareArrayFromForm($array, $index = false, $allowed = true, $mq_fix = true, $schema = null)\r
+ {\r
+ if ($index !== false) {\r
+ $array = (isset($array[$index]) && is_array($array[$index])) ? $array[$index] : array();\r
+ }\r
+ $mq = $mq_fix && function_exists('get_magic_quotes_gpc') && get_magic_quotes_gpc();\r
+\r
+ $allowed = HTMLPurifier_Config::getAllowedDirectivesForForm($allowed, $schema);\r
+ $ret = array();\r
+ foreach ($allowed as $key) {\r
+ list($ns, $directive) = $key;\r
+ $skey = "$ns.$directive";\r
+ if (!empty($array["Null_$skey"])) {\r
+ $ret[$ns][$directive] = null;\r
+ continue;\r
+ }\r
+ if (!isset($array[$skey])) {\r
+ continue;\r
+ }\r
+ $value = $mq ? stripslashes($array[$skey]) : $array[$skey];\r
+ $ret[$ns][$directive] = $value;\r
+ }\r
+ return $ret;\r
+ }\r
+\r
+ /**\r
+ * Loads configuration values from an ini file\r
+ *\r
+ * @param string $filename Name of ini file\r
+ */\r
+ public function loadIni($filename)\r
+ {\r
+ if ($this->isFinalized('Cannot load directives after finalization')) {\r
+ return;\r
+ }\r
+ $array = parse_ini_file($filename, true);\r
+ $this->loadArray($array);\r
+ }\r
+\r
+ /**\r
+ * Checks whether or not the configuration object is finalized.\r
+ *\r
+ * @param string|bool $error String error message, or false for no error\r
+ *\r
+ * @return bool\r
+ */\r
+ public function isFinalized($error = false)\r
+ {\r
+ if ($this->finalized && $error) {\r
+ $this->triggerError($error, E_USER_ERROR);\r
+ }\r
+ return $this->finalized;\r
+ }\r
+\r
+ /**\r
+ * Finalizes configuration only if auto finalize is on and not\r
+ * already finalized\r
+ */\r
+ public function autoFinalize()\r
+ {\r
+ if ($this->autoFinalize) {\r
+ $this->finalize();\r
+ } else {\r
+ $this->plist->squash(true);\r
+ }\r
+ }\r
+\r
+ /**\r
+ * Finalizes a configuration object, prohibiting further change\r
+ */\r
+ public function finalize()\r
+ {\r
+ $this->finalized = true;\r
+ $this->parser = null;\r
+ }\r
+\r
+ /**\r
+ * Produces a nicely formatted error message by supplying the\r
+ * stack frame information OUTSIDE of HTMLPurifier_Config.\r
+ *\r
+ * @param string $msg An error message\r
+ * @param int $no An error number\r
+ */\r
+ protected function triggerError($msg, $no)\r
+ {\r
+ // determine previous stack frame\r
+ $extra = '';\r
+ if ($this->chatty) {\r
+ $trace = debug_backtrace();\r
+ // zip(tail(trace), trace) -- but PHP is not Haskell har har\r
+ for ($i = 0, $c = count($trace); $i < $c - 1; $i++) {\r
+ // XXX this is not correct on some versions of HTML Purifier\r
+ if (isset($trace[$i + 1]['class']) && $trace[$i + 1]['class'] === 'HTMLPurifier_Config') {\r
+ continue;\r
+ }\r
+ $frame = $trace[$i];\r
+ $extra = " invoked on line {$frame['line']} in file {$frame['file']}";\r
+ break;\r
+ }\r
+ }\r
+ trigger_error($msg . $extra, $no);\r
+ }\r
+\r
+ /**\r
+ * Returns a serialized form of the configuration object that can\r
+ * be reconstituted.\r
+ *\r
+ * @return string\r
+ */\r
+ public function serialize()\r
+ {\r
+ $this->getDefinition('HTML');\r
+ $this->getDefinition('CSS');\r
+ $this->getDefinition('URI');\r
+ return serialize($this);\r
+ }\r
+\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Configuration definition, defines directives and their defaults.\r
+ */\r
+class HTMLPurifier_ConfigSchema\r
+{\r
+ /**\r
+ * Defaults of the directives and namespaces.\r
+ * @type array\r
+ * @note This shares the exact same structure as HTMLPurifier_Config::$conf\r
+ */\r
+ public $defaults = array();\r
+\r
+ /**\r
+ * The default property list. Do not edit this property list.\r
+ * @type array\r
+ */\r
+ public $defaultPlist;\r
+\r
+ /**\r
+ * Definition of the directives.\r
+ * The structure of this is:\r
+ *\r
+ * array(\r
+ * 'Namespace' => array(\r
+ * 'Directive' => new stdClass(),\r
+ * )\r
+ * )\r
+ *\r
+ * The stdClass may have the following properties:\r
+ *\r
+ * - If isAlias isn't set:\r
+ * - type: Integer type of directive, see HTMLPurifier_VarParser for definitions\r
+ * - allow_null: If set, this directive allows null values\r
+ * - aliases: If set, an associative array of value aliases to real values\r
+ * - allowed: If set, a lookup array of allowed (string) values\r
+ * - If isAlias is set:\r
+ * - namespace: Namespace this directive aliases to\r
+ * - name: Directive name this directive aliases to\r
+ *\r
+ * In certain degenerate cases, stdClass will actually be an integer. In\r
+ * that case, the value is equivalent to an stdClass with the type\r
+ * property set to the integer. If the integer is negative, type is\r
+ * equal to the absolute value of integer, and allow_null is true.\r
+ *\r
+ * This class is friendly with HTMLPurifier_Config. If you need introspection\r
+ * about the schema, you're better of using the ConfigSchema_Interchange,\r
+ * which uses more memory but has much richer information.\r
+ * @type array\r
+ */\r
+ public $info = array();\r
+\r
+ /**\r
+ * Application-wide singleton\r
+ * @type HTMLPurifier_ConfigSchema\r
+ */\r
+ protected static $singleton;\r
+\r
+ public function __construct()\r
+ {\r
+ $this->defaultPlist = new HTMLPurifier_PropertyList();\r
+ }\r
+\r
+ /**\r
+ * Unserializes the default ConfigSchema.\r
+ * @return HTMLPurifier_ConfigSchema\r
+ */\r
+ public static function makeFromSerial()\r
+ {\r
+ $contents = file_get_contents(HTMLPURIFIER_PREFIX . '/HTMLPurifier/ConfigSchema/schema.ser');\r
+ $r = unserialize($contents);\r
+ if (!$r) {\r
+ $hash = sha1($contents);\r
+ trigger_error("Unserialization of configuration schema failed, sha1 of file was $hash", E_USER_ERROR);\r
+ }\r
+ return $r;\r
+ }\r
+\r
+ /**\r
+ * Retrieves an instance of the application-wide configuration definition.\r
+ * @param HTMLPurifier_ConfigSchema $prototype\r
+ * @return HTMLPurifier_ConfigSchema\r
+ */\r
+ public static function instance($prototype = null)\r
+ {\r
+ if ($prototype !== null) {\r
+ HTMLPurifier_ConfigSchema::$singleton = $prototype;\r
+ } elseif (HTMLPurifier_ConfigSchema::$singleton === null || $prototype === true) {\r
+ HTMLPurifier_ConfigSchema::$singleton = HTMLPurifier_ConfigSchema::makeFromSerial();\r
+ }\r
+ return HTMLPurifier_ConfigSchema::$singleton;\r
+ }\r
+\r
+ /**\r
+ * Defines a directive for configuration\r
+ * @warning Will fail of directive's namespace is defined.\r
+ * @warning This method's signature is slightly different from the legacy\r
+ * define() static method! Beware!\r
+ * @param string $key Name of directive\r
+ * @param mixed $default Default value of directive\r
+ * @param string $type Allowed type of the directive. See\r
+ * HTMLPurifier_VarParser::$types for allowed values\r
+ * @param bool $allow_null Whether or not to allow null values\r
+ */\r
+ public function add($key, $default, $type, $allow_null)\r
+ {\r
+ $obj = new stdClass();\r
+ $obj->type = is_int($type) ? $type : HTMLPurifier_VarParser::$types[$type];\r
+ if ($allow_null) {\r
+ $obj->allow_null = true;\r
+ }\r
+ $this->info[$key] = $obj;\r
+ $this->defaults[$key] = $default;\r
+ $this->defaultPlist->set($key, $default);\r
+ }\r
+\r
+ /**\r
+ * Defines a directive value alias.\r
+ *\r
+ * Directive value aliases are convenient for developers because it lets\r
+ * them set a directive to several values and get the same result.\r
+ * @param string $key Name of Directive\r
+ * @param array $aliases Hash of aliased values to the real alias\r
+ */\r
+ public function addValueAliases($key, $aliases)\r
+ {\r
+ if (!isset($this->info[$key]->aliases)) {\r
+ $this->info[$key]->aliases = array();\r
+ }\r
+ foreach ($aliases as $alias => $real) {\r
+ $this->info[$key]->aliases[$alias] = $real;\r
+ }\r
+ }\r
+\r
+ /**\r
+ * Defines a set of allowed values for a directive.\r
+ * @warning This is slightly different from the corresponding static\r
+ * method definition.\r
+ * @param string $key Name of directive\r
+ * @param array $allowed Lookup array of allowed values\r
+ */\r
+ public function addAllowedValues($key, $allowed)\r
+ {\r
+ $this->info[$key]->allowed = $allowed;\r
+ }\r
+\r
+ /**\r
+ * Defines a directive alias for backwards compatibility\r
+ * @param string $key Directive that will be aliased\r
+ * @param string $new_key Directive that the alias will be to\r
+ */\r
+ public function addAlias($key, $new_key)\r
+ {\r
+ $obj = new stdClass;\r
+ $obj->key = $new_key;\r
+ $obj->isAlias = true;\r
+ $this->info[$key] = $obj;\r
+ }\r
+\r
+ /**\r
+ * Replaces any stdClass that only has the type property with type integer.\r
+ */\r
+ public function postProcess()\r
+ {\r
+ foreach ($this->info as $key => $v) {\r
+ if (count((array) $v) == 1) {\r
+ $this->info[$key] = $v->type;\r
+ } elseif (count((array) $v) == 2 && isset($v->allow_null)) {\r
+ $this->info[$key] = -$v->type;\r
+ }\r
+ }\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * @todo Unit test\r
+ */\r
+class HTMLPurifier_ContentSets\r
+{\r
+\r
+ /**\r
+ * List of content set strings (pipe separators) indexed by name.\r
+ * @type array\r
+ */\r
+ public $info = array();\r
+\r
+ /**\r
+ * List of content set lookups (element => true) indexed by name.\r
+ * @type array\r
+ * @note This is in HTMLPurifier_HTMLDefinition->info_content_sets\r
+ */\r
+ public $lookup = array();\r
+\r
+ /**\r
+ * Synchronized list of defined content sets (keys of info).\r
+ * @type array\r
+ */\r
+ protected $keys = array();\r
+ /**\r
+ * Synchronized list of defined content values (values of info).\r
+ * @type array\r
+ */\r
+ protected $values = array();\r
+\r
+ /**\r
+ * Merges in module's content sets, expands identifiers in the content\r
+ * sets and populates the keys, values and lookup member variables.\r
+ * @param HTMLPurifier_HTMLModule[] $modules List of HTMLPurifier_HTMLModule\r
+ */\r
+ public function __construct($modules)\r
+ {\r
+ if (!is_array($modules)) {\r
+ $modules = array($modules);\r
+ }\r
+ // populate content_sets based on module hints\r
+ // sorry, no way of overloading\r
+ foreach ($modules as $module) {\r
+ foreach ($module->content_sets as $key => $value) {\r
+ $temp = $this->convertToLookup($value);\r
+ if (isset($this->lookup[$key])) {\r
+ // add it into the existing content set\r
+ $this->lookup[$key] = array_merge($this->lookup[$key], $temp);\r
+ } else {\r
+ $this->lookup[$key] = $temp;\r
+ }\r
+ }\r
+ }\r
+ $old_lookup = false;\r
+ while ($old_lookup !== $this->lookup) {\r
+ $old_lookup = $this->lookup;\r
+ foreach ($this->lookup as $i => $set) {\r
+ $add = array();\r
+ foreach ($set as $element => $x) {\r
+ if (isset($this->lookup[$element])) {\r
+ $add += $this->lookup[$element];\r
+ unset($this->lookup[$i][$element]);\r
+ }\r
+ }\r
+ $this->lookup[$i] += $add;\r
+ }\r
+ }\r
+\r
+ foreach ($this->lookup as $key => $lookup) {\r
+ $this->info[$key] = implode(' | ', array_keys($lookup));\r
+ }\r
+ $this->keys = array_keys($this->info);\r
+ $this->values = array_values($this->info);\r
+ }\r
+\r
+ /**\r
+ * Accepts a definition; generates and assigns a ChildDef for it\r
+ * @param HTMLPurifier_ElementDef $def HTMLPurifier_ElementDef reference\r
+ * @param HTMLPurifier_HTMLModule $module Module that defined the ElementDef\r
+ */\r
+ public function generateChildDef(&$def, $module)\r
+ {\r
+ if (!empty($def->child)) { // already done!\r
+ return;\r
+ }\r
+ $content_model = $def->content_model;\r
+ if (is_string($content_model)) {\r
+ // Assume that $this->keys is alphanumeric\r
+ $def->content_model = preg_replace_callback(\r
+ '/\b(' . implode('|', $this->keys) . ')\b/',\r
+ array($this, 'generateChildDefCallback'),\r
+ $content_model\r
+ );\r
+ //$def->content_model = str_replace(\r
+ // $this->keys, $this->values, $content_model);\r
+ }\r
+ $def->child = $this->getChildDef($def, $module);\r
+ }\r
+\r
+ public function generateChildDefCallback($matches)\r
+ {\r
+ return $this->info[$matches[0]];\r
+ }\r
+\r
+ /**\r
+ * Instantiates a ChildDef based on content_model and content_model_type\r
+ * member variables in HTMLPurifier_ElementDef\r
+ * @note This will also defer to modules for custom HTMLPurifier_ChildDef\r
+ * subclasses that need content set expansion\r
+ * @param HTMLPurifier_ElementDef $def HTMLPurifier_ElementDef to have ChildDef extracted\r
+ * @param HTMLPurifier_HTMLModule $module Module that defined the ElementDef\r
+ * @return HTMLPurifier_ChildDef corresponding to ElementDef\r
+ */\r
+ public function getChildDef($def, $module)\r
+ {\r
+ $value = $def->content_model;\r
+ if (is_object($value)) {\r
+ trigger_error(\r
+ 'Literal object child definitions should be stored in '.\r
+ 'ElementDef->child not ElementDef->content_model',\r
+ E_USER_NOTICE\r
+ );\r
+ return $value;\r
+ }\r
+ switch ($def->content_model_type) {\r
+ case 'required':\r
+ return new HTMLPurifier_ChildDef_Required($value);\r
+ case 'optional':\r
+ return new HTMLPurifier_ChildDef_Optional($value);\r
+ case 'empty':\r
+ return new HTMLPurifier_ChildDef_Empty();\r
+ case 'custom':\r
+ return new HTMLPurifier_ChildDef_Custom($value);\r
+ }\r
+ // defer to its module\r
+ $return = false;\r
+ if ($module->defines_child_def) { // save a func call\r
+ $return = $module->getChildDef($def);\r
+ }\r
+ if ($return !== false) {\r
+ return $return;\r
+ }\r
+ // error-out\r
+ trigger_error(\r
+ 'Could not determine which ChildDef class to instantiate',\r
+ E_USER_ERROR\r
+ );\r
+ return false;\r
+ }\r
+\r
+ /**\r
+ * Converts a string list of elements separated by pipes into\r
+ * a lookup array.\r
+ * @param string $string List of elements\r
+ * @return array Lookup array of elements\r
+ */\r
+ protected function convertToLookup($string)\r
+ {\r
+ $array = explode('|', str_replace(' ', '', $string));\r
+ $ret = array();\r
+ foreach ($array as $k) {\r
+ $ret[$k] = true;\r
+ }\r
+ return $ret;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Registry object that contains information about the current context.\r
+ * @warning Is a bit buggy when variables are set to null: it thinks\r
+ * they don't exist! So use false instead, please.\r
+ * @note Since the variables Context deals with may not be objects,\r
+ * references are very important here! Do not remove!\r
+ */\r
+class HTMLPurifier_Context\r
+{\r
+\r
+ /**\r
+ * Private array that stores the references.\r
+ * @type array\r
+ */\r
+ private $_storage = array();\r
+\r
+ /**\r
+ * Registers a variable into the context.\r
+ * @param string $name String name\r
+ * @param mixed $ref Reference to variable to be registered\r
+ */\r
+ public function register($name, &$ref)\r
+ {\r
+ if (array_key_exists($name, $this->_storage)) {\r
+ trigger_error(\r
+ "Name $name produces collision, cannot re-register",\r
+ E_USER_ERROR\r
+ );\r
+ return;\r
+ }\r
+ $this->_storage[$name] =& $ref;\r
+ }\r
+\r
+ /**\r
+ * Retrieves a variable reference from the context.\r
+ * @param string $name String name\r
+ * @param bool $ignore_error Boolean whether or not to ignore error\r
+ * @return mixed\r
+ */\r
+ public function &get($name, $ignore_error = false)\r
+ {\r
+ if (!array_key_exists($name, $this->_storage)) {\r
+ if (!$ignore_error) {\r
+ trigger_error(\r
+ "Attempted to retrieve non-existent variable $name",\r
+ E_USER_ERROR\r
+ );\r
+ }\r
+ $var = null; // so we can return by reference\r
+ return $var;\r
+ }\r
+ return $this->_storage[$name];\r
+ }\r
+\r
+ /**\r
+ * Destroys a variable in the context.\r
+ * @param string $name String name\r
+ */\r
+ public function destroy($name)\r
+ {\r
+ if (!array_key_exists($name, $this->_storage)) {\r
+ trigger_error(\r
+ "Attempted to destroy non-existent variable $name",\r
+ E_USER_ERROR\r
+ );\r
+ return;\r
+ }\r
+ unset($this->_storage[$name]);\r
+ }\r
+\r
+ /**\r
+ * Checks whether or not the variable exists.\r
+ * @param string $name String name\r
+ * @return bool\r
+ */\r
+ public function exists($name)\r
+ {\r
+ return array_key_exists($name, $this->_storage);\r
+ }\r
+\r
+ /**\r
+ * Loads a series of variables from an associative array\r
+ * @param array $context_array Assoc array of variables to load\r
+ */\r
+ public function loadArray($context_array)\r
+ {\r
+ foreach ($context_array as $key => $discard) {\r
+ $this->register($key, $context_array[$key]);\r
+ }\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Abstract class representing Definition cache managers that implements\r
+ * useful common methods and is a factory.\r
+ * @todo Create a separate maintenance file advanced users can use to\r
+ * cache their custom HTMLDefinition, which can be loaded\r
+ * via a configuration directive\r
+ * @todo Implement memcached\r
+ */\r
+abstract class HTMLPurifier_DefinitionCache\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $type;\r
+\r
+ /**\r
+ * @param string $type Type of definition objects this instance of the\r
+ * cache will handle.\r
+ */\r
+ public function __construct($type)\r
+ {\r
+ $this->type = $type;\r
+ }\r
+\r
+ /**\r
+ * Generates a unique identifier for a particular configuration\r
+ * @param HTMLPurifier_Config $config Instance of HTMLPurifier_Config\r
+ * @return string\r
+ */\r
+ public function generateKey($config)\r
+ {\r
+ return $config->version . ',' . // possibly replace with function calls\r
+ $config->getBatchSerial($this->type) . ',' .\r
+ $config->get($this->type . '.DefinitionRev');\r
+ }\r
+\r
+ /**\r
+ * Tests whether or not a key is old with respect to the configuration's\r
+ * version and revision number.\r
+ * @param string $key Key to test\r
+ * @param HTMLPurifier_Config $config Instance of HTMLPurifier_Config to test against\r
+ * @return bool\r
+ */\r
+ public function isOld($key, $config)\r
+ {\r
+ if (substr_count($key, ',') < 2) {\r
+ return true;\r
+ }\r
+ list($version, $hash, $revision) = explode(',', $key, 3);\r
+ $compare = version_compare($version, $config->version);\r
+ // version mismatch, is always old\r
+ if ($compare != 0) {\r
+ return true;\r
+ }\r
+ // versions match, ids match, check revision number\r
+ if ($hash == $config->getBatchSerial($this->type) &&\r
+ $revision < $config->get($this->type . '.DefinitionRev')) {\r
+ return true;\r
+ }\r
+ return false;\r
+ }\r
+\r
+ /**\r
+ * Checks if a definition's type jives with the cache's type\r
+ * @note Throws an error on failure\r
+ * @param HTMLPurifier_Definition $def Definition object to check\r
+ * @return bool true if good, false if not\r
+ */\r
+ public function checkDefType($def)\r
+ {\r
+ if ($def->type !== $this->type) {\r
+ trigger_error("Cannot use definition of type {$def->type} in cache for {$this->type}");\r
+ return false;\r
+ }\r
+ return true;\r
+ }\r
+\r
+ /**\r
+ * Adds a definition object to the cache\r
+ * @param HTMLPurifier_Definition $def\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ abstract public function add($def, $config);\r
+\r
+ /**\r
+ * Unconditionally saves a definition object to the cache\r
+ * @param HTMLPurifier_Definition $def\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ abstract public function set($def, $config);\r
+\r
+ /**\r
+ * Replace an object in the cache\r
+ * @param HTMLPurifier_Definition $def\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ abstract public function replace($def, $config);\r
+\r
+ /**\r
+ * Retrieves a definition object from the cache\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ abstract public function get($config);\r
+\r
+ /**\r
+ * Removes a definition object to the cache\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ abstract public function remove($config);\r
+\r
+ /**\r
+ * Clears all objects from cache\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ abstract public function flush($config);\r
+\r
+ /**\r
+ * Clears all expired (older version or revision) objects from cache\r
+ * @note Be careful implementing this method as flush. Flush must\r
+ * not interfere with other Definition types, and cleanup()\r
+ * should not be repeatedly called by userland code.\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ abstract public function cleanup($config);\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Responsible for creating definition caches.\r
+ */\r
+class HTMLPurifier_DefinitionCacheFactory\r
+{\r
+ /**\r
+ * @type array\r
+ */\r
+ protected $caches = array('Serializer' => array());\r
+\r
+ /**\r
+ * @type array\r
+ */\r
+ protected $implementations = array();\r
+\r
+ /**\r
+ * @type HTMLPurifier_DefinitionCache_Decorator[]\r
+ */\r
+ protected $decorators = array();\r
+\r
+ /**\r
+ * Initialize default decorators\r
+ */\r
+ public function setup()\r
+ {\r
+ $this->addDecorator('Cleanup');\r
+ }\r
+\r
+ /**\r
+ * Retrieves an instance of global definition cache factory.\r
+ * @param HTMLPurifier_DefinitionCacheFactory $prototype\r
+ * @return HTMLPurifier_DefinitionCacheFactory\r
+ */\r
+ public static function instance($prototype = null)\r
+ {\r
+ static $instance;\r
+ if ($prototype !== null) {\r
+ $instance = $prototype;\r
+ } elseif ($instance === null || $prototype === true) {\r
+ $instance = new HTMLPurifier_DefinitionCacheFactory();\r
+ $instance->setup();\r
+ }\r
+ return $instance;\r
+ }\r
+\r
+ /**\r
+ * Registers a new definition cache object\r
+ * @param string $short Short name of cache object, for reference\r
+ * @param string $long Full class name of cache object, for construction\r
+ */\r
+ public function register($short, $long)\r
+ {\r
+ $this->implementations[$short] = $long;\r
+ }\r
+\r
+ /**\r
+ * Factory method that creates a cache object based on configuration\r
+ * @param string $type Name of definitions handled by cache\r
+ * @param HTMLPurifier_Config $config Config instance\r
+ * @return mixed\r
+ */\r
+ public function create($type, $config)\r
+ {\r
+ $method = $config->get('Cache.DefinitionImpl');\r
+ if ($method === null) {\r
+ return new HTMLPurifier_DefinitionCache_Null($type);\r
+ }\r
+ if (!empty($this->caches[$method][$type])) {\r
+ return $this->caches[$method][$type];\r
+ }\r
+ if (isset($this->implementations[$method]) &&\r
+ class_exists($class = $this->implementations[$method], false)) {\r
+ $cache = new $class($type);\r
+ } else {\r
+ if ($method != 'Serializer') {\r
+ trigger_error("Unrecognized DefinitionCache $method, using Serializer instead", E_USER_WARNING);\r
+ }\r
+ $cache = new HTMLPurifier_DefinitionCache_Serializer($type);\r
+ }\r
+ foreach ($this->decorators as $decorator) {\r
+ $new_cache = $decorator->decorate($cache);\r
+ // prevent infinite recursion in PHP 4\r
+ unset($cache);\r
+ $cache = $new_cache;\r
+ }\r
+ $this->caches[$method][$type] = $cache;\r
+ return $this->caches[$method][$type];\r
+ }\r
+\r
+ /**\r
+ * Registers a decorator to add to all new cache objects\r
+ * @param HTMLPurifier_DefinitionCache_Decorator|string $decorator An instance or the name of a decorator\r
+ */\r
+ public function addDecorator($decorator)\r
+ {\r
+ if (is_string($decorator)) {\r
+ $class = "HTMLPurifier_DefinitionCache_Decorator_$decorator";\r
+ $decorator = new $class;\r
+ }\r
+ $this->decorators[$decorator->name] = $decorator;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Represents a document type, contains information on which modules\r
+ * need to be loaded.\r
+ * @note This class is inspected by Printer_HTMLDefinition->renderDoctype.\r
+ * If structure changes, please update that function.\r
+ */\r
+class HTMLPurifier_Doctype\r
+{\r
+ /**\r
+ * Full name of doctype\r
+ * @type string\r
+ */\r
+ public $name;\r
+\r
+ /**\r
+ * List of standard modules (string identifiers or literal objects)\r
+ * that this doctype uses\r
+ * @type array\r
+ */\r
+ public $modules = array();\r
+\r
+ /**\r
+ * List of modules to use for tidying up code\r
+ * @type array\r
+ */\r
+ public $tidyModules = array();\r
+\r
+ /**\r
+ * Is the language derived from XML (i.e. XHTML)?\r
+ * @type bool\r
+ */\r
+ public $xml = true;\r
+\r
+ /**\r
+ * List of aliases for this doctype\r
+ * @type array\r
+ */\r
+ public $aliases = array();\r
+\r
+ /**\r
+ * Public DTD identifier\r
+ * @type string\r
+ */\r
+ public $dtdPublic;\r
+\r
+ /**\r
+ * System DTD identifier\r
+ * @type string\r
+ */\r
+ public $dtdSystem;\r
+\r
+ public function __construct(\r
+ $name = null,\r
+ $xml = true,\r
+ $modules = array(),\r
+ $tidyModules = array(),\r
+ $aliases = array(),\r
+ $dtd_public = null,\r
+ $dtd_system = null\r
+ ) {\r
+ $this->name = $name;\r
+ $this->xml = $xml;\r
+ $this->modules = $modules;\r
+ $this->tidyModules = $tidyModules;\r
+ $this->aliases = $aliases;\r
+ $this->dtdPublic = $dtd_public;\r
+ $this->dtdSystem = $dtd_system;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+class HTMLPurifier_DoctypeRegistry\r
+{\r
+\r
+ /**\r
+ * Hash of doctype names to doctype objects.\r
+ * @type array\r
+ */\r
+ protected $doctypes;\r
+\r
+ /**\r
+ * Lookup table of aliases to real doctype names.\r
+ * @type array\r
+ */\r
+ protected $aliases;\r
+\r
+ /**\r
+ * Registers a doctype to the registry\r
+ * @note Accepts a fully-formed doctype object, or the\r
+ * parameters for constructing a doctype object\r
+ * @param string $doctype Name of doctype or literal doctype object\r
+ * @param bool $xml\r
+ * @param array $modules Modules doctype will load\r
+ * @param array $tidy_modules Modules doctype will load for certain modes\r
+ * @param array $aliases Alias names for doctype\r
+ * @param string $dtd_public\r
+ * @param string $dtd_system\r
+ * @return HTMLPurifier_Doctype Editable registered doctype\r
+ */\r
+ public function register(\r
+ $doctype,\r
+ $xml = true,\r
+ $modules = array(),\r
+ $tidy_modules = array(),\r
+ $aliases = array(),\r
+ $dtd_public = null,\r
+ $dtd_system = null\r
+ ) {\r
+ if (!is_array($modules)) {\r
+ $modules = array($modules);\r
+ }\r
+ if (!is_array($tidy_modules)) {\r
+ $tidy_modules = array($tidy_modules);\r
+ }\r
+ if (!is_array($aliases)) {\r
+ $aliases = array($aliases);\r
+ }\r
+ if (!is_object($doctype)) {\r
+ $doctype = new HTMLPurifier_Doctype(\r
+ $doctype,\r
+ $xml,\r
+ $modules,\r
+ $tidy_modules,\r
+ $aliases,\r
+ $dtd_public,\r
+ $dtd_system\r
+ );\r
+ }\r
+ $this->doctypes[$doctype->name] = $doctype;\r
+ $name = $doctype->name;\r
+ // hookup aliases\r
+ foreach ($doctype->aliases as $alias) {\r
+ if (isset($this->doctypes[$alias])) {\r
+ continue;\r
+ }\r
+ $this->aliases[$alias] = $name;\r
+ }\r
+ // remove old aliases\r
+ if (isset($this->aliases[$name])) {\r
+ unset($this->aliases[$name]);\r
+ }\r
+ return $doctype;\r
+ }\r
+\r
+ /**\r
+ * Retrieves reference to a doctype of a certain name\r
+ * @note This function resolves aliases\r
+ * @note When possible, use the more fully-featured make()\r
+ * @param string $doctype Name of doctype\r
+ * @return HTMLPurifier_Doctype Editable doctype object\r
+ */\r
+ public function get($doctype)\r
+ {\r
+ if (isset($this->aliases[$doctype])) {\r
+ $doctype = $this->aliases[$doctype];\r
+ }\r
+ if (!isset($this->doctypes[$doctype])) {\r
+ trigger_error('Doctype ' . htmlspecialchars($doctype) . ' does not exist', E_USER_ERROR);\r
+ $anon = new HTMLPurifier_Doctype($doctype);\r
+ return $anon;\r
+ }\r
+ return $this->doctypes[$doctype];\r
+ }\r
+\r
+ /**\r
+ * Creates a doctype based on a configuration object,\r
+ * will perform initialization on the doctype\r
+ * @note Use this function to get a copy of doctype that config\r
+ * can hold on to (this is necessary in order to tell\r
+ * Generator whether or not the current document is XML\r
+ * based or not).\r
+ * @param HTMLPurifier_Config $config\r
+ * @return HTMLPurifier_Doctype\r
+ */\r
+ public function make($config)\r
+ {\r
+ return clone $this->get($this->getDoctypeFromConfig($config));\r
+ }\r
+\r
+ /**\r
+ * Retrieves the doctype from the configuration object\r
+ * @param HTMLPurifier_Config $config\r
+ * @return string\r
+ */\r
+ public function getDoctypeFromConfig($config)\r
+ {\r
+ // recommended test\r
+ $doctype = $config->get('HTML.Doctype');\r
+ if (!empty($doctype)) {\r
+ return $doctype;\r
+ }\r
+ $doctype = $config->get('HTML.CustomDoctype');\r
+ if (!empty($doctype)) {\r
+ return $doctype;\r
+ }\r
+ // backwards-compatibility\r
+ if ($config->get('HTML.XHTML')) {\r
+ $doctype = 'XHTML 1.0';\r
+ } else {\r
+ $doctype = 'HTML 4.01';\r
+ }\r
+ if ($config->get('HTML.Strict')) {\r
+ $doctype .= ' Strict';\r
+ } else {\r
+ $doctype .= ' Transitional';\r
+ }\r
+ return $doctype;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Structure that stores an HTML element definition. Used by\r
+ * HTMLPurifier_HTMLDefinition and HTMLPurifier_HTMLModule.\r
+ * @note This class is inspected by HTMLPurifier_Printer_HTMLDefinition.\r
+ * Please update that class too.\r
+ * @warning If you add new properties to this class, you MUST update\r
+ * the mergeIn() method.\r
+ */\r
+class HTMLPurifier_ElementDef\r
+{\r
+ /**\r
+ * Does the definition work by itself, or is it created solely\r
+ * for the purpose of merging into another definition?\r
+ * @type bool\r
+ */\r
+ public $standalone = true;\r
+\r
+ /**\r
+ * Associative array of attribute name to HTMLPurifier_AttrDef.\r
+ * @type array\r
+ * @note Before being processed by HTMLPurifier_AttrCollections\r
+ * when modules are finalized during\r
+ * HTMLPurifier_HTMLDefinition->setup(), this array may also\r
+ * contain an array at index 0 that indicates which attribute\r
+ * collections to load into the full array. It may also\r
+ * contain string indentifiers in lieu of HTMLPurifier_AttrDef,\r
+ * see HTMLPurifier_AttrTypes on how they are expanded during\r
+ * HTMLPurifier_HTMLDefinition->setup() processing.\r
+ */\r
+ public $attr = array();\r
+\r
+ // XXX: Design note: currently, it's not possible to override\r
+ // previously defined AttrTransforms without messing around with\r
+ // the final generated config. This is by design; a previous version\r
+ // used an associated list of attr_transform, but it was extremely\r
+ // easy to accidentally override other attribute transforms by\r
+ // forgetting to specify an index (and just using 0.) While we\r
+ // could check this by checking the index number and complaining,\r
+ // there is a second problem which is that it is not at all easy to\r
+ // tell when something is getting overridden. Combine this with a\r
+ // codebase where this isn't really being used, and it's perfect for\r
+ // nuking.\r
+\r
+ /**\r
+ * List of tags HTMLPurifier_AttrTransform to be done before validation.\r
+ * @type array\r
+ */\r
+ public $attr_transform_pre = array();\r
+\r
+ /**\r
+ * List of tags HTMLPurifier_AttrTransform to be done after validation.\r
+ * @type array\r
+ */\r
+ public $attr_transform_post = array();\r
+\r
+ /**\r
+ * HTMLPurifier_ChildDef of this tag.\r
+ * @type HTMLPurifier_ChildDef\r
+ */\r
+ public $child;\r
+\r
+ /**\r
+ * Abstract string representation of internal ChildDef rules.\r
+ * @see HTMLPurifier_ContentSets for how this is parsed and then transformed\r
+ * into an HTMLPurifier_ChildDef.\r
+ * @warning This is a temporary variable that is not available after\r
+ * being processed by HTMLDefinition\r
+ * @type string\r
+ */\r
+ public $content_model;\r
+\r
+ /**\r
+ * Value of $child->type, used to determine which ChildDef to use,\r
+ * used in combination with $content_model.\r
+ * @warning This must be lowercase\r
+ * @warning This is a temporary variable that is not available after\r
+ * being processed by HTMLDefinition\r
+ * @type string\r
+ */\r
+ public $content_model_type;\r
+\r
+ /**\r
+ * Does the element have a content model (#PCDATA | Inline)*? This\r
+ * is important for chameleon ins and del processing in\r
+ * HTMLPurifier_ChildDef_Chameleon. Dynamically set: modules don't\r
+ * have to worry about this one.\r
+ * @type bool\r
+ */\r
+ public $descendants_are_inline = false;\r
+\r
+ /**\r
+ * List of the names of required attributes this element has.\r
+ * Dynamically populated by HTMLPurifier_HTMLDefinition::getElement()\r
+ * @type array\r
+ */\r
+ public $required_attr = array();\r
+\r
+ /**\r
+ * Lookup table of tags excluded from all descendants of this tag.\r
+ * @type array\r
+ * @note SGML permits exclusions for all descendants, but this is\r
+ * not possible with DTDs or XML Schemas. W3C has elected to\r
+ * use complicated compositions of content_models to simulate\r
+ * exclusion for children, but we go the simpler, SGML-style\r
+ * route of flat-out exclusions, which correctly apply to\r
+ * all descendants and not just children. Note that the XHTML\r
+ * Modularization Abstract Modules are blithely unaware of such\r
+ * distinctions.\r
+ */\r
+ public $excludes = array();\r
+\r
+ /**\r
+ * This tag is explicitly auto-closed by the following tags.\r
+ * @type array\r
+ */\r
+ public $autoclose = array();\r
+\r
+ /**\r
+ * If a foreign element is found in this element, test if it is\r
+ * allowed by this sub-element; if it is, instead of closing the\r
+ * current element, place it inside this element.\r
+ * @type string\r
+ */\r
+ public $wrap;\r
+\r
+ /**\r
+ * Whether or not this is a formatting element affected by the\r
+ * "Active Formatting Elements" algorithm.\r
+ * @type bool\r
+ */\r
+ public $formatting;\r
+\r
+ /**\r
+ * Low-level factory constructor for creating new standalone element defs\r
+ */\r
+ public static function create($content_model, $content_model_type, $attr)\r
+ {\r
+ $def = new HTMLPurifier_ElementDef();\r
+ $def->content_model = $content_model;\r
+ $def->content_model_type = $content_model_type;\r
+ $def->attr = $attr;\r
+ return $def;\r
+ }\r
+\r
+ /**\r
+ * Merges the values of another element definition into this one.\r
+ * Values from the new element def take precedence if a value is\r
+ * not mergeable.\r
+ * @param HTMLPurifier_ElementDef $def\r
+ */\r
+ public function mergeIn($def)\r
+ {\r
+ // later keys takes precedence\r
+ foreach ($def->attr as $k => $v) {\r
+ if ($k === 0) {\r
+ // merge in the includes\r
+ // sorry, no way to override an include\r
+ foreach ($v as $v2) {\r
+ $this->attr[0][] = $v2;\r
+ }\r
+ continue;\r
+ }\r
+ if ($v === false) {\r
+ if (isset($this->attr[$k])) {\r
+ unset($this->attr[$k]);\r
+ }\r
+ continue;\r
+ }\r
+ $this->attr[$k] = $v;\r
+ }\r
+ $this->_mergeAssocArray($this->excludes, $def->excludes);\r
+ $this->attr_transform_pre = array_merge($this->attr_transform_pre, $def->attr_transform_pre);\r
+ $this->attr_transform_post = array_merge($this->attr_transform_post, $def->attr_transform_post);\r
+\r
+ if (!empty($def->content_model)) {\r
+ $this->content_model =\r
+ str_replace("#SUPER", $this->content_model, $def->content_model);\r
+ $this->child = false;\r
+ }\r
+ if (!empty($def->content_model_type)) {\r
+ $this->content_model_type = $def->content_model_type;\r
+ $this->child = false;\r
+ }\r
+ if (!is_null($def->child)) {\r
+ $this->child = $def->child;\r
+ }\r
+ if (!is_null($def->formatting)) {\r
+ $this->formatting = $def->formatting;\r
+ }\r
+ if ($def->descendants_are_inline) {\r
+ $this->descendants_are_inline = $def->descendants_are_inline;\r
+ }\r
+ }\r
+\r
+ /**\r
+ * Merges one array into another, removes values which equal false\r
+ * @param $a1 Array by reference that is merged into\r
+ * @param $a2 Array that merges into $a1\r
+ */\r
+ private function _mergeAssocArray(&$a1, $a2)\r
+ {\r
+ foreach ($a2 as $k => $v) {\r
+ if ($v === false) {\r
+ if (isset($a1[$k])) {\r
+ unset($a1[$k]);\r
+ }\r
+ continue;\r
+ }\r
+ $a1[$k] = $v;\r
+ }\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * A UTF-8 specific character encoder that handles cleaning and transforming.\r
+ * @note All functions in this class should be static.\r
+ */\r
+class HTMLPurifier_Encoder\r
+{\r
+\r
+ /**\r
+ * Constructor throws fatal error if you attempt to instantiate class\r
+ */\r
+ private function __construct()\r
+ {\r
+ trigger_error('Cannot instantiate encoder, call methods statically', E_USER_ERROR);\r
+ }\r
+\r
+ /**\r
+ * Error-handler that mutes errors, alternative to shut-up operator.\r
+ */\r
+ public static function muteErrorHandler()\r
+ {\r
+ }\r
+\r
+ /**\r
+ * iconv wrapper which mutes errors, but doesn't work around bugs.\r
+ * @param string $in Input encoding\r
+ * @param string $out Output encoding\r
+ * @param string $text The text to convert\r
+ * @return string\r
+ */\r
+ public static function unsafeIconv($in, $out, $text)\r
+ {\r
+ set_error_handler(array('HTMLPurifier_Encoder', 'muteErrorHandler'));\r
+ $r = iconv($in, $out, $text);\r
+ restore_error_handler();\r
+ return $r;\r
+ }\r
+\r
+ /**\r
+ * iconv wrapper which mutes errors and works around bugs.\r
+ * @param string $in Input encoding\r
+ * @param string $out Output encoding\r
+ * @param string $text The text to convert\r
+ * @param int $max_chunk_size\r
+ * @return string\r
+ */\r
+ public static function iconv($in, $out, $text, $max_chunk_size = 8000)\r
+ {\r
+ $code = self::testIconvTruncateBug();\r
+ if ($code == self::ICONV_OK) {\r
+ return self::unsafeIconv($in, $out, $text);\r
+ } elseif ($code == self::ICONV_TRUNCATES) {\r
+ // we can only work around this if the input character set\r
+ // is utf-8\r
+ if ($in == 'utf-8') {\r
+ if ($max_chunk_size < 4) {\r
+ trigger_error('max_chunk_size is too small', E_USER_WARNING);\r
+ return false;\r
+ }\r
+ // split into 8000 byte chunks, but be careful to handle\r
+ // multibyte boundaries properly\r
+ if (($c = strlen($text)) <= $max_chunk_size) {\r
+ return self::unsafeIconv($in, $out, $text);\r
+ }\r
+ $r = '';\r
+ $i = 0;\r
+ while (true) {\r
+ if ($i + $max_chunk_size >= $c) {\r
+ $r .= self::unsafeIconv($in, $out, substr($text, $i));\r
+ break;\r
+ }\r
+ // wibble the boundary\r
+ if (0x80 != (0xC0 & ord($text[$i + $max_chunk_size]))) {\r
+ $chunk_size = $max_chunk_size;\r
+ } elseif (0x80 != (0xC0 & ord($text[$i + $max_chunk_size - 1]))) {\r
+ $chunk_size = $max_chunk_size - 1;\r
+ } elseif (0x80 != (0xC0 & ord($text[$i + $max_chunk_size - 2]))) {\r
+ $chunk_size = $max_chunk_size - 2;\r
+ } elseif (0x80 != (0xC0 & ord($text[$i + $max_chunk_size - 3]))) {\r
+ $chunk_size = $max_chunk_size - 3;\r
+ } else {\r
+ return false; // rather confusing UTF-8...\r
+ }\r
+ $chunk = substr($text, $i, $chunk_size); // substr doesn't mind overlong lengths\r
+ $r .= self::unsafeIconv($in, $out, $chunk);\r
+ $i += $chunk_size;\r
+ }\r
+ return $r;\r
+ } else {\r
+ return false;\r
+ }\r
+ } else {\r
+ return false;\r
+ }\r
+ }\r
+\r
+ /**\r
+ * Cleans a UTF-8 string for well-formedness and SGML validity\r
+ *\r
+ * It will parse according to UTF-8 and return a valid UTF8 string, with\r
+ * non-SGML codepoints excluded.\r
+ *\r
+ * Specifically, it will permit:\r
+ * \x{9}\x{A}\x{D}\x{20}-\x{7E}\x{A0}-\x{D7FF}\x{E000}-\x{FFFD}\x{10000}-\x{10FFFF}\r
+ * Source: https://www.w3.org/TR/REC-xml/#NT-Char\r
+ * Arguably this function should be modernized to the HTML5 set\r
+ * of allowed characters:\r
+ * https://www.w3.org/TR/html5/syntax.html#preprocessing-the-input-stream\r
+ * which simultaneously expand and restrict the set of allowed characters.\r
+ *\r
+ * @param string $str The string to clean\r
+ * @param bool $force_php\r
+ * @return string\r
+ *\r
+ * @note Just for reference, the non-SGML code points are 0 to 31 and\r
+ * 127 to 159, inclusive. However, we allow code points 9, 10\r
+ * and 13, which are the tab, line feed and carriage return\r
+ * respectively. 128 and above the code points map to multibyte\r
+ * UTF-8 representations.\r
+ *\r
+ * @note Fallback code adapted from utf8ToUnicode by Henri Sivonen and\r
+ * hsivonen@iki.fi at <http://iki.fi/hsivonen/php-utf8/> under the\r
+ * LGPL license. Notes on what changed are inside, but in general,\r
+ * the original code transformed UTF-8 text into an array of integer\r
+ * Unicode codepoints. Understandably, transforming that back to\r
+ * a string would be somewhat expensive, so the function was modded to\r
+ * directly operate on the string. However, this discourages code\r
+ * reuse, and the logic enumerated here would be useful for any\r
+ * function that needs to be able to understand UTF-8 characters.\r
+ * As of right now, only smart lossless character encoding converters\r
+ * would need that, and I'm probably not going to implement them.\r
+ */\r
+ public static function cleanUTF8($str, $force_php = false)\r
+ {\r
+ // UTF-8 validity is checked since PHP 4.3.5\r
+ // This is an optimization: if the string is already valid UTF-8, no\r
+ // need to do PHP stuff. 99% of the time, this will be the case.\r
+ if (preg_match(\r
+ '/^[\x{9}\x{A}\x{D}\x{20}-\x{7E}\x{A0}-\x{D7FF}\x{E000}-\x{FFFD}\x{10000}-\x{10FFFF}]*$/Du',\r
+ $str\r
+ )) {\r
+ return $str;\r
+ }\r
+\r
+ $mState = 0; // cached expected number of octets after the current octet\r
+ // until the beginning of the next UTF8 character sequence\r
+ $mUcs4 = 0; // cached Unicode character\r
+ $mBytes = 1; // cached expected number of octets in the current sequence\r
+\r
+ // original code involved an $out that was an array of Unicode\r
+ // codepoints. Instead of having to convert back into UTF-8, we've\r
+ // decided to directly append valid UTF-8 characters onto a string\r
+ // $out once they're done. $char accumulates raw bytes, while $mUcs4\r
+ // turns into the Unicode code point, so there's some redundancy.\r
+\r
+ $out = '';\r
+ $char = '';\r
+\r
+ $len = strlen($str);\r
+ for ($i = 0; $i < $len; $i++) {\r
+ $in = ord($str{$i});\r
+ $char .= $str[$i]; // append byte to char\r
+ if (0 == $mState) {\r
+ // When mState is zero we expect either a US-ASCII character\r
+ // or a multi-octet sequence.\r
+ if (0 == (0x80 & ($in))) {\r
+ // US-ASCII, pass straight through.\r
+ if (($in <= 31 || $in == 127) &&\r
+ !($in == 9 || $in == 13 || $in == 10) // save \r\t\n\r
+ ) {\r
+ // control characters, remove\r
+ } else {\r
+ $out .= $char;\r
+ }\r
+ // reset\r
+ $char = '';\r
+ $mBytes = 1;\r
+ } elseif (0xC0 == (0xE0 & ($in))) {\r
+ // First octet of 2 octet sequence\r
+ $mUcs4 = ($in);\r
+ $mUcs4 = ($mUcs4 & 0x1F) << 6;\r
+ $mState = 1;\r
+ $mBytes = 2;\r
+ } elseif (0xE0 == (0xF0 & ($in))) {\r
+ // First octet of 3 octet sequence\r
+ $mUcs4 = ($in);\r
+ $mUcs4 = ($mUcs4 & 0x0F) << 12;\r
+ $mState = 2;\r
+ $mBytes = 3;\r
+ } elseif (0xF0 == (0xF8 & ($in))) {\r
+ // First octet of 4 octet sequence\r
+ $mUcs4 = ($in);\r
+ $mUcs4 = ($mUcs4 & 0x07) << 18;\r
+ $mState = 3;\r
+ $mBytes = 4;\r
+ } elseif (0xF8 == (0xFC & ($in))) {\r
+ // First octet of 5 octet sequence.\r
+ //\r
+ // This is illegal because the encoded codepoint must be\r
+ // either:\r
+ // (a) not the shortest form or\r
+ // (b) outside the Unicode range of 0-0x10FFFF.\r
+ // Rather than trying to resynchronize, we will carry on\r
+ // until the end of the sequence and let the later error\r
+ // handling code catch it.\r
+ $mUcs4 = ($in);\r
+ $mUcs4 = ($mUcs4 & 0x03) << 24;\r
+ $mState = 4;\r
+ $mBytes = 5;\r
+ } elseif (0xFC == (0xFE & ($in))) {\r
+ // First octet of 6 octet sequence, see comments for 5\r
+ // octet sequence.\r
+ $mUcs4 = ($in);\r
+ $mUcs4 = ($mUcs4 & 1) << 30;\r
+ $mState = 5;\r
+ $mBytes = 6;\r
+ } else {\r
+ // Current octet is neither in the US-ASCII range nor a\r
+ // legal first octet of a multi-octet sequence.\r
+ $mState = 0;\r
+ $mUcs4 = 0;\r
+ $mBytes = 1;\r
+ $char = '';\r
+ }\r
+ } else {\r
+ // When mState is non-zero, we expect a continuation of the\r
+ // multi-octet sequence\r
+ if (0x80 == (0xC0 & ($in))) {\r
+ // Legal continuation.\r
+ $shift = ($mState - 1) * 6;\r
+ $tmp = $in;\r
+ $tmp = ($tmp & 0x0000003F) << $shift;\r
+ $mUcs4 |= $tmp;\r
+\r
+ if (0 == --$mState) {\r
+ // End of the multi-octet sequence. mUcs4 now contains\r
+ // the final Unicode codepoint to be output\r
+\r
+ // Check for illegal sequences and codepoints.\r
+\r
+ // From Unicode 3.1, non-shortest form is illegal\r
+ if (((2 == $mBytes) && ($mUcs4 < 0x0080)) ||\r
+ ((3 == $mBytes) && ($mUcs4 < 0x0800)) ||\r
+ ((4 == $mBytes) && ($mUcs4 < 0x10000)) ||\r
+ (4 < $mBytes) ||\r
+ // From Unicode 3.2, surrogate characters = illegal\r
+ (($mUcs4 & 0xFFFFF800) == 0xD800) ||\r
+ // Codepoints outside the Unicode range are illegal\r
+ ($mUcs4 > 0x10FFFF)\r
+ ) {\r
+\r
+ } elseif (0xFEFF != $mUcs4 && // omit BOM\r
+ // check for valid Char unicode codepoints\r
+ (\r
+ 0x9 == $mUcs4 ||\r
+ 0xA == $mUcs4 ||\r
+ 0xD == $mUcs4 ||\r
+ (0x20 <= $mUcs4 && 0x7E >= $mUcs4) ||\r
+ // 7F-9F is not strictly prohibited by XML,\r
+ // but it is non-SGML, and thus we don't allow it\r
+ (0xA0 <= $mUcs4 && 0xD7FF >= $mUcs4) ||\r
+ (0xE000 <= $mUcs4 && 0xFFFD >= $mUcs4) ||\r
+ (0x10000 <= $mUcs4 && 0x10FFFF >= $mUcs4)\r
+ )\r
+ ) {\r
+ $out .= $char;\r
+ }\r
+ // initialize UTF8 cache (reset)\r
+ $mState = 0;\r
+ $mUcs4 = 0;\r
+ $mBytes = 1;\r
+ $char = '';\r
+ }\r
+ } else {\r
+ // ((0xC0 & (*in) != 0x80) && (mState != 0))\r
+ // Incomplete multi-octet sequence.\r
+ // used to result in complete fail, but we'll reset\r
+ $mState = 0;\r
+ $mUcs4 = 0;\r
+ $mBytes = 1;\r
+ $char ='';\r
+ }\r
+ }\r
+ }\r
+ return $out;\r
+ }\r
+\r
+ /**\r
+ * Translates a Unicode codepoint into its corresponding UTF-8 character.\r
+ * @note Based on Feyd's function at\r
+ * <http://forums.devnetwork.net/viewtopic.php?p=191404#191404>,\r
+ * which is in public domain.\r
+ * @note While we're going to do code point parsing anyway, a good\r
+ * optimization would be to refuse to translate code points that\r
+ * are non-SGML characters. However, this could lead to duplication.\r
+ * @note This is very similar to the unichr function in\r
+ * maintenance/generate-entity-file.php (although this is superior,\r
+ * due to its sanity checks).\r
+ */\r
+\r
+ // +----------+----------+----------+----------+\r
+ // | 33222222 | 22221111 | 111111 | |\r
+ // | 10987654 | 32109876 | 54321098 | 76543210 | bit\r
+ // +----------+----------+----------+----------+\r
+ // | | | | 0xxxxxxx | 1 byte 0x00000000..0x0000007F\r
+ // | | | 110yyyyy | 10xxxxxx | 2 byte 0x00000080..0x000007FF\r
+ // | | 1110zzzz | 10yyyyyy | 10xxxxxx | 3 byte 0x00000800..0x0000FFFF\r
+ // | 11110www | 10wwzzzz | 10yyyyyy | 10xxxxxx | 4 byte 0x00010000..0x0010FFFF\r
+ // +----------+----------+----------+----------+\r
+ // | 00000000 | 00011111 | 11111111 | 11111111 | Theoretical upper limit of legal scalars: 2097151 (0x001FFFFF)\r
+ // | 00000000 | 00010000 | 11111111 | 11111111 | Defined upper limit of legal scalar codes\r
+ // +----------+----------+----------+----------+\r
+\r
+ public static function unichr($code)\r
+ {\r
+ if ($code > 1114111 or $code < 0 or\r
+ ($code >= 55296 and $code <= 57343) ) {\r
+ // bits are set outside the "valid" range as defined\r
+ // by UNICODE 4.1.0\r
+ return '';\r
+ }\r
+\r
+ $x = $y = $z = $w = 0;\r
+ if ($code < 128) {\r
+ // regular ASCII character\r
+ $x = $code;\r
+ } else {\r
+ // set up bits for UTF-8\r
+ $x = ($code & 63) | 128;\r
+ if ($code < 2048) {\r
+ $y = (($code & 2047) >> 6) | 192;\r
+ } else {\r
+ $y = (($code & 4032) >> 6) | 128;\r
+ if ($code < 65536) {\r
+ $z = (($code >> 12) & 15) | 224;\r
+ } else {\r
+ $z = (($code >> 12) & 63) | 128;\r
+ $w = (($code >> 18) & 7) | 240;\r
+ }\r
+ }\r
+ }\r
+ // set up the actual character\r
+ $ret = '';\r
+ if ($w) {\r
+ $ret .= chr($w);\r
+ }\r
+ if ($z) {\r
+ $ret .= chr($z);\r
+ }\r
+ if ($y) {\r
+ $ret .= chr($y);\r
+ }\r
+ $ret .= chr($x);\r
+\r
+ return $ret;\r
+ }\r
+\r
+ /**\r
+ * @return bool\r
+ */\r
+ public static function iconvAvailable()\r
+ {\r
+ static $iconv = null;\r
+ if ($iconv === null) {\r
+ $iconv = function_exists('iconv') && self::testIconvTruncateBug() != self::ICONV_UNUSABLE;\r
+ }\r
+ return $iconv;\r
+ }\r
+\r
+ /**\r
+ * Convert a string to UTF-8 based on configuration.\r
+ * @param string $str The string to convert\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return string\r
+ */\r
+ public static function convertToUTF8($str, $config, $context)\r
+ {\r
+ $encoding = $config->get('Core.Encoding');\r
+ if ($encoding === 'utf-8') {\r
+ return $str;\r
+ }\r
+ static $iconv = null;\r
+ if ($iconv === null) {\r
+ $iconv = self::iconvAvailable();\r
+ }\r
+ if ($iconv && !$config->get('Test.ForceNoIconv')) {\r
+ // unaffected by bugs, since UTF-8 support all characters\r
+ $str = self::unsafeIconv($encoding, 'utf-8//IGNORE', $str);\r
+ if ($str === false) {\r
+ // $encoding is not a valid encoding\r
+ trigger_error('Invalid encoding ' . $encoding, E_USER_ERROR);\r
+ return '';\r
+ }\r
+ // If the string is bjorked by Shift_JIS or a similar encoding\r
+ // that doesn't support all of ASCII, convert the naughty\r
+ // characters to their true byte-wise ASCII/UTF-8 equivalents.\r
+ $str = strtr($str, self::testEncodingSupportsASCII($encoding));\r
+ return $str;\r
+ } elseif ($encoding === 'iso-8859-1') {\r
+ $str = utf8_encode($str);\r
+ return $str;\r
+ }\r
+ $bug = HTMLPurifier_Encoder::testIconvTruncateBug();\r
+ if ($bug == self::ICONV_OK) {\r
+ trigger_error('Encoding not supported, please install iconv', E_USER_ERROR);\r
+ } else {\r
+ trigger_error(\r
+ 'You have a buggy version of iconv, see https://bugs.php.net/bug.php?id=48147 ' .\r
+ 'and http://sourceware.org/bugzilla/show_bug.cgi?id=13541',\r
+ E_USER_ERROR\r
+ );\r
+ }\r
+ }\r
+\r
+ /**\r
+ * Converts a string from UTF-8 based on configuration.\r
+ * @param string $str The string to convert\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return string\r
+ * @note Currently, this is a lossy conversion, with unexpressable\r
+ * characters being omitted.\r
+ */\r
+ public static function convertFromUTF8($str, $config, $context)\r
+ {\r
+ $encoding = $config->get('Core.Encoding');\r
+ if ($escape = $config->get('Core.EscapeNonASCIICharacters')) {\r
+ $str = self::convertToASCIIDumbLossless($str);\r
+ }\r
+ if ($encoding === 'utf-8') {\r
+ return $str;\r
+ }\r
+ static $iconv = null;\r
+ if ($iconv === null) {\r
+ $iconv = self::iconvAvailable();\r
+ }\r
+ if ($iconv && !$config->get('Test.ForceNoIconv')) {\r
+ // Undo our previous fix in convertToUTF8, otherwise iconv will barf\r
+ $ascii_fix = self::testEncodingSupportsASCII($encoding);\r
+ if (!$escape && !empty($ascii_fix)) {\r
+ $clear_fix = array();\r
+ foreach ($ascii_fix as $utf8 => $native) {\r
+ $clear_fix[$utf8] = '';\r
+ }\r
+ $str = strtr($str, $clear_fix);\r
+ }\r
+ $str = strtr($str, array_flip($ascii_fix));\r
+ // Normal stuff\r
+ $str = self::iconv('utf-8', $encoding . '//IGNORE', $str);\r
+ return $str;\r
+ } elseif ($encoding === 'iso-8859-1') {\r
+ $str = utf8_decode($str);\r
+ return $str;\r
+ }\r
+ trigger_error('Encoding not supported', E_USER_ERROR);\r
+ // You might be tempted to assume that the ASCII representation\r
+ // might be OK, however, this is *not* universally true over all\r
+ // encodings. So we take the conservative route here, rather\r
+ // than forcibly turn on %Core.EscapeNonASCIICharacters\r
+ }\r
+\r
+ /**\r
+ * Lossless (character-wise) conversion of HTML to ASCII\r
+ * @param string $str UTF-8 string to be converted to ASCII\r
+ * @return string ASCII encoded string with non-ASCII character entity-ized\r
+ * @warning Adapted from MediaWiki, claiming fair use: this is a common\r
+ * algorithm. If you disagree with this license fudgery,\r
+ * implement it yourself.\r
+ * @note Uses decimal numeric entities since they are best supported.\r
+ * @note This is a DUMB function: it has no concept of keeping\r
+ * character entities that the projected character encoding\r
+ * can allow. We could possibly implement a smart version\r
+ * but that would require it to also know which Unicode\r
+ * codepoints the charset supported (not an easy task).\r
+ * @note Sort of with cleanUTF8() but it assumes that $str is\r
+ * well-formed UTF-8\r
+ */\r
+ public static function convertToASCIIDumbLossless($str)\r
+ {\r
+ $bytesleft = 0;\r
+ $result = '';\r
+ $working = 0;\r
+ $len = strlen($str);\r
+ for ($i = 0; $i < $len; $i++) {\r
+ $bytevalue = ord($str[$i]);\r
+ if ($bytevalue <= 0x7F) { //0xxx xxxx\r
+ $result .= chr($bytevalue);\r
+ $bytesleft = 0;\r
+ } elseif ($bytevalue <= 0xBF) { //10xx xxxx\r
+ $working = $working << 6;\r
+ $working += ($bytevalue & 0x3F);\r
+ $bytesleft--;\r
+ if ($bytesleft <= 0) {\r
+ $result .= "&#" . $working . ";";\r
+ }\r
+ } elseif ($bytevalue <= 0xDF) { //110x xxxx\r
+ $working = $bytevalue & 0x1F;\r
+ $bytesleft = 1;\r
+ } elseif ($bytevalue <= 0xEF) { //1110 xxxx\r
+ $working = $bytevalue & 0x0F;\r
+ $bytesleft = 2;\r
+ } else { //1111 0xxx\r
+ $working = $bytevalue & 0x07;\r
+ $bytesleft = 3;\r
+ }\r
+ }\r
+ return $result;\r
+ }\r
+\r
+ /** No bugs detected in iconv. */\r
+ const ICONV_OK = 0;\r
+\r
+ /** Iconv truncates output if converting from UTF-8 to another\r
+ * character set with //IGNORE, and a non-encodable character is found */\r
+ const ICONV_TRUNCATES = 1;\r
+\r
+ /** Iconv does not support //IGNORE, making it unusable for\r
+ * transcoding purposes */\r
+ const ICONV_UNUSABLE = 2;\r
+\r
+ /**\r
+ * glibc iconv has a known bug where it doesn't handle the magic\r
+ * //IGNORE stanza correctly. In particular, rather than ignore\r
+ * characters, it will return an EILSEQ after consuming some number\r
+ * of characters, and expect you to restart iconv as if it were\r
+ * an E2BIG. Old versions of PHP did not respect the errno, and\r
+ * returned the fragment, so as a result you would see iconv\r
+ * mysteriously truncating output. We can work around this by\r
+ * manually chopping our input into segments of about 8000\r
+ * characters, as long as PHP ignores the error code. If PHP starts\r
+ * paying attention to the error code, iconv becomes unusable.\r
+ *\r
+ * @return int Error code indicating severity of bug.\r
+ */\r
+ public static function testIconvTruncateBug()\r
+ {\r
+ static $code = null;\r
+ if ($code === null) {\r
+ // better not use iconv, otherwise infinite loop!\r
+ $r = self::unsafeIconv('utf-8', 'ascii//IGNORE', "\xCE\xB1" . str_repeat('a', 9000));\r
+ if ($r === false) {\r
+ $code = self::ICONV_UNUSABLE;\r
+ } elseif (($c = strlen($r)) < 9000) {\r
+ $code = self::ICONV_TRUNCATES;\r
+ } elseif ($c > 9000) {\r
+ trigger_error(\r
+ 'Your copy of iconv is extremely buggy. Please notify HTML Purifier maintainers: ' .\r
+ 'include your iconv version as per phpversion()',\r
+ E_USER_ERROR\r
+ );\r
+ } else {\r
+ $code = self::ICONV_OK;\r
+ }\r
+ }\r
+ return $code;\r
+ }\r
+\r
+ /**\r
+ * This expensive function tests whether or not a given character\r
+ * encoding supports ASCII. 7/8-bit encodings like Shift_JIS will\r
+ * fail this test, and require special processing. Variable width\r
+ * encodings shouldn't ever fail.\r
+ *\r
+ * @param string $encoding Encoding name to test, as per iconv format\r
+ * @param bool $bypass Whether or not to bypass the precompiled arrays.\r
+ * @return Array of UTF-8 characters to their corresponding ASCII,\r
+ * which can be used to "undo" any overzealous iconv action.\r
+ */\r
+ public static function testEncodingSupportsASCII($encoding, $bypass = false)\r
+ {\r
+ // All calls to iconv here are unsafe, proof by case analysis:\r
+ // If ICONV_OK, no difference.\r
+ // If ICONV_TRUNCATE, all calls involve one character inputs,\r
+ // so bug is not triggered.\r
+ // If ICONV_UNUSABLE, this call is irrelevant\r
+ static $encodings = array();\r
+ if (!$bypass) {\r
+ if (isset($encodings[$encoding])) {\r
+ return $encodings[$encoding];\r
+ }\r
+ $lenc = strtolower($encoding);\r
+ switch ($lenc) {\r
+ case 'shift_jis':\r
+ return array("\xC2\xA5" => '\\', "\xE2\x80\xBE" => '~');\r
+ case 'johab':\r
+ return array("\xE2\x82\xA9" => '\\');\r
+ }\r
+ if (strpos($lenc, 'iso-8859-') === 0) {\r
+ return array();\r
+ }\r
+ }\r
+ $ret = array();\r
+ if (self::unsafeIconv('UTF-8', $encoding, 'a') === false) {\r
+ return false;\r
+ }\r
+ for ($i = 0x20; $i <= 0x7E; $i++) { // all printable ASCII chars\r
+ $c = chr($i); // UTF-8 char\r
+ $r = self::unsafeIconv('UTF-8', "$encoding//IGNORE", $c); // initial conversion\r
+ if ($r === '' ||\r
+ // This line is needed for iconv implementations that do not\r
+ // omit characters that do not exist in the target character set\r
+ ($r === $c && self::unsafeIconv($encoding, 'UTF-8//IGNORE', $r) !== $c)\r
+ ) {\r
+ // Reverse engineer: what's the UTF-8 equiv of this byte\r
+ // sequence? This assumes that there's no variable width\r
+ // encoding that doesn't support ASCII.\r
+ $ret[self::unsafeIconv($encoding, 'UTF-8//IGNORE', $c)] = $c;\r
+ }\r
+ }\r
+ $encodings[$encoding] = $ret;\r
+ return $ret;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Object that provides entity lookup table from entity name to character\r
+ */\r
+class HTMLPurifier_EntityLookup\r
+{\r
+ /**\r
+ * Assoc array of entity name to character represented.\r
+ * @type array\r
+ */\r
+ public $table;\r
+\r
+ /**\r
+ * Sets up the entity lookup table from the serialized file contents.\r
+ * @param bool $file\r
+ * @note The serialized contents are versioned, but were generated\r
+ * using the maintenance script generate_entity_file.php\r
+ * @warning This is not in constructor to help enforce the Singleton\r
+ */\r
+ public function setup($file = false)\r
+ {\r
+ if (!$file) {\r
+ $file = HTMLPURIFIER_PREFIX . '/HTMLPurifier/EntityLookup/entities.ser';\r
+ }\r
+ $this->table = unserialize(file_get_contents($file));\r
+ }\r
+\r
+ /**\r
+ * Retrieves sole instance of the object.\r
+ * @param bool|HTMLPurifier_EntityLookup $prototype Optional prototype of custom lookup table to overload with.\r
+ * @return HTMLPurifier_EntityLookup\r
+ */\r
+ public static function instance($prototype = false)\r
+ {\r
+ // no references, since PHP doesn't copy unless modified\r
+ static $instance = null;\r
+ if ($prototype) {\r
+ $instance = $prototype;\r
+ } elseif (!$instance) {\r
+ $instance = new HTMLPurifier_EntityLookup();\r
+ $instance->setup();\r
+ }\r
+ return $instance;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+// if want to implement error collecting here, we'll need to use some sort\r
+// of global data (probably trigger_error) because it's impossible to pass\r
+// $config or $context to the callback functions.\r
+\r
+/**\r
+ * Handles referencing and derefencing character entities\r
+ */\r
+class HTMLPurifier_EntityParser\r
+{\r
+\r
+ /**\r
+ * Reference to entity lookup table.\r
+ * @type HTMLPurifier_EntityLookup\r
+ */\r
+ protected $_entity_lookup;\r
+\r
+ /**\r
+ * Callback regex string for entities in text.\r
+ * @type string\r
+ */\r
+ protected $_textEntitiesRegex;\r
+\r
+ /**\r
+ * Callback regex string for entities in attributes.\r
+ * @type string\r
+ */\r
+ protected $_attrEntitiesRegex;\r
+\r
+ /**\r
+ * Tests if the beginning of a string is a semi-optional regex\r
+ */\r
+ protected $_semiOptionalPrefixRegex;\r
+\r
+ public function __construct() {\r
+ // From\r
+ // http://stackoverflow.com/questions/15532252/why-is-reg-being-rendered-as-without-the-bounding-semicolon\r
+ $semi_optional = "quot|QUOT|lt|LT|gt|GT|amp|AMP|AElig|Aacute|Acirc|Agrave|Aring|Atilde|Auml|COPY|Ccedil|ETH|Eacute|Ecirc|Egrave|Euml|Iacute|Icirc|Igrave|Iuml|Ntilde|Oacute|Ocirc|Ograve|Oslash|Otilde|Ouml|REG|THORN|Uacute|Ucirc|Ugrave|Uuml|Yacute|aacute|acirc|acute|aelig|agrave|aring|atilde|auml|brvbar|ccedil|cedil|cent|copy|curren|deg|divide|eacute|ecirc|egrave|eth|euml|frac12|frac14|frac34|iacute|icirc|iexcl|igrave|iquest|iuml|laquo|macr|micro|middot|nbsp|not|ntilde|oacute|ocirc|ograve|ordf|ordm|oslash|otilde|ouml|para|plusmn|pound|raquo|reg|sect|shy|sup1|sup2|sup3|szlig|thorn|times|uacute|ucirc|ugrave|uml|uuml|yacute|yen|yuml";\r
+\r
+ // NB: three empty captures to put the fourth match in the right\r
+ // place\r
+ $this->_semiOptionalPrefixRegex = "/&()()()($semi_optional)/";\r
+\r
+ $this->_textEntitiesRegex =\r
+ '/&(?:'.\r
+ // hex\r
+ '[#]x([a-fA-F0-9]+);?|'.\r
+ // dec\r
+ '[#]0*(\d+);?|'.\r
+ // string (mandatory semicolon)\r
+ // NB: order matters: match semicolon preferentially\r
+ '([A-Za-z_:][A-Za-z0-9.\-_:]*);|'.\r
+ // string (optional semicolon)\r
+ "($semi_optional)".\r
+ ')/';\r
+\r
+ $this->_attrEntitiesRegex =\r
+ '/&(?:'.\r
+ // hex\r
+ '[#]x([a-fA-F0-9]+);?|'.\r
+ // dec\r
+ '[#]0*(\d+);?|'.\r
+ // string (mandatory semicolon)\r
+ // NB: order matters: match semicolon preferentially\r
+ '([A-Za-z_:][A-Za-z0-9.\-_:]*);|'.\r
+ // string (optional semicolon)\r
+ // don't match if trailing is equals or alphanumeric (URL\r
+ // like)\r
+ "($semi_optional)(?![=;A-Za-z0-9])".\r
+ ')/';\r
+\r
+ }\r
+\r
+ /**\r
+ * Substitute entities with the parsed equivalents. Use this on\r
+ * textual data in an HTML document (as opposed to attributes.)\r
+ *\r
+ * @param string $string String to have entities parsed.\r
+ * @return string Parsed string.\r
+ */\r
+ public function substituteTextEntities($string)\r
+ {\r
+ return preg_replace_callback(\r
+ $this->_textEntitiesRegex,\r
+ array($this, 'entityCallback'),\r
+ $string\r
+ );\r
+ }\r
+\r
+ /**\r
+ * Substitute entities with the parsed equivalents. Use this on\r
+ * attribute contents in documents.\r
+ *\r
+ * @param string $string String to have entities parsed.\r
+ * @return string Parsed string.\r
+ */\r
+ public function substituteAttrEntities($string)\r
+ {\r
+ return preg_replace_callback(\r
+ $this->_attrEntitiesRegex,\r
+ array($this, 'entityCallback'),\r
+ $string\r
+ );\r
+ }\r
+\r
+ /**\r
+ * Callback function for substituteNonSpecialEntities() that does the work.\r
+ *\r
+ * @param array $matches PCRE matches array, with 0 the entire match, and\r
+ * either index 1, 2 or 3 set with a hex value, dec value,\r
+ * or string (respectively).\r
+ * @return string Replacement string.\r
+ */\r
+\r
+ protected function entityCallback($matches)\r
+ {\r
+ $entity = $matches[0];\r
+ $hex_part = @$matches[1];\r
+ $dec_part = @$matches[2];\r
+ $named_part = empty($matches[3]) ? (empty($matches[4]) ? "" : $matches[4]) : $matches[3];\r
+ if ($hex_part !== NULL && $hex_part !== "") {\r
+ return HTMLPurifier_Encoder::unichr(hexdec($hex_part));\r
+ } elseif ($dec_part !== NULL && $dec_part !== "") {\r
+ return HTMLPurifier_Encoder::unichr((int) $dec_part);\r
+ } else {\r
+ if (!$this->_entity_lookup) {\r
+ $this->_entity_lookup = HTMLPurifier_EntityLookup::instance();\r
+ }\r
+ if (isset($this->_entity_lookup->table[$named_part])) {\r
+ return $this->_entity_lookup->table[$named_part];\r
+ } else {\r
+ // exact match didn't match anything, so test if\r
+ // any of the semicolon optional match the prefix.\r
+ // Test that this is an EXACT match is important to\r
+ // prevent infinite loop\r
+ if (!empty($matches[3])) {\r
+ return preg_replace_callback(\r
+ $this->_semiOptionalPrefixRegex,\r
+ array($this, 'entityCallback'),\r
+ $entity\r
+ );\r
+ }\r
+ return $entity;\r
+ }\r
+ }\r
+ }\r
+\r
+ // LEGACY CODE BELOW\r
+\r
+ /**\r
+ * Callback regex string for parsing entities.\r
+ * @type string\r
+ */\r
+ protected $_substituteEntitiesRegex =\r
+ '/&(?:[#]x([a-fA-F0-9]+)|[#]0*(\d+)|([A-Za-z_:][A-Za-z0-9.\-_:]*));?/';\r
+ // 1. hex 2. dec 3. string (XML style)\r
+\r
+ /**\r
+ * Decimal to parsed string conversion table for special entities.\r
+ * @type array\r
+ */\r
+ protected $_special_dec2str =\r
+ array(\r
+ 34 => '"',\r
+ 38 => '&',\r
+ 39 => "'",\r
+ 60 => '<',\r
+ 62 => '>'\r
+ );\r
+\r
+ /**\r
+ * Stripped entity names to decimal conversion table for special entities.\r
+ * @type array\r
+ */\r
+ protected $_special_ent2dec =\r
+ array(\r
+ 'quot' => 34,\r
+ 'amp' => 38,\r
+ 'lt' => 60,\r
+ 'gt' => 62\r
+ );\r
+\r
+ /**\r
+ * Substitutes non-special entities with their parsed equivalents. Since\r
+ * running this whenever you have parsed character is t3h 5uck, we run\r
+ * it before everything else.\r
+ *\r
+ * @param string $string String to have non-special entities parsed.\r
+ * @return string Parsed string.\r
+ */\r
+ public function substituteNonSpecialEntities($string)\r
+ {\r
+ // it will try to detect missing semicolons, but don't rely on it\r
+ return preg_replace_callback(\r
+ $this->_substituteEntitiesRegex,\r
+ array($this, 'nonSpecialEntityCallback'),\r
+ $string\r
+ );\r
+ }\r
+\r
+ /**\r
+ * Callback function for substituteNonSpecialEntities() that does the work.\r
+ *\r
+ * @param array $matches PCRE matches array, with 0 the entire match, and\r
+ * either index 1, 2 or 3 set with a hex value, dec value,\r
+ * or string (respectively).\r
+ * @return string Replacement string.\r
+ */\r
+\r
+ protected function nonSpecialEntityCallback($matches)\r
+ {\r
+ // replaces all but big five\r
+ $entity = $matches[0];\r
+ $is_num = (@$matches[0][1] === '#');\r
+ if ($is_num) {\r
+ $is_hex = (@$entity[2] === 'x');\r
+ $code = $is_hex ? hexdec($matches[1]) : (int) $matches[2];\r
+ // abort for special characters\r
+ if (isset($this->_special_dec2str[$code])) {\r
+ return $entity;\r
+ }\r
+ return HTMLPurifier_Encoder::unichr($code);\r
+ } else {\r
+ if (isset($this->_special_ent2dec[$matches[3]])) {\r
+ return $entity;\r
+ }\r
+ if (!$this->_entity_lookup) {\r
+ $this->_entity_lookup = HTMLPurifier_EntityLookup::instance();\r
+ }\r
+ if (isset($this->_entity_lookup->table[$matches[3]])) {\r
+ return $this->_entity_lookup->table[$matches[3]];\r
+ } else {\r
+ return $entity;\r
+ }\r
+ }\r
+ }\r
+\r
+ /**\r
+ * Substitutes only special entities with their parsed equivalents.\r
+ *\r
+ * @notice We try to avoid calling this function because otherwise, it\r
+ * would have to be called a lot (for every parsed section).\r
+ *\r
+ * @param string $string String to have non-special entities parsed.\r
+ * @return string Parsed string.\r
+ */\r
+ public function substituteSpecialEntities($string)\r
+ {\r
+ return preg_replace_callback(\r
+ $this->_substituteEntitiesRegex,\r
+ array($this, 'specialEntityCallback'),\r
+ $string\r
+ );\r
+ }\r
+\r
+ /**\r
+ * Callback function for substituteSpecialEntities() that does the work.\r
+ *\r
+ * This callback has same syntax as nonSpecialEntityCallback().\r
+ *\r
+ * @param array $matches PCRE-style matches array, with 0 the entire match, and\r
+ * either index 1, 2 or 3 set with a hex value, dec value,\r
+ * or string (respectively).\r
+ * @return string Replacement string.\r
+ */\r
+ protected function specialEntityCallback($matches)\r
+ {\r
+ $entity = $matches[0];\r
+ $is_num = (@$matches[0][1] === '#');\r
+ if ($is_num) {\r
+ $is_hex = (@$entity[2] === 'x');\r
+ $int = $is_hex ? hexdec($matches[1]) : (int) $matches[2];\r
+ return isset($this->_special_dec2str[$int]) ?\r
+ $this->_special_dec2str[$int] :\r
+ $entity;\r
+ } else {\r
+ return isset($this->_special_ent2dec[$matches[3]]) ?\r
+ $this->_special_dec2str[$this->_special_ent2dec[$matches[3]]] :\r
+ $entity;\r
+ }\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Error collection class that enables HTML Purifier to report HTML\r
+ * problems back to the user\r
+ */\r
+class HTMLPurifier_ErrorCollector\r
+{\r
+\r
+ /**\r
+ * Identifiers for the returned error array. These are purposely numeric\r
+ * so list() can be used.\r
+ */\r
+ const LINENO = 0;\r
+ const SEVERITY = 1;\r
+ const MESSAGE = 2;\r
+ const CHILDREN = 3;\r
+\r
+ /**\r
+ * @type array\r
+ */\r
+ protected $errors;\r
+\r
+ /**\r
+ * @type array\r
+ */\r
+ protected $_current;\r
+\r
+ /**\r
+ * @type array\r
+ */\r
+ protected $_stacks = array(array());\r
+\r
+ /**\r
+ * @type HTMLPurifier_Language\r
+ */\r
+ protected $locale;\r
+\r
+ /**\r
+ * @type HTMLPurifier_Generator\r
+ */\r
+ protected $generator;\r
+\r
+ /**\r
+ * @type HTMLPurifier_Context\r
+ */\r
+ protected $context;\r
+\r
+ /**\r
+ * @type array\r
+ */\r
+ protected $lines = array();\r
+\r
+ /**\r
+ * @param HTMLPurifier_Context $context\r
+ */\r
+ public function __construct($context)\r
+ {\r
+ $this->locale =& $context->get('Locale');\r
+ $this->context = $context;\r
+ $this->_current =& $this->_stacks[0];\r
+ $this->errors =& $this->_stacks[0];\r
+ }\r
+\r
+ /**\r
+ * Sends an error message to the collector for later use\r
+ * @param int $severity Error severity, PHP error style (don't use E_USER_)\r
+ * @param string $msg Error message text\r
+ */\r
+ public function send($severity, $msg)\r
+ {\r
+ $args = array();\r
+ if (func_num_args() > 2) {\r
+ $args = func_get_args();\r
+ array_shift($args);\r
+ unset($args[0]);\r
+ }\r
+\r
+ $token = $this->context->get('CurrentToken', true);\r
+ $line = $token ? $token->line : $this->context->get('CurrentLine', true);\r
+ $col = $token ? $token->col : $this->context->get('CurrentCol', true);\r
+ $attr = $this->context->get('CurrentAttr', true);\r
+\r
+ // perform special substitutions, also add custom parameters\r
+ $subst = array();\r
+ if (!is_null($token)) {\r
+ $args['CurrentToken'] = $token;\r
+ }\r
+ if (!is_null($attr)) {\r
+ $subst['$CurrentAttr.Name'] = $attr;\r
+ if (isset($token->attr[$attr])) {\r
+ $subst['$CurrentAttr.Value'] = $token->attr[$attr];\r
+ }\r
+ }\r
+\r
+ if (empty($args)) {\r
+ $msg = $this->locale->getMessage($msg);\r
+ } else {\r
+ $msg = $this->locale->formatMessage($msg, $args);\r
+ }\r
+\r
+ if (!empty($subst)) {\r
+ $msg = strtr($msg, $subst);\r
+ }\r
+\r
+ // (numerically indexed)\r
+ $error = array(\r
+ self::LINENO => $line,\r
+ self::SEVERITY => $severity,\r
+ self::MESSAGE => $msg,\r
+ self::CHILDREN => array()\r
+ );\r
+ $this->_current[] = $error;\r
+\r
+ // NEW CODE BELOW ...\r
+ // Top-level errors are either:\r
+ // TOKEN type, if $value is set appropriately, or\r
+ // "syntax" type, if $value is null\r
+ $new_struct = new HTMLPurifier_ErrorStruct();\r
+ $new_struct->type = HTMLPurifier_ErrorStruct::TOKEN;\r
+ if ($token) {\r
+ $new_struct->value = clone $token;\r
+ }\r
+ if (is_int($line) && is_int($col)) {\r
+ if (isset($this->lines[$line][$col])) {\r
+ $struct = $this->lines[$line][$col];\r
+ } else {\r
+ $struct = $this->lines[$line][$col] = $new_struct;\r
+ }\r
+ // These ksorts may present a performance problem\r
+ ksort($this->lines[$line], SORT_NUMERIC);\r
+ } else {\r
+ if (isset($this->lines[-1])) {\r
+ $struct = $this->lines[-1];\r
+ } else {\r
+ $struct = $this->lines[-1] = $new_struct;\r
+ }\r
+ }\r
+ ksort($this->lines, SORT_NUMERIC);\r
+\r
+ // Now, check if we need to operate on a lower structure\r
+ if (!empty($attr)) {\r
+ $struct = $struct->getChild(HTMLPurifier_ErrorStruct::ATTR, $attr);\r
+ if (!$struct->value) {\r
+ $struct->value = array($attr, 'PUT VALUE HERE');\r
+ }\r
+ }\r
+ if (!empty($cssprop)) {\r
+ $struct = $struct->getChild(HTMLPurifier_ErrorStruct::CSSPROP, $cssprop);\r
+ if (!$struct->value) {\r
+ // if we tokenize CSS this might be a little more difficult to do\r
+ $struct->value = array($cssprop, 'PUT VALUE HERE');\r
+ }\r
+ }\r
+\r
+ // Ok, structs are all setup, now time to register the error\r
+ $struct->addError($severity, $msg);\r
+ }\r
+\r
+ /**\r
+ * Retrieves raw error data for custom formatter to use\r
+ */\r
+ public function getRaw()\r
+ {\r
+ return $this->errors;\r
+ }\r
+\r
+ /**\r
+ * Default HTML formatting implementation for error messages\r
+ * @param HTMLPurifier_Config $config Configuration, vital for HTML output nature\r
+ * @param array $errors Errors array to display; used for recursion.\r
+ * @return string\r
+ */\r
+ public function getHTMLFormatted($config, $errors = null)\r
+ {\r
+ $ret = array();\r
+\r
+ $this->generator = new HTMLPurifier_Generator($config, $this->context);\r
+ if ($errors === null) {\r
+ $errors = $this->errors;\r
+ }\r
+\r
+ // 'At line' message needs to be removed\r
+\r
+ // generation code for new structure goes here. It needs to be recursive.\r
+ foreach ($this->lines as $line => $col_array) {\r
+ if ($line == -1) {\r
+ continue;\r
+ }\r
+ foreach ($col_array as $col => $struct) {\r
+ $this->_renderStruct($ret, $struct, $line, $col);\r
+ }\r
+ }\r
+ if (isset($this->lines[-1])) {\r
+ $this->_renderStruct($ret, $this->lines[-1]);\r
+ }\r
+\r
+ if (empty($errors)) {\r
+ return '<p>' . $this->locale->getMessage('ErrorCollector: No errors') . '</p>';\r
+ } else {\r
+ return '<ul><li>' . implode('</li><li>', $ret) . '</li></ul>';\r
+ }\r
+\r
+ }\r
+\r
+ private function _renderStruct(&$ret, $struct, $line = null, $col = null)\r
+ {\r
+ $stack = array($struct);\r
+ $context_stack = array(array());\r
+ while ($current = array_pop($stack)) {\r
+ $context = array_pop($context_stack);\r
+ foreach ($current->errors as $error) {\r
+ list($severity, $msg) = $error;\r
+ $string = '';\r
+ $string .= '<div>';\r
+ // W3C uses an icon to indicate the severity of the error.\r
+ $error = $this->locale->getErrorName($severity);\r
+ $string .= "<span class=\"error e$severity\"><strong>$error</strong></span> ";\r
+ if (!is_null($line) && !is_null($col)) {\r
+ $string .= "<em class=\"location\">Line $line, Column $col: </em> ";\r
+ } else {\r
+ $string .= '<em class="location">End of Document: </em> ';\r
+ }\r
+ $string .= '<strong class="description">' . $this->generator->escape($msg) . '</strong> ';\r
+ $string .= '</div>';\r
+ // Here, have a marker for the character on the column appropriate.\r
+ // Be sure to clip extremely long lines.\r
+ //$string .= '<pre>';\r
+ //$string .= '';\r
+ //$string .= '</pre>';\r
+ $ret[] = $string;\r
+ }\r
+ foreach ($current->children as $array) {\r
+ $context[] = $current;\r
+ $stack = array_merge($stack, array_reverse($array, true));\r
+ for ($i = count($array); $i > 0; $i--) {\r
+ $context_stack[] = $context;\r
+ }\r
+ }\r
+ }\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Records errors for particular segments of an HTML document such as tokens,\r
+ * attributes or CSS properties. They can contain error structs (which apply\r
+ * to components of what they represent), but their main purpose is to hold\r
+ * errors applying to whatever struct is being used.\r
+ */\r
+class HTMLPurifier_ErrorStruct\r
+{\r
+\r
+ /**\r
+ * Possible values for $children first-key. Note that top-level structures\r
+ * are automatically token-level.\r
+ */\r
+ const TOKEN = 0;\r
+ const ATTR = 1;\r
+ const CSSPROP = 2;\r
+\r
+ /**\r
+ * Type of this struct.\r
+ * @type string\r
+ */\r
+ public $type;\r
+\r
+ /**\r
+ * Value of the struct we are recording errors for. There are various\r
+ * values for this:\r
+ * - TOKEN: Instance of HTMLPurifier_Token\r
+ * - ATTR: array('attr-name', 'value')\r
+ * - CSSPROP: array('prop-name', 'value')\r
+ * @type mixed\r
+ */\r
+ public $value;\r
+\r
+ /**\r
+ * Errors registered for this structure.\r
+ * @type array\r
+ */\r
+ public $errors = array();\r
+\r
+ /**\r
+ * Child ErrorStructs that are from this structure. For example, a TOKEN\r
+ * ErrorStruct would contain ATTR ErrorStructs. This is a multi-dimensional\r
+ * array in structure: [TYPE]['identifier']\r
+ * @type array\r
+ */\r
+ public $children = array();\r
+\r
+ /**\r
+ * @param string $type\r
+ * @param string $id\r
+ * @return mixed\r
+ */\r
+ public function getChild($type, $id)\r
+ {\r
+ if (!isset($this->children[$type][$id])) {\r
+ $this->children[$type][$id] = new HTMLPurifier_ErrorStruct();\r
+ $this->children[$type][$id]->type = $type;\r
+ }\r
+ return $this->children[$type][$id];\r
+ }\r
+\r
+ /**\r
+ * @param int $severity\r
+ * @param string $message\r
+ */\r
+ public function addError($severity, $message)\r
+ {\r
+ $this->errors[] = array($severity, $message);\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Global exception class for HTML Purifier; any exceptions we throw\r
+ * are from here.\r
+ */\r
+class HTMLPurifier_Exception extends Exception\r
+{\r
+\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Represents a pre or post processing filter on HTML Purifier's output\r
+ *\r
+ * Sometimes, a little ad-hoc fixing of HTML has to be done before\r
+ * it gets sent through HTML Purifier: you can use filters to acheive\r
+ * this effect. For instance, YouTube videos can be preserved using\r
+ * this manner. You could have used a decorator for this task, but\r
+ * PHP's support for them is not terribly robust, so we're going\r
+ * to just loop through the filters.\r
+ *\r
+ * Filters should be exited first in, last out. If there are three filters,\r
+ * named 1, 2 and 3, the order of execution should go 1->preFilter,\r
+ * 2->preFilter, 3->preFilter, purify, 3->postFilter, 2->postFilter,\r
+ * 1->postFilter.\r
+ *\r
+ * @note Methods are not declared abstract as it is perfectly legitimate\r
+ * for an implementation not to want anything to happen on a step\r
+ */\r
+\r
+class HTMLPurifier_Filter\r
+{\r
+\r
+ /**\r
+ * Name of the filter for identification purposes.\r
+ * @type string\r
+ */\r
+ public $name;\r
+\r
+ /**\r
+ * Pre-processor function, handles HTML before HTML Purifier\r
+ * @param string $html\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return string\r
+ */\r
+ public function preFilter($html, $config, $context)\r
+ {\r
+ return $html;\r
+ }\r
+\r
+ /**\r
+ * Post-processor function, handles HTML after HTML Purifier\r
+ * @param string $html\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return string\r
+ */\r
+ public function postFilter($html, $config, $context)\r
+ {\r
+ return $html;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Generates HTML from tokens.\r
+ * @todo Refactor interface so that configuration/context is determined\r
+ * upon instantiation, no need for messy generateFromTokens() calls\r
+ * @todo Make some of the more internal functions protected, and have\r
+ * unit tests work around that\r
+ */\r
+class HTMLPurifier_Generator\r
+{\r
+\r
+ /**\r
+ * Whether or not generator should produce XML output.\r
+ * @type bool\r
+ */\r
+ private $_xhtml = true;\r
+\r
+ /**\r
+ * :HACK: Whether or not generator should comment the insides of <script> tags.\r
+ * @type bool\r
+ */\r
+ private $_scriptFix = false;\r
+\r
+ /**\r
+ * Cache of HTMLDefinition during HTML output to determine whether or\r
+ * not attributes should be minimized.\r
+ * @type HTMLPurifier_HTMLDefinition\r
+ */\r
+ private $_def;\r
+\r
+ /**\r
+ * Cache of %Output.SortAttr.\r
+ * @type bool\r
+ */\r
+ private $_sortAttr;\r
+\r
+ /**\r
+ * Cache of %Output.FlashCompat.\r
+ * @type bool\r
+ */\r
+ private $_flashCompat;\r
+\r
+ /**\r
+ * Cache of %Output.FixInnerHTML.\r
+ * @type bool\r
+ */\r
+ private $_innerHTMLFix;\r
+\r
+ /**\r
+ * Stack for keeping track of object information when outputting IE\r
+ * compatibility code.\r
+ * @type array\r
+ */\r
+ private $_flashStack = array();\r
+\r
+ /**\r
+ * Configuration for the generator\r
+ * @type HTMLPurifier_Config\r
+ */\r
+ protected $config;\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ */\r
+ public function __construct($config, $context)\r
+ {\r
+ $this->config = $config;\r
+ $this->_scriptFix = $config->get('Output.CommentScriptContents');\r
+ $this->_innerHTMLFix = $config->get('Output.FixInnerHTML');\r
+ $this->_sortAttr = $config->get('Output.SortAttr');\r
+ $this->_flashCompat = $config->get('Output.FlashCompat');\r
+ $this->_def = $config->getHTMLDefinition();\r
+ $this->_xhtml = $this->_def->doctype->xml;\r
+ }\r
+\r
+ /**\r
+ * Generates HTML from an array of tokens.\r
+ * @param HTMLPurifier_Token[] $tokens Array of HTMLPurifier_Token\r
+ * @return string Generated HTML\r
+ */\r
+ public function generateFromTokens($tokens)\r
+ {\r
+ if (!$tokens) {\r
+ return '';\r
+ }\r
+\r
+ // Basic algorithm\r
+ $html = '';\r
+ for ($i = 0, $size = count($tokens); $i < $size; $i++) {\r
+ if ($this->_scriptFix && $tokens[$i]->name === 'script'\r
+ && $i + 2 < $size && $tokens[$i+2] instanceof HTMLPurifier_Token_End) {\r
+ // script special case\r
+ // the contents of the script block must be ONE token\r
+ // for this to work.\r
+ $html .= $this->generateFromToken($tokens[$i++]);\r
+ $html .= $this->generateScriptFromToken($tokens[$i++]);\r
+ }\r
+ $html .= $this->generateFromToken($tokens[$i]);\r
+ }\r
+\r
+ // Tidy cleanup\r
+ if (extension_loaded('tidy') && $this->config->get('Output.TidyFormat')) {\r
+ $tidy = new Tidy;\r
+ $tidy->parseString(\r
+ $html,\r
+ array(\r
+ 'indent'=> true,\r
+ 'output-xhtml' => $this->_xhtml,\r
+ 'show-body-only' => true,\r
+ 'indent-spaces' => 2,\r
+ 'wrap' => 68,\r
+ ),\r
+ 'utf8'\r
+ );\r
+ $tidy->cleanRepair();\r
+ $html = (string) $tidy; // explicit cast necessary\r
+ }\r
+\r
+ // Normalize newlines to system defined value\r
+ if ($this->config->get('Core.NormalizeNewlines')) {\r
+ $nl = $this->config->get('Output.Newline');\r
+ if ($nl === null) {\r
+ $nl = PHP_EOL;\r
+ }\r
+ if ($nl !== "\n") {\r
+ $html = str_replace("\n", $nl, $html);\r
+ }\r
+ }\r
+ return $html;\r
+ }\r
+\r
+ /**\r
+ * Generates HTML from a single token.\r
+ * @param HTMLPurifier_Token $token HTMLPurifier_Token object.\r
+ * @return string Generated HTML\r
+ */\r
+ public function generateFromToken($token)\r
+ {\r
+ if (!$token instanceof HTMLPurifier_Token) {\r
+ trigger_error('Cannot generate HTML from non-HTMLPurifier_Token object', E_USER_WARNING);\r
+ return '';\r
+\r
+ } elseif ($token instanceof HTMLPurifier_Token_Start) {\r
+ $attr = $this->generateAttributes($token->attr, $token->name);\r
+ if ($this->_flashCompat) {\r
+ if ($token->name == "object") {\r
+ $flash = new stdClass();\r
+ $flash->attr = $token->attr;\r
+ $flash->param = array();\r
+ $this->_flashStack[] = $flash;\r
+ }\r
+ }\r
+ return '<' . $token->name . ($attr ? ' ' : '') . $attr . '>';\r
+\r
+ } elseif ($token instanceof HTMLPurifier_Token_End) {\r
+ $_extra = '';\r
+ if ($this->_flashCompat) {\r
+ if ($token->name == "object" && !empty($this->_flashStack)) {\r
+ // doesn't do anything for now\r
+ }\r
+ }\r
+ return $_extra . '</' . $token->name . '>';\r
+\r
+ } elseif ($token instanceof HTMLPurifier_Token_Empty) {\r
+ if ($this->_flashCompat && $token->name == "param" && !empty($this->_flashStack)) {\r
+ $this->_flashStack[count($this->_flashStack)-1]->param[$token->attr['name']] = $token->attr['value'];\r
+ }\r
+ $attr = $this->generateAttributes($token->attr, $token->name);\r
+ return '<' . $token->name . ($attr ? ' ' : '') . $attr .\r
+ ( $this->_xhtml ? ' /': '' ) // <br /> v. <br>\r
+ . '>';\r
+\r
+ } elseif ($token instanceof HTMLPurifier_Token_Text) {\r
+ return $this->escape($token->data, ENT_NOQUOTES);\r
+\r
+ } elseif ($token instanceof HTMLPurifier_Token_Comment) {\r
+ return '<!--' . $token->data . '-->';\r
+ } else {\r
+ return '';\r
+\r
+ }\r
+ }\r
+\r
+ /**\r
+ * Special case processor for the contents of script tags\r
+ * @param HTMLPurifier_Token $token HTMLPurifier_Token object.\r
+ * @return string\r
+ * @warning This runs into problems if there's already a literal\r
+ * --> somewhere inside the script contents.\r
+ */\r
+ public function generateScriptFromToken($token)\r
+ {\r
+ if (!$token instanceof HTMLPurifier_Token_Text) {\r
+ return $this->generateFromToken($token);\r
+ }\r
+ // Thanks <http://lachy.id.au/log/2005/05/script-comments>\r
+ $data = preg_replace('#//\s*$#', '', $token->data);\r
+ return '<!--//--><![CDATA[//><!--' . "\n" . trim($data) . "\n" . '//--><!]]>';\r
+ }\r
+\r
+ /**\r
+ * Generates attribute declarations from attribute array.\r
+ * @note This does not include the leading or trailing space.\r
+ * @param array $assoc_array_of_attributes Attribute array\r
+ * @param string $element Name of element attributes are for, used to check\r
+ * attribute minimization.\r
+ * @return string Generated HTML fragment for insertion.\r
+ */\r
+ public function generateAttributes($assoc_array_of_attributes, $element = '')\r
+ {\r
+ $html = '';\r
+ if ($this->_sortAttr) {\r
+ ksort($assoc_array_of_attributes);\r
+ }\r
+ foreach ($assoc_array_of_attributes as $key => $value) {\r
+ if (!$this->_xhtml) {\r
+ // Remove namespaced attributes\r
+ if (strpos($key, ':') !== false) {\r
+ continue;\r
+ }\r
+ // Check if we should minimize the attribute: val="val" -> val\r
+ if ($element && !empty($this->_def->info[$element]->attr[$key]->minimized)) {\r
+ $html .= $key . ' ';\r
+ continue;\r
+ }\r
+ }\r
+ // Workaround for Internet Explorer innerHTML bug.\r
+ // Essentially, Internet Explorer, when calculating\r
+ // innerHTML, omits quotes if there are no instances of\r
+ // angled brackets, quotes or spaces. However, when parsing\r
+ // HTML (for example, when you assign to innerHTML), it\r
+ // treats backticks as quotes. Thus,\r
+ // <img alt="``" />\r
+ // becomes\r
+ // <img alt=`` />\r
+ // becomes\r
+ // <img alt='' />\r
+ // Fortunately, all we need to do is trigger an appropriate\r
+ // quoting style, which we do by adding an extra space.\r
+ // This also is consistent with the W3C spec, which states\r
+ // that user agents may ignore leading or trailing\r
+ // whitespace (in fact, most don't, at least for attributes\r
+ // like alt, but an extra space at the end is barely\r
+ // noticeable). Still, we have a configuration knob for\r
+ // this, since this transformation is not necesary if you\r
+ // don't process user input with innerHTML or you don't plan\r
+ // on supporting Internet Explorer.\r
+ if ($this->_innerHTMLFix) {\r
+ if (strpos($value, '`') !== false) {\r
+ // check if correct quoting style would not already be\r
+ // triggered\r
+ if (strcspn($value, '"\' <>') === strlen($value)) {\r
+ // protect!\r
+ $value .= ' ';\r
+ }\r
+ }\r
+ }\r
+ $html .= $key.'="'.$this->escape($value).'" ';\r
+ }\r
+ return rtrim($html);\r
+ }\r
+\r
+ /**\r
+ * Escapes raw text data.\r
+ * @todo This really ought to be protected, but until we have a facility\r
+ * for properly generating HTML here w/o using tokens, it stays\r
+ * public.\r
+ * @param string $string String data to escape for HTML.\r
+ * @param int $quote Quoting style, like htmlspecialchars. ENT_NOQUOTES is\r
+ * permissible for non-attribute output.\r
+ * @return string escaped data.\r
+ */\r
+ public function escape($string, $quote = null)\r
+ {\r
+ // Workaround for APC bug on Mac Leopard reported by sidepodcast\r
+ // http://htmlpurifier.org/phorum/read.php?3,4823,4846\r
+ if ($quote === null) {\r
+ $quote = ENT_COMPAT;\r
+ }\r
+ return htmlspecialchars($string, $quote, 'UTF-8');\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Definition of the purified HTML that describes allowed children,\r
+ * attributes, and many other things.\r
+ *\r
+ * Conventions:\r
+ *\r
+ * All member variables that are prefixed with info\r
+ * (including the main $info array) are used by HTML Purifier internals\r
+ * and should not be directly edited when customizing the HTMLDefinition.\r
+ * They can usually be set via configuration directives or custom\r
+ * modules.\r
+ *\r
+ * On the other hand, member variables without the info prefix are used\r
+ * internally by the HTMLDefinition and MUST NOT be used by other HTML\r
+ * Purifier internals. Many of them, however, are public, and may be\r
+ * edited by userspace code to tweak the behavior of HTMLDefinition.\r
+ *\r
+ * @note This class is inspected by Printer_HTMLDefinition; please\r
+ * update that class if things here change.\r
+ *\r
+ * @warning Directives that change this object's structure must be in\r
+ * the HTML or Attr namespace!\r
+ */\r
+class HTMLPurifier_HTMLDefinition extends HTMLPurifier_Definition\r
+{\r
+\r
+ // FULLY-PUBLIC VARIABLES ---------------------------------------------\r
+\r
+ /**\r
+ * Associative array of element names to HTMLPurifier_ElementDef.\r
+ * @type HTMLPurifier_ElementDef[]\r
+ */\r
+ public $info = array();\r
+\r
+ /**\r
+ * Associative array of global attribute name to attribute definition.\r
+ * @type array\r
+ */\r
+ public $info_global_attr = array();\r
+\r
+ /**\r
+ * String name of parent element HTML will be going into.\r
+ * @type string\r
+ */\r
+ public $info_parent = 'div';\r
+\r
+ /**\r
+ * Definition for parent element, allows parent element to be a\r
+ * tag that's not allowed inside the HTML fragment.\r
+ * @type HTMLPurifier_ElementDef\r
+ */\r
+ public $info_parent_def;\r
+\r
+ /**\r
+ * String name of element used to wrap inline elements in block context.\r
+ * @type string\r
+ * @note This is rarely used except for BLOCKQUOTEs in strict mode\r
+ */\r
+ public $info_block_wrapper = 'p';\r
+\r
+ /**\r
+ * Associative array of deprecated tag name to HTMLPurifier_TagTransform.\r
+ * @type array\r
+ */\r
+ public $info_tag_transform = array();\r
+\r
+ /**\r
+ * Indexed list of HTMLPurifier_AttrTransform to be performed before validation.\r
+ * @type HTMLPurifier_AttrTransform[]\r
+ */\r
+ public $info_attr_transform_pre = array();\r
+\r
+ /**\r
+ * Indexed list of HTMLPurifier_AttrTransform to be performed after validation.\r
+ * @type HTMLPurifier_AttrTransform[]\r
+ */\r
+ public $info_attr_transform_post = array();\r
+\r
+ /**\r
+ * Nested lookup array of content set name (Block, Inline) to\r
+ * element name to whether or not it belongs in that content set.\r
+ * @type array\r
+ */\r
+ public $info_content_sets = array();\r
+\r
+ /**\r
+ * Indexed list of HTMLPurifier_Injector to be used.\r
+ * @type HTMLPurifier_Injector[]\r
+ */\r
+ public $info_injector = array();\r
+\r
+ /**\r
+ * Doctype object\r
+ * @type HTMLPurifier_Doctype\r
+ */\r
+ public $doctype;\r
+\r
+\r
+\r
+ // RAW CUSTOMIZATION STUFF --------------------------------------------\r
+\r
+ /**\r
+ * Adds a custom attribute to a pre-existing element\r
+ * @note This is strictly convenience, and does not have a corresponding\r
+ * method in HTMLPurifier_HTMLModule\r
+ * @param string $element_name Element name to add attribute to\r
+ * @param string $attr_name Name of attribute\r
+ * @param mixed $def Attribute definition, can be string or object, see\r
+ * HTMLPurifier_AttrTypes for details\r
+ */\r
+ public function addAttribute($element_name, $attr_name, $def)\r
+ {\r
+ $module = $this->getAnonymousModule();\r
+ if (!isset($module->info[$element_name])) {\r
+ $element = $module->addBlankElement($element_name);\r
+ } else {\r
+ $element = $module->info[$element_name];\r
+ }\r
+ $element->attr[$attr_name] = $def;\r
+ }\r
+\r
+ /**\r
+ * Adds a custom element to your HTML definition\r
+ * @see HTMLPurifier_HTMLModule::addElement() for detailed\r
+ * parameter and return value descriptions.\r
+ */\r
+ public function addElement($element_name, $type, $contents, $attr_collections, $attributes = array())\r
+ {\r
+ $module = $this->getAnonymousModule();\r
+ // assume that if the user is calling this, the element\r
+ // is safe. This may not be a good idea\r
+ $element = $module->addElement($element_name, $type, $contents, $attr_collections, $attributes);\r
+ return $element;\r
+ }\r
+\r
+ /**\r
+ * Adds a blank element to your HTML definition, for overriding\r
+ * existing behavior\r
+ * @param string $element_name\r
+ * @return HTMLPurifier_ElementDef\r
+ * @see HTMLPurifier_HTMLModule::addBlankElement() for detailed\r
+ * parameter and return value descriptions.\r
+ */\r
+ public function addBlankElement($element_name)\r
+ {\r
+ $module = $this->getAnonymousModule();\r
+ $element = $module->addBlankElement($element_name);\r
+ return $element;\r
+ }\r
+\r
+ /**\r
+ * Retrieves a reference to the anonymous module, so you can\r
+ * bust out advanced features without having to make your own\r
+ * module.\r
+ * @return HTMLPurifier_HTMLModule\r
+ */\r
+ public function getAnonymousModule()\r
+ {\r
+ if (!$this->_anonModule) {\r
+ $this->_anonModule = new HTMLPurifier_HTMLModule();\r
+ $this->_anonModule->name = 'Anonymous';\r
+ }\r
+ return $this->_anonModule;\r
+ }\r
+\r
+ private $_anonModule = null;\r
+\r
+ // PUBLIC BUT INTERNAL VARIABLES --------------------------------------\r
+\r
+ /**\r
+ * @type string\r
+ */\r
+ public $type = 'HTML';\r
+\r
+ /**\r
+ * @type HTMLPurifier_HTMLModuleManager\r
+ */\r
+ public $manager;\r
+\r
+ /**\r
+ * Performs low-cost, preliminary initialization.\r
+ */\r
+ public function __construct()\r
+ {\r
+ $this->manager = new HTMLPurifier_HTMLModuleManager();\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ protected function doSetup($config)\r
+ {\r
+ $this->processModules($config);\r
+ $this->setupConfigStuff($config);\r
+ unset($this->manager);\r
+\r
+ // cleanup some of the element definitions\r
+ foreach ($this->info as $k => $v) {\r
+ unset($this->info[$k]->content_model);\r
+ unset($this->info[$k]->content_model_type);\r
+ }\r
+ }\r
+\r
+ /**\r
+ * Extract out the information from the manager\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ protected function processModules($config)\r
+ {\r
+ if ($this->_anonModule) {\r
+ // for user specific changes\r
+ // this is late-loaded so we don't have to deal with PHP4\r
+ // reference wonky-ness\r
+ $this->manager->addModule($this->_anonModule);\r
+ unset($this->_anonModule);\r
+ }\r
+\r
+ $this->manager->setup($config);\r
+ $this->doctype = $this->manager->doctype;\r
+\r
+ foreach ($this->manager->modules as $module) {\r
+ foreach ($module->info_tag_transform as $k => $v) {\r
+ if ($v === false) {\r
+ unset($this->info_tag_transform[$k]);\r
+ } else {\r
+ $this->info_tag_transform[$k] = $v;\r
+ }\r
+ }\r
+ foreach ($module->info_attr_transform_pre as $k => $v) {\r
+ if ($v === false) {\r
+ unset($this->info_attr_transform_pre[$k]);\r
+ } else {\r
+ $this->info_attr_transform_pre[$k] = $v;\r
+ }\r
+ }\r
+ foreach ($module->info_attr_transform_post as $k => $v) {\r
+ if ($v === false) {\r
+ unset($this->info_attr_transform_post[$k]);\r
+ } else {\r
+ $this->info_attr_transform_post[$k] = $v;\r
+ }\r
+ }\r
+ foreach ($module->info_injector as $k => $v) {\r
+ if ($v === false) {\r
+ unset($this->info_injector[$k]);\r
+ } else {\r
+ $this->info_injector[$k] = $v;\r
+ }\r
+ }\r
+ }\r
+ $this->info = $this->manager->getElements();\r
+ $this->info_content_sets = $this->manager->contentSets->lookup;\r
+ }\r
+\r
+ /**\r
+ * Sets up stuff based on config. We need a better way of doing this.\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ protected function setupConfigStuff($config)\r
+ {\r
+ $block_wrapper = $config->get('HTML.BlockWrapper');\r
+ if (isset($this->info_content_sets['Block'][$block_wrapper])) {\r
+ $this->info_block_wrapper = $block_wrapper;\r
+ } else {\r
+ trigger_error(\r
+ 'Cannot use non-block element as block wrapper',\r
+ E_USER_ERROR\r
+ );\r
+ }\r
+\r
+ $parent = $config->get('HTML.Parent');\r
+ $def = $this->manager->getElement($parent, true);\r
+ if ($def) {\r
+ $this->info_parent = $parent;\r
+ $this->info_parent_def = $def;\r
+ } else {\r
+ trigger_error(\r
+ 'Cannot use unrecognized element as parent',\r
+ E_USER_ERROR\r
+ );\r
+ $this->info_parent_def = $this->manager->getElement($this->info_parent, true);\r
+ }\r
+\r
+ // support template text\r
+ $support = "(for information on implementing this, see the support forums) ";\r
+\r
+ // setup allowed elements -----------------------------------------\r
+\r
+ $allowed_elements = $config->get('HTML.AllowedElements');\r
+ $allowed_attributes = $config->get('HTML.AllowedAttributes'); // retrieve early\r
+\r
+ if (!is_array($allowed_elements) && !is_array($allowed_attributes)) {\r
+ $allowed = $config->get('HTML.Allowed');\r
+ if (is_string($allowed)) {\r
+ list($allowed_elements, $allowed_attributes) = $this->parseTinyMCEAllowedList($allowed);\r
+ }\r
+ }\r
+\r
+ if (is_array($allowed_elements)) {\r
+ foreach ($this->info as $name => $d) {\r
+ if (!isset($allowed_elements[$name])) {\r
+ unset($this->info[$name]);\r
+ }\r
+ unset($allowed_elements[$name]);\r
+ }\r
+ // emit errors\r
+ foreach ($allowed_elements as $element => $d) {\r
+ $element = htmlspecialchars($element); // PHP doesn't escape errors, be careful!\r
+ trigger_error("Element '$element' is not supported $support", E_USER_WARNING);\r
+ }\r
+ }\r
+\r
+ // setup allowed attributes ---------------------------------------\r
+\r
+ $allowed_attributes_mutable = $allowed_attributes; // by copy!\r
+ if (is_array($allowed_attributes)) {\r
+ // This actually doesn't do anything, since we went away from\r
+ // global attributes. It's possible that userland code uses\r
+ // it, but HTMLModuleManager doesn't!\r
+ foreach ($this->info_global_attr as $attr => $x) {\r
+ $keys = array($attr, "*@$attr", "*.$attr");\r
+ $delete = true;\r
+ foreach ($keys as $key) {\r
+ if ($delete && isset($allowed_attributes[$key])) {\r
+ $delete = false;\r
+ }\r
+ if (isset($allowed_attributes_mutable[$key])) {\r
+ unset($allowed_attributes_mutable[$key]);\r
+ }\r
+ }\r
+ if ($delete) {\r
+ unset($this->info_global_attr[$attr]);\r
+ }\r
+ }\r
+\r
+ foreach ($this->info as $tag => $info) {\r
+ foreach ($info->attr as $attr => $x) {\r
+ $keys = array("$tag@$attr", $attr, "*@$attr", "$tag.$attr", "*.$attr");\r
+ $delete = true;\r
+ foreach ($keys as $key) {\r
+ if ($delete && isset($allowed_attributes[$key])) {\r
+ $delete = false;\r
+ }\r
+ if (isset($allowed_attributes_mutable[$key])) {\r
+ unset($allowed_attributes_mutable[$key]);\r
+ }\r
+ }\r
+ if ($delete) {\r
+ if ($this->info[$tag]->attr[$attr]->required) {\r
+ trigger_error(\r
+ "Required attribute '$attr' in element '$tag' " .\r
+ "was not allowed, which means '$tag' will not be allowed either",\r
+ E_USER_WARNING\r
+ );\r
+ }\r
+ unset($this->info[$tag]->attr[$attr]);\r
+ }\r
+ }\r
+ }\r
+ // emit errors\r
+ foreach ($allowed_attributes_mutable as $elattr => $d) {\r
+ $bits = preg_split('/[.@]/', $elattr, 2);\r
+ $c = count($bits);\r
+ switch ($c) {\r
+ case 2:\r
+ if ($bits[0] !== '*') {\r
+ $element = htmlspecialchars($bits[0]);\r
+ $attribute = htmlspecialchars($bits[1]);\r
+ if (!isset($this->info[$element])) {\r
+ trigger_error(\r
+ "Cannot allow attribute '$attribute' if element " .\r
+ "'$element' is not allowed/supported $support"\r
+ );\r
+ } else {\r
+ trigger_error(\r
+ "Attribute '$attribute' in element '$element' not supported $support",\r
+ E_USER_WARNING\r
+ );\r
+ }\r
+ break;\r
+ }\r
+ // otherwise fall through\r
+ case 1:\r
+ $attribute = htmlspecialchars($bits[0]);\r
+ trigger_error(\r
+ "Global attribute '$attribute' is not ".\r
+ "supported in any elements $support",\r
+ E_USER_WARNING\r
+ );\r
+ break;\r
+ }\r
+ }\r
+ }\r
+\r
+ // setup forbidden elements ---------------------------------------\r
+\r
+ $forbidden_elements = $config->get('HTML.ForbiddenElements');\r
+ $forbidden_attributes = $config->get('HTML.ForbiddenAttributes');\r
+\r
+ foreach ($this->info as $tag => $info) {\r
+ if (isset($forbidden_elements[$tag])) {\r
+ unset($this->info[$tag]);\r
+ continue;\r
+ }\r
+ foreach ($info->attr as $attr => $x) {\r
+ if (isset($forbidden_attributes["$tag@$attr"]) ||\r
+ isset($forbidden_attributes["*@$attr"]) ||\r
+ isset($forbidden_attributes[$attr])\r
+ ) {\r
+ unset($this->info[$tag]->attr[$attr]);\r
+ continue;\r
+ } elseif (isset($forbidden_attributes["$tag.$attr"])) { // this segment might get removed eventually\r
+ // $tag.$attr are not user supplied, so no worries!\r
+ trigger_error(\r
+ "Error with $tag.$attr: tag.attr syntax not supported for " .\r
+ "HTML.ForbiddenAttributes; use tag@attr instead",\r
+ E_USER_WARNING\r
+ );\r
+ }\r
+ }\r
+ }\r
+ foreach ($forbidden_attributes as $key => $v) {\r
+ if (strlen($key) < 2) {\r
+ continue;\r
+ }\r
+ if ($key[0] != '*') {\r
+ continue;\r
+ }\r
+ if ($key[1] == '.') {\r
+ trigger_error(\r
+ "Error with $key: *.attr syntax not supported for HTML.ForbiddenAttributes; use attr instead",\r
+ E_USER_WARNING\r
+ );\r
+ }\r
+ }\r
+\r
+ // setup injectors -----------------------------------------------------\r
+ foreach ($this->info_injector as $i => $injector) {\r
+ if ($injector->checkNeeded($config) !== false) {\r
+ // remove injector that does not have it's required\r
+ // elements/attributes present, and is thus not needed.\r
+ unset($this->info_injector[$i]);\r
+ }\r
+ }\r
+ }\r
+\r
+ /**\r
+ * Parses a TinyMCE-flavored Allowed Elements and Attributes list into\r
+ * separate lists for processing. Format is element[attr1|attr2],element2...\r
+ * @warning Although it's largely drawn from TinyMCE's implementation,\r
+ * it is different, and you'll probably have to modify your lists\r
+ * @param array $list String list to parse\r
+ * @return array\r
+ * @todo Give this its own class, probably static interface\r
+ */\r
+ public function parseTinyMCEAllowedList($list)\r
+ {\r
+ $list = str_replace(array(' ', "\t"), '', $list);\r
+\r
+ $elements = array();\r
+ $attributes = array();\r
+\r
+ $chunks = preg_split('/(,|[\n\r]+)/', $list);\r
+ foreach ($chunks as $chunk) {\r
+ if (empty($chunk)) {\r
+ continue;\r
+ }\r
+ // remove TinyMCE element control characters\r
+ if (!strpos($chunk, '[')) {\r
+ $element = $chunk;\r
+ $attr = false;\r
+ } else {\r
+ list($element, $attr) = explode('[', $chunk);\r
+ }\r
+ if ($element !== '*') {\r
+ $elements[$element] = true;\r
+ }\r
+ if (!$attr) {\r
+ continue;\r
+ }\r
+ $attr = substr($attr, 0, strlen($attr) - 1); // remove trailing ]\r
+ $attr = explode('|', $attr);\r
+ foreach ($attr as $key) {\r
+ $attributes["$element.$key"] = true;\r
+ }\r
+ }\r
+ return array($elements, $attributes);\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Represents an XHTML 1.1 module, with information on elements, tags\r
+ * and attributes.\r
+ * @note Even though this is technically XHTML 1.1, it is also used for\r
+ * regular HTML parsing. We are using modulization as a convenient\r
+ * way to represent the internals of HTMLDefinition, and our\r
+ * implementation is by no means conforming and does not directly\r
+ * use the normative DTDs or XML schemas.\r
+ * @note The public variables in a module should almost directly\r
+ * correspond to the variables in HTMLPurifier_HTMLDefinition.\r
+ * However, the prefix info carries no special meaning in these\r
+ * objects (include it anyway if that's the correspondence though).\r
+ * @todo Consider making some member functions protected\r
+ */\r
+\r
+class HTMLPurifier_HTMLModule\r
+{\r
+\r
+ // -- Overloadable ----------------------------------------------------\r
+\r
+ /**\r
+ * Short unique string identifier of the module.\r
+ * @type string\r
+ */\r
+ public $name;\r
+\r
+ /**\r
+ * Informally, a list of elements this module changes.\r
+ * Not used in any significant way.\r
+ * @type array\r
+ */\r
+ public $elements = array();\r
+\r
+ /**\r
+ * Associative array of element names to element definitions.\r
+ * Some definitions may be incomplete, to be merged in later\r
+ * with the full definition.\r
+ * @type array\r
+ */\r
+ public $info = array();\r
+\r
+ /**\r
+ * Associative array of content set names to content set additions.\r
+ * This is commonly used to, say, add an A element to the Inline\r
+ * content set. This corresponds to an internal variable $content_sets\r
+ * and NOT info_content_sets member variable of HTMLDefinition.\r
+ * @type array\r
+ */\r
+ public $content_sets = array();\r
+\r
+ /**\r
+ * Associative array of attribute collection names to attribute\r
+ * collection additions. More rarely used for adding attributes to\r
+ * the global collections. Example is the StyleAttribute module adding\r
+ * the style attribute to the Core. Corresponds to HTMLDefinition's\r
+ * attr_collections->info, since the object's data is only info,\r
+ * with extra behavior associated with it.\r
+ * @type array\r
+ */\r
+ public $attr_collections = array();\r
+\r
+ /**\r
+ * Associative array of deprecated tag name to HTMLPurifier_TagTransform.\r
+ * @type array\r
+ */\r
+ public $info_tag_transform = array();\r
+\r
+ /**\r
+ * List of HTMLPurifier_AttrTransform to be performed before validation.\r
+ * @type array\r
+ */\r
+ public $info_attr_transform_pre = array();\r
+\r
+ /**\r
+ * List of HTMLPurifier_AttrTransform to be performed after validation.\r
+ * @type array\r
+ */\r
+ public $info_attr_transform_post = array();\r
+\r
+ /**\r
+ * List of HTMLPurifier_Injector to be performed during well-formedness fixing.\r
+ * An injector will only be invoked if all of it's pre-requisites are met;\r
+ * if an injector fails setup, there will be no error; it will simply be\r
+ * silently disabled.\r
+ * @type array\r
+ */\r
+ public $info_injector = array();\r
+\r
+ /**\r
+ * Boolean flag that indicates whether or not getChildDef is implemented.\r
+ * For optimization reasons: may save a call to a function. Be sure\r
+ * to set it if you do implement getChildDef(), otherwise it will have\r
+ * no effect!\r
+ * @type bool\r
+ */\r
+ public $defines_child_def = false;\r
+\r
+ /**\r
+ * Boolean flag whether or not this module is safe. If it is not safe, all\r
+ * of its members are unsafe. Modules are safe by default (this might be\r
+ * slightly dangerous, but it doesn't make much sense to force HTML Purifier,\r
+ * which is based off of safe HTML, to explicitly say, "This is safe," even\r
+ * though there are modules which are "unsafe")\r
+ *\r
+ * @type bool\r
+ * @note Previously, safety could be applied at an element level granularity.\r
+ * We've removed this ability, so in order to add "unsafe" elements\r
+ * or attributes, a dedicated module with this property set to false\r
+ * must be used.\r
+ */\r
+ public $safe = true;\r
+\r
+ /**\r
+ * Retrieves a proper HTMLPurifier_ChildDef subclass based on\r
+ * content_model and content_model_type member variables of\r
+ * the HTMLPurifier_ElementDef class. There is a similar function\r
+ * in HTMLPurifier_HTMLDefinition.\r
+ * @param HTMLPurifier_ElementDef $def\r
+ * @return HTMLPurifier_ChildDef subclass\r
+ */\r
+ public function getChildDef($def)\r
+ {\r
+ return false;\r
+ }\r
+\r
+ // -- Convenience -----------------------------------------------------\r
+\r
+ /**\r
+ * Convenience function that sets up a new element\r
+ * @param string $element Name of element to add\r
+ * @param string|bool $type What content set should element be registered to?\r
+ * Set as false to skip this step.\r
+ * @param string $contents Allowed children in form of:\r
+ * "$content_model_type: $content_model"\r
+ * @param array $attr_includes What attribute collections to register to\r
+ * element?\r
+ * @param array $attr What unique attributes does the element define?\r
+ * @see HTMLPurifier_ElementDef:: for in-depth descriptions of these parameters.\r
+ * @return HTMLPurifier_ElementDef Created element definition object, so you\r
+ * can set advanced parameters\r
+ */\r
+ public function addElement($element, $type, $contents, $attr_includes = array(), $attr = array())\r
+ {\r
+ $this->elements[] = $element;\r
+ // parse content_model\r
+ list($content_model_type, $content_model) = $this->parseContents($contents);\r
+ // merge in attribute inclusions\r
+ $this->mergeInAttrIncludes($attr, $attr_includes);\r
+ // add element to content sets\r
+ if ($type) {\r
+ $this->addElementToContentSet($element, $type);\r
+ }\r
+ // create element\r
+ $this->info[$element] = HTMLPurifier_ElementDef::create(\r
+ $content_model,\r
+ $content_model_type,\r
+ $attr\r
+ );\r
+ // literal object $contents means direct child manipulation\r
+ if (!is_string($contents)) {\r
+ $this->info[$element]->child = $contents;\r
+ }\r
+ return $this->info[$element];\r
+ }\r
+\r
+ /**\r
+ * Convenience function that creates a totally blank, non-standalone\r
+ * element.\r
+ * @param string $element Name of element to create\r
+ * @return HTMLPurifier_ElementDef Created element\r
+ */\r
+ public function addBlankElement($element)\r
+ {\r
+ if (!isset($this->info[$element])) {\r
+ $this->elements[] = $element;\r
+ $this->info[$element] = new HTMLPurifier_ElementDef();\r
+ $this->info[$element]->standalone = false;\r
+ } else {\r
+ trigger_error("Definition for $element already exists in module, cannot redefine");\r
+ }\r
+ return $this->info[$element];\r
+ }\r
+\r
+ /**\r
+ * Convenience function that registers an element to a content set\r
+ * @param string $element Element to register\r
+ * @param string $type Name content set (warning: case sensitive, usually upper-case\r
+ * first letter)\r
+ */\r
+ public function addElementToContentSet($element, $type)\r
+ {\r
+ if (!isset($this->content_sets[$type])) {\r
+ $this->content_sets[$type] = '';\r
+ } else {\r
+ $this->content_sets[$type] .= ' | ';\r
+ }\r
+ $this->content_sets[$type] .= $element;\r
+ }\r
+\r
+ /**\r
+ * Convenience function that transforms single-string contents\r
+ * into separate content model and content model type\r
+ * @param string $contents Allowed children in form of:\r
+ * "$content_model_type: $content_model"\r
+ * @return array\r
+ * @note If contents is an object, an array of two nulls will be\r
+ * returned, and the callee needs to take the original $contents\r
+ * and use it directly.\r
+ */\r
+ public function parseContents($contents)\r
+ {\r
+ if (!is_string($contents)) {\r
+ return array(null, null);\r
+ } // defer\r
+ switch ($contents) {\r
+ // check for shorthand content model forms\r
+ case 'Empty':\r
+ return array('empty', '');\r
+ case 'Inline':\r
+ return array('optional', 'Inline | #PCDATA');\r
+ case 'Flow':\r
+ return array('optional', 'Flow | #PCDATA');\r
+ }\r
+ list($content_model_type, $content_model) = explode(':', $contents);\r
+ $content_model_type = strtolower(trim($content_model_type));\r
+ $content_model = trim($content_model);\r
+ return array($content_model_type, $content_model);\r
+ }\r
+\r
+ /**\r
+ * Convenience function that merges a list of attribute includes into\r
+ * an attribute array.\r
+ * @param array $attr Reference to attr array to modify\r
+ * @param array $attr_includes Array of includes / string include to merge in\r
+ */\r
+ public function mergeInAttrIncludes(&$attr, $attr_includes)\r
+ {\r
+ if (!is_array($attr_includes)) {\r
+ if (empty($attr_includes)) {\r
+ $attr_includes = array();\r
+ } else {\r
+ $attr_includes = array($attr_includes);\r
+ }\r
+ }\r
+ $attr[0] = $attr_includes;\r
+ }\r
+\r
+ /**\r
+ * Convenience function that generates a lookup table with boolean\r
+ * true as value.\r
+ * @param string $list List of values to turn into a lookup\r
+ * @note You can also pass an arbitrary number of arguments in\r
+ * place of the regular argument\r
+ * @return array array equivalent of list\r
+ */\r
+ public function makeLookup($list)\r
+ {\r
+ if (is_string($list)) {\r
+ $list = func_get_args();\r
+ }\r
+ $ret = array();\r
+ foreach ($list as $value) {\r
+ if (is_null($value)) {\r
+ continue;\r
+ }\r
+ $ret[$value] = true;\r
+ }\r
+ return $ret;\r
+ }\r
+\r
+ /**\r
+ * Lazy load construction of the module after determining whether\r
+ * or not it's needed, and also when a finalized configuration object\r
+ * is available.\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ public function setup($config)\r
+ {\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+class HTMLPurifier_HTMLModuleManager\r
+{\r
+\r
+ /**\r
+ * @type HTMLPurifier_DoctypeRegistry\r
+ */\r
+ public $doctypes;\r
+\r
+ /**\r
+ * Instance of current doctype.\r
+ * @type string\r
+ */\r
+ public $doctype;\r
+\r
+ /**\r
+ * @type HTMLPurifier_AttrTypes\r
+ */\r
+ public $attrTypes;\r
+\r
+ /**\r
+ * Active instances of modules for the specified doctype are\r
+ * indexed, by name, in this array.\r
+ * @type HTMLPurifier_HTMLModule[]\r
+ */\r
+ public $modules = array();\r
+\r
+ /**\r
+ * Array of recognized HTMLPurifier_HTMLModule instances,\r
+ * indexed by module's class name. This array is usually lazy loaded, but a\r
+ * user can overload a module by pre-emptively registering it.\r
+ * @type HTMLPurifier_HTMLModule[]\r
+ */\r
+ public $registeredModules = array();\r
+\r
+ /**\r
+ * List of extra modules that were added by the user\r
+ * using addModule(). These get unconditionally merged into the current doctype, whatever\r
+ * it may be.\r
+ * @type HTMLPurifier_HTMLModule[]\r
+ */\r
+ public $userModules = array();\r
+\r
+ /**\r
+ * Associative array of element name to list of modules that have\r
+ * definitions for the element; this array is dynamically filled.\r
+ * @type array\r
+ */\r
+ public $elementLookup = array();\r
+\r
+ /**\r
+ * List of prefixes we should use for registering small names.\r
+ * @type array\r
+ */\r
+ public $prefixes = array('HTMLPurifier_HTMLModule_');\r
+\r
+ /**\r
+ * @type HTMLPurifier_ContentSets\r
+ */\r
+ public $contentSets;\r
+\r
+ /**\r
+ * @type HTMLPurifier_AttrCollections\r
+ */\r
+ public $attrCollections;\r
+\r
+ /**\r
+ * If set to true, unsafe elements and attributes will be allowed.\r
+ * @type bool\r
+ */\r
+ public $trusted = false;\r
+\r
+ public function __construct()\r
+ {\r
+ // editable internal objects\r
+ $this->attrTypes = new HTMLPurifier_AttrTypes();\r
+ $this->doctypes = new HTMLPurifier_DoctypeRegistry();\r
+\r
+ // setup basic modules\r
+ $common = array(\r
+ 'CommonAttributes', 'Text', 'Hypertext', 'List',\r
+ 'Presentation', 'Edit', 'Bdo', 'Tables', 'Image',\r
+ 'StyleAttribute',\r
+ // Unsafe:\r
+ 'Scripting', 'Object', 'Forms',\r
+ // Sorta legacy, but present in strict:\r
+ 'Name',\r
+ );\r
+ $transitional = array('Legacy', 'Target', 'Iframe');\r
+ $xml = array('XMLCommonAttributes');\r
+ $non_xml = array('NonXMLCommonAttributes');\r
+\r
+ // setup basic doctypes\r
+ $this->doctypes->register(\r
+ 'HTML 4.01 Transitional',\r
+ false,\r
+ array_merge($common, $transitional, $non_xml),\r
+ array('Tidy_Transitional', 'Tidy_Proprietary'),\r
+ array(),\r
+ '-//W3C//DTD HTML 4.01 Transitional//EN',\r
+ 'http://www.w3.org/TR/html4/loose.dtd'\r
+ );\r
+\r
+ $this->doctypes->register(\r
+ 'HTML 4.01 Strict',\r
+ false,\r
+ array_merge($common, $non_xml),\r
+ array('Tidy_Strict', 'Tidy_Proprietary', 'Tidy_Name'),\r
+ array(),\r
+ '-//W3C//DTD HTML 4.01//EN',\r
+ 'http://www.w3.org/TR/html4/strict.dtd'\r
+ );\r
+\r
+ $this->doctypes->register(\r
+ 'XHTML 1.0 Transitional',\r
+ true,\r
+ array_merge($common, $transitional, $xml, $non_xml),\r
+ array('Tidy_Transitional', 'Tidy_XHTML', 'Tidy_Proprietary', 'Tidy_Name'),\r
+ array(),\r
+ '-//W3C//DTD XHTML 1.0 Transitional//EN',\r
+ 'http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd'\r
+ );\r
+\r
+ $this->doctypes->register(\r
+ 'XHTML 1.0 Strict',\r
+ true,\r
+ array_merge($common, $xml, $non_xml),\r
+ array('Tidy_Strict', 'Tidy_XHTML', 'Tidy_Strict', 'Tidy_Proprietary', 'Tidy_Name'),\r
+ array(),\r
+ '-//W3C//DTD XHTML 1.0 Strict//EN',\r
+ 'http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd'\r
+ );\r
+\r
+ $this->doctypes->register(\r
+ 'XHTML 1.1',\r
+ true,\r
+ // Iframe is a real XHTML 1.1 module, despite being\r
+ // "transitional"!\r
+ array_merge($common, $xml, array('Ruby', 'Iframe')),\r
+ array('Tidy_Strict', 'Tidy_XHTML', 'Tidy_Proprietary', 'Tidy_Strict', 'Tidy_Name'), // Tidy_XHTML1_1\r
+ array(),\r
+ '-//W3C//DTD XHTML 1.1//EN',\r
+ 'http://www.w3.org/TR/xhtml11/DTD/xhtml11.dtd'\r
+ );\r
+\r
+ }\r
+\r
+ /**\r
+ * Registers a module to the recognized module list, useful for\r
+ * overloading pre-existing modules.\r
+ * @param $module Mixed: string module name, with or without\r
+ * HTMLPurifier_HTMLModule prefix, or instance of\r
+ * subclass of HTMLPurifier_HTMLModule.\r
+ * @param $overload Boolean whether or not to overload previous modules.\r
+ * If this is not set, and you do overload a module,\r
+ * HTML Purifier will complain with a warning.\r
+ * @note This function will not call autoload, you must instantiate\r
+ * (and thus invoke) autoload outside the method.\r
+ * @note If a string is passed as a module name, different variants\r
+ * will be tested in this order:\r
+ * - Check for HTMLPurifier_HTMLModule_$name\r
+ * - Check all prefixes with $name in order they were added\r
+ * - Check for literal object name\r
+ * - Throw fatal error\r
+ * If your object name collides with an internal class, specify\r
+ * your module manually. All modules must have been included\r
+ * externally: registerModule will not perform inclusions for you!\r
+ */\r
+ public function registerModule($module, $overload = false)\r
+ {\r
+ if (is_string($module)) {\r
+ // attempt to load the module\r
+ $original_module = $module;\r
+ $ok = false;\r
+ foreach ($this->prefixes as $prefix) {\r
+ $module = $prefix . $original_module;\r
+ if (class_exists($module)) {\r
+ $ok = true;\r
+ break;\r
+ }\r
+ }\r
+ if (!$ok) {\r
+ $module = $original_module;\r
+ if (!class_exists($module)) {\r
+ trigger_error(\r
+ $original_module . ' module does not exist',\r
+ E_USER_ERROR\r
+ );\r
+ return;\r
+ }\r
+ }\r
+ $module = new $module();\r
+ }\r
+ if (empty($module->name)) {\r
+ trigger_error('Module instance of ' . get_class($module) . ' must have name');\r
+ return;\r
+ }\r
+ if (!$overload && isset($this->registeredModules[$module->name])) {\r
+ trigger_error('Overloading ' . $module->name . ' without explicit overload parameter', E_USER_WARNING);\r
+ }\r
+ $this->registeredModules[$module->name] = $module;\r
+ }\r
+\r
+ /**\r
+ * Adds a module to the current doctype by first registering it,\r
+ * and then tacking it on to the active doctype\r
+ */\r
+ public function addModule($module)\r
+ {\r
+ $this->registerModule($module);\r
+ if (is_object($module)) {\r
+ $module = $module->name;\r
+ }\r
+ $this->userModules[] = $module;\r
+ }\r
+\r
+ /**\r
+ * Adds a class prefix that registerModule() will use to resolve a\r
+ * string name to a concrete class\r
+ */\r
+ public function addPrefix($prefix)\r
+ {\r
+ $this->prefixes[] = $prefix;\r
+ }\r
+\r
+ /**\r
+ * Performs processing on modules, after being called you may\r
+ * use getElement() and getElements()\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ public function setup($config)\r
+ {\r
+ $this->trusted = $config->get('HTML.Trusted');\r
+\r
+ // generate\r
+ $this->doctype = $this->doctypes->make($config);\r
+ $modules = $this->doctype->modules;\r
+\r
+ // take out the default modules that aren't allowed\r
+ $lookup = $config->get('HTML.AllowedModules');\r
+ $special_cases = $config->get('HTML.CoreModules');\r
+\r
+ if (is_array($lookup)) {\r
+ foreach ($modules as $k => $m) {\r
+ if (isset($special_cases[$m])) {\r
+ continue;\r
+ }\r
+ if (!isset($lookup[$m])) {\r
+ unset($modules[$k]);\r
+ }\r
+ }\r
+ }\r
+\r
+ // custom modules\r
+ if ($config->get('HTML.Proprietary')) {\r
+ $modules[] = 'Proprietary';\r
+ }\r
+ if ($config->get('HTML.SafeObject')) {\r
+ $modules[] = 'SafeObject';\r
+ }\r
+ if ($config->get('HTML.SafeEmbed')) {\r
+ $modules[] = 'SafeEmbed';\r
+ }\r
+ if ($config->get('HTML.SafeScripting') !== array()) {\r
+ $modules[] = 'SafeScripting';\r
+ }\r
+ if ($config->get('HTML.Nofollow')) {\r
+ $modules[] = 'Nofollow';\r
+ }\r
+ if ($config->get('HTML.TargetBlank')) {\r
+ $modules[] = 'TargetBlank';\r
+ }\r
+ // NB: HTML.TargetNoreferrer and HTML.TargetNoopener must be AFTER HTML.TargetBlank\r
+ // so that its post-attr-transform gets run afterwards.\r
+ if ($config->get('HTML.TargetNoreferrer')) {\r
+ $modules[] = 'TargetNoreferrer';\r
+ }\r
+ if ($config->get('HTML.TargetNoopener')) {\r
+ $modules[] = 'TargetNoopener';\r
+ }\r
+\r
+ // merge in custom modules\r
+ $modules = array_merge($modules, $this->userModules);\r
+\r
+ foreach ($modules as $module) {\r
+ $this->processModule($module);\r
+ $this->modules[$module]->setup($config);\r
+ }\r
+\r
+ foreach ($this->doctype->tidyModules as $module) {\r
+ $this->processModule($module);\r
+ $this->modules[$module]->setup($config);\r
+ }\r
+\r
+ // prepare any injectors\r
+ foreach ($this->modules as $module) {\r
+ $n = array();\r
+ foreach ($module->info_injector as $injector) {\r
+ if (!is_object($injector)) {\r
+ $class = "HTMLPurifier_Injector_$injector";\r
+ $injector = new $class;\r
+ }\r
+ $n[$injector->name] = $injector;\r
+ }\r
+ $module->info_injector = $n;\r
+ }\r
+\r
+ // setup lookup table based on all valid modules\r
+ foreach ($this->modules as $module) {\r
+ foreach ($module->info as $name => $def) {\r
+ if (!isset($this->elementLookup[$name])) {\r
+ $this->elementLookup[$name] = array();\r
+ }\r
+ $this->elementLookup[$name][] = $module->name;\r
+ }\r
+ }\r
+\r
+ // note the different choice\r
+ $this->contentSets = new HTMLPurifier_ContentSets(\r
+ // content set assembly deals with all possible modules,\r
+ // not just ones deemed to be "safe"\r
+ $this->modules\r
+ );\r
+ $this->attrCollections = new HTMLPurifier_AttrCollections(\r
+ $this->attrTypes,\r
+ // there is no way to directly disable a global attribute,\r
+ // but using AllowedAttributes or simply not including\r
+ // the module in your custom doctype should be sufficient\r
+ $this->modules\r
+ );\r
+ }\r
+\r
+ /**\r
+ * Takes a module and adds it to the active module collection,\r
+ * registering it if necessary.\r
+ */\r
+ public function processModule($module)\r
+ {\r
+ if (!isset($this->registeredModules[$module]) || is_object($module)) {\r
+ $this->registerModule($module);\r
+ }\r
+ $this->modules[$module] = $this->registeredModules[$module];\r
+ }\r
+\r
+ /**\r
+ * Retrieves merged element definitions.\r
+ * @return Array of HTMLPurifier_ElementDef\r
+ */\r
+ public function getElements()\r
+ {\r
+ $elements = array();\r
+ foreach ($this->modules as $module) {\r
+ if (!$this->trusted && !$module->safe) {\r
+ continue;\r
+ }\r
+ foreach ($module->info as $name => $v) {\r
+ if (isset($elements[$name])) {\r
+ continue;\r
+ }\r
+ $elements[$name] = $this->getElement($name);\r
+ }\r
+ }\r
+\r
+ // remove dud elements, this happens when an element that\r
+ // appeared to be safe actually wasn't\r
+ foreach ($elements as $n => $v) {\r
+ if ($v === false) {\r
+ unset($elements[$n]);\r
+ }\r
+ }\r
+\r
+ return $elements;\r
+\r
+ }\r
+\r
+ /**\r
+ * Retrieves a single merged element definition\r
+ * @param string $name Name of element\r
+ * @param bool $trusted Boolean trusted overriding parameter: set to true\r
+ * if you want the full version of an element\r
+ * @return HTMLPurifier_ElementDef Merged HTMLPurifier_ElementDef\r
+ * @note You may notice that modules are getting iterated over twice (once\r
+ * in getElements() and once here). This\r
+ * is because\r
+ */\r
+ public function getElement($name, $trusted = null)\r
+ {\r
+ if (!isset($this->elementLookup[$name])) {\r
+ return false;\r
+ }\r
+\r
+ // setup global state variables\r
+ $def = false;\r
+ if ($trusted === null) {\r
+ $trusted = $this->trusted;\r
+ }\r
+\r
+ // iterate through each module that has registered itself to this\r
+ // element\r
+ foreach ($this->elementLookup[$name] as $module_name) {\r
+ $module = $this->modules[$module_name];\r
+\r
+ // refuse to create/merge from a module that is deemed unsafe--\r
+ // pretend the module doesn't exist--when trusted mode is not on.\r
+ if (!$trusted && !$module->safe) {\r
+ continue;\r
+ }\r
+\r
+ // clone is used because, ideally speaking, the original\r
+ // definition should not be modified. Usually, this will\r
+ // make no difference, but for consistency's sake\r
+ $new_def = clone $module->info[$name];\r
+\r
+ if (!$def && $new_def->standalone) {\r
+ $def = $new_def;\r
+ } elseif ($def) {\r
+ // This will occur even if $new_def is standalone. In practice,\r
+ // this will usually result in a full replacement.\r
+ $def->mergeIn($new_def);\r
+ } else {\r
+ // :TODO:\r
+ // non-standalone definitions that don't have a standalone\r
+ // to merge into could be deferred to the end\r
+ // HOWEVER, it is perfectly valid for a non-standalone\r
+ // definition to lack a standalone definition, even\r
+ // after all processing: this allows us to safely\r
+ // specify extra attributes for elements that may not be\r
+ // enabled all in one place. In particular, this might\r
+ // be the case for trusted elements. WARNING: care must\r
+ // be taken that the /extra/ definitions are all safe.\r
+ continue;\r
+ }\r
+\r
+ // attribute value expansions\r
+ $this->attrCollections->performInclusions($def->attr);\r
+ $this->attrCollections->expandIdentifiers($def->attr, $this->attrTypes);\r
+\r
+ // descendants_are_inline, for ChildDef_Chameleon\r
+ if (is_string($def->content_model) &&\r
+ strpos($def->content_model, 'Inline') !== false) {\r
+ if ($name != 'del' && $name != 'ins') {\r
+ // this is for you, ins/del\r
+ $def->descendants_are_inline = true;\r
+ }\r
+ }\r
+\r
+ $this->contentSets->generateChildDef($def, $module);\r
+ }\r
+\r
+ // This can occur if there is a blank definition, but no base to\r
+ // mix it in with\r
+ if (!$def) {\r
+ return false;\r
+ }\r
+\r
+ // add information on required attributes\r
+ foreach ($def->attr as $attr_name => $attr_def) {\r
+ if ($attr_def->required) {\r
+ $def->required_attr[] = $attr_name;\r
+ }\r
+ }\r
+ return $def;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Component of HTMLPurifier_AttrContext that accumulates IDs to prevent dupes\r
+ * @note In Slashdot-speak, dupe means duplicate.\r
+ * @note The default constructor does not accept $config or $context objects:\r
+ * use must use the static build() factory method to perform initialization.\r
+ */\r
+class HTMLPurifier_IDAccumulator\r
+{\r
+\r
+ /**\r
+ * Lookup table of IDs we've accumulated.\r
+ * @public\r
+ */\r
+ public $ids = array();\r
+\r
+ /**\r
+ * Builds an IDAccumulator, also initializing the default blacklist\r
+ * @param HTMLPurifier_Config $config Instance of HTMLPurifier_Config\r
+ * @param HTMLPurifier_Context $context Instance of HTMLPurifier_Context\r
+ * @return HTMLPurifier_IDAccumulator Fully initialized HTMLPurifier_IDAccumulator\r
+ */\r
+ public static function build($config, $context)\r
+ {\r
+ $id_accumulator = new HTMLPurifier_IDAccumulator();\r
+ $id_accumulator->load($config->get('Attr.IDBlacklist'));\r
+ return $id_accumulator;\r
+ }\r
+\r
+ /**\r
+ * Add an ID to the lookup table.\r
+ * @param string $id ID to be added.\r
+ * @return bool status, true if success, false if there's a dupe\r
+ */\r
+ public function add($id)\r
+ {\r
+ if (isset($this->ids[$id])) {\r
+ return false;\r
+ }\r
+ return $this->ids[$id] = true;\r
+ }\r
+\r
+ /**\r
+ * Load a list of IDs into the lookup table\r
+ * @param $array_of_ids Array of IDs to load\r
+ * @note This function doesn't care about duplicates\r
+ */\r
+ public function load($array_of_ids)\r
+ {\r
+ foreach ($array_of_ids as $id) {\r
+ $this->ids[$id] = true;\r
+ }\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Injects tokens into the document while parsing for well-formedness.\r
+ * This enables "formatter-like" functionality such as auto-paragraphing,\r
+ * smiley-ification and linkification to take place.\r
+ *\r
+ * A note on how handlers create changes; this is done by assigning a new\r
+ * value to the $token reference. These values can take a variety of forms and\r
+ * are best described HTMLPurifier_Strategy_MakeWellFormed->processToken()\r
+ * documentation.\r
+ *\r
+ * @todo Allow injectors to request a re-run on their output. This\r
+ * would help if an operation is recursive.\r
+ */\r
+abstract class HTMLPurifier_Injector\r
+{\r
+\r
+ /**\r
+ * Advisory name of injector, this is for friendly error messages.\r
+ * @type string\r
+ */\r
+ public $name;\r
+\r
+ /**\r
+ * @type HTMLPurifier_HTMLDefinition\r
+ */\r
+ protected $htmlDefinition;\r
+\r
+ /**\r
+ * Reference to CurrentNesting variable in Context. This is an array\r
+ * list of tokens that we are currently "inside"\r
+ * @type array\r
+ */\r
+ protected $currentNesting;\r
+\r
+ /**\r
+ * Reference to current token.\r
+ * @type HTMLPurifier_Token\r
+ */\r
+ protected $currentToken;\r
+\r
+ /**\r
+ * Reference to InputZipper variable in Context.\r
+ * @type HTMLPurifier_Zipper\r
+ */\r
+ protected $inputZipper;\r
+\r
+ /**\r
+ * Array of elements and attributes this injector creates and therefore\r
+ * need to be allowed by the definition. Takes form of\r
+ * array('element' => array('attr', 'attr2'), 'element2')\r
+ * @type array\r
+ */\r
+ public $needed = array();\r
+\r
+ /**\r
+ * Number of elements to rewind backwards (relative).\r
+ * @type bool|int\r
+ */\r
+ protected $rewindOffset = false;\r
+\r
+ /**\r
+ * Rewind to a spot to re-perform processing. This is useful if you\r
+ * deleted a node, and now need to see if this change affected any\r
+ * earlier nodes. Rewinding does not affect other injectors, and can\r
+ * result in infinite loops if not used carefully.\r
+ * @param bool|int $offset\r
+ * @warning HTML Purifier will prevent you from fast-forwarding with this\r
+ * function.\r
+ */\r
+ public function rewindOffset($offset)\r
+ {\r
+ $this->rewindOffset = $offset;\r
+ }\r
+\r
+ /**\r
+ * Retrieves rewind offset, and then unsets it.\r
+ * @return bool|int\r
+ */\r
+ public function getRewindOffset()\r
+ {\r
+ $r = $this->rewindOffset;\r
+ $this->rewindOffset = false;\r
+ return $r;\r
+ }\r
+\r
+ /**\r
+ * Prepares the injector by giving it the config and context objects:\r
+ * this allows references to important variables to be made within\r
+ * the injector. This function also checks if the HTML environment\r
+ * will work with the Injector (see checkNeeded()).\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string Boolean false if success, string of missing needed element/attribute if failure\r
+ */\r
+ public function prepare($config, $context)\r
+ {\r
+ $this->htmlDefinition = $config->getHTMLDefinition();\r
+ // Even though this might fail, some unit tests ignore this and\r
+ // still test checkNeeded, so be careful. Maybe get rid of that\r
+ // dependency.\r
+ $result = $this->checkNeeded($config);\r
+ if ($result !== false) {\r
+ return $result;\r
+ }\r
+ $this->currentNesting =& $context->get('CurrentNesting');\r
+ $this->currentToken =& $context->get('CurrentToken');\r
+ $this->inputZipper =& $context->get('InputZipper');\r
+ return false;\r
+ }\r
+\r
+ /**\r
+ * This function checks if the HTML environment\r
+ * will work with the Injector: if p tags are not allowed, the\r
+ * Auto-Paragraphing injector should not be enabled.\r
+ * @param HTMLPurifier_Config $config\r
+ * @return bool|string Boolean false if success, string of missing needed element/attribute if failure\r
+ */\r
+ public function checkNeeded($config)\r
+ {\r
+ $def = $config->getHTMLDefinition();\r
+ foreach ($this->needed as $element => $attributes) {\r
+ if (is_int($element)) {\r
+ $element = $attributes;\r
+ }\r
+ if (!isset($def->info[$element])) {\r
+ return $element;\r
+ }\r
+ if (!is_array($attributes)) {\r
+ continue;\r
+ }\r
+ foreach ($attributes as $name) {\r
+ if (!isset($def->info[$element]->attr[$name])) {\r
+ return "$element.$name";\r
+ }\r
+ }\r
+ }\r
+ return false;\r
+ }\r
+\r
+ /**\r
+ * Tests if the context node allows a certain element\r
+ * @param string $name Name of element to test for\r
+ * @return bool True if element is allowed, false if it is not\r
+ */\r
+ public function allowsElement($name)\r
+ {\r
+ if (!empty($this->currentNesting)) {\r
+ $parent_token = array_pop($this->currentNesting);\r
+ $this->currentNesting[] = $parent_token;\r
+ $parent = $this->htmlDefinition->info[$parent_token->name];\r
+ } else {\r
+ $parent = $this->htmlDefinition->info_parent_def;\r
+ }\r
+ if (!isset($parent->child->elements[$name]) || isset($parent->excludes[$name])) {\r
+ return false;\r
+ }\r
+ // check for exclusion\r
+ if (!empty($this->currentNesting)) {\r
+ for ($i = count($this->currentNesting) - 2; $i >= 0; $i--) {\r
+ $node = $this->currentNesting[$i];\r
+ $def = $this->htmlDefinition->info[$node->name];\r
+ if (isset($def->excludes[$name])) {\r
+ return false;\r
+ }\r
+ }\r
+ }\r
+ return true;\r
+ }\r
+\r
+ /**\r
+ * Iterator function, which starts with the next token and continues until\r
+ * you reach the end of the input tokens.\r
+ * @warning Please prevent previous references from interfering with this\r
+ * functions by setting $i = null beforehand!\r
+ * @param int $i Current integer index variable for inputTokens\r
+ * @param HTMLPurifier_Token $current Current token variable.\r
+ * Do NOT use $token, as that variable is also a reference\r
+ * @return bool\r
+ */\r
+ protected function forward(&$i, &$current)\r
+ {\r
+ if ($i === null) {\r
+ $i = count($this->inputZipper->back) - 1;\r
+ } else {\r
+ $i--;\r
+ }\r
+ if ($i < 0) {\r
+ return false;\r
+ }\r
+ $current = $this->inputZipper->back[$i];\r
+ return true;\r
+ }\r
+\r
+ /**\r
+ * Similar to _forward, but accepts a third parameter $nesting (which\r
+ * should be initialized at 0) and stops when we hit the end tag\r
+ * for the node $this->inputIndex starts in.\r
+ * @param int $i Current integer index variable for inputTokens\r
+ * @param HTMLPurifier_Token $current Current token variable.\r
+ * Do NOT use $token, as that variable is also a reference\r
+ * @param int $nesting\r
+ * @return bool\r
+ */\r
+ protected function forwardUntilEndToken(&$i, &$current, &$nesting)\r
+ {\r
+ $result = $this->forward($i, $current);\r
+ if (!$result) {\r
+ return false;\r
+ }\r
+ if ($nesting === null) {\r
+ $nesting = 0;\r
+ }\r
+ if ($current instanceof HTMLPurifier_Token_Start) {\r
+ $nesting++;\r
+ } elseif ($current instanceof HTMLPurifier_Token_End) {\r
+ if ($nesting <= 0) {\r
+ return false;\r
+ }\r
+ $nesting--;\r
+ }\r
+ return true;\r
+ }\r
+\r
+ /**\r
+ * Iterator function, starts with the previous token and continues until\r
+ * you reach the beginning of input tokens.\r
+ * @warning Please prevent previous references from interfering with this\r
+ * functions by setting $i = null beforehand!\r
+ * @param int $i Current integer index variable for inputTokens\r
+ * @param HTMLPurifier_Token $current Current token variable.\r
+ * Do NOT use $token, as that variable is also a reference\r
+ * @return bool\r
+ */\r
+ protected function backward(&$i, &$current)\r
+ {\r
+ if ($i === null) {\r
+ $i = count($this->inputZipper->front) - 1;\r
+ } else {\r
+ $i--;\r
+ }\r
+ if ($i < 0) {\r
+ return false;\r
+ }\r
+ $current = $this->inputZipper->front[$i];\r
+ return true;\r
+ }\r
+\r
+ /**\r
+ * Handler that is called when a text token is processed\r
+ */\r
+ public function handleText(&$token)\r
+ {\r
+ }\r
+\r
+ /**\r
+ * Handler that is called when a start or empty token is processed\r
+ */\r
+ public function handleElement(&$token)\r
+ {\r
+ }\r
+\r
+ /**\r
+ * Handler that is called when an end token is processed\r
+ */\r
+ public function handleEnd(&$token)\r
+ {\r
+ $this->notifyEnd($token);\r
+ }\r
+\r
+ /**\r
+ * Notifier that is called when an end token is processed\r
+ * @param HTMLPurifier_Token $token Current token variable.\r
+ * @note This differs from handlers in that the token is read-only\r
+ * @deprecated\r
+ */\r
+ public function notifyEnd($token)\r
+ {\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Represents a language and defines localizable string formatting and\r
+ * other functions, as well as the localized messages for HTML Purifier.\r
+ */\r
+class HTMLPurifier_Language\r
+{\r
+\r
+ /**\r
+ * ISO 639 language code of language. Prefers shortest possible version.\r
+ * @type string\r
+ */\r
+ public $code = 'en';\r
+\r
+ /**\r
+ * Fallback language code.\r
+ * @type bool|string\r
+ */\r
+ public $fallback = false;\r
+\r
+ /**\r
+ * Array of localizable messages.\r
+ * @type array\r
+ */\r
+ public $messages = array();\r
+\r
+ /**\r
+ * Array of localizable error codes.\r
+ * @type array\r
+ */\r
+ public $errorNames = array();\r
+\r
+ /**\r
+ * True if no message file was found for this language, so English\r
+ * is being used instead. Check this if you'd like to notify the\r
+ * user that they've used a non-supported language.\r
+ * @type bool\r
+ */\r
+ public $error = false;\r
+\r
+ /**\r
+ * Has the language object been loaded yet?\r
+ * @type bool\r
+ * @todo Make it private, fix usage in HTMLPurifier_LanguageTest\r
+ */\r
+ public $_loaded = false;\r
+\r
+ /**\r
+ * @type HTMLPurifier_Config\r
+ */\r
+ protected $config;\r
+\r
+ /**\r
+ * @type HTMLPurifier_Context\r
+ */\r
+ protected $context;\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ */\r
+ public function __construct($config, $context)\r
+ {\r
+ $this->config = $config;\r
+ $this->context = $context;\r
+ }\r
+\r
+ /**\r
+ * Loads language object with necessary info from factory cache\r
+ * @note This is a lazy loader\r
+ */\r
+ public function load()\r
+ {\r
+ if ($this->_loaded) {\r
+ return;\r
+ }\r
+ $factory = HTMLPurifier_LanguageFactory::instance();\r
+ $factory->loadLanguage($this->code);\r
+ foreach ($factory->keys as $key) {\r
+ $this->$key = $factory->cache[$this->code][$key];\r
+ }\r
+ $this->_loaded = true;\r
+ }\r
+\r
+ /**\r
+ * Retrieves a localised message.\r
+ * @param string $key string identifier of message\r
+ * @return string localised message\r
+ */\r
+ public function getMessage($key)\r
+ {\r
+ if (!$this->_loaded) {\r
+ $this->load();\r
+ }\r
+ if (!isset($this->messages[$key])) {\r
+ return "[$key]";\r
+ }\r
+ return $this->messages[$key];\r
+ }\r
+\r
+ /**\r
+ * Retrieves a localised error name.\r
+ * @param int $int error number, corresponding to PHP's error reporting\r
+ * @return string localised message\r
+ */\r
+ public function getErrorName($int)\r
+ {\r
+ if (!$this->_loaded) {\r
+ $this->load();\r
+ }\r
+ if (!isset($this->errorNames[$int])) {\r
+ return "[Error: $int]";\r
+ }\r
+ return $this->errorNames[$int];\r
+ }\r
+\r
+ /**\r
+ * Converts an array list into a string readable representation\r
+ * @param array $array\r
+ * @return string\r
+ */\r
+ public function listify($array)\r
+ {\r
+ $sep = $this->getMessage('Item separator');\r
+ $sep_last = $this->getMessage('Item separator last');\r
+ $ret = '';\r
+ for ($i = 0, $c = count($array); $i < $c; $i++) {\r
+ if ($i == 0) {\r
+ } elseif ($i + 1 < $c) {\r
+ $ret .= $sep;\r
+ } else {\r
+ $ret .= $sep_last;\r
+ }\r
+ $ret .= $array[$i];\r
+ }\r
+ return $ret;\r
+ }\r
+\r
+ /**\r
+ * Formats a localised message with passed parameters\r
+ * @param string $key string identifier of message\r
+ * @param array $args Parameters to substitute in\r
+ * @return string localised message\r
+ * @todo Implement conditionals? Right now, some messages make\r
+ * reference to line numbers, but those aren't always available\r
+ */\r
+ public function formatMessage($key, $args = array())\r
+ {\r
+ if (!$this->_loaded) {\r
+ $this->load();\r
+ }\r
+ if (!isset($this->messages[$key])) {\r
+ return "[$key]";\r
+ }\r
+ $raw = $this->messages[$key];\r
+ $subst = array();\r
+ $generator = false;\r
+ foreach ($args as $i => $value) {\r
+ if (is_object($value)) {\r
+ if ($value instanceof HTMLPurifier_Token) {\r
+ // factor this out some time\r
+ if (!$generator) {\r
+ $generator = $this->context->get('Generator');\r
+ }\r
+ if (isset($value->name)) {\r
+ $subst['$'.$i.'.Name'] = $value->name;\r
+ }\r
+ if (isset($value->data)) {\r
+ $subst['$'.$i.'.Data'] = $value->data;\r
+ }\r
+ $subst['$'.$i.'.Compact'] =\r
+ $subst['$'.$i.'.Serialized'] = $generator->generateFromToken($value);\r
+ // a more complex algorithm for compact representation\r
+ // could be introduced for all types of tokens. This\r
+ // may need to be factored out into a dedicated class\r
+ if (!empty($value->attr)) {\r
+ $stripped_token = clone $value;\r
+ $stripped_token->attr = array();\r
+ $subst['$'.$i.'.Compact'] = $generator->generateFromToken($stripped_token);\r
+ }\r
+ $subst['$'.$i.'.Line'] = $value->line ? $value->line : 'unknown';\r
+ }\r
+ continue;\r
+ } elseif (is_array($value)) {\r
+ $keys = array_keys($value);\r
+ if (array_keys($keys) === $keys) {\r
+ // list\r
+ $subst['$'.$i] = $this->listify($value);\r
+ } else {\r
+ // associative array\r
+ // no $i implementation yet, sorry\r
+ $subst['$'.$i.'.Keys'] = $this->listify($keys);\r
+ $subst['$'.$i.'.Values'] = $this->listify(array_values($value));\r
+ }\r
+ continue;\r
+ }\r
+ $subst['$' . $i] = $value;\r
+ }\r
+ return strtr($raw, $subst);\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Class responsible for generating HTMLPurifier_Language objects, managing\r
+ * caching and fallbacks.\r
+ * @note Thanks to MediaWiki for the general logic, although this version\r
+ * has been entirely rewritten\r
+ * @todo Serialized cache for languages\r
+ */\r
+class HTMLPurifier_LanguageFactory\r
+{\r
+\r
+ /**\r
+ * Cache of language code information used to load HTMLPurifier_Language objects.\r
+ * Structure is: $factory->cache[$language_code][$key] = $value\r
+ * @type array\r
+ */\r
+ public $cache;\r
+\r
+ /**\r
+ * Valid keys in the HTMLPurifier_Language object. Designates which\r
+ * variables to slurp out of a message file.\r
+ * @type array\r
+ */\r
+ public $keys = array('fallback', 'messages', 'errorNames');\r
+\r
+ /**\r
+ * Instance to validate language codes.\r
+ * @type HTMLPurifier_AttrDef_Lang\r
+ *\r
+ */\r
+ protected $validator;\r
+\r
+ /**\r
+ * Cached copy of dirname(__FILE__), directory of current file without\r
+ * trailing slash.\r
+ * @type string\r
+ */\r
+ protected $dir;\r
+\r
+ /**\r
+ * Keys whose contents are a hash map and can be merged.\r
+ * @type array\r
+ */\r
+ protected $mergeable_keys_map = array('messages' => true, 'errorNames' => true);\r
+\r
+ /**\r
+ * Keys whose contents are a list and can be merged.\r
+ * @value array lookup\r
+ */\r
+ protected $mergeable_keys_list = array();\r
+\r
+ /**\r
+ * Retrieve sole instance of the factory.\r
+ * @param HTMLPurifier_LanguageFactory $prototype Optional prototype to overload sole instance with,\r
+ * or bool true to reset to default factory.\r
+ * @return HTMLPurifier_LanguageFactory\r
+ */\r
+ public static function instance($prototype = null)\r
+ {\r
+ static $instance = null;\r
+ if ($prototype !== null) {\r
+ $instance = $prototype;\r
+ } elseif ($instance === null || $prototype == true) {\r
+ $instance = new HTMLPurifier_LanguageFactory();\r
+ $instance->setup();\r
+ }\r
+ return $instance;\r
+ }\r
+\r
+ /**\r
+ * Sets up the singleton, much like a constructor\r
+ * @note Prevents people from getting this outside of the singleton\r
+ */\r
+ public function setup()\r
+ {\r
+ $this->validator = new HTMLPurifier_AttrDef_Lang();\r
+ $this->dir = HTMLPURIFIER_PREFIX . '/HTMLPurifier';\r
+ }\r
+\r
+ /**\r
+ * Creates a language object, handles class fallbacks\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @param bool|string $code Code to override configuration with. Private parameter.\r
+ * @return HTMLPurifier_Language\r
+ */\r
+ public function create($config, $context, $code = false)\r
+ {\r
+ // validate language code\r
+ if ($code === false) {\r
+ $code = $this->validator->validate(\r
+ $config->get('Core.Language'),\r
+ $config,\r
+ $context\r
+ );\r
+ } else {\r
+ $code = $this->validator->validate($code, $config, $context);\r
+ }\r
+ if ($code === false) {\r
+ $code = 'en'; // malformed code becomes English\r
+ }\r
+\r
+ $pcode = str_replace('-', '_', $code); // make valid PHP classname\r
+ static $depth = 0; // recursion protection\r
+\r
+ if ($code == 'en') {\r
+ $lang = new HTMLPurifier_Language($config, $context);\r
+ } else {\r
+ $class = 'HTMLPurifier_Language_' . $pcode;\r
+ $file = $this->dir . '/Language/classes/' . $code . '.php';\r
+ if (file_exists($file) || class_exists($class, false)) {\r
+ $lang = new $class($config, $context);\r
+ } else {\r
+ // Go fallback\r
+ $raw_fallback = $this->getFallbackFor($code);\r
+ $fallback = $raw_fallback ? $raw_fallback : 'en';\r
+ $depth++;\r
+ $lang = $this->create($config, $context, $fallback);\r
+ if (!$raw_fallback) {\r
+ $lang->error = true;\r
+ }\r
+ $depth--;\r
+ }\r
+ }\r
+ $lang->code = $code;\r
+ return $lang;\r
+ }\r
+\r
+ /**\r
+ * Returns the fallback language for language\r
+ * @note Loads the original language into cache\r
+ * @param string $code language code\r
+ * @return string|bool\r
+ */\r
+ public function getFallbackFor($code)\r
+ {\r
+ $this->loadLanguage($code);\r
+ return $this->cache[$code]['fallback'];\r
+ }\r
+\r
+ /**\r
+ * Loads language into the cache, handles message file and fallbacks\r
+ * @param string $code language code\r
+ */\r
+ public function loadLanguage($code)\r
+ {\r
+ static $languages_seen = array(); // recursion guard\r
+\r
+ // abort if we've already loaded it\r
+ if (isset($this->cache[$code])) {\r
+ return;\r
+ }\r
+\r
+ // generate filename\r
+ $filename = $this->dir . '/Language/messages/' . $code . '.php';\r
+\r
+ // default fallback : may be overwritten by the ensuing include\r
+ $fallback = ($code != 'en') ? 'en' : false;\r
+\r
+ // load primary localisation\r
+ if (!file_exists($filename)) {\r
+ // skip the include: will rely solely on fallback\r
+ $filename = $this->dir . '/Language/messages/en.php';\r
+ $cache = array();\r
+ } else {\r
+ include $filename;\r
+ $cache = compact($this->keys);\r
+ }\r
+\r
+ // load fallback localisation\r
+ if (!empty($fallback)) {\r
+\r
+ // infinite recursion guard\r
+ if (isset($languages_seen[$code])) {\r
+ trigger_error(\r
+ 'Circular fallback reference in language ' .\r
+ $code,\r
+ E_USER_ERROR\r
+ );\r
+ $fallback = 'en';\r
+ }\r
+ $language_seen[$code] = true;\r
+\r
+ // load the fallback recursively\r
+ $this->loadLanguage($fallback);\r
+ $fallback_cache = $this->cache[$fallback];\r
+\r
+ // merge fallback with current language\r
+ foreach ($this->keys as $key) {\r
+ if (isset($cache[$key]) && isset($fallback_cache[$key])) {\r
+ if (isset($this->mergeable_keys_map[$key])) {\r
+ $cache[$key] = $cache[$key] + $fallback_cache[$key];\r
+ } elseif (isset($this->mergeable_keys_list[$key])) {\r
+ $cache[$key] = array_merge($fallback_cache[$key], $cache[$key]);\r
+ }\r
+ } else {\r
+ $cache[$key] = $fallback_cache[$key];\r
+ }\r
+ }\r
+ }\r
+\r
+ // save to cache for later retrieval\r
+ $this->cache[$code] = $cache;\r
+ return;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Represents a measurable length, with a string numeric magnitude\r
+ * and a unit. This object is immutable.\r
+ */\r
+class HTMLPurifier_Length\r
+{\r
+\r
+ /**\r
+ * String numeric magnitude.\r
+ * @type string\r
+ */\r
+ protected $n;\r
+\r
+ /**\r
+ * String unit. False is permitted if $n = 0.\r
+ * @type string|bool\r
+ */\r
+ protected $unit;\r
+\r
+ /**\r
+ * Whether or not this length is valid. Null if not calculated yet.\r
+ * @type bool\r
+ */\r
+ protected $isValid;\r
+\r
+ /**\r
+ * Array Lookup array of units recognized by CSS 3\r
+ * @type array\r
+ */\r
+ protected static $allowedUnits = array(\r
+ 'em' => true, 'ex' => true, 'px' => true, 'in' => true,\r
+ 'cm' => true, 'mm' => true, 'pt' => true, 'pc' => true,\r
+ 'ch' => true, 'rem' => true, 'vw' => true, 'vh' => true,\r
+ 'vmin' => true, 'vmax' => true\r
+ );\r
+\r
+ /**\r
+ * @param string $n Magnitude\r
+ * @param bool|string $u Unit\r
+ */\r
+ public function __construct($n = '0', $u = false)\r
+ {\r
+ $this->n = (string) $n;\r
+ $this->unit = $u !== false ? (string) $u : false;\r
+ }\r
+\r
+ /**\r
+ * @param string $s Unit string, like '2em' or '3.4in'\r
+ * @return HTMLPurifier_Length\r
+ * @warning Does not perform validation.\r
+ */\r
+ public static function make($s)\r
+ {\r
+ if ($s instanceof HTMLPurifier_Length) {\r
+ return $s;\r
+ }\r
+ $n_length = strspn($s, '1234567890.+-');\r
+ $n = substr($s, 0, $n_length);\r
+ $unit = substr($s, $n_length);\r
+ if ($unit === '') {\r
+ $unit = false;\r
+ }\r
+ return new HTMLPurifier_Length($n, $unit);\r
+ }\r
+\r
+ /**\r
+ * Validates the number and unit.\r
+ * @return bool\r
+ */\r
+ protected function validate()\r
+ {\r
+ // Special case:\r
+ if ($this->n === '+0' || $this->n === '-0') {\r
+ $this->n = '0';\r
+ }\r
+ if ($this->n === '0' && $this->unit === false) {\r
+ return true;\r
+ }\r
+ if (!ctype_lower($this->unit)) {\r
+ $this->unit = strtolower($this->unit);\r
+ }\r
+ if (!isset(HTMLPurifier_Length::$allowedUnits[$this->unit])) {\r
+ return false;\r
+ }\r
+ // Hack:\r
+ $def = new HTMLPurifier_AttrDef_CSS_Number();\r
+ $result = $def->validate($this->n, false, false);\r
+ if ($result === false) {\r
+ return false;\r
+ }\r
+ $this->n = $result;\r
+ return true;\r
+ }\r
+\r
+ /**\r
+ * Returns string representation of number.\r
+ * @return string\r
+ */\r
+ public function toString()\r
+ {\r
+ if (!$this->isValid()) {\r
+ return false;\r
+ }\r
+ return $this->n . $this->unit;\r
+ }\r
+\r
+ /**\r
+ * Retrieves string numeric magnitude.\r
+ * @return string\r
+ */\r
+ public function getN()\r
+ {\r
+ return $this->n;\r
+ }\r
+\r
+ /**\r
+ * Retrieves string unit.\r
+ * @return string\r
+ */\r
+ public function getUnit()\r
+ {\r
+ return $this->unit;\r
+ }\r
+\r
+ /**\r
+ * Returns true if this length unit is valid.\r
+ * @return bool\r
+ */\r
+ public function isValid()\r
+ {\r
+ if ($this->isValid === null) {\r
+ $this->isValid = $this->validate();\r
+ }\r
+ return $this->isValid;\r
+ }\r
+\r
+ /**\r
+ * Compares two lengths, and returns 1 if greater, -1 if less and 0 if equal.\r
+ * @param HTMLPurifier_Length $l\r
+ * @return int\r
+ * @warning If both values are too large or small, this calculation will\r
+ * not work properly\r
+ */\r
+ public function compareTo($l)\r
+ {\r
+ if ($l === false) {\r
+ return false;\r
+ }\r
+ if ($l->unit !== $this->unit) {\r
+ $converter = new HTMLPurifier_UnitConverter();\r
+ $l = $converter->convert($l, $this->unit);\r
+ if ($l === false) {\r
+ return false;\r
+ }\r
+ }\r
+ return $this->n - $l->n;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Forgivingly lexes HTML (SGML-style) markup into tokens.\r
+ *\r
+ * A lexer parses a string of SGML-style markup and converts them into\r
+ * corresponding tokens. It doesn't check for well-formedness, although its\r
+ * internal mechanism may make this automatic (such as the case of\r
+ * HTMLPurifier_Lexer_DOMLex). There are several implementations to choose\r
+ * from.\r
+ *\r
+ * A lexer is HTML-oriented: it might work with XML, but it's not\r
+ * recommended, as we adhere to a subset of the specification for optimization\r
+ * reasons. This might change in the future. Also, most tokenizers are not\r
+ * expected to handle DTDs or PIs.\r
+ *\r
+ * This class should not be directly instantiated, but you may use create() to\r
+ * retrieve a default copy of the lexer. Being a supertype, this class\r
+ * does not actually define any implementation, but offers commonly used\r
+ * convenience functions for subclasses.\r
+ *\r
+ * @note The unit tests will instantiate this class for testing purposes, as\r
+ * many of the utility functions require a class to be instantiated.\r
+ * This means that, even though this class is not runnable, it will\r
+ * not be declared abstract.\r
+ *\r
+ * @par\r
+ *\r
+ * @note\r
+ * We use tokens rather than create a DOM representation because DOM would:\r
+ *\r
+ * @par\r
+ * -# Require more processing and memory to create,\r
+ * -# Is not streamable, and\r
+ * -# Has the entire document structure (html and body not needed).\r
+ *\r
+ * @par\r
+ * However, DOM is helpful in that it makes it easy to move around nodes\r
+ * without a lot of lookaheads to see when a tag is closed. This is a\r
+ * limitation of the token system and some workarounds would be nice.\r
+ */\r
+class HTMLPurifier_Lexer\r
+{\r
+\r
+ /**\r
+ * Whether or not this lexer implements line-number/column-number tracking.\r
+ * If it does, set to true.\r
+ */\r
+ public $tracksLineNumbers = false;\r
+\r
+ // -- STATIC ----------------------------------------------------------\r
+\r
+ /**\r
+ * Retrieves or sets the default Lexer as a Prototype Factory.\r
+ *\r
+ * By default HTMLPurifier_Lexer_DOMLex will be returned. There are\r
+ * a few exceptions involving special features that only DirectLex\r
+ * implements.\r
+ *\r
+ * @note The behavior of this class has changed, rather than accepting\r
+ * a prototype object, it now accepts a configuration object.\r
+ * To specify your own prototype, set %Core.LexerImpl to it.\r
+ * This change in behavior de-singletonizes the lexer object.\r
+ *\r
+ * @param HTMLPurifier_Config $config\r
+ * @return HTMLPurifier_Lexer\r
+ * @throws HTMLPurifier_Exception\r
+ */\r
+ public static function create($config)\r
+ {\r
+ if (!($config instanceof HTMLPurifier_Config)) {\r
+ $lexer = $config;\r
+ trigger_error(\r
+ "Passing a prototype to\r
+ HTMLPurifier_Lexer::create() is deprecated, please instead\r
+ use %Core.LexerImpl",\r
+ E_USER_WARNING\r
+ );\r
+ } else {\r
+ $lexer = $config->get('Core.LexerImpl');\r
+ }\r
+\r
+ $needs_tracking =\r
+ $config->get('Core.MaintainLineNumbers') ||\r
+ $config->get('Core.CollectErrors');\r
+\r
+ $inst = null;\r
+ if (is_object($lexer)) {\r
+ $inst = $lexer;\r
+ } else {\r
+ if (is_null($lexer)) {\r
+ do {\r
+ // auto-detection algorithm\r
+ if ($needs_tracking) {\r
+ $lexer = 'DirectLex';\r
+ break;\r
+ }\r
+\r
+ if (class_exists('DOMDocument', false) &&\r
+ method_exists('DOMDocument', 'loadHTML') &&\r
+ !extension_loaded('domxml')\r
+ ) {\r
+ // check for DOM support, because while it's part of the\r
+ // core, it can be disabled compile time. Also, the PECL\r
+ // domxml extension overrides the default DOM, and is evil\r
+ // and nasty and we shan't bother to support it\r
+ $lexer = 'DOMLex';\r
+ } else {\r
+ $lexer = 'DirectLex';\r
+ }\r
+ } while (0);\r
+ } // do..while so we can break\r
+\r
+ // instantiate recognized string names\r
+ switch ($lexer) {\r
+ case 'DOMLex':\r
+ $inst = new HTMLPurifier_Lexer_DOMLex();\r
+ break;\r
+ case 'DirectLex':\r
+ $inst = new HTMLPurifier_Lexer_DirectLex();\r
+ break;\r
+ case 'PH5P':\r
+ $inst = new HTMLPurifier_Lexer_PH5P();\r
+ break;\r
+ default:\r
+ throw new HTMLPurifier_Exception(\r
+ "Cannot instantiate unrecognized Lexer type " .\r
+ htmlspecialchars($lexer)\r
+ );\r
+ }\r
+ }\r
+\r
+ if (!$inst) {\r
+ throw new HTMLPurifier_Exception('No lexer was instantiated');\r
+ }\r
+\r
+ // once PHP DOM implements native line numbers, or we\r
+ // hack out something using XSLT, remove this stipulation\r
+ if ($needs_tracking && !$inst->tracksLineNumbers) {\r
+ throw new HTMLPurifier_Exception(\r
+ 'Cannot use lexer that does not support line numbers with ' .\r
+ 'Core.MaintainLineNumbers or Core.CollectErrors (use DirectLex instead)'\r
+ );\r
+ }\r
+\r
+ return $inst;\r
+\r
+ }\r
+\r
+ // -- CONVENIENCE MEMBERS ---------------------------------------------\r
+\r
+ public function __construct()\r
+ {\r
+ $this->_entity_parser = new HTMLPurifier_EntityParser();\r
+ }\r
+\r
+ /**\r
+ * Most common entity to raw value conversion table for special entities.\r
+ * @type array\r
+ */\r
+ protected $_special_entity2str =\r
+ array(\r
+ '"' => '"',\r
+ '&' => '&',\r
+ '<' => '<',\r
+ '>' => '>',\r
+ ''' => "'",\r
+ ''' => "'",\r
+ ''' => "'"\r
+ );\r
+\r
+ public function parseText($string, $config) {\r
+ return $this->parseData($string, false, $config);\r
+ }\r
+\r
+ public function parseAttr($string, $config) {\r
+ return $this->parseData($string, true, $config);\r
+ }\r
+\r
+ /**\r
+ * Parses special entities into the proper characters.\r
+ *\r
+ * This string will translate escaped versions of the special characters\r
+ * into the correct ones.\r
+ *\r
+ * @param string $string String character data to be parsed.\r
+ * @return string Parsed character data.\r
+ */\r
+ public function parseData($string, $is_attr, $config)\r
+ {\r
+ // following functions require at least one character\r
+ if ($string === '') {\r
+ return '';\r
+ }\r
+\r
+ // subtracts amps that cannot possibly be escaped\r
+ $num_amp = substr_count($string, '&') - substr_count($string, '& ') -\r
+ ($string[strlen($string) - 1] === '&' ? 1 : 0);\r
+\r
+ if (!$num_amp) {\r
+ return $string;\r
+ } // abort if no entities\r
+ $num_esc_amp = substr_count($string, '&');\r
+ $string = strtr($string, $this->_special_entity2str);\r
+\r
+ // code duplication for sake of optimization, see above\r
+ $num_amp_2 = substr_count($string, '&') - substr_count($string, '& ') -\r
+ ($string[strlen($string) - 1] === '&' ? 1 : 0);\r
+\r
+ if ($num_amp_2 <= $num_esc_amp) {\r
+ return $string;\r
+ }\r
+\r
+ // hmm... now we have some uncommon entities. Use the callback.\r
+ if ($config->get('Core.LegacyEntityDecoder')) {\r
+ $string = $this->_entity_parser->substituteSpecialEntities($string);\r
+ } else {\r
+ if ($is_attr) {\r
+ $string = $this->_entity_parser->substituteAttrEntities($string);\r
+ } else {\r
+ $string = $this->_entity_parser->substituteTextEntities($string);\r
+ }\r
+ }\r
+ return $string;\r
+ }\r
+\r
+ /**\r
+ * Lexes an HTML string into tokens.\r
+ * @param $string String HTML.\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return HTMLPurifier_Token[] array representation of HTML.\r
+ */\r
+ public function tokenizeHTML($string, $config, $context)\r
+ {\r
+ trigger_error('Call to abstract class', E_USER_ERROR);\r
+ }\r
+\r
+ /**\r
+ * Translates CDATA sections into regular sections (through escaping).\r
+ * @param string $string HTML string to process.\r
+ * @return string HTML with CDATA sections escaped.\r
+ */\r
+ protected static function escapeCDATA($string)\r
+ {\r
+ return preg_replace_callback(\r
+ '/<!\[CDATA\[(.+?)\]\]>/s',\r
+ array('HTMLPurifier_Lexer', 'CDATACallback'),\r
+ $string\r
+ );\r
+ }\r
+\r
+ /**\r
+ * Special CDATA case that is especially convoluted for <script>\r
+ * @param string $string HTML string to process.\r
+ * @return string HTML with CDATA sections escaped.\r
+ */\r
+ protected static function escapeCommentedCDATA($string)\r
+ {\r
+ return preg_replace_callback(\r
+ '#<!--//--><!\[CDATA\[//><!--(.+?)//--><!\]\]>#s',\r
+ array('HTMLPurifier_Lexer', 'CDATACallback'),\r
+ $string\r
+ );\r
+ }\r
+\r
+ /**\r
+ * Special Internet Explorer conditional comments should be removed.\r
+ * @param string $string HTML string to process.\r
+ * @return string HTML with conditional comments removed.\r
+ */\r
+ protected static function removeIEConditional($string)\r
+ {\r
+ return preg_replace(\r
+ '#<!--\[if [^>]+\]>.*?<!\[endif\]-->#si', // probably should generalize for all strings\r
+ '',\r
+ $string\r
+ );\r
+ }\r
+\r
+ /**\r
+ * Callback function for escapeCDATA() that does the work.\r
+ *\r
+ * @warning Though this is public in order to let the callback happen,\r
+ * calling it directly is not recommended.\r
+ * @param array $matches PCRE matches array, with index 0 the entire match\r
+ * and 1 the inside of the CDATA section.\r
+ * @return string Escaped internals of the CDATA section.\r
+ */\r
+ protected static function CDATACallback($matches)\r
+ {\r
+ // not exactly sure why the character set is needed, but whatever\r
+ return htmlspecialchars($matches[1], ENT_COMPAT, 'UTF-8');\r
+ }\r
+\r
+ /**\r
+ * Takes a piece of HTML and normalizes it by converting entities, fixing\r
+ * encoding, extracting bits, and other good stuff.\r
+ * @param string $html HTML.\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return string\r
+ * @todo Consider making protected\r
+ */\r
+ public function normalize($html, $config, $context)\r
+ {\r
+ // normalize newlines to \n\r
+ if ($config->get('Core.NormalizeNewlines')) {\r
+ $html = str_replace("\r\n", "\n", $html);\r
+ $html = str_replace("\r", "\n", $html);\r
+ }\r
+\r
+ if ($config->get('HTML.Trusted')) {\r
+ // escape convoluted CDATA\r
+ $html = $this->escapeCommentedCDATA($html);\r
+ }\r
+\r
+ // escape CDATA\r
+ $html = $this->escapeCDATA($html);\r
+\r
+ $html = $this->removeIEConditional($html);\r
+\r
+ // extract body from document if applicable\r
+ if ($config->get('Core.ConvertDocumentToFragment')) {\r
+ $e = false;\r
+ if ($config->get('Core.CollectErrors')) {\r
+ $e =& $context->get('ErrorCollector');\r
+ }\r
+ $new_html = $this->extractBody($html);\r
+ if ($e && $new_html != $html) {\r
+ $e->send(E_WARNING, 'Lexer: Extracted body');\r
+ }\r
+ $html = $new_html;\r
+ }\r
+\r
+ // expand entities that aren't the big five\r
+ if ($config->get('Core.LegacyEntityDecoder')) {\r
+ $html = $this->_entity_parser->substituteNonSpecialEntities($html);\r
+ }\r
+\r
+ // clean into wellformed UTF-8 string for an SGML context: this has\r
+ // to be done after entity expansion because the entities sometimes\r
+ // represent non-SGML characters (horror, horror!)\r
+ $html = HTMLPurifier_Encoder::cleanUTF8($html);\r
+\r
+ // if processing instructions are to removed, remove them now\r
+ if ($config->get('Core.RemoveProcessingInstructions')) {\r
+ $html = preg_replace('#<\?.+?\?>#s', '', $html);\r
+ }\r
+\r
+ $hidden_elements = $config->get('Core.HiddenElements');\r
+ if ($config->get('Core.AggressivelyRemoveScript') &&\r
+ !($config->get('HTML.Trusted') || !$config->get('Core.RemoveScriptContents')\r
+ || empty($hidden_elements["script"]))) {\r
+ $html = preg_replace('#<script[^>]*>.*?</script>#i', '', $html);\r
+ }\r
+\r
+ return $html;\r
+ }\r
+\r
+ /**\r
+ * Takes a string of HTML (fragment or document) and returns the content\r
+ * @todo Consider making protected\r
+ */\r
+ public function extractBody($html)\r
+ {\r
+ $matches = array();\r
+ $result = preg_match('|(.*?)<body[^>]*>(.*)</body>|is', $html, $matches);\r
+ if ($result) {\r
+ // Make sure it's not in a comment\r
+ $comment_start = strrpos($matches[1], '<!--');\r
+ $comment_end = strrpos($matches[1], '-->');\r
+ if ($comment_start === false ||\r
+ ($comment_end !== false && $comment_end > $comment_start)) {\r
+ return $matches[2];\r
+ }\r
+ }\r
+ return $html;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Abstract base node class that all others inherit from.\r
+ *\r
+ * Why do we not use the DOM extension? (1) It is not always available,\r
+ * (2) it has funny constraints on the data it can represent,\r
+ * whereas we want a maximally flexible representation, and (3) its\r
+ * interface is a bit cumbersome.\r
+ */\r
+abstract class HTMLPurifier_Node\r
+{\r
+ /**\r
+ * Line number of the start token in the source document\r
+ * @type int\r
+ */\r
+ public $line;\r
+\r
+ /**\r
+ * Column number of the start token in the source document. Null if unknown.\r
+ * @type int\r
+ */\r
+ public $col;\r
+\r
+ /**\r
+ * Lookup array of processing that this token is exempt from.\r
+ * Currently, valid values are "ValidateAttributes".\r
+ * @type array\r
+ */\r
+ public $armor = array();\r
+\r
+ /**\r
+ * When true, this node should be ignored as non-existent.\r
+ *\r
+ * Who is responsible for ignoring dead nodes? FixNesting is\r
+ * responsible for removing them before passing on to child\r
+ * validators.\r
+ */\r
+ public $dead = false;\r
+\r
+ /**\r
+ * Returns a pair of start and end tokens, where the end token\r
+ * is null if it is not necessary. Does not include children.\r
+ * @type array\r
+ */\r
+ abstract public function toTokenPair();\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Class that handles operations involving percent-encoding in URIs.\r
+ *\r
+ * @warning\r
+ * Be careful when reusing instances of PercentEncoder. The object\r
+ * you use for normalize() SHOULD NOT be used for encode(), or\r
+ * vice-versa.\r
+ */\r
+class HTMLPurifier_PercentEncoder\r
+{\r
+\r
+ /**\r
+ * Reserved characters to preserve when using encode().\r
+ * @type array\r
+ */\r
+ protected $preserve = array();\r
+\r
+ /**\r
+ * String of characters that should be preserved while using encode().\r
+ * @param bool $preserve\r
+ */\r
+ public function __construct($preserve = false)\r
+ {\r
+ // unreserved letters, ought to const-ify\r
+ for ($i = 48; $i <= 57; $i++) { // digits\r
+ $this->preserve[$i] = true;\r
+ }\r
+ for ($i = 65; $i <= 90; $i++) { // upper-case\r
+ $this->preserve[$i] = true;\r
+ }\r
+ for ($i = 97; $i <= 122; $i++) { // lower-case\r
+ $this->preserve[$i] = true;\r
+ }\r
+ $this->preserve[45] = true; // Dash -\r
+ $this->preserve[46] = true; // Period .\r
+ $this->preserve[95] = true; // Underscore _\r
+ $this->preserve[126]= true; // Tilde ~\r
+\r
+ // extra letters not to escape\r
+ if ($preserve !== false) {\r
+ for ($i = 0, $c = strlen($preserve); $i < $c; $i++) {\r
+ $this->preserve[ord($preserve[$i])] = true;\r
+ }\r
+ }\r
+ }\r
+\r
+ /**\r
+ * Our replacement for urlencode, it encodes all non-reserved characters,\r
+ * as well as any extra characters that were instructed to be preserved.\r
+ * @note\r
+ * Assumes that the string has already been normalized, making any\r
+ * and all percent escape sequences valid. Percents will not be\r
+ * re-escaped, regardless of their status in $preserve\r
+ * @param string $string String to be encoded\r
+ * @return string Encoded string.\r
+ */\r
+ public function encode($string)\r
+ {\r
+ $ret = '';\r
+ for ($i = 0, $c = strlen($string); $i < $c; $i++) {\r
+ if ($string[$i] !== '%' && !isset($this->preserve[$int = ord($string[$i])])) {\r
+ $ret .= '%' . sprintf('%02X', $int);\r
+ } else {\r
+ $ret .= $string[$i];\r
+ }\r
+ }\r
+ return $ret;\r
+ }\r
+\r
+ /**\r
+ * Fix up percent-encoding by decoding unreserved characters and normalizing.\r
+ * @warning This function is affected by $preserve, even though the\r
+ * usual desired behavior is for this not to preserve those\r
+ * characters. Be careful when reusing instances of PercentEncoder!\r
+ * @param string $string String to normalize\r
+ * @return string\r
+ */\r
+ public function normalize($string)\r
+ {\r
+ if ($string == '') {\r
+ return '';\r
+ }\r
+ $parts = explode('%', $string);\r
+ $ret = array_shift($parts);\r
+ foreach ($parts as $part) {\r
+ $length = strlen($part);\r
+ if ($length < 2) {\r
+ $ret .= '%25' . $part;\r
+ continue;\r
+ }\r
+ $encoding = substr($part, 0, 2);\r
+ $text = substr($part, 2);\r
+ if (!ctype_xdigit($encoding)) {\r
+ $ret .= '%25' . $part;\r
+ continue;\r
+ }\r
+ $int = hexdec($encoding);\r
+ if (isset($this->preserve[$int])) {\r
+ $ret .= chr($int) . $text;\r
+ continue;\r
+ }\r
+ $encoding = strtoupper($encoding);\r
+ $ret .= '%' . $encoding . $text;\r
+ }\r
+ return $ret;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Generic property list implementation\r
+ */\r
+class HTMLPurifier_PropertyList\r
+{\r
+ /**\r
+ * Internal data-structure for properties.\r
+ * @type array\r
+ */\r
+ protected $data = array();\r
+\r
+ /**\r
+ * Parent plist.\r
+ * @type HTMLPurifier_PropertyList\r
+ */\r
+ protected $parent;\r
+\r
+ /**\r
+ * Cache.\r
+ * @type array\r
+ */\r
+ protected $cache;\r
+\r
+ /**\r
+ * @param HTMLPurifier_PropertyList $parent Parent plist\r
+ */\r
+ public function __construct($parent = null)\r
+ {\r
+ $this->parent = $parent;\r
+ }\r
+\r
+ /**\r
+ * Recursively retrieves the value for a key\r
+ * @param string $name\r
+ * @throws HTMLPurifier_Exception\r
+ */\r
+ public function get($name)\r
+ {\r
+ if ($this->has($name)) {\r
+ return $this->data[$name];\r
+ }\r
+ // possible performance bottleneck, convert to iterative if necessary\r
+ if ($this->parent) {\r
+ return $this->parent->get($name);\r
+ }\r
+ throw new HTMLPurifier_Exception("Key '$name' not found");\r
+ }\r
+\r
+ /**\r
+ * Sets the value of a key, for this plist\r
+ * @param string $name\r
+ * @param mixed $value\r
+ */\r
+ public function set($name, $value)\r
+ {\r
+ $this->data[$name] = $value;\r
+ }\r
+\r
+ /**\r
+ * Returns true if a given key exists\r
+ * @param string $name\r
+ * @return bool\r
+ */\r
+ public function has($name)\r
+ {\r
+ return array_key_exists($name, $this->data);\r
+ }\r
+\r
+ /**\r
+ * Resets a value to the value of it's parent, usually the default. If\r
+ * no value is specified, the entire plist is reset.\r
+ * @param string $name\r
+ */\r
+ public function reset($name = null)\r
+ {\r
+ if ($name == null) {\r
+ $this->data = array();\r
+ } else {\r
+ unset($this->data[$name]);\r
+ }\r
+ }\r
+\r
+ /**\r
+ * Squashes this property list and all of its property lists into a single\r
+ * array, and returns the array. This value is cached by default.\r
+ * @param bool $force If true, ignores the cache and regenerates the array.\r
+ * @return array\r
+ */\r
+ public function squash($force = false)\r
+ {\r
+ if ($this->cache !== null && !$force) {\r
+ return $this->cache;\r
+ }\r
+ if ($this->parent) {\r
+ return $this->cache = array_merge($this->parent->squash($force), $this->data);\r
+ } else {\r
+ return $this->cache = $this->data;\r
+ }\r
+ }\r
+\r
+ /**\r
+ * Returns the parent plist.\r
+ * @return HTMLPurifier_PropertyList\r
+ */\r
+ public function getParent()\r
+ {\r
+ return $this->parent;\r
+ }\r
+\r
+ /**\r
+ * Sets the parent plist.\r
+ * @param HTMLPurifier_PropertyList $plist Parent plist\r
+ */\r
+ public function setParent($plist)\r
+ {\r
+ $this->parent = $plist;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Property list iterator. Do not instantiate this class directly.\r
+ */\r
+class HTMLPurifier_PropertyListIterator extends FilterIterator\r
+{\r
+\r
+ /**\r
+ * @type int\r
+ */\r
+ protected $l;\r
+ /**\r
+ * @type string\r
+ */\r
+ protected $filter;\r
+\r
+ /**\r
+ * @param Iterator $iterator Array of data to iterate over\r
+ * @param string $filter Optional prefix to only allow values of\r
+ */\r
+ public function __construct(Iterator $iterator, $filter = null)\r
+ {\r
+ parent::__construct($iterator);\r
+ $this->l = strlen($filter);\r
+ $this->filter = $filter;\r
+ }\r
+\r
+ /**\r
+ * @return bool\r
+ */\r
+ public function accept()\r
+ {\r
+ $key = $this->getInnerIterator()->key();\r
+ if (strncmp($key, $this->filter, $this->l) !== 0) {\r
+ return false;\r
+ }\r
+ return true;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * A simple array-backed queue, based off of the classic Okasaki\r
+ * persistent amortized queue. The basic idea is to maintain two\r
+ * stacks: an input stack and an output stack. When the output\r
+ * stack runs out, reverse the input stack and use it as the output\r
+ * stack.\r
+ *\r
+ * We don't use the SPL implementation because it's only supported\r
+ * on PHP 5.3 and later.\r
+ *\r
+ * Exercise: Prove that push/pop on this queue take amortized O(1) time.\r
+ *\r
+ * Exercise: Extend this queue to be a deque, while preserving amortized\r
+ * O(1) time. Some care must be taken on rebalancing to avoid quadratic\r
+ * behaviour caused by repeatedly shuffling data from the input stack\r
+ * to the output stack and back.\r
+ */\r
+class HTMLPurifier_Queue {\r
+ private $input;\r
+ private $output;\r
+\r
+ public function __construct($input = array()) {\r
+ $this->input = $input;\r
+ $this->output = array();\r
+ }\r
+\r
+ /**\r
+ * Shifts an element off the front of the queue.\r
+ */\r
+ public function shift() {\r
+ if (empty($this->output)) {\r
+ $this->output = array_reverse($this->input);\r
+ $this->input = array();\r
+ }\r
+ if (empty($this->output)) {\r
+ return NULL;\r
+ }\r
+ return array_pop($this->output);\r
+ }\r
+\r
+ /**\r
+ * Pushes an element onto the front of the queue.\r
+ */\r
+ public function push($x) {\r
+ array_push($this->input, $x);\r
+ }\r
+\r
+ /**\r
+ * Checks if it's empty.\r
+ */\r
+ public function isEmpty() {\r
+ return empty($this->input) && empty($this->output);\r
+ }\r
+}\r
+
+\r
+\r
+/**\r
+ * Supertype for classes that define a strategy for modifying/purifying tokens.\r
+ *\r
+ * While HTMLPurifier's core purpose is fixing HTML into something proper,\r
+ * strategies provide plug points for extra configuration or even extra\r
+ * features, such as custom tags, custom parsing of text, etc.\r
+ */\r
+\r
+\r
+abstract class HTMLPurifier_Strategy\r
+{\r
+\r
+ /**\r
+ * Executes the strategy on the tokens.\r
+ *\r
+ * @param HTMLPurifier_Token[] $tokens Array of HTMLPurifier_Token objects to be operated on.\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return HTMLPurifier_Token[] Processed array of token objects.\r
+ */\r
+ abstract public function execute($tokens, $config, $context);\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * This is in almost every respect equivalent to an array except\r
+ * that it keeps track of which keys were accessed.\r
+ *\r
+ * @warning For the sake of backwards compatibility with early versions\r
+ * of PHP 5, you must not use the $hash[$key] syntax; if you do\r
+ * our version of offsetGet is never called.\r
+ */\r
+class HTMLPurifier_StringHash extends ArrayObject\r
+{\r
+ /**\r
+ * @type array\r
+ */\r
+ protected $accessed = array();\r
+\r
+ /**\r
+ * Retrieves a value, and logs the access.\r
+ * @param mixed $index\r
+ * @return mixed\r
+ */\r
+ public function offsetGet($index)\r
+ {\r
+ $this->accessed[$index] = true;\r
+ return parent::offsetGet($index);\r
+ }\r
+\r
+ /**\r
+ * Returns a lookup array of all array indexes that have been accessed.\r
+ * @return array in form array($index => true).\r
+ */\r
+ public function getAccessed()\r
+ {\r
+ return $this->accessed;\r
+ }\r
+\r
+ /**\r
+ * Resets the access array.\r
+ */\r
+ public function resetAccessed()\r
+ {\r
+ $this->accessed = array();\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Parses string hash files. File format is as such:\r
+ *\r
+ * DefaultKeyValue\r
+ * KEY: Value\r
+ * KEY2: Value2\r
+ * --MULTILINE-KEY--\r
+ * Multiline\r
+ * value.\r
+ *\r
+ * Which would output something similar to:\r
+ *\r
+ * array(\r
+ * 'ID' => 'DefaultKeyValue',\r
+ * 'KEY' => 'Value',\r
+ * 'KEY2' => 'Value2',\r
+ * 'MULTILINE-KEY' => "Multiline\nvalue.\n",\r
+ * )\r
+ *\r
+ * We use this as an easy to use file-format for configuration schema\r
+ * files, but the class itself is usage agnostic.\r
+ *\r
+ * You can use ---- to forcibly terminate parsing of a single string-hash;\r
+ * this marker is used in multi string-hashes to delimit boundaries.\r
+ */\r
+class HTMLPurifier_StringHashParser\r
+{\r
+\r
+ /**\r
+ * @type string\r
+ */\r
+ public $default = 'ID';\r
+\r
+ /**\r
+ * Parses a file that contains a single string-hash.\r
+ * @param string $file\r
+ * @return array\r
+ */\r
+ public function parseFile($file)\r
+ {\r
+ if (!file_exists($file)) {\r
+ return false;\r
+ }\r
+ $fh = fopen($file, 'r');\r
+ if (!$fh) {\r
+ return false;\r
+ }\r
+ $ret = $this->parseHandle($fh);\r
+ fclose($fh);\r
+ return $ret;\r
+ }\r
+\r
+ /**\r
+ * Parses a file that contains multiple string-hashes delimited by '----'\r
+ * @param string $file\r
+ * @return array\r
+ */\r
+ public function parseMultiFile($file)\r
+ {\r
+ if (!file_exists($file)) {\r
+ return false;\r
+ }\r
+ $ret = array();\r
+ $fh = fopen($file, 'r');\r
+ if (!$fh) {\r
+ return false;\r
+ }\r
+ while (!feof($fh)) {\r
+ $ret[] = $this->parseHandle($fh);\r
+ }\r
+ fclose($fh);\r
+ return $ret;\r
+ }\r
+\r
+ /**\r
+ * Internal parser that acepts a file handle.\r
+ * @note While it's possible to simulate in-memory parsing by using\r
+ * custom stream wrappers, if such a use-case arises we should\r
+ * factor out the file handle into its own class.\r
+ * @param resource $fh File handle with pointer at start of valid string-hash\r
+ * block.\r
+ * @return array\r
+ */\r
+ protected function parseHandle($fh)\r
+ {\r
+ $state = false;\r
+ $single = false;\r
+ $ret = array();\r
+ do {\r
+ $line = fgets($fh);\r
+ if ($line === false) {\r
+ break;\r
+ }\r
+ $line = rtrim($line, "\n\r");\r
+ if (!$state && $line === '') {\r
+ continue;\r
+ }\r
+ if ($line === '----') {\r
+ break;\r
+ }\r
+ if (strncmp('--#', $line, 3) === 0) {\r
+ // Comment\r
+ continue;\r
+ } elseif (strncmp('--', $line, 2) === 0) {\r
+ // Multiline declaration\r
+ $state = trim($line, '- ');\r
+ if (!isset($ret[$state])) {\r
+ $ret[$state] = '';\r
+ }\r
+ continue;\r
+ } elseif (!$state) {\r
+ $single = true;\r
+ if (strpos($line, ':') !== false) {\r
+ // Single-line declaration\r
+ list($state, $line) = explode(':', $line, 2);\r
+ $line = trim($line);\r
+ } else {\r
+ // Use default declaration\r
+ $state = $this->default;\r
+ }\r
+ }\r
+ if ($single) {\r
+ $ret[$state] = $line;\r
+ $single = false;\r
+ $state = false;\r
+ } else {\r
+ $ret[$state] .= "$line\n";\r
+ }\r
+ } while (!feof($fh));\r
+ return $ret;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Defines a mutation of an obsolete tag into a valid tag.\r
+ */\r
+abstract class HTMLPurifier_TagTransform\r
+{\r
+\r
+ /**\r
+ * Tag name to transform the tag to.\r
+ * @type string\r
+ */\r
+ public $transform_to;\r
+\r
+ /**\r
+ * Transforms the obsolete tag into the valid tag.\r
+ * @param HTMLPurifier_Token_Tag $tag Tag to be transformed.\r
+ * @param HTMLPurifier_Config $config Mandatory HTMLPurifier_Config object\r
+ * @param HTMLPurifier_Context $context Mandatory HTMLPurifier_Context object\r
+ */\r
+ abstract public function transform($tag, $config, $context);\r
+\r
+ /**\r
+ * Prepends CSS properties to the style attribute, creating the\r
+ * attribute if it doesn't exist.\r
+ * @warning Copied over from AttrTransform, be sure to keep in sync\r
+ * @param array $attr Attribute array to process (passed by reference)\r
+ * @param string $css CSS to prepend\r
+ */\r
+ protected function prependCSS(&$attr, $css)\r
+ {\r
+ $attr['style'] = isset($attr['style']) ? $attr['style'] : '';\r
+ $attr['style'] = $css . $attr['style'];\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Abstract base token class that all others inherit from.\r
+ */\r
+abstract class HTMLPurifier_Token\r
+{\r
+ /**\r
+ * Line number node was on in source document. Null if unknown.\r
+ * @type int\r
+ */\r
+ public $line;\r
+\r
+ /**\r
+ * Column of line node was on in source document. Null if unknown.\r
+ * @type int\r
+ */\r
+ public $col;\r
+\r
+ /**\r
+ * Lookup array of processing that this token is exempt from.\r
+ * Currently, valid values are "ValidateAttributes" and\r
+ * "MakeWellFormed_TagClosedError"\r
+ * @type array\r
+ */\r
+ public $armor = array();\r
+\r
+ /**\r
+ * Used during MakeWellFormed. See Note [Injector skips]\r
+ * @type\r
+ */\r
+ public $skip;\r
+\r
+ /**\r
+ * @type\r
+ */\r
+ public $rewind;\r
+\r
+ /**\r
+ * @type\r
+ */\r
+ public $carryover;\r
+\r
+ /**\r
+ * @param string $n\r
+ * @return null|string\r
+ */\r
+ public function __get($n)\r
+ {\r
+ if ($n === 'type') {\r
+ trigger_error('Deprecated type property called; use instanceof', E_USER_NOTICE);\r
+ switch (get_class($this)) {\r
+ case 'HTMLPurifier_Token_Start':\r
+ return 'start';\r
+ case 'HTMLPurifier_Token_Empty':\r
+ return 'empty';\r
+ case 'HTMLPurifier_Token_End':\r
+ return 'end';\r
+ case 'HTMLPurifier_Token_Text':\r
+ return 'text';\r
+ case 'HTMLPurifier_Token_Comment':\r
+ return 'comment';\r
+ default:\r
+ return null;\r
+ }\r
+ }\r
+ }\r
+\r
+ /**\r
+ * Sets the position of the token in the source document.\r
+ * @param int $l\r
+ * @param int $c\r
+ */\r
+ public function position($l = null, $c = null)\r
+ {\r
+ $this->line = $l;\r
+ $this->col = $c;\r
+ }\r
+\r
+ /**\r
+ * Convenience function for DirectLex settings line/col position.\r
+ * @param int $l\r
+ * @param int $c\r
+ */\r
+ public function rawPosition($l, $c)\r
+ {\r
+ if ($c === -1) {\r
+ $l++;\r
+ }\r
+ $this->line = $l;\r
+ $this->col = $c;\r
+ }\r
+\r
+ /**\r
+ * Converts a token into its corresponding node.\r
+ */\r
+ abstract public function toNode();\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Factory for token generation.\r
+ *\r
+ * @note Doing some benchmarking indicates that the new operator is much\r
+ * slower than the clone operator (even discounting the cost of the\r
+ * constructor). This class is for that optimization.\r
+ * Other then that, there's not much point as we don't\r
+ * maintain parallel HTMLPurifier_Token hierarchies (the main reason why\r
+ * you'd want to use an abstract factory).\r
+ * @todo Port DirectLex to use this\r
+ */\r
+class HTMLPurifier_TokenFactory\r
+{\r
+ // p stands for prototype\r
+\r
+ /**\r
+ * @type HTMLPurifier_Token_Start\r
+ */\r
+ private $p_start;\r
+\r
+ /**\r
+ * @type HTMLPurifier_Token_End\r
+ */\r
+ private $p_end;\r
+\r
+ /**\r
+ * @type HTMLPurifier_Token_Empty\r
+ */\r
+ private $p_empty;\r
+\r
+ /**\r
+ * @type HTMLPurifier_Token_Text\r
+ */\r
+ private $p_text;\r
+\r
+ /**\r
+ * @type HTMLPurifier_Token_Comment\r
+ */\r
+ private $p_comment;\r
+\r
+ /**\r
+ * Generates blank prototypes for cloning.\r
+ */\r
+ public function __construct()\r
+ {\r
+ $this->p_start = new HTMLPurifier_Token_Start('', array());\r
+ $this->p_end = new HTMLPurifier_Token_End('');\r
+ $this->p_empty = new HTMLPurifier_Token_Empty('', array());\r
+ $this->p_text = new HTMLPurifier_Token_Text('');\r
+ $this->p_comment = new HTMLPurifier_Token_Comment('');\r
+ }\r
+\r
+ /**\r
+ * Creates a HTMLPurifier_Token_Start.\r
+ * @param string $name Tag name\r
+ * @param array $attr Associative array of attributes\r
+ * @return HTMLPurifier_Token_Start Generated HTMLPurifier_Token_Start\r
+ */\r
+ public function createStart($name, $attr = array())\r
+ {\r
+ $p = clone $this->p_start;\r
+ $p->__construct($name, $attr);\r
+ return $p;\r
+ }\r
+\r
+ /**\r
+ * Creates a HTMLPurifier_Token_End.\r
+ * @param string $name Tag name\r
+ * @return HTMLPurifier_Token_End Generated HTMLPurifier_Token_End\r
+ */\r
+ public function createEnd($name)\r
+ {\r
+ $p = clone $this->p_end;\r
+ $p->__construct($name);\r
+ return $p;\r
+ }\r
+\r
+ /**\r
+ * Creates a HTMLPurifier_Token_Empty.\r
+ * @param string $name Tag name\r
+ * @param array $attr Associative array of attributes\r
+ * @return HTMLPurifier_Token_Empty Generated HTMLPurifier_Token_Empty\r
+ */\r
+ public function createEmpty($name, $attr = array())\r
+ {\r
+ $p = clone $this->p_empty;\r
+ $p->__construct($name, $attr);\r
+ return $p;\r
+ }\r
+\r
+ /**\r
+ * Creates a HTMLPurifier_Token_Text.\r
+ * @param string $data Data of text token\r
+ * @return HTMLPurifier_Token_Text Generated HTMLPurifier_Token_Text\r
+ */\r
+ public function createText($data)\r
+ {\r
+ $p = clone $this->p_text;\r
+ $p->__construct($data);\r
+ return $p;\r
+ }\r
+\r
+ /**\r
+ * Creates a HTMLPurifier_Token_Comment.\r
+ * @param string $data Data of comment token\r
+ * @return HTMLPurifier_Token_Comment Generated HTMLPurifier_Token_Comment\r
+ */\r
+ public function createComment($data)\r
+ {\r
+ $p = clone $this->p_comment;\r
+ $p->__construct($data);\r
+ return $p;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * HTML Purifier's internal representation of a URI.\r
+ * @note\r
+ * Internal data-structures are completely escaped. If the data needs\r
+ * to be used in a non-URI context (which is very unlikely), be sure\r
+ * to decode it first. The URI may not necessarily be well-formed until\r
+ * validate() is called.\r
+ */\r
+class HTMLPurifier_URI\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $scheme;\r
+\r
+ /**\r
+ * @type string\r
+ */\r
+ public $userinfo;\r
+\r
+ /**\r
+ * @type string\r
+ */\r
+ public $host;\r
+\r
+ /**\r
+ * @type int\r
+ */\r
+ public $port;\r
+\r
+ /**\r
+ * @type string\r
+ */\r
+ public $path;\r
+\r
+ /**\r
+ * @type string\r
+ */\r
+ public $query;\r
+\r
+ /**\r
+ * @type string\r
+ */\r
+ public $fragment;\r
+\r
+ /**\r
+ * @param string $scheme\r
+ * @param string $userinfo\r
+ * @param string $host\r
+ * @param int $port\r
+ * @param string $path\r
+ * @param string $query\r
+ * @param string $fragment\r
+ * @note Automatically normalizes scheme and port\r
+ */\r
+ public function __construct($scheme, $userinfo, $host, $port, $path, $query, $fragment)\r
+ {\r
+ $this->scheme = is_null($scheme) || ctype_lower($scheme) ? $scheme : strtolower($scheme);\r
+ $this->userinfo = $userinfo;\r
+ $this->host = $host;\r
+ $this->port = is_null($port) ? $port : (int)$port;\r
+ $this->path = $path;\r
+ $this->query = $query;\r
+ $this->fragment = $fragment;\r
+ }\r
+\r
+ /**\r
+ * Retrieves a scheme object corresponding to the URI's scheme/default\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return HTMLPurifier_URIScheme Scheme object appropriate for validating this URI\r
+ */\r
+ public function getSchemeObj($config, $context)\r
+ {\r
+ $registry = HTMLPurifier_URISchemeRegistry::instance();\r
+ if ($this->scheme !== null) {\r
+ $scheme_obj = $registry->getScheme($this->scheme, $config, $context);\r
+ if (!$scheme_obj) {\r
+ return false;\r
+ } // invalid scheme, clean it out\r
+ } else {\r
+ // no scheme: retrieve the default one\r
+ $def = $config->getDefinition('URI');\r
+ $scheme_obj = $def->getDefaultScheme($config, $context);\r
+ if (!$scheme_obj) {\r
+ if ($def->defaultScheme !== null) {\r
+ // something funky happened to the default scheme object\r
+ trigger_error(\r
+ 'Default scheme object "' . $def->defaultScheme . '" was not readable',\r
+ E_USER_WARNING\r
+ );\r
+ } // suppress error if it's null\r
+ return false;\r
+ }\r
+ }\r
+ return $scheme_obj;\r
+ }\r
+\r
+ /**\r
+ * Generic validation method applicable for all schemes. May modify\r
+ * this URI in order to get it into a compliant form.\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool True if validation/filtering succeeds, false if failure\r
+ */\r
+ public function validate($config, $context)\r
+ {\r
+ // ABNF definitions from RFC 3986\r
+ $chars_sub_delims = '!$&\'()*+,;=';\r
+ $chars_gen_delims = ':/?#[]@';\r
+ $chars_pchar = $chars_sub_delims . ':@';\r
+\r
+ // validate host\r
+ if (!is_null($this->host)) {\r
+ $host_def = new HTMLPurifier_AttrDef_URI_Host();\r
+ $this->host = $host_def->validate($this->host, $config, $context);\r
+ if ($this->host === false) {\r
+ $this->host = null;\r
+ }\r
+ }\r
+\r
+ // validate scheme\r
+ // NOTE: It's not appropriate to check whether or not this\r
+ // scheme is in our registry, since a URIFilter may convert a\r
+ // URI that we don't allow into one we do. So instead, we just\r
+ // check if the scheme can be dropped because there is no host\r
+ // and it is our default scheme.\r
+ if (!is_null($this->scheme) && is_null($this->host) || $this->host === '') {\r
+ // support for relative paths is pretty abysmal when the\r
+ // scheme is present, so axe it when possible\r
+ $def = $config->getDefinition('URI');\r
+ if ($def->defaultScheme === $this->scheme) {\r
+ $this->scheme = null;\r
+ }\r
+ }\r
+\r
+ // validate username\r
+ if (!is_null($this->userinfo)) {\r
+ $encoder = new HTMLPurifier_PercentEncoder($chars_sub_delims . ':');\r
+ $this->userinfo = $encoder->encode($this->userinfo);\r
+ }\r
+\r
+ // validate port\r
+ if (!is_null($this->port)) {\r
+ if ($this->port < 1 || $this->port > 65535) {\r
+ $this->port = null;\r
+ }\r
+ }\r
+\r
+ // validate path\r
+ $segments_encoder = new HTMLPurifier_PercentEncoder($chars_pchar . '/');\r
+ if (!is_null($this->host)) { // this catches $this->host === ''\r
+ // path-abempty (hier and relative)\r
+ // http://www.example.com/my/path\r
+ // //www.example.com/my/path (looks odd, but works, and\r
+ // recognized by most browsers)\r
+ // (this set is valid or invalid on a scheme by scheme\r
+ // basis, so we'll deal with it later)\r
+ // file:///my/path\r
+ // ///my/path\r
+ $this->path = $segments_encoder->encode($this->path);\r
+ } elseif ($this->path !== '') {\r
+ if ($this->path[0] === '/') {\r
+ // path-absolute (hier and relative)\r
+ // http:/my/path\r
+ // /my/path\r
+ if (strlen($this->path) >= 2 && $this->path[1] === '/') {\r
+ // This could happen if both the host gets stripped\r
+ // out\r
+ // http://my/path\r
+ // //my/path\r
+ $this->path = '';\r
+ } else {\r
+ $this->path = $segments_encoder->encode($this->path);\r
+ }\r
+ } elseif (!is_null($this->scheme)) {\r
+ // path-rootless (hier)\r
+ // http:my/path\r
+ // Short circuit evaluation means we don't need to check nz\r
+ $this->path = $segments_encoder->encode($this->path);\r
+ } else {\r
+ // path-noscheme (relative)\r
+ // my/path\r
+ // (once again, not checking nz)\r
+ $segment_nc_encoder = new HTMLPurifier_PercentEncoder($chars_sub_delims . '@');\r
+ $c = strpos($this->path, '/');\r
+ if ($c !== false) {\r
+ $this->path =\r
+ $segment_nc_encoder->encode(substr($this->path, 0, $c)) .\r
+ $segments_encoder->encode(substr($this->path, $c));\r
+ } else {\r
+ $this->path = $segment_nc_encoder->encode($this->path);\r
+ }\r
+ }\r
+ } else {\r
+ // path-empty (hier and relative)\r
+ $this->path = ''; // just to be safe\r
+ }\r
+\r
+ // qf = query and fragment\r
+ $qf_encoder = new HTMLPurifier_PercentEncoder($chars_pchar . '/?');\r
+\r
+ if (!is_null($this->query)) {\r
+ $this->query = $qf_encoder->encode($this->query);\r
+ }\r
+\r
+ if (!is_null($this->fragment)) {\r
+ $this->fragment = $qf_encoder->encode($this->fragment);\r
+ }\r
+ return true;\r
+ }\r
+\r
+ /**\r
+ * Convert URI back to string\r
+ * @return string URI appropriate for output\r
+ */\r
+ public function toString()\r
+ {\r
+ // reconstruct authority\r
+ $authority = null;\r
+ // there is a rendering difference between a null authority\r
+ // (http:foo-bar) and an empty string authority\r
+ // (http:///foo-bar).\r
+ if (!is_null($this->host)) {\r
+ $authority = '';\r
+ if (!is_null($this->userinfo)) {\r
+ $authority .= $this->userinfo . '@';\r
+ }\r
+ $authority .= $this->host;\r
+ if (!is_null($this->port)) {\r
+ $authority .= ':' . $this->port;\r
+ }\r
+ }\r
+\r
+ // Reconstruct the result\r
+ // One might wonder about parsing quirks from browsers after\r
+ // this reconstruction. Unfortunately, parsing behavior depends\r
+ // on what *scheme* was employed (file:///foo is handled *very*\r
+ // differently than http:///foo), so unfortunately we have to\r
+ // defer to the schemes to do the right thing.\r
+ $result = '';\r
+ if (!is_null($this->scheme)) {\r
+ $result .= $this->scheme . ':';\r
+ }\r
+ if (!is_null($authority)) {\r
+ $result .= '//' . $authority;\r
+ }\r
+ $result .= $this->path;\r
+ if (!is_null($this->query)) {\r
+ $result .= '?' . $this->query;\r
+ }\r
+ if (!is_null($this->fragment)) {\r
+ $result .= '#' . $this->fragment;\r
+ }\r
+\r
+ return $result;\r
+ }\r
+\r
+ /**\r
+ * Returns true if this URL might be considered a 'local' URL given\r
+ * the current context. This is true when the host is null, or\r
+ * when it matches the host supplied to the configuration.\r
+ *\r
+ * Note that this does not do any scheme checking, so it is mostly\r
+ * only appropriate for metadata that doesn't care about protocol\r
+ * security. isBenign is probably what you actually want.\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool\r
+ */\r
+ public function isLocal($config, $context)\r
+ {\r
+ if ($this->host === null) {\r
+ return true;\r
+ }\r
+ $uri_def = $config->getDefinition('URI');\r
+ if ($uri_def->host === $this->host) {\r
+ return true;\r
+ }\r
+ return false;\r
+ }\r
+\r
+ /**\r
+ * Returns true if this URL should be considered a 'benign' URL,\r
+ * that is:\r
+ *\r
+ * - It is a local URL (isLocal), and\r
+ * - It has a equal or better level of security\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool\r
+ */\r
+ public function isBenign($config, $context)\r
+ {\r
+ if (!$this->isLocal($config, $context)) {\r
+ return false;\r
+ }\r
+\r
+ $scheme_obj = $this->getSchemeObj($config, $context);\r
+ if (!$scheme_obj) {\r
+ return false;\r
+ } // conservative approach\r
+\r
+ $current_scheme_obj = $config->getDefinition('URI')->getDefaultScheme($config, $context);\r
+ if ($current_scheme_obj->secure) {\r
+ if (!$scheme_obj->secure) {\r
+ return false;\r
+ }\r
+ }\r
+ return true;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+class HTMLPurifier_URIDefinition extends HTMLPurifier_Definition\r
+{\r
+\r
+ public $type = 'URI';\r
+ protected $filters = array();\r
+ protected $postFilters = array();\r
+ protected $registeredFilters = array();\r
+\r
+ /**\r
+ * HTMLPurifier_URI object of the base specified at %URI.Base\r
+ */\r
+ public $base;\r
+\r
+ /**\r
+ * String host to consider "home" base, derived off of $base\r
+ */\r
+ public $host;\r
+\r
+ /**\r
+ * Name of default scheme based on %URI.DefaultScheme and %URI.Base\r
+ */\r
+ public $defaultScheme;\r
+\r
+ public function __construct()\r
+ {\r
+ $this->registerFilter(new HTMLPurifier_URIFilter_DisableExternal());\r
+ $this->registerFilter(new HTMLPurifier_URIFilter_DisableExternalResources());\r
+ $this->registerFilter(new HTMLPurifier_URIFilter_DisableResources());\r
+ $this->registerFilter(new HTMLPurifier_URIFilter_HostBlacklist());\r
+ $this->registerFilter(new HTMLPurifier_URIFilter_SafeIframe());\r
+ $this->registerFilter(new HTMLPurifier_URIFilter_MakeAbsolute());\r
+ $this->registerFilter(new HTMLPurifier_URIFilter_Munge());\r
+ }\r
+\r
+ public function registerFilter($filter)\r
+ {\r
+ $this->registeredFilters[$filter->name] = $filter;\r
+ }\r
+\r
+ public function addFilter($filter, $config)\r
+ {\r
+ $r = $filter->prepare($config);\r
+ if ($r === false) return; // null is ok, for backwards compat\r
+ if ($filter->post) {\r
+ $this->postFilters[$filter->name] = $filter;\r
+ } else {\r
+ $this->filters[$filter->name] = $filter;\r
+ }\r
+ }\r
+\r
+ protected function doSetup($config)\r
+ {\r
+ $this->setupMemberVariables($config);\r
+ $this->setupFilters($config);\r
+ }\r
+\r
+ protected function setupFilters($config)\r
+ {\r
+ foreach ($this->registeredFilters as $name => $filter) {\r
+ if ($filter->always_load) {\r
+ $this->addFilter($filter, $config);\r
+ } else {\r
+ $conf = $config->get('URI.' . $name);\r
+ if ($conf !== false && $conf !== null) {\r
+ $this->addFilter($filter, $config);\r
+ }\r
+ }\r
+ }\r
+ unset($this->registeredFilters);\r
+ }\r
+\r
+ protected function setupMemberVariables($config)\r
+ {\r
+ $this->host = $config->get('URI.Host');\r
+ $base_uri = $config->get('URI.Base');\r
+ if (!is_null($base_uri)) {\r
+ $parser = new HTMLPurifier_URIParser();\r
+ $this->base = $parser->parse($base_uri);\r
+ $this->defaultScheme = $this->base->scheme;\r
+ if (is_null($this->host)) $this->host = $this->base->host;\r
+ }\r
+ if (is_null($this->defaultScheme)) $this->defaultScheme = $config->get('URI.DefaultScheme');\r
+ }\r
+\r
+ public function getDefaultScheme($config, $context)\r
+ {\r
+ return HTMLPurifier_URISchemeRegistry::instance()->getScheme($this->defaultScheme, $config, $context);\r
+ }\r
+\r
+ public function filter(&$uri, $config, $context)\r
+ {\r
+ foreach ($this->filters as $name => $f) {\r
+ $result = $f->filter($uri, $config, $context);\r
+ if (!$result) return false;\r
+ }\r
+ return true;\r
+ }\r
+\r
+ public function postFilter(&$uri, $config, $context)\r
+ {\r
+ foreach ($this->postFilters as $name => $f) {\r
+ $result = $f->filter($uri, $config, $context);\r
+ if (!$result) return false;\r
+ }\r
+ return true;\r
+ }\r
+\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Chainable filters for custom URI processing.\r
+ *\r
+ * These filters can perform custom actions on a URI filter object,\r
+ * including transformation or blacklisting. A filter named Foo\r
+ * must have a corresponding configuration directive %URI.Foo,\r
+ * unless always_load is specified to be true.\r
+ *\r
+ * The following contexts may be available while URIFilters are being\r
+ * processed:\r
+ *\r
+ * - EmbeddedURI: true if URI is an embedded resource that will\r
+ * be loaded automatically on page load\r
+ * - CurrentToken: a reference to the token that is currently\r
+ * being processed\r
+ * - CurrentAttr: the name of the attribute that is currently being\r
+ * processed\r
+ * - CurrentCSSProperty: the name of the CSS property that is\r
+ * currently being processed (if applicable)\r
+ *\r
+ * @warning This filter is called before scheme object validation occurs.\r
+ * Make sure, if you require a specific scheme object, you\r
+ * you check that it exists. This allows filters to convert\r
+ * proprietary URI schemes into regular ones.\r
+ */\r
+abstract class HTMLPurifier_URIFilter\r
+{\r
+\r
+ /**\r
+ * Unique identifier of filter.\r
+ * @type string\r
+ */\r
+ public $name;\r
+\r
+ /**\r
+ * True if this filter should be run after scheme validation.\r
+ * @type bool\r
+ */\r
+ public $post = false;\r
+\r
+ /**\r
+ * True if this filter should always be loaded.\r
+ * This permits a filter to be named Foo without the corresponding\r
+ * %URI.Foo directive existing.\r
+ * @type bool\r
+ */\r
+ public $always_load = false;\r
+\r
+ /**\r
+ * Performs initialization for the filter. If the filter returns\r
+ * false, this means that it shouldn't be considered active.\r
+ * @param HTMLPurifier_Config $config\r
+ * @return bool\r
+ */\r
+ public function prepare($config)\r
+ {\r
+ return true;\r
+ }\r
+\r
+ /**\r
+ * Filter a URI object\r
+ * @param HTMLPurifier_URI $uri Reference to URI object variable\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool Whether or not to continue processing: false indicates\r
+ * URL is no good, true indicates continue processing. Note that\r
+ * all changes are committed directly on the URI object\r
+ */\r
+ abstract public function filter(&$uri, $config, $context);\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Parses a URI into the components and fragment identifier as specified\r
+ * by RFC 3986.\r
+ */\r
+class HTMLPurifier_URIParser\r
+{\r
+\r
+ /**\r
+ * Instance of HTMLPurifier_PercentEncoder to do normalization with.\r
+ */\r
+ protected $percentEncoder;\r
+\r
+ public function __construct()\r
+ {\r
+ $this->percentEncoder = new HTMLPurifier_PercentEncoder();\r
+ }\r
+\r
+ /**\r
+ * Parses a URI.\r
+ * @param $uri string URI to parse\r
+ * @return HTMLPurifier_URI representation of URI. This representation has\r
+ * not been validated yet and may not conform to RFC.\r
+ */\r
+ public function parse($uri)\r
+ {\r
+ $uri = $this->percentEncoder->normalize($uri);\r
+\r
+ // Regexp is as per Appendix B.\r
+ // Note that ["<>] are an addition to the RFC's recommended\r
+ // characters, because they represent external delimeters.\r
+ $r_URI = '!'.\r
+ '(([a-zA-Z0-9\.\+\-]+):)?'. // 2. Scheme\r
+ '(//([^/?#"<>]*))?'. // 4. Authority\r
+ '([^?#"<>]*)'. // 5. Path\r
+ '(\?([^#"<>]*))?'. // 7. Query\r
+ '(#([^"<>]*))?'. // 8. Fragment\r
+ '!';\r
+\r
+ $matches = array();\r
+ $result = preg_match($r_URI, $uri, $matches);\r
+\r
+ if (!$result) return false; // *really* invalid URI\r
+\r
+ // seperate out parts\r
+ $scheme = !empty($matches[1]) ? $matches[2] : null;\r
+ $authority = !empty($matches[3]) ? $matches[4] : null;\r
+ $path = $matches[5]; // always present, can be empty\r
+ $query = !empty($matches[6]) ? $matches[7] : null;\r
+ $fragment = !empty($matches[8]) ? $matches[9] : null;\r
+\r
+ // further parse authority\r
+ if ($authority !== null) {\r
+ $r_authority = "/^((.+?)@)?(\[[^\]]+\]|[^:]*)(:(\d*))?/";\r
+ $matches = array();\r
+ preg_match($r_authority, $authority, $matches);\r
+ $userinfo = !empty($matches[1]) ? $matches[2] : null;\r
+ $host = !empty($matches[3]) ? $matches[3] : '';\r
+ $port = !empty($matches[4]) ? (int) $matches[5] : null;\r
+ } else {\r
+ $port = $host = $userinfo = null;\r
+ }\r
+\r
+ return new HTMLPurifier_URI(\r
+ $scheme, $userinfo, $host, $port, $path, $query, $fragment);\r
+ }\r
+\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Validator for the components of a URI for a specific scheme\r
+ */\r
+abstract class HTMLPurifier_URIScheme\r
+{\r
+\r
+ /**\r
+ * Scheme's default port (integer). If an explicit port number is\r
+ * specified that coincides with the default port, it will be\r
+ * elided.\r
+ * @type int\r
+ */\r
+ public $default_port = null;\r
+\r
+ /**\r
+ * Whether or not URIs of this scheme are locatable by a browser\r
+ * http and ftp are accessible, while mailto and news are not.\r
+ * @type bool\r
+ */\r
+ public $browsable = false;\r
+\r
+ /**\r
+ * Whether or not data transmitted over this scheme is encrypted.\r
+ * https is secure, http is not.\r
+ * @type bool\r
+ */\r
+ public $secure = false;\r
+\r
+ /**\r
+ * Whether or not the URI always uses <hier_part>, resolves edge cases\r
+ * with making relative URIs absolute\r
+ * @type bool\r
+ */\r
+ public $hierarchical = false;\r
+\r
+ /**\r
+ * Whether or not the URI may omit a hostname when the scheme is\r
+ * explicitly specified, ala file:///path/to/file. As of writing,\r
+ * 'file' is the only scheme that browsers support his properly.\r
+ * @type bool\r
+ */\r
+ public $may_omit_host = false;\r
+\r
+ /**\r
+ * Validates the components of a URI for a specific scheme.\r
+ * @param HTMLPurifier_URI $uri Reference to a HTMLPurifier_URI object\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool success or failure\r
+ */\r
+ abstract public function doValidate(&$uri, $config, $context);\r
+\r
+ /**\r
+ * Public interface for validating components of a URI. Performs a\r
+ * bunch of default actions. Don't overload this method.\r
+ * @param HTMLPurifier_URI $uri Reference to a HTMLPurifier_URI object\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool success or failure\r
+ */\r
+ public function validate(&$uri, $config, $context)\r
+ {\r
+ if ($this->default_port == $uri->port) {\r
+ $uri->port = null;\r
+ }\r
+ // kludge: browsers do funny things when the scheme but not the\r
+ // authority is set\r
+ if (!$this->may_omit_host &&\r
+ // if the scheme is present, a missing host is always in error\r
+ (!is_null($uri->scheme) && ($uri->host === '' || is_null($uri->host))) ||\r
+ // if the scheme is not present, a *blank* host is in error,\r
+ // since this translates into '///path' which most browsers\r
+ // interpret as being 'http://path'.\r
+ (is_null($uri->scheme) && $uri->host === '')\r
+ ) {\r
+ do {\r
+ if (is_null($uri->scheme)) {\r
+ if (substr($uri->path, 0, 2) != '//') {\r
+ $uri->host = null;\r
+ break;\r
+ }\r
+ // URI is '////path', so we cannot nullify the\r
+ // host to preserve semantics. Try expanding the\r
+ // hostname instead (fall through)\r
+ }\r
+ // first see if we can manually insert a hostname\r
+ $host = $config->get('URI.Host');\r
+ if (!is_null($host)) {\r
+ $uri->host = $host;\r
+ } else {\r
+ // we can't do anything sensible, reject the URL.\r
+ return false;\r
+ }\r
+ } while (false);\r
+ }\r
+ return $this->doValidate($uri, $config, $context);\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Registry for retrieving specific URI scheme validator objects.\r
+ */\r
+class HTMLPurifier_URISchemeRegistry\r
+{\r
+\r
+ /**\r
+ * Retrieve sole instance of the registry.\r
+ * @param HTMLPurifier_URISchemeRegistry $prototype Optional prototype to overload sole instance with,\r
+ * or bool true to reset to default registry.\r
+ * @return HTMLPurifier_URISchemeRegistry\r
+ * @note Pass a registry object $prototype with a compatible interface and\r
+ * the function will copy it and return it all further times.\r
+ */\r
+ public static function instance($prototype = null)\r
+ {\r
+ static $instance = null;\r
+ if ($prototype !== null) {\r
+ $instance = $prototype;\r
+ } elseif ($instance === null || $prototype == true) {\r
+ $instance = new HTMLPurifier_URISchemeRegistry();\r
+ }\r
+ return $instance;\r
+ }\r
+\r
+ /**\r
+ * Cache of retrieved schemes.\r
+ * @type HTMLPurifier_URIScheme[]\r
+ */\r
+ protected $schemes = array();\r
+\r
+ /**\r
+ * Retrieves a scheme validator object\r
+ * @param string $scheme String scheme name like http or mailto\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return HTMLPurifier_URIScheme\r
+ */\r
+ public function getScheme($scheme, $config, $context)\r
+ {\r
+ if (!$config) {\r
+ $config = HTMLPurifier_Config::createDefault();\r
+ }\r
+\r
+ // important, otherwise attacker could include arbitrary file\r
+ $allowed_schemes = $config->get('URI.AllowedSchemes');\r
+ if (!$config->get('URI.OverrideAllowedSchemes') &&\r
+ !isset($allowed_schemes[$scheme])\r
+ ) {\r
+ return;\r
+ }\r
+\r
+ if (isset($this->schemes[$scheme])) {\r
+ return $this->schemes[$scheme];\r
+ }\r
+ if (!isset($allowed_schemes[$scheme])) {\r
+ return;\r
+ }\r
+\r
+ $class = 'HTMLPurifier_URIScheme_' . $scheme;\r
+ if (!class_exists($class)) {\r
+ return;\r
+ }\r
+ $this->schemes[$scheme] = new $class();\r
+ return $this->schemes[$scheme];\r
+ }\r
+\r
+ /**\r
+ * Registers a custom scheme to the cache, bypassing reflection.\r
+ * @param string $scheme Scheme name\r
+ * @param HTMLPurifier_URIScheme $scheme_obj\r
+ */\r
+ public function register($scheme, $scheme_obj)\r
+ {\r
+ $this->schemes[$scheme] = $scheme_obj;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Class for converting between different unit-lengths as specified by\r
+ * CSS.\r
+ */\r
+class HTMLPurifier_UnitConverter\r
+{\r
+\r
+ const ENGLISH = 1;\r
+ const METRIC = 2;\r
+ const DIGITAL = 3;\r
+\r
+ /**\r
+ * Units information array. Units are grouped into measuring systems\r
+ * (English, Metric), and are assigned an integer representing\r
+ * the conversion factor between that unit and the smallest unit in\r
+ * the system. Numeric indexes are actually magical constants that\r
+ * encode conversion data from one system to the next, with a O(n^2)\r
+ * constraint on memory (this is generally not a problem, since\r
+ * the number of measuring systems is small.)\r
+ */\r
+ protected static $units = array(\r
+ self::ENGLISH => array(\r
+ 'px' => 3, // This is as per CSS 2.1 and Firefox. Your mileage may vary\r
+ 'pt' => 4,\r
+ 'pc' => 48,\r
+ 'in' => 288,\r
+ self::METRIC => array('pt', '0.352777778', 'mm'),\r
+ ),\r
+ self::METRIC => array(\r
+ 'mm' => 1,\r
+ 'cm' => 10,\r
+ self::ENGLISH => array('mm', '2.83464567', 'pt'),\r
+ ),\r
+ );\r
+\r
+ /**\r
+ * Minimum bcmath precision for output.\r
+ * @type int\r
+ */\r
+ protected $outputPrecision;\r
+\r
+ /**\r
+ * Bcmath precision for internal calculations.\r
+ * @type int\r
+ */\r
+ protected $internalPrecision;\r
+\r
+ /**\r
+ * Whether or not BCMath is available.\r
+ * @type bool\r
+ */\r
+ private $bcmath;\r
+\r
+ public function __construct($output_precision = 4, $internal_precision = 10, $force_no_bcmath = false)\r
+ {\r
+ $this->outputPrecision = $output_precision;\r
+ $this->internalPrecision = $internal_precision;\r
+ $this->bcmath = !$force_no_bcmath && function_exists('bcmul');\r
+ }\r
+\r
+ /**\r
+ * Converts a length object of one unit into another unit.\r
+ * @param HTMLPurifier_Length $length\r
+ * Instance of HTMLPurifier_Length to convert. You must validate()\r
+ * it before passing it here!\r
+ * @param string $to_unit\r
+ * Unit to convert to.\r
+ * @return HTMLPurifier_Length|bool\r
+ * @note\r
+ * About precision: This conversion function pays very special\r
+ * attention to the incoming precision of values and attempts\r
+ * to maintain a number of significant figure. Results are\r
+ * fairly accurate up to nine digits. Some caveats:\r
+ * - If a number is zero-padded as a result of this significant\r
+ * figure tracking, the zeroes will be eliminated.\r
+ * - If a number contains less than four sigfigs ($outputPrecision)\r
+ * and this causes some decimals to be excluded, those\r
+ * decimals will be added on.\r
+ */\r
+ public function convert($length, $to_unit)\r
+ {\r
+ if (!$length->isValid()) {\r
+ return false;\r
+ }\r
+\r
+ $n = $length->getN();\r
+ $unit = $length->getUnit();\r
+\r
+ if ($n === '0' || $unit === false) {\r
+ return new HTMLPurifier_Length('0', false);\r
+ }\r
+\r
+ $state = $dest_state = false;\r
+ foreach (self::$units as $k => $x) {\r
+ if (isset($x[$unit])) {\r
+ $state = $k;\r
+ }\r
+ if (isset($x[$to_unit])) {\r
+ $dest_state = $k;\r
+ }\r
+ }\r
+ if (!$state || !$dest_state) {\r
+ return false;\r
+ }\r
+\r
+ // Some calculations about the initial precision of the number;\r
+ // this will be useful when we need to do final rounding.\r
+ $sigfigs = $this->getSigFigs($n);\r
+ if ($sigfigs < $this->outputPrecision) {\r
+ $sigfigs = $this->outputPrecision;\r
+ }\r
+\r
+ // BCMath's internal precision deals only with decimals. Use\r
+ // our default if the initial number has no decimals, or increase\r
+ // it by how ever many decimals, thus, the number of guard digits\r
+ // will always be greater than or equal to internalPrecision.\r
+ $log = (int)floor(log(abs($n), 10));\r
+ $cp = ($log < 0) ? $this->internalPrecision - $log : $this->internalPrecision; // internal precision\r
+\r
+ for ($i = 0; $i < 2; $i++) {\r
+\r
+ // Determine what unit IN THIS SYSTEM we need to convert to\r
+ if ($dest_state === $state) {\r
+ // Simple conversion\r
+ $dest_unit = $to_unit;\r
+ } else {\r
+ // Convert to the smallest unit, pending a system shift\r
+ $dest_unit = self::$units[$state][$dest_state][0];\r
+ }\r
+\r
+ // Do the conversion if necessary\r
+ if ($dest_unit !== $unit) {\r
+ $factor = $this->div(self::$units[$state][$unit], self::$units[$state][$dest_unit], $cp);\r
+ $n = $this->mul($n, $factor, $cp);\r
+ $unit = $dest_unit;\r
+ }\r
+\r
+ // Output was zero, so bail out early. Shouldn't ever happen.\r
+ if ($n === '') {\r
+ $n = '0';\r
+ $unit = $to_unit;\r
+ break;\r
+ }\r
+\r
+ // It was a simple conversion, so bail out\r
+ if ($dest_state === $state) {\r
+ break;\r
+ }\r
+\r
+ if ($i !== 0) {\r
+ // Conversion failed! Apparently, the system we forwarded\r
+ // to didn't have this unit. This should never happen!\r
+ return false;\r
+ }\r
+\r
+ // Pre-condition: $i == 0\r
+\r
+ // Perform conversion to next system of units\r
+ $n = $this->mul($n, self::$units[$state][$dest_state][1], $cp);\r
+ $unit = self::$units[$state][$dest_state][2];\r
+ $state = $dest_state;\r
+\r
+ // One more loop around to convert the unit in the new system.\r
+\r
+ }\r
+\r
+ // Post-condition: $unit == $to_unit\r
+ if ($unit !== $to_unit) {\r
+ return false;\r
+ }\r
+\r
+ // Useful for debugging:\r
+ //echo "<pre>n";\r
+ //echo "$n\nsigfigs = $sigfigs\nnew_log = $new_log\nlog = $log\nrp = $rp\n</pre>\n";\r
+\r
+ $n = $this->round($n, $sigfigs);\r
+ if (strpos($n, '.') !== false) {\r
+ $n = rtrim($n, '0');\r
+ }\r
+ $n = rtrim($n, '.');\r
+\r
+ return new HTMLPurifier_Length($n, $unit);\r
+ }\r
+\r
+ /**\r
+ * Returns the number of significant figures in a string number.\r
+ * @param string $n Decimal number\r
+ * @return int number of sigfigs\r
+ */\r
+ public function getSigFigs($n)\r
+ {\r
+ $n = ltrim($n, '0+-');\r
+ $dp = strpos($n, '.'); // decimal position\r
+ if ($dp === false) {\r
+ $sigfigs = strlen(rtrim($n, '0'));\r
+ } else {\r
+ $sigfigs = strlen(ltrim($n, '0.')); // eliminate extra decimal character\r
+ if ($dp !== 0) {\r
+ $sigfigs--;\r
+ }\r
+ }\r
+ return $sigfigs;\r
+ }\r
+\r
+ /**\r
+ * Adds two numbers, using arbitrary precision when available.\r
+ * @param string $s1\r
+ * @param string $s2\r
+ * @param int $scale\r
+ * @return string\r
+ */\r
+ private function add($s1, $s2, $scale)\r
+ {\r
+ if ($this->bcmath) {\r
+ return bcadd($s1, $s2, $scale);\r
+ } else {\r
+ return $this->scale((float)$s1 + (float)$s2, $scale);\r
+ }\r
+ }\r
+\r
+ /**\r
+ * Multiples two numbers, using arbitrary precision when available.\r
+ * @param string $s1\r
+ * @param string $s2\r
+ * @param int $scale\r
+ * @return string\r
+ */\r
+ private function mul($s1, $s2, $scale)\r
+ {\r
+ if ($this->bcmath) {\r
+ return bcmul($s1, $s2, $scale);\r
+ } else {\r
+ return $this->scale((float)$s1 * (float)$s2, $scale);\r
+ }\r
+ }\r
+\r
+ /**\r
+ * Divides two numbers, using arbitrary precision when available.\r
+ * @param string $s1\r
+ * @param string $s2\r
+ * @param int $scale\r
+ * @return string\r
+ */\r
+ private function div($s1, $s2, $scale)\r
+ {\r
+ if ($this->bcmath) {\r
+ return bcdiv($s1, $s2, $scale);\r
+ } else {\r
+ return $this->scale((float)$s1 / (float)$s2, $scale);\r
+ }\r
+ }\r
+\r
+ /**\r
+ * Rounds a number according to the number of sigfigs it should have,\r
+ * using arbitrary precision when available.\r
+ * @param float $n\r
+ * @param int $sigfigs\r
+ * @return string\r
+ */\r
+ private function round($n, $sigfigs)\r
+ {\r
+ $new_log = (int)floor(log(abs($n), 10)); // Number of digits left of decimal - 1\r
+ $rp = $sigfigs - $new_log - 1; // Number of decimal places needed\r
+ $neg = $n < 0 ? '-' : ''; // Negative sign\r
+ if ($this->bcmath) {\r
+ if ($rp >= 0) {\r
+ $n = bcadd($n, $neg . '0.' . str_repeat('0', $rp) . '5', $rp + 1);\r
+ $n = bcdiv($n, '1', $rp);\r
+ } else {\r
+ // This algorithm partially depends on the standardized\r
+ // form of numbers that comes out of bcmath.\r
+ $n = bcadd($n, $neg . '5' . str_repeat('0', $new_log - $sigfigs), 0);\r
+ $n = substr($n, 0, $sigfigs + strlen($neg)) . str_repeat('0', $new_log - $sigfigs + 1);\r
+ }\r
+ return $n;\r
+ } else {\r
+ return $this->scale(round($n, $sigfigs - $new_log - 1), $rp + 1);\r
+ }\r
+ }\r
+\r
+ /**\r
+ * Scales a float to $scale digits right of decimal point, like BCMath.\r
+ * @param float $r\r
+ * @param int $scale\r
+ * @return string\r
+ */\r
+ private function scale($r, $scale)\r
+ {\r
+ if ($scale < 0) {\r
+ // The f sprintf type doesn't support negative numbers, so we\r
+ // need to cludge things manually. First get the string.\r
+ $r = sprintf('%.0f', (float)$r);\r
+ // Due to floating point precision loss, $r will more than likely\r
+ // look something like 4652999999999.9234. We grab one more digit\r
+ // than we need to precise from $r and then use that to round\r
+ // appropriately.\r
+ $precise = (string)round(substr($r, 0, strlen($r) + $scale), -1);\r
+ // Now we return it, truncating the zero that was rounded off.\r
+ return substr($precise, 0, -1) . str_repeat('0', -$scale + 1);\r
+ }\r
+ return sprintf('%.' . $scale . 'f', (float)$r);\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Parses string representations into their corresponding native PHP\r
+ * variable type. The base implementation does a simple type-check.\r
+ */\r
+class HTMLPurifier_VarParser\r
+{\r
+\r
+ const C_STRING = 1;\r
+ const ISTRING = 2;\r
+ const TEXT = 3;\r
+ const ITEXT = 4;\r
+ const C_INT = 5;\r
+ const C_FLOAT = 6;\r
+ const C_BOOL = 7;\r
+ const LOOKUP = 8;\r
+ const ALIST = 9;\r
+ const HASH = 10;\r
+ const C_MIXED = 11;\r
+\r
+ /**\r
+ * Lookup table of allowed types. Mainly for backwards compatibility, but\r
+ * also convenient for transforming string type names to the integer constants.\r
+ */\r
+ public static $types = array(\r
+ 'string' => self::C_STRING,\r
+ 'istring' => self::ISTRING,\r
+ 'text' => self::TEXT,\r
+ 'itext' => self::ITEXT,\r
+ 'int' => self::C_INT,\r
+ 'float' => self::C_FLOAT,\r
+ 'bool' => self::C_BOOL,\r
+ 'lookup' => self::LOOKUP,\r
+ 'list' => self::ALIST,\r
+ 'hash' => self::HASH,\r
+ 'mixed' => self::C_MIXED\r
+ );\r
+\r
+ /**\r
+ * Lookup table of types that are string, and can have aliases or\r
+ * allowed value lists.\r
+ */\r
+ public static $stringTypes = array(\r
+ self::C_STRING => true,\r
+ self::ISTRING => true,\r
+ self::TEXT => true,\r
+ self::ITEXT => true,\r
+ );\r
+\r
+ /**\r
+ * Validate a variable according to type.\r
+ * It may return NULL as a valid type if $allow_null is true.\r
+ *\r
+ * @param mixed $var Variable to validate\r
+ * @param int $type Type of variable, see HTMLPurifier_VarParser->types\r
+ * @param bool $allow_null Whether or not to permit null as a value\r
+ * @return string Validated and type-coerced variable\r
+ * @throws HTMLPurifier_VarParserException\r
+ */\r
+ final public function parse($var, $type, $allow_null = false)\r
+ {\r
+ if (is_string($type)) {\r
+ if (!isset(HTMLPurifier_VarParser::$types[$type])) {\r
+ throw new HTMLPurifier_VarParserException("Invalid type '$type'");\r
+ } else {\r
+ $type = HTMLPurifier_VarParser::$types[$type];\r
+ }\r
+ }\r
+ $var = $this->parseImplementation($var, $type, $allow_null);\r
+ if ($allow_null && $var === null) {\r
+ return null;\r
+ }\r
+ // These are basic checks, to make sure nothing horribly wrong\r
+ // happened in our implementations.\r
+ switch ($type) {\r
+ case (self::C_STRING):\r
+ case (self::ISTRING):\r
+ case (self::TEXT):\r
+ case (self::ITEXT):\r
+ if (!is_string($var)) {\r
+ break;\r
+ }\r
+ if ($type == self::ISTRING || $type == self::ITEXT) {\r
+ $var = strtolower($var);\r
+ }\r
+ return $var;\r
+ case (self::C_INT):\r
+ if (!is_int($var)) {\r
+ break;\r
+ }\r
+ return $var;\r
+ case (self::C_FLOAT):\r
+ if (!is_float($var)) {\r
+ break;\r
+ }\r
+ return $var;\r
+ case (self::C_BOOL):\r
+ if (!is_bool($var)) {\r
+ break;\r
+ }\r
+ return $var;\r
+ case (self::LOOKUP):\r
+ case (self::ALIST):\r
+ case (self::HASH):\r
+ if (!is_array($var)) {\r
+ break;\r
+ }\r
+ if ($type === self::LOOKUP) {\r
+ foreach ($var as $k) {\r
+ if ($k !== true) {\r
+ $this->error('Lookup table contains value other than true');\r
+ }\r
+ }\r
+ } elseif ($type === self::ALIST) {\r
+ $keys = array_keys($var);\r
+ if (array_keys($keys) !== $keys) {\r
+ $this->error('Indices for list are not uniform');\r
+ }\r
+ }\r
+ return $var;\r
+ case (self::C_MIXED):\r
+ return $var;\r
+ default:\r
+ $this->errorInconsistent(get_class($this), $type);\r
+ }\r
+ $this->errorGeneric($var, $type);\r
+ }\r
+\r
+ /**\r
+ * Actually implements the parsing. Base implementation does not\r
+ * do anything to $var. Subclasses should overload this!\r
+ * @param mixed $var\r
+ * @param int $type\r
+ * @param bool $allow_null\r
+ * @return string\r
+ */\r
+ protected function parseImplementation($var, $type, $allow_null)\r
+ {\r
+ return $var;\r
+ }\r
+\r
+ /**\r
+ * Throws an exception.\r
+ * @throws HTMLPurifier_VarParserException\r
+ */\r
+ protected function error($msg)\r
+ {\r
+ throw new HTMLPurifier_VarParserException($msg);\r
+ }\r
+\r
+ /**\r
+ * Throws an inconsistency exception.\r
+ * @note This should not ever be called. It would be called if we\r
+ * extend the allowed values of HTMLPurifier_VarParser without\r
+ * updating subclasses.\r
+ * @param string $class\r
+ * @param int $type\r
+ * @throws HTMLPurifier_Exception\r
+ */\r
+ protected function errorInconsistent($class, $type)\r
+ {\r
+ throw new HTMLPurifier_Exception(\r
+ "Inconsistency in $class: " . HTMLPurifier_VarParser::getTypeName($type) .\r
+ " not implemented"\r
+ );\r
+ }\r
+\r
+ /**\r
+ * Generic error for if a type didn't work.\r
+ * @param mixed $var\r
+ * @param int $type\r
+ */\r
+ protected function errorGeneric($var, $type)\r
+ {\r
+ $vtype = gettype($var);\r
+ $this->error("Expected type " . HTMLPurifier_VarParser::getTypeName($type) . ", got $vtype");\r
+ }\r
+\r
+ /**\r
+ * @param int $type\r
+ * @return string\r
+ */\r
+ public static function getTypeName($type)\r
+ {\r
+ static $lookup;\r
+ if (!$lookup) {\r
+ // Lazy load the alternative lookup table\r
+ $lookup = array_flip(HTMLPurifier_VarParser::$types);\r
+ }\r
+ if (!isset($lookup[$type])) {\r
+ return 'unknown';\r
+ }\r
+ return $lookup[$type];\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Exception type for HTMLPurifier_VarParser\r
+ */\r
+class HTMLPurifier_VarParserException extends HTMLPurifier_Exception\r
+{\r
+\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * A zipper is a purely-functional data structure which contains\r
+ * a focus that can be efficiently manipulated. It is known as\r
+ * a "one-hole context". This mutable variant implements a zipper\r
+ * for a list as a pair of two arrays, laid out as follows:\r
+ *\r
+ * Base list: 1 2 3 4 [ ] 6 7 8 9\r
+ * Front list: 1 2 3 4\r
+ * Back list: 9 8 7 6\r
+ *\r
+ * User is expected to keep track of the "current element" and properly\r
+ * fill it back in as necessary. (ToDo: Maybe it's more user friendly\r
+ * to implicitly track the current element?)\r
+ *\r
+ * Nota bene: the current class gets confused if you try to store NULLs\r
+ * in the list.\r
+ */\r
+\r
+class HTMLPurifier_Zipper\r
+{\r
+ public $front, $back;\r
+\r
+ public function __construct($front, $back) {\r
+ $this->front = $front;\r
+ $this->back = $back;\r
+ }\r
+\r
+ /**\r
+ * Creates a zipper from an array, with a hole in the\r
+ * 0-index position.\r
+ * @param Array to zipper-ify.\r
+ * @return Tuple of zipper and element of first position.\r
+ */\r
+ static public function fromArray($array) {\r
+ $z = new self(array(), array_reverse($array));\r
+ $t = $z->delete(); // delete the "dummy hole"\r
+ return array($z, $t);\r
+ }\r
+\r
+ /**\r
+ * Convert zipper back into a normal array, optionally filling in\r
+ * the hole with a value. (Usually you should supply a $t, unless you\r
+ * are at the end of the array.)\r
+ */\r
+ public function toArray($t = NULL) {\r
+ $a = $this->front;\r
+ if ($t !== NULL) $a[] = $t;\r
+ for ($i = count($this->back)-1; $i >= 0; $i--) {\r
+ $a[] = $this->back[$i];\r
+ }\r
+ return $a;\r
+ }\r
+\r
+ /**\r
+ * Move hole to the next element.\r
+ * @param $t Element to fill hole with\r
+ * @return Original contents of new hole.\r
+ */\r
+ public function next($t) {\r
+ if ($t !== NULL) array_push($this->front, $t);\r
+ return empty($this->back) ? NULL : array_pop($this->back);\r
+ }\r
+\r
+ /**\r
+ * Iterated hole advancement.\r
+ * @param $t Element to fill hole with\r
+ * @param $i How many forward to advance hole\r
+ * @return Original contents of new hole, i away\r
+ */\r
+ public function advance($t, $n) {\r
+ for ($i = 0; $i < $n; $i++) {\r
+ $t = $this->next($t);\r
+ }\r
+ return $t;\r
+ }\r
+\r
+ /**\r
+ * Move hole to the previous element\r
+ * @param $t Element to fill hole with\r
+ * @return Original contents of new hole.\r
+ */\r
+ public function prev($t) {\r
+ if ($t !== NULL) array_push($this->back, $t);\r
+ return empty($this->front) ? NULL : array_pop($this->front);\r
+ }\r
+\r
+ /**\r
+ * Delete contents of current hole, shifting hole to\r
+ * next element.\r
+ * @return Original contents of new hole.\r
+ */\r
+ public function delete() {\r
+ return empty($this->back) ? NULL : array_pop($this->back);\r
+ }\r
+\r
+ /**\r
+ * Returns true if we are at the end of the list.\r
+ * @return bool\r
+ */\r
+ public function done() {\r
+ return empty($this->back);\r
+ }\r
+\r
+ /**\r
+ * Insert element before hole.\r
+ * @param Element to insert\r
+ */\r
+ public function insertBefore($t) {\r
+ if ($t !== NULL) array_push($this->front, $t);\r
+ }\r
+\r
+ /**\r
+ * Insert element after hole.\r
+ * @param Element to insert\r
+ */\r
+ public function insertAfter($t) {\r
+ if ($t !== NULL) array_push($this->back, $t);\r
+ }\r
+\r
+ /**\r
+ * Splice in multiple elements at hole. Functional specification\r
+ * in terms of array_splice:\r
+ *\r
+ * $arr1 = $arr;\r
+ * $old1 = array_splice($arr1, $i, $delete, $replacement);\r
+ *\r
+ * list($z, $t) = HTMLPurifier_Zipper::fromArray($arr);\r
+ * $t = $z->advance($t, $i);\r
+ * list($old2, $t) = $z->splice($t, $delete, $replacement);\r
+ * $arr2 = $z->toArray($t);\r
+ *\r
+ * assert($old1 === $old2);\r
+ * assert($arr1 === $arr2);\r
+ *\r
+ * NB: the absolute index location after this operation is\r
+ * *unchanged!*\r
+ *\r
+ * @param Current contents of hole.\r
+ */\r
+ public function splice($t, $delete, $replacement) {\r
+ // delete\r
+ $old = array();\r
+ $r = $t;\r
+ for ($i = $delete; $i > 0; $i--) {\r
+ $old[] = $r;\r
+ $r = $this->delete();\r
+ }\r
+ // insert\r
+ for ($i = count($replacement)-1; $i >= 0; $i--) {\r
+ $this->insertAfter($r);\r
+ $r = $replacement[$i];\r
+ }\r
+ return array($old, $r);\r
+ }\r
+}\r
+
+\r
+\r
+/**\r
+ * Validates the HTML attribute style, otherwise known as CSS.\r
+ * @note We don't implement the whole CSS specification, so it might be\r
+ * difficult to reuse this component in the context of validating\r
+ * actual stylesheet declarations.\r
+ * @note If we were really serious about validating the CSS, we would\r
+ * tokenize the styles and then parse the tokens. Obviously, we\r
+ * are not doing that. Doing that could seriously harm performance,\r
+ * but would make these components a lot more viable for a CSS\r
+ * filtering solution.\r
+ */\r
+class HTMLPurifier_AttrDef_CSS extends HTMLPurifier_AttrDef\r
+{\r
+\r
+ /**\r
+ * @param string $css\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string\r
+ */\r
+ public function validate($css, $config, $context)\r
+ {\r
+ $css = $this->parseCDATA($css);\r
+\r
+ $definition = $config->getCSSDefinition();\r
+ $allow_duplicates = $config->get("CSS.AllowDuplicates");\r
+\r
+\r
+ // According to the CSS2.1 spec, the places where a\r
+ // non-delimiting semicolon can appear are in strings\r
+ // escape sequences. So here is some dumb hack to\r
+ // handle quotes.\r
+ $len = strlen($css);\r
+ $accum = "";\r
+ $declarations = array();\r
+ $quoted = false;\r
+ for ($i = 0; $i < $len; $i++) {\r
+ $c = strcspn($css, ";'\"", $i);\r
+ $accum .= substr($css, $i, $c);\r
+ $i += $c;\r
+ if ($i == $len) break;\r
+ $d = $css[$i];\r
+ if ($quoted) {\r
+ $accum .= $d;\r
+ if ($d == $quoted) {\r
+ $quoted = false;\r
+ }\r
+ } else {\r
+ if ($d == ";") {\r
+ $declarations[] = $accum;\r
+ $accum = "";\r
+ } else {\r
+ $accum .= $d;\r
+ $quoted = $d;\r
+ }\r
+ }\r
+ }\r
+ if ($accum != "") $declarations[] = $accum;\r
+\r
+ $propvalues = array();\r
+ $new_declarations = '';\r
+\r
+ /**\r
+ * Name of the current CSS property being validated.\r
+ */\r
+ $property = false;\r
+ $context->register('CurrentCSSProperty', $property);\r
+\r
+ foreach ($declarations as $declaration) {\r
+ if (!$declaration) {\r
+ continue;\r
+ }\r
+ if (!strpos($declaration, ':')) {\r
+ continue;\r
+ }\r
+ list($property, $value) = explode(':', $declaration, 2);\r
+ $property = trim($property);\r
+ $value = trim($value);\r
+ $ok = false;\r
+ do {\r
+ if (isset($definition->info[$property])) {\r
+ $ok = true;\r
+ break;\r
+ }\r
+ if (ctype_lower($property)) {\r
+ break;\r
+ }\r
+ $property = strtolower($property);\r
+ if (isset($definition->info[$property])) {\r
+ $ok = true;\r
+ break;\r
+ }\r
+ } while (0);\r
+ if (!$ok) {\r
+ continue;\r
+ }\r
+ // inefficient call, since the validator will do this again\r
+ if (strtolower(trim($value)) !== 'inherit') {\r
+ // inherit works for everything (but only on the base property)\r
+ $result = $definition->info[$property]->validate(\r
+ $value,\r
+ $config,\r
+ $context\r
+ );\r
+ } else {\r
+ $result = 'inherit';\r
+ }\r
+ if ($result === false) {\r
+ continue;\r
+ }\r
+ if ($allow_duplicates) {\r
+ $new_declarations .= "$property:$result;";\r
+ } else {\r
+ $propvalues[$property] = $result;\r
+ }\r
+ }\r
+\r
+ $context->destroy('CurrentCSSProperty');\r
+\r
+ // procedure does not write the new CSS simultaneously, so it's\r
+ // slightly inefficient, but it's the only way of getting rid of\r
+ // duplicates. Perhaps config to optimize it, but not now.\r
+\r
+ foreach ($propvalues as $prop => $value) {\r
+ $new_declarations .= "$prop:$value;";\r
+ }\r
+\r
+ return $new_declarations ? $new_declarations : false;\r
+\r
+ }\r
+\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Dummy AttrDef that mimics another AttrDef, BUT it generates clones\r
+ * with make.\r
+ */\r
+class HTMLPurifier_AttrDef_Clone extends HTMLPurifier_AttrDef\r
+{\r
+ /**\r
+ * What we're cloning.\r
+ * @type HTMLPurifier_AttrDef\r
+ */\r
+ protected $clone;\r
+\r
+ /**\r
+ * @param HTMLPurifier_AttrDef $clone\r
+ */\r
+ public function __construct($clone)\r
+ {\r
+ $this->clone = $clone;\r
+ }\r
+\r
+ /**\r
+ * @param string $v\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string\r
+ */\r
+ public function validate($v, $config, $context)\r
+ {\r
+ return $this->clone->validate($v, $config, $context);\r
+ }\r
+\r
+ /**\r
+ * @param string $string\r
+ * @return HTMLPurifier_AttrDef\r
+ */\r
+ public function make($string)\r
+ {\r
+ return clone $this->clone;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+// Enum = Enumerated\r
+/**\r
+ * Validates a keyword against a list of valid values.\r
+ * @warning The case-insensitive compare of this function uses PHP's\r
+ * built-in strtolower and ctype_lower functions, which may\r
+ * cause problems with international comparisons\r
+ */\r
+class HTMLPurifier_AttrDef_Enum extends HTMLPurifier_AttrDef\r
+{\r
+\r
+ /**\r
+ * Lookup table of valid values.\r
+ * @type array\r
+ * @todo Make protected\r
+ */\r
+ public $valid_values = array();\r
+\r
+ /**\r
+ * Bool indicating whether or not enumeration is case sensitive.\r
+ * @note In general this is always case insensitive.\r
+ */\r
+ protected $case_sensitive = false; // values according to W3C spec\r
+\r
+ /**\r
+ * @param array $valid_values List of valid values\r
+ * @param bool $case_sensitive Whether or not case sensitive\r
+ */\r
+ public function __construct($valid_values = array(), $case_sensitive = false)\r
+ {\r
+ $this->valid_values = array_flip($valid_values);\r
+ $this->case_sensitive = $case_sensitive;\r
+ }\r
+\r
+ /**\r
+ * @param string $string\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string\r
+ */\r
+ public function validate($string, $config, $context)\r
+ {\r
+ $string = trim($string);\r
+ if (!$this->case_sensitive) {\r
+ // we may want to do full case-insensitive libraries\r
+ $string = ctype_lower($string) ? $string : strtolower($string);\r
+ }\r
+ $result = isset($this->valid_values[$string]);\r
+\r
+ return $result ? $string : false;\r
+ }\r
+\r
+ /**\r
+ * @param string $string In form of comma-delimited list of case-insensitive\r
+ * valid values. Example: "foo,bar,baz". Prepend "s:" to make\r
+ * case sensitive\r
+ * @return HTMLPurifier_AttrDef_Enum\r
+ */\r
+ public function make($string)\r
+ {\r
+ if (strlen($string) > 2 && $string[0] == 's' && $string[1] == ':') {\r
+ $string = substr($string, 2);\r
+ $sensitive = true;\r
+ } else {\r
+ $sensitive = false;\r
+ }\r
+ $values = explode(',', $string);\r
+ return new HTMLPurifier_AttrDef_Enum($values, $sensitive);\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Validates an integer.\r
+ * @note While this class was modeled off the CSS definition, no currently\r
+ * allowed CSS uses this type. The properties that do are: widows,\r
+ * orphans, z-index, counter-increment, counter-reset. Some of the\r
+ * HTML attributes, however, find use for a non-negative version of this.\r
+ */\r
+class HTMLPurifier_AttrDef_Integer extends HTMLPurifier_AttrDef\r
+{\r
+\r
+ /**\r
+ * Whether or not negative values are allowed.\r
+ * @type bool\r
+ */\r
+ protected $negative = true;\r
+\r
+ /**\r
+ * Whether or not zero is allowed.\r
+ * @type bool\r
+ */\r
+ protected $zero = true;\r
+\r
+ /**\r
+ * Whether or not positive values are allowed.\r
+ * @type bool\r
+ */\r
+ protected $positive = true;\r
+\r
+ /**\r
+ * @param $negative Bool indicating whether or not negative values are allowed\r
+ * @param $zero Bool indicating whether or not zero is allowed\r
+ * @param $positive Bool indicating whether or not positive values are allowed\r
+ */\r
+ public function __construct($negative = true, $zero = true, $positive = true)\r
+ {\r
+ $this->negative = $negative;\r
+ $this->zero = $zero;\r
+ $this->positive = $positive;\r
+ }\r
+\r
+ /**\r
+ * @param string $integer\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string\r
+ */\r
+ public function validate($integer, $config, $context)\r
+ {\r
+ $integer = $this->parseCDATA($integer);\r
+ if ($integer === '') {\r
+ return false;\r
+ }\r
+\r
+ // we could possibly simply typecast it to integer, but there are\r
+ // certain fringe cases that must not return an integer.\r
+\r
+ // clip leading sign\r
+ if ($this->negative && $integer[0] === '-') {\r
+ $digits = substr($integer, 1);\r
+ if ($digits === '0') {\r
+ $integer = '0';\r
+ } // rm minus sign for zero\r
+ } elseif ($this->positive && $integer[0] === '+') {\r
+ $digits = $integer = substr($integer, 1); // rm unnecessary plus\r
+ } else {\r
+ $digits = $integer;\r
+ }\r
+\r
+ // test if it's numeric\r
+ if (!ctype_digit($digits)) {\r
+ return false;\r
+ }\r
+\r
+ // perform scope tests\r
+ if (!$this->zero && $integer == 0) {\r
+ return false;\r
+ }\r
+ if (!$this->positive && $integer > 0) {\r
+ return false;\r
+ }\r
+ if (!$this->negative && $integer < 0) {\r
+ return false;\r
+ }\r
+\r
+ return $integer;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Validates the HTML attribute lang, effectively a language code.\r
+ * @note Built according to RFC 3066, which obsoleted RFC 1766\r
+ */\r
+class HTMLPurifier_AttrDef_Lang extends HTMLPurifier_AttrDef\r
+{\r
+\r
+ /**\r
+ * @param string $string\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string\r
+ */\r
+ public function validate($string, $config, $context)\r
+ {\r
+ $string = trim($string);\r
+ if (!$string) {\r
+ return false;\r
+ }\r
+\r
+ $subtags = explode('-', $string);\r
+ $num_subtags = count($subtags);\r
+\r
+ if ($num_subtags == 0) { // sanity check\r
+ return false;\r
+ }\r
+\r
+ // process primary subtag : $subtags[0]\r
+ $length = strlen($subtags[0]);\r
+ switch ($length) {\r
+ case 0:\r
+ return false;\r
+ case 1:\r
+ if (!($subtags[0] == 'x' || $subtags[0] == 'i')) {\r
+ return false;\r
+ }\r
+ break;\r
+ case 2:\r
+ case 3:\r
+ if (!ctype_alpha($subtags[0])) {\r
+ return false;\r
+ } elseif (!ctype_lower($subtags[0])) {\r
+ $subtags[0] = strtolower($subtags[0]);\r
+ }\r
+ break;\r
+ default:\r
+ return false;\r
+ }\r
+\r
+ $new_string = $subtags[0];\r
+ if ($num_subtags == 1) {\r
+ return $new_string;\r
+ }\r
+\r
+ // process second subtag : $subtags[1]\r
+ $length = strlen($subtags[1]);\r
+ if ($length == 0 || ($length == 1 && $subtags[1] != 'x') || $length > 8 || !ctype_alnum($subtags[1])) {\r
+ return $new_string;\r
+ }\r
+ if (!ctype_lower($subtags[1])) {\r
+ $subtags[1] = strtolower($subtags[1]);\r
+ }\r
+\r
+ $new_string .= '-' . $subtags[1];\r
+ if ($num_subtags == 2) {\r
+ return $new_string;\r
+ }\r
+\r
+ // process all other subtags, index 2 and up\r
+ for ($i = 2; $i < $num_subtags; $i++) {\r
+ $length = strlen($subtags[$i]);\r
+ if ($length == 0 || $length > 8 || !ctype_alnum($subtags[$i])) {\r
+ return $new_string;\r
+ }\r
+ if (!ctype_lower($subtags[$i])) {\r
+ $subtags[$i] = strtolower($subtags[$i]);\r
+ }\r
+ $new_string .= '-' . $subtags[$i];\r
+ }\r
+ return $new_string;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Decorator that, depending on a token, switches between two definitions.\r
+ */\r
+class HTMLPurifier_AttrDef_Switch\r
+{\r
+\r
+ /**\r
+ * @type string\r
+ */\r
+ protected $tag;\r
+\r
+ /**\r
+ * @type HTMLPurifier_AttrDef\r
+ */\r
+ protected $withTag;\r
+\r
+ /**\r
+ * @type HTMLPurifier_AttrDef\r
+ */\r
+ protected $withoutTag;\r
+\r
+ /**\r
+ * @param string $tag Tag name to switch upon\r
+ * @param HTMLPurifier_AttrDef $with_tag Call if token matches tag\r
+ * @param HTMLPurifier_AttrDef $without_tag Call if token doesn't match, or there is no token\r
+ */\r
+ public function __construct($tag, $with_tag, $without_tag)\r
+ {\r
+ $this->tag = $tag;\r
+ $this->withTag = $with_tag;\r
+ $this->withoutTag = $without_tag;\r
+ }\r
+\r
+ /**\r
+ * @param string $string\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string\r
+ */\r
+ public function validate($string, $config, $context)\r
+ {\r
+ $token = $context->get('CurrentToken', true);\r
+ if (!$token || $token->name !== $this->tag) {\r
+ return $this->withoutTag->validate($string, $config, $context);\r
+ } else {\r
+ return $this->withTag->validate($string, $config, $context);\r
+ }\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Validates arbitrary text according to the HTML spec.\r
+ */\r
+class HTMLPurifier_AttrDef_Text extends HTMLPurifier_AttrDef\r
+{\r
+\r
+ /**\r
+ * @param string $string\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string\r
+ */\r
+ public function validate($string, $config, $context)\r
+ {\r
+ return $this->parseCDATA($string);\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Validates a URI as defined by RFC 3986.\r
+ * @note Scheme-specific mechanics deferred to HTMLPurifier_URIScheme\r
+ */\r
+class HTMLPurifier_AttrDef_URI extends HTMLPurifier_AttrDef\r
+{\r
+\r
+ /**\r
+ * @type HTMLPurifier_URIParser\r
+ */\r
+ protected $parser;\r
+\r
+ /**\r
+ * @type bool\r
+ */\r
+ protected $embedsResource;\r
+\r
+ /**\r
+ * @param bool $embeds_resource Does the URI here result in an extra HTTP request?\r
+ */\r
+ public function __construct($embeds_resource = false)\r
+ {\r
+ $this->parser = new HTMLPurifier_URIParser();\r
+ $this->embedsResource = (bool)$embeds_resource;\r
+ }\r
+\r
+ /**\r
+ * @param string $string\r
+ * @return HTMLPurifier_AttrDef_URI\r
+ */\r
+ public function make($string)\r
+ {\r
+ $embeds = ($string === 'embedded');\r
+ return new HTMLPurifier_AttrDef_URI($embeds);\r
+ }\r
+\r
+ /**\r
+ * @param string $uri\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string\r
+ */\r
+ public function validate($uri, $config, $context)\r
+ {\r
+ if ($config->get('URI.Disable')) {\r
+ return false;\r
+ }\r
+\r
+ $uri = $this->parseCDATA($uri);\r
+\r
+ // parse the URI\r
+ $uri = $this->parser->parse($uri);\r
+ if ($uri === false) {\r
+ return false;\r
+ }\r
+\r
+ // add embedded flag to context for validators\r
+ $context->register('EmbeddedURI', $this->embedsResource);\r
+\r
+ $ok = false;\r
+ do {\r
+\r
+ // generic validation\r
+ $result = $uri->validate($config, $context);\r
+ if (!$result) {\r
+ break;\r
+ }\r
+\r
+ // chained filtering\r
+ $uri_def = $config->getDefinition('URI');\r
+ $result = $uri_def->filter($uri, $config, $context);\r
+ if (!$result) {\r
+ break;\r
+ }\r
+\r
+ // scheme-specific validation\r
+ $scheme_obj = $uri->getSchemeObj($config, $context);\r
+ if (!$scheme_obj) {\r
+ break;\r
+ }\r
+ if ($this->embedsResource && !$scheme_obj->browsable) {\r
+ break;\r
+ }\r
+ $result = $scheme_obj->validate($uri, $config, $context);\r
+ if (!$result) {\r
+ break;\r
+ }\r
+\r
+ // Post chained filtering\r
+ $result = $uri_def->postFilter($uri, $config, $context);\r
+ if (!$result) {\r
+ break;\r
+ }\r
+\r
+ // survived gauntlet\r
+ $ok = true;\r
+\r
+ } while (false);\r
+\r
+ $context->destroy('EmbeddedURI');\r
+ if (!$ok) {\r
+ return false;\r
+ }\r
+ // back to string\r
+ return $uri->toString();\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Validates a number as defined by the CSS spec.\r
+ */\r
+class HTMLPurifier_AttrDef_CSS_Number extends HTMLPurifier_AttrDef\r
+{\r
+\r
+ /**\r
+ * Indicates whether or not only positive values are allowed.\r
+ * @type bool\r
+ */\r
+ protected $non_negative = false;\r
+\r
+ /**\r
+ * @param bool $non_negative indicates whether negatives are forbidden\r
+ */\r
+ public function __construct($non_negative = false)\r
+ {\r
+ $this->non_negative = $non_negative;\r
+ }\r
+\r
+ /**\r
+ * @param string $number\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return string|bool\r
+ * @warning Some contexts do not pass $config, $context. These\r
+ * variables should not be used without checking HTMLPurifier_Length\r
+ */\r
+ public function validate($number, $config, $context)\r
+ {\r
+ $number = $this->parseCDATA($number);\r
+\r
+ if ($number === '') {\r
+ return false;\r
+ }\r
+ if ($number === '0') {\r
+ return '0';\r
+ }\r
+\r
+ $sign = '';\r
+ switch ($number[0]) {\r
+ case '-':\r
+ if ($this->non_negative) {\r
+ return false;\r
+ }\r
+ $sign = '-';\r
+ case '+':\r
+ $number = substr($number, 1);\r
+ }\r
+\r
+ if (ctype_digit($number)) {\r
+ $number = ltrim($number, '0');\r
+ return $number ? $sign . $number : '0';\r
+ }\r
+\r
+ // Period is the only non-numeric character allowed\r
+ if (strpos($number, '.') === false) {\r
+ return false;\r
+ }\r
+\r
+ list($left, $right) = explode('.', $number, 2);\r
+\r
+ if ($left === '' && $right === '') {\r
+ return false;\r
+ }\r
+ if ($left !== '' && !ctype_digit($left)) {\r
+ return false;\r
+ }\r
+\r
+ $left = ltrim($left, '0');\r
+ $right = rtrim($right, '0');\r
+\r
+ if ($right === '') {\r
+ return $left ? $sign . $left : '0';\r
+ } elseif (!ctype_digit($right)) {\r
+ return false;\r
+ }\r
+ return $sign . $left . '.' . $right;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+class HTMLPurifier_AttrDef_CSS_AlphaValue extends HTMLPurifier_AttrDef_CSS_Number\r
+{\r
+\r
+ public function __construct()\r
+ {\r
+ parent::__construct(false); // opacity is non-negative, but we will clamp it\r
+ }\r
+\r
+ /**\r
+ * @param string $number\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return string\r
+ */\r
+ public function validate($number, $config, $context)\r
+ {\r
+ $result = parent::validate($number, $config, $context);\r
+ if ($result === false) {\r
+ return $result;\r
+ }\r
+ $float = (float)$result;\r
+ if ($float < 0.0) {\r
+ $result = '0';\r
+ }\r
+ if ($float > 1.0) {\r
+ $result = '1';\r
+ }\r
+ return $result;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Validates shorthand CSS property background.\r
+ * @warning Does not support url tokens that have internal spaces.\r
+ */\r
+class HTMLPurifier_AttrDef_CSS_Background extends HTMLPurifier_AttrDef\r
+{\r
+\r
+ /**\r
+ * Local copy of component validators.\r
+ * @type HTMLPurifier_AttrDef[]\r
+ * @note See HTMLPurifier_AttrDef_Font::$info for a similar impl.\r
+ */\r
+ protected $info;\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ public function __construct($config)\r
+ {\r
+ $def = $config->getCSSDefinition();\r
+ $this->info['background-color'] = $def->info['background-color'];\r
+ $this->info['background-image'] = $def->info['background-image'];\r
+ $this->info['background-repeat'] = $def->info['background-repeat'];\r
+ $this->info['background-attachment'] = $def->info['background-attachment'];\r
+ $this->info['background-position'] = $def->info['background-position'];\r
+ }\r
+\r
+ /**\r
+ * @param string $string\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string\r
+ */\r
+ public function validate($string, $config, $context)\r
+ {\r
+ // regular pre-processing\r
+ $string = $this->parseCDATA($string);\r
+ if ($string === '') {\r
+ return false;\r
+ }\r
+\r
+ // munge rgb() decl if necessary\r
+ $string = $this->mungeRgb($string);\r
+\r
+ // assumes URI doesn't have spaces in it\r
+ $bits = explode(' ', $string); // bits to process\r
+\r
+ $caught = array();\r
+ $caught['color'] = false;\r
+ $caught['image'] = false;\r
+ $caught['repeat'] = false;\r
+ $caught['attachment'] = false;\r
+ $caught['position'] = false;\r
+\r
+ $i = 0; // number of catches\r
+\r
+ foreach ($bits as $bit) {\r
+ if ($bit === '') {\r
+ continue;\r
+ }\r
+ foreach ($caught as $key => $status) {\r
+ if ($key != 'position') {\r
+ if ($status !== false) {\r
+ continue;\r
+ }\r
+ $r = $this->info['background-' . $key]->validate($bit, $config, $context);\r
+ } else {\r
+ $r = $bit;\r
+ }\r
+ if ($r === false) {\r
+ continue;\r
+ }\r
+ if ($key == 'position') {\r
+ if ($caught[$key] === false) {\r
+ $caught[$key] = '';\r
+ }\r
+ $caught[$key] .= $r . ' ';\r
+ } else {\r
+ $caught[$key] = $r;\r
+ }\r
+ $i++;\r
+ break;\r
+ }\r
+ }\r
+\r
+ if (!$i) {\r
+ return false;\r
+ }\r
+ if ($caught['position'] !== false) {\r
+ $caught['position'] = $this->info['background-position']->\r
+ validate($caught['position'], $config, $context);\r
+ }\r
+\r
+ $ret = array();\r
+ foreach ($caught as $value) {\r
+ if ($value === false) {\r
+ continue;\r
+ }\r
+ $ret[] = $value;\r
+ }\r
+\r
+ if (empty($ret)) {\r
+ return false;\r
+ }\r
+ return implode(' ', $ret);\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/* W3C says:\r
+ [ // adjective and number must be in correct order, even if\r
+ // you could switch them without introducing ambiguity.\r
+ // some browsers support that syntax\r
+ [\r
+ <percentage> | <length> | left | center | right\r
+ ]\r
+ [\r
+ <percentage> | <length> | top | center | bottom\r
+ ]?\r
+ ] |\r
+ [ // this signifies that the vertical and horizontal adjectives\r
+ // can be arbitrarily ordered, however, there can only be two,\r
+ // one of each, or none at all\r
+ [\r
+ left | center | right\r
+ ] ||\r
+ [\r
+ top | center | bottom\r
+ ]\r
+ ]\r
+ top, left = 0%\r
+ center, (none) = 50%\r
+ bottom, right = 100%\r
+*/\r
+\r
+/* QuirksMode says:\r
+ keyword + length/percentage must be ordered correctly, as per W3C\r
+\r
+ Internet Explorer and Opera, however, support arbitrary ordering. We\r
+ should fix it up.\r
+\r
+ Minor issue though, not strictly necessary.\r
+*/\r
+\r
+// control freaks may appreciate the ability to convert these to\r
+// percentages or something, but it's not necessary\r
+\r
+/**\r
+ * Validates the value of background-position.\r
+ */\r
+class HTMLPurifier_AttrDef_CSS_BackgroundPosition extends HTMLPurifier_AttrDef\r
+{\r
+\r
+ /**\r
+ * @type HTMLPurifier_AttrDef_CSS_Length\r
+ */\r
+ protected $length;\r
+\r
+ /**\r
+ * @type HTMLPurifier_AttrDef_CSS_Percentage\r
+ */\r
+ protected $percentage;\r
+\r
+ public function __construct()\r
+ {\r
+ $this->length = new HTMLPurifier_AttrDef_CSS_Length();\r
+ $this->percentage = new HTMLPurifier_AttrDef_CSS_Percentage();\r
+ }\r
+\r
+ /**\r
+ * @param string $string\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string\r
+ */\r
+ public function validate($string, $config, $context)\r
+ {\r
+ $string = $this->parseCDATA($string);\r
+ $bits = explode(' ', $string);\r
+\r
+ $keywords = array();\r
+ $keywords['h'] = false; // left, right\r
+ $keywords['v'] = false; // top, bottom\r
+ $keywords['ch'] = false; // center (first word)\r
+ $keywords['cv'] = false; // center (second word)\r
+ $measures = array();\r
+\r
+ $i = 0;\r
+\r
+ $lookup = array(\r
+ 'top' => 'v',\r
+ 'bottom' => 'v',\r
+ 'left' => 'h',\r
+ 'right' => 'h',\r
+ 'center' => 'c'\r
+ );\r
+\r
+ foreach ($bits as $bit) {\r
+ if ($bit === '') {\r
+ continue;\r
+ }\r
+\r
+ // test for keyword\r
+ $lbit = ctype_lower($bit) ? $bit : strtolower($bit);\r
+ if (isset($lookup[$lbit])) {\r
+ $status = $lookup[$lbit];\r
+ if ($status == 'c') {\r
+ if ($i == 0) {\r
+ $status = 'ch';\r
+ } else {\r
+ $status = 'cv';\r
+ }\r
+ }\r
+ $keywords[$status] = $lbit;\r
+ $i++;\r
+ }\r
+\r
+ // test for length\r
+ $r = $this->length->validate($bit, $config, $context);\r
+ if ($r !== false) {\r
+ $measures[] = $r;\r
+ $i++;\r
+ }\r
+\r
+ // test for percentage\r
+ $r = $this->percentage->validate($bit, $config, $context);\r
+ if ($r !== false) {\r
+ $measures[] = $r;\r
+ $i++;\r
+ }\r
+ }\r
+\r
+ if (!$i) {\r
+ return false;\r
+ } // no valid values were caught\r
+\r
+ $ret = array();\r
+\r
+ // first keyword\r
+ if ($keywords['h']) {\r
+ $ret[] = $keywords['h'];\r
+ } elseif ($keywords['ch']) {\r
+ $ret[] = $keywords['ch'];\r
+ $keywords['cv'] = false; // prevent re-use: center = center center\r
+ } elseif (count($measures)) {\r
+ $ret[] = array_shift($measures);\r
+ }\r
+\r
+ if ($keywords['v']) {\r
+ $ret[] = $keywords['v'];\r
+ } elseif ($keywords['cv']) {\r
+ $ret[] = $keywords['cv'];\r
+ } elseif (count($measures)) {\r
+ $ret[] = array_shift($measures);\r
+ }\r
+\r
+ if (empty($ret)) {\r
+ return false;\r
+ }\r
+ return implode(' ', $ret);\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Validates the border property as defined by CSS.\r
+ */\r
+class HTMLPurifier_AttrDef_CSS_Border extends HTMLPurifier_AttrDef\r
+{\r
+\r
+ /**\r
+ * Local copy of properties this property is shorthand for.\r
+ * @type HTMLPurifier_AttrDef[]\r
+ */\r
+ protected $info = array();\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ public function __construct($config)\r
+ {\r
+ $def = $config->getCSSDefinition();\r
+ $this->info['border-width'] = $def->info['border-width'];\r
+ $this->info['border-style'] = $def->info['border-style'];\r
+ $this->info['border-top-color'] = $def->info['border-top-color'];\r
+ }\r
+\r
+ /**\r
+ * @param string $string\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string\r
+ */\r
+ public function validate($string, $config, $context)\r
+ {\r
+ $string = $this->parseCDATA($string);\r
+ $string = $this->mungeRgb($string);\r
+ $bits = explode(' ', $string);\r
+ $done = array(); // segments we've finished\r
+ $ret = ''; // return value\r
+ foreach ($bits as $bit) {\r
+ foreach ($this->info as $propname => $validator) {\r
+ if (isset($done[$propname])) {\r
+ continue;\r
+ }\r
+ $r = $validator->validate($bit, $config, $context);\r
+ if ($r !== false) {\r
+ $ret .= $r . ' ';\r
+ $done[$propname] = true;\r
+ break;\r
+ }\r
+ }\r
+ }\r
+ return rtrim($ret);\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Validates Color as defined by CSS.\r
+ */\r
+class HTMLPurifier_AttrDef_CSS_Color extends HTMLPurifier_AttrDef\r
+{\r
+\r
+ /**\r
+ * @type HTMLPurifier_AttrDef_CSS_AlphaValue\r
+ */\r
+ protected $alpha;\r
+\r
+ public function __construct()\r
+ {\r
+ $this->alpha = new HTMLPurifier_AttrDef_CSS_AlphaValue();\r
+ }\r
+\r
+ /**\r
+ * @param string $color\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string\r
+ */\r
+ public function validate($color, $config, $context)\r
+ {\r
+ static $colors = null;\r
+ if ($colors === null) {\r
+ $colors = $config->get('Core.ColorKeywords');\r
+ }\r
+\r
+ $color = trim($color);\r
+ if ($color === '') {\r
+ return false;\r
+ }\r
+\r
+ $lower = strtolower($color);\r
+ if (isset($colors[$lower])) {\r
+ return $colors[$lower];\r
+ }\r
+\r
+ if (preg_match('#(rgb|rgba|hsl|hsla)\(#', $color, $matches) === 1) {\r
+ $length = strlen($color);\r
+ if (strpos($color, ')') !== $length - 1) {\r
+ return false;\r
+ }\r
+\r
+ // get used function : rgb, rgba, hsl or hsla\r
+ $function = $matches[1];\r
+\r
+ $parameters_size = 3;\r
+ $alpha_channel = false;\r
+ if (substr($function, -1) === 'a') {\r
+ $parameters_size = 4;\r
+ $alpha_channel = true;\r
+ }\r
+\r
+ /*\r
+ * Allowed types for values :\r
+ * parameter_position => [type => max_value]\r
+ */\r
+ $allowed_types = array(\r
+ 1 => array('percentage' => 100, 'integer' => 255),\r
+ 2 => array('percentage' => 100, 'integer' => 255),\r
+ 3 => array('percentage' => 100, 'integer' => 255),\r
+ );\r
+ $allow_different_types = false;\r
+\r
+ if (strpos($function, 'hsl') !== false) {\r
+ $allowed_types = array(\r
+ 1 => array('integer' => 360),\r
+ 2 => array('percentage' => 100),\r
+ 3 => array('percentage' => 100),\r
+ );\r
+ $allow_different_types = true;\r
+ }\r
+\r
+ $values = trim(str_replace($function, '', $color), ' ()');\r
+\r
+ $parts = explode(',', $values);\r
+ if (count($parts) !== $parameters_size) {\r
+ return false;\r
+ }\r
+\r
+ $type = false;\r
+ $new_parts = array();\r
+ $i = 0;\r
+\r
+ foreach ($parts as $part) {\r
+ $i++;\r
+ $part = trim($part);\r
+\r
+ if ($part === '') {\r
+ return false;\r
+ }\r
+\r
+ // different check for alpha channel\r
+ if ($alpha_channel === true && $i === count($parts)) {\r
+ $result = $this->alpha->validate($part, $config, $context);\r
+\r
+ if ($result === false) {\r
+ return false;\r
+ }\r
+\r
+ $new_parts[] = (string)$result;\r
+ continue;\r
+ }\r
+\r
+ if (substr($part, -1) === '%') {\r
+ $current_type = 'percentage';\r
+ } else {\r
+ $current_type = 'integer';\r
+ }\r
+\r
+ if (!array_key_exists($current_type, $allowed_types[$i])) {\r
+ return false;\r
+ }\r
+\r
+ if (!$type) {\r
+ $type = $current_type;\r
+ }\r
+\r
+ if ($allow_different_types === false && $type != $current_type) {\r
+ return false;\r
+ }\r
+\r
+ $max_value = $allowed_types[$i][$current_type];\r
+\r
+ if ($current_type == 'integer') {\r
+ // Return value between range 0 -> $max_value\r
+ $new_parts[] = (int)max(min($part, $max_value), 0);\r
+ } elseif ($current_type == 'percentage') {\r
+ $new_parts[] = (float)max(min(rtrim($part, '%'), $max_value), 0) . '%';\r
+ }\r
+ }\r
+\r
+ $new_values = implode(',', $new_parts);\r
+\r
+ $color = $function . '(' . $new_values . ')';\r
+ } else {\r
+ // hexadecimal handling\r
+ if ($color[0] === '#') {\r
+ $hex = substr($color, 1);\r
+ } else {\r
+ $hex = $color;\r
+ $color = '#' . $color;\r
+ }\r
+ $length = strlen($hex);\r
+ if ($length !== 3 && $length !== 6) {\r
+ return false;\r
+ }\r
+ if (!ctype_xdigit($hex)) {\r
+ return false;\r
+ }\r
+ }\r
+ return $color;\r
+ }\r
+\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Allows multiple validators to attempt to validate attribute.\r
+ *\r
+ * Composite is just what it sounds like: a composite of many validators.\r
+ * This means that multiple HTMLPurifier_AttrDef objects will have a whack\r
+ * at the string. If one of them passes, that's what is returned. This is\r
+ * especially useful for CSS values, which often are a choice between\r
+ * an enumerated set of predefined values or a flexible data type.\r
+ */\r
+class HTMLPurifier_AttrDef_CSS_Composite extends HTMLPurifier_AttrDef\r
+{\r
+\r
+ /**\r
+ * List of objects that may process strings.\r
+ * @type HTMLPurifier_AttrDef[]\r
+ * @todo Make protected\r
+ */\r
+ public $defs;\r
+\r
+ /**\r
+ * @param HTMLPurifier_AttrDef[] $defs List of HTMLPurifier_AttrDef objects\r
+ */\r
+ public function __construct($defs)\r
+ {\r
+ $this->defs = $defs;\r
+ }\r
+\r
+ /**\r
+ * @param string $string\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string\r
+ */\r
+ public function validate($string, $config, $context)\r
+ {\r
+ foreach ($this->defs as $i => $def) {\r
+ $result = $this->defs[$i]->validate($string, $config, $context);\r
+ if ($result !== false) {\r
+ return $result;\r
+ }\r
+ }\r
+ return false;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Decorator which enables CSS properties to be disabled for specific elements.\r
+ */\r
+class HTMLPurifier_AttrDef_CSS_DenyElementDecorator extends HTMLPurifier_AttrDef\r
+{\r
+ /**\r
+ * @type HTMLPurifier_AttrDef\r
+ */\r
+ public $def;\r
+ /**\r
+ * @type string\r
+ */\r
+ public $element;\r
+\r
+ /**\r
+ * @param HTMLPurifier_AttrDef $def Definition to wrap\r
+ * @param string $element Element to deny\r
+ */\r
+ public function __construct($def, $element)\r
+ {\r
+ $this->def = $def;\r
+ $this->element = $element;\r
+ }\r
+\r
+ /**\r
+ * Checks if CurrentToken is set and equal to $this->element\r
+ * @param string $string\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string\r
+ */\r
+ public function validate($string, $config, $context)\r
+ {\r
+ $token = $context->get('CurrentToken', true);\r
+ if ($token && $token->name == $this->element) {\r
+ return false;\r
+ }\r
+ return $this->def->validate($string, $config, $context);\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Microsoft's proprietary filter: CSS property\r
+ * @note Currently supports the alpha filter. In the future, this will\r
+ * probably need an extensible framework\r
+ */\r
+class HTMLPurifier_AttrDef_CSS_Filter extends HTMLPurifier_AttrDef\r
+{\r
+ /**\r
+ * @type HTMLPurifier_AttrDef_Integer\r
+ */\r
+ protected $intValidator;\r
+\r
+ public function __construct()\r
+ {\r
+ $this->intValidator = new HTMLPurifier_AttrDef_Integer();\r
+ }\r
+\r
+ /**\r
+ * @param string $value\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string\r
+ */\r
+ public function validate($value, $config, $context)\r
+ {\r
+ $value = $this->parseCDATA($value);\r
+ if ($value === 'none') {\r
+ return $value;\r
+ }\r
+ // if we looped this we could support multiple filters\r
+ $function_length = strcspn($value, '(');\r
+ $function = trim(substr($value, 0, $function_length));\r
+ if ($function !== 'alpha' &&\r
+ $function !== 'Alpha' &&\r
+ $function !== 'progid:DXImageTransform.Microsoft.Alpha'\r
+ ) {\r
+ return false;\r
+ }\r
+ $cursor = $function_length + 1;\r
+ $parameters_length = strcspn($value, ')', $cursor);\r
+ $parameters = substr($value, $cursor, $parameters_length);\r
+ $params = explode(',', $parameters);\r
+ $ret_params = array();\r
+ $lookup = array();\r
+ foreach ($params as $param) {\r
+ list($key, $value) = explode('=', $param);\r
+ $key = trim($key);\r
+ $value = trim($value);\r
+ if (isset($lookup[$key])) {\r
+ continue;\r
+ }\r
+ if ($key !== 'opacity') {\r
+ continue;\r
+ }\r
+ $value = $this->intValidator->validate($value, $config, $context);\r
+ if ($value === false) {\r
+ continue;\r
+ }\r
+ $int = (int)$value;\r
+ if ($int > 100) {\r
+ $value = '100';\r
+ }\r
+ if ($int < 0) {\r
+ $value = '0';\r
+ }\r
+ $ret_params[] = "$key=$value";\r
+ $lookup[$key] = true;\r
+ }\r
+ $ret_parameters = implode(',', $ret_params);\r
+ $ret_function = "$function($ret_parameters)";\r
+ return $ret_function;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Validates shorthand CSS property font.\r
+ */\r
+class HTMLPurifier_AttrDef_CSS_Font extends HTMLPurifier_AttrDef\r
+{\r
+\r
+ /**\r
+ * Local copy of validators\r
+ * @type HTMLPurifier_AttrDef[]\r
+ * @note If we moved specific CSS property definitions to their own\r
+ * classes instead of having them be assembled at run time by\r
+ * CSSDefinition, this wouldn't be necessary. We'd instantiate\r
+ * our own copies.\r
+ */\r
+ protected $info = array();\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ public function __construct($config)\r
+ {\r
+ $def = $config->getCSSDefinition();\r
+ $this->info['font-style'] = $def->info['font-style'];\r
+ $this->info['font-variant'] = $def->info['font-variant'];\r
+ $this->info['font-weight'] = $def->info['font-weight'];\r
+ $this->info['font-size'] = $def->info['font-size'];\r
+ $this->info['line-height'] = $def->info['line-height'];\r
+ $this->info['font-family'] = $def->info['font-family'];\r
+ }\r
+\r
+ /**\r
+ * @param string $string\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string\r
+ */\r
+ public function validate($string, $config, $context)\r
+ {\r
+ static $system_fonts = array(\r
+ 'caption' => true,\r
+ 'icon' => true,\r
+ 'menu' => true,\r
+ 'message-box' => true,\r
+ 'small-caption' => true,\r
+ 'status-bar' => true\r
+ );\r
+\r
+ // regular pre-processing\r
+ $string = $this->parseCDATA($string);\r
+ if ($string === '') {\r
+ return false;\r
+ }\r
+\r
+ // check if it's one of the keywords\r
+ $lowercase_string = strtolower($string);\r
+ if (isset($system_fonts[$lowercase_string])) {\r
+ return $lowercase_string;\r
+ }\r
+\r
+ $bits = explode(' ', $string); // bits to process\r
+ $stage = 0; // this indicates what we're looking for\r
+ $caught = array(); // which stage 0 properties have we caught?\r
+ $stage_1 = array('font-style', 'font-variant', 'font-weight');\r
+ $final = ''; // output\r
+\r
+ for ($i = 0, $size = count($bits); $i < $size; $i++) {\r
+ if ($bits[$i] === '') {\r
+ continue;\r
+ }\r
+ switch ($stage) {\r
+ case 0: // attempting to catch font-style, font-variant or font-weight\r
+ foreach ($stage_1 as $validator_name) {\r
+ if (isset($caught[$validator_name])) {\r
+ continue;\r
+ }\r
+ $r = $this->info[$validator_name]->validate(\r
+ $bits[$i],\r
+ $config,\r
+ $context\r
+ );\r
+ if ($r !== false) {\r
+ $final .= $r . ' ';\r
+ $caught[$validator_name] = true;\r
+ break;\r
+ }\r
+ }\r
+ // all three caught, continue on\r
+ if (count($caught) >= 3) {\r
+ $stage = 1;\r
+ }\r
+ if ($r !== false) {\r
+ break;\r
+ }\r
+ case 1: // attempting to catch font-size and perhaps line-height\r
+ $found_slash = false;\r
+ if (strpos($bits[$i], '/') !== false) {\r
+ list($font_size, $line_height) =\r
+ explode('/', $bits[$i]);\r
+ if ($line_height === '') {\r
+ // ooh, there's a space after the slash!\r
+ $line_height = false;\r
+ $found_slash = true;\r
+ }\r
+ } else {\r
+ $font_size = $bits[$i];\r
+ $line_height = false;\r
+ }\r
+ $r = $this->info['font-size']->validate(\r
+ $font_size,\r
+ $config,\r
+ $context\r
+ );\r
+ if ($r !== false) {\r
+ $final .= $r;\r
+ // attempt to catch line-height\r
+ if ($line_height === false) {\r
+ // we need to scroll forward\r
+ for ($j = $i + 1; $j < $size; $j++) {\r
+ if ($bits[$j] === '') {\r
+ continue;\r
+ }\r
+ if ($bits[$j] === '/') {\r
+ if ($found_slash) {\r
+ return false;\r
+ } else {\r
+ $found_slash = true;\r
+ continue;\r
+ }\r
+ }\r
+ $line_height = $bits[$j];\r
+ break;\r
+ }\r
+ } else {\r
+ // slash already found\r
+ $found_slash = true;\r
+ $j = $i;\r
+ }\r
+ if ($found_slash) {\r
+ $i = $j;\r
+ $r = $this->info['line-height']->validate(\r
+ $line_height,\r
+ $config,\r
+ $context\r
+ );\r
+ if ($r !== false) {\r
+ $final .= '/' . $r;\r
+ }\r
+ }\r
+ $final .= ' ';\r
+ $stage = 2;\r
+ break;\r
+ }\r
+ return false;\r
+ case 2: // attempting to catch font-family\r
+ $font_family =\r
+ implode(' ', array_slice($bits, $i, $size - $i));\r
+ $r = $this->info['font-family']->validate(\r
+ $font_family,\r
+ $config,\r
+ $context\r
+ );\r
+ if ($r !== false) {\r
+ $final .= $r . ' ';\r
+ // processing completed successfully\r
+ return rtrim($final);\r
+ }\r
+ return false;\r
+ }\r
+ }\r
+ return false;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Validates a font family list according to CSS spec\r
+ */\r
+class HTMLPurifier_AttrDef_CSS_FontFamily extends HTMLPurifier_AttrDef\r
+{\r
+\r
+ protected $mask = null;\r
+\r
+ public function __construct()\r
+ {\r
+ $this->mask = '_- ';\r
+ for ($c = 'a'; $c <= 'z'; $c++) {\r
+ $this->mask .= $c;\r
+ }\r
+ for ($c = 'A'; $c <= 'Z'; $c++) {\r
+ $this->mask .= $c;\r
+ }\r
+ for ($c = '0'; $c <= '9'; $c++) {\r
+ $this->mask .= $c;\r
+ } // cast-y, but should be fine\r
+ // special bytes used by UTF-8\r
+ for ($i = 0x80; $i <= 0xFF; $i++) {\r
+ // We don't bother excluding invalid bytes in this range,\r
+ // because the our restriction of well-formed UTF-8 will\r
+ // prevent these from ever occurring.\r
+ $this->mask .= chr($i);\r
+ }\r
+\r
+ /*\r
+ PHP's internal strcspn implementation is\r
+ O(length of string * length of mask), making it inefficient\r
+ for large masks. However, it's still faster than\r
+ preg_match 8)\r
+ for (p = s1;;) {\r
+ spanp = s2;\r
+ do {\r
+ if (*spanp == c || p == s1_end) {\r
+ return p - s1;\r
+ }\r
+ } while (spanp++ < (s2_end - 1));\r
+ c = *++p;\r
+ }\r
+ */\r
+ // possible optimization: invert the mask.\r
+ }\r
+\r
+ /**\r
+ * @param string $string\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string\r
+ */\r
+ public function validate($string, $config, $context)\r
+ {\r
+ static $generic_names = array(\r
+ 'serif' => true,\r
+ 'sans-serif' => true,\r
+ 'monospace' => true,\r
+ 'fantasy' => true,\r
+ 'cursive' => true\r
+ );\r
+ $allowed_fonts = $config->get('CSS.AllowedFonts');\r
+\r
+ // assume that no font names contain commas in them\r
+ $fonts = explode(',', $string);\r
+ $final = '';\r
+ foreach ($fonts as $font) {\r
+ $font = trim($font);\r
+ if ($font === '') {\r
+ continue;\r
+ }\r
+ // match a generic name\r
+ if (isset($generic_names[$font])) {\r
+ if ($allowed_fonts === null || isset($allowed_fonts[$font])) {\r
+ $final .= $font . ', ';\r
+ }\r
+ continue;\r
+ }\r
+ // match a quoted name\r
+ if ($font[0] === '"' || $font[0] === "'") {\r
+ $length = strlen($font);\r
+ if ($length <= 2) {\r
+ continue;\r
+ }\r
+ $quote = $font[0];\r
+ if ($font[$length - 1] !== $quote) {\r
+ continue;\r
+ }\r
+ $font = substr($font, 1, $length - 2);\r
+ }\r
+\r
+ $font = $this->expandCSSEscape($font);\r
+\r
+ // $font is a pure representation of the font name\r
+\r
+ if ($allowed_fonts !== null && !isset($allowed_fonts[$font])) {\r
+ continue;\r
+ }\r
+\r
+ if (ctype_alnum($font) && $font !== '') {\r
+ // very simple font, allow it in unharmed\r
+ $final .= $font . ', ';\r
+ continue;\r
+ }\r
+\r
+ // bugger out on whitespace. form feed (0C) really\r
+ // shouldn't show up regardless\r
+ $font = str_replace(array("\n", "\t", "\r", "\x0C"), ' ', $font);\r
+\r
+ // Here, there are various classes of characters which need\r
+ // to be treated differently:\r
+ // - Alphanumeric characters are essentially safe. We\r
+ // handled these above.\r
+ // - Spaces require quoting, though most parsers will do\r
+ // the right thing if there aren't any characters that\r
+ // can be misinterpreted\r
+ // - Dashes rarely occur, but they fairly unproblematic\r
+ // for parsing/rendering purposes.\r
+ // The above characters cover the majority of Western font\r
+ // names.\r
+ // - Arbitrary Unicode characters not in ASCII. Because\r
+ // most parsers give little thought to Unicode, treatment\r
+ // of these codepoints is basically uniform, even for\r
+ // punctuation-like codepoints. These characters can\r
+ // show up in non-Western pages and are supported by most\r
+ // major browsers, for example: "ＭＳ 明朝" is a\r
+ // legitimate font-name\r
+ // <http://ja.wikipedia.org/wiki/MS_明朝>. See\r
+ // the CSS3 spec for more examples:\r
+ // <http://www.w3.org/TR/2011/WD-css3-fonts-20110324/localizedfamilynames.png>\r
+ // You can see live samples of these on the Internet:\r
+ // <http://www.google.co.jp/search?q=font-family+ＭＳ+明朝|ゴシック>\r
+ // However, most of these fonts have ASCII equivalents:\r
+ // for example, 'MS Mincho', and it's considered\r
+ // professional to use ASCII font names instead of\r
+ // Unicode font names. Thanks Takeshi Terada for\r
+ // providing this information.\r
+ // The following characters, to my knowledge, have not been\r
+ // used to name font names.\r
+ // - Single quote. While theoretically you might find a\r
+ // font name that has a single quote in its name (serving\r
+ // as an apostrophe, e.g. Dave's Scribble), I haven't\r
+ // been able to find any actual examples of this.\r
+ // Internet Explorer's cssText translation (which I\r
+ // believe is invoked by innerHTML) normalizes any\r
+ // quoting to single quotes, and fails to escape single\r
+ // quotes. (Note that this is not IE's behavior for all\r
+ // CSS properties, just some sort of special casing for\r
+ // font-family). So a single quote *cannot* be used\r
+ // safely in the font-family context if there will be an\r
+ // innerHTML/cssText translation. Note that Firefox 3.x\r
+ // does this too.\r
+ // - Double quote. In IE, these get normalized to\r
+ // single-quotes, no matter what the encoding. (Fun\r
+ // fact, in IE8, the 'content' CSS property gained\r
+ // support, where they special cased to preserve encoded\r
+ // double quotes, but still translate unadorned double\r
+ // quotes into single quotes.) So, because their\r
+ // fixpoint behavior is identical to single quotes, they\r
+ // cannot be allowed either. Firefox 3.x displays\r
+ // single-quote style behavior.\r
+ // - Backslashes are reduced by one (so \\ -> \) every\r
+ // iteration, so they cannot be used safely. This shows\r
+ // up in IE7, IE8 and FF3\r
+ // - Semicolons, commas and backticks are handled properly.\r
+ // - The rest of the ASCII punctuation is handled properly.\r
+ // We haven't checked what browsers do to unadorned\r
+ // versions, but this is not important as long as the\r
+ // browser doesn't /remove/ surrounding quotes (as IE does\r
+ // for HTML).\r
+ //\r
+ // With these results in hand, we conclude that there are\r
+ // various levels of safety:\r
+ // - Paranoid: alphanumeric, spaces and dashes(?)\r
+ // - International: Paranoid + non-ASCII Unicode\r
+ // - Edgy: Everything except quotes, backslashes\r
+ // - NoJS: Standards compliance, e.g. sod IE. Note that\r
+ // with some judicious character escaping (since certain\r
+ // types of escaping doesn't work) this is theoretically\r
+ // OK as long as innerHTML/cssText is not called.\r
+ // We believe that international is a reasonable default\r
+ // (that we will implement now), and once we do more\r
+ // extensive research, we may feel comfortable with dropping\r
+ // it down to edgy.\r
+\r
+ // Edgy: alphanumeric, spaces, dashes, underscores and Unicode. Use of\r
+ // str(c)spn assumes that the string was already well formed\r
+ // Unicode (which of course it is).\r
+ if (strspn($font, $this->mask) !== strlen($font)) {\r
+ continue;\r
+ }\r
+\r
+ // Historical:\r
+ // In the absence of innerHTML/cssText, these ugly\r
+ // transforms don't pose a security risk (as \\ and \"\r
+ // might--these escapes are not supported by most browsers).\r
+ // We could try to be clever and use single-quote wrapping\r
+ // when there is a double quote present, but I have choosen\r
+ // not to implement that. (NOTE: you can reduce the amount\r
+ // of escapes by one depending on what quoting style you use)\r
+ // $font = str_replace('\\', '\\5C ', $font);\r
+ // $font = str_replace('"', '\\22 ', $font);\r
+ // $font = str_replace("'", '\\27 ', $font);\r
+\r
+ // font possibly with spaces, requires quoting\r
+ $final .= "'$font', ";\r
+ }\r
+ $final = rtrim($final, ', ');\r
+ if ($final === '') {\r
+ return false;\r
+ }\r
+ return $final;\r
+ }\r
+\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Validates based on {ident} CSS grammar production\r
+ */\r
+class HTMLPurifier_AttrDef_CSS_Ident extends HTMLPurifier_AttrDef\r
+{\r
+\r
+ /**\r
+ * @param string $string\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string\r
+ */\r
+ public function validate($string, $config, $context)\r
+ {\r
+ $string = trim($string);\r
+\r
+ // early abort: '' and '0' (strings that convert to false) are invalid\r
+ if (!$string) {\r
+ return false;\r
+ }\r
+\r
+ $pattern = '/^(-?[A-Za-z_][A-Za-z_\-0-9]*)$/';\r
+ if (!preg_match($pattern, $string)) {\r
+ return false;\r
+ }\r
+ return $string;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Decorator which enables !important to be used in CSS values.\r
+ */\r
+class HTMLPurifier_AttrDef_CSS_ImportantDecorator extends HTMLPurifier_AttrDef\r
+{\r
+ /**\r
+ * @type HTMLPurifier_AttrDef\r
+ */\r
+ public $def;\r
+ /**\r
+ * @type bool\r
+ */\r
+ public $allow;\r
+\r
+ /**\r
+ * @param HTMLPurifier_AttrDef $def Definition to wrap\r
+ * @param bool $allow Whether or not to allow !important\r
+ */\r
+ public function __construct($def, $allow = false)\r
+ {\r
+ $this->def = $def;\r
+ $this->allow = $allow;\r
+ }\r
+\r
+ /**\r
+ * Intercepts and removes !important if necessary\r
+ * @param string $string\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string\r
+ */\r
+ public function validate($string, $config, $context)\r
+ {\r
+ // test for ! and important tokens\r
+ $string = trim($string);\r
+ $is_important = false;\r
+ // :TODO: optimization: test directly for !important and ! important\r
+ if (strlen($string) >= 9 && substr($string, -9) === 'important') {\r
+ $temp = rtrim(substr($string, 0, -9));\r
+ // use a temp, because we might want to restore important\r
+ if (strlen($temp) >= 1 && substr($temp, -1) === '!') {\r
+ $string = rtrim(substr($temp, 0, -1));\r
+ $is_important = true;\r
+ }\r
+ }\r
+ $string = $this->def->validate($string, $config, $context);\r
+ if ($this->allow && $is_important) {\r
+ $string .= ' !important';\r
+ }\r
+ return $string;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Represents a Length as defined by CSS.\r
+ */\r
+class HTMLPurifier_AttrDef_CSS_Length extends HTMLPurifier_AttrDef\r
+{\r
+\r
+ /**\r
+ * @type HTMLPurifier_Length|string\r
+ */\r
+ protected $min;\r
+\r
+ /**\r
+ * @type HTMLPurifier_Length|string\r
+ */\r
+ protected $max;\r
+\r
+ /**\r
+ * @param HTMLPurifier_Length|string $min Minimum length, or null for no bound. String is also acceptable.\r
+ * @param HTMLPurifier_Length|string $max Maximum length, or null for no bound. String is also acceptable.\r
+ */\r
+ public function __construct($min = null, $max = null)\r
+ {\r
+ $this->min = $min !== null ? HTMLPurifier_Length::make($min) : null;\r
+ $this->max = $max !== null ? HTMLPurifier_Length::make($max) : null;\r
+ }\r
+\r
+ /**\r
+ * @param string $string\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string\r
+ */\r
+ public function validate($string, $config, $context)\r
+ {\r
+ $string = $this->parseCDATA($string);\r
+\r
+ // Optimizations\r
+ if ($string === '') {\r
+ return false;\r
+ }\r
+ if ($string === '0') {\r
+ return '0';\r
+ }\r
+ if (strlen($string) === 1) {\r
+ return false;\r
+ }\r
+\r
+ $length = HTMLPurifier_Length::make($string);\r
+ if (!$length->isValid()) {\r
+ return false;\r
+ }\r
+\r
+ if ($this->min) {\r
+ $c = $length->compareTo($this->min);\r
+ if ($c === false) {\r
+ return false;\r
+ }\r
+ if ($c < 0) {\r
+ return false;\r
+ }\r
+ }\r
+ if ($this->max) {\r
+ $c = $length->compareTo($this->max);\r
+ if ($c === false) {\r
+ return false;\r
+ }\r
+ if ($c > 0) {\r
+ return false;\r
+ }\r
+ }\r
+ return $length->toString();\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Validates shorthand CSS property list-style.\r
+ * @warning Does not support url tokens that have internal spaces.\r
+ */\r
+class HTMLPurifier_AttrDef_CSS_ListStyle extends HTMLPurifier_AttrDef\r
+{\r
+\r
+ /**\r
+ * Local copy of validators.\r
+ * @type HTMLPurifier_AttrDef[]\r
+ * @note See HTMLPurifier_AttrDef_CSS_Font::$info for a similar impl.\r
+ */\r
+ protected $info;\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ public function __construct($config)\r
+ {\r
+ $def = $config->getCSSDefinition();\r
+ $this->info['list-style-type'] = $def->info['list-style-type'];\r
+ $this->info['list-style-position'] = $def->info['list-style-position'];\r
+ $this->info['list-style-image'] = $def->info['list-style-image'];\r
+ }\r
+\r
+ /**\r
+ * @param string $string\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string\r
+ */\r
+ public function validate($string, $config, $context)\r
+ {\r
+ // regular pre-processing\r
+ $string = $this->parseCDATA($string);\r
+ if ($string === '') {\r
+ return false;\r
+ }\r
+\r
+ // assumes URI doesn't have spaces in it\r
+ $bits = explode(' ', strtolower($string)); // bits to process\r
+\r
+ $caught = array();\r
+ $caught['type'] = false;\r
+ $caught['position'] = false;\r
+ $caught['image'] = false;\r
+\r
+ $i = 0; // number of catches\r
+ $none = false;\r
+\r
+ foreach ($bits as $bit) {\r
+ if ($i >= 3) {\r
+ return;\r
+ } // optimization bit\r
+ if ($bit === '') {\r
+ continue;\r
+ }\r
+ foreach ($caught as $key => $status) {\r
+ if ($status !== false) {\r
+ continue;\r
+ }\r
+ $r = $this->info['list-style-' . $key]->validate($bit, $config, $context);\r
+ if ($r === false) {\r
+ continue;\r
+ }\r
+ if ($r === 'none') {\r
+ if ($none) {\r
+ continue;\r
+ } else {\r
+ $none = true;\r
+ }\r
+ if ($key == 'image') {\r
+ continue;\r
+ }\r
+ }\r
+ $caught[$key] = $r;\r
+ $i++;\r
+ break;\r
+ }\r
+ }\r
+\r
+ if (!$i) {\r
+ return false;\r
+ }\r
+\r
+ $ret = array();\r
+\r
+ // construct type\r
+ if ($caught['type']) {\r
+ $ret[] = $caught['type'];\r
+ }\r
+\r
+ // construct image\r
+ if ($caught['image']) {\r
+ $ret[] = $caught['image'];\r
+ }\r
+\r
+ // construct position\r
+ if ($caught['position']) {\r
+ $ret[] = $caught['position'];\r
+ }\r
+\r
+ if (empty($ret)) {\r
+ return false;\r
+ }\r
+ return implode(' ', $ret);\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Framework class for strings that involve multiple values.\r
+ *\r
+ * Certain CSS properties such as border-width and margin allow multiple\r
+ * lengths to be specified. This class can take a vanilla border-width\r
+ * definition and multiply it, usually into a max of four.\r
+ *\r
+ * @note Even though the CSS specification isn't clear about it, inherit\r
+ * can only be used alone: it will never manifest as part of a multi\r
+ * shorthand declaration. Thus, this class does not allow inherit.\r
+ */\r
+class HTMLPurifier_AttrDef_CSS_Multiple extends HTMLPurifier_AttrDef\r
+{\r
+ /**\r
+ * Instance of component definition to defer validation to.\r
+ * @type HTMLPurifier_AttrDef\r
+ * @todo Make protected\r
+ */\r
+ public $single;\r
+\r
+ /**\r
+ * Max number of values allowed.\r
+ * @todo Make protected\r
+ */\r
+ public $max;\r
+\r
+ /**\r
+ * @param HTMLPurifier_AttrDef $single HTMLPurifier_AttrDef to multiply\r
+ * @param int $max Max number of values allowed (usually four)\r
+ */\r
+ public function __construct($single, $max = 4)\r
+ {\r
+ $this->single = $single;\r
+ $this->max = $max;\r
+ }\r
+\r
+ /**\r
+ * @param string $string\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string\r
+ */\r
+ public function validate($string, $config, $context)\r
+ {\r
+ $string = $this->mungeRgb($this->parseCDATA($string));\r
+ if ($string === '') {\r
+ return false;\r
+ }\r
+ $parts = explode(' ', $string); // parseCDATA replaced \r, \t and \n\r
+ $length = count($parts);\r
+ $final = '';\r
+ for ($i = 0, $num = 0; $i < $length && $num < $this->max; $i++) {\r
+ if (ctype_space($parts[$i])) {\r
+ continue;\r
+ }\r
+ $result = $this->single->validate($parts[$i], $config, $context);\r
+ if ($result !== false) {\r
+ $final .= $result . ' ';\r
+ $num++;\r
+ }\r
+ }\r
+ if ($final === '') {\r
+ return false;\r
+ }\r
+ return rtrim($final);\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Validates a Percentage as defined by the CSS spec.\r
+ */\r
+class HTMLPurifier_AttrDef_CSS_Percentage extends HTMLPurifier_AttrDef\r
+{\r
+\r
+ /**\r
+ * Instance to defer number validation to.\r
+ * @type HTMLPurifier_AttrDef_CSS_Number\r
+ */\r
+ protected $number_def;\r
+\r
+ /**\r
+ * @param bool $non_negative Whether to forbid negative values\r
+ */\r
+ public function __construct($non_negative = false)\r
+ {\r
+ $this->number_def = new HTMLPurifier_AttrDef_CSS_Number($non_negative);\r
+ }\r
+\r
+ /**\r
+ * @param string $string\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string\r
+ */\r
+ public function validate($string, $config, $context)\r
+ {\r
+ $string = $this->parseCDATA($string);\r
+\r
+ if ($string === '') {\r
+ return false;\r
+ }\r
+ $length = strlen($string);\r
+ if ($length === 1) {\r
+ return false;\r
+ }\r
+ if ($string[$length - 1] !== '%') {\r
+ return false;\r
+ }\r
+\r
+ $number = substr($string, 0, $length - 1);\r
+ $number = $this->number_def->validate($number, $config, $context);\r
+\r
+ if ($number === false) {\r
+ return false;\r
+ }\r
+ return "$number%";\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Validates the value for the CSS property text-decoration\r
+ * @note This class could be generalized into a version that acts sort of\r
+ * like Enum except you can compound the allowed values.\r
+ */\r
+class HTMLPurifier_AttrDef_CSS_TextDecoration extends HTMLPurifier_AttrDef\r
+{\r
+\r
+ /**\r
+ * @param string $string\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string\r
+ */\r
+ public function validate($string, $config, $context)\r
+ {\r
+ static $allowed_values = array(\r
+ 'line-through' => true,\r
+ 'overline' => true,\r
+ 'underline' => true,\r
+ );\r
+\r
+ $string = strtolower($this->parseCDATA($string));\r
+\r
+ if ($string === 'none') {\r
+ return $string;\r
+ }\r
+\r
+ $parts = explode(' ', $string);\r
+ $final = '';\r
+ foreach ($parts as $part) {\r
+ if (isset($allowed_values[$part])) {\r
+ $final .= $part . ' ';\r
+ }\r
+ }\r
+ $final = rtrim($final);\r
+ if ($final === '') {\r
+ return false;\r
+ }\r
+ return $final;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Validates a URI in CSS syntax, which uses url('http://example.com')\r
+ * @note While theoretically speaking a URI in a CSS document could\r
+ * be non-embedded, as of CSS2 there is no such usage so we're\r
+ * generalizing it. This may need to be changed in the future.\r
+ * @warning Since HTMLPurifier_AttrDef_CSS blindly uses semicolons as\r
+ * the separator, you cannot put a literal semicolon in\r
+ * in the URI. Try percent encoding it, in that case.\r
+ */\r
+class HTMLPurifier_AttrDef_CSS_URI extends HTMLPurifier_AttrDef_URI\r
+{\r
+\r
+ public function __construct()\r
+ {\r
+ parent::__construct(true); // always embedded\r
+ }\r
+\r
+ /**\r
+ * @param string $uri_string\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string\r
+ */\r
+ public function validate($uri_string, $config, $context)\r
+ {\r
+ // parse the URI out of the string and then pass it onto\r
+ // the parent object\r
+\r
+ $uri_string = $this->parseCDATA($uri_string);\r
+ if (strpos($uri_string, 'url(') !== 0) {\r
+ return false;\r
+ }\r
+ $uri_string = substr($uri_string, 4);\r
+ if (strlen($uri_string) == 0) {\r
+ return false;\r
+ }\r
+ $new_length = strlen($uri_string) - 1;\r
+ if ($uri_string[$new_length] != ')') {\r
+ return false;\r
+ }\r
+ $uri = trim(substr($uri_string, 0, $new_length));\r
+\r
+ if (!empty($uri) && ($uri[0] == "'" || $uri[0] == '"')) {\r
+ $quote = $uri[0];\r
+ $new_length = strlen($uri) - 1;\r
+ if ($uri[$new_length] !== $quote) {\r
+ return false;\r
+ }\r
+ $uri = substr($uri, 1, $new_length - 1);\r
+ }\r
+\r
+ $uri = $this->expandCSSEscape($uri);\r
+\r
+ $result = parent::validate($uri, $config, $context);\r
+\r
+ if ($result === false) {\r
+ return false;\r
+ }\r
+\r
+ // extra sanity check; should have been done by URI\r
+ $result = str_replace(array('"', "\\", "\n", "\x0c", "\r"), "", $result);\r
+\r
+ // suspicious characters are ()'; we're going to percent encode\r
+ // them for safety.\r
+ $result = str_replace(array('(', ')', "'"), array('%28', '%29', '%27'), $result);\r
+\r
+ // there's an extra bug where ampersands lose their escaping on\r
+ // an innerHTML cycle, so a very unlucky query parameter could\r
+ // then change the meaning of the URL. Unfortunately, there's\r
+ // not much we can do about that...\r
+ return "url(\"$result\")";\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Validates a boolean attribute\r
+ */\r
+class HTMLPurifier_AttrDef_HTML_Bool extends HTMLPurifier_AttrDef\r
+{\r
+\r
+ /**\r
+ * @type bool\r
+ */\r
+ protected $name;\r
+\r
+ /**\r
+ * @type bool\r
+ */\r
+ public $minimized = true;\r
+\r
+ /**\r
+ * @param bool $name\r
+ */\r
+ public function __construct($name = false)\r
+ {\r
+ $this->name = $name;\r
+ }\r
+\r
+ /**\r
+ * @param string $string\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string\r
+ */\r
+ public function validate($string, $config, $context)\r
+ {\r
+ return $this->name;\r
+ }\r
+\r
+ /**\r
+ * @param string $string Name of attribute\r
+ * @return HTMLPurifier_AttrDef_HTML_Bool\r
+ */\r
+ public function make($string)\r
+ {\r
+ return new HTMLPurifier_AttrDef_HTML_Bool($string);\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Validates contents based on NMTOKENS attribute type.\r
+ */\r
+class HTMLPurifier_AttrDef_HTML_Nmtokens extends HTMLPurifier_AttrDef\r
+{\r
+\r
+ /**\r
+ * @param string $string\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string\r
+ */\r
+ public function validate($string, $config, $context)\r
+ {\r
+ $string = trim($string);\r
+\r
+ // early abort: '' and '0' (strings that convert to false) are invalid\r
+ if (!$string) {\r
+ return false;\r
+ }\r
+\r
+ $tokens = $this->split($string, $config, $context);\r
+ $tokens = $this->filter($tokens, $config, $context);\r
+ if (empty($tokens)) {\r
+ return false;\r
+ }\r
+ return implode(' ', $tokens);\r
+ }\r
+\r
+ /**\r
+ * Splits a space separated list of tokens into its constituent parts.\r
+ * @param string $string\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return array\r
+ */\r
+ protected function split($string, $config, $context)\r
+ {\r
+ // OPTIMIZABLE!\r
+ // do the preg_match, capture all subpatterns for reformulation\r
+\r
+ // we don't support U+00A1 and up codepoints or\r
+ // escaping because I don't know how to do that with regexps\r
+ // and plus it would complicate optimization efforts (you never\r
+ // see that anyway).\r
+ $pattern = '/(?:(?<=\s)|\A)' . // look behind for space or string start\r
+ '((?:--|-?[A-Za-z_])[A-Za-z_\-0-9]*)' .\r
+ '(?:(?=\s)|\z)/'; // look ahead for space or string end\r
+ preg_match_all($pattern, $string, $matches);\r
+ return $matches[1];\r
+ }\r
+\r
+ /**\r
+ * Template method for removing certain tokens based on arbitrary criteria.\r
+ * @note If we wanted to be really functional, we'd do an array_filter\r
+ * with a callback. But... we're not.\r
+ * @param array $tokens\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return array\r
+ */\r
+ protected function filter($tokens, $config, $context)\r
+ {\r
+ return $tokens;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Implements special behavior for class attribute (normally NMTOKENS)\r
+ */\r
+class HTMLPurifier_AttrDef_HTML_Class extends HTMLPurifier_AttrDef_HTML_Nmtokens\r
+{\r
+ /**\r
+ * @param string $string\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string\r
+ */\r
+ protected function split($string, $config, $context)\r
+ {\r
+ // really, this twiddle should be lazy loaded\r
+ $name = $config->getDefinition('HTML')->doctype->name;\r
+ if ($name == "XHTML 1.1" || $name == "XHTML 2.0") {\r
+ return parent::split($string, $config, $context);\r
+ } else {\r
+ return preg_split('/\s+/', $string);\r
+ }\r
+ }\r
+\r
+ /**\r
+ * @param array $tokens\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return array\r
+ */\r
+ protected function filter($tokens, $config, $context)\r
+ {\r
+ $allowed = $config->get('Attr.AllowedClasses');\r
+ $forbidden = $config->get('Attr.ForbiddenClasses');\r
+ $ret = array();\r
+ foreach ($tokens as $token) {\r
+ if (($allowed === null || isset($allowed[$token])) &&\r
+ !isset($forbidden[$token]) &&\r
+ // We need this O(n) check because of PHP's array\r
+ // implementation that casts -0 to 0.\r
+ !in_array($token, $ret, true)\r
+ ) {\r
+ $ret[] = $token;\r
+ }\r
+ }\r
+ return $ret;\r
+ }\r
+}\r
+
+\r
+\r
+/**\r
+ * Validates a color according to the HTML spec.\r
+ */\r
+class HTMLPurifier_AttrDef_HTML_Color extends HTMLPurifier_AttrDef\r
+{\r
+\r
+ /**\r
+ * @param string $string\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string\r
+ */\r
+ public function validate($string, $config, $context)\r
+ {\r
+ static $colors = null;\r
+ if ($colors === null) {\r
+ $colors = $config->get('Core.ColorKeywords');\r
+ }\r
+\r
+ $string = trim($string);\r
+\r
+ if (empty($string)) {\r
+ return false;\r
+ }\r
+ $lower = strtolower($string);\r
+ if (isset($colors[$lower])) {\r
+ return $colors[$lower];\r
+ }\r
+ if ($string[0] === '#') {\r
+ $hex = substr($string, 1);\r
+ } else {\r
+ $hex = $string;\r
+ }\r
+\r
+ $length = strlen($hex);\r
+ if ($length !== 3 && $length !== 6) {\r
+ return false;\r
+ }\r
+ if (!ctype_xdigit($hex)) {\r
+ return false;\r
+ }\r
+ if ($length === 3) {\r
+ $hex = $hex[0] . $hex[0] . $hex[1] . $hex[1] . $hex[2] . $hex[2];\r
+ }\r
+ return "#$hex";\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Special-case enum attribute definition that lazy loads allowed frame targets\r
+ */\r
+class HTMLPurifier_AttrDef_HTML_FrameTarget extends HTMLPurifier_AttrDef_Enum\r
+{\r
+\r
+ /**\r
+ * @type array\r
+ */\r
+ public $valid_values = false; // uninitialized value\r
+\r
+ /**\r
+ * @type bool\r
+ */\r
+ protected $case_sensitive = false;\r
+\r
+ public function __construct()\r
+ {\r
+ }\r
+\r
+ /**\r
+ * @param string $string\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string\r
+ */\r
+ public function validate($string, $config, $context)\r
+ {\r
+ if ($this->valid_values === false) {\r
+ $this->valid_values = $config->get('Attr.AllowedFrameTargets');\r
+ }\r
+ return parent::validate($string, $config, $context);\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Validates the HTML attribute ID.\r
+ * @warning Even though this is the id processor, it\r
+ * will ignore the directive Attr:IDBlacklist, since it will only\r
+ * go according to the ID accumulator. Since the accumulator is\r
+ * automatically generated, it will have already absorbed the\r
+ * blacklist. If you're hacking around, make sure you use load()!\r
+ */\r
+\r
+class HTMLPurifier_AttrDef_HTML_ID extends HTMLPurifier_AttrDef\r
+{\r
+\r
+ // selector is NOT a valid thing to use for IDREFs, because IDREFs\r
+ // *must* target IDs that exist, whereas selector #ids do not.\r
+\r
+ /**\r
+ * Determines whether or not we're validating an ID in a CSS\r
+ * selector context.\r
+ * @type bool\r
+ */\r
+ protected $selector;\r
+\r
+ /**\r
+ * @param bool $selector\r
+ */\r
+ public function __construct($selector = false)\r
+ {\r
+ $this->selector = $selector;\r
+ }\r
+\r
+ /**\r
+ * @param string $id\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string\r
+ */\r
+ public function validate($id, $config, $context)\r
+ {\r
+ if (!$this->selector && !$config->get('Attr.EnableID')) {\r
+ return false;\r
+ }\r
+\r
+ $id = trim($id); // trim it first\r
+\r
+ if ($id === '') {\r
+ return false;\r
+ }\r
+\r
+ $prefix = $config->get('Attr.IDPrefix');\r
+ if ($prefix !== '') {\r
+ $prefix .= $config->get('Attr.IDPrefixLocal');\r
+ // prevent re-appending the prefix\r
+ if (strpos($id, $prefix) !== 0) {\r
+ $id = $prefix . $id;\r
+ }\r
+ } elseif ($config->get('Attr.IDPrefixLocal') !== '') {\r
+ trigger_error(\r
+ '%Attr.IDPrefixLocal cannot be used unless ' .\r
+ '%Attr.IDPrefix is set',\r
+ E_USER_WARNING\r
+ );\r
+ }\r
+\r
+ if (!$this->selector) {\r
+ $id_accumulator =& $context->get('IDAccumulator');\r
+ if (isset($id_accumulator->ids[$id])) {\r
+ return false;\r
+ }\r
+ }\r
+\r
+ // we purposely avoid using regex, hopefully this is faster\r
+\r
+ if ($config->get('Attr.ID.HTML5') === true) {\r
+ if (preg_match('/[\t\n\x0b\x0c ]/', $id)) {\r
+ return false;\r
+ }\r
+ } else {\r
+ if (ctype_alpha($id)) {\r
+ // OK\r
+ } else {\r
+ if (!ctype_alpha(@$id[0])) {\r
+ return false;\r
+ }\r
+ // primitive style of regexps, I suppose\r
+ $trim = trim(\r
+ $id,\r
+ 'A..Za..z0..9:-._'\r
+ );\r
+ if ($trim !== '') {\r
+ return false;\r
+ }\r
+ }\r
+ }\r
+\r
+ $regexp = $config->get('Attr.IDBlacklistRegexp');\r
+ if ($regexp && preg_match($regexp, $id)) {\r
+ return false;\r
+ }\r
+\r
+ if (!$this->selector) {\r
+ $id_accumulator->add($id);\r
+ }\r
+\r
+ // if no change was made to the ID, return the result\r
+ // else, return the new id if stripping whitespace made it\r
+ // valid, or return false.\r
+ return $id;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Validates an integer representation of pixels according to the HTML spec.\r
+ */\r
+class HTMLPurifier_AttrDef_HTML_Pixels extends HTMLPurifier_AttrDef\r
+{\r
+\r
+ /**\r
+ * @type int\r
+ */\r
+ protected $max;\r
+\r
+ /**\r
+ * @param int $max\r
+ */\r
+ public function __construct($max = null)\r
+ {\r
+ $this->max = $max;\r
+ }\r
+\r
+ /**\r
+ * @param string $string\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string\r
+ */\r
+ public function validate($string, $config, $context)\r
+ {\r
+ $string = trim($string);\r
+ if ($string === '0') {\r
+ return $string;\r
+ }\r
+ if ($string === '') {\r
+ return false;\r
+ }\r
+ $length = strlen($string);\r
+ if (substr($string, $length - 2) == 'px') {\r
+ $string = substr($string, 0, $length - 2);\r
+ }\r
+ if (!is_numeric($string)) {\r
+ return false;\r
+ }\r
+ $int = (int)$string;\r
+\r
+ if ($int < 0) {\r
+ return '0';\r
+ }\r
+\r
+ // upper-bound value, extremely high values can\r
+ // crash operating systems, see <http://ha.ckers.org/imagecrash.html>\r
+ // WARNING, above link WILL crash you if you're using Windows\r
+\r
+ if ($this->max !== null && $int > $this->max) {\r
+ return (string)$this->max;\r
+ }\r
+ return (string)$int;\r
+ }\r
+\r
+ /**\r
+ * @param string $string\r
+ * @return HTMLPurifier_AttrDef\r
+ */\r
+ public function make($string)\r
+ {\r
+ if ($string === '') {\r
+ $max = null;\r
+ } else {\r
+ $max = (int)$string;\r
+ }\r
+ $class = get_class($this);\r
+ return new $class($max);\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Validates the HTML type length (not to be confused with CSS's length).\r
+ *\r
+ * This accepts integer pixels or percentages as lengths for certain\r
+ * HTML attributes.\r
+ */\r
+\r
+class HTMLPurifier_AttrDef_HTML_Length extends HTMLPurifier_AttrDef_HTML_Pixels\r
+{\r
+\r
+ /**\r
+ * @param string $string\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string\r
+ */\r
+ public function validate($string, $config, $context)\r
+ {\r
+ $string = trim($string);\r
+ if ($string === '') {\r
+ return false;\r
+ }\r
+\r
+ $parent_result = parent::validate($string, $config, $context);\r
+ if ($parent_result !== false) {\r
+ return $parent_result;\r
+ }\r
+\r
+ $length = strlen($string);\r
+ $last_char = $string[$length - 1];\r
+\r
+ if ($last_char !== '%') {\r
+ return false;\r
+ }\r
+\r
+ $points = substr($string, 0, $length - 1);\r
+\r
+ if (!is_numeric($points)) {\r
+ return false;\r
+ }\r
+\r
+ $points = (int)$points;\r
+\r
+ if ($points < 0) {\r
+ return '0%';\r
+ }\r
+ if ($points > 100) {\r
+ return '100%';\r
+ }\r
+ return ((string)$points) . '%';\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Validates a rel/rev link attribute against a directive of allowed values\r
+ * @note We cannot use Enum because link types allow multiple\r
+ * values.\r
+ * @note Assumes link types are ASCII text\r
+ */\r
+class HTMLPurifier_AttrDef_HTML_LinkTypes extends HTMLPurifier_AttrDef\r
+{\r
+\r
+ /**\r
+ * Name config attribute to pull.\r
+ * @type string\r
+ */\r
+ protected $name;\r
+\r
+ /**\r
+ * @param string $name\r
+ */\r
+ public function __construct($name)\r
+ {\r
+ $configLookup = array(\r
+ 'rel' => 'AllowedRel',\r
+ 'rev' => 'AllowedRev'\r
+ );\r
+ if (!isset($configLookup[$name])) {\r
+ trigger_error(\r
+ 'Unrecognized attribute name for link ' .\r
+ 'relationship.',\r
+ E_USER_ERROR\r
+ );\r
+ return;\r
+ }\r
+ $this->name = $configLookup[$name];\r
+ }\r
+\r
+ /**\r
+ * @param string $string\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string\r
+ */\r
+ public function validate($string, $config, $context)\r
+ {\r
+ $allowed = $config->get('Attr.' . $this->name);\r
+ if (empty($allowed)) {\r
+ return false;\r
+ }\r
+\r
+ $string = $this->parseCDATA($string);\r
+ $parts = explode(' ', $string);\r
+\r
+ // lookup to prevent duplicates\r
+ $ret_lookup = array();\r
+ foreach ($parts as $part) {\r
+ $part = strtolower(trim($part));\r
+ if (!isset($allowed[$part])) {\r
+ continue;\r
+ }\r
+ $ret_lookup[$part] = true;\r
+ }\r
+\r
+ if (empty($ret_lookup)) {\r
+ return false;\r
+ }\r
+ $string = implode(' ', array_keys($ret_lookup));\r
+ return $string;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Validates a MultiLength as defined by the HTML spec.\r
+ *\r
+ * A multilength is either a integer (pixel count), a percentage, or\r
+ * a relative number.\r
+ */\r
+class HTMLPurifier_AttrDef_HTML_MultiLength extends HTMLPurifier_AttrDef_HTML_Length\r
+{\r
+\r
+ /**\r
+ * @param string $string\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string\r
+ */\r
+ public function validate($string, $config, $context)\r
+ {\r
+ $string = trim($string);\r
+ if ($string === '') {\r
+ return false;\r
+ }\r
+\r
+ $parent_result = parent::validate($string, $config, $context);\r
+ if ($parent_result !== false) {\r
+ return $parent_result;\r
+ }\r
+\r
+ $length = strlen($string);\r
+ $last_char = $string[$length - 1];\r
+\r
+ if ($last_char !== '*') {\r
+ return false;\r
+ }\r
+\r
+ $int = substr($string, 0, $length - 1);\r
+\r
+ if ($int == '') {\r
+ return '*';\r
+ }\r
+ if (!is_numeric($int)) {\r
+ return false;\r
+ }\r
+\r
+ $int = (int)$int;\r
+ if ($int < 0) {\r
+ return false;\r
+ }\r
+ if ($int == 0) {\r
+ return '0';\r
+ }\r
+ if ($int == 1) {\r
+ return '*';\r
+ }\r
+ return ((string)$int) . '*';\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+abstract class HTMLPurifier_AttrDef_URI_Email extends HTMLPurifier_AttrDef\r
+{\r
+\r
+ /**\r
+ * Unpacks a mailbox into its display-name and address\r
+ * @param string $string\r
+ * @return mixed\r
+ */\r
+ public function unpack($string)\r
+ {\r
+ // needs to be implemented\r
+ }\r
+\r
+}\r
+\r
+// sub-implementations\r
+\r
+
+
+\r
+\r
+/**\r
+ * Validates a host according to the IPv4, IPv6 and DNS (future) specifications.\r
+ */\r
+class HTMLPurifier_AttrDef_URI_Host extends HTMLPurifier_AttrDef\r
+{\r
+\r
+ /**\r
+ * IPv4 sub-validator.\r
+ * @type HTMLPurifier_AttrDef_URI_IPv4\r
+ */\r
+ protected $ipv4;\r
+\r
+ /**\r
+ * IPv6 sub-validator.\r
+ * @type HTMLPurifier_AttrDef_URI_IPv6\r
+ */\r
+ protected $ipv6;\r
+\r
+ public function __construct()\r
+ {\r
+ $this->ipv4 = new HTMLPurifier_AttrDef_URI_IPv4();\r
+ $this->ipv6 = new HTMLPurifier_AttrDef_URI_IPv6();\r
+ }\r
+\r
+ /**\r
+ * @param string $string\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string\r
+ */\r
+ public function validate($string, $config, $context)\r
+ {\r
+ $length = strlen($string);\r
+ // empty hostname is OK; it's usually semantically equivalent:\r
+ // the default host as defined by a URI scheme is used:\r
+ //\r
+ // If the URI scheme defines a default for host, then that\r
+ // default applies when the host subcomponent is undefined\r
+ // or when the registered name is empty (zero length).\r
+ if ($string === '') {\r
+ return '';\r
+ }\r
+ if ($length > 1 && $string[0] === '[' && $string[$length - 1] === ']') {\r
+ //IPv6\r
+ $ip = substr($string, 1, $length - 2);\r
+ $valid = $this->ipv6->validate($ip, $config, $context);\r
+ if ($valid === false) {\r
+ return false;\r
+ }\r
+ return '[' . $valid . ']';\r
+ }\r
+\r
+ // need to do checks on unusual encodings too\r
+ $ipv4 = $this->ipv4->validate($string, $config, $context);\r
+ if ($ipv4 !== false) {\r
+ return $ipv4;\r
+ }\r
+\r
+ // A regular domain name.\r
+\r
+ // This doesn't match I18N domain names, but we don't have proper IRI support,\r
+ // so force users to insert Punycode.\r
+\r
+ // There is not a good sense in which underscores should be\r
+ // allowed, since it's technically not! (And if you go as\r
+ // far to allow everything as specified by the DNS spec...\r
+ // well, that's literally everything, modulo some space limits\r
+ // for the components and the overall name (which, by the way,\r
+ // we are NOT checking!). So we (arbitrarily) decide this:\r
+ // let's allow underscores wherever we would have allowed\r
+ // hyphens, if they are enabled. This is a pretty good match\r
+ // for browser behavior, for example, a large number of browsers\r
+ // cannot handle foo_.example.com, but foo_bar.example.com is\r
+ // fairly well supported.\r
+ $underscore = $config->get('Core.AllowHostnameUnderscore') ? '_' : '';\r
+\r
+ // Based off of RFC 1738, but amended so that\r
+ // as per RFC 3696, the top label need only not be all numeric.\r
+ // The productions describing this are:\r
+ $a = '[a-z]'; // alpha\r
+ $an = '[a-z0-9]'; // alphanum\r
+ $and = "[a-z0-9-$underscore]"; // alphanum | "-"\r
+ // domainlabel = alphanum | alphanum *( alphanum | "-" ) alphanum\r
+ $domainlabel = "$an(?:$and*$an)?";\r
+ // AMENDED as per RFC 3696\r
+ // toplabel = alphanum | alphanum *( alphanum | "-" ) alphanum\r
+ // side condition: not all numeric\r
+ $toplabel = "$an(?:$and*$an)?";\r
+ // hostname = *( domainlabel "." ) toplabel [ "." ]\r
+ if (preg_match("/^(?:$domainlabel\.)*($toplabel)\.?$/i", $string, $matches)) {\r
+ if (!ctype_digit($matches[1])) {\r
+ return $string;\r
+ }\r
+ }\r
+\r
+ // PHP 5.3 and later support this functionality natively\r
+ if (function_exists('idn_to_ascii')) {\r
+ if (defined('IDNA_NONTRANSITIONAL_TO_ASCII') && defined('INTL_IDNA_VARIANT_UTS46')) {\r
+ $string = idn_to_ascii($string, IDNA_NONTRANSITIONAL_TO_ASCII, INTL_IDNA_VARIANT_UTS46);\r
+ } else {\r
+ $string = idn_to_ascii($string);\r
+ }\r
+\r
+ // If we have Net_IDNA2 support, we can support IRIs by\r
+ // punycoding them. (This is the most portable thing to do,\r
+ // since otherwise we have to assume browsers support\r
+ } elseif ($config->get('Core.EnableIDNA')) {\r
+ $idna = new Net_IDNA2(array('encoding' => 'utf8', 'overlong' => false, 'strict' => true));\r
+ // we need to encode each period separately\r
+ $parts = explode('.', $string);\r
+ try {\r
+ $new_parts = array();\r
+ foreach ($parts as $part) {\r
+ $encodable = false;\r
+ for ($i = 0, $c = strlen($part); $i < $c; $i++) {\r
+ if (ord($part[$i]) > 0x7a) {\r
+ $encodable = true;\r
+ break;\r
+ }\r
+ }\r
+ if (!$encodable) {\r
+ $new_parts[] = $part;\r
+ } else {\r
+ $new_parts[] = $idna->encode($part);\r
+ }\r
+ }\r
+ $string = implode('.', $new_parts);\r
+ } catch (Exception $e) {\r
+ // XXX error reporting\r
+ }\r
+ }\r
+ // Try again\r
+ if (preg_match("/^($domainlabel\.)*$toplabel\.?$/i", $string)) {\r
+ return $string;\r
+ }\r
+ return false;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Validates an IPv4 address\r
+ * @author Feyd @ forums.devnetwork.net (public domain)\r
+ */\r
+class HTMLPurifier_AttrDef_URI_IPv4 extends HTMLPurifier_AttrDef\r
+{\r
+\r
+ /**\r
+ * IPv4 regex, protected so that IPv6 can reuse it.\r
+ * @type string\r
+ */\r
+ protected $ip4;\r
+\r
+ /**\r
+ * @param string $aIP\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string\r
+ */\r
+ public function validate($aIP, $config, $context)\r
+ {\r
+ if (!$this->ip4) {\r
+ $this->_loadRegex();\r
+ }\r
+\r
+ if (preg_match('#^' . $this->ip4 . '$#s', $aIP)) {\r
+ return $aIP;\r
+ }\r
+ return false;\r
+ }\r
+\r
+ /**\r
+ * Lazy load function to prevent regex from being stuffed in\r
+ * cache.\r
+ */\r
+ protected function _loadRegex()\r
+ {\r
+ $oct = '(?:25[0-5]|2[0-4][0-9]|1[0-9]{2}|[1-9][0-9]|[0-9])'; // 0-255\r
+ $this->ip4 = "(?:{$oct}\\.{$oct}\\.{$oct}\\.{$oct})";\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Validates an IPv6 address.\r
+ * @author Feyd @ forums.devnetwork.net (public domain)\r
+ * @note This function requires brackets to have been removed from address\r
+ * in URI.\r
+ */\r
+class HTMLPurifier_AttrDef_URI_IPv6 extends HTMLPurifier_AttrDef_URI_IPv4\r
+{\r
+\r
+ /**\r
+ * @param string $aIP\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string\r
+ */\r
+ public function validate($aIP, $config, $context)\r
+ {\r
+ if (!$this->ip4) {\r
+ $this->_loadRegex();\r
+ }\r
+\r
+ $original = $aIP;\r
+\r
+ $hex = '[0-9a-fA-F]';\r
+ $blk = '(?:' . $hex . '{1,4})';\r
+ $pre = '(?:/(?:12[0-8]|1[0-1][0-9]|[1-9][0-9]|[0-9]))'; // /0 - /128\r
+\r
+ // prefix check\r
+ if (strpos($aIP, '/') !== false) {\r
+ if (preg_match('#' . $pre . '$#s', $aIP, $find)) {\r
+ $aIP = substr($aIP, 0, 0 - strlen($find[0]));\r
+ unset($find);\r
+ } else {\r
+ return false;\r
+ }\r
+ }\r
+\r
+ // IPv4-compatiblity check\r
+ if (preg_match('#(?<=:' . ')' . $this->ip4 . '$#s', $aIP, $find)) {\r
+ $aIP = substr($aIP, 0, 0 - strlen($find[0]));\r
+ $ip = explode('.', $find[0]);\r
+ $ip = array_map('dechex', $ip);\r
+ $aIP .= $ip[0] . $ip[1] . ':' . $ip[2] . $ip[3];\r
+ unset($find, $ip);\r
+ }\r
+\r
+ // compression check\r
+ $aIP = explode('::', $aIP);\r
+ $c = count($aIP);\r
+ if ($c > 2) {\r
+ return false;\r
+ } elseif ($c == 2) {\r
+ list($first, $second) = $aIP;\r
+ $first = explode(':', $first);\r
+ $second = explode(':', $second);\r
+\r
+ if (count($first) + count($second) > 8) {\r
+ return false;\r
+ }\r
+\r
+ while (count($first) < 8) {\r
+ array_push($first, '0');\r
+ }\r
+\r
+ array_splice($first, 8 - count($second), 8, $second);\r
+ $aIP = $first;\r
+ unset($first, $second);\r
+ } else {\r
+ $aIP = explode(':', $aIP[0]);\r
+ }\r
+ $c = count($aIP);\r
+\r
+ if ($c != 8) {\r
+ return false;\r
+ }\r
+\r
+ // All the pieces should be 16-bit hex strings. Are they?\r
+ foreach ($aIP as $piece) {\r
+ if (!preg_match('#^[0-9a-fA-F]{4}$#s', sprintf('%04s', $piece))) {\r
+ return false;\r
+ }\r
+ }\r
+ return $original;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Primitive email validation class based on the regexp found at\r
+ * http://www.regular-expressions.info/email.html\r
+ */\r
+class HTMLPurifier_AttrDef_URI_Email_SimpleCheck extends HTMLPurifier_AttrDef_URI_Email\r
+{\r
+\r
+ /**\r
+ * @param string $string\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool|string\r
+ */\r
+ public function validate($string, $config, $context)\r
+ {\r
+ // no support for named mailboxes i.e. "Bob <bob@example.com>"\r
+ // that needs more percent encoding to be done\r
+ if ($string == '') {\r
+ return false;\r
+ }\r
+ $string = trim($string);\r
+ $result = preg_match('/^[A-Z0-9._%-]+@[A-Z0-9.-]+\.[A-Z]{2,4}$/i', $string);\r
+ return $result ? $string : false;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Pre-transform that changes proprietary background attribute to CSS.\r
+ */\r
+class HTMLPurifier_AttrTransform_Background extends HTMLPurifier_AttrTransform\r
+{\r
+ /**\r
+ * @param array $attr\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return array\r
+ */\r
+ public function transform($attr, $config, $context)\r
+ {\r
+ if (!isset($attr['background'])) {\r
+ return $attr;\r
+ }\r
+\r
+ $background = $this->confiscateAttr($attr, 'background');\r
+ // some validation should happen here\r
+\r
+ $this->prependCSS($attr, "background-image:url($background);");\r
+ return $attr;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+// this MUST be placed in post, as it assumes that any value in dir is valid\r
+\r
+/**\r
+ * Post-trasnform that ensures that bdo tags have the dir attribute set.\r
+ */\r
+class HTMLPurifier_AttrTransform_BdoDir extends HTMLPurifier_AttrTransform\r
+{\r
+\r
+ /**\r
+ * @param array $attr\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return array\r
+ */\r
+ public function transform($attr, $config, $context)\r
+ {\r
+ if (isset($attr['dir'])) {\r
+ return $attr;\r
+ }\r
+ $attr['dir'] = $config->get('Attr.DefaultTextDir');\r
+ return $attr;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Pre-transform that changes deprecated bgcolor attribute to CSS.\r
+ */\r
+class HTMLPurifier_AttrTransform_BgColor extends HTMLPurifier_AttrTransform\r
+{\r
+ /**\r
+ * @param array $attr\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return array\r
+ */\r
+ public function transform($attr, $config, $context)\r
+ {\r
+ if (!isset($attr['bgcolor'])) {\r
+ return $attr;\r
+ }\r
+\r
+ $bgcolor = $this->confiscateAttr($attr, 'bgcolor');\r
+ // some validation should happen here\r
+\r
+ $this->prependCSS($attr, "background-color:$bgcolor;");\r
+ return $attr;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Pre-transform that changes converts a boolean attribute to fixed CSS\r
+ */\r
+class HTMLPurifier_AttrTransform_BoolToCSS extends HTMLPurifier_AttrTransform\r
+{\r
+ /**\r
+ * Name of boolean attribute that is trigger.\r
+ * @type string\r
+ */\r
+ protected $attr;\r
+\r
+ /**\r
+ * CSS declarations to add to style, needs trailing semicolon.\r
+ * @type string\r
+ */\r
+ protected $css;\r
+\r
+ /**\r
+ * @param string $attr attribute name to convert from\r
+ * @param string $css CSS declarations to add to style (needs semicolon)\r
+ */\r
+ public function __construct($attr, $css)\r
+ {\r
+ $this->attr = $attr;\r
+ $this->css = $css;\r
+ }\r
+\r
+ /**\r
+ * @param array $attr\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return array\r
+ */\r
+ public function transform($attr, $config, $context)\r
+ {\r
+ if (!isset($attr[$this->attr])) {\r
+ return $attr;\r
+ }\r
+ unset($attr[$this->attr]);\r
+ $this->prependCSS($attr, $this->css);\r
+ return $attr;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Pre-transform that changes deprecated border attribute to CSS.\r
+ */\r
+class HTMLPurifier_AttrTransform_Border extends HTMLPurifier_AttrTransform\r
+{\r
+ /**\r
+ * @param array $attr\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return array\r
+ */\r
+ public function transform($attr, $config, $context)\r
+ {\r
+ if (!isset($attr['border'])) {\r
+ return $attr;\r
+ }\r
+ $border_width = $this->confiscateAttr($attr, 'border');\r
+ // some validation should happen here\r
+ $this->prependCSS($attr, "border:{$border_width}px solid;");\r
+ return $attr;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Generic pre-transform that converts an attribute with a fixed number of\r
+ * values (enumerated) to CSS.\r
+ */\r
+class HTMLPurifier_AttrTransform_EnumToCSS extends HTMLPurifier_AttrTransform\r
+{\r
+ /**\r
+ * Name of attribute to transform from.\r
+ * @type string\r
+ */\r
+ protected $attr;\r
+\r
+ /**\r
+ * Lookup array of attribute values to CSS.\r
+ * @type array\r
+ */\r
+ protected $enumToCSS = array();\r
+\r
+ /**\r
+ * Case sensitivity of the matching.\r
+ * @type bool\r
+ * @warning Currently can only be guaranteed to work with ASCII\r
+ * values.\r
+ */\r
+ protected $caseSensitive = false;\r
+\r
+ /**\r
+ * @param string $attr Attribute name to transform from\r
+ * @param array $enum_to_css Lookup array of attribute values to CSS\r
+ * @param bool $case_sensitive Case sensitivity indicator, default false\r
+ */\r
+ public function __construct($attr, $enum_to_css, $case_sensitive = false)\r
+ {\r
+ $this->attr = $attr;\r
+ $this->enumToCSS = $enum_to_css;\r
+ $this->caseSensitive = (bool)$case_sensitive;\r
+ }\r
+\r
+ /**\r
+ * @param array $attr\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return array\r
+ */\r
+ public function transform($attr, $config, $context)\r
+ {\r
+ if (!isset($attr[$this->attr])) {\r
+ return $attr;\r
+ }\r
+\r
+ $value = trim($attr[$this->attr]);\r
+ unset($attr[$this->attr]);\r
+\r
+ if (!$this->caseSensitive) {\r
+ $value = strtolower($value);\r
+ }\r
+\r
+ if (!isset($this->enumToCSS[$value])) {\r
+ return $attr;\r
+ }\r
+ $this->prependCSS($attr, $this->enumToCSS[$value]);\r
+ return $attr;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+// must be called POST validation\r
+\r
+/**\r
+ * Transform that supplies default values for the src and alt attributes\r
+ * in img tags, as well as prevents the img tag from being removed\r
+ * because of a missing alt tag. This needs to be registered as both\r
+ * a pre and post attribute transform.\r
+ */\r
+class HTMLPurifier_AttrTransform_ImgRequired extends HTMLPurifier_AttrTransform\r
+{\r
+\r
+ /**\r
+ * @param array $attr\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return array\r
+ */\r
+ public function transform($attr, $config, $context)\r
+ {\r
+ $src = true;\r
+ if (!isset($attr['src'])) {\r
+ if ($config->get('Core.RemoveInvalidImg')) {\r
+ return $attr;\r
+ }\r
+ $attr['src'] = $config->get('Attr.DefaultInvalidImage');\r
+ $src = false;\r
+ }\r
+\r
+ if (!isset($attr['alt'])) {\r
+ if ($src) {\r
+ $alt = $config->get('Attr.DefaultImageAlt');\r
+ if ($alt === null) {\r
+ $attr['alt'] = basename($attr['src']);\r
+ } else {\r
+ $attr['alt'] = $alt;\r
+ }\r
+ } else {\r
+ $attr['alt'] = $config->get('Attr.DefaultInvalidImageAlt');\r
+ }\r
+ }\r
+ return $attr;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Pre-transform that changes deprecated hspace and vspace attributes to CSS\r
+ */\r
+class HTMLPurifier_AttrTransform_ImgSpace extends HTMLPurifier_AttrTransform\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ protected $attr;\r
+\r
+ /**\r
+ * @type array\r
+ */\r
+ protected $css = array(\r
+ 'hspace' => array('left', 'right'),\r
+ 'vspace' => array('top', 'bottom')\r
+ );\r
+\r
+ /**\r
+ * @param string $attr\r
+ */\r
+ public function __construct($attr)\r
+ {\r
+ $this->attr = $attr;\r
+ if (!isset($this->css[$attr])) {\r
+ trigger_error(htmlspecialchars($attr) . ' is not valid space attribute');\r
+ }\r
+ }\r
+\r
+ /**\r
+ * @param array $attr\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return array\r
+ */\r
+ public function transform($attr, $config, $context)\r
+ {\r
+ if (!isset($attr[$this->attr])) {\r
+ return $attr;\r
+ }\r
+\r
+ $width = $this->confiscateAttr($attr, $this->attr);\r
+ // some validation could happen here\r
+\r
+ if (!isset($this->css[$this->attr])) {\r
+ return $attr;\r
+ }\r
+\r
+ $style = '';\r
+ foreach ($this->css[$this->attr] as $suffix) {\r
+ $property = "margin-$suffix";\r
+ $style .= "$property:{$width}px;";\r
+ }\r
+ $this->prependCSS($attr, $style);\r
+ return $attr;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Performs miscellaneous cross attribute validation and filtering for\r
+ * input elements. This is meant to be a post-transform.\r
+ */\r
+class HTMLPurifier_AttrTransform_Input extends HTMLPurifier_AttrTransform\r
+{\r
+ /**\r
+ * @type HTMLPurifier_AttrDef_HTML_Pixels\r
+ */\r
+ protected $pixels;\r
+\r
+ public function __construct()\r
+ {\r
+ $this->pixels = new HTMLPurifier_AttrDef_HTML_Pixels();\r
+ }\r
+\r
+ /**\r
+ * @param array $attr\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return array\r
+ */\r
+ public function transform($attr, $config, $context)\r
+ {\r
+ if (!isset($attr['type'])) {\r
+ $t = 'text';\r
+ } else {\r
+ $t = strtolower($attr['type']);\r
+ }\r
+ if (isset($attr['checked']) && $t !== 'radio' && $t !== 'checkbox') {\r
+ unset($attr['checked']);\r
+ }\r
+ if (isset($attr['maxlength']) && $t !== 'text' && $t !== 'password') {\r
+ unset($attr['maxlength']);\r
+ }\r
+ if (isset($attr['size']) && $t !== 'text' && $t !== 'password') {\r
+ $result = $this->pixels->validate($attr['size'], $config, $context);\r
+ if ($result === false) {\r
+ unset($attr['size']);\r
+ } else {\r
+ $attr['size'] = $result;\r
+ }\r
+ }\r
+ if (isset($attr['src']) && $t !== 'image') {\r
+ unset($attr['src']);\r
+ }\r
+ if (!isset($attr['value']) && ($t === 'radio' || $t === 'checkbox')) {\r
+ $attr['value'] = '';\r
+ }\r
+ return $attr;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Post-transform that copies lang's value to xml:lang (and vice-versa)\r
+ * @note Theoretically speaking, this could be a pre-transform, but putting\r
+ * post is more efficient.\r
+ */\r
+class HTMLPurifier_AttrTransform_Lang extends HTMLPurifier_AttrTransform\r
+{\r
+\r
+ /**\r
+ * @param array $attr\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return array\r
+ */\r
+ public function transform($attr, $config, $context)\r
+ {\r
+ $lang = isset($attr['lang']) ? $attr['lang'] : false;\r
+ $xml_lang = isset($attr['xml:lang']) ? $attr['xml:lang'] : false;\r
+\r
+ if ($lang !== false && $xml_lang === false) {\r
+ $attr['xml:lang'] = $lang;\r
+ } elseif ($xml_lang !== false) {\r
+ $attr['lang'] = $xml_lang;\r
+ }\r
+ return $attr;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Class for handling width/height length attribute transformations to CSS\r
+ */\r
+class HTMLPurifier_AttrTransform_Length extends HTMLPurifier_AttrTransform\r
+{\r
+\r
+ /**\r
+ * @type string\r
+ */\r
+ protected $name;\r
+\r
+ /**\r
+ * @type string\r
+ */\r
+ protected $cssName;\r
+\r
+ public function __construct($name, $css_name = null)\r
+ {\r
+ $this->name = $name;\r
+ $this->cssName = $css_name ? $css_name : $name;\r
+ }\r
+\r
+ /**\r
+ * @param array $attr\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return array\r
+ */\r
+ public function transform($attr, $config, $context)\r
+ {\r
+ if (!isset($attr[$this->name])) {\r
+ return $attr;\r
+ }\r
+ $length = $this->confiscateAttr($attr, $this->name);\r
+ if (ctype_digit($length)) {\r
+ $length .= 'px';\r
+ }\r
+ $this->prependCSS($attr, $this->cssName . ":$length;");\r
+ return $attr;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Pre-transform that changes deprecated name attribute to ID if necessary\r
+ */\r
+class HTMLPurifier_AttrTransform_Name extends HTMLPurifier_AttrTransform\r
+{\r
+\r
+ /**\r
+ * @param array $attr\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return array\r
+ */\r
+ public function transform($attr, $config, $context)\r
+ {\r
+ // Abort early if we're using relaxed definition of name\r
+ if ($config->get('HTML.Attr.Name.UseCDATA')) {\r
+ return $attr;\r
+ }\r
+ if (!isset($attr['name'])) {\r
+ return $attr;\r
+ }\r
+ $id = $this->confiscateAttr($attr, 'name');\r
+ if (isset($attr['id'])) {\r
+ return $attr;\r
+ }\r
+ $attr['id'] = $id;\r
+ return $attr;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Post-transform that performs validation to the name attribute; if\r
+ * it is present with an equivalent id attribute, it is passed through;\r
+ * otherwise validation is performed.\r
+ */\r
+class HTMLPurifier_AttrTransform_NameSync extends HTMLPurifier_AttrTransform\r
+{\r
+\r
+ public function __construct()\r
+ {\r
+ $this->idDef = new HTMLPurifier_AttrDef_HTML_ID();\r
+ }\r
+\r
+ /**\r
+ * @param array $attr\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return array\r
+ */\r
+ public function transform($attr, $config, $context)\r
+ {\r
+ if (!isset($attr['name'])) {\r
+ return $attr;\r
+ }\r
+ $name = $attr['name'];\r
+ if (isset($attr['id']) && $attr['id'] === $name) {\r
+ return $attr;\r
+ }\r
+ $result = $this->idDef->validate($name, $config, $context);\r
+ if ($result === false) {\r
+ unset($attr['name']);\r
+ } else {\r
+ $attr['name'] = $result;\r
+ }\r
+ return $attr;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+// must be called POST validation\r
+\r
+/**\r
+ * Adds rel="nofollow" to all outbound links. This transform is\r
+ * only attached if Attr.Nofollow is TRUE.\r
+ */\r
+class HTMLPurifier_AttrTransform_Nofollow extends HTMLPurifier_AttrTransform\r
+{\r
+ /**\r
+ * @type HTMLPurifier_URIParser\r
+ */\r
+ private $parser;\r
+\r
+ public function __construct()\r
+ {\r
+ $this->parser = new HTMLPurifier_URIParser();\r
+ }\r
+\r
+ /**\r
+ * @param array $attr\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return array\r
+ */\r
+ public function transform($attr, $config, $context)\r
+ {\r
+ if (!isset($attr['href'])) {\r
+ return $attr;\r
+ }\r
+\r
+ // XXX Kind of inefficient\r
+ $url = $this->parser->parse($attr['href']);\r
+ $scheme = $url->getSchemeObj($config, $context);\r
+\r
+ if ($scheme->browsable && !$url->isLocal($config, $context)) {\r
+ if (isset($attr['rel'])) {\r
+ $rels = explode(' ', $attr['rel']);\r
+ if (!in_array('nofollow', $rels)) {\r
+ $rels[] = 'nofollow';\r
+ }\r
+ $attr['rel'] = implode(' ', $rels);\r
+ } else {\r
+ $attr['rel'] = 'nofollow';\r
+ }\r
+ }\r
+ return $attr;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+class HTMLPurifier_AttrTransform_SafeEmbed extends HTMLPurifier_AttrTransform\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = "SafeEmbed";\r
+\r
+ /**\r
+ * @param array $attr\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return array\r
+ */\r
+ public function transform($attr, $config, $context)\r
+ {\r
+ $attr['allowscriptaccess'] = 'never';\r
+ $attr['allownetworking'] = 'internal';\r
+ $attr['type'] = 'application/x-shockwave-flash';\r
+ return $attr;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Writes default type for all objects. Currently only supports flash.\r
+ */\r
+class HTMLPurifier_AttrTransform_SafeObject extends HTMLPurifier_AttrTransform\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = "SafeObject";\r
+\r
+ /**\r
+ * @param array $attr\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return array\r
+ */\r
+ public function transform($attr, $config, $context)\r
+ {\r
+ if (!isset($attr['type'])) {\r
+ $attr['type'] = 'application/x-shockwave-flash';\r
+ }\r
+ return $attr;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Validates name/value pairs in param tags to be used in safe objects. This\r
+ * will only allow name values it recognizes, and pre-fill certain attributes\r
+ * with required values.\r
+ *\r
+ * @note\r
+ * This class only supports Flash. In the future, Quicktime support\r
+ * may be added.\r
+ *\r
+ * @warning\r
+ * This class expects an injector to add the necessary parameters tags.\r
+ */\r
+class HTMLPurifier_AttrTransform_SafeParam extends HTMLPurifier_AttrTransform\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = "SafeParam";\r
+\r
+ /**\r
+ * @type HTMLPurifier_AttrDef_URI\r
+ */\r
+ private $uri;\r
+\r
+ public function __construct()\r
+ {\r
+ $this->uri = new HTMLPurifier_AttrDef_URI(true); // embedded\r
+ $this->wmode = new HTMLPurifier_AttrDef_Enum(array('window', 'opaque', 'transparent'));\r
+ }\r
+\r
+ /**\r
+ * @param array $attr\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return array\r
+ */\r
+ public function transform($attr, $config, $context)\r
+ {\r
+ // If we add support for other objects, we'll need to alter the\r
+ // transforms.\r
+ switch ($attr['name']) {\r
+ // application/x-shockwave-flash\r
+ // Keep this synchronized with Injector/SafeObject.php\r
+ case 'allowScriptAccess':\r
+ $attr['value'] = 'never';\r
+ break;\r
+ case 'allowNetworking':\r
+ $attr['value'] = 'internal';\r
+ break;\r
+ case 'allowFullScreen':\r
+ if ($config->get('HTML.FlashAllowFullScreen')) {\r
+ $attr['value'] = ($attr['value'] == 'true') ? 'true' : 'false';\r
+ } else {\r
+ $attr['value'] = 'false';\r
+ }\r
+ break;\r
+ case 'wmode':\r
+ $attr['value'] = $this->wmode->validate($attr['value'], $config, $context);\r
+ break;\r
+ case 'movie':\r
+ case 'src':\r
+ $attr['name'] = "movie";\r
+ $attr['value'] = $this->uri->validate($attr['value'], $config, $context);\r
+ break;\r
+ case 'flashvars':\r
+ // we're going to allow arbitrary inputs to the SWF, on\r
+ // the reasoning that it could only hack the SWF, not us.\r
+ break;\r
+ // add other cases to support other param name/value pairs\r
+ default:\r
+ $attr['name'] = $attr['value'] = null;\r
+ }\r
+ return $attr;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Implements required attribute stipulation for <script>\r
+ */\r
+class HTMLPurifier_AttrTransform_ScriptRequired extends HTMLPurifier_AttrTransform\r
+{\r
+ /**\r
+ * @param array $attr\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return array\r
+ */\r
+ public function transform($attr, $config, $context)\r
+ {\r
+ if (!isset($attr['type'])) {\r
+ $attr['type'] = 'text/javascript';\r
+ }\r
+ return $attr;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+// must be called POST validation\r
+\r
+/**\r
+ * Adds target="blank" to all outbound links. This transform is\r
+ * only attached if Attr.TargetBlank is TRUE. This works regardless\r
+ * of whether or not Attr.AllowedFrameTargets\r
+ */\r
+class HTMLPurifier_AttrTransform_TargetBlank extends HTMLPurifier_AttrTransform\r
+{\r
+ /**\r
+ * @type HTMLPurifier_URIParser\r
+ */\r
+ private $parser;\r
+\r
+ public function __construct()\r
+ {\r
+ $this->parser = new HTMLPurifier_URIParser();\r
+ }\r
+\r
+ /**\r
+ * @param array $attr\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return array\r
+ */\r
+ public function transform($attr, $config, $context)\r
+ {\r
+ if (!isset($attr['href'])) {\r
+ return $attr;\r
+ }\r
+\r
+ // XXX Kind of inefficient\r
+ $url = $this->parser->parse($attr['href']);\r
+ $scheme = $url->getSchemeObj($config, $context);\r
+\r
+ if ($scheme->browsable && !$url->isBenign($config, $context)) {\r
+ $attr['target'] = '_blank';\r
+ }\r
+ return $attr;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+// must be called POST validation\r
+\r
+/**\r
+ * Adds rel="noopener" to any links which target a different window\r
+ * than the current one. This is used to prevent malicious websites\r
+ * from silently replacing the original window, which could be used\r
+ * to do phishing.\r
+ * This transform is controlled by %HTML.TargetNoopener.\r
+ */\r
+class HTMLPurifier_AttrTransform_TargetNoopener extends HTMLPurifier_AttrTransform\r
+{\r
+ /**\r
+ * @param array $attr\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return array\r
+ */\r
+ public function transform($attr, $config, $context)\r
+ {\r
+ if (isset($attr['rel'])) {\r
+ $rels = explode(' ', $attr['rel']);\r
+ } else {\r
+ $rels = array();\r
+ }\r
+ if (isset($attr['target']) && !in_array('noopener', $rels)) {\r
+ $rels[] = 'noopener';\r
+ }\r
+ if (!empty($rels) || isset($attr['rel'])) {\r
+ $attr['rel'] = implode(' ', $rels);\r
+ }\r
+\r
+ return $attr;\r
+ }\r
+}\r
+\r
+
+\r
+\r
+// must be called POST validation\r
+\r
+/**\r
+ * Adds rel="noreferrer" to any links which target a different window\r
+ * than the current one. This is used to prevent malicious websites\r
+ * from silently replacing the original window, which could be used\r
+ * to do phishing.\r
+ * This transform is controlled by %HTML.TargetNoreferrer.\r
+ */\r
+class HTMLPurifier_AttrTransform_TargetNoreferrer extends HTMLPurifier_AttrTransform\r
+{\r
+ /**\r
+ * @param array $attr\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return array\r
+ */\r
+ public function transform($attr, $config, $context)\r
+ {\r
+ if (isset($attr['rel'])) {\r
+ $rels = explode(' ', $attr['rel']);\r
+ } else {\r
+ $rels = array();\r
+ }\r
+ if (isset($attr['target']) && !in_array('noreferrer', $rels)) {\r
+ $rels[] = 'noreferrer';\r
+ }\r
+ if (!empty($rels) || isset($attr['rel'])) {\r
+ $attr['rel'] = implode(' ', $rels);\r
+ }\r
+\r
+ return $attr;\r
+ }\r
+}\r
+\r
+
+\r
+\r
+/**\r
+ * Sets height/width defaults for <textarea>\r
+ */\r
+class HTMLPurifier_AttrTransform_Textarea extends HTMLPurifier_AttrTransform\r
+{\r
+ /**\r
+ * @param array $attr\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return array\r
+ */\r
+ public function transform($attr, $config, $context)\r
+ {\r
+ // Calculated from Firefox\r
+ if (!isset($attr['cols'])) {\r
+ $attr['cols'] = '22';\r
+ }\r
+ if (!isset($attr['rows'])) {\r
+ $attr['rows'] = '3';\r
+ }\r
+ return $attr;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Definition that uses different definitions depending on context.\r
+ *\r
+ * The del and ins tags are notable because they allow different types of\r
+ * elements depending on whether or not they're in a block or inline context.\r
+ * Chameleon allows this behavior to happen by using two different\r
+ * definitions depending on context. While this somewhat generalized,\r
+ * it is specifically intended for those two tags.\r
+ */\r
+class HTMLPurifier_ChildDef_Chameleon extends HTMLPurifier_ChildDef\r
+{\r
+\r
+ /**\r
+ * Instance of the definition object to use when inline. Usually stricter.\r
+ * @type HTMLPurifier_ChildDef_Optional\r
+ */\r
+ public $inline;\r
+\r
+ /**\r
+ * Instance of the definition object to use when block.\r
+ * @type HTMLPurifier_ChildDef_Optional\r
+ */\r
+ public $block;\r
+\r
+ /**\r
+ * @type string\r
+ */\r
+ public $type = 'chameleon';\r
+\r
+ /**\r
+ * @param array $inline List of elements to allow when inline.\r
+ * @param array $block List of elements to allow when block.\r
+ */\r
+ public function __construct($inline, $block)\r
+ {\r
+ $this->inline = new HTMLPurifier_ChildDef_Optional($inline);\r
+ $this->block = new HTMLPurifier_ChildDef_Optional($block);\r
+ $this->elements = $this->block->elements;\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Node[] $children\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool\r
+ */\r
+ public function validateChildren($children, $config, $context)\r
+ {\r
+ if ($context->get('IsInline') === false) {\r
+ return $this->block->validateChildren(\r
+ $children,\r
+ $config,\r
+ $context\r
+ );\r
+ } else {\r
+ return $this->inline->validateChildren(\r
+ $children,\r
+ $config,\r
+ $context\r
+ );\r
+ }\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Custom validation class, accepts DTD child definitions\r
+ *\r
+ * @warning Currently this class is an all or nothing proposition, that is,\r
+ * it will only give a bool return value.\r
+ */\r
+class HTMLPurifier_ChildDef_Custom extends HTMLPurifier_ChildDef\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $type = 'custom';\r
+\r
+ /**\r
+ * @type bool\r
+ */\r
+ public $allow_empty = false;\r
+\r
+ /**\r
+ * Allowed child pattern as defined by the DTD.\r
+ * @type string\r
+ */\r
+ public $dtd_regex;\r
+\r
+ /**\r
+ * PCRE regex derived from $dtd_regex.\r
+ * @type string\r
+ */\r
+ private $_pcre_regex;\r
+\r
+ /**\r
+ * @param $dtd_regex Allowed child pattern from the DTD\r
+ */\r
+ public function __construct($dtd_regex)\r
+ {\r
+ $this->dtd_regex = $dtd_regex;\r
+ $this->_compileRegex();\r
+ }\r
+\r
+ /**\r
+ * Compiles the PCRE regex from a DTD regex ($dtd_regex to $_pcre_regex)\r
+ */\r
+ protected function _compileRegex()\r
+ {\r
+ $raw = str_replace(' ', '', $this->dtd_regex);\r
+ if ($raw{0} != '(') {\r
+ $raw = "($raw)";\r
+ }\r
+ $el = '[#a-zA-Z0-9_.-]+';\r
+ $reg = $raw;\r
+\r
+ // COMPLICATED! AND MIGHT BE BUGGY! I HAVE NO CLUE WHAT I'M\r
+ // DOING! Seriously: if there's problems, please report them.\r
+\r
+ // collect all elements into the $elements array\r
+ preg_match_all("/$el/", $reg, $matches);\r
+ foreach ($matches[0] as $match) {\r
+ $this->elements[$match] = true;\r
+ }\r
+\r
+ // setup all elements as parentheticals with leading commas\r
+ $reg = preg_replace("/$el/", '(,\\0)', $reg);\r
+\r
+ // remove commas when they were not solicited\r
+ $reg = preg_replace("/([^,(|]\(+),/", '\\1', $reg);\r
+\r
+ // remove all non-paranthetical commas: they are handled by first regex\r
+ $reg = preg_replace("/,\(/", '(', $reg);\r
+\r
+ $this->_pcre_regex = $reg;\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Node[] $children\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool\r
+ */\r
+ public function validateChildren($children, $config, $context)\r
+ {\r
+ $list_of_children = '';\r
+ $nesting = 0; // depth into the nest\r
+ foreach ($children as $node) {\r
+ if (!empty($node->is_whitespace)) {\r
+ continue;\r
+ }\r
+ $list_of_children .= $node->name . ',';\r
+ }\r
+ // add leading comma to deal with stray comma declarations\r
+ $list_of_children = ',' . rtrim($list_of_children, ',');\r
+ $okay =\r
+ preg_match(\r
+ '/^,?' . $this->_pcre_regex . '$/',\r
+ $list_of_children\r
+ );\r
+ return (bool)$okay;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Definition that disallows all elements.\r
+ * @warning validateChildren() in this class is actually never called, because\r
+ * empty elements are corrected in HTMLPurifier_Strategy_MakeWellFormed\r
+ * before child definitions are parsed in earnest by\r
+ * HTMLPurifier_Strategy_FixNesting.\r
+ */\r
+class HTMLPurifier_ChildDef_Empty extends HTMLPurifier_ChildDef\r
+{\r
+ /**\r
+ * @type bool\r
+ */\r
+ public $allow_empty = true;\r
+\r
+ /**\r
+ * @type string\r
+ */\r
+ public $type = 'empty';\r
+\r
+ public function __construct()\r
+ {\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Node[] $children\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return array\r
+ */\r
+ public function validateChildren($children, $config, $context)\r
+ {\r
+ return array();\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Definition for list containers ul and ol.\r
+ *\r
+ * What does this do? The big thing is to handle ol/ul at the top\r
+ * level of list nodes, which should be handled specially by /folding/\r
+ * them into the previous list node. We generally shouldn't ever\r
+ * see other disallowed elements, because the autoclose behavior\r
+ * in MakeWellFormed handles it.\r
+ */\r
+class HTMLPurifier_ChildDef_List extends HTMLPurifier_ChildDef\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $type = 'list';\r
+ /**\r
+ * @type array\r
+ */\r
+ // lying a little bit, so that we can handle ul and ol ourselves\r
+ // XXX: This whole business with 'wrap' is all a bit unsatisfactory\r
+ public $elements = array('li' => true, 'ul' => true, 'ol' => true);\r
+\r
+ /**\r
+ * @param array $children\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return array\r
+ */\r
+ public function validateChildren($children, $config, $context)\r
+ {\r
+ // Flag for subclasses\r
+ $this->whitespace = false;\r
+\r
+ // if there are no tokens, delete parent node\r
+ if (empty($children)) {\r
+ return false;\r
+ }\r
+\r
+ // if li is not allowed, delete parent node\r
+ if (!isset($config->getHTMLDefinition()->info['li'])) {\r
+ trigger_error("Cannot allow ul/ol without allowing li", E_USER_WARNING);\r
+ return false;\r
+ }\r
+\r
+ // the new set of children\r
+ $result = array();\r
+\r
+ // a little sanity check to make sure it's not ALL whitespace\r
+ $all_whitespace = true;\r
+\r
+ $current_li = null;\r
+\r
+ foreach ($children as $node) {\r
+ if (!empty($node->is_whitespace)) {\r
+ $result[] = $node;\r
+ continue;\r
+ }\r
+ $all_whitespace = false; // phew, we're not talking about whitespace\r
+\r
+ if ($node->name === 'li') {\r
+ // good\r
+ $current_li = $node;\r
+ $result[] = $node;\r
+ } else {\r
+ // we want to tuck this into the previous li\r
+ // Invariant: we expect the node to be ol/ul\r
+ // ToDo: Make this more robust in the case of not ol/ul\r
+ // by distinguishing between existing li and li created\r
+ // to handle non-list elements; non-list elements should\r
+ // not be appended to an existing li; only li created\r
+ // for non-list. This distinction is not currently made.\r
+ if ($current_li === null) {\r
+ $current_li = new HTMLPurifier_Node_Element('li');\r
+ $result[] = $current_li;\r
+ }\r
+ $current_li->children[] = $node;\r
+ $current_li->empty = false; // XXX fascinating! Check for this error elsewhere ToDo\r
+ }\r
+ }\r
+ if (empty($result)) {\r
+ return false;\r
+ }\r
+ if ($all_whitespace) {\r
+ return false;\r
+ }\r
+ return $result;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Definition that allows a set of elements, but disallows empty children.\r
+ */\r
+class HTMLPurifier_ChildDef_Required extends HTMLPurifier_ChildDef\r
+{\r
+ /**\r
+ * Lookup table of allowed elements.\r
+ * @type array\r
+ */\r
+ public $elements = array();\r
+\r
+ /**\r
+ * Whether or not the last passed node was all whitespace.\r
+ * @type bool\r
+ */\r
+ protected $whitespace = false;\r
+\r
+ /**\r
+ * @param array|string $elements List of allowed element names (lowercase).\r
+ */\r
+ public function __construct($elements)\r
+ {\r
+ if (is_string($elements)) {\r
+ $elements = str_replace(' ', '', $elements);\r
+ $elements = explode('|', $elements);\r
+ }\r
+ $keys = array_keys($elements);\r
+ if ($keys == array_keys($keys)) {\r
+ $elements = array_flip($elements);\r
+ foreach ($elements as $i => $x) {\r
+ $elements[$i] = true;\r
+ if (empty($i)) {\r
+ unset($elements[$i]);\r
+ } // remove blank\r
+ }\r
+ }\r
+ $this->elements = $elements;\r
+ }\r
+\r
+ /**\r
+ * @type bool\r
+ */\r
+ public $allow_empty = false;\r
+\r
+ /**\r
+ * @type string\r
+ */\r
+ public $type = 'required';\r
+\r
+ /**\r
+ * @param array $children\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return array\r
+ */\r
+ public function validateChildren($children, $config, $context)\r
+ {\r
+ // Flag for subclasses\r
+ $this->whitespace = false;\r
+\r
+ // if there are no tokens, delete parent node\r
+ if (empty($children)) {\r
+ return false;\r
+ }\r
+\r
+ // the new set of children\r
+ $result = array();\r
+\r
+ // whether or not parsed character data is allowed\r
+ // this controls whether or not we silently drop a tag\r
+ // or generate escaped HTML from it\r
+ $pcdata_allowed = isset($this->elements['#PCDATA']);\r
+\r
+ // a little sanity check to make sure it's not ALL whitespace\r
+ $all_whitespace = true;\r
+\r
+ $stack = array_reverse($children);\r
+ while (!empty($stack)) {\r
+ $node = array_pop($stack);\r
+ if (!empty($node->is_whitespace)) {\r
+ $result[] = $node;\r
+ continue;\r
+ }\r
+ $all_whitespace = false; // phew, we're not talking about whitespace\r
+\r
+ if (!isset($this->elements[$node->name])) {\r
+ // special case text\r
+ // XXX One of these ought to be redundant or something\r
+ if ($pcdata_allowed && $node instanceof HTMLPurifier_Node_Text) {\r
+ $result[] = $node;\r
+ continue;\r
+ }\r
+ // spill the child contents in\r
+ // ToDo: Make configurable\r
+ if ($node instanceof HTMLPurifier_Node_Element) {\r
+ for ($i = count($node->children) - 1; $i >= 0; $i--) {\r
+ $stack[] = $node->children[$i];\r
+ }\r
+ continue;\r
+ }\r
+ continue;\r
+ }\r
+ $result[] = $node;\r
+ }\r
+ if (empty($result)) {\r
+ return false;\r
+ }\r
+ if ($all_whitespace) {\r
+ $this->whitespace = true;\r
+ return false;\r
+ }\r
+ return $result;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Definition that allows a set of elements, and allows no children.\r
+ * @note This is a hack to reuse code from HTMLPurifier_ChildDef_Required,\r
+ * really, one shouldn't inherit from the other. Only altered behavior\r
+ * is to overload a returned false with an array. Thus, it will never\r
+ * return false.\r
+ */\r
+class HTMLPurifier_ChildDef_Optional extends HTMLPurifier_ChildDef_Required\r
+{\r
+ /**\r
+ * @type bool\r
+ */\r
+ public $allow_empty = true;\r
+\r
+ /**\r
+ * @type string\r
+ */\r
+ public $type = 'optional';\r
+\r
+ /**\r
+ * @param array $children\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return array\r
+ */\r
+ public function validateChildren($children, $config, $context)\r
+ {\r
+ $result = parent::validateChildren($children, $config, $context);\r
+ // we assume that $children is not modified\r
+ if ($result === false) {\r
+ if (empty($children)) {\r
+ return true;\r
+ } elseif ($this->whitespace) {\r
+ return $children;\r
+ } else {\r
+ return array();\r
+ }\r
+ }\r
+ return $result;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Takes the contents of blockquote when in strict and reformats for validation.\r
+ */\r
+class HTMLPurifier_ChildDef_StrictBlockquote extends HTMLPurifier_ChildDef_Required\r
+{\r
+ /**\r
+ * @type array\r
+ */\r
+ protected $real_elements;\r
+\r
+ /**\r
+ * @type array\r
+ */\r
+ protected $fake_elements;\r
+\r
+ /**\r
+ * @type bool\r
+ */\r
+ public $allow_empty = true;\r
+\r
+ /**\r
+ * @type string\r
+ */\r
+ public $type = 'strictblockquote';\r
+\r
+ /**\r
+ * @type bool\r
+ */\r
+ protected $init = false;\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ * @return array\r
+ * @note We don't want MakeWellFormed to auto-close inline elements since\r
+ * they might be allowed.\r
+ */\r
+ public function getAllowedElements($config)\r
+ {\r
+ $this->init($config);\r
+ return $this->fake_elements;\r
+ }\r
+\r
+ /**\r
+ * @param array $children\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return array\r
+ */\r
+ public function validateChildren($children, $config, $context)\r
+ {\r
+ $this->init($config);\r
+\r
+ // trick the parent class into thinking it allows more\r
+ $this->elements = $this->fake_elements;\r
+ $result = parent::validateChildren($children, $config, $context);\r
+ $this->elements = $this->real_elements;\r
+\r
+ if ($result === false) {\r
+ return array();\r
+ }\r
+ if ($result === true) {\r
+ $result = $children;\r
+ }\r
+\r
+ $def = $config->getHTMLDefinition();\r
+ $block_wrap_name = $def->info_block_wrapper;\r
+ $block_wrap = false;\r
+ $ret = array();\r
+\r
+ foreach ($result as $node) {\r
+ if ($block_wrap === false) {\r
+ if (($node instanceof HTMLPurifier_Node_Text && !$node->is_whitespace) ||\r
+ ($node instanceof HTMLPurifier_Node_Element && !isset($this->elements[$node->name]))) {\r
+ $block_wrap = new HTMLPurifier_Node_Element($def->info_block_wrapper);\r
+ $ret[] = $block_wrap;\r
+ }\r
+ } else {\r
+ if ($node instanceof HTMLPurifier_Node_Element && isset($this->elements[$node->name])) {\r
+ $block_wrap = false;\r
+\r
+ }\r
+ }\r
+ if ($block_wrap) {\r
+ $block_wrap->children[] = $node;\r
+ } else {\r
+ $ret[] = $node;\r
+ }\r
+ }\r
+ return $ret;\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ private function init($config)\r
+ {\r
+ if (!$this->init) {\r
+ $def = $config->getHTMLDefinition();\r
+ // allow all inline elements\r
+ $this->real_elements = $this->elements;\r
+ $this->fake_elements = $def->info_content_sets['Flow'];\r
+ $this->fake_elements['#PCDATA'] = true;\r
+ $this->init = true;\r
+ }\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Definition for tables. The general idea is to extract out all of the\r
+ * essential bits, and then reconstruct it later.\r
+ *\r
+ * This is a bit confusing, because the DTDs and the W3C\r
+ * validators seem to disagree on the appropriate definition. The\r
+ * DTD claims:\r
+ *\r
+ * (CAPTION?, (COL*|COLGROUP*), THEAD?, TFOOT?, TBODY+)\r
+ *\r
+ * But actually, the HTML4 spec then has this to say:\r
+ *\r
+ * The TBODY start tag is always required except when the table\r
+ * contains only one table body and no table head or foot sections.\r
+ * The TBODY end tag may always be safely omitted.\r
+ *\r
+ * So the DTD is kind of wrong. The validator is, unfortunately, kind\r
+ * of on crack.\r
+ *\r
+ * The definition changed again in XHTML1.1; and in my opinion, this\r
+ * formulation makes the most sense.\r
+ *\r
+ * caption?, ( col* | colgroup* ), (( thead?, tfoot?, tbody+ ) | ( tr+ ))\r
+ *\r
+ * Essentially, we have two modes: thead/tfoot/tbody mode, and tr mode.\r
+ * If we encounter a thead, tfoot or tbody, we are placed in the former\r
+ * mode, and we *must* wrap any stray tr segments with a tbody. But if\r
+ * we don't run into any of them, just have tr tags is OK.\r
+ */\r
+class HTMLPurifier_ChildDef_Table extends HTMLPurifier_ChildDef\r
+{\r
+ /**\r
+ * @type bool\r
+ */\r
+ public $allow_empty = false;\r
+\r
+ /**\r
+ * @type string\r
+ */\r
+ public $type = 'table';\r
+\r
+ /**\r
+ * @type array\r
+ */\r
+ public $elements = array(\r
+ 'tr' => true,\r
+ 'tbody' => true,\r
+ 'thead' => true,\r
+ 'tfoot' => true,\r
+ 'caption' => true,\r
+ 'colgroup' => true,\r
+ 'col' => true\r
+ );\r
+\r
+ public function __construct()\r
+ {\r
+ }\r
+\r
+ /**\r
+ * @param array $children\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return array\r
+ */\r
+ public function validateChildren($children, $config, $context)\r
+ {\r
+ if (empty($children)) {\r
+ return false;\r
+ }\r
+\r
+ // only one of these elements is allowed in a table\r
+ $caption = false;\r
+ $thead = false;\r
+ $tfoot = false;\r
+\r
+ // whitespace\r
+ $initial_ws = array();\r
+ $after_caption_ws = array();\r
+ $after_thead_ws = array();\r
+ $after_tfoot_ws = array();\r
+\r
+ // as many of these as you want\r
+ $cols = array();\r
+ $content = array();\r
+\r
+ $tbody_mode = false; // if true, then we need to wrap any stray\r
+ // <tr>s with a <tbody>.\r
+\r
+ $ws_accum =& $initial_ws;\r
+\r
+ foreach ($children as $node) {\r
+ if ($node instanceof HTMLPurifier_Node_Comment) {\r
+ $ws_accum[] = $node;\r
+ continue;\r
+ }\r
+ switch ($node->name) {\r
+ case 'tbody':\r
+ $tbody_mode = true;\r
+ // fall through\r
+ case 'tr':\r
+ $content[] = $node;\r
+ $ws_accum =& $content;\r
+ break;\r
+ case 'caption':\r
+ // there can only be one caption!\r
+ if ($caption !== false) break;\r
+ $caption = $node;\r
+ $ws_accum =& $after_caption_ws;\r
+ break;\r
+ case 'thead':\r
+ $tbody_mode = true;\r
+ // XXX This breaks rendering properties with\r
+ // Firefox, which never floats a <thead> to\r
+ // the top. Ever. (Our scheme will float the\r
+ // first <thead> to the top.) So maybe\r
+ // <thead>s that are not first should be\r
+ // turned into <tbody>? Very tricky, indeed.\r
+ if ($thead === false) {\r
+ $thead = $node;\r
+ $ws_accum =& $after_thead_ws;\r
+ } else {\r
+ // Oops, there's a second one! What\r
+ // should we do? Current behavior is to\r
+ // transmutate the first and last entries into\r
+ // tbody tags, and then put into content.\r
+ // Maybe a better idea is to *attach\r
+ // it* to the existing thead or tfoot?\r
+ // We don't do this, because Firefox\r
+ // doesn't float an extra tfoot to the\r
+ // bottom like it does for the first one.\r
+ $node->name = 'tbody';\r
+ $content[] = $node;\r
+ $ws_accum =& $content;\r
+ }\r
+ break;\r
+ case 'tfoot':\r
+ // see above for some aveats\r
+ $tbody_mode = true;\r
+ if ($tfoot === false) {\r
+ $tfoot = $node;\r
+ $ws_accum =& $after_tfoot_ws;\r
+ } else {\r
+ $node->name = 'tbody';\r
+ $content[] = $node;\r
+ $ws_accum =& $content;\r
+ }\r
+ break;\r
+ case 'colgroup':\r
+ case 'col':\r
+ $cols[] = $node;\r
+ $ws_accum =& $cols;\r
+ break;\r
+ case '#PCDATA':\r
+ // How is whitespace handled? We treat is as sticky to\r
+ // the *end* of the previous element. So all of the\r
+ // nonsense we have worked on is to keep things\r
+ // together.\r
+ if (!empty($node->is_whitespace)) {\r
+ $ws_accum[] = $node;\r
+ }\r
+ break;\r
+ }\r
+ }\r
+\r
+ if (empty($content)) {\r
+ return false;\r
+ }\r
+\r
+ $ret = $initial_ws;\r
+ if ($caption !== false) {\r
+ $ret[] = $caption;\r
+ $ret = array_merge($ret, $after_caption_ws);\r
+ }\r
+ if ($cols !== false) {\r
+ $ret = array_merge($ret, $cols);\r
+ }\r
+ if ($thead !== false) {\r
+ $ret[] = $thead;\r
+ $ret = array_merge($ret, $after_thead_ws);\r
+ }\r
+ if ($tfoot !== false) {\r
+ $ret[] = $tfoot;\r
+ $ret = array_merge($ret, $after_tfoot_ws);\r
+ }\r
+\r
+ if ($tbody_mode) {\r
+ // we have to shuffle tr into tbody\r
+ $current_tr_tbody = null;\r
+\r
+ foreach($content as $node) {\r
+ switch ($node->name) {\r
+ case 'tbody':\r
+ $current_tr_tbody = null;\r
+ $ret[] = $node;\r
+ break;\r
+ case 'tr':\r
+ if ($current_tr_tbody === null) {\r
+ $current_tr_tbody = new HTMLPurifier_Node_Element('tbody');\r
+ $ret[] = $current_tr_tbody;\r
+ }\r
+ $current_tr_tbody->children[] = $node;\r
+ break;\r
+ case '#PCDATA':\r
+ //assert($node->is_whitespace);\r
+ if ($current_tr_tbody === null) {\r
+ $ret[] = $node;\r
+ } else {\r
+ $current_tr_tbody->children[] = $node;\r
+ }\r
+ break;\r
+ }\r
+ }\r
+ } else {\r
+ $ret = array_merge($ret, $content);\r
+ }\r
+\r
+ return $ret;\r
+\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+class HTMLPurifier_DefinitionCache_Decorator extends HTMLPurifier_DefinitionCache\r
+{\r
+\r
+ /**\r
+ * Cache object we are decorating\r
+ * @type HTMLPurifier_DefinitionCache\r
+ */\r
+ public $cache;\r
+\r
+ /**\r
+ * The name of the decorator\r
+ * @var string\r
+ */\r
+ public $name;\r
+\r
+ public function __construct()\r
+ {\r
+ }\r
+\r
+ /**\r
+ * Lazy decorator function\r
+ * @param HTMLPurifier_DefinitionCache $cache Reference to cache object to decorate\r
+ * @return HTMLPurifier_DefinitionCache_Decorator\r
+ */\r
+ public function decorate(&$cache)\r
+ {\r
+ $decorator = $this->copy();\r
+ // reference is necessary for mocks in PHP 4\r
+ $decorator->cache =& $cache;\r
+ $decorator->type = $cache->type;\r
+ return $decorator;\r
+ }\r
+\r
+ /**\r
+ * Cross-compatible clone substitute\r
+ * @return HTMLPurifier_DefinitionCache_Decorator\r
+ */\r
+ public function copy()\r
+ {\r
+ return new HTMLPurifier_DefinitionCache_Decorator();\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Definition $def\r
+ * @param HTMLPurifier_Config $config\r
+ * @return mixed\r
+ */\r
+ public function add($def, $config)\r
+ {\r
+ return $this->cache->add($def, $config);\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Definition $def\r
+ * @param HTMLPurifier_Config $config\r
+ * @return mixed\r
+ */\r
+ public function set($def, $config)\r
+ {\r
+ return $this->cache->set($def, $config);\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Definition $def\r
+ * @param HTMLPurifier_Config $config\r
+ * @return mixed\r
+ */\r
+ public function replace($def, $config)\r
+ {\r
+ return $this->cache->replace($def, $config);\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ * @return mixed\r
+ */\r
+ public function get($config)\r
+ {\r
+ return $this->cache->get($config);\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ * @return mixed\r
+ */\r
+ public function remove($config)\r
+ {\r
+ return $this->cache->remove($config);\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ * @return mixed\r
+ */\r
+ public function flush($config)\r
+ {\r
+ return $this->cache->flush($config);\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ * @return mixed\r
+ */\r
+ public function cleanup($config)\r
+ {\r
+ return $this->cache->cleanup($config);\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Null cache object to use when no caching is on.\r
+ */\r
+class HTMLPurifier_DefinitionCache_Null extends HTMLPurifier_DefinitionCache\r
+{\r
+\r
+ /**\r
+ * @param HTMLPurifier_Definition $def\r
+ * @param HTMLPurifier_Config $config\r
+ * @return bool\r
+ */\r
+ public function add($def, $config)\r
+ {\r
+ return false;\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Definition $def\r
+ * @param HTMLPurifier_Config $config\r
+ * @return bool\r
+ */\r
+ public function set($def, $config)\r
+ {\r
+ return false;\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Definition $def\r
+ * @param HTMLPurifier_Config $config\r
+ * @return bool\r
+ */\r
+ public function replace($def, $config)\r
+ {\r
+ return false;\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ * @return bool\r
+ */\r
+ public function remove($config)\r
+ {\r
+ return false;\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ * @return bool\r
+ */\r
+ public function get($config)\r
+ {\r
+ return false;\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ * @return bool\r
+ */\r
+ public function flush($config)\r
+ {\r
+ return false;\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ * @return bool\r
+ */\r
+ public function cleanup($config)\r
+ {\r
+ return false;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+class HTMLPurifier_DefinitionCache_Serializer extends HTMLPurifier_DefinitionCache\r
+{\r
+\r
+ /**\r
+ * @param HTMLPurifier_Definition $def\r
+ * @param HTMLPurifier_Config $config\r
+ * @return int|bool\r
+ */\r
+ public function add($def, $config)\r
+ {\r
+ if (!$this->checkDefType($def)) {\r
+ return;\r
+ }\r
+ $file = $this->generateFilePath($config);\r
+ if (file_exists($file)) {\r
+ return false;\r
+ }\r
+ if (!$this->_prepareDir($config)) {\r
+ return false;\r
+ }\r
+ return $this->_write($file, serialize($def), $config);\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Definition $def\r
+ * @param HTMLPurifier_Config $config\r
+ * @return int|bool\r
+ */\r
+ public function set($def, $config)\r
+ {\r
+ if (!$this->checkDefType($def)) {\r
+ return;\r
+ }\r
+ $file = $this->generateFilePath($config);\r
+ if (!$this->_prepareDir($config)) {\r
+ return false;\r
+ }\r
+ return $this->_write($file, serialize($def), $config);\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Definition $def\r
+ * @param HTMLPurifier_Config $config\r
+ * @return int|bool\r
+ */\r
+ public function replace($def, $config)\r
+ {\r
+ if (!$this->checkDefType($def)) {\r
+ return;\r
+ }\r
+ $file = $this->generateFilePath($config);\r
+ if (!file_exists($file)) {\r
+ return false;\r
+ }\r
+ if (!$this->_prepareDir($config)) {\r
+ return false;\r
+ }\r
+ return $this->_write($file, serialize($def), $config);\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ * @return bool|HTMLPurifier_Config\r
+ */\r
+ public function get($config)\r
+ {\r
+ $file = $this->generateFilePath($config);\r
+ if (!file_exists($file)) {\r
+ return false;\r
+ }\r
+ return unserialize(file_get_contents($file));\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ * @return bool\r
+ */\r
+ public function remove($config)\r
+ {\r
+ $file = $this->generateFilePath($config);\r
+ if (!file_exists($file)) {\r
+ return false;\r
+ }\r
+ return unlink($file);\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ * @return bool\r
+ */\r
+ public function flush($config)\r
+ {\r
+ if (!$this->_prepareDir($config)) {\r
+ return false;\r
+ }\r
+ $dir = $this->generateDirectoryPath($config);\r
+ $dh = opendir($dir);\r
+ // Apparently, on some versions of PHP, readdir will return\r
+ // an empty string if you pass an invalid argument to readdir.\r
+ // So you need this test. See #49.\r
+ if (false === $dh) {\r
+ return false;\r
+ }\r
+ while (false !== ($filename = readdir($dh))) {\r
+ if (empty($filename)) {\r
+ continue;\r
+ }\r
+ if ($filename[0] === '.') {\r
+ continue;\r
+ }\r
+ unlink($dir . '/' . $filename);\r
+ }\r
+ closedir($dh);\r
+ return true;\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ * @return bool\r
+ */\r
+ public function cleanup($config)\r
+ {\r
+ if (!$this->_prepareDir($config)) {\r
+ return false;\r
+ }\r
+ $dir = $this->generateDirectoryPath($config);\r
+ $dh = opendir($dir);\r
+ // See #49 (and above).\r
+ if (false === $dh) {\r
+ return false;\r
+ }\r
+ while (false !== ($filename = readdir($dh))) {\r
+ if (empty($filename)) {\r
+ continue;\r
+ }\r
+ if ($filename[0] === '.') {\r
+ continue;\r
+ }\r
+ $key = substr($filename, 0, strlen($filename) - 4);\r
+ if ($this->isOld($key, $config)) {\r
+ unlink($dir . '/' . $filename);\r
+ }\r
+ }\r
+ closedir($dh);\r
+ return true;\r
+ }\r
+\r
+ /**\r
+ * Generates the file path to the serial file corresponding to\r
+ * the configuration and definition name\r
+ * @param HTMLPurifier_Config $config\r
+ * @return string\r
+ * @todo Make protected\r
+ */\r
+ public function generateFilePath($config)\r
+ {\r
+ $key = $this->generateKey($config);\r
+ return $this->generateDirectoryPath($config) . '/' . $key . '.ser';\r
+ }\r
+\r
+ /**\r
+ * Generates the path to the directory contain this cache's serial files\r
+ * @param HTMLPurifier_Config $config\r
+ * @return string\r
+ * @note No trailing slash\r
+ * @todo Make protected\r
+ */\r
+ public function generateDirectoryPath($config)\r
+ {\r
+ $base = $this->generateBaseDirectoryPath($config);\r
+ return $base . '/' . $this->type;\r
+ }\r
+\r
+ /**\r
+ * Generates path to base directory that contains all definition type\r
+ * serials\r
+ * @param HTMLPurifier_Config $config\r
+ * @return mixed|string\r
+ * @todo Make protected\r
+ */\r
+ public function generateBaseDirectoryPath($config)\r
+ {\r
+ $base = $config->get('Cache.SerializerPath');\r
+ $base = is_null($base) ? HTMLPURIFIER_PREFIX . '/HTMLPurifier/DefinitionCache/Serializer' : $base;\r
+ return $base;\r
+ }\r
+\r
+ /**\r
+ * Convenience wrapper function for file_put_contents\r
+ * @param string $file File name to write to\r
+ * @param string $data Data to write into file\r
+ * @param HTMLPurifier_Config $config\r
+ * @return int|bool Number of bytes written if success, or false if failure.\r
+ */\r
+ private function _write($file, $data, $config)\r
+ {\r
+ $result = file_put_contents($file, $data);\r
+ if ($result !== false) {\r
+ // set permissions of the new file (no execute)\r
+ $chmod = $config->get('Cache.SerializerPermissions');\r
+ if ($chmod !== null) {\r
+ chmod($file, $chmod & 0666);\r
+ }\r
+ }\r
+ return $result;\r
+ }\r
+\r
+ /**\r
+ * Prepares the directory that this type stores the serials in\r
+ * @param HTMLPurifier_Config $config\r
+ * @return bool True if successful\r
+ */\r
+ private function _prepareDir($config)\r
+ {\r
+ $directory = $this->generateDirectoryPath($config);\r
+ $chmod = $config->get('Cache.SerializerPermissions');\r
+ if ($chmod === null) {\r
+ if (!@mkdir($directory) && !is_dir($directory)) {\r
+ trigger_error(\r
+ 'Could not create directory ' . $directory . '',\r
+ E_USER_WARNING\r
+ );\r
+ return false;\r
+ }\r
+ return true;\r
+ }\r
+ if (!is_dir($directory)) {\r
+ $base = $this->generateBaseDirectoryPath($config);\r
+ if (!is_dir($base)) {\r
+ trigger_error(\r
+ 'Base directory ' . $base . ' does not exist,\r
+ please create or change using %Cache.SerializerPath',\r
+ E_USER_WARNING\r
+ );\r
+ return false;\r
+ } elseif (!$this->_testPermissions($base, $chmod)) {\r
+ return false;\r
+ }\r
+ if (!@mkdir($directory, $chmod) && !is_dir($directory)) {\r
+ trigger_error(\r
+ 'Could not create directory ' . $directory . '',\r
+ E_USER_WARNING\r
+ );\r
+ return false;\r
+ }\r
+ if (!$this->_testPermissions($directory, $chmod)) {\r
+ return false;\r
+ }\r
+ } elseif (!$this->_testPermissions($directory, $chmod)) {\r
+ return false;\r
+ }\r
+ return true;\r
+ }\r
+\r
+ /**\r
+ * Tests permissions on a directory and throws out friendly\r
+ * error messages and attempts to chmod it itself if possible\r
+ * @param string $dir Directory path\r
+ * @param int $chmod Permissions\r
+ * @return bool True if directory is writable\r
+ */\r
+ private function _testPermissions($dir, $chmod)\r
+ {\r
+ // early abort, if it is writable, everything is hunky-dory\r
+ if (is_writable($dir)) {\r
+ return true;\r
+ }\r
+ if (!is_dir($dir)) {\r
+ // generally, you'll want to handle this beforehand\r
+ // so a more specific error message can be given\r
+ trigger_error(\r
+ 'Directory ' . $dir . ' does not exist',\r
+ E_USER_WARNING\r
+ );\r
+ return false;\r
+ }\r
+ if (function_exists('posix_getuid') && $chmod !== null) {\r
+ // POSIX system, we can give more specific advice\r
+ if (fileowner($dir) === posix_getuid()) {\r
+ // we can chmod it ourselves\r
+ $chmod = $chmod | 0700;\r
+ if (chmod($dir, $chmod)) {\r
+ return true;\r
+ }\r
+ } elseif (filegroup($dir) === posix_getgid()) {\r
+ $chmod = $chmod | 0070;\r
+ } else {\r
+ // PHP's probably running as nobody, so we'll\r
+ // need to give global permissions\r
+ $chmod = $chmod | 0777;\r
+ }\r
+ trigger_error(\r
+ 'Directory ' . $dir . ' not writable, ' .\r
+ 'please chmod to ' . decoct($chmod),\r
+ E_USER_WARNING\r
+ );\r
+ } else {\r
+ // generic error message\r
+ trigger_error(\r
+ 'Directory ' . $dir . ' not writable, ' .\r
+ 'please alter file permissions',\r
+ E_USER_WARNING\r
+ );\r
+ }\r
+ return false;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Definition cache decorator class that cleans up the cache\r
+ * whenever there is a cache miss.\r
+ */\r
+class HTMLPurifier_DefinitionCache_Decorator_Cleanup extends HTMLPurifier_DefinitionCache_Decorator\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'Cleanup';\r
+\r
+ /**\r
+ * @return HTMLPurifier_DefinitionCache_Decorator_Cleanup\r
+ */\r
+ public function copy()\r
+ {\r
+ return new HTMLPurifier_DefinitionCache_Decorator_Cleanup();\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Definition $def\r
+ * @param HTMLPurifier_Config $config\r
+ * @return mixed\r
+ */\r
+ public function add($def, $config)\r
+ {\r
+ $status = parent::add($def, $config);\r
+ if (!$status) {\r
+ parent::cleanup($config);\r
+ }\r
+ return $status;\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Definition $def\r
+ * @param HTMLPurifier_Config $config\r
+ * @return mixed\r
+ */\r
+ public function set($def, $config)\r
+ {\r
+ $status = parent::set($def, $config);\r
+ if (!$status) {\r
+ parent::cleanup($config);\r
+ }\r
+ return $status;\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Definition $def\r
+ * @param HTMLPurifier_Config $config\r
+ * @return mixed\r
+ */\r
+ public function replace($def, $config)\r
+ {\r
+ $status = parent::replace($def, $config);\r
+ if (!$status) {\r
+ parent::cleanup($config);\r
+ }\r
+ return $status;\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ * @return mixed\r
+ */\r
+ public function get($config)\r
+ {\r
+ $ret = parent::get($config);\r
+ if (!$ret) {\r
+ parent::cleanup($config);\r
+ }\r
+ return $ret;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Definition cache decorator class that saves all cache retrievals\r
+ * to PHP's memory; good for unit tests or circumstances where\r
+ * there are lots of configuration objects floating around.\r
+ */\r
+class HTMLPurifier_DefinitionCache_Decorator_Memory extends HTMLPurifier_DefinitionCache_Decorator\r
+{\r
+ /**\r
+ * @type array\r
+ */\r
+ protected $definitions;\r
+\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'Memory';\r
+\r
+ /**\r
+ * @return HTMLPurifier_DefinitionCache_Decorator_Memory\r
+ */\r
+ public function copy()\r
+ {\r
+ return new HTMLPurifier_DefinitionCache_Decorator_Memory();\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Definition $def\r
+ * @param HTMLPurifier_Config $config\r
+ * @return mixed\r
+ */\r
+ public function add($def, $config)\r
+ {\r
+ $status = parent::add($def, $config);\r
+ if ($status) {\r
+ $this->definitions[$this->generateKey($config)] = $def;\r
+ }\r
+ return $status;\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Definition $def\r
+ * @param HTMLPurifier_Config $config\r
+ * @return mixed\r
+ */\r
+ public function set($def, $config)\r
+ {\r
+ $status = parent::set($def, $config);\r
+ if ($status) {\r
+ $this->definitions[$this->generateKey($config)] = $def;\r
+ }\r
+ return $status;\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Definition $def\r
+ * @param HTMLPurifier_Config $config\r
+ * @return mixed\r
+ */\r
+ public function replace($def, $config)\r
+ {\r
+ $status = parent::replace($def, $config);\r
+ if ($status) {\r
+ $this->definitions[$this->generateKey($config)] = $def;\r
+ }\r
+ return $status;\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ * @return mixed\r
+ */\r
+ public function get($config)\r
+ {\r
+ $key = $this->generateKey($config);\r
+ if (isset($this->definitions[$key])) {\r
+ return $this->definitions[$key];\r
+ }\r
+ $this->definitions[$key] = parent::get($config);\r
+ return $this->definitions[$key];\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * XHTML 1.1 Bi-directional Text Module, defines elements that\r
+ * declare directionality of content. Text Extension Module.\r
+ */\r
+class HTMLPurifier_HTMLModule_Bdo extends HTMLPurifier_HTMLModule\r
+{\r
+\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'Bdo';\r
+\r
+ /**\r
+ * @type array\r
+ */\r
+ public $attr_collections = array(\r
+ 'I18N' => array('dir' => false)\r
+ );\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ public function setup($config)\r
+ {\r
+ $bdo = $this->addElement(\r
+ 'bdo',\r
+ 'Inline',\r
+ 'Inline',\r
+ array('Core', 'Lang'),\r
+ array(\r
+ 'dir' => 'Enum#ltr,rtl', // required\r
+ // The Abstract Module specification has the attribute\r
+ // inclusions wrong for bdo: bdo allows Lang\r
+ )\r
+ );\r
+ $bdo->attr_transform_post[] = new HTMLPurifier_AttrTransform_BdoDir();\r
+\r
+ $this->attr_collections['I18N']['dir'] = 'Enum#ltr,rtl';\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+class HTMLPurifier_HTMLModule_CommonAttributes extends HTMLPurifier_HTMLModule\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'CommonAttributes';\r
+\r
+ /**\r
+ * @type array\r
+ */\r
+ public $attr_collections = array(\r
+ 'Core' => array(\r
+ 0 => array('Style'),\r
+ // 'xml:space' => false,\r
+ 'class' => 'Class',\r
+ 'id' => 'ID',\r
+ 'title' => 'CDATA',\r
+ ),\r
+ 'Lang' => array(),\r
+ 'I18N' => array(\r
+ 0 => array('Lang'), // proprietary, for xml:lang/lang\r
+ ),\r
+ 'Common' => array(\r
+ 0 => array('Core', 'I18N')\r
+ )\r
+ );\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * XHTML 1.1 Edit Module, defines editing-related elements. Text Extension\r
+ * Module.\r
+ */\r
+class HTMLPurifier_HTMLModule_Edit extends HTMLPurifier_HTMLModule\r
+{\r
+\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'Edit';\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ public function setup($config)\r
+ {\r
+ $contents = 'Chameleon: #PCDATA | Inline ! #PCDATA | Flow';\r
+ $attr = array(\r
+ 'cite' => 'URI',\r
+ // 'datetime' => 'Datetime', // not implemented\r
+ );\r
+ $this->addElement('del', 'Inline', $contents, 'Common', $attr);\r
+ $this->addElement('ins', 'Inline', $contents, 'Common', $attr);\r
+ }\r
+\r
+ // HTML 4.01 specifies that ins/del must not contain block\r
+ // elements when used in an inline context, chameleon is\r
+ // a complicated workaround to acheive this effect\r
+\r
+ // Inline context ! Block context (exclamation mark is\r
+ // separator, see getChildDef for parsing)\r
+\r
+ /**\r
+ * @type bool\r
+ */\r
+ public $defines_child_def = true;\r
+\r
+ /**\r
+ * @param HTMLPurifier_ElementDef $def\r
+ * @return HTMLPurifier_ChildDef_Chameleon\r
+ */\r
+ public function getChildDef($def)\r
+ {\r
+ if ($def->content_model_type != 'chameleon') {\r
+ return false;\r
+ }\r
+ $value = explode('!', $def->content_model);\r
+ return new HTMLPurifier_ChildDef_Chameleon($value[0], $value[1]);\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * XHTML 1.1 Forms module, defines all form-related elements found in HTML 4.\r
+ */\r
+class HTMLPurifier_HTMLModule_Forms extends HTMLPurifier_HTMLModule\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'Forms';\r
+\r
+ /**\r
+ * @type bool\r
+ */\r
+ public $safe = false;\r
+\r
+ /**\r
+ * @type array\r
+ */\r
+ public $content_sets = array(\r
+ 'Block' => 'Form',\r
+ 'Inline' => 'Formctrl',\r
+ );\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ public function setup($config)\r
+ {\r
+ $form = $this->addElement(\r
+ 'form',\r
+ 'Form',\r
+ 'Required: Heading | List | Block | fieldset',\r
+ 'Common',\r
+ array(\r
+ 'accept' => 'ContentTypes',\r
+ 'accept-charset' => 'Charsets',\r
+ 'action*' => 'URI',\r
+ 'method' => 'Enum#get,post',\r
+ // really ContentType, but these two are the only ones used today\r
+ 'enctype' => 'Enum#application/x-www-form-urlencoded,multipart/form-data',\r
+ )\r
+ );\r
+ $form->excludes = array('form' => true);\r
+\r
+ $input = $this->addElement(\r
+ 'input',\r
+ 'Formctrl',\r
+ 'Empty',\r
+ 'Common',\r
+ array(\r
+ 'accept' => 'ContentTypes',\r
+ 'accesskey' => 'Character',\r
+ 'alt' => 'Text',\r
+ 'checked' => 'Bool#checked',\r
+ 'disabled' => 'Bool#disabled',\r
+ 'maxlength' => 'Number',\r
+ 'name' => 'CDATA',\r
+ 'readonly' => 'Bool#readonly',\r
+ 'size' => 'Number',\r
+ 'src' => 'URI#embedded',\r
+ 'tabindex' => 'Number',\r
+ 'type' => 'Enum#text,password,checkbox,button,radio,submit,reset,file,hidden,image',\r
+ 'value' => 'CDATA',\r
+ )\r
+ );\r
+ $input->attr_transform_post[] = new HTMLPurifier_AttrTransform_Input();\r
+\r
+ $this->addElement(\r
+ 'select',\r
+ 'Formctrl',\r
+ 'Required: optgroup | option',\r
+ 'Common',\r
+ array(\r
+ 'disabled' => 'Bool#disabled',\r
+ 'multiple' => 'Bool#multiple',\r
+ 'name' => 'CDATA',\r
+ 'size' => 'Number',\r
+ 'tabindex' => 'Number',\r
+ )\r
+ );\r
+\r
+ $this->addElement(\r
+ 'option',\r
+ false,\r
+ 'Optional: #PCDATA',\r
+ 'Common',\r
+ array(\r
+ 'disabled' => 'Bool#disabled',\r
+ 'label' => 'Text',\r
+ 'selected' => 'Bool#selected',\r
+ 'value' => 'CDATA',\r
+ )\r
+ );\r
+ // It's illegal for there to be more than one selected, but not\r
+ // be multiple. Also, no selected means undefined behavior. This might\r
+ // be difficult to implement; perhaps an injector, or a context variable.\r
+\r
+ $textarea = $this->addElement(\r
+ 'textarea',\r
+ 'Formctrl',\r
+ 'Optional: #PCDATA',\r
+ 'Common',\r
+ array(\r
+ 'accesskey' => 'Character',\r
+ 'cols*' => 'Number',\r
+ 'disabled' => 'Bool#disabled',\r
+ 'name' => 'CDATA',\r
+ 'readonly' => 'Bool#readonly',\r
+ 'rows*' => 'Number',\r
+ 'tabindex' => 'Number',\r
+ )\r
+ );\r
+ $textarea->attr_transform_pre[] = new HTMLPurifier_AttrTransform_Textarea();\r
+\r
+ $button = $this->addElement(\r
+ 'button',\r
+ 'Formctrl',\r
+ 'Optional: #PCDATA | Heading | List | Block | Inline',\r
+ 'Common',\r
+ array(\r
+ 'accesskey' => 'Character',\r
+ 'disabled' => 'Bool#disabled',\r
+ 'name' => 'CDATA',\r
+ 'tabindex' => 'Number',\r
+ 'type' => 'Enum#button,submit,reset',\r
+ 'value' => 'CDATA',\r
+ )\r
+ );\r
+\r
+ // For exclusions, ideally we'd specify content sets, not literal elements\r
+ $button->excludes = $this->makeLookup(\r
+ 'form',\r
+ 'fieldset', // Form\r
+ 'input',\r
+ 'select',\r
+ 'textarea',\r
+ 'label',\r
+ 'button', // Formctrl\r
+ 'a', // as per HTML 4.01 spec, this is omitted by modularization\r
+ 'isindex',\r
+ 'iframe' // legacy items\r
+ );\r
+\r
+ // Extra exclusion: img usemap="" is not permitted within this element.\r
+ // We'll omit this for now, since we don't have any good way of\r
+ // indicating it yet.\r
+\r
+ // This is HIGHLY user-unfriendly; we need a custom child-def for this\r
+ $this->addElement('fieldset', 'Form', 'Custom: (#WS?,legend,(Flow|#PCDATA)*)', 'Common');\r
+\r
+ $label = $this->addElement(\r
+ 'label',\r
+ 'Formctrl',\r
+ 'Optional: #PCDATA | Inline',\r
+ 'Common',\r
+ array(\r
+ 'accesskey' => 'Character',\r
+ // 'for' => 'IDREF', // IDREF not implemented, cannot allow\r
+ )\r
+ );\r
+ $label->excludes = array('label' => true);\r
+\r
+ $this->addElement(\r
+ 'legend',\r
+ false,\r
+ 'Optional: #PCDATA | Inline',\r
+ 'Common',\r
+ array(\r
+ 'accesskey' => 'Character',\r
+ )\r
+ );\r
+\r
+ $this->addElement(\r
+ 'optgroup',\r
+ false,\r
+ 'Required: option',\r
+ 'Common',\r
+ array(\r
+ 'disabled' => 'Bool#disabled',\r
+ 'label*' => 'Text',\r
+ )\r
+ );\r
+ // Don't forget an injector for <isindex>. This one's a little complex\r
+ // because it maps to multiple elements.\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * XHTML 1.1 Hypertext Module, defines hypertext links. Core Module.\r
+ */\r
+class HTMLPurifier_HTMLModule_Hypertext extends HTMLPurifier_HTMLModule\r
+{\r
+\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'Hypertext';\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ public function setup($config)\r
+ {\r
+ $a = $this->addElement(\r
+ 'a',\r
+ 'Inline',\r
+ 'Inline',\r
+ 'Common',\r
+ array(\r
+ // 'accesskey' => 'Character',\r
+ // 'charset' => 'Charset',\r
+ 'href' => 'URI',\r
+ // 'hreflang' => 'LanguageCode',\r
+ 'rel' => new HTMLPurifier_AttrDef_HTML_LinkTypes('rel'),\r
+ 'rev' => new HTMLPurifier_AttrDef_HTML_LinkTypes('rev'),\r
+ // 'tabindex' => 'Number',\r
+ // 'type' => 'ContentType',\r
+ )\r
+ );\r
+ $a->formatting = true;\r
+ $a->excludes = array('a' => true);\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * XHTML 1.1 Iframe Module provides inline frames.\r
+ *\r
+ * @note This module is not considered safe unless an Iframe\r
+ * whitelisting mechanism is specified. Currently, the only\r
+ * such mechanism is %URL.SafeIframeRegexp\r
+ */\r
+class HTMLPurifier_HTMLModule_Iframe extends HTMLPurifier_HTMLModule\r
+{\r
+\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'Iframe';\r
+\r
+ /**\r
+ * @type bool\r
+ */\r
+ public $safe = false;\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ public function setup($config)\r
+ {\r
+ if ($config->get('HTML.SafeIframe')) {\r
+ $this->safe = true;\r
+ }\r
+ $this->addElement(\r
+ 'iframe',\r
+ 'Inline',\r
+ 'Flow',\r
+ 'Common',\r
+ array(\r
+ 'src' => 'URI#embedded',\r
+ 'width' => 'Length',\r
+ 'height' => 'Length',\r
+ 'name' => 'ID',\r
+ 'scrolling' => 'Enum#yes,no,auto',\r
+ 'frameborder' => 'Enum#0,1',\r
+ 'longdesc' => 'URI',\r
+ 'marginheight' => 'Pixels',\r
+ 'marginwidth' => 'Pixels',\r
+ )\r
+ );\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * XHTML 1.1 Image Module provides basic image embedding.\r
+ * @note There is specialized code for removing empty images in\r
+ * HTMLPurifier_Strategy_RemoveForeignElements\r
+ */\r
+class HTMLPurifier_HTMLModule_Image extends HTMLPurifier_HTMLModule\r
+{\r
+\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'Image';\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ public function setup($config)\r
+ {\r
+ $max = $config->get('HTML.MaxImgLength');\r
+ $img = $this->addElement(\r
+ 'img',\r
+ 'Inline',\r
+ 'Empty',\r
+ 'Common',\r
+ array(\r
+ 'alt*' => 'Text',\r
+ // According to the spec, it's Length, but percents can\r
+ // be abused, so we allow only Pixels.\r
+ 'height' => 'Pixels#' . $max,\r
+ 'width' => 'Pixels#' . $max,\r
+ 'longdesc' => 'URI',\r
+ 'src*' => new HTMLPurifier_AttrDef_URI(true), // embedded\r
+ )\r
+ );\r
+ if ($max === null || $config->get('HTML.Trusted')) {\r
+ $img->attr['height'] =\r
+ $img->attr['width'] = 'Length';\r
+ }\r
+\r
+ // kind of strange, but splitting things up would be inefficient\r
+ $img->attr_transform_pre[] =\r
+ $img->attr_transform_post[] =\r
+ new HTMLPurifier_AttrTransform_ImgRequired();\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * XHTML 1.1 Legacy module defines elements that were previously\r
+ * deprecated.\r
+ *\r
+ * @note Not all legacy elements have been implemented yet, which\r
+ * is a bit of a reverse problem as compared to browsers! In\r
+ * addition, this legacy module may implement a bit more than\r
+ * mandated by XHTML 1.1.\r
+ *\r
+ * This module can be used in combination with TransformToStrict in order\r
+ * to transform as many deprecated elements as possible, but retain\r
+ * questionably deprecated elements that do not have good alternatives\r
+ * as well as transform elements that don't have an implementation.\r
+ * See docs/ref-strictness.txt for more details.\r
+ */\r
+\r
+class HTMLPurifier_HTMLModule_Legacy extends HTMLPurifier_HTMLModule\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'Legacy';\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ public function setup($config)\r
+ {\r
+ $this->addElement(\r
+ 'basefont',\r
+ 'Inline',\r
+ 'Empty',\r
+ null,\r
+ array(\r
+ 'color' => 'Color',\r
+ 'face' => 'Text', // extremely broad, we should\r
+ 'size' => 'Text', // tighten it\r
+ 'id' => 'ID'\r
+ )\r
+ );\r
+ $this->addElement('center', 'Block', 'Flow', 'Common');\r
+ $this->addElement(\r
+ 'dir',\r
+ 'Block',\r
+ 'Required: li',\r
+ 'Common',\r
+ array(\r
+ 'compact' => 'Bool#compact'\r
+ )\r
+ );\r
+ $this->addElement(\r
+ 'font',\r
+ 'Inline',\r
+ 'Inline',\r
+ array('Core', 'I18N'),\r
+ array(\r
+ 'color' => 'Color',\r
+ 'face' => 'Text', // extremely broad, we should\r
+ 'size' => 'Text', // tighten it\r
+ )\r
+ );\r
+ $this->addElement(\r
+ 'menu',\r
+ 'Block',\r
+ 'Required: li',\r
+ 'Common',\r
+ array(\r
+ 'compact' => 'Bool#compact'\r
+ )\r
+ );\r
+\r
+ $s = $this->addElement('s', 'Inline', 'Inline', 'Common');\r
+ $s->formatting = true;\r
+\r
+ $strike = $this->addElement('strike', 'Inline', 'Inline', 'Common');\r
+ $strike->formatting = true;\r
+\r
+ $u = $this->addElement('u', 'Inline', 'Inline', 'Common');\r
+ $u->formatting = true;\r
+\r
+ // setup modifications to old elements\r
+\r
+ $align = 'Enum#left,right,center,justify';\r
+\r
+ $address = $this->addBlankElement('address');\r
+ $address->content_model = 'Inline | #PCDATA | p';\r
+ $address->content_model_type = 'optional';\r
+ $address->child = false;\r
+\r
+ $blockquote = $this->addBlankElement('blockquote');\r
+ $blockquote->content_model = 'Flow | #PCDATA';\r
+ $blockquote->content_model_type = 'optional';\r
+ $blockquote->child = false;\r
+\r
+ $br = $this->addBlankElement('br');\r
+ $br->attr['clear'] = 'Enum#left,all,right,none';\r
+\r
+ $caption = $this->addBlankElement('caption');\r
+ $caption->attr['align'] = 'Enum#top,bottom,left,right';\r
+\r
+ $div = $this->addBlankElement('div');\r
+ $div->attr['align'] = $align;\r
+\r
+ $dl = $this->addBlankElement('dl');\r
+ $dl->attr['compact'] = 'Bool#compact';\r
+\r
+ for ($i = 1; $i <= 6; $i++) {\r
+ $h = $this->addBlankElement("h$i");\r
+ $h->attr['align'] = $align;\r
+ }\r
+\r
+ $hr = $this->addBlankElement('hr');\r
+ $hr->attr['align'] = $align;\r
+ $hr->attr['noshade'] = 'Bool#noshade';\r
+ $hr->attr['size'] = 'Pixels';\r
+ $hr->attr['width'] = 'Length';\r
+\r
+ $img = $this->addBlankElement('img');\r
+ $img->attr['align'] = 'IAlign';\r
+ $img->attr['border'] = 'Pixels';\r
+ $img->attr['hspace'] = 'Pixels';\r
+ $img->attr['vspace'] = 'Pixels';\r
+\r
+ // figure out this integer business\r
+\r
+ $li = $this->addBlankElement('li');\r
+ $li->attr['value'] = new HTMLPurifier_AttrDef_Integer();\r
+ $li->attr['type'] = 'Enum#s:1,i,I,a,A,disc,square,circle';\r
+\r
+ $ol = $this->addBlankElement('ol');\r
+ $ol->attr['compact'] = 'Bool#compact';\r
+ $ol->attr['start'] = new HTMLPurifier_AttrDef_Integer();\r
+ $ol->attr['type'] = 'Enum#s:1,i,I,a,A';\r
+\r
+ $p = $this->addBlankElement('p');\r
+ $p->attr['align'] = $align;\r
+\r
+ $pre = $this->addBlankElement('pre');\r
+ $pre->attr['width'] = 'Number';\r
+\r
+ // script omitted\r
+\r
+ $table = $this->addBlankElement('table');\r
+ $table->attr['align'] = 'Enum#left,center,right';\r
+ $table->attr['bgcolor'] = 'Color';\r
+\r
+ $tr = $this->addBlankElement('tr');\r
+ $tr->attr['bgcolor'] = 'Color';\r
+\r
+ $th = $this->addBlankElement('th');\r
+ $th->attr['bgcolor'] = 'Color';\r
+ $th->attr['height'] = 'Length';\r
+ $th->attr['nowrap'] = 'Bool#nowrap';\r
+ $th->attr['width'] = 'Length';\r
+\r
+ $td = $this->addBlankElement('td');\r
+ $td->attr['bgcolor'] = 'Color';\r
+ $td->attr['height'] = 'Length';\r
+ $td->attr['nowrap'] = 'Bool#nowrap';\r
+ $td->attr['width'] = 'Length';\r
+\r
+ $ul = $this->addBlankElement('ul');\r
+ $ul->attr['compact'] = 'Bool#compact';\r
+ $ul->attr['type'] = 'Enum#square,disc,circle';\r
+\r
+ // "safe" modifications to "unsafe" elements\r
+ // WARNING: If you want to add support for an unsafe, legacy\r
+ // attribute, make a new TrustedLegacy module with the trusted\r
+ // bit set appropriately\r
+\r
+ $form = $this->addBlankElement('form');\r
+ $form->content_model = 'Flow | #PCDATA';\r
+ $form->content_model_type = 'optional';\r
+ $form->attr['target'] = 'FrameTarget';\r
+\r
+ $input = $this->addBlankElement('input');\r
+ $input->attr['align'] = 'IAlign';\r
+\r
+ $legend = $this->addBlankElement('legend');\r
+ $legend->attr['align'] = 'LAlign';\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * XHTML 1.1 List Module, defines list-oriented elements. Core Module.\r
+ */\r
+class HTMLPurifier_HTMLModule_List extends HTMLPurifier_HTMLModule\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'List';\r
+\r
+ // According to the abstract schema, the List content set is a fully formed\r
+ // one or more expr, but it invariably occurs in an optional declaration\r
+ // so we're not going to do that subtlety. It might cause trouble\r
+ // if a user defines "List" and expects that multiple lists are\r
+ // allowed to be specified, but then again, that's not very intuitive.\r
+ // Furthermore, the actual XML Schema may disagree. Regardless,\r
+ // we don't have support for such nested expressions without using\r
+ // the incredibly inefficient and draconic Custom ChildDef.\r
+\r
+ /**\r
+ * @type array\r
+ */\r
+ public $content_sets = array('Flow' => 'List');\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ public function setup($config)\r
+ {\r
+ $ol = $this->addElement('ol', 'List', new HTMLPurifier_ChildDef_List(), 'Common');\r
+ $ul = $this->addElement('ul', 'List', new HTMLPurifier_ChildDef_List(), 'Common');\r
+ // XXX The wrap attribute is handled by MakeWellFormed. This is all\r
+ // quite unsatisfactory, because we generated this\r
+ // *specifically* for lists, and now a big chunk of the handling\r
+ // is done properly by the List ChildDef. So actually, we just\r
+ // want enough information to make autoclosing work properly,\r
+ // and then hand off the tricky stuff to the ChildDef.\r
+ $ol->wrap = 'li';\r
+ $ul->wrap = 'li';\r
+ $this->addElement('dl', 'List', 'Required: dt | dd', 'Common');\r
+\r
+ $this->addElement('li', false, 'Flow', 'Common');\r
+\r
+ $this->addElement('dd', false, 'Flow', 'Common');\r
+ $this->addElement('dt', false, 'Inline', 'Common');\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+class HTMLPurifier_HTMLModule_Name extends HTMLPurifier_HTMLModule\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'Name';\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ public function setup($config)\r
+ {\r
+ $elements = array('a', 'applet', 'form', 'frame', 'iframe', 'img', 'map');\r
+ foreach ($elements as $name) {\r
+ $element = $this->addBlankElement($name);\r
+ $element->attr['name'] = 'CDATA';\r
+ if (!$config->get('HTML.Attr.Name.UseCDATA')) {\r
+ $element->attr_transform_post[] = new HTMLPurifier_AttrTransform_NameSync();\r
+ }\r
+ }\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Module adds the nofollow attribute transformation to a tags. It\r
+ * is enabled by HTML.Nofollow\r
+ */\r
+class HTMLPurifier_HTMLModule_Nofollow extends HTMLPurifier_HTMLModule\r
+{\r
+\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'Nofollow';\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ public function setup($config)\r
+ {\r
+ $a = $this->addBlankElement('a');\r
+ $a->attr_transform_post[] = new HTMLPurifier_AttrTransform_Nofollow();\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+class HTMLPurifier_HTMLModule_NonXMLCommonAttributes extends HTMLPurifier_HTMLModule\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'NonXMLCommonAttributes';\r
+\r
+ /**\r
+ * @type array\r
+ */\r
+ public $attr_collections = array(\r
+ 'Lang' => array(\r
+ 'lang' => 'LanguageCode',\r
+ )\r
+ );\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * XHTML 1.1 Object Module, defines elements for generic object inclusion\r
+ * @warning Users will commonly use <embed> to cater to legacy browsers: this\r
+ * module does not allow this sort of behavior\r
+ */\r
+class HTMLPurifier_HTMLModule_Object extends HTMLPurifier_HTMLModule\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'Object';\r
+\r
+ /**\r
+ * @type bool\r
+ */\r
+ public $safe = false;\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ public function setup($config)\r
+ {\r
+ $this->addElement(\r
+ 'object',\r
+ 'Inline',\r
+ 'Optional: #PCDATA | Flow | param',\r
+ 'Common',\r
+ array(\r
+ 'archive' => 'URI',\r
+ 'classid' => 'URI',\r
+ 'codebase' => 'URI',\r
+ 'codetype' => 'Text',\r
+ 'data' => 'URI',\r
+ 'declare' => 'Bool#declare',\r
+ 'height' => 'Length',\r
+ 'name' => 'CDATA',\r
+ 'standby' => 'Text',\r
+ 'tabindex' => 'Number',\r
+ 'type' => 'ContentType',\r
+ 'width' => 'Length'\r
+ )\r
+ );\r
+\r
+ $this->addElement(\r
+ 'param',\r
+ false,\r
+ 'Empty',\r
+ null,\r
+ array(\r
+ 'id' => 'ID',\r
+ 'name*' => 'Text',\r
+ 'type' => 'Text',\r
+ 'value' => 'Text',\r
+ 'valuetype' => 'Enum#data,ref,object'\r
+ )\r
+ );\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * XHTML 1.1 Presentation Module, defines simple presentation-related\r
+ * markup. Text Extension Module.\r
+ * @note The official XML Schema and DTD specs further divide this into\r
+ * two modules:\r
+ * - Block Presentation (hr)\r
+ * - Inline Presentation (b, big, i, small, sub, sup, tt)\r
+ * We have chosen not to heed this distinction, as content_sets\r
+ * provides satisfactory disambiguation.\r
+ */\r
+class HTMLPurifier_HTMLModule_Presentation extends HTMLPurifier_HTMLModule\r
+{\r
+\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'Presentation';\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ public function setup($config)\r
+ {\r
+ $this->addElement('hr', 'Block', 'Empty', 'Common');\r
+ $this->addElement('sub', 'Inline', 'Inline', 'Common');\r
+ $this->addElement('sup', 'Inline', 'Inline', 'Common');\r
+ $b = $this->addElement('b', 'Inline', 'Inline', 'Common');\r
+ $b->formatting = true;\r
+ $big = $this->addElement('big', 'Inline', 'Inline', 'Common');\r
+ $big->formatting = true;\r
+ $i = $this->addElement('i', 'Inline', 'Inline', 'Common');\r
+ $i->formatting = true;\r
+ $small = $this->addElement('small', 'Inline', 'Inline', 'Common');\r
+ $small->formatting = true;\r
+ $tt = $this->addElement('tt', 'Inline', 'Inline', 'Common');\r
+ $tt->formatting = true;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Module defines proprietary tags and attributes in HTML.\r
+ * @warning If this module is enabled, standards-compliance is off!\r
+ */\r
+class HTMLPurifier_HTMLModule_Proprietary extends HTMLPurifier_HTMLModule\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'Proprietary';\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ public function setup($config)\r
+ {\r
+ $this->addElement(\r
+ 'marquee',\r
+ 'Inline',\r
+ 'Flow',\r
+ 'Common',\r
+ array(\r
+ 'direction' => 'Enum#left,right,up,down',\r
+ 'behavior' => 'Enum#alternate',\r
+ 'width' => 'Length',\r
+ 'height' => 'Length',\r
+ 'scrolldelay' => 'Number',\r
+ 'scrollamount' => 'Number',\r
+ 'loop' => 'Number',\r
+ 'bgcolor' => 'Color',\r
+ 'hspace' => 'Pixels',\r
+ 'vspace' => 'Pixels',\r
+ )\r
+ );\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * XHTML 1.1 Ruby Annotation Module, defines elements that indicate\r
+ * short runs of text alongside base text for annotation or pronounciation.\r
+ */\r
+class HTMLPurifier_HTMLModule_Ruby extends HTMLPurifier_HTMLModule\r
+{\r
+\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'Ruby';\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ public function setup($config)\r
+ {\r
+ $this->addElement(\r
+ 'ruby',\r
+ 'Inline',\r
+ 'Custom: ((rb, (rt | (rp, rt, rp))) | (rbc, rtc, rtc?))',\r
+ 'Common'\r
+ );\r
+ $this->addElement('rbc', false, 'Required: rb', 'Common');\r
+ $this->addElement('rtc', false, 'Required: rt', 'Common');\r
+ $rb = $this->addElement('rb', false, 'Inline', 'Common');\r
+ $rb->excludes = array('ruby' => true);\r
+ $rt = $this->addElement('rt', false, 'Inline', 'Common', array('rbspan' => 'Number'));\r
+ $rt->excludes = array('ruby' => true);\r
+ $this->addElement('rp', false, 'Optional: #PCDATA', 'Common');\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * A "safe" embed module. See SafeObject. This is a proprietary element.\r
+ */\r
+class HTMLPurifier_HTMLModule_SafeEmbed extends HTMLPurifier_HTMLModule\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'SafeEmbed';\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ public function setup($config)\r
+ {\r
+ $max = $config->get('HTML.MaxImgLength');\r
+ $embed = $this->addElement(\r
+ 'embed',\r
+ 'Inline',\r
+ 'Empty',\r
+ 'Common',\r
+ array(\r
+ 'src*' => 'URI#embedded',\r
+ 'type' => 'Enum#application/x-shockwave-flash',\r
+ 'width' => 'Pixels#' . $max,\r
+ 'height' => 'Pixels#' . $max,\r
+ 'allowscriptaccess' => 'Enum#never',\r
+ 'allownetworking' => 'Enum#internal',\r
+ 'flashvars' => 'Text',\r
+ 'wmode' => 'Enum#window,transparent,opaque',\r
+ 'name' => 'ID',\r
+ )\r
+ );\r
+ $embed->attr_transform_post[] = new HTMLPurifier_AttrTransform_SafeEmbed();\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * A "safe" object module. In theory, objects permitted by this module will\r
+ * be safe, and untrusted users can be allowed to embed arbitrary flash objects\r
+ * (maybe other types too, but only Flash is supported as of right now).\r
+ * Highly experimental.\r
+ */\r
+class HTMLPurifier_HTMLModule_SafeObject extends HTMLPurifier_HTMLModule\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'SafeObject';\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ public function setup($config)\r
+ {\r
+ // These definitions are not intrinsically safe: the attribute transforms\r
+ // are a vital part of ensuring safety.\r
+\r
+ $max = $config->get('HTML.MaxImgLength');\r
+ $object = $this->addElement(\r
+ 'object',\r
+ 'Inline',\r
+ 'Optional: param | Flow | #PCDATA',\r
+ 'Common',\r
+ array(\r
+ // While technically not required by the spec, we're forcing\r
+ // it to this value.\r
+ 'type' => 'Enum#application/x-shockwave-flash',\r
+ 'width' => 'Pixels#' . $max,\r
+ 'height' => 'Pixels#' . $max,\r
+ 'data' => 'URI#embedded',\r
+ 'codebase' => new HTMLPurifier_AttrDef_Enum(\r
+ array(\r
+ 'http://download.macromedia.com/pub/shockwave/cabs/flash/swflash.cab#version=6,0,40,0'\r
+ )\r
+ ),\r
+ )\r
+ );\r
+ $object->attr_transform_post[] = new HTMLPurifier_AttrTransform_SafeObject();\r
+\r
+ $param = $this->addElement(\r
+ 'param',\r
+ false,\r
+ 'Empty',\r
+ false,\r
+ array(\r
+ 'id' => 'ID',\r
+ 'name*' => 'Text',\r
+ 'value' => 'Text'\r
+ )\r
+ );\r
+ $param->attr_transform_post[] = new HTMLPurifier_AttrTransform_SafeParam();\r
+ $this->info_injector[] = 'SafeObject';\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * A "safe" script module. No inline JS is allowed, and pointed to JS\r
+ * files must match whitelist.\r
+ */\r
+class HTMLPurifier_HTMLModule_SafeScripting extends HTMLPurifier_HTMLModule\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'SafeScripting';\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ public function setup($config)\r
+ {\r
+ // These definitions are not intrinsically safe: the attribute transforms\r
+ // are a vital part of ensuring safety.\r
+\r
+ $allowed = $config->get('HTML.SafeScripting');\r
+ $script = $this->addElement(\r
+ 'script',\r
+ 'Inline',\r
+ 'Optional:', // Not `Empty` to not allow to autoclose the <script /> tag @see https://www.w3.org/TR/html4/interact/scripts.html\r
+ null,\r
+ array(\r
+ // While technically not required by the spec, we're forcing\r
+ // it to this value.\r
+ 'type' => 'Enum#text/javascript',\r
+ 'src*' => new HTMLPurifier_AttrDef_Enum(array_keys($allowed), /*case sensitive*/ true)\r
+ )\r
+ );\r
+ $script->attr_transform_pre[] =\r
+ $script->attr_transform_post[] = new HTMLPurifier_AttrTransform_ScriptRequired();\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/*\r
+\r
+WARNING: THIS MODULE IS EXTREMELY DANGEROUS AS IT ENABLES INLINE SCRIPTING\r
+INSIDE HTML PURIFIER DOCUMENTS. USE ONLY WITH TRUSTED USER INPUT!!!\r
+\r
+*/\r
+\r
+/**\r
+ * XHTML 1.1 Scripting module, defines elements that are used to contain\r
+ * information pertaining to executable scripts or the lack of support\r
+ * for executable scripts.\r
+ * @note This module does not contain inline scripting elements\r
+ */\r
+class HTMLPurifier_HTMLModule_Scripting extends HTMLPurifier_HTMLModule\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'Scripting';\r
+\r
+ /**\r
+ * @type array\r
+ */\r
+ public $elements = array('script', 'noscript');\r
+\r
+ /**\r
+ * @type array\r
+ */\r
+ public $content_sets = array('Block' => 'script | noscript', 'Inline' => 'script | noscript');\r
+\r
+ /**\r
+ * @type bool\r
+ */\r
+ public $safe = false;\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ public function setup($config)\r
+ {\r
+ // TODO: create custom child-definition for noscript that\r
+ // auto-wraps stray #PCDATA in a similar manner to\r
+ // blockquote's custom definition (we would use it but\r
+ // blockquote's contents are optional while noscript's contents\r
+ // are required)\r
+\r
+ // TODO: convert this to new syntax, main problem is getting\r
+ // both content sets working\r
+\r
+ // In theory, this could be safe, but I don't see any reason to\r
+ // allow it.\r
+ $this->info['noscript'] = new HTMLPurifier_ElementDef();\r
+ $this->info['noscript']->attr = array(0 => array('Common'));\r
+ $this->info['noscript']->content_model = 'Heading | List | Block';\r
+ $this->info['noscript']->content_model_type = 'required';\r
+\r
+ $this->info['script'] = new HTMLPurifier_ElementDef();\r
+ $this->info['script']->attr = array(\r
+ 'defer' => new HTMLPurifier_AttrDef_Enum(array('defer')),\r
+ 'src' => new HTMLPurifier_AttrDef_URI(true),\r
+ 'type' => new HTMLPurifier_AttrDef_Enum(array('text/javascript'))\r
+ );\r
+ $this->info['script']->content_model = '#PCDATA';\r
+ $this->info['script']->content_model_type = 'optional';\r
+ $this->info['script']->attr_transform_pre[] =\r
+ $this->info['script']->attr_transform_post[] =\r
+ new HTMLPurifier_AttrTransform_ScriptRequired();\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * XHTML 1.1 Edit Module, defines editing-related elements. Text Extension\r
+ * Module.\r
+ */\r
+class HTMLPurifier_HTMLModule_StyleAttribute extends HTMLPurifier_HTMLModule\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'StyleAttribute';\r
+\r
+ /**\r
+ * @type array\r
+ */\r
+ public $attr_collections = array(\r
+ // The inclusion routine differs from the Abstract Modules but\r
+ // is in line with the DTD and XML Schemas.\r
+ 'Style' => array('style' => false), // see constructor\r
+ 'Core' => array(0 => array('Style'))\r
+ );\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ public function setup($config)\r
+ {\r
+ $this->attr_collections['Style']['style'] = new HTMLPurifier_AttrDef_CSS();\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * XHTML 1.1 Tables Module, fully defines accessible table elements.\r
+ */\r
+class HTMLPurifier_HTMLModule_Tables extends HTMLPurifier_HTMLModule\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'Tables';\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ public function setup($config)\r
+ {\r
+ $this->addElement('caption', false, 'Inline', 'Common');\r
+\r
+ $this->addElement(\r
+ 'table',\r
+ 'Block',\r
+ new HTMLPurifier_ChildDef_Table(),\r
+ 'Common',\r
+ array(\r
+ 'border' => 'Pixels',\r
+ 'cellpadding' => 'Length',\r
+ 'cellspacing' => 'Length',\r
+ 'frame' => 'Enum#void,above,below,hsides,lhs,rhs,vsides,box,border',\r
+ 'rules' => 'Enum#none,groups,rows,cols,all',\r
+ 'summary' => 'Text',\r
+ 'width' => 'Length'\r
+ )\r
+ );\r
+\r
+ // common attributes\r
+ $cell_align = array(\r
+ 'align' => 'Enum#left,center,right,justify,char',\r
+ 'charoff' => 'Length',\r
+ 'valign' => 'Enum#top,middle,bottom,baseline',\r
+ );\r
+\r
+ $cell_t = array_merge(\r
+ array(\r
+ 'abbr' => 'Text',\r
+ 'colspan' => 'Number',\r
+ 'rowspan' => 'Number',\r
+ // Apparently, as of HTML5 this attribute only applies\r
+ // to 'th' elements.\r
+ 'scope' => 'Enum#row,col,rowgroup,colgroup',\r
+ ),\r
+ $cell_align\r
+ );\r
+ $this->addElement('td', false, 'Flow', 'Common', $cell_t);\r
+ $this->addElement('th', false, 'Flow', 'Common', $cell_t);\r
+\r
+ $this->addElement('tr', false, 'Required: td | th', 'Common', $cell_align);\r
+\r
+ $cell_col = array_merge(\r
+ array(\r
+ 'span' => 'Number',\r
+ 'width' => 'MultiLength',\r
+ ),\r
+ $cell_align\r
+ );\r
+ $this->addElement('col', false, 'Empty', 'Common', $cell_col);\r
+ $this->addElement('colgroup', false, 'Optional: col', 'Common', $cell_col);\r
+\r
+ $this->addElement('tbody', false, 'Required: tr', 'Common', $cell_align);\r
+ $this->addElement('thead', false, 'Required: tr', 'Common', $cell_align);\r
+ $this->addElement('tfoot', false, 'Required: tr', 'Common', $cell_align);\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * XHTML 1.1 Target Module, defines target attribute in link elements.\r
+ */\r
+class HTMLPurifier_HTMLModule_Target extends HTMLPurifier_HTMLModule\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'Target';\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ public function setup($config)\r
+ {\r
+ $elements = array('a');\r
+ foreach ($elements as $name) {\r
+ $e = $this->addBlankElement($name);\r
+ $e->attr = array(\r
+ 'target' => new HTMLPurifier_AttrDef_HTML_FrameTarget()\r
+ );\r
+ }\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Module adds the target=blank attribute transformation to a tags. It\r
+ * is enabled by HTML.TargetBlank\r
+ */\r
+class HTMLPurifier_HTMLModule_TargetBlank extends HTMLPurifier_HTMLModule\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'TargetBlank';\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ public function setup($config)\r
+ {\r
+ $a = $this->addBlankElement('a');\r
+ $a->attr_transform_post[] = new HTMLPurifier_AttrTransform_TargetBlank();\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Module adds the target-based noopener attribute transformation to a tags. It\r
+ * is enabled by HTML.TargetNoopener\r
+ */\r
+class HTMLPurifier_HTMLModule_TargetNoopener extends HTMLPurifier_HTMLModule\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'TargetNoopener';\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ public function setup($config) {\r
+ $a = $this->addBlankElement('a');\r
+ $a->attr_transform_post[] = new HTMLPurifier_AttrTransform_TargetNoopener();\r
+ }\r
+}\r
+
+\r
+\r
+/**\r
+ * Module adds the target-based noreferrer attribute transformation to a tags. It\r
+ * is enabled by HTML.TargetNoreferrer\r
+ */\r
+class HTMLPurifier_HTMLModule_TargetNoreferrer extends HTMLPurifier_HTMLModule\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'TargetNoreferrer';\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ public function setup($config) {\r
+ $a = $this->addBlankElement('a');\r
+ $a->attr_transform_post[] = new HTMLPurifier_AttrTransform_TargetNoreferrer();\r
+ }\r
+}\r
+
+\r
+\r
+/**\r
+ * XHTML 1.1 Text Module, defines basic text containers. Core Module.\r
+ * @note In the normative XML Schema specification, this module\r
+ * is further abstracted into the following modules:\r
+ * - Block Phrasal (address, blockquote, pre, h1, h2, h3, h4, h5, h6)\r
+ * - Block Structural (div, p)\r
+ * - Inline Phrasal (abbr, acronym, cite, code, dfn, em, kbd, q, samp, strong, var)\r
+ * - Inline Structural (br, span)\r
+ * This module, functionally, does not distinguish between these\r
+ * sub-modules, but the code is internally structured to reflect\r
+ * these distinctions.\r
+ */\r
+class HTMLPurifier_HTMLModule_Text extends HTMLPurifier_HTMLModule\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'Text';\r
+\r
+ /**\r
+ * @type array\r
+ */\r
+ public $content_sets = array(\r
+ 'Flow' => 'Heading | Block | Inline'\r
+ );\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ */\r
+ public function setup($config)\r
+ {\r
+ // Inline Phrasal -------------------------------------------------\r
+ $this->addElement('abbr', 'Inline', 'Inline', 'Common');\r
+ $this->addElement('acronym', 'Inline', 'Inline', 'Common');\r
+ $this->addElement('cite', 'Inline', 'Inline', 'Common');\r
+ $this->addElement('dfn', 'Inline', 'Inline', 'Common');\r
+ $this->addElement('kbd', 'Inline', 'Inline', 'Common');\r
+ $this->addElement('q', 'Inline', 'Inline', 'Common', array('cite' => 'URI'));\r
+ $this->addElement('samp', 'Inline', 'Inline', 'Common');\r
+ $this->addElement('var', 'Inline', 'Inline', 'Common');\r
+\r
+ $em = $this->addElement('em', 'Inline', 'Inline', 'Common');\r
+ $em->formatting = true;\r
+\r
+ $strong = $this->addElement('strong', 'Inline', 'Inline', 'Common');\r
+ $strong->formatting = true;\r
+\r
+ $code = $this->addElement('code', 'Inline', 'Inline', 'Common');\r
+ $code->formatting = true;\r
+\r
+ // Inline Structural ----------------------------------------------\r
+ $this->addElement('span', 'Inline', 'Inline', 'Common');\r
+ $this->addElement('br', 'Inline', 'Empty', 'Core');\r
+\r
+ // Block Phrasal --------------------------------------------------\r
+ $this->addElement('address', 'Block', 'Inline', 'Common');\r
+ $this->addElement('blockquote', 'Block', 'Optional: Heading | Block | List', 'Common', array('cite' => 'URI'));\r
+ $pre = $this->addElement('pre', 'Block', 'Inline', 'Common');\r
+ $pre->excludes = $this->makeLookup(\r
+ 'img',\r
+ 'big',\r
+ 'small',\r
+ 'object',\r
+ 'applet',\r
+ 'font',\r
+ 'basefont'\r
+ );\r
+ $this->addElement('h1', 'Heading', 'Inline', 'Common');\r
+ $this->addElement('h2', 'Heading', 'Inline', 'Common');\r
+ $this->addElement('h3', 'Heading', 'Inline', 'Common');\r
+ $this->addElement('h4', 'Heading', 'Inline', 'Common');\r
+ $this->addElement('h5', 'Heading', 'Inline', 'Common');\r
+ $this->addElement('h6', 'Heading', 'Inline', 'Common');\r
+\r
+ // Block Structural -----------------------------------------------\r
+ $p = $this->addElement('p', 'Block', 'Inline', 'Common');\r
+ $p->autoclose = array_flip(\r
+ array("address", "blockquote", "center", "dir", "div", "dl", "fieldset", "ol", "p", "ul")\r
+ );\r
+\r
+ $this->addElement('div', 'Block', 'Flow', 'Common');\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Abstract class for a set of proprietary modules that clean up (tidy)\r
+ * poorly written HTML.\r
+ * @todo Figure out how to protect some of these methods/properties\r
+ */\r
+class HTMLPurifier_HTMLModule_Tidy extends HTMLPurifier_HTMLModule\r
+{\r
+ /**\r
+ * List of supported levels.\r
+ * Index zero is a special case "no fixes" level.\r
+ * @type array\r
+ */\r
+ public $levels = array(0 => 'none', 'light', 'medium', 'heavy');\r
+\r
+ /**\r
+ * Default level to place all fixes in.\r
+ * Disabled by default.\r
+ * @type string\r
+ */\r
+ public $defaultLevel = null;\r
+\r
+ /**\r
+ * Lists of fixes used by getFixesForLevel().\r
+ * Format is:\r
+ * HTMLModule_Tidy->fixesForLevel[$level] = array('fix-1', 'fix-2');\r
+ * @type array\r
+ */\r
+ public $fixesForLevel = array(\r
+ 'light' => array(),\r
+ 'medium' => array(),\r
+ 'heavy' => array()\r
+ );\r
+\r
+ /**\r
+ * Lazy load constructs the module by determining the necessary\r
+ * fixes to create and then delegating to the populate() function.\r
+ * @param HTMLPurifier_Config $config\r
+ * @todo Wildcard matching and error reporting when an added or\r
+ * subtracted fix has no effect.\r
+ */\r
+ public function setup($config)\r
+ {\r
+ // create fixes, initialize fixesForLevel\r
+ $fixes = $this->makeFixes();\r
+ $this->makeFixesForLevel($fixes);\r
+\r
+ // figure out which fixes to use\r
+ $level = $config->get('HTML.TidyLevel');\r
+ $fixes_lookup = $this->getFixesForLevel($level);\r
+\r
+ // get custom fix declarations: these need namespace processing\r
+ $add_fixes = $config->get('HTML.TidyAdd');\r
+ $remove_fixes = $config->get('HTML.TidyRemove');\r
+\r
+ foreach ($fixes as $name => $fix) {\r
+ // needs to be refactored a little to implement globbing\r
+ if (isset($remove_fixes[$name]) ||\r
+ (!isset($add_fixes[$name]) && !isset($fixes_lookup[$name]))) {\r
+ unset($fixes[$name]);\r
+ }\r
+ }\r
+\r
+ // populate this module with necessary fixes\r
+ $this->populate($fixes);\r
+ }\r
+\r
+ /**\r
+ * Retrieves all fixes per a level, returning fixes for that specific\r
+ * level as well as all levels below it.\r
+ * @param string $level level identifier, see $levels for valid values\r
+ * @return array Lookup up table of fixes\r
+ */\r
+ public function getFixesForLevel($level)\r
+ {\r
+ if ($level == $this->levels[0]) {\r
+ return array();\r
+ }\r
+ $activated_levels = array();\r
+ for ($i = 1, $c = count($this->levels); $i < $c; $i++) {\r
+ $activated_levels[] = $this->levels[$i];\r
+ if ($this->levels[$i] == $level) {\r
+ break;\r
+ }\r
+ }\r
+ if ($i == $c) {\r
+ trigger_error(\r
+ 'Tidy level ' . htmlspecialchars($level) . ' not recognized',\r
+ E_USER_WARNING\r
+ );\r
+ return array();\r
+ }\r
+ $ret = array();\r
+ foreach ($activated_levels as $level) {\r
+ foreach ($this->fixesForLevel[$level] as $fix) {\r
+ $ret[$fix] = true;\r
+ }\r
+ }\r
+ return $ret;\r
+ }\r
+\r
+ /**\r
+ * Dynamically populates the $fixesForLevel member variable using\r
+ * the fixes array. It may be custom overloaded, used in conjunction\r
+ * with $defaultLevel, or not used at all.\r
+ * @param array $fixes\r
+ */\r
+ public function makeFixesForLevel($fixes)\r
+ {\r
+ if (!isset($this->defaultLevel)) {\r
+ return;\r
+ }\r
+ if (!isset($this->fixesForLevel[$this->defaultLevel])) {\r
+ trigger_error(\r
+ 'Default level ' . $this->defaultLevel . ' does not exist',\r
+ E_USER_ERROR\r
+ );\r
+ return;\r
+ }\r
+ $this->fixesForLevel[$this->defaultLevel] = array_keys($fixes);\r
+ }\r
+\r
+ /**\r
+ * Populates the module with transforms and other special-case code\r
+ * based on a list of fixes passed to it\r
+ * @param array $fixes Lookup table of fixes to activate\r
+ */\r
+ public function populate($fixes)\r
+ {\r
+ foreach ($fixes as $name => $fix) {\r
+ // determine what the fix is for\r
+ list($type, $params) = $this->getFixType($name);\r
+ switch ($type) {\r
+ case 'attr_transform_pre':\r
+ case 'attr_transform_post':\r
+ $attr = $params['attr'];\r
+ if (isset($params['element'])) {\r
+ $element = $params['element'];\r
+ if (empty($this->info[$element])) {\r
+ $e = $this->addBlankElement($element);\r
+ } else {\r
+ $e = $this->info[$element];\r
+ }\r
+ } else {\r
+ $type = "info_$type";\r
+ $e = $this;\r
+ }\r
+ // PHP does some weird parsing when I do\r
+ // $e->$type[$attr], so I have to assign a ref.\r
+ $f =& $e->$type;\r
+ $f[$attr] = $fix;\r
+ break;\r
+ case 'tag_transform':\r
+ $this->info_tag_transform[$params['element']] = $fix;\r
+ break;\r
+ case 'child':\r
+ case 'content_model_type':\r
+ $element = $params['element'];\r
+ if (empty($this->info[$element])) {\r
+ $e = $this->addBlankElement($element);\r
+ } else {\r
+ $e = $this->info[$element];\r
+ }\r
+ $e->$type = $fix;\r
+ break;\r
+ default:\r
+ trigger_error("Fix type $type not supported", E_USER_ERROR);\r
+ break;\r
+ }\r
+ }\r
+ }\r
+\r
+ /**\r
+ * Parses a fix name and determines what kind of fix it is, as well\r
+ * as other information defined by the fix\r
+ * @param $name String name of fix\r
+ * @return array(string $fix_type, array $fix_parameters)\r
+ * @note $fix_parameters is type dependant, see populate() for usage\r
+ * of these parameters\r
+ */\r
+ public function getFixType($name)\r
+ {\r
+ // parse it\r
+ $property = $attr = null;\r
+ if (strpos($name, '#') !== false) {\r
+ list($name, $property) = explode('#', $name);\r
+ }\r
+ if (strpos($name, '@') !== false) {\r
+ list($name, $attr) = explode('@', $name);\r
+ }\r
+\r
+ // figure out the parameters\r
+ $params = array();\r
+ if ($name !== '') {\r
+ $params['element'] = $name;\r
+ }\r
+ if (!is_null($attr)) {\r
+ $params['attr'] = $attr;\r
+ }\r
+\r
+ // special case: attribute transform\r
+ if (!is_null($attr)) {\r
+ if (is_null($property)) {\r
+ $property = 'pre';\r
+ }\r
+ $type = 'attr_transform_' . $property;\r
+ return array($type, $params);\r
+ }\r
+\r
+ // special case: tag transform\r
+ if (is_null($property)) {\r
+ return array('tag_transform', $params);\r
+ }\r
+\r
+ return array($property, $params);\r
+\r
+ }\r
+\r
+ /**\r
+ * Defines all fixes the module will perform in a compact\r
+ * associative array of fix name to fix implementation.\r
+ * @return array\r
+ */\r
+ public function makeFixes()\r
+ {\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+class HTMLPurifier_HTMLModule_XMLCommonAttributes extends HTMLPurifier_HTMLModule\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'XMLCommonAttributes';\r
+\r
+ /**\r
+ * @type array\r
+ */\r
+ public $attr_collections = array(\r
+ 'Lang' => array(\r
+ 'xml:lang' => 'LanguageCode',\r
+ )\r
+ );\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Name is deprecated, but allowed in strict doctypes, so onl\r
+ */\r
+class HTMLPurifier_HTMLModule_Tidy_Name extends HTMLPurifier_HTMLModule_Tidy\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'Tidy_Name';\r
+\r
+ /**\r
+ * @type string\r
+ */\r
+ public $defaultLevel = 'heavy';\r
+\r
+ /**\r
+ * @return array\r
+ */\r
+ public function makeFixes()\r
+ {\r
+ $r = array();\r
+ // @name for img, a -----------------------------------------------\r
+ // Technically, it's allowed even on strict, so we allow authors to use\r
+ // it. However, it's deprecated in future versions of XHTML.\r
+ $r['img@name'] =\r
+ $r['a@name'] = new HTMLPurifier_AttrTransform_Name();\r
+ return $r;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+class HTMLPurifier_HTMLModule_Tidy_Proprietary extends HTMLPurifier_HTMLModule_Tidy\r
+{\r
+\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'Tidy_Proprietary';\r
+\r
+ /**\r
+ * @type string\r
+ */\r
+ public $defaultLevel = 'light';\r
+\r
+ /**\r
+ * @return array\r
+ */\r
+ public function makeFixes()\r
+ {\r
+ $r = array();\r
+ $r['table@background'] = new HTMLPurifier_AttrTransform_Background();\r
+ $r['td@background'] = new HTMLPurifier_AttrTransform_Background();\r
+ $r['th@background'] = new HTMLPurifier_AttrTransform_Background();\r
+ $r['tr@background'] = new HTMLPurifier_AttrTransform_Background();\r
+ $r['thead@background'] = new HTMLPurifier_AttrTransform_Background();\r
+ $r['tfoot@background'] = new HTMLPurifier_AttrTransform_Background();\r
+ $r['tbody@background'] = new HTMLPurifier_AttrTransform_Background();\r
+ $r['table@height'] = new HTMLPurifier_AttrTransform_Length('height');\r
+ return $r;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+class HTMLPurifier_HTMLModule_Tidy_XHTMLAndHTML4 extends HTMLPurifier_HTMLModule_Tidy\r
+{\r
+\r
+ /**\r
+ * @return array\r
+ */\r
+ public function makeFixes()\r
+ {\r
+ $r = array();\r
+\r
+ // == deprecated tag transforms ===================================\r
+\r
+ $r['font'] = new HTMLPurifier_TagTransform_Font();\r
+ $r['menu'] = new HTMLPurifier_TagTransform_Simple('ul');\r
+ $r['dir'] = new HTMLPurifier_TagTransform_Simple('ul');\r
+ $r['center'] = new HTMLPurifier_TagTransform_Simple('div', 'text-align:center;');\r
+ $r['u'] = new HTMLPurifier_TagTransform_Simple('span', 'text-decoration:underline;');\r
+ $r['s'] = new HTMLPurifier_TagTransform_Simple('span', 'text-decoration:line-through;');\r
+ $r['strike'] = new HTMLPurifier_TagTransform_Simple('span', 'text-decoration:line-through;');\r
+\r
+ // == deprecated attribute transforms =============================\r
+\r
+ $r['caption@align'] =\r
+ new HTMLPurifier_AttrTransform_EnumToCSS(\r
+ 'align',\r
+ array(\r
+ // we're following IE's behavior, not Firefox's, due\r
+ // to the fact that no one supports caption-side:right,\r
+ // W3C included (with CSS 2.1). This is a slightly\r
+ // unreasonable attribute!\r
+ 'left' => 'text-align:left;',\r
+ 'right' => 'text-align:right;',\r
+ 'top' => 'caption-side:top;',\r
+ 'bottom' => 'caption-side:bottom;' // not supported by IE\r
+ )\r
+ );\r
+\r
+ // @align for img -------------------------------------------------\r
+ $r['img@align'] =\r
+ new HTMLPurifier_AttrTransform_EnumToCSS(\r
+ 'align',\r
+ array(\r
+ 'left' => 'float:left;',\r
+ 'right' => 'float:right;',\r
+ 'top' => 'vertical-align:top;',\r
+ 'middle' => 'vertical-align:middle;',\r
+ 'bottom' => 'vertical-align:baseline;',\r
+ )\r
+ );\r
+\r
+ // @align for table -----------------------------------------------\r
+ $r['table@align'] =\r
+ new HTMLPurifier_AttrTransform_EnumToCSS(\r
+ 'align',\r
+ array(\r
+ 'left' => 'float:left;',\r
+ 'center' => 'margin-left:auto;margin-right:auto;',\r
+ 'right' => 'float:right;'\r
+ )\r
+ );\r
+\r
+ // @align for hr -----------------------------------------------\r
+ $r['hr@align'] =\r
+ new HTMLPurifier_AttrTransform_EnumToCSS(\r
+ 'align',\r
+ array(\r
+ // we use both text-align and margin because these work\r
+ // for different browsers (IE and Firefox, respectively)\r
+ // and the melange makes for a pretty cross-compatible\r
+ // solution\r
+ 'left' => 'margin-left:0;margin-right:auto;text-align:left;',\r
+ 'center' => 'margin-left:auto;margin-right:auto;text-align:center;',\r
+ 'right' => 'margin-left:auto;margin-right:0;text-align:right;'\r
+ )\r
+ );\r
+\r
+ // @align for h1, h2, h3, h4, h5, h6, p, div ----------------------\r
+ // {{{\r
+ $align_lookup = array();\r
+ $align_values = array('left', 'right', 'center', 'justify');\r
+ foreach ($align_values as $v) {\r
+ $align_lookup[$v] = "text-align:$v;";\r
+ }\r
+ // }}}\r
+ $r['h1@align'] =\r
+ $r['h2@align'] =\r
+ $r['h3@align'] =\r
+ $r['h4@align'] =\r
+ $r['h5@align'] =\r
+ $r['h6@align'] =\r
+ $r['p@align'] =\r
+ $r['div@align'] =\r
+ new HTMLPurifier_AttrTransform_EnumToCSS('align', $align_lookup);\r
+\r
+ // @bgcolor for table, tr, td, th ---------------------------------\r
+ $r['table@bgcolor'] =\r
+ $r['td@bgcolor'] =\r
+ $r['th@bgcolor'] =\r
+ new HTMLPurifier_AttrTransform_BgColor();\r
+\r
+ // @border for img ------------------------------------------------\r
+ $r['img@border'] = new HTMLPurifier_AttrTransform_Border();\r
+\r
+ // @clear for br --------------------------------------------------\r
+ $r['br@clear'] =\r
+ new HTMLPurifier_AttrTransform_EnumToCSS(\r
+ 'clear',\r
+ array(\r
+ 'left' => 'clear:left;',\r
+ 'right' => 'clear:right;',\r
+ 'all' => 'clear:both;',\r
+ 'none' => 'clear:none;',\r
+ )\r
+ );\r
+\r
+ // @height for td, th ---------------------------------------------\r
+ $r['td@height'] =\r
+ $r['th@height'] =\r
+ new HTMLPurifier_AttrTransform_Length('height');\r
+\r
+ // @hspace for img ------------------------------------------------\r
+ $r['img@hspace'] = new HTMLPurifier_AttrTransform_ImgSpace('hspace');\r
+\r
+ // @noshade for hr ------------------------------------------------\r
+ // this transformation is not precise but often good enough.\r
+ // different browsers use different styles to designate noshade\r
+ $r['hr@noshade'] =\r
+ new HTMLPurifier_AttrTransform_BoolToCSS(\r
+ 'noshade',\r
+ 'color:#808080;background-color:#808080;border:0;'\r
+ );\r
+\r
+ // @nowrap for td, th ---------------------------------------------\r
+ $r['td@nowrap'] =\r
+ $r['th@nowrap'] =\r
+ new HTMLPurifier_AttrTransform_BoolToCSS(\r
+ 'nowrap',\r
+ 'white-space:nowrap;'\r
+ );\r
+\r
+ // @size for hr --------------------------------------------------\r
+ $r['hr@size'] = new HTMLPurifier_AttrTransform_Length('size', 'height');\r
+\r
+ // @type for li, ol, ul -------------------------------------------\r
+ // {{{\r
+ $ul_types = array(\r
+ 'disc' => 'list-style-type:disc;',\r
+ 'square' => 'list-style-type:square;',\r
+ 'circle' => 'list-style-type:circle;'\r
+ );\r
+ $ol_types = array(\r
+ '1' => 'list-style-type:decimal;',\r
+ 'i' => 'list-style-type:lower-roman;',\r
+ 'I' => 'list-style-type:upper-roman;',\r
+ 'a' => 'list-style-type:lower-alpha;',\r
+ 'A' => 'list-style-type:upper-alpha;'\r
+ );\r
+ $li_types = $ul_types + $ol_types;\r
+ // }}}\r
+\r
+ $r['ul@type'] = new HTMLPurifier_AttrTransform_EnumToCSS('type', $ul_types);\r
+ $r['ol@type'] = new HTMLPurifier_AttrTransform_EnumToCSS('type', $ol_types, true);\r
+ $r['li@type'] = new HTMLPurifier_AttrTransform_EnumToCSS('type', $li_types, true);\r
+\r
+ // @vspace for img ------------------------------------------------\r
+ $r['img@vspace'] = new HTMLPurifier_AttrTransform_ImgSpace('vspace');\r
+\r
+ // @width for hr, td, th ------------------------------------------\r
+ $r['td@width'] =\r
+ $r['th@width'] =\r
+ $r['hr@width'] = new HTMLPurifier_AttrTransform_Length('width');\r
+\r
+ return $r;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+class HTMLPurifier_HTMLModule_Tidy_Strict extends HTMLPurifier_HTMLModule_Tidy_XHTMLAndHTML4\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'Tidy_Strict';\r
+\r
+ /**\r
+ * @type string\r
+ */\r
+ public $defaultLevel = 'light';\r
+\r
+ /**\r
+ * @return array\r
+ */\r
+ public function makeFixes()\r
+ {\r
+ $r = parent::makeFixes();\r
+ $r['blockquote#content_model_type'] = 'strictblockquote';\r
+ return $r;\r
+ }\r
+\r
+ /**\r
+ * @type bool\r
+ */\r
+ public $defines_child_def = true;\r
+\r
+ /**\r
+ * @param HTMLPurifier_ElementDef $def\r
+ * @return HTMLPurifier_ChildDef_StrictBlockquote\r
+ */\r
+ public function getChildDef($def)\r
+ {\r
+ if ($def->content_model_type != 'strictblockquote') {\r
+ return parent::getChildDef($def);\r
+ }\r
+ return new HTMLPurifier_ChildDef_StrictBlockquote($def->content_model);\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+class HTMLPurifier_HTMLModule_Tidy_Transitional extends HTMLPurifier_HTMLModule_Tidy_XHTMLAndHTML4\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'Tidy_Transitional';\r
+\r
+ /**\r
+ * @type string\r
+ */\r
+ public $defaultLevel = 'heavy';\r
+}\r
+\r
+
+
+\r
+\r
+class HTMLPurifier_HTMLModule_Tidy_XHTML extends HTMLPurifier_HTMLModule_Tidy\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'Tidy_XHTML';\r
+\r
+ /**\r
+ * @type string\r
+ */\r
+ public $defaultLevel = 'medium';\r
+\r
+ /**\r
+ * @return array\r
+ */\r
+ public function makeFixes()\r
+ {\r
+ $r = array();\r
+ $r['@lang'] = new HTMLPurifier_AttrTransform_Lang();\r
+ return $r;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Injector that auto paragraphs text in the root node based on\r
+ * double-spacing.\r
+ * @todo Ensure all states are unit tested, including variations as well.\r
+ * @todo Make a graph of the flow control for this Injector.\r
+ */\r
+class HTMLPurifier_Injector_AutoParagraph extends HTMLPurifier_Injector\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'AutoParagraph';\r
+\r
+ /**\r
+ * @type array\r
+ */\r
+ public $needed = array('p');\r
+\r
+ /**\r
+ * @return HTMLPurifier_Token_Start\r
+ */\r
+ private function _pStart()\r
+ {\r
+ $par = new HTMLPurifier_Token_Start('p');\r
+ $par->armor['MakeWellFormed_TagClosedError'] = true;\r
+ return $par;\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Token_Text $token\r
+ */\r
+ public function handleText(&$token)\r
+ {\r
+ $text = $token->data;\r
+ // Does the current parent allow <p> tags?\r
+ if ($this->allowsElement('p')) {\r
+ if (empty($this->currentNesting) || strpos($text, "\n\n") !== false) {\r
+ // Note that we have differing behavior when dealing with text\r
+ // in the anonymous root node, or a node inside the document.\r
+ // If the text as a double-newline, the treatment is the same;\r
+ // if it doesn't, see the next if-block if you're in the document.\r
+\r
+ $i = $nesting = null;\r
+ if (!$this->forwardUntilEndToken($i, $current, $nesting) && $token->is_whitespace) {\r
+ // State 1.1: ... ^ (whitespace, then document end)\r
+ // ----\r
+ // This is a degenerate case\r
+ } else {\r
+ if (!$token->is_whitespace || $this->_isInline($current)) {\r
+ // State 1.2: PAR1\r
+ // ----\r
+\r
+ // State 1.3: PAR1\n\nPAR2\r
+ // ------------\r
+\r
+ // State 1.4: <div>PAR1\n\nPAR2 (see State 2)\r
+ // ------------\r
+ $token = array($this->_pStart());\r
+ $this->_splitText($text, $token);\r
+ } else {\r
+ // State 1.5: \n<hr />\r
+ // --\r
+ }\r
+ }\r
+ } else {\r
+ // State 2: <div>PAR1... (similar to 1.4)\r
+ // ----\r
+\r
+ // We're in an element that allows paragraph tags, but we're not\r
+ // sure if we're going to need them.\r
+ if ($this->_pLookAhead()) {\r
+ // State 2.1: <div>PAR1<b>PAR1\n\nPAR2\r
+ // ----\r
+ // Note: This will always be the first child, since any\r
+ // previous inline element would have triggered this very\r
+ // same routine, and found the double newline. One possible\r
+ // exception would be a comment.\r
+ $token = array($this->_pStart(), $token);\r
+ } else {\r
+ // State 2.2.1: <div>PAR1<div>\r
+ // ----\r
+\r
+ // State 2.2.2: <div>PAR1<b>PAR1</b></div>\r
+ // ----\r
+ }\r
+ }\r
+ // Is the current parent a <p> tag?\r
+ } elseif (!empty($this->currentNesting) &&\r
+ $this->currentNesting[count($this->currentNesting) - 1]->name == 'p') {\r
+ // State 3.1: ...<p>PAR1\r
+ // ----\r
+\r
+ // State 3.2: ...<p>PAR1\n\nPAR2\r
+ // ------------\r
+ $token = array();\r
+ $this->_splitText($text, $token);\r
+ // Abort!\r
+ } else {\r
+ // State 4.1: ...<b>PAR1\r
+ // ----\r
+\r
+ // State 4.2: ...<b>PAR1\n\nPAR2\r
+ // ------------\r
+ }\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Token $token\r
+ */\r
+ public function handleElement(&$token)\r
+ {\r
+ // We don't have to check if we're already in a <p> tag for block\r
+ // tokens, because the tag would have been autoclosed by MakeWellFormed.\r
+ if ($this->allowsElement('p')) {\r
+ if (!empty($this->currentNesting)) {\r
+ if ($this->_isInline($token)) {\r
+ // State 1: <div>...<b>\r
+ // ---\r
+ // Check if this token is adjacent to the parent token\r
+ // (seek backwards until token isn't whitespace)\r
+ $i = null;\r
+ $this->backward($i, $prev);\r
+\r
+ if (!$prev instanceof HTMLPurifier_Token_Start) {\r
+ // Token wasn't adjacent\r
+ if ($prev instanceof HTMLPurifier_Token_Text &&\r
+ substr($prev->data, -2) === "\n\n"\r
+ ) {\r
+ // State 1.1.4: <div><p>PAR1</p>\n\n<b>\r
+ // ---\r
+ // Quite frankly, this should be handled by splitText\r
+ $token = array($this->_pStart(), $token);\r
+ } else {\r
+ // State 1.1.1: <div><p>PAR1</p><b>\r
+ // ---\r
+ // State 1.1.2: <div><br /><b>\r
+ // ---\r
+ // State 1.1.3: <div>PAR<b>\r
+ // ---\r
+ }\r
+ } else {\r
+ // State 1.2.1: <div><b>\r
+ // ---\r
+ // Lookahead to see if <p> is needed.\r
+ if ($this->_pLookAhead()) {\r
+ // State 1.3.1: <div><b>PAR1\n\nPAR2\r
+ // ---\r
+ $token = array($this->_pStart(), $token);\r
+ } else {\r
+ // State 1.3.2: <div><b>PAR1</b></div>\r
+ // ---\r
+\r
+ // State 1.3.3: <div><b>PAR1</b><div></div>\n\n</div>\r
+ // ---\r
+ }\r
+ }\r
+ } else {\r
+ // State 2.3: ...<div>\r
+ // -----\r
+ }\r
+ } else {\r
+ if ($this->_isInline($token)) {\r
+ // State 3.1: <b>\r
+ // ---\r
+ // This is where the {p} tag is inserted, not reflected in\r
+ // inputTokens yet, however.\r
+ $token = array($this->_pStart(), $token);\r
+ } else {\r
+ // State 3.2: <div>\r
+ // -----\r
+ }\r
+\r
+ $i = null;\r
+ if ($this->backward($i, $prev)) {\r
+ if (!$prev instanceof HTMLPurifier_Token_Text) {\r
+ // State 3.1.1: ...</p>{p}<b>\r
+ // ---\r
+ // State 3.2.1: ...</p><div>\r
+ // -----\r
+ if (!is_array($token)) {\r
+ $token = array($token);\r
+ }\r
+ array_unshift($token, new HTMLPurifier_Token_Text("\n\n"));\r
+ } else {\r
+ // State 3.1.2: ...</p>\n\n{p}<b>\r
+ // ---\r
+ // State 3.2.2: ...</p>\n\n<div>\r
+ // -----\r
+ // Note: PAR<ELEM> cannot occur because PAR would have been\r
+ // wrapped in <p> tags.\r
+ }\r
+ }\r
+ }\r
+ } else {\r
+ // State 2.2: <ul><li>\r
+ // ----\r
+ // State 2.4: <p><b>\r
+ // ---\r
+ }\r
+ }\r
+\r
+ /**\r
+ * Splits up a text in paragraph tokens and appends them\r
+ * to the result stream that will replace the original\r
+ * @param string $data String text data that will be processed\r
+ * into paragraphs\r
+ * @param HTMLPurifier_Token[] $result Reference to array of tokens that the\r
+ * tags will be appended onto\r
+ */\r
+ private function _splitText($data, &$result)\r
+ {\r
+ $raw_paragraphs = explode("\n\n", $data);\r
+ $paragraphs = array(); // without empty paragraphs\r
+ $needs_start = false;\r
+ $needs_end = false;\r
+\r
+ $c = count($raw_paragraphs);\r
+ if ($c == 1) {\r
+ // There were no double-newlines, abort quickly. In theory this\r
+ // should never happen.\r
+ $result[] = new HTMLPurifier_Token_Text($data);\r
+ return;\r
+ }\r
+ for ($i = 0; $i < $c; $i++) {\r
+ $par = $raw_paragraphs[$i];\r
+ if (trim($par) !== '') {\r
+ $paragraphs[] = $par;\r
+ } else {\r
+ if ($i == 0) {\r
+ // Double newline at the front\r
+ if (empty($result)) {\r
+ // The empty result indicates that the AutoParagraph\r
+ // injector did not add any start paragraph tokens.\r
+ // This means that we have been in a paragraph for\r
+ // a while, and the newline means we should start a new one.\r
+ $result[] = new HTMLPurifier_Token_End('p');\r
+ $result[] = new HTMLPurifier_Token_Text("\n\n");\r
+ // However, the start token should only be added if\r
+ // there is more processing to be done (i.e. there are\r
+ // real paragraphs in here). If there are none, the\r
+ // next start paragraph tag will be handled by the\r
+ // next call to the injector\r
+ $needs_start = true;\r
+ } else {\r
+ // We just started a new paragraph!\r
+ // Reinstate a double-newline for presentation's sake, since\r
+ // it was in the source code.\r
+ array_unshift($result, new HTMLPurifier_Token_Text("\n\n"));\r
+ }\r
+ } elseif ($i + 1 == $c) {\r
+ // Double newline at the end\r
+ // There should be a trailing </p> when we're finally done.\r
+ $needs_end = true;\r
+ }\r
+ }\r
+ }\r
+\r
+ // Check if this was just a giant blob of whitespace. Move this earlier,\r
+ // perhaps?\r
+ if (empty($paragraphs)) {\r
+ return;\r
+ }\r
+\r
+ // Add the start tag indicated by \n\n at the beginning of $data\r
+ if ($needs_start) {\r
+ $result[] = $this->_pStart();\r
+ }\r
+\r
+ // Append the paragraphs onto the result\r
+ foreach ($paragraphs as $par) {\r
+ $result[] = new HTMLPurifier_Token_Text($par);\r
+ $result[] = new HTMLPurifier_Token_End('p');\r
+ $result[] = new HTMLPurifier_Token_Text("\n\n");\r
+ $result[] = $this->_pStart();\r
+ }\r
+\r
+ // Remove trailing start token; Injector will handle this later if\r
+ // it was indeed needed. This prevents from needing to do a lookahead,\r
+ // at the cost of a lookbehind later.\r
+ array_pop($result);\r
+\r
+ // If there is no need for an end tag, remove all of it and let\r
+ // MakeWellFormed close it later.\r
+ if (!$needs_end) {\r
+ array_pop($result); // removes \n\n\r
+ array_pop($result); // removes </p>\r
+ }\r
+ }\r
+\r
+ /**\r
+ * Returns true if passed token is inline (and, ergo, allowed in\r
+ * paragraph tags)\r
+ * @param HTMLPurifier_Token $token\r
+ * @return bool\r
+ */\r
+ private function _isInline($token)\r
+ {\r
+ return isset($this->htmlDefinition->info['p']->child->elements[$token->name]);\r
+ }\r
+\r
+ /**\r
+ * Looks ahead in the token list and determines whether or not we need\r
+ * to insert a <p> tag.\r
+ * @return bool\r
+ */\r
+ private function _pLookAhead()\r
+ {\r
+ if ($this->currentToken instanceof HTMLPurifier_Token_Start) {\r
+ $nesting = 1;\r
+ } else {\r
+ $nesting = 0;\r
+ }\r
+ $ok = false;\r
+ $i = null;\r
+ while ($this->forwardUntilEndToken($i, $current, $nesting)) {\r
+ $result = $this->_checkNeedsP($current);\r
+ if ($result !== null) {\r
+ $ok = $result;\r
+ break;\r
+ }\r
+ }\r
+ return $ok;\r
+ }\r
+\r
+ /**\r
+ * Determines if a particular token requires an earlier inline token\r
+ * to get a paragraph. This should be used with _forwardUntilEndToken\r
+ * @param HTMLPurifier_Token $current\r
+ * @return bool\r
+ */\r
+ private function _checkNeedsP($current)\r
+ {\r
+ if ($current instanceof HTMLPurifier_Token_Start) {\r
+ if (!$this->_isInline($current)) {\r
+ // <div>PAR1<div>\r
+ // ----\r
+ // Terminate early, since we hit a block element\r
+ return false;\r
+ }\r
+ } elseif ($current instanceof HTMLPurifier_Token_Text) {\r
+ if (strpos($current->data, "\n\n") !== false) {\r
+ // <div>PAR1<b>PAR1\n\nPAR2\r
+ // ----\r
+ return true;\r
+ } else {\r
+ // <div>PAR1<b>PAR1...\r
+ // ----\r
+ }\r
+ }\r
+ return null;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Injector that displays the URL of an anchor instead of linking to it, in addition to showing the text of the link.\r
+ */\r
+class HTMLPurifier_Injector_DisplayLinkURI extends HTMLPurifier_Injector\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'DisplayLinkURI';\r
+\r
+ /**\r
+ * @type array\r
+ */\r
+ public $needed = array('a');\r
+\r
+ /**\r
+ * @param $token\r
+ */\r
+ public function handleElement(&$token)\r
+ {\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Token $token\r
+ */\r
+ public function handleEnd(&$token)\r
+ {\r
+ if (isset($token->start->attr['href'])) {\r
+ $url = $token->start->attr['href'];\r
+ unset($token->start->attr['href']);\r
+ $token = array($token, new HTMLPurifier_Token_Text(" ($url)"));\r
+ } else {\r
+ // nothing to display\r
+ }\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Injector that converts http, https and ftp text URLs to actual links.\r
+ */\r
+class HTMLPurifier_Injector_Linkify extends HTMLPurifier_Injector\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'Linkify';\r
+\r
+ /**\r
+ * @type array\r
+ */\r
+ public $needed = array('a' => array('href'));\r
+\r
+ /**\r
+ * @param HTMLPurifier_Token $token\r
+ */\r
+ public function handleText(&$token)\r
+ {\r
+ if (!$this->allowsElement('a')) {\r
+ return;\r
+ }\r
+\r
+ if (strpos($token->data, '://') === false) {\r
+ // our really quick heuristic failed, abort\r
+ // this may not work so well if we want to match things like\r
+ // "google.com", but then again, most people don't\r
+ return;\r
+ }\r
+\r
+ // there is/are URL(s). Let's split the string.\r
+ // We use this regex:\r
+ // https://gist.github.com/gruber/249502\r
+ // but with @cscott's backtracking fix and also\r
+ // the Unicode characters un-Unicodified.\r
+ $bits = preg_split(\r
+ '/\\b((?:[a-z][\\w\\-]+:(?:\\/{1,3}|[a-z0-9%])|www\\d{0,3}[.]|[a-z0-9.\\-]+[.][a-z]{2,4}\\/)(?:[^\\s()<>]|\\((?:[^\\s()<>]|(?:\\([^\\s()<>]+\\)))*\\))+(?:\\((?:[^\\s()<>]|(?:\\([^\\s()<>]+\\)))*\\)|[^\\s`!()\\[\\]{};:\'".,<>?\x{00ab}\x{00bb}\x{201c}\x{201d}\x{2018}\x{2019}]))/iu',\r
+ $token->data, -1, PREG_SPLIT_DELIM_CAPTURE);\r
+\r
+\r
+ $token = array();\r
+\r
+ // $i = index\r
+ // $c = count\r
+ // $l = is link\r
+ for ($i = 0, $c = count($bits), $l = false; $i < $c; $i++, $l = !$l) {\r
+ if (!$l) {\r
+ if ($bits[$i] === '') {\r
+ continue;\r
+ }\r
+ $token[] = new HTMLPurifier_Token_Text($bits[$i]);\r
+ } else {\r
+ $token[] = new HTMLPurifier_Token_Start('a', array('href' => $bits[$i]));\r
+ $token[] = new HTMLPurifier_Token_Text($bits[$i]);\r
+ $token[] = new HTMLPurifier_Token_End('a');\r
+ }\r
+ }\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Injector that converts configuration directive syntax %Namespace.Directive\r
+ * to links\r
+ */\r
+class HTMLPurifier_Injector_PurifierLinkify extends HTMLPurifier_Injector\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'PurifierLinkify';\r
+\r
+ /**\r
+ * @type string\r
+ */\r
+ public $docURL;\r
+\r
+ /**\r
+ * @type array\r
+ */\r
+ public $needed = array('a' => array('href'));\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return string\r
+ */\r
+ public function prepare($config, $context)\r
+ {\r
+ $this->docURL = $config->get('AutoFormat.PurifierLinkify.DocURL');\r
+ return parent::prepare($config, $context);\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Token $token\r
+ */\r
+ public function handleText(&$token)\r
+ {\r
+ if (!$this->allowsElement('a')) {\r
+ return;\r
+ }\r
+ if (strpos($token->data, '%') === false) {\r
+ return;\r
+ }\r
+\r
+ $bits = preg_split('#%([a-z0-9]+\.[a-z0-9]+)#Si', $token->data, -1, PREG_SPLIT_DELIM_CAPTURE);\r
+ $token = array();\r
+\r
+ // $i = index\r
+ // $c = count\r
+ // $l = is link\r
+ for ($i = 0, $c = count($bits), $l = false; $i < $c; $i++, $l = !$l) {\r
+ if (!$l) {\r
+ if ($bits[$i] === '') {\r
+ continue;\r
+ }\r
+ $token[] = new HTMLPurifier_Token_Text($bits[$i]);\r
+ } else {\r
+ $token[] = new HTMLPurifier_Token_Start(\r
+ 'a',\r
+ array('href' => str_replace('%s', $bits[$i], $this->docURL))\r
+ );\r
+ $token[] = new HTMLPurifier_Token_Text('%' . $bits[$i]);\r
+ $token[] = new HTMLPurifier_Token_End('a');\r
+ }\r
+ }\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+class HTMLPurifier_Injector_RemoveEmpty extends HTMLPurifier_Injector\r
+{\r
+ /**\r
+ * @type HTMLPurifier_Context\r
+ */\r
+ private $context;\r
+\r
+ /**\r
+ * @type HTMLPurifier_Config\r
+ */\r
+ private $config;\r
+\r
+ /**\r
+ * @type HTMLPurifier_AttrValidator\r
+ */\r
+ private $attrValidator;\r
+\r
+ /**\r
+ * @type bool\r
+ */\r
+ private $removeNbsp;\r
+\r
+ /**\r
+ * @type bool\r
+ */\r
+ private $removeNbspExceptions;\r
+\r
+ /**\r
+ * Cached contents of %AutoFormat.RemoveEmpty.Predicate\r
+ * @type array\r
+ */\r
+ private $exclude;\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return void\r
+ */\r
+ public function prepare($config, $context)\r
+ {\r
+ parent::prepare($config, $context);\r
+ $this->config = $config;\r
+ $this->context = $context;\r
+ $this->removeNbsp = $config->get('AutoFormat.RemoveEmpty.RemoveNbsp');\r
+ $this->removeNbspExceptions = $config->get('AutoFormat.RemoveEmpty.RemoveNbsp.Exceptions');\r
+ $this->exclude = $config->get('AutoFormat.RemoveEmpty.Predicate');\r
+ foreach ($this->exclude as $key => $attrs) {\r
+ if (!is_array($attrs)) {\r
+ // HACK, see HTMLPurifier/Printer/ConfigForm.php\r
+ $this->exclude[$key] = explode(';', $attrs);\r
+ }\r
+ }\r
+ $this->attrValidator = new HTMLPurifier_AttrValidator();\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Token $token\r
+ */\r
+ public function handleElement(&$token)\r
+ {\r
+ if (!$token instanceof HTMLPurifier_Token_Start) {\r
+ return;\r
+ }\r
+ $next = false;\r
+ $deleted = 1; // the current tag\r
+ for ($i = count($this->inputZipper->back) - 1; $i >= 0; $i--, $deleted++) {\r
+ $next = $this->inputZipper->back[$i];\r
+ if ($next instanceof HTMLPurifier_Token_Text) {\r
+ if ($next->is_whitespace) {\r
+ continue;\r
+ }\r
+ if ($this->removeNbsp && !isset($this->removeNbspExceptions[$token->name])) {\r
+ $plain = str_replace("\xC2\xA0", "", $next->data);\r
+ $isWsOrNbsp = $plain === '' || ctype_space($plain);\r
+ if ($isWsOrNbsp) {\r
+ continue;\r
+ }\r
+ }\r
+ }\r
+ break;\r
+ }\r
+ if (!$next || ($next instanceof HTMLPurifier_Token_End && $next->name == $token->name)) {\r
+ $this->attrValidator->validateToken($token, $this->config, $this->context);\r
+ $token->armor['ValidateAttributes'] = true;\r
+ if (isset($this->exclude[$token->name])) {\r
+ $r = true;\r
+ foreach ($this->exclude[$token->name] as $elem) {\r
+ if (!isset($token->attr[$elem])) $r = false;\r
+ }\r
+ if ($r) return;\r
+ }\r
+ if (isset($token->attr['id']) || isset($token->attr['name'])) {\r
+ return;\r
+ }\r
+ $token = $deleted + 1;\r
+ for ($b = 0, $c = count($this->inputZipper->front); $b < $c; $b++) {\r
+ $prev = $this->inputZipper->front[$b];\r
+ if ($prev instanceof HTMLPurifier_Token_Text && $prev->is_whitespace) {\r
+ continue;\r
+ }\r
+ break;\r
+ }\r
+ // This is safe because we removed the token that triggered this.\r
+ $this->rewindOffset($b+$deleted);\r
+ return;\r
+ }\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Injector that removes spans with no attributes\r
+ */\r
+class HTMLPurifier_Injector_RemoveSpansWithoutAttributes extends HTMLPurifier_Injector\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'RemoveSpansWithoutAttributes';\r
+\r
+ /**\r
+ * @type array\r
+ */\r
+ public $needed = array('span');\r
+\r
+ /**\r
+ * @type HTMLPurifier_AttrValidator\r
+ */\r
+ private $attrValidator;\r
+\r
+ /**\r
+ * Used by AttrValidator.\r
+ * @type HTMLPurifier_Config\r
+ */\r
+ private $config;\r
+\r
+ /**\r
+ * @type HTMLPurifier_Context\r
+ */\r
+ private $context;\r
+\r
+ public function prepare($config, $context)\r
+ {\r
+ $this->attrValidator = new HTMLPurifier_AttrValidator();\r
+ $this->config = $config;\r
+ $this->context = $context;\r
+ return parent::prepare($config, $context);\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Token $token\r
+ */\r
+ public function handleElement(&$token)\r
+ {\r
+ if ($token->name !== 'span' || !$token instanceof HTMLPurifier_Token_Start) {\r
+ return;\r
+ }\r
+\r
+ // We need to validate the attributes now since this doesn't normally\r
+ // happen until after MakeWellFormed. If all the attributes are removed\r
+ // the span needs to be removed too.\r
+ $this->attrValidator->validateToken($token, $this->config, $this->context);\r
+ $token->armor['ValidateAttributes'] = true;\r
+\r
+ if (!empty($token->attr)) {\r
+ return;\r
+ }\r
+\r
+ $nesting = 0;\r
+ while ($this->forwardUntilEndToken($i, $current, $nesting)) {\r
+ }\r
+\r
+ if ($current instanceof HTMLPurifier_Token_End && $current->name === 'span') {\r
+ // Mark closing span tag for deletion\r
+ $current->markForDeletion = true;\r
+ // Delete open span tag\r
+ $token = false;\r
+ }\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Token $token\r
+ */\r
+ public function handleEnd(&$token)\r
+ {\r
+ if ($token->markForDeletion) {\r
+ $token = false;\r
+ }\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Adds important param elements to inside of object in order to make\r
+ * things safe.\r
+ */\r
+class HTMLPurifier_Injector_SafeObject extends HTMLPurifier_Injector\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'SafeObject';\r
+\r
+ /**\r
+ * @type array\r
+ */\r
+ public $needed = array('object', 'param');\r
+\r
+ /**\r
+ * @type array\r
+ */\r
+ protected $objectStack = array();\r
+\r
+ /**\r
+ * @type array\r
+ */\r
+ protected $paramStack = array();\r
+\r
+ /**\r
+ * Keep this synchronized with AttrTransform/SafeParam.php.\r
+ * @type array\r
+ */\r
+ protected $addParam = array(\r
+ 'allowScriptAccess' => 'never',\r
+ 'allowNetworking' => 'internal',\r
+ );\r
+\r
+ /**\r
+ * These are all lower-case keys.\r
+ * @type array\r
+ */\r
+ protected $allowedParam = array(\r
+ 'wmode' => true,\r
+ 'movie' => true,\r
+ 'flashvars' => true,\r
+ 'src' => true,\r
+ 'allowfullscreen' => true, // if omitted, assume to be 'false'\r
+ );\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return void\r
+ */\r
+ public function prepare($config, $context)\r
+ {\r
+ parent::prepare($config, $context);\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Token $token\r
+ */\r
+ public function handleElement(&$token)\r
+ {\r
+ if ($token->name == 'object') {\r
+ $this->objectStack[] = $token;\r
+ $this->paramStack[] = array();\r
+ $new = array($token);\r
+ foreach ($this->addParam as $name => $value) {\r
+ $new[] = new HTMLPurifier_Token_Empty('param', array('name' => $name, 'value' => $value));\r
+ }\r
+ $token = $new;\r
+ } elseif ($token->name == 'param') {\r
+ $nest = count($this->currentNesting) - 1;\r
+ if ($nest >= 0 && $this->currentNesting[$nest]->name === 'object') {\r
+ $i = count($this->objectStack) - 1;\r
+ if (!isset($token->attr['name'])) {\r
+ $token = false;\r
+ return;\r
+ }\r
+ $n = $token->attr['name'];\r
+ // We need this fix because YouTube doesn't supply a data\r
+ // attribute, which we need if a type is specified. This is\r
+ // *very* Flash specific.\r
+ if (!isset($this->objectStack[$i]->attr['data']) &&\r
+ ($token->attr['name'] == 'movie' || $token->attr['name'] == 'src')\r
+ ) {\r
+ $this->objectStack[$i]->attr['data'] = $token->attr['value'];\r
+ }\r
+ // Check if the parameter is the correct value but has not\r
+ // already been added\r
+ if (!isset($this->paramStack[$i][$n]) &&\r
+ isset($this->addParam[$n]) &&\r
+ $token->attr['name'] === $this->addParam[$n]) {\r
+ // keep token, and add to param stack\r
+ $this->paramStack[$i][$n] = true;\r
+ } elseif (isset($this->allowedParam[strtolower($n)])) {\r
+ // keep token, don't do anything to it\r
+ // (could possibly check for duplicates here)\r
+ // Note: In principle, parameters should be case sensitive.\r
+ // But it seems they are not really; so accept any case.\r
+ } else {\r
+ $token = false;\r
+ }\r
+ } else {\r
+ // not directly inside an object, DENY!\r
+ $token = false;\r
+ }\r
+ }\r
+ }\r
+\r
+ public function handleEnd(&$token)\r
+ {\r
+ // This is the WRONG way of handling the object and param stacks;\r
+ // we should be inserting them directly on the relevant object tokens\r
+ // so that the global stack handling handles it.\r
+ if ($token->name == 'object') {\r
+ array_pop($this->objectStack);\r
+ array_pop($this->paramStack);\r
+ }\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Parser that uses PHP 5's DOM extension (part of the core).\r
+ *\r
+ * In PHP 5, the DOM XML extension was revamped into DOM and added to the core.\r
+ * It gives us a forgiving HTML parser, which we use to transform the HTML\r
+ * into a DOM, and then into the tokens. It is blazingly fast (for large\r
+ * documents, it performs twenty times faster than\r
+ * HTMLPurifier_Lexer_DirectLex,and is the default choice for PHP 5.\r
+ *\r
+ * @note Any empty elements will have empty tokens associated with them, even if\r
+ * this is prohibited by the spec. This is cannot be fixed until the spec\r
+ * comes into play.\r
+ *\r
+ * @note PHP's DOM extension does not actually parse any entities, we use\r
+ * our own function to do that.\r
+ *\r
+ * @warning DOM tends to drop whitespace, which may wreak havoc on indenting.\r
+ * If this is a huge problem, due to the fact that HTML is hand\r
+ * edited and you are unable to get a parser cache that caches the\r
+ * the output of HTML Purifier while keeping the original HTML lying\r
+ * around, you may want to run Tidy on the resulting output or use\r
+ * HTMLPurifier_DirectLex\r
+ */\r
+\r
+class HTMLPurifier_Lexer_DOMLex extends HTMLPurifier_Lexer\r
+{\r
+\r
+ /**\r
+ * @type HTMLPurifier_TokenFactory\r
+ */\r
+ private $factory;\r
+\r
+ public function __construct()\r
+ {\r
+ // setup the factory\r
+ parent::__construct();\r
+ $this->factory = new HTMLPurifier_TokenFactory();\r
+ }\r
+\r
+ /**\r
+ * @param string $html\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return HTMLPurifier_Token[]\r
+ */\r
+ public function tokenizeHTML($html, $config, $context)\r
+ {\r
+ $html = $this->normalize($html, $config, $context);\r
+\r
+ // attempt to armor stray angled brackets that cannot possibly\r
+ // form tags and thus are probably being used as emoticons\r
+ if ($config->get('Core.AggressivelyFixLt')) {\r
+ $char = '[^a-z!\/]';\r
+ $comment = "/<!--(.*?)(-->|\z)/is";\r
+ $html = preg_replace_callback($comment, array($this, 'callbackArmorCommentEntities'), $html);\r
+ do {\r
+ $old = $html;\r
+ $html = preg_replace("/<($char)/i", '<\\1', $html);\r
+ } while ($html !== $old);\r
+ $html = preg_replace_callback($comment, array($this, 'callbackUndoCommentSubst'), $html); // fix comments\r
+ }\r
+\r
+ // preprocess html, essential for UTF-8\r
+ $html = $this->wrapHTML($html, $config, $context);\r
+\r
+ $doc = new DOMDocument();\r
+ $doc->encoding = 'UTF-8'; // theoretically, the above has this covered\r
+\r
+ $options = 0;\r
+ if ($config->get('Core.AllowParseManyTags') && defined('LIBXML_PARSEHUGE')) {\r
+ $options |= LIBXML_PARSEHUGE;\r
+ }\r
+\r
+ set_error_handler(array($this, 'muteErrorHandler'));\r
+ $doc->loadHTML($html, $options);\r
+ restore_error_handler();\r
+\r
+ $body = $doc->getElementsByTagName('html')->item(0)-> // <html>\r
+ getElementsByTagName('body')->item(0); // <body>\r
+\r
+ $div = $body->getElementsByTagName('div')->item(0); // <div>\r
+ $tokens = array();\r
+ $this->tokenizeDOM($div, $tokens, $config);\r
+ // If the div has a sibling, that means we tripped across\r
+ // a premature </div> tag. So remove the div we parsed,\r
+ // and then tokenize the rest of body. We can't tokenize\r
+ // the sibling directly as we'll lose the tags in that case.\r
+ if ($div->nextSibling) {\r
+ $body->removeChild($div);\r
+ $this->tokenizeDOM($body, $tokens, $config);\r
+ }\r
+ return $tokens;\r
+ }\r
+\r
+ /**\r
+ * Iterative function that tokenizes a node, putting it into an accumulator.\r
+ * To iterate is human, to recurse divine - L. Peter Deutsch\r
+ * @param DOMNode $node DOMNode to be tokenized.\r
+ * @param HTMLPurifier_Token[] $tokens Array-list of already tokenized tokens.\r
+ * @return HTMLPurifier_Token of node appended to previously passed tokens.\r
+ */\r
+ protected function tokenizeDOM($node, &$tokens, $config)\r
+ {\r
+ $level = 0;\r
+ $nodes = array($level => new HTMLPurifier_Queue(array($node)));\r
+ $closingNodes = array();\r
+ do {\r
+ while (!$nodes[$level]->isEmpty()) {\r
+ $node = $nodes[$level]->shift(); // FIFO\r
+ $collect = $level > 0 ? true : false;\r
+ $needEndingTag = $this->createStartNode($node, $tokens, $collect, $config);\r
+ if ($needEndingTag) {\r
+ $closingNodes[$level][] = $node;\r
+ }\r
+ if ($node->childNodes && $node->childNodes->length) {\r
+ $level++;\r
+ $nodes[$level] = new HTMLPurifier_Queue();\r
+ foreach ($node->childNodes as $childNode) {\r
+ $nodes[$level]->push($childNode);\r
+ }\r
+ }\r
+ }\r
+ $level--;\r
+ if ($level && isset($closingNodes[$level])) {\r
+ while ($node = array_pop($closingNodes[$level])) {\r
+ $this->createEndNode($node, $tokens);\r
+ }\r
+ }\r
+ } while ($level > 0);\r
+ }\r
+\r
+ /**\r
+ * Portably retrieve the tag name of a node; deals with older versions\r
+ * of libxml like 2.7.6\r
+ * @param DOMNode $node\r
+ */\r
+ protected function getTagName($node)\r
+ {\r
+ if (isset($node->tagName)) {\r
+ return $node->tagName;\r
+ } else if (isset($node->nodeName)) {\r
+ return $node->nodeName;\r
+ } else if (isset($node->localName)) {\r
+ return $node->localName;\r
+ }\r
+ return null;\r
+ }\r
+\r
+ /**\r
+ * Portably retrieve the data of a node; deals with older versions\r
+ * of libxml like 2.7.6\r
+ * @param DOMNode $node\r
+ */\r
+ protected function getData($node)\r
+ {\r
+ if (isset($node->data)) {\r
+ return $node->data;\r
+ } else if (isset($node->nodeValue)) {\r
+ return $node->nodeValue;\r
+ } else if (isset($node->textContent)) {\r
+ return $node->textContent;\r
+ }\r
+ return null;\r
+ }\r
+\r
+\r
+ /**\r
+ * @param DOMNode $node DOMNode to be tokenized.\r
+ * @param HTMLPurifier_Token[] $tokens Array-list of already tokenized tokens.\r
+ * @param bool $collect Says whether or start and close are collected, set to\r
+ * false at first recursion because it's the implicit DIV\r
+ * tag you're dealing with.\r
+ * @return bool if the token needs an endtoken\r
+ * @todo data and tagName properties don't seem to exist in DOMNode?\r
+ */\r
+ protected function createStartNode($node, &$tokens, $collect, $config)\r
+ {\r
+ // intercept non element nodes. WE MUST catch all of them,\r
+ // but we're not getting the character reference nodes because\r
+ // those should have been preprocessed\r
+ if ($node->nodeType === XML_TEXT_NODE) {\r
+ $data = $this->getData($node); // Handle variable data property\r
+ if ($data !== null) {\r
+ $tokens[] = $this->factory->createText($data);\r
+ }\r
+ return false;\r
+ } elseif ($node->nodeType === XML_CDATA_SECTION_NODE) {\r
+ // undo libxml's special treatment of <script> and <style> tags\r
+ $last = end($tokens);\r
+ $data = $node->data;\r
+ // (note $node->tagname is already normalized)\r
+ if ($last instanceof HTMLPurifier_Token_Start && ($last->name == 'script' || $last->name == 'style')) {\r
+ $new_data = trim($data);\r
+ if (substr($new_data, 0, 4) === '<!--') {\r
+ $data = substr($new_data, 4);\r
+ if (substr($data, -3) === '-->') {\r
+ $data = substr($data, 0, -3);\r
+ } else {\r
+ // Highly suspicious! Not sure what to do...\r
+ }\r
+ }\r
+ }\r
+ $tokens[] = $this->factory->createText($this->parseText($data, $config));\r
+ return false;\r
+ } elseif ($node->nodeType === XML_COMMENT_NODE) {\r
+ // this is code is only invoked for comments in script/style in versions\r
+ // of libxml pre-2.6.28 (regular comments, of course, are still\r
+ // handled regularly)\r
+ $tokens[] = $this->factory->createComment($node->data);\r
+ return false;\r
+ } elseif ($node->nodeType !== XML_ELEMENT_NODE) {\r
+ // not-well tested: there may be other nodes we have to grab\r
+ return false;\r
+ }\r
+ $attr = $node->hasAttributes() ? $this->transformAttrToAssoc($node->attributes) : array();\r
+ $tag_name = $this->getTagName($node); // Handle variable tagName property\r
+ if (empty($tag_name)) {\r
+ return (bool) $node->childNodes->length;\r
+ }\r
+ // We still have to make sure that the element actually IS empty\r
+ if (!$node->childNodes->length) {\r
+ if ($collect) {\r
+ $tokens[] = $this->factory->createEmpty($tag_name, $attr);\r
+ }\r
+ return false;\r
+ } else {\r
+ if ($collect) {\r
+ $tokens[] = $this->factory->createStart($tag_name, $attr);\r
+ }\r
+ return true;\r
+ }\r
+ }\r
+\r
+ /**\r
+ * @param DOMNode $node\r
+ * @param HTMLPurifier_Token[] $tokens\r
+ */\r
+ protected function createEndNode($node, &$tokens)\r
+ {\r
+ $tag_name = $this->getTagName($node); // Handle variable tagName property\r
+ $tokens[] = $this->factory->createEnd($tag_name);\r
+ }\r
+\r
+ /**\r
+ * Converts a DOMNamedNodeMap of DOMAttr objects into an assoc array.\r
+ *\r
+ * @param DOMNamedNodeMap $node_map DOMNamedNodeMap of DOMAttr objects.\r
+ * @return array Associative array of attributes.\r
+ */\r
+ protected function transformAttrToAssoc($node_map)\r
+ {\r
+ // NamedNodeMap is documented very well, so we're using undocumented\r
+ // features, namely, the fact that it implements Iterator and\r
+ // has a ->length attribute\r
+ if ($node_map->length === 0) {\r
+ return array();\r
+ }\r
+ $array = array();\r
+ foreach ($node_map as $attr) {\r
+ $array[$attr->name] = $attr->value;\r
+ }\r
+ return $array;\r
+ }\r
+\r
+ /**\r
+ * An error handler that mutes all errors\r
+ * @param int $errno\r
+ * @param string $errstr\r
+ */\r
+ public function muteErrorHandler($errno, $errstr)\r
+ {\r
+ }\r
+\r
+ /**\r
+ * Callback function for undoing escaping of stray angled brackets\r
+ * in comments\r
+ * @param array $matches\r
+ * @return string\r
+ */\r
+ public function callbackUndoCommentSubst($matches)\r
+ {\r
+ return '<!--' . strtr($matches[1], array('&' => '&', '<' => '<')) . $matches[2];\r
+ }\r
+\r
+ /**\r
+ * Callback function that entity-izes ampersands in comments so that\r
+ * callbackUndoCommentSubst doesn't clobber them\r
+ * @param array $matches\r
+ * @return string\r
+ */\r
+ public function callbackArmorCommentEntities($matches)\r
+ {\r
+ return '<!--' . str_replace('&', '&', $matches[1]) . $matches[2];\r
+ }\r
+\r
+ /**\r
+ * Wraps an HTML fragment in the necessary HTML\r
+ * @param string $html\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return string\r
+ */\r
+ protected function wrapHTML($html, $config, $context, $use_div = true)\r
+ {\r
+ $def = $config->getDefinition('HTML');\r
+ $ret = '';\r
+\r
+ if (!empty($def->doctype->dtdPublic) || !empty($def->doctype->dtdSystem)) {\r
+ $ret .= '<!DOCTYPE html ';\r
+ if (!empty($def->doctype->dtdPublic)) {\r
+ $ret .= 'PUBLIC "' . $def->doctype->dtdPublic . '" ';\r
+ }\r
+ if (!empty($def->doctype->dtdSystem)) {\r
+ $ret .= '"' . $def->doctype->dtdSystem . '" ';\r
+ }\r
+ $ret .= '>';\r
+ }\r
+\r
+ $ret .= '<html><head>';\r
+ $ret .= '<meta http-equiv="Content-Type" content="text/html; charset=utf-8" />';\r
+ // No protection if $html contains a stray </div>!\r
+ $ret .= '</head><body>';\r
+ if ($use_div) $ret .= '<div>';\r
+ $ret .= $html;\r
+ if ($use_div) $ret .= '</div>';\r
+ $ret .= '</body></html>';\r
+ return $ret;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Our in-house implementation of a parser.\r
+ *\r
+ * A pure PHP parser, DirectLex has absolutely no dependencies, making\r
+ * it a reasonably good default for PHP4. Written with efficiency in mind,\r
+ * it can be four times faster than HTMLPurifier_Lexer_PEARSax3, although it\r
+ * pales in comparison to HTMLPurifier_Lexer_DOMLex.\r
+ *\r
+ * @todo Reread XML spec and document differences.\r
+ */\r
+class HTMLPurifier_Lexer_DirectLex extends HTMLPurifier_Lexer\r
+{\r
+ /**\r
+ * @type bool\r
+ */\r
+ public $tracksLineNumbers = true;\r
+\r
+ /**\r
+ * Whitespace characters for str(c)spn.\r
+ * @type string\r
+ */\r
+ protected $_whitespace = "\x20\x09\x0D\x0A";\r
+\r
+ /**\r
+ * Callback function for script CDATA fudge\r
+ * @param array $matches, in form of array(opening tag, contents, closing tag)\r
+ * @return string\r
+ */\r
+ protected function scriptCallback($matches)\r
+ {\r
+ return $matches[1] . htmlspecialchars($matches[2], ENT_COMPAT, 'UTF-8') . $matches[3];\r
+ }\r
+\r
+ /**\r
+ * @param String $html\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return array|HTMLPurifier_Token[]\r
+ */\r
+ public function tokenizeHTML($html, $config, $context)\r
+ {\r
+ // special normalization for script tags without any armor\r
+ // our "armor" heurstic is a < sign any number of whitespaces after\r
+ // the first script tag\r
+ if ($config->get('HTML.Trusted')) {\r
+ $html = preg_replace_callback(\r
+ '#(<script[^>]*>)(\s*[^<].+?)(</script>)#si',\r
+ array($this, 'scriptCallback'),\r
+ $html\r
+ );\r
+ }\r
+\r
+ $html = $this->normalize($html, $config, $context);\r
+\r
+ $cursor = 0; // our location in the text\r
+ $inside_tag = false; // whether or not we're parsing the inside of a tag\r
+ $array = array(); // result array\r
+\r
+ // This is also treated to mean maintain *column* numbers too\r
+ $maintain_line_numbers = $config->get('Core.MaintainLineNumbers');\r
+\r
+ if ($maintain_line_numbers === null) {\r
+ // automatically determine line numbering by checking\r
+ // if error collection is on\r
+ $maintain_line_numbers = $config->get('Core.CollectErrors');\r
+ }\r
+\r
+ if ($maintain_line_numbers) {\r
+ $current_line = 1;\r
+ $current_col = 0;\r
+ $length = strlen($html);\r
+ } else {\r
+ $current_line = false;\r
+ $current_col = false;\r
+ $length = false;\r
+ }\r
+ $context->register('CurrentLine', $current_line);\r
+ $context->register('CurrentCol', $current_col);\r
+ $nl = "\n";\r
+ // how often to manually recalculate. This will ALWAYS be right,\r
+ // but it's pretty wasteful. Set to 0 to turn off\r
+ $synchronize_interval = $config->get('Core.DirectLexLineNumberSyncInterval');\r
+\r
+ $e = false;\r
+ if ($config->get('Core.CollectErrors')) {\r
+ $e =& $context->get('ErrorCollector');\r
+ }\r
+\r
+ // for testing synchronization\r
+ $loops = 0;\r
+\r
+ while (++$loops) {\r
+ // $cursor is either at the start of a token, or inside of\r
+ // a tag (i.e. there was a < immediately before it), as indicated\r
+ // by $inside_tag\r
+\r
+ if ($maintain_line_numbers) {\r
+ // $rcursor, however, is always at the start of a token.\r
+ $rcursor = $cursor - (int)$inside_tag;\r
+\r
+ // Column number is cheap, so we calculate it every round.\r
+ // We're interested at the *end* of the newline string, so\r
+ // we need to add strlen($nl) == 1 to $nl_pos before subtracting it\r
+ // from our "rcursor" position.\r
+ $nl_pos = strrpos($html, $nl, $rcursor - $length);\r
+ $current_col = $rcursor - (is_bool($nl_pos) ? 0 : $nl_pos + 1);\r
+\r
+ // recalculate lines\r
+ if ($synchronize_interval && // synchronization is on\r
+ $cursor > 0 && // cursor is further than zero\r
+ $loops % $synchronize_interval === 0) { // time to synchronize!\r
+ $current_line = 1 + $this->substrCount($html, $nl, 0, $cursor);\r
+ }\r
+ }\r
+\r
+ $position_next_lt = strpos($html, '<', $cursor);\r
+ $position_next_gt = strpos($html, '>', $cursor);\r
+\r
+ // triggers on "<b>asdf</b>" but not "asdf <b></b>"\r
+ // special case to set up context\r
+ if ($position_next_lt === $cursor) {\r
+ $inside_tag = true;\r
+ $cursor++;\r
+ }\r
+\r
+ if (!$inside_tag && $position_next_lt !== false) {\r
+ // We are not inside tag and there still is another tag to parse\r
+ $token = new\r
+ HTMLPurifier_Token_Text(\r
+ $this->parseText(\r
+ substr(\r
+ $html,\r
+ $cursor,\r
+ $position_next_lt - $cursor\r
+ ), $config\r
+ )\r
+ );\r
+ if ($maintain_line_numbers) {\r
+ $token->rawPosition($current_line, $current_col);\r
+ $current_line += $this->substrCount($html, $nl, $cursor, $position_next_lt - $cursor);\r
+ }\r
+ $array[] = $token;\r
+ $cursor = $position_next_lt + 1;\r
+ $inside_tag = true;\r
+ continue;\r
+ } elseif (!$inside_tag) {\r
+ // We are not inside tag but there are no more tags\r
+ // If we're already at the end, break\r
+ if ($cursor === strlen($html)) {\r
+ break;\r
+ }\r
+ // Create Text of rest of string\r
+ $token = new\r
+ HTMLPurifier_Token_Text(\r
+ $this->parseText(\r
+ substr(\r
+ $html,\r
+ $cursor\r
+ ), $config\r
+ )\r
+ );\r
+ if ($maintain_line_numbers) {\r
+ $token->rawPosition($current_line, $current_col);\r
+ }\r
+ $array[] = $token;\r
+ break;\r
+ } elseif ($inside_tag && $position_next_gt !== false) {\r
+ // We are in tag and it is well formed\r
+ // Grab the internals of the tag\r
+ $strlen_segment = $position_next_gt - $cursor;\r
+\r
+ if ($strlen_segment < 1) {\r
+ // there's nothing to process!\r
+ $token = new HTMLPurifier_Token_Text('<');\r
+ $cursor++;\r
+ continue;\r
+ }\r
+\r
+ $segment = substr($html, $cursor, $strlen_segment);\r
+\r
+ if ($segment === false) {\r
+ // somehow, we attempted to access beyond the end of\r
+ // the string, defense-in-depth, reported by Nate Abele\r
+ break;\r
+ }\r
+\r
+ // Check if it's a comment\r
+ if (substr($segment, 0, 3) === '!--') {\r
+ // re-determine segment length, looking for -->\r
+ $position_comment_end = strpos($html, '-->', $cursor);\r
+ if ($position_comment_end === false) {\r
+ // uh oh, we have a comment that extends to\r
+ // infinity. Can't be helped: set comment\r
+ // end position to end of string\r
+ if ($e) {\r
+ $e->send(E_WARNING, 'Lexer: Unclosed comment');\r
+ }\r
+ $position_comment_end = strlen($html);\r
+ $end = true;\r
+ } else {\r
+ $end = false;\r
+ }\r
+ $strlen_segment = $position_comment_end - $cursor;\r
+ $segment = substr($html, $cursor, $strlen_segment);\r
+ $token = new\r
+ HTMLPurifier_Token_Comment(\r
+ substr(\r
+ $segment,\r
+ 3,\r
+ $strlen_segment - 3\r
+ )\r
+ );\r
+ if ($maintain_line_numbers) {\r
+ $token->rawPosition($current_line, $current_col);\r
+ $current_line += $this->substrCount($html, $nl, $cursor, $strlen_segment);\r
+ }\r
+ $array[] = $token;\r
+ $cursor = $end ? $position_comment_end : $position_comment_end + 3;\r
+ $inside_tag = false;\r
+ continue;\r
+ }\r
+\r
+ // Check if it's an end tag\r
+ $is_end_tag = (strpos($segment, '/') === 0);\r
+ if ($is_end_tag) {\r
+ $type = substr($segment, 1);\r
+ $token = new HTMLPurifier_Token_End($type);\r
+ if ($maintain_line_numbers) {\r
+ $token->rawPosition($current_line, $current_col);\r
+ $current_line += $this->substrCount($html, $nl, $cursor, $position_next_gt - $cursor);\r
+ }\r
+ $array[] = $token;\r
+ $inside_tag = false;\r
+ $cursor = $position_next_gt + 1;\r
+ continue;\r
+ }\r
+\r
+ // Check leading character is alnum, if not, we may\r
+ // have accidently grabbed an emoticon. Translate into\r
+ // text and go our merry way\r
+ if (!ctype_alpha($segment[0])) {\r
+ // XML: $segment[0] !== '_' && $segment[0] !== ':'\r
+ if ($e) {\r
+ $e->send(E_NOTICE, 'Lexer: Unescaped lt');\r
+ }\r
+ $token = new HTMLPurifier_Token_Text('<');\r
+ if ($maintain_line_numbers) {\r
+ $token->rawPosition($current_line, $current_col);\r
+ $current_line += $this->substrCount($html, $nl, $cursor, $position_next_gt - $cursor);\r
+ }\r
+ $array[] = $token;\r
+ $inside_tag = false;\r
+ continue;\r
+ }\r
+\r
+ // Check if it is explicitly self closing, if so, remove\r
+ // trailing slash. Remember, we could have a tag like <br>, so\r
+ // any later token processing scripts must convert improperly\r
+ // classified EmptyTags from StartTags.\r
+ $is_self_closing = (strrpos($segment, '/') === $strlen_segment - 1);\r
+ if ($is_self_closing) {\r
+ $strlen_segment--;\r
+ $segment = substr($segment, 0, $strlen_segment);\r
+ }\r
+\r
+ // Check if there are any attributes\r
+ $position_first_space = strcspn($segment, $this->_whitespace);\r
+\r
+ if ($position_first_space >= $strlen_segment) {\r
+ if ($is_self_closing) {\r
+ $token = new HTMLPurifier_Token_Empty($segment);\r
+ } else {\r
+ $token = new HTMLPurifier_Token_Start($segment);\r
+ }\r
+ if ($maintain_line_numbers) {\r
+ $token->rawPosition($current_line, $current_col);\r
+ $current_line += $this->substrCount($html, $nl, $cursor, $position_next_gt - $cursor);\r
+ }\r
+ $array[] = $token;\r
+ $inside_tag = false;\r
+ $cursor = $position_next_gt + 1;\r
+ continue;\r
+ }\r
+\r
+ // Grab out all the data\r
+ $type = substr($segment, 0, $position_first_space);\r
+ $attribute_string =\r
+ trim(\r
+ substr(\r
+ $segment,\r
+ $position_first_space\r
+ )\r
+ );\r
+ if ($attribute_string) {\r
+ $attr = $this->parseAttributeString(\r
+ $attribute_string,\r
+ $config,\r
+ $context\r
+ );\r
+ } else {\r
+ $attr = array();\r
+ }\r
+\r
+ if ($is_self_closing) {\r
+ $token = new HTMLPurifier_Token_Empty($type, $attr);\r
+ } else {\r
+ $token = new HTMLPurifier_Token_Start($type, $attr);\r
+ }\r
+ if ($maintain_line_numbers) {\r
+ $token->rawPosition($current_line, $current_col);\r
+ $current_line += $this->substrCount($html, $nl, $cursor, $position_next_gt - $cursor);\r
+ }\r
+ $array[] = $token;\r
+ $cursor = $position_next_gt + 1;\r
+ $inside_tag = false;\r
+ continue;\r
+ } else {\r
+ // inside tag, but there's no ending > sign\r
+ if ($e) {\r
+ $e->send(E_WARNING, 'Lexer: Missing gt');\r
+ }\r
+ $token = new\r
+ HTMLPurifier_Token_Text(\r
+ '<' .\r
+ $this->parseText(\r
+ substr($html, $cursor), $config\r
+ )\r
+ );\r
+ if ($maintain_line_numbers) {\r
+ $token->rawPosition($current_line, $current_col);\r
+ }\r
+ // no cursor scroll? Hmm...\r
+ $array[] = $token;\r
+ break;\r
+ }\r
+ break;\r
+ }\r
+\r
+ $context->destroy('CurrentLine');\r
+ $context->destroy('CurrentCol');\r
+ return $array;\r
+ }\r
+\r
+ /**\r
+ * PHP 5.0.x compatible substr_count that implements offset and length\r
+ * @param string $haystack\r
+ * @param string $needle\r
+ * @param int $offset\r
+ * @param int $length\r
+ * @return int\r
+ */\r
+ protected function substrCount($haystack, $needle, $offset, $length)\r
+ {\r
+ static $oldVersion;\r
+ if ($oldVersion === null) {\r
+ $oldVersion = version_compare(PHP_VERSION, '5.1', '<');\r
+ }\r
+ if ($oldVersion) {\r
+ $haystack = substr($haystack, $offset, $length);\r
+ return substr_count($haystack, $needle);\r
+ } else {\r
+ return substr_count($haystack, $needle, $offset, $length);\r
+ }\r
+ }\r
+\r
+ /**\r
+ * Takes the inside of an HTML tag and makes an assoc array of attributes.\r
+ *\r
+ * @param string $string Inside of tag excluding name.\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return array Assoc array of attributes.\r
+ */\r
+ public function parseAttributeString($string, $config, $context)\r
+ {\r
+ $string = (string)$string; // quick typecast\r
+\r
+ if ($string == '') {\r
+ return array();\r
+ } // no attributes\r
+\r
+ $e = false;\r
+ if ($config->get('Core.CollectErrors')) {\r
+ $e =& $context->get('ErrorCollector');\r
+ }\r
+\r
+ // let's see if we can abort as quickly as possible\r
+ // one equal sign, no spaces => one attribute\r
+ $num_equal = substr_count($string, '=');\r
+ $has_space = strpos($string, ' ');\r
+ if ($num_equal === 0 && !$has_space) {\r
+ // bool attribute\r
+ return array($string => $string);\r
+ } elseif ($num_equal === 1 && !$has_space) {\r
+ // only one attribute\r
+ list($key, $quoted_value) = explode('=', $string);\r
+ $quoted_value = trim($quoted_value);\r
+ if (!$key) {\r
+ if ($e) {\r
+ $e->send(E_ERROR, 'Lexer: Missing attribute key');\r
+ }\r
+ return array();\r
+ }\r
+ if (!$quoted_value) {\r
+ return array($key => '');\r
+ }\r
+ $first_char = @$quoted_value[0];\r
+ $last_char = @$quoted_value[strlen($quoted_value) - 1];\r
+\r
+ $same_quote = ($first_char == $last_char);\r
+ $open_quote = ($first_char == '"' || $first_char == "'");\r
+\r
+ if ($same_quote && $open_quote) {\r
+ // well behaved\r
+ $value = substr($quoted_value, 1, strlen($quoted_value) - 2);\r
+ } else {\r
+ // not well behaved\r
+ if ($open_quote) {\r
+ if ($e) {\r
+ $e->send(E_ERROR, 'Lexer: Missing end quote');\r
+ }\r
+ $value = substr($quoted_value, 1);\r
+ } else {\r
+ $value = $quoted_value;\r
+ }\r
+ }\r
+ if ($value === false) {\r
+ $value = '';\r
+ }\r
+ return array($key => $this->parseAttr($value, $config));\r
+ }\r
+\r
+ // setup loop environment\r
+ $array = array(); // return assoc array of attributes\r
+ $cursor = 0; // current position in string (moves forward)\r
+ $size = strlen($string); // size of the string (stays the same)\r
+\r
+ // if we have unquoted attributes, the parser expects a terminating\r
+ // space, so let's guarantee that there's always a terminating space.\r
+ $string .= ' ';\r
+\r
+ $old_cursor = -1;\r
+ while ($cursor < $size) {\r
+ if ($old_cursor >= $cursor) {\r
+ throw new Exception("Infinite loop detected");\r
+ }\r
+ $old_cursor = $cursor;\r
+\r
+ $cursor += ($value = strspn($string, $this->_whitespace, $cursor));\r
+ // grab the key\r
+\r
+ $key_begin = $cursor; //we're currently at the start of the key\r
+\r
+ // scroll past all characters that are the key (not whitespace or =)\r
+ $cursor += strcspn($string, $this->_whitespace . '=', $cursor);\r
+\r
+ $key_end = $cursor; // now at the end of the key\r
+\r
+ $key = substr($string, $key_begin, $key_end - $key_begin);\r
+\r
+ if (!$key) {\r
+ if ($e) {\r
+ $e->send(E_ERROR, 'Lexer: Missing attribute key');\r
+ }\r
+ $cursor += 1 + strcspn($string, $this->_whitespace, $cursor + 1); // prevent infinite loop\r
+ continue; // empty key\r
+ }\r
+\r
+ // scroll past all whitespace\r
+ $cursor += strspn($string, $this->_whitespace, $cursor);\r
+\r
+ if ($cursor >= $size) {\r
+ $array[$key] = $key;\r
+ break;\r
+ }\r
+\r
+ // if the next character is an equal sign, we've got a regular\r
+ // pair, otherwise, it's a bool attribute\r
+ $first_char = @$string[$cursor];\r
+\r
+ if ($first_char == '=') {\r
+ // key="value"\r
+\r
+ $cursor++;\r
+ $cursor += strspn($string, $this->_whitespace, $cursor);\r
+\r
+ if ($cursor === false) {\r
+ $array[$key] = '';\r
+ break;\r
+ }\r
+\r
+ // we might be in front of a quote right now\r
+\r
+ $char = @$string[$cursor];\r
+\r
+ if ($char == '"' || $char == "'") {\r
+ // it's quoted, end bound is $char\r
+ $cursor++;\r
+ $value_begin = $cursor;\r
+ $cursor = strpos($string, $char, $cursor);\r
+ $value_end = $cursor;\r
+ } else {\r
+ // it's not quoted, end bound is whitespace\r
+ $value_begin = $cursor;\r
+ $cursor += strcspn($string, $this->_whitespace, $cursor);\r
+ $value_end = $cursor;\r
+ }\r
+\r
+ // we reached a premature end\r
+ if ($cursor === false) {\r
+ $cursor = $size;\r
+ $value_end = $cursor;\r
+ }\r
+\r
+ $value = substr($string, $value_begin, $value_end - $value_begin);\r
+ if ($value === false) {\r
+ $value = '';\r
+ }\r
+ $array[$key] = $this->parseAttr($value, $config);\r
+ $cursor++;\r
+ } else {\r
+ // boolattr\r
+ if ($key !== '') {\r
+ $array[$key] = $key;\r
+ } else {\r
+ // purely theoretical\r
+ if ($e) {\r
+ $e->send(E_ERROR, 'Lexer: Missing attribute key');\r
+ }\r
+ }\r
+ }\r
+ }\r
+ return $array;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Concrete comment node class.\r
+ */\r
+class HTMLPurifier_Node_Comment extends HTMLPurifier_Node\r
+{\r
+ /**\r
+ * Character data within comment.\r
+ * @type string\r
+ */\r
+ public $data;\r
+\r
+ /**\r
+ * @type bool\r
+ */\r
+ public $is_whitespace = true;\r
+\r
+ /**\r
+ * Transparent constructor.\r
+ *\r
+ * @param string $data String comment data.\r
+ * @param int $line\r
+ * @param int $col\r
+ */\r
+ public function __construct($data, $line = null, $col = null)\r
+ {\r
+ $this->data = $data;\r
+ $this->line = $line;\r
+ $this->col = $col;\r
+ }\r
+\r
+ public function toTokenPair() {\r
+ return array(new HTMLPurifier_Token_Comment($this->data, $this->line, $this->col), null);\r
+ }\r
+}\r
+
+\r
+\r
+/**\r
+ * Concrete element node class.\r
+ */\r
+class HTMLPurifier_Node_Element extends HTMLPurifier_Node\r
+{\r
+ /**\r
+ * The lower-case name of the tag, like 'a', 'b' or 'blockquote'.\r
+ *\r
+ * @note Strictly speaking, XML tags are case sensitive, so we shouldn't\r
+ * be lower-casing them, but these tokens cater to HTML tags, which are\r
+ * insensitive.\r
+ * @type string\r
+ */\r
+ public $name;\r
+\r
+ /**\r
+ * Associative array of the node's attributes.\r
+ * @type array\r
+ */\r
+ public $attr = array();\r
+\r
+ /**\r
+ * List of child elements.\r
+ * @type array\r
+ */\r
+ public $children = array();\r
+\r
+ /**\r
+ * Does this use the <a></a> form or the </a> form, i.e.\r
+ * is it a pair of start/end tokens or an empty token.\r
+ * @bool\r
+ */\r
+ public $empty = false;\r
+\r
+ public $endCol = null, $endLine = null, $endArmor = array();\r
+\r
+ public function __construct($name, $attr = array(), $line = null, $col = null, $armor = array()) {\r
+ $this->name = $name;\r
+ $this->attr = $attr;\r
+ $this->line = $line;\r
+ $this->col = $col;\r
+ $this->armor = $armor;\r
+ }\r
+\r
+ public function toTokenPair() {\r
+ // XXX inefficiency here, normalization is not necessary\r
+ if ($this->empty) {\r
+ return array(new HTMLPurifier_Token_Empty($this->name, $this->attr, $this->line, $this->col, $this->armor), null);\r
+ } else {\r
+ $start = new HTMLPurifier_Token_Start($this->name, $this->attr, $this->line, $this->col, $this->armor);\r
+ $end = new HTMLPurifier_Token_End($this->name, array(), $this->endLine, $this->endCol, $this->endArmor);\r
+ //$end->start = $start;\r
+ return array($start, $end);\r
+ }\r
+ }\r
+}\r
+\r
+
+\r
+\r
+/**\r
+ * Concrete text token class.\r
+ *\r
+ * Text tokens comprise of regular parsed character data (PCDATA) and raw\r
+ * character data (from the CDATA sections). Internally, their\r
+ * data is parsed with all entities expanded. Surprisingly, the text token\r
+ * does have a "tag name" called #PCDATA, which is how the DTD represents it\r
+ * in permissible child nodes.\r
+ */\r
+class HTMLPurifier_Node_Text extends HTMLPurifier_Node\r
+{\r
+\r
+ /**\r
+ * PCDATA tag name compatible with DTD, see\r
+ * HTMLPurifier_ChildDef_Custom for details.\r
+ * @type string\r
+ */\r
+ public $name = '#PCDATA';\r
+\r
+ /**\r
+ * @type string\r
+ */\r
+ public $data;\r
+ /**< Parsed character data of text. */\r
+\r
+ /**\r
+ * @type bool\r
+ */\r
+ public $is_whitespace;\r
+\r
+ /**< Bool indicating if node is whitespace. */\r
+\r
+ /**\r
+ * Constructor, accepts data and determines if it is whitespace.\r
+ * @param string $data String parsed character data.\r
+ * @param int $line\r
+ * @param int $col\r
+ */\r
+ public function __construct($data, $is_whitespace, $line = null, $col = null)\r
+ {\r
+ $this->data = $data;\r
+ $this->is_whitespace = $is_whitespace;\r
+ $this->line = $line;\r
+ $this->col = $col;\r
+ }\r
+\r
+ public function toTokenPair() {\r
+ return array(new HTMLPurifier_Token_Text($this->data, $this->line, $this->col), null);\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Composite strategy that runs multiple strategies on tokens.\r
+ */\r
+abstract class HTMLPurifier_Strategy_Composite extends HTMLPurifier_Strategy\r
+{\r
+\r
+ /**\r
+ * List of strategies to run tokens through.\r
+ * @type HTMLPurifier_Strategy[]\r
+ */\r
+ protected $strategies = array();\r
+\r
+ /**\r
+ * @param HTMLPurifier_Token[] $tokens\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return HTMLPurifier_Token[]\r
+ */\r
+ public function execute($tokens, $config, $context)\r
+ {\r
+ foreach ($this->strategies as $strategy) {\r
+ $tokens = $strategy->execute($tokens, $config, $context);\r
+ }\r
+ return $tokens;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Core strategy composed of the big four strategies.\r
+ */\r
+class HTMLPurifier_Strategy_Core extends HTMLPurifier_Strategy_Composite\r
+{\r
+ public function __construct()\r
+ {\r
+ $this->strategies[] = new HTMLPurifier_Strategy_RemoveForeignElements();\r
+ $this->strategies[] = new HTMLPurifier_Strategy_MakeWellFormed();\r
+ $this->strategies[] = new HTMLPurifier_Strategy_FixNesting();\r
+ $this->strategies[] = new HTMLPurifier_Strategy_ValidateAttributes();\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Takes a well formed list of tokens and fixes their nesting.\r
+ *\r
+ * HTML elements dictate which elements are allowed to be their children,\r
+ * for example, you can't have a p tag in a span tag. Other elements have\r
+ * much more rigorous definitions: tables, for instance, require a specific\r
+ * order for their elements. There are also constraints not expressible by\r
+ * document type definitions, such as the chameleon nature of ins/del\r
+ * tags and global child exclusions.\r
+ *\r
+ * The first major objective of this strategy is to iterate through all\r
+ * the nodes and determine whether or not their children conform to the\r
+ * element's definition. If they do not, the child definition may\r
+ * optionally supply an amended list of elements that is valid or\r
+ * require that the entire node be deleted (and the previous node\r
+ * rescanned).\r
+ *\r
+ * The second objective is to ensure that explicitly excluded elements of\r
+ * an element do not appear in its children. Code that accomplishes this\r
+ * task is pervasive through the strategy, though the two are distinct tasks\r
+ * and could, theoretically, be seperated (although it's not recommended).\r
+ *\r
+ * @note Whether or not unrecognized children are silently dropped or\r
+ * translated into text depends on the child definitions.\r
+ *\r
+ * @todo Enable nodes to be bubbled out of the structure. This is\r
+ * easier with our new algorithm.\r
+ */\r
+\r
+class HTMLPurifier_Strategy_FixNesting extends HTMLPurifier_Strategy\r
+{\r
+\r
+ /**\r
+ * @param HTMLPurifier_Token[] $tokens\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return array|HTMLPurifier_Token[]\r
+ */\r
+ public function execute($tokens, $config, $context)\r
+ {\r
+\r
+ //####################################################################//\r
+ // Pre-processing\r
+\r
+ // O(n) pass to convert to a tree, so that we can efficiently\r
+ // refer to substrings\r
+ $top_node = HTMLPurifier_Arborize::arborize($tokens, $config, $context);\r
+\r
+ // get a copy of the HTML definition\r
+ $definition = $config->getHTMLDefinition();\r
+\r
+ $excludes_enabled = !$config->get('Core.DisableExcludes');\r
+\r
+ // setup the context variable 'IsInline', for chameleon processing\r
+ // is 'false' when we are not inline, 'true' when it must always\r
+ // be inline, and an integer when it is inline for a certain\r
+ // branch of the document tree\r
+ $is_inline = $definition->info_parent_def->descendants_are_inline;\r
+ $context->register('IsInline', $is_inline);\r
+\r
+ // setup error collector\r
+ $e =& $context->get('ErrorCollector', true);\r
+\r
+ //####################################################################//\r
+ // Loop initialization\r
+\r
+ // stack that contains all elements that are excluded\r
+ // it is organized by parent elements, similar to $stack,\r
+ // but it is only populated when an element with exclusions is\r
+ // processed, i.e. there won't be empty exclusions.\r
+ $exclude_stack = array($definition->info_parent_def->excludes);\r
+\r
+ // variable that contains the start token while we are processing\r
+ // nodes. This enables error reporting to do its job\r
+ $node = $top_node;\r
+ // dummy token\r
+ list($token, $d) = $node->toTokenPair();\r
+ $context->register('CurrentNode', $node);\r
+ $context->register('CurrentToken', $token);\r
+\r
+ //####################################################################//\r
+ // Loop\r
+\r
+ // We need to implement a post-order traversal iteratively, to\r
+ // avoid running into stack space limits. This is pretty tricky\r
+ // to reason about, so we just manually stack-ify the recursive\r
+ // variant:\r
+ //\r
+ // function f($node) {\r
+ // foreach ($node->children as $child) {\r
+ // f($child);\r
+ // }\r
+ // validate($node);\r
+ // }\r
+ //\r
+ // Thus, we will represent a stack frame as array($node,\r
+ // $is_inline, stack of children)\r
+ // e.g. array_reverse($node->children) - already processed\r
+ // children.\r
+\r
+ $parent_def = $definition->info_parent_def;\r
+ $stack = array(\r
+ array($top_node,\r
+ $parent_def->descendants_are_inline,\r
+ $parent_def->excludes, // exclusions\r
+ 0)\r
+ );\r
+\r
+ while (!empty($stack)) {\r
+ list($node, $is_inline, $excludes, $ix) = array_pop($stack);\r
+ // recursive call\r
+ $go = false;\r
+ $def = empty($stack) ? $definition->info_parent_def : $definition->info[$node->name];\r
+ while (isset($node->children[$ix])) {\r
+ $child = $node->children[$ix++];\r
+ if ($child instanceof HTMLPurifier_Node_Element) {\r
+ $go = true;\r
+ $stack[] = array($node, $is_inline, $excludes, $ix);\r
+ $stack[] = array($child,\r
+ // ToDo: I don't think it matters if it's def or\r
+ // child_def, but double check this...\r
+ $is_inline || $def->descendants_are_inline,\r
+ empty($def->excludes) ? $excludes\r
+ : array_merge($excludes, $def->excludes),\r
+ 0);\r
+ break;\r
+ }\r
+ };\r
+ if ($go) continue;\r
+ list($token, $d) = $node->toTokenPair();\r
+ // base case\r
+ if ($excludes_enabled && isset($excludes[$node->name])) {\r
+ $node->dead = true;\r
+ if ($e) $e->send(E_ERROR, 'Strategy_FixNesting: Node excluded');\r
+ } else {\r
+ // XXX I suppose it would be slightly more efficient to\r
+ // avoid the allocation here and have children\r
+ // strategies handle it\r
+ $children = array();\r
+ foreach ($node->children as $child) {\r
+ if (!$child->dead) $children[] = $child;\r
+ }\r
+ $result = $def->child->validateChildren($children, $config, $context);\r
+ if ($result === true) {\r
+ // nop\r
+ $node->children = $children;\r
+ } elseif ($result === false) {\r
+ $node->dead = true;\r
+ if ($e) $e->send(E_ERROR, 'Strategy_FixNesting: Node removed');\r
+ } else {\r
+ $node->children = $result;\r
+ if ($e) {\r
+ // XXX This will miss mutations of internal nodes. Perhaps defer to the child validators\r
+ if (empty($result) && !empty($children)) {\r
+ $e->send(E_ERROR, 'Strategy_FixNesting: Node contents removed');\r
+ } else if ($result != $children) {\r
+ $e->send(E_WARNING, 'Strategy_FixNesting: Node reorganized');\r
+ }\r
+ }\r
+ }\r
+ }\r
+ }\r
+\r
+ //####################################################################//\r
+ // Post-processing\r
+\r
+ // remove context variables\r
+ $context->destroy('IsInline');\r
+ $context->destroy('CurrentNode');\r
+ $context->destroy('CurrentToken');\r
+\r
+ //####################################################################//\r
+ // Return\r
+\r
+ return HTMLPurifier_Arborize::flatten($node, $config, $context);\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Takes tokens makes them well-formed (balance end tags, etc.)\r
+ *\r
+ * Specification of the armor attributes this strategy uses:\r
+ *\r
+ * - MakeWellFormed_TagClosedError: This armor field is used to\r
+ * suppress tag closed errors for certain tokens [TagClosedSuppress],\r
+ * in particular, if a tag was generated automatically by HTML\r
+ * Purifier, we may rely on our infrastructure to close it for us\r
+ * and shouldn't report an error to the user [TagClosedAuto].\r
+ */\r
+class HTMLPurifier_Strategy_MakeWellFormed extends HTMLPurifier_Strategy\r
+{\r
+\r
+ /**\r
+ * Array stream of tokens being processed.\r
+ * @type HTMLPurifier_Token[]\r
+ */\r
+ protected $tokens;\r
+\r
+ /**\r
+ * Current token.\r
+ * @type HTMLPurifier_Token\r
+ */\r
+ protected $token;\r
+\r
+ /**\r
+ * Zipper managing the true state.\r
+ * @type HTMLPurifier_Zipper\r
+ */\r
+ protected $zipper;\r
+\r
+ /**\r
+ * Current nesting of elements.\r
+ * @type array\r
+ */\r
+ protected $stack;\r
+\r
+ /**\r
+ * Injectors active in this stream processing.\r
+ * @type HTMLPurifier_Injector[]\r
+ */\r
+ protected $injectors;\r
+\r
+ /**\r
+ * Current instance of HTMLPurifier_Config.\r
+ * @type HTMLPurifier_Config\r
+ */\r
+ protected $config;\r
+\r
+ /**\r
+ * Current instance of HTMLPurifier_Context.\r
+ * @type HTMLPurifier_Context\r
+ */\r
+ protected $context;\r
+\r
+ /**\r
+ * @param HTMLPurifier_Token[] $tokens\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return HTMLPurifier_Token[]\r
+ * @throws HTMLPurifier_Exception\r
+ */\r
+ public function execute($tokens, $config, $context)\r
+ {\r
+ $definition = $config->getHTMLDefinition();\r
+\r
+ // local variables\r
+ $generator = new HTMLPurifier_Generator($config, $context);\r
+ $escape_invalid_tags = $config->get('Core.EscapeInvalidTags');\r
+ // used for autoclose early abortion\r
+ $global_parent_allowed_elements = $definition->info_parent_def->child->getAllowedElements($config);\r
+ $e = $context->get('ErrorCollector', true);\r
+ $i = false; // injector index\r
+ list($zipper, $token) = HTMLPurifier_Zipper::fromArray($tokens);\r
+ if ($token === NULL) {\r
+ return array();\r
+ }\r
+ $reprocess = false; // whether or not to reprocess the same token\r
+ $stack = array();\r
+\r
+ // member variables\r
+ $this->stack =& $stack;\r
+ $this->tokens =& $tokens;\r
+ $this->token =& $token;\r
+ $this->zipper =& $zipper;\r
+ $this->config = $config;\r
+ $this->context = $context;\r
+\r
+ // context variables\r
+ $context->register('CurrentNesting', $stack);\r
+ $context->register('InputZipper', $zipper);\r
+ $context->register('CurrentToken', $token);\r
+\r
+ // -- begin INJECTOR --\r
+\r
+ $this->injectors = array();\r
+\r
+ $injectors = $config->getBatch('AutoFormat');\r
+ $def_injectors = $definition->info_injector;\r
+ $custom_injectors = $injectors['Custom'];\r
+ unset($injectors['Custom']); // special case\r
+ foreach ($injectors as $injector => $b) {\r
+ // XXX: Fix with a legitimate lookup table of enabled filters\r
+ if (strpos($injector, '.') !== false) {\r
+ continue;\r
+ }\r
+ $injector = "HTMLPurifier_Injector_$injector";\r
+ if (!$b) {\r
+ continue;\r
+ }\r
+ $this->injectors[] = new $injector;\r
+ }\r
+ foreach ($def_injectors as $injector) {\r
+ // assumed to be objects\r
+ $this->injectors[] = $injector;\r
+ }\r
+ foreach ($custom_injectors as $injector) {\r
+ if (!$injector) {\r
+ continue;\r
+ }\r
+ if (is_string($injector)) {\r
+ $injector = "HTMLPurifier_Injector_$injector";\r
+ $injector = new $injector;\r
+ }\r
+ $this->injectors[] = $injector;\r
+ }\r
+\r
+ // give the injectors references to the definition and context\r
+ // variables for performance reasons\r
+ foreach ($this->injectors as $ix => $injector) {\r
+ $error = $injector->prepare($config, $context);\r
+ if (!$error) {\r
+ continue;\r
+ }\r
+ array_splice($this->injectors, $ix, 1); // rm the injector\r
+ trigger_error("Cannot enable {$injector->name} injector because $error is not allowed", E_USER_WARNING);\r
+ }\r
+\r
+ // -- end INJECTOR --\r
+\r
+ // a note on reprocessing:\r
+ // In order to reduce code duplication, whenever some code needs\r
+ // to make HTML changes in order to make things "correct", the\r
+ // new HTML gets sent through the purifier, regardless of its\r
+ // status. This means that if we add a start token, because it\r
+ // was totally necessary, we don't have to update nesting; we just\r
+ // punt ($reprocess = true; continue;) and it does that for us.\r
+\r
+ // isset is in loop because $tokens size changes during loop exec\r
+ for (;;\r
+ // only increment if we don't need to reprocess\r
+ $reprocess ? $reprocess = false : $token = $zipper->next($token)) {\r
+\r
+ // check for a rewind\r
+ if (is_int($i)) {\r
+ // possibility: disable rewinding if the current token has a\r
+ // rewind set on it already. This would offer protection from\r
+ // infinite loop, but might hinder some advanced rewinding.\r
+ $rewind_offset = $this->injectors[$i]->getRewindOffset();\r
+ if (is_int($rewind_offset)) {\r
+ for ($j = 0; $j < $rewind_offset; $j++) {\r
+ if (empty($zipper->front)) break;\r
+ $token = $zipper->prev($token);\r
+ // indicate that other injectors should not process this token,\r
+ // but we need to reprocess it. See Note [Injector skips]\r
+ unset($token->skip[$i]);\r
+ $token->rewind = $i;\r
+ if ($token instanceof HTMLPurifier_Token_Start) {\r
+ array_pop($this->stack);\r
+ } elseif ($token instanceof HTMLPurifier_Token_End) {\r
+ $this->stack[] = $token->start;\r
+ }\r
+ }\r
+ }\r
+ $i = false;\r
+ }\r
+\r
+ // handle case of document end\r
+ if ($token === NULL) {\r
+ // kill processing if stack is empty\r
+ if (empty($this->stack)) {\r
+ break;\r
+ }\r
+\r
+ // peek\r
+ $top_nesting = array_pop($this->stack);\r
+ $this->stack[] = $top_nesting;\r
+\r
+ // send error [TagClosedSuppress]\r
+ if ($e && !isset($top_nesting->armor['MakeWellFormed_TagClosedError'])) {\r
+ $e->send(E_NOTICE, 'Strategy_MakeWellFormed: Tag closed by document end', $top_nesting);\r
+ }\r
+\r
+ // append, don't splice, since this is the end\r
+ $token = new HTMLPurifier_Token_End($top_nesting->name);\r
+\r
+ // punt!\r
+ $reprocess = true;\r
+ continue;\r
+ }\r
+\r
+ //echo '<br>'; printZipper($zipper, $token);//printTokens($this->stack);\r
+ //flush();\r
+\r
+ // quick-check: if it's not a tag, no need to process\r
+ if (empty($token->is_tag)) {\r
+ if ($token instanceof HTMLPurifier_Token_Text) {\r
+ foreach ($this->injectors as $i => $injector) {\r
+ if (isset($token->skip[$i])) {\r
+ // See Note [Injector skips]\r
+ continue;\r
+ }\r
+ if ($token->rewind !== null && $token->rewind !== $i) {\r
+ continue;\r
+ }\r
+ // XXX fuckup\r
+ $r = $token;\r
+ $injector->handleText($r);\r
+ $token = $this->processToken($r, $i);\r
+ $reprocess = true;\r
+ break;\r
+ }\r
+ }\r
+ // another possibility is a comment\r
+ continue;\r
+ }\r
+\r
+ if (isset($definition->info[$token->name])) {\r
+ $type = $definition->info[$token->name]->child->type;\r
+ } else {\r
+ $type = false; // Type is unknown, treat accordingly\r
+ }\r
+\r
+ // quick tag checks: anything that's *not* an end tag\r
+ $ok = false;\r
+ if ($type === 'empty' && $token instanceof HTMLPurifier_Token_Start) {\r
+ // claims to be a start tag but is empty\r
+ $token = new HTMLPurifier_Token_Empty(\r
+ $token->name,\r
+ $token->attr,\r
+ $token->line,\r
+ $token->col,\r
+ $token->armor\r
+ );\r
+ $ok = true;\r
+ } elseif ($type && $type !== 'empty' && $token instanceof HTMLPurifier_Token_Empty) {\r
+ // claims to be empty but really is a start tag\r
+ // NB: this assignment is required\r
+ $old_token = $token;\r
+ $token = new HTMLPurifier_Token_End($token->name);\r
+ $token = $this->insertBefore(\r
+ new HTMLPurifier_Token_Start($old_token->name, $old_token->attr, $old_token->line, $old_token->col, $old_token->armor)\r
+ );\r
+ // punt (since we had to modify the input stream in a non-trivial way)\r
+ $reprocess = true;\r
+ continue;\r
+ } elseif ($token instanceof HTMLPurifier_Token_Empty) {\r
+ // real empty token\r
+ $ok = true;\r
+ } elseif ($token instanceof HTMLPurifier_Token_Start) {\r
+ // start tag\r
+\r
+ // ...unless they also have to close their parent\r
+ if (!empty($this->stack)) {\r
+\r
+ // Performance note: you might think that it's rather\r
+ // inefficient, recalculating the autoclose information\r
+ // for every tag that a token closes (since when we\r
+ // do an autoclose, we push a new token into the\r
+ // stream and then /process/ that, before\r
+ // re-processing this token.) But this is\r
+ // necessary, because an injector can make an\r
+ // arbitrary transformations to the autoclosing\r
+ // tokens we introduce, so things may have changed\r
+ // in the meantime. Also, doing the inefficient thing is\r
+ // "easy" to reason about (for certain perverse definitions\r
+ // of "easy")\r
+\r
+ $parent = array_pop($this->stack);\r
+ $this->stack[] = $parent;\r
+\r
+ $parent_def = null;\r
+ $parent_elements = null;\r
+ $autoclose = false;\r
+ if (isset($definition->info[$parent->name])) {\r
+ $parent_def = $definition->info[$parent->name];\r
+ $parent_elements = $parent_def->child->getAllowedElements($config);\r
+ $autoclose = !isset($parent_elements[$token->name]);\r
+ }\r
+\r
+ if ($autoclose && $definition->info[$token->name]->wrap) {\r
+ // Check if an element can be wrapped by another\r
+ // element to make it valid in a context (for\r
+ // example, <ul><ul> needs a <li> in between)\r
+ $wrapname = $definition->info[$token->name]->wrap;\r
+ $wrapdef = $definition->info[$wrapname];\r
+ $elements = $wrapdef->child->getAllowedElements($config);\r
+ if (isset($elements[$token->name]) && isset($parent_elements[$wrapname])) {\r
+ $newtoken = new HTMLPurifier_Token_Start($wrapname);\r
+ $token = $this->insertBefore($newtoken);\r
+ $reprocess = true;\r
+ continue;\r
+ }\r
+ }\r
+\r
+ $carryover = false;\r
+ if ($autoclose && $parent_def->formatting) {\r
+ $carryover = true;\r
+ }\r
+\r
+ if ($autoclose) {\r
+ // check if this autoclose is doomed to fail\r
+ // (this rechecks $parent, which his harmless)\r
+ $autoclose_ok = isset($global_parent_allowed_elements[$token->name]);\r
+ if (!$autoclose_ok) {\r
+ foreach ($this->stack as $ancestor) {\r
+ $elements = $definition->info[$ancestor->name]->child->getAllowedElements($config);\r
+ if (isset($elements[$token->name])) {\r
+ $autoclose_ok = true;\r
+ break;\r
+ }\r
+ if ($definition->info[$token->name]->wrap) {\r
+ $wrapname = $definition->info[$token->name]->wrap;\r
+ $wrapdef = $definition->info[$wrapname];\r
+ $wrap_elements = $wrapdef->child->getAllowedElements($config);\r
+ if (isset($wrap_elements[$token->name]) && isset($elements[$wrapname])) {\r
+ $autoclose_ok = true;\r
+ break;\r
+ }\r
+ }\r
+ }\r
+ }\r
+ if ($autoclose_ok) {\r
+ // errors need to be updated\r
+ $new_token = new HTMLPurifier_Token_End($parent->name);\r
+ $new_token->start = $parent;\r
+ // [TagClosedSuppress]\r
+ if ($e && !isset($parent->armor['MakeWellFormed_TagClosedError'])) {\r
+ if (!$carryover) {\r
+ $e->send(E_NOTICE, 'Strategy_MakeWellFormed: Tag auto closed', $parent);\r
+ } else {\r
+ $e->send(E_NOTICE, 'Strategy_MakeWellFormed: Tag carryover', $parent);\r
+ }\r
+ }\r
+ if ($carryover) {\r
+ $element = clone $parent;\r
+ // [TagClosedAuto]\r
+ $element->armor['MakeWellFormed_TagClosedError'] = true;\r
+ $element->carryover = true;\r
+ $token = $this->processToken(array($new_token, $token, $element));\r
+ } else {\r
+ $token = $this->insertBefore($new_token);\r
+ }\r
+ } else {\r
+ $token = $this->remove();\r
+ }\r
+ $reprocess = true;\r
+ continue;\r
+ }\r
+\r
+ }\r
+ $ok = true;\r
+ }\r
+\r
+ if ($ok) {\r
+ foreach ($this->injectors as $i => $injector) {\r
+ if (isset($token->skip[$i])) {\r
+ // See Note [Injector skips]\r
+ continue;\r
+ }\r
+ if ($token->rewind !== null && $token->rewind !== $i) {\r
+ continue;\r
+ }\r
+ $r = $token;\r
+ $injector->handleElement($r);\r
+ $token = $this->processToken($r, $i);\r
+ $reprocess = true;\r
+ break;\r
+ }\r
+ if (!$reprocess) {\r
+ // ah, nothing interesting happened; do normal processing\r
+ if ($token instanceof HTMLPurifier_Token_Start) {\r
+ $this->stack[] = $token;\r
+ } elseif ($token instanceof HTMLPurifier_Token_End) {\r
+ throw new HTMLPurifier_Exception(\r
+ 'Improper handling of end tag in start code; possible error in MakeWellFormed'\r
+ );\r
+ }\r
+ }\r
+ continue;\r
+ }\r
+\r
+ // sanity check: we should be dealing with a closing tag\r
+ if (!$token instanceof HTMLPurifier_Token_End) {\r
+ throw new HTMLPurifier_Exception('Unaccounted for tag token in input stream, bug in HTML Purifier');\r
+ }\r
+\r
+ // make sure that we have something open\r
+ if (empty($this->stack)) {\r
+ if ($escape_invalid_tags) {\r
+ if ($e) {\r
+ $e->send(E_WARNING, 'Strategy_MakeWellFormed: Unnecessary end tag to text');\r
+ }\r
+ $token = new HTMLPurifier_Token_Text($generator->generateFromToken($token));\r
+ } else {\r
+ if ($e) {\r
+ $e->send(E_WARNING, 'Strategy_MakeWellFormed: Unnecessary end tag removed');\r
+ }\r
+ $token = $this->remove();\r
+ }\r
+ $reprocess = true;\r
+ continue;\r
+ }\r
+\r
+ // first, check for the simplest case: everything closes neatly.\r
+ // Eventually, everything passes through here; if there are problems\r
+ // we modify the input stream accordingly and then punt, so that\r
+ // the tokens get processed again.\r
+ $current_parent = array_pop($this->stack);\r
+ if ($current_parent->name == $token->name) {\r
+ $token->start = $current_parent;\r
+ foreach ($this->injectors as $i => $injector) {\r
+ if (isset($token->skip[$i])) {\r
+ // See Note [Injector skips]\r
+ continue;\r
+ }\r
+ if ($token->rewind !== null && $token->rewind !== $i) {\r
+ continue;\r
+ }\r
+ $r = $token;\r
+ $injector->handleEnd($r);\r
+ $token = $this->processToken($r, $i);\r
+ $this->stack[] = $current_parent;\r
+ $reprocess = true;\r
+ break;\r
+ }\r
+ continue;\r
+ }\r
+\r
+ // okay, so we're trying to close the wrong tag\r
+\r
+ // undo the pop previous pop\r
+ $this->stack[] = $current_parent;\r
+\r
+ // scroll back the entire nest, trying to find our tag.\r
+ // (feature could be to specify how far you'd like to go)\r
+ $size = count($this->stack);\r
+ // -2 because -1 is the last element, but we already checked that\r
+ $skipped_tags = false;\r
+ for ($j = $size - 2; $j >= 0; $j--) {\r
+ if ($this->stack[$j]->name == $token->name) {\r
+ $skipped_tags = array_slice($this->stack, $j);\r
+ break;\r
+ }\r
+ }\r
+\r
+ // we didn't find the tag, so remove\r
+ if ($skipped_tags === false) {\r
+ if ($escape_invalid_tags) {\r
+ if ($e) {\r
+ $e->send(E_WARNING, 'Strategy_MakeWellFormed: Stray end tag to text');\r
+ }\r
+ $token = new HTMLPurifier_Token_Text($generator->generateFromToken($token));\r
+ } else {\r
+ if ($e) {\r
+ $e->send(E_WARNING, 'Strategy_MakeWellFormed: Stray end tag removed');\r
+ }\r
+ $token = $this->remove();\r
+ }\r
+ $reprocess = true;\r
+ continue;\r
+ }\r
+\r
+ // do errors, in REVERSE $j order: a,b,c with </a></b></c>\r
+ $c = count($skipped_tags);\r
+ if ($e) {\r
+ for ($j = $c - 1; $j > 0; $j--) {\r
+ // notice we exclude $j == 0, i.e. the current ending tag, from\r
+ // the errors... [TagClosedSuppress]\r
+ if (!isset($skipped_tags[$j]->armor['MakeWellFormed_TagClosedError'])) {\r
+ $e->send(E_NOTICE, 'Strategy_MakeWellFormed: Tag closed by element end', $skipped_tags[$j]);\r
+ }\r
+ }\r
+ }\r
+\r
+ // insert tags, in FORWARD $j order: c,b,a with </a></b></c>\r
+ $replace = array($token);\r
+ for ($j = 1; $j < $c; $j++) {\r
+ // ...as well as from the insertions\r
+ $new_token = new HTMLPurifier_Token_End($skipped_tags[$j]->name);\r
+ $new_token->start = $skipped_tags[$j];\r
+ array_unshift($replace, $new_token);\r
+ if (isset($definition->info[$new_token->name]) && $definition->info[$new_token->name]->formatting) {\r
+ // [TagClosedAuto]\r
+ $element = clone $skipped_tags[$j];\r
+ $element->carryover = true;\r
+ $element->armor['MakeWellFormed_TagClosedError'] = true;\r
+ $replace[] = $element;\r
+ }\r
+ }\r
+ $token = $this->processToken($replace);\r
+ $reprocess = true;\r
+ continue;\r
+ }\r
+\r
+ $context->destroy('CurrentToken');\r
+ $context->destroy('CurrentNesting');\r
+ $context->destroy('InputZipper');\r
+\r
+ unset($this->injectors, $this->stack, $this->tokens);\r
+ return $zipper->toArray($token);\r
+ }\r
+\r
+ /**\r
+ * Processes arbitrary token values for complicated substitution patterns.\r
+ * In general:\r
+ *\r
+ * If $token is an array, it is a list of tokens to substitute for the\r
+ * current token. These tokens then get individually processed. If there\r
+ * is a leading integer in the list, that integer determines how many\r
+ * tokens from the stream should be removed.\r
+ *\r
+ * If $token is a regular token, it is swapped with the current token.\r
+ *\r
+ * If $token is false, the current token is deleted.\r
+ *\r
+ * If $token is an integer, that number of tokens (with the first token\r
+ * being the current one) will be deleted.\r
+ *\r
+ * @param HTMLPurifier_Token|array|int|bool $token Token substitution value\r
+ * @param HTMLPurifier_Injector|int $injector Injector that performed the substitution; default is if\r
+ * this is not an injector related operation.\r
+ * @throws HTMLPurifier_Exception\r
+ */\r
+ protected function processToken($token, $injector = -1)\r
+ {\r
+ // Zend OpCache miscompiles $token = array($token), so\r
+ // avoid this pattern. See: https://github.com/ezyang/htmlpurifier/issues/108\r
+\r
+ // normalize forms of token\r
+ if (is_object($token)) {\r
+ $tmp = $token;\r
+ $token = array(1, $tmp);\r
+ }\r
+ if (is_int($token)) {\r
+ $tmp = $token;\r
+ $token = array($tmp);\r
+ }\r
+ if ($token === false) {\r
+ $token = array(1);\r
+ }\r
+ if (!is_array($token)) {\r
+ throw new HTMLPurifier_Exception('Invalid token type from injector');\r
+ }\r
+ if (!is_int($token[0])) {\r
+ array_unshift($token, 1);\r
+ }\r
+ if ($token[0] === 0) {\r
+ throw new HTMLPurifier_Exception('Deleting zero tokens is not valid');\r
+ }\r
+\r
+ // $token is now an array with the following form:\r
+ // array(number nodes to delete, new node 1, new node 2, ...)\r
+\r
+ $delete = array_shift($token);\r
+ list($old, $r) = $this->zipper->splice($this->token, $delete, $token);\r
+\r
+ if ($injector > -1) {\r
+ // See Note [Injector skips]\r
+ // Determine appropriate skips. Here's what the code does:\r
+ // *If* we deleted one or more tokens, copy the skips\r
+ // of those tokens into the skips of the new tokens (in $token).\r
+ // Also, mark the newly inserted tokens as having come from\r
+ // $injector.\r
+ $oldskip = isset($old[0]) ? $old[0]->skip : array();\r
+ foreach ($token as $object) {\r
+ $object->skip = $oldskip;\r
+ $object->skip[$injector] = true;\r
+ }\r
+ }\r
+\r
+ return $r;\r
+\r
+ }\r
+\r
+ /**\r
+ * Inserts a token before the current token. Cursor now points to\r
+ * this token. You must reprocess after this.\r
+ * @param HTMLPurifier_Token $token\r
+ */\r
+ private function insertBefore($token)\r
+ {\r
+ // NB not $this->zipper->insertBefore(), due to positioning\r
+ // differences\r
+ $splice = $this->zipper->splice($this->token, 0, array($token));\r
+\r
+ return $splice[1];\r
+ }\r
+\r
+ /**\r
+ * Removes current token. Cursor now points to new token occupying previously\r
+ * occupied space. You must reprocess after this.\r
+ */\r
+ private function remove()\r
+ {\r
+ return $this->zipper->delete();\r
+ }\r
+}\r
+\r
+// Note [Injector skips]\r
+// ~~~~~~~~~~~~~~~~~~~~~\r
+// When I originally designed this class, the idea behind the 'skip'\r
+// property of HTMLPurifier_Token was to help avoid infinite loops\r
+// in injector processing. For example, suppose you wrote an injector\r
+// that bolded swear words. Naively, you might write it so that\r
+// whenever you saw ****, you replaced it with <strong>****</strong>.\r
+//\r
+// When this happens, we will reprocess all of the tokens with the\r
+// other injectors. Now there is an opportunity for infinite loop:\r
+// if we rerun the swear-word injector on these tokens, we might\r
+// see **** and then reprocess again to get\r
+// <strong><strong>****</strong></strong> ad infinitum.\r
+//\r
+// Thus, the idea of a skip is that once we process a token with\r
+// an injector, we mark all of those tokens as having "come from"\r
+// the injector, and we never run the injector again on these\r
+// tokens.\r
+//\r
+// There were two more complications, however:\r
+//\r
+// - With HTMLPurifier_Injector_RemoveEmpty, we noticed that if\r
+// you had <b><i></i></b>, after you removed the <i></i>, you\r
+// really would like this injector to go back and reprocess\r
+// the <b> tag, discovering that it is now empty and can be\r
+// removed. So we reintroduced the possibility of infinite looping\r
+// by adding a "rewind" function, which let you go back to an\r
+// earlier point in the token stream and reprocess it with injectors.\r
+// Needless to say, we need to UN-skip the token so it gets\r
+// reprocessed.\r
+//\r
+// - Suppose that you successfuly process a token, replace it with\r
+// one with your skip mark, but now another injector wants to\r
+// process the skipped token with another token. Should you continue\r
+// to skip that new token, or reprocess it? If you reprocess,\r
+// you can end up with an infinite loop where one injector converts\r
+// <a> to <b>, and then another injector converts it back. So\r
+// we inherit the skips, but for some reason, I thought that we\r
+// should inherit the skip from the first token of the token\r
+// that we deleted. Why? Well, it seems to work OK.\r
+//\r
+// If I were to redesign this functionality, I would absolutely not\r
+// go about doing it this way: the semantics are just not very well\r
+// defined, and in any case you probably wanted to operate on trees,\r
+// not token streams.\r
+\r
+
+
+\r
+\r
+/**\r
+ * Removes all unrecognized tags from the list of tokens.\r
+ *\r
+ * This strategy iterates through all the tokens and removes unrecognized\r
+ * tokens. If a token is not recognized but a TagTransform is defined for\r
+ * that element, the element will be transformed accordingly.\r
+ */\r
+\r
+class HTMLPurifier_Strategy_RemoveForeignElements extends HTMLPurifier_Strategy\r
+{\r
+\r
+ /**\r
+ * @param HTMLPurifier_Token[] $tokens\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return array|HTMLPurifier_Token[]\r
+ */\r
+ public function execute($tokens, $config, $context)\r
+ {\r
+ $definition = $config->getHTMLDefinition();\r
+ $generator = new HTMLPurifier_Generator($config, $context);\r
+ $result = array();\r
+\r
+ $escape_invalid_tags = $config->get('Core.EscapeInvalidTags');\r
+ $remove_invalid_img = $config->get('Core.RemoveInvalidImg');\r
+\r
+ // currently only used to determine if comments should be kept\r
+ $trusted = $config->get('HTML.Trusted');\r
+ $comment_lookup = $config->get('HTML.AllowedComments');\r
+ $comment_regexp = $config->get('HTML.AllowedCommentsRegexp');\r
+ $check_comments = $comment_lookup !== array() || $comment_regexp !== null;\r
+\r
+ $remove_script_contents = $config->get('Core.RemoveScriptContents');\r
+ $hidden_elements = $config->get('Core.HiddenElements');\r
+\r
+ // remove script contents compatibility\r
+ if ($remove_script_contents === true) {\r
+ $hidden_elements['script'] = true;\r
+ } elseif ($remove_script_contents === false && isset($hidden_elements['script'])) {\r
+ unset($hidden_elements['script']);\r
+ }\r
+\r
+ $attr_validator = new HTMLPurifier_AttrValidator();\r
+\r
+ // removes tokens until it reaches a closing tag with its value\r
+ $remove_until = false;\r
+\r
+ // converts comments into text tokens when this is equal to a tag name\r
+ $textify_comments = false;\r
+\r
+ $token = false;\r
+ $context->register('CurrentToken', $token);\r
+\r
+ $e = false;\r
+ if ($config->get('Core.CollectErrors')) {\r
+ $e =& $context->get('ErrorCollector');\r
+ }\r
+\r
+ foreach ($tokens as $token) {\r
+ if ($remove_until) {\r
+ if (empty($token->is_tag) || $token->name !== $remove_until) {\r
+ continue;\r
+ }\r
+ }\r
+ if (!empty($token->is_tag)) {\r
+ // DEFINITION CALL\r
+\r
+ // before any processing, try to transform the element\r
+ if (isset($definition->info_tag_transform[$token->name])) {\r
+ $original_name = $token->name;\r
+ // there is a transformation for this tag\r
+ // DEFINITION CALL\r
+ $token = $definition->\r
+ info_tag_transform[$token->name]->transform($token, $config, $context);\r
+ if ($e) {\r
+ $e->send(E_NOTICE, 'Strategy_RemoveForeignElements: Tag transform', $original_name);\r
+ }\r
+ }\r
+\r
+ if (isset($definition->info[$token->name])) {\r
+ // mostly everything's good, but\r
+ // we need to make sure required attributes are in order\r
+ if (($token instanceof HTMLPurifier_Token_Start || $token instanceof HTMLPurifier_Token_Empty) &&\r
+ $definition->info[$token->name]->required_attr &&\r
+ ($token->name != 'img' || $remove_invalid_img) // ensure config option still works\r
+ ) {\r
+ $attr_validator->validateToken($token, $config, $context);\r
+ $ok = true;\r
+ foreach ($definition->info[$token->name]->required_attr as $name) {\r
+ if (!isset($token->attr[$name])) {\r
+ $ok = false;\r
+ break;\r
+ }\r
+ }\r
+ if (!$ok) {\r
+ if ($e) {\r
+ $e->send(\r
+ E_ERROR,\r
+ 'Strategy_RemoveForeignElements: Missing required attribute',\r
+ $name\r
+ );\r
+ }\r
+ continue;\r
+ }\r
+ $token->armor['ValidateAttributes'] = true;\r
+ }\r
+\r
+ if (isset($hidden_elements[$token->name]) && $token instanceof HTMLPurifier_Token_Start) {\r
+ $textify_comments = $token->name;\r
+ } elseif ($token->name === $textify_comments && $token instanceof HTMLPurifier_Token_End) {\r
+ $textify_comments = false;\r
+ }\r
+\r
+ } elseif ($escape_invalid_tags) {\r
+ // invalid tag, generate HTML representation and insert in\r
+ if ($e) {\r
+ $e->send(E_WARNING, 'Strategy_RemoveForeignElements: Foreign element to text');\r
+ }\r
+ $token = new HTMLPurifier_Token_Text(\r
+ $generator->generateFromToken($token)\r
+ );\r
+ } else {\r
+ // check if we need to destroy all of the tag's children\r
+ // CAN BE GENERICIZED\r
+ if (isset($hidden_elements[$token->name])) {\r
+ if ($token instanceof HTMLPurifier_Token_Start) {\r
+ $remove_until = $token->name;\r
+ } elseif ($token instanceof HTMLPurifier_Token_Empty) {\r
+ // do nothing: we're still looking\r
+ } else {\r
+ $remove_until = false;\r
+ }\r
+ if ($e) {\r
+ $e->send(E_ERROR, 'Strategy_RemoveForeignElements: Foreign meta element removed');\r
+ }\r
+ } else {\r
+ if ($e) {\r
+ $e->send(E_ERROR, 'Strategy_RemoveForeignElements: Foreign element removed');\r
+ }\r
+ }\r
+ continue;\r
+ }\r
+ } elseif ($token instanceof HTMLPurifier_Token_Comment) {\r
+ // textify comments in script tags when they are allowed\r
+ if ($textify_comments !== false) {\r
+ $data = $token->data;\r
+ $token = new HTMLPurifier_Token_Text($data);\r
+ } elseif ($trusted || $check_comments) {\r
+ // always cleanup comments\r
+ $trailing_hyphen = false;\r
+ if ($e) {\r
+ // perform check whether or not there's a trailing hyphen\r
+ if (substr($token->data, -1) == '-') {\r
+ $trailing_hyphen = true;\r
+ }\r
+ }\r
+ $token->data = rtrim($token->data, '-');\r
+ $found_double_hyphen = false;\r
+ while (strpos($token->data, '--') !== false) {\r
+ $found_double_hyphen = true;\r
+ $token->data = str_replace('--', '-', $token->data);\r
+ }\r
+ if ($trusted || !empty($comment_lookup[trim($token->data)]) ||\r
+ ($comment_regexp !== null && preg_match($comment_regexp, trim($token->data)))) {\r
+ // OK good\r
+ if ($e) {\r
+ if ($trailing_hyphen) {\r
+ $e->send(\r
+ E_NOTICE,\r
+ 'Strategy_RemoveForeignElements: Trailing hyphen in comment removed'\r
+ );\r
+ }\r
+ if ($found_double_hyphen) {\r
+ $e->send(E_NOTICE, 'Strategy_RemoveForeignElements: Hyphens in comment collapsed');\r
+ }\r
+ }\r
+ } else {\r
+ if ($e) {\r
+ $e->send(E_NOTICE, 'Strategy_RemoveForeignElements: Comment removed');\r
+ }\r
+ continue;\r
+ }\r
+ } else {\r
+ // strip comments\r
+ if ($e) {\r
+ $e->send(E_NOTICE, 'Strategy_RemoveForeignElements: Comment removed');\r
+ }\r
+ continue;\r
+ }\r
+ } elseif ($token instanceof HTMLPurifier_Token_Text) {\r
+ } else {\r
+ continue;\r
+ }\r
+ $result[] = $token;\r
+ }\r
+ if ($remove_until && $e) {\r
+ // we removed tokens until the end, throw error\r
+ $e->send(E_ERROR, 'Strategy_RemoveForeignElements: Token removed to end', $remove_until);\r
+ }\r
+ $context->destroy('CurrentToken');\r
+ return $result;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Validate all attributes in the tokens.\r
+ */\r
+\r
+class HTMLPurifier_Strategy_ValidateAttributes extends HTMLPurifier_Strategy\r
+{\r
+\r
+ /**\r
+ * @param HTMLPurifier_Token[] $tokens\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return HTMLPurifier_Token[]\r
+ */\r
+ public function execute($tokens, $config, $context)\r
+ {\r
+ // setup validator\r
+ $validator = new HTMLPurifier_AttrValidator();\r
+\r
+ $token = false;\r
+ $context->register('CurrentToken', $token);\r
+\r
+ foreach ($tokens as $key => $token) {\r
+\r
+ // only process tokens that have attributes,\r
+ // namely start and empty tags\r
+ if (!$token instanceof HTMLPurifier_Token_Start && !$token instanceof HTMLPurifier_Token_Empty) {\r
+ continue;\r
+ }\r
+\r
+ // skip tokens that are armored\r
+ if (!empty($token->armor['ValidateAttributes'])) {\r
+ continue;\r
+ }\r
+\r
+ // note that we have no facilities here for removing tokens\r
+ $validator->validateToken($token, $config, $context);\r
+ }\r
+ $context->destroy('CurrentToken');\r
+ return $tokens;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Transforms FONT tags to the proper form (SPAN with CSS styling)\r
+ *\r
+ * This transformation takes the three proprietary attributes of FONT and\r
+ * transforms them into their corresponding CSS attributes. These are color,\r
+ * face, and size.\r
+ *\r
+ * @note Size is an interesting case because it doesn't map cleanly to CSS.\r
+ * Thanks to\r
+ * http://style.cleverchimp.com/font_size_intervals/altintervals.html\r
+ * for reasonable mappings.\r
+ * @warning This doesn't work completely correctly; specifically, this\r
+ * TagTransform operates before well-formedness is enforced, so\r
+ * the "active formatting elements" algorithm doesn't get applied.\r
+ */\r
+class HTMLPurifier_TagTransform_Font extends HTMLPurifier_TagTransform\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $transform_to = 'span';\r
+\r
+ /**\r
+ * @type array\r
+ */\r
+ protected $_size_lookup = array(\r
+ '0' => 'xx-small',\r
+ '1' => 'xx-small',\r
+ '2' => 'small',\r
+ '3' => 'medium',\r
+ '4' => 'large',\r
+ '5' => 'x-large',\r
+ '6' => 'xx-large',\r
+ '7' => '300%',\r
+ '-1' => 'smaller',\r
+ '-2' => '60%',\r
+ '+1' => 'larger',\r
+ '+2' => '150%',\r
+ '+3' => '200%',\r
+ '+4' => '300%'\r
+ );\r
+\r
+ /**\r
+ * @param HTMLPurifier_Token_Tag $tag\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return HTMLPurifier_Token_End|string\r
+ */\r
+ public function transform($tag, $config, $context)\r
+ {\r
+ if ($tag instanceof HTMLPurifier_Token_End) {\r
+ $new_tag = clone $tag;\r
+ $new_tag->name = $this->transform_to;\r
+ return $new_tag;\r
+ }\r
+\r
+ $attr = $tag->attr;\r
+ $prepend_style = '';\r
+\r
+ // handle color transform\r
+ if (isset($attr['color'])) {\r
+ $prepend_style .= 'color:' . $attr['color'] . ';';\r
+ unset($attr['color']);\r
+ }\r
+\r
+ // handle face transform\r
+ if (isset($attr['face'])) {\r
+ $prepend_style .= 'font-family:' . $attr['face'] . ';';\r
+ unset($attr['face']);\r
+ }\r
+\r
+ // handle size transform\r
+ if (isset($attr['size'])) {\r
+ // normalize large numbers\r
+ if ($attr['size'] !== '') {\r
+ if ($attr['size']{0} == '+' || $attr['size']{0} == '-') {\r
+ $size = (int)$attr['size'];\r
+ if ($size < -2) {\r
+ $attr['size'] = '-2';\r
+ }\r
+ if ($size > 4) {\r
+ $attr['size'] = '+4';\r
+ }\r
+ } else {\r
+ $size = (int)$attr['size'];\r
+ if ($size > 7) {\r
+ $attr['size'] = '7';\r
+ }\r
+ }\r
+ }\r
+ if (isset($this->_size_lookup[$attr['size']])) {\r
+ $prepend_style .= 'font-size:' .\r
+ $this->_size_lookup[$attr['size']] . ';';\r
+ }\r
+ unset($attr['size']);\r
+ }\r
+\r
+ if ($prepend_style) {\r
+ $attr['style'] = isset($attr['style']) ?\r
+ $prepend_style . $attr['style'] :\r
+ $prepend_style;\r
+ }\r
+\r
+ $new_tag = clone $tag;\r
+ $new_tag->name = $this->transform_to;\r
+ $new_tag->attr = $attr;\r
+\r
+ return $new_tag;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Simple transformation, just change tag name to something else,\r
+ * and possibly add some styling. This will cover most of the deprecated\r
+ * tag cases.\r
+ */\r
+class HTMLPurifier_TagTransform_Simple extends HTMLPurifier_TagTransform\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ protected $style;\r
+\r
+ /**\r
+ * @param string $transform_to Tag name to transform to.\r
+ * @param string $style CSS style to add to the tag\r
+ */\r
+ public function __construct($transform_to, $style = null)\r
+ {\r
+ $this->transform_to = $transform_to;\r
+ $this->style = $style;\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_Token_Tag $tag\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return string\r
+ */\r
+ public function transform($tag, $config, $context)\r
+ {\r
+ $new_tag = clone $tag;\r
+ $new_tag->name = $this->transform_to;\r
+ if (!is_null($this->style) &&\r
+ ($new_tag instanceof HTMLPurifier_Token_Start || $new_tag instanceof HTMLPurifier_Token_Empty)\r
+ ) {\r
+ $this->prependCSS($new_tag->attr, $this->style);\r
+ }\r
+ return $new_tag;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Concrete comment token class. Generally will be ignored.\r
+ */\r
+class HTMLPurifier_Token_Comment extends HTMLPurifier_Token\r
+{\r
+ /**\r
+ * Character data within comment.\r
+ * @type string\r
+ */\r
+ public $data;\r
+\r
+ /**\r
+ * @type bool\r
+ */\r
+ public $is_whitespace = true;\r
+\r
+ /**\r
+ * Transparent constructor.\r
+ *\r
+ * @param string $data String comment data.\r
+ * @param int $line\r
+ * @param int $col\r
+ */\r
+ public function __construct($data, $line = null, $col = null)\r
+ {\r
+ $this->data = $data;\r
+ $this->line = $line;\r
+ $this->col = $col;\r
+ }\r
+\r
+ public function toNode() {\r
+ return new HTMLPurifier_Node_Comment($this->data, $this->line, $this->col);\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Abstract class of a tag token (start, end or empty), and its behavior.\r
+ */\r
+abstract class HTMLPurifier_Token_Tag extends HTMLPurifier_Token\r
+{\r
+ /**\r
+ * Static bool marker that indicates the class is a tag.\r
+ *\r
+ * This allows us to check objects with <tt>!empty($obj->is_tag)</tt>\r
+ * without having to use a function call <tt>is_a()</tt>.\r
+ * @type bool\r
+ */\r
+ public $is_tag = true;\r
+\r
+ /**\r
+ * The lower-case name of the tag, like 'a', 'b' or 'blockquote'.\r
+ *\r
+ * @note Strictly speaking, XML tags are case sensitive, so we shouldn't\r
+ * be lower-casing them, but these tokens cater to HTML tags, which are\r
+ * insensitive.\r
+ * @type string\r
+ */\r
+ public $name;\r
+\r
+ /**\r
+ * Associative array of the tag's attributes.\r
+ * @type array\r
+ */\r
+ public $attr = array();\r
+\r
+ /**\r
+ * Non-overloaded constructor, which lower-cases passed tag name.\r
+ *\r
+ * @param string $name String name.\r
+ * @param array $attr Associative array of attributes.\r
+ * @param int $line\r
+ * @param int $col\r
+ * @param array $armor\r
+ */\r
+ public function __construct($name, $attr = array(), $line = null, $col = null, $armor = array())\r
+ {\r
+ $this->name = ctype_lower($name) ? $name : strtolower($name);\r
+ foreach ($attr as $key => $value) {\r
+ // normalization only necessary when key is not lowercase\r
+ if (!ctype_lower($key)) {\r
+ $new_key = strtolower($key);\r
+ if (!isset($attr[$new_key])) {\r
+ $attr[$new_key] = $attr[$key];\r
+ }\r
+ if ($new_key !== $key) {\r
+ unset($attr[$key]);\r
+ }\r
+ }\r
+ }\r
+ $this->attr = $attr;\r
+ $this->line = $line;\r
+ $this->col = $col;\r
+ $this->armor = $armor;\r
+ }\r
+\r
+ public function toNode() {\r
+ return new HTMLPurifier_Node_Element($this->name, $this->attr, $this->line, $this->col, $this->armor);\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Concrete empty token class.\r
+ */\r
+class HTMLPurifier_Token_Empty extends HTMLPurifier_Token_Tag\r
+{\r
+ public function toNode() {\r
+ $n = parent::toNode();\r
+ $n->empty = true;\r
+ return $n;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Concrete end token class.\r
+ *\r
+ * @warning This class accepts attributes even though end tags cannot. This\r
+ * is for optimization reasons, as under normal circumstances, the Lexers\r
+ * do not pass attributes.\r
+ */\r
+class HTMLPurifier_Token_End extends HTMLPurifier_Token_Tag\r
+{\r
+ /**\r
+ * Token that started this node.\r
+ * Added by MakeWellFormed. Please do not edit this!\r
+ * @type HTMLPurifier_Token\r
+ */\r
+ public $start;\r
+\r
+ public function toNode() {\r
+ throw new Exception("HTMLPurifier_Token_End->toNode not supported!");\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Concrete start token class.\r
+ */\r
+class HTMLPurifier_Token_Start extends HTMLPurifier_Token_Tag\r
+{\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Concrete text token class.\r
+ *\r
+ * Text tokens comprise of regular parsed character data (PCDATA) and raw\r
+ * character data (from the CDATA sections). Internally, their\r
+ * data is parsed with all entities expanded. Surprisingly, the text token\r
+ * does have a "tag name" called #PCDATA, which is how the DTD represents it\r
+ * in permissible child nodes.\r
+ */\r
+class HTMLPurifier_Token_Text extends HTMLPurifier_Token\r
+{\r
+\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = '#PCDATA';\r
+ /**< PCDATA tag name compatible with DTD. */\r
+\r
+ /**\r
+ * @type string\r
+ */\r
+ public $data;\r
+ /**< Parsed character data of text. */\r
+\r
+ /**\r
+ * @type bool\r
+ */\r
+ public $is_whitespace;\r
+\r
+ /**< Bool indicating if node is whitespace. */\r
+\r
+ /**\r
+ * Constructor, accepts data and determines if it is whitespace.\r
+ * @param string $data String parsed character data.\r
+ * @param int $line\r
+ * @param int $col\r
+ */\r
+ public function __construct($data, $line = null, $col = null)\r
+ {\r
+ $this->data = $data;\r
+ $this->is_whitespace = ctype_space($data);\r
+ $this->line = $line;\r
+ $this->col = $col;\r
+ }\r
+\r
+ public function toNode() {\r
+ return new HTMLPurifier_Node_Text($this->data, $this->is_whitespace, $this->line, $this->col);\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+class HTMLPurifier_URIFilter_DisableExternal extends HTMLPurifier_URIFilter\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'DisableExternal';\r
+\r
+ /**\r
+ * @type array\r
+ */\r
+ protected $ourHostParts = false;\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ * @return void\r
+ */\r
+ public function prepare($config)\r
+ {\r
+ $our_host = $config->getDefinition('URI')->host;\r
+ if ($our_host !== null) {\r
+ $this->ourHostParts = array_reverse(explode('.', $our_host));\r
+ }\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_URI $uri Reference\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool\r
+ */\r
+ public function filter(&$uri, $config, $context)\r
+ {\r
+ if (is_null($uri->host)) {\r
+ return true;\r
+ }\r
+ if ($this->ourHostParts === false) {\r
+ return false;\r
+ }\r
+ $host_parts = array_reverse(explode('.', $uri->host));\r
+ foreach ($this->ourHostParts as $i => $x) {\r
+ if (!isset($host_parts[$i])) {\r
+ return false;\r
+ }\r
+ if ($host_parts[$i] != $this->ourHostParts[$i]) {\r
+ return false;\r
+ }\r
+ }\r
+ return true;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+class HTMLPurifier_URIFilter_DisableExternalResources extends HTMLPurifier_URIFilter_DisableExternal\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'DisableExternalResources';\r
+\r
+ /**\r
+ * @param HTMLPurifier_URI $uri\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool\r
+ */\r
+ public function filter(&$uri, $config, $context)\r
+ {\r
+ if (!$context->get('EmbeddedURI', true)) {\r
+ return true;\r
+ }\r
+ return parent::filter($uri, $config, $context);\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+class HTMLPurifier_URIFilter_DisableResources extends HTMLPurifier_URIFilter\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'DisableResources';\r
+\r
+ /**\r
+ * @param HTMLPurifier_URI $uri\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool\r
+ */\r
+ public function filter(&$uri, $config, $context)\r
+ {\r
+ return !$context->get('EmbeddedURI', true);\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+// It's not clear to me whether or not Punycode means that hostnames\r
+// do not have canonical forms anymore. As far as I can tell, it's\r
+// not a problem (punycoding should be identity when no Unicode\r
+// points are involved), but I'm not 100% sure\r
+class HTMLPurifier_URIFilter_HostBlacklist extends HTMLPurifier_URIFilter\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'HostBlacklist';\r
+\r
+ /**\r
+ * @type array\r
+ */\r
+ protected $blacklist = array();\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ * @return bool\r
+ */\r
+ public function prepare($config)\r
+ {\r
+ $this->blacklist = $config->get('URI.HostBlacklist');\r
+ return true;\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_URI $uri\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool\r
+ */\r
+ public function filter(&$uri, $config, $context)\r
+ {\r
+ foreach ($this->blacklist as $blacklisted_host_fragment) {\r
+ if (strpos($uri->host, $blacklisted_host_fragment) !== false) {\r
+ return false;\r
+ }\r
+ }\r
+ return true;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+// does not support network paths\r
+\r
+class HTMLPurifier_URIFilter_MakeAbsolute extends HTMLPurifier_URIFilter\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'MakeAbsolute';\r
+\r
+ /**\r
+ * @type\r
+ */\r
+ protected $base;\r
+\r
+ /**\r
+ * @type array\r
+ */\r
+ protected $basePathStack = array();\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ * @return bool\r
+ */\r
+ public function prepare($config)\r
+ {\r
+ $def = $config->getDefinition('URI');\r
+ $this->base = $def->base;\r
+ if (is_null($this->base)) {\r
+ trigger_error(\r
+ 'URI.MakeAbsolute is being ignored due to lack of ' .\r
+ 'value for URI.Base configuration',\r
+ E_USER_WARNING\r
+ );\r
+ return false;\r
+ }\r
+ $this->base->fragment = null; // fragment is invalid for base URI\r
+ $stack = explode('/', $this->base->path);\r
+ array_pop($stack); // discard last segment\r
+ $stack = $this->_collapseStack($stack); // do pre-parsing\r
+ $this->basePathStack = $stack;\r
+ return true;\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_URI $uri\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool\r
+ */\r
+ public function filter(&$uri, $config, $context)\r
+ {\r
+ if (is_null($this->base)) {\r
+ return true;\r
+ } // abort early\r
+ if ($uri->path === '' && is_null($uri->scheme) &&\r
+ is_null($uri->host) && is_null($uri->query) && is_null($uri->fragment)) {\r
+ // reference to current document\r
+ $uri = clone $this->base;\r
+ return true;\r
+ }\r
+ if (!is_null($uri->scheme)) {\r
+ // absolute URI already: don't change\r
+ if (!is_null($uri->host)) {\r
+ return true;\r
+ }\r
+ $scheme_obj = $uri->getSchemeObj($config, $context);\r
+ if (!$scheme_obj) {\r
+ // scheme not recognized\r
+ return false;\r
+ }\r
+ if (!$scheme_obj->hierarchical) {\r
+ // non-hierarchal URI with explicit scheme, don't change\r
+ return true;\r
+ }\r
+ // special case: had a scheme but always is hierarchical and had no authority\r
+ }\r
+ if (!is_null($uri->host)) {\r
+ // network path, don't bother\r
+ return true;\r
+ }\r
+ if ($uri->path === '') {\r
+ $uri->path = $this->base->path;\r
+ } elseif ($uri->path[0] !== '/') {\r
+ // relative path, needs more complicated processing\r
+ $stack = explode('/', $uri->path);\r
+ $new_stack = array_merge($this->basePathStack, $stack);\r
+ if ($new_stack[0] !== '' && !is_null($this->base->host)) {\r
+ array_unshift($new_stack, '');\r
+ }\r
+ $new_stack = $this->_collapseStack($new_stack);\r
+ $uri->path = implode('/', $new_stack);\r
+ } else {\r
+ // absolute path, but still we should collapse\r
+ $uri->path = implode('/', $this->_collapseStack(explode('/', $uri->path)));\r
+ }\r
+ // re-combine\r
+ $uri->scheme = $this->base->scheme;\r
+ if (is_null($uri->userinfo)) {\r
+ $uri->userinfo = $this->base->userinfo;\r
+ }\r
+ if (is_null($uri->host)) {\r
+ $uri->host = $this->base->host;\r
+ }\r
+ if (is_null($uri->port)) {\r
+ $uri->port = $this->base->port;\r
+ }\r
+ return true;\r
+ }\r
+\r
+ /**\r
+ * Resolve dots and double-dots in a path stack\r
+ * @param array $stack\r
+ * @return array\r
+ */\r
+ private function _collapseStack($stack)\r
+ {\r
+ $result = array();\r
+ $is_folder = false;\r
+ for ($i = 0; isset($stack[$i]); $i++) {\r
+ $is_folder = false;\r
+ // absorb an internally duplicated slash\r
+ if ($stack[$i] == '' && $i && isset($stack[$i + 1])) {\r
+ continue;\r
+ }\r
+ if ($stack[$i] == '..') {\r
+ if (!empty($result)) {\r
+ $segment = array_pop($result);\r
+ if ($segment === '' && empty($result)) {\r
+ // error case: attempted to back out too far:\r
+ // restore the leading slash\r
+ $result[] = '';\r
+ } elseif ($segment === '..') {\r
+ $result[] = '..'; // cannot remove .. with ..\r
+ }\r
+ } else {\r
+ // relative path, preserve the double-dots\r
+ $result[] = '..';\r
+ }\r
+ $is_folder = true;\r
+ continue;\r
+ }\r
+ if ($stack[$i] == '.') {\r
+ // silently absorb\r
+ $is_folder = true;\r
+ continue;\r
+ }\r
+ $result[] = $stack[$i];\r
+ }\r
+ if ($is_folder) {\r
+ $result[] = '';\r
+ }\r
+ return $result;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+class HTMLPurifier_URIFilter_Munge extends HTMLPurifier_URIFilter\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'Munge';\r
+\r
+ /**\r
+ * @type bool\r
+ */\r
+ public $post = true;\r
+\r
+ /**\r
+ * @type string\r
+ */\r
+ private $target;\r
+\r
+ /**\r
+ * @type HTMLPurifier_URIParser\r
+ */\r
+ private $parser;\r
+\r
+ /**\r
+ * @type bool\r
+ */\r
+ private $doEmbed;\r
+\r
+ /**\r
+ * @type string\r
+ */\r
+ private $secretKey;\r
+\r
+ /**\r
+ * @type array\r
+ */\r
+ protected $replace = array();\r
+\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ * @return bool\r
+ */\r
+ public function prepare($config)\r
+ {\r
+ $this->target = $config->get('URI.' . $this->name);\r
+ $this->parser = new HTMLPurifier_URIParser();\r
+ $this->doEmbed = $config->get('URI.MungeResources');\r
+ $this->secretKey = $config->get('URI.MungeSecretKey');\r
+ if ($this->secretKey && !function_exists('hash_hmac')) {\r
+ throw new Exception("Cannot use %URI.MungeSecretKey without hash_hmac support.");\r
+ }\r
+ return true;\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_URI $uri\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool\r
+ */\r
+ public function filter(&$uri, $config, $context)\r
+ {\r
+ if ($context->get('EmbeddedURI', true) && !$this->doEmbed) {\r
+ return true;\r
+ }\r
+\r
+ $scheme_obj = $uri->getSchemeObj($config, $context);\r
+ if (!$scheme_obj) {\r
+ return true;\r
+ } // ignore unknown schemes, maybe another postfilter did it\r
+ if (!$scheme_obj->browsable) {\r
+ return true;\r
+ } // ignore non-browseable schemes, since we can't munge those in a reasonable way\r
+ if ($uri->isBenign($config, $context)) {\r
+ return true;\r
+ } // don't redirect if a benign URL\r
+\r
+ $this->makeReplace($uri, $config, $context);\r
+ $this->replace = array_map('rawurlencode', $this->replace);\r
+\r
+ $new_uri = strtr($this->target, $this->replace);\r
+ $new_uri = $this->parser->parse($new_uri);\r
+ // don't redirect if the target host is the same as the\r
+ // starting host\r
+ if ($uri->host === $new_uri->host) {\r
+ return true;\r
+ }\r
+ $uri = $new_uri; // overwrite\r
+ return true;\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_URI $uri\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ */\r
+ protected function makeReplace($uri, $config, $context)\r
+ {\r
+ $string = $uri->toString();\r
+ // always available\r
+ $this->replace['%s'] = $string;\r
+ $this->replace['%r'] = $context->get('EmbeddedURI', true);\r
+ $token = $context->get('CurrentToken', true);\r
+ $this->replace['%n'] = $token ? $token->name : null;\r
+ $this->replace['%m'] = $context->get('CurrentAttr', true);\r
+ $this->replace['%p'] = $context->get('CurrentCSSProperty', true);\r
+ // not always available\r
+ if ($this->secretKey) {\r
+ $this->replace['%t'] = hash_hmac("sha256", $string, $this->secretKey);\r
+ }\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Implements safety checks for safe iframes.\r
+ *\r
+ * @warning This filter is *critical* for ensuring that %HTML.SafeIframe\r
+ * works safely.\r
+ */\r
+class HTMLPurifier_URIFilter_SafeIframe extends HTMLPurifier_URIFilter\r
+{\r
+ /**\r
+ * @type string\r
+ */\r
+ public $name = 'SafeIframe';\r
+\r
+ /**\r
+ * @type bool\r
+ */\r
+ public $always_load = true;\r
+\r
+ /**\r
+ * @type string\r
+ */\r
+ protected $regexp = null;\r
+\r
+ // XXX: The not so good bit about how this is all set up now is we\r
+ // can't check HTML.SafeIframe in the 'prepare' step: we have to\r
+ // defer till the actual filtering.\r
+ /**\r
+ * @param HTMLPurifier_Config $config\r
+ * @return bool\r
+ */\r
+ public function prepare($config)\r
+ {\r
+ $this->regexp = $config->get('URI.SafeIframeRegexp');\r
+ return true;\r
+ }\r
+\r
+ /**\r
+ * @param HTMLPurifier_URI $uri\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool\r
+ */\r
+ public function filter(&$uri, $config, $context)\r
+ {\r
+ // check if filter not applicable\r
+ if (!$config->get('HTML.SafeIframe')) {\r
+ return true;\r
+ }\r
+ // check if the filter should actually trigger\r
+ if (!$context->get('EmbeddedURI', true)) {\r
+ return true;\r
+ }\r
+ $token = $context->get('CurrentToken', true);\r
+ if (!($token && $token->name == 'iframe')) {\r
+ return true;\r
+ }\r
+ // check if we actually have some whitelists enabled\r
+ if ($this->regexp === null) {\r
+ return false;\r
+ }\r
+ // actually check the whitelists\r
+ return preg_match($this->regexp, $uri->toString());\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Implements data: URI for base64 encoded images supported by GD.\r
+ */\r
+class HTMLPurifier_URIScheme_data extends HTMLPurifier_URIScheme\r
+{\r
+ /**\r
+ * @type bool\r
+ */\r
+ public $browsable = true;\r
+\r
+ /**\r
+ * @type array\r
+ */\r
+ public $allowed_types = array(\r
+ // you better write validation code for other types if you\r
+ // decide to allow them\r
+ 'image/jpeg' => true,\r
+ 'image/gif' => true,\r
+ 'image/png' => true,\r
+ );\r
+ // this is actually irrelevant since we only write out the path\r
+ // component\r
+ /**\r
+ * @type bool\r
+ */\r
+ public $may_omit_host = true;\r
+\r
+ /**\r
+ * @param HTMLPurifier_URI $uri\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool\r
+ */\r
+ public function doValidate(&$uri, $config, $context)\r
+ {\r
+ $result = explode(',', $uri->path, 2);\r
+ $is_base64 = false;\r
+ $charset = null;\r
+ $content_type = null;\r
+ if (count($result) == 2) {\r
+ list($metadata, $data) = $result;\r
+ // do some legwork on the metadata\r
+ $metas = explode(';', $metadata);\r
+ while (!empty($metas)) {\r
+ $cur = array_shift($metas);\r
+ if ($cur == 'base64') {\r
+ $is_base64 = true;\r
+ break;\r
+ }\r
+ if (substr($cur, 0, 8) == 'charset=') {\r
+ // doesn't match if there are arbitrary spaces, but\r
+ // whatever dude\r
+ if ($charset !== null) {\r
+ continue;\r
+ } // garbage\r
+ $charset = substr($cur, 8); // not used\r
+ } else {\r
+ if ($content_type !== null) {\r
+ continue;\r
+ } // garbage\r
+ $content_type = $cur;\r
+ }\r
+ }\r
+ } else {\r
+ $data = $result[0];\r
+ }\r
+ if ($content_type !== null && empty($this->allowed_types[$content_type])) {\r
+ return false;\r
+ }\r
+ if ($charset !== null) {\r
+ // error; we don't allow plaintext stuff\r
+ $charset = null;\r
+ }\r
+ $data = rawurldecode($data);\r
+ if ($is_base64) {\r
+ $raw_data = base64_decode($data);\r
+ } else {\r
+ $raw_data = $data;\r
+ }\r
+ if ( strlen($raw_data) < 12 ) {\r
+ // error; exif_imagetype throws exception with small files,\r
+ // and this likely indicates a corrupt URI/failed parse anyway\r
+ return false;\r
+ }\r
+ // XXX probably want to refactor this into a general mechanism\r
+ // for filtering arbitrary content types\r
+ if (function_exists('sys_get_temp_dir')) {\r
+ $file = tempnam(sys_get_temp_dir(), "");\r
+ } else {\r
+ $file = tempnam("/tmp", "");\r
+ }\r
+ file_put_contents($file, $raw_data);\r
+ if (function_exists('exif_imagetype')) {\r
+ $image_code = exif_imagetype($file);\r
+ unlink($file);\r
+ } elseif (function_exists('getimagesize')) {\r
+ set_error_handler(array($this, 'muteErrorHandler'));\r
+ $info = getimagesize($file);\r
+ restore_error_handler();\r
+ unlink($file);\r
+ if ($info == false) {\r
+ return false;\r
+ }\r
+ $image_code = $info[2];\r
+ } else {\r
+ trigger_error("could not find exif_imagetype or getimagesize functions", E_USER_ERROR);\r
+ }\r
+ $real_content_type = image_type_to_mime_type($image_code);\r
+ if ($real_content_type != $content_type) {\r
+ // we're nice guys; if the content type is something else we\r
+ // support, change it over\r
+ if (empty($this->allowed_types[$real_content_type])) {\r
+ return false;\r
+ }\r
+ $content_type = $real_content_type;\r
+ }\r
+ // ok, it's kosher, rewrite what we need\r
+ $uri->userinfo = null;\r
+ $uri->host = null;\r
+ $uri->port = null;\r
+ $uri->fragment = null;\r
+ $uri->query = null;\r
+ $uri->path = "$content_type;base64," . base64_encode($raw_data);\r
+ return true;\r
+ }\r
+\r
+ /**\r
+ * @param int $errno\r
+ * @param string $errstr\r
+ */\r
+ public function muteErrorHandler($errno, $errstr)\r
+ {\r
+ }\r
+}\r
+
+\r
+\r
+/**\r
+ * Validates file as defined by RFC 1630 and RFC 1738.\r
+ */\r
+class HTMLPurifier_URIScheme_file extends HTMLPurifier_URIScheme\r
+{\r
+ /**\r
+ * Generally file:// URLs are not accessible from most\r
+ * machines, so placing them as an img src is incorrect.\r
+ * @type bool\r
+ */\r
+ public $browsable = false;\r
+\r
+ /**\r
+ * Basically the *only* URI scheme for which this is true, since\r
+ * accessing files on the local machine is very common. In fact,\r
+ * browsers on some operating systems don't understand the\r
+ * authority, though I hear it is used on Windows to refer to\r
+ * network shares.\r
+ * @type bool\r
+ */\r
+ public $may_omit_host = true;\r
+\r
+ /**\r
+ * @param HTMLPurifier_URI $uri\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool\r
+ */\r
+ public function doValidate(&$uri, $config, $context)\r
+ {\r
+ // Authentication method is not supported\r
+ $uri->userinfo = null;\r
+ // file:// makes no provisions for accessing the resource\r
+ $uri->port = null;\r
+ // While it seems to work on Firefox, the querystring has\r
+ // no possible effect and is thus stripped.\r
+ $uri->query = null;\r
+ return true;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Validates ftp (File Transfer Protocol) URIs as defined by generic RFC 1738.\r
+ */\r
+class HTMLPurifier_URIScheme_ftp extends HTMLPurifier_URIScheme\r
+{\r
+ /**\r
+ * @type int\r
+ */\r
+ public $default_port = 21;\r
+\r
+ /**\r
+ * @type bool\r
+ */\r
+ public $browsable = true; // usually\r
+\r
+ /**\r
+ * @type bool\r
+ */\r
+ public $hierarchical = true;\r
+\r
+ /**\r
+ * @param HTMLPurifier_URI $uri\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool\r
+ */\r
+ public function doValidate(&$uri, $config, $context)\r
+ {\r
+ $uri->query = null;\r
+\r
+ // typecode check\r
+ $semicolon_pos = strrpos($uri->path, ';'); // reverse\r
+ if ($semicolon_pos !== false) {\r
+ $type = substr($uri->path, $semicolon_pos + 1); // no semicolon\r
+ $uri->path = substr($uri->path, 0, $semicolon_pos);\r
+ $type_ret = '';\r
+ if (strpos($type, '=') !== false) {\r
+ // figure out whether or not the declaration is correct\r
+ list($key, $typecode) = explode('=', $type, 2);\r
+ if ($key !== 'type') {\r
+ // invalid key, tack it back on encoded\r
+ $uri->path .= '%3B' . $type;\r
+ } elseif ($typecode === 'a' || $typecode === 'i' || $typecode === 'd') {\r
+ $type_ret = ";type=$typecode";\r
+ }\r
+ } else {\r
+ $uri->path .= '%3B' . $type;\r
+ }\r
+ $uri->path = str_replace(';', '%3B', $uri->path);\r
+ $uri->path .= $type_ret;\r
+ }\r
+ return true;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Validates http (HyperText Transfer Protocol) as defined by RFC 2616\r
+ */\r
+class HTMLPurifier_URIScheme_http extends HTMLPurifier_URIScheme\r
+{\r
+ /**\r
+ * @type int\r
+ */\r
+ public $default_port = 80;\r
+\r
+ /**\r
+ * @type bool\r
+ */\r
+ public $browsable = true;\r
+\r
+ /**\r
+ * @type bool\r
+ */\r
+ public $hierarchical = true;\r
+\r
+ /**\r
+ * @param HTMLPurifier_URI $uri\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool\r
+ */\r
+ public function doValidate(&$uri, $config, $context)\r
+ {\r
+ $uri->userinfo = null;\r
+ return true;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Validates https (Secure HTTP) according to http scheme.\r
+ */\r
+class HTMLPurifier_URIScheme_https extends HTMLPurifier_URIScheme_http\r
+{\r
+ /**\r
+ * @type int\r
+ */\r
+ public $default_port = 443;\r
+ /**\r
+ * @type bool\r
+ */\r
+ public $secure = true;\r
+}\r
+\r
+
+
+\r
+\r
+// VERY RELAXED! Shouldn't cause problems, not even Firefox checks if the\r
+// email is valid, but be careful!\r
+\r
+/**\r
+ * Validates mailto (for E-mail) according to RFC 2368\r
+ * @todo Validate the email address\r
+ * @todo Filter allowed query parameters\r
+ */\r
+\r
+class HTMLPurifier_URIScheme_mailto extends HTMLPurifier_URIScheme\r
+{\r
+ /**\r
+ * @type bool\r
+ */\r
+ public $browsable = false;\r
+\r
+ /**\r
+ * @type bool\r
+ */\r
+ public $may_omit_host = true;\r
+\r
+ /**\r
+ * @param HTMLPurifier_URI $uri\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool\r
+ */\r
+ public function doValidate(&$uri, $config, $context)\r
+ {\r
+ $uri->userinfo = null;\r
+ $uri->host = null;\r
+ $uri->port = null;\r
+ // we need to validate path against RFC 2368's addr-spec\r
+ return true;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Validates news (Usenet) as defined by generic RFC 1738\r
+ */\r
+class HTMLPurifier_URIScheme_news extends HTMLPurifier_URIScheme\r
+{\r
+ /**\r
+ * @type bool\r
+ */\r
+ public $browsable = false;\r
+\r
+ /**\r
+ * @type bool\r
+ */\r
+ public $may_omit_host = true;\r
+\r
+ /**\r
+ * @param HTMLPurifier_URI $uri\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool\r
+ */\r
+ public function doValidate(&$uri, $config, $context)\r
+ {\r
+ $uri->userinfo = null;\r
+ $uri->host = null;\r
+ $uri->port = null;\r
+ $uri->query = null;\r
+ // typecode check needed on path\r
+ return true;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Validates nntp (Network News Transfer Protocol) as defined by generic RFC 1738\r
+ */\r
+class HTMLPurifier_URIScheme_nntp extends HTMLPurifier_URIScheme\r
+{\r
+ /**\r
+ * @type int\r
+ */\r
+ public $default_port = 119;\r
+\r
+ /**\r
+ * @type bool\r
+ */\r
+ public $browsable = false;\r
+\r
+ /**\r
+ * @param HTMLPurifier_URI $uri\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool\r
+ */\r
+ public function doValidate(&$uri, $config, $context)\r
+ {\r
+ $uri->userinfo = null;\r
+ $uri->query = null;\r
+ return true;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Validates tel (for phone numbers).\r
+ *\r
+ * The relevant specifications for this protocol are RFC 3966 and RFC 5341,\r
+ * but this class takes a much simpler approach: we normalize phone\r
+ * numbers so that they only include (possibly) a leading plus,\r
+ * and then any number of digits and x'es.\r
+ */\r
+\r
+class HTMLPurifier_URIScheme_tel extends HTMLPurifier_URIScheme\r
+{\r
+ /**\r
+ * @type bool\r
+ */\r
+ public $browsable = false;\r
+\r
+ /**\r
+ * @type bool\r
+ */\r
+ public $may_omit_host = true;\r
+\r
+ /**\r
+ * @param HTMLPurifier_URI $uri\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return bool\r
+ */\r
+ public function doValidate(&$uri, $config, $context)\r
+ {\r
+ $uri->userinfo = null;\r
+ $uri->host = null;\r
+ $uri->port = null;\r
+\r
+ // Delete all non-numeric characters, non-x characters\r
+ // from phone number, EXCEPT for a leading plus sign.\r
+ $uri->path = preg_replace('/(?!^\+)[^\dx]/', '',\r
+ // Normalize e(x)tension to lower-case\r
+ str_replace('X', 'x', $uri->path));\r
+\r
+ return true;\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * Performs safe variable parsing based on types which can be used by\r
+ * users. This may not be able to represent all possible data inputs,\r
+ * however.\r
+ */\r
+class HTMLPurifier_VarParser_Flexible extends HTMLPurifier_VarParser\r
+{\r
+ /**\r
+ * @param mixed $var\r
+ * @param int $type\r
+ * @param bool $allow_null\r
+ * @return array|bool|float|int|mixed|null|string\r
+ * @throws HTMLPurifier_VarParserException\r
+ */\r
+ protected function parseImplementation($var, $type, $allow_null)\r
+ {\r
+ if ($allow_null && $var === null) {\r
+ return null;\r
+ }\r
+ switch ($type) {\r
+ // Note: if code "breaks" from the switch, it triggers a generic\r
+ // exception to be thrown. Specific errors can be specifically\r
+ // done here.\r
+ case self::C_MIXED:\r
+ case self::ISTRING:\r
+ case self::C_STRING:\r
+ case self::TEXT:\r
+ case self::ITEXT:\r
+ return $var;\r
+ case self::C_INT:\r
+ if (is_string($var) && ctype_digit($var)) {\r
+ $var = (int)$var;\r
+ }\r
+ return $var;\r
+ case self::C_FLOAT:\r
+ if ((is_string($var) && is_numeric($var)) || is_int($var)) {\r
+ $var = (float)$var;\r
+ }\r
+ return $var;\r
+ case self::C_BOOL:\r
+ if (is_int($var) && ($var === 0 || $var === 1)) {\r
+ $var = (bool)$var;\r
+ } elseif (is_string($var)) {\r
+ if ($var == 'on' || $var == 'true' || $var == '1') {\r
+ $var = true;\r
+ } elseif ($var == 'off' || $var == 'false' || $var == '0') {\r
+ $var = false;\r
+ } else {\r
+ throw new HTMLPurifier_VarParserException("Unrecognized value '$var' for $type");\r
+ }\r
+ }\r
+ return $var;\r
+ case self::ALIST:\r
+ case self::HASH:\r
+ case self::LOOKUP:\r
+ if (is_string($var)) {\r
+ // special case: technically, this is an array with\r
+ // a single empty string item, but having an empty\r
+ // array is more intuitive\r
+ if ($var == '') {\r
+ return array();\r
+ }\r
+ if (strpos($var, "\n") === false && strpos($var, "\r") === false) {\r
+ // simplistic string to array method that only works\r
+ // for simple lists of tag names or alphanumeric characters\r
+ $var = explode(',', $var);\r
+ } else {\r
+ $var = preg_split('/(,|[\n\r]+)/', $var);\r
+ }\r
+ // remove spaces\r
+ foreach ($var as $i => $j) {\r
+ $var[$i] = trim($j);\r
+ }\r
+ if ($type === self::HASH) {\r
+ // key:value,key2:value2\r
+ $nvar = array();\r
+ foreach ($var as $keypair) {\r
+ $c = explode(':', $keypair, 2);\r
+ if (!isset($c[1])) {\r
+ continue;\r
+ }\r
+ $nvar[trim($c[0])] = trim($c[1]);\r
+ }\r
+ $var = $nvar;\r
+ }\r
+ }\r
+ if (!is_array($var)) {\r
+ break;\r
+ }\r
+ $keys = array_keys($var);\r
+ if ($keys === array_keys($keys)) {\r
+ if ($type == self::ALIST) {\r
+ return $var;\r
+ } elseif ($type == self::LOOKUP) {\r
+ $new = array();\r
+ foreach ($var as $key) {\r
+ $new[$key] = true;\r
+ }\r
+ return $new;\r
+ } else {\r
+ break;\r
+ }\r
+ }\r
+ if ($type === self::ALIST) {\r
+ trigger_error("Array list did not have consecutive integer indexes", E_USER_WARNING);\r
+ return array_values($var);\r
+ }\r
+ if ($type === self::LOOKUP) {\r
+ foreach ($var as $key => $value) {\r
+ if ($value !== true) {\r
+ trigger_error(\r
+ "Lookup array has non-true value at key '$key'; " .\r
+ "maybe your input array was not indexed numerically",\r
+ E_USER_WARNING\r
+ );\r
+ }\r
+ $var[$key] = true;\r
+ }\r
+ }\r
+ return $var;\r
+ default:\r
+ $this->errorInconsistent(__CLASS__, $type);\r
+ }\r
+ $this->errorGeneric($var, $type);\r
+ }\r
+}\r
+\r
+
+
+\r
+\r
+/**\r
+ * This variable parser uses PHP's internal code engine. Because it does\r
+ * this, it can represent all inputs; however, it is dangerous and cannot\r
+ * be used by users.\r
+ */\r
+class HTMLPurifier_VarParser_Native extends HTMLPurifier_VarParser\r
+{\r
+\r
+ /**\r
+ * @param mixed $var\r
+ * @param int $type\r
+ * @param bool $allow_null\r
+ * @return null|string\r
+ */\r
+ protected function parseImplementation($var, $type, $allow_null)\r
+ {\r
+ return $this->evalExpression($var);\r
+ }\r
+\r
+ /**\r
+ * @param string $expr\r
+ * @return mixed\r
+ * @throws HTMLPurifier_VarParserException\r
+ */\r
+ protected function evalExpression($expr)\r
+ {\r
+ $var = null;\r
+ $result = eval("\$var = $expr;");\r
+ if ($result === false) {\r
+ throw new HTMLPurifier_VarParserException("Fatal error in evaluated code");\r
+ }\r
+ return $var;\r
+ }\r
+}\r
+\r
+
+
--- /dev/null
+<?php\r
+\r
+/**\r
+ * Experimental HTML5-based parser using Jeroen van der Meer's PH5P library.\r
+ * Occupies space in the HTML5 pseudo-namespace, which may cause conflicts.\r
+ *\r
+ * @note\r
+ * Recent changes to PHP's DOM extension have resulted in some fatal\r
+ * error conditions with the original version of PH5P. Pending changes,\r
+ * this lexer will punt to DirectLex if DOM throws an exception.\r
+ */\r
+\r
+class HTMLPurifier_Lexer_PH5P extends HTMLPurifier_Lexer_DOMLex\r
+{\r
+ /**\r
+ * @param string $html\r
+ * @param HTMLPurifier_Config $config\r
+ * @param HTMLPurifier_Context $context\r
+ * @return HTMLPurifier_Token[]\r
+ */\r
+ public function tokenizeHTML($html, $config, $context)\r
+ {\r
+ $new_html = $this->normalize($html, $config, $context);\r
+ $new_html = $this->wrapHTML($new_html, $config, $context, false /* no div */);\r
+ try {\r
+ $parser = new HTML5($new_html);\r
+ $doc = $parser->save();\r
+ } catch (DOMException $e) {\r
+ // Uh oh, it failed. Punt to DirectLex.\r
+ $lexer = new HTMLPurifier_Lexer_DirectLex();\r
+ $context->register('PH5PError', $e); // save the error, so we can detect it\r
+ return $lexer->tokenizeHTML($html, $config, $context); // use original HTML\r
+ }\r
+ $tokens = array();\r
+ $this->tokenizeDOM(\r
+ $doc->getElementsByTagName('html')->item(0)-> // <html>\r
+ getElementsByTagName('body')->item(0) // <body>\r
+ ,\r
+ $tokens, $config\r
+ );\r
+ return $tokens;\r
+ }\r
+}\r
+\r
+/*\r
+\r
+Copyright 2007 Jeroen van der Meer <http://jero.net/>\r
+\r
+Permission is hereby granted, free of charge, to any person obtaining a\r
+copy of this software and associated documentation files (the\r
+"Software"), to deal in the Software without restriction, including\r
+without limitation the rights to use, copy, modify, merge, publish,\r
+distribute, sublicense, and/or sell copies of the Software, and to\r
+permit persons to whom the Software is furnished to do so, subject to\r
+the following conditions:\r
+\r
+The above copyright notice and this permission notice shall be included\r
+in all copies or substantial portions of the Software.\r
+\r
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS\r
+OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF\r
+MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.\r
+IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY\r
+CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,\r
+TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE\r
+SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.\r
+\r
+*/\r
+\r
+class HTML5\r
+{\r
+ private $data;\r
+ private $char;\r
+ private $EOF;\r
+ private $state;\r
+ private $tree;\r
+ private $token;\r
+ private $content_model;\r
+ private $escape = false;\r
+ private $entities = array(\r
+ 'AElig;',\r
+ 'AElig',\r
+ 'AMP;',\r
+ 'AMP',\r
+ 'Aacute;',\r
+ 'Aacute',\r
+ 'Acirc;',\r
+ 'Acirc',\r
+ 'Agrave;',\r
+ 'Agrave',\r
+ 'Alpha;',\r
+ 'Aring;',\r
+ 'Aring',\r
+ 'Atilde;',\r
+ 'Atilde',\r
+ 'Auml;',\r
+ 'Auml',\r
+ 'Beta;',\r
+ 'COPY;',\r
+ 'COPY',\r
+ 'Ccedil;',\r
+ 'Ccedil',\r
+ 'Chi;',\r
+ 'Dagger;',\r
+ 'Delta;',\r
+ 'ETH;',\r
+ 'ETH',\r
+ 'Eacute;',\r
+ 'Eacute',\r
+ 'Ecirc;',\r
+ 'Ecirc',\r
+ 'Egrave;',\r
+ 'Egrave',\r
+ 'Epsilon;',\r
+ 'Eta;',\r
+ 'Euml;',\r
+ 'Euml',\r
+ 'GT;',\r
+ 'GT',\r
+ 'Gamma;',\r
+ 'Iacute;',\r
+ 'Iacute',\r
+ 'Icirc;',\r
+ 'Icirc',\r
+ 'Igrave;',\r
+ 'Igrave',\r
+ 'Iota;',\r
+ 'Iuml;',\r
+ 'Iuml',\r
+ 'Kappa;',\r
+ 'LT;',\r
+ 'LT',\r
+ 'Lambda;',\r
+ 'Mu;',\r
+ 'Ntilde;',\r
+ 'Ntilde',\r
+ 'Nu;',\r
+ 'OElig;',\r
+ 'Oacute;',\r
+ 'Oacute',\r
+ 'Ocirc;',\r
+ 'Ocirc',\r
+ 'Ograve;',\r
+ 'Ograve',\r
+ 'Omega;',\r
+ 'Omicron;',\r
+ 'Oslash;',\r
+ 'Oslash',\r
+ 'Otilde;',\r
+ 'Otilde',\r
+ 'Ouml;',\r
+ 'Ouml',\r
+ 'Phi;',\r
+ 'Pi;',\r
+ 'Prime;',\r
+ 'Psi;',\r
+ 'QUOT;',\r
+ 'QUOT',\r
+ 'REG;',\r
+ 'REG',\r
+ 'Rho;',\r
+ 'Scaron;',\r
+ 'Sigma;',\r
+ 'THORN;',\r
+ 'THORN',\r
+ 'TRADE;',\r
+ 'Tau;',\r
+ 'Theta;',\r
+ 'Uacute;',\r
+ 'Uacute',\r
+ 'Ucirc;',\r
+ 'Ucirc',\r
+ 'Ugrave;',\r
+ 'Ugrave',\r
+ 'Upsilon;',\r
+ 'Uuml;',\r
+ 'Uuml',\r
+ 'Xi;',\r
+ 'Yacute;',\r
+ 'Yacute',\r
+ 'Yuml;',\r
+ 'Zeta;',\r
+ 'aacute;',\r
+ 'aacute',\r
+ 'acirc;',\r
+ 'acirc',\r
+ 'acute;',\r
+ 'acute',\r
+ 'aelig;',\r
+ 'aelig',\r
+ 'agrave;',\r
+ 'agrave',\r
+ 'alefsym;',\r
+ 'alpha;',\r
+ 'amp;',\r
+ 'amp',\r
+ 'and;',\r
+ 'ang;',\r
+ 'apos;',\r
+ 'aring;',\r
+ 'aring',\r
+ 'asymp;',\r
+ 'atilde;',\r
+ 'atilde',\r
+ 'auml;',\r
+ 'auml',\r
+ 'bdquo;',\r
+ 'beta;',\r
+ 'brvbar;',\r
+ 'brvbar',\r
+ 'bull;',\r
+ 'cap;',\r
+ 'ccedil;',\r
+ 'ccedil',\r
+ 'cedil;',\r
+ 'cedil',\r
+ 'cent;',\r
+ 'cent',\r
+ 'chi;',\r
+ 'circ;',\r
+ 'clubs;',\r
+ 'cong;',\r
+ 'copy;',\r
+ 'copy',\r
+ 'crarr;',\r
+ 'cup;',\r
+ 'curren;',\r
+ 'curren',\r
+ 'dArr;',\r
+ 'dagger;',\r
+ 'darr;',\r
+ 'deg;',\r
+ 'deg',\r
+ 'delta;',\r
+ 'diams;',\r
+ 'divide;',\r
+ 'divide',\r
+ 'eacute;',\r
+ 'eacute',\r
+ 'ecirc;',\r
+ 'ecirc',\r
+ 'egrave;',\r
+ 'egrave',\r
+ 'empty;',\r
+ 'emsp;',\r
+ 'ensp;',\r
+ 'epsilon;',\r
+ 'equiv;',\r
+ 'eta;',\r
+ 'eth;',\r
+ 'eth',\r
+ 'euml;',\r
+ 'euml',\r
+ 'euro;',\r
+ 'exist;',\r
+ 'fnof;',\r
+ 'forall;',\r
+ 'frac12;',\r
+ 'frac12',\r
+ 'frac14;',\r
+ 'frac14',\r
+ 'frac34;',\r
+ 'frac34',\r
+ 'frasl;',\r
+ 'gamma;',\r
+ 'ge;',\r
+ 'gt;',\r
+ 'gt',\r
+ 'hArr;',\r
+ 'harr;',\r
+ 'hearts;',\r
+ 'hellip;',\r
+ 'iacute;',\r
+ 'iacute',\r
+ 'icirc;',\r
+ 'icirc',\r
+ 'iexcl;',\r
+ 'iexcl',\r
+ 'igrave;',\r
+ 'igrave',\r
+ 'image;',\r
+ 'infin;',\r
+ 'int;',\r
+ 'iota;',\r
+ 'iquest;',\r
+ 'iquest',\r
+ 'isin;',\r
+ 'iuml;',\r
+ 'iuml',\r
+ 'kappa;',\r
+ 'lArr;',\r
+ 'lambda;',\r
+ 'lang;',\r
+ 'laquo;',\r
+ 'laquo',\r
+ 'larr;',\r
+ 'lceil;',\r
+ 'ldquo;',\r
+ 'le;',\r
+ 'lfloor;',\r
+ 'lowast;',\r
+ 'loz;',\r
+ 'lrm;',\r
+ 'lsaquo;',\r
+ 'lsquo;',\r
+ 'lt;',\r
+ 'lt',\r
+ 'macr;',\r
+ 'macr',\r
+ 'mdash;',\r
+ 'micro;',\r
+ 'micro',\r
+ 'middot;',\r
+ 'middot',\r
+ 'minus;',\r
+ 'mu;',\r
+ 'nabla;',\r
+ 'nbsp;',\r
+ 'nbsp',\r
+ 'ndash;',\r
+ 'ne;',\r
+ 'ni;',\r
+ 'not;',\r
+ 'not',\r
+ 'notin;',\r
+ 'nsub;',\r
+ 'ntilde;',\r
+ 'ntilde',\r
+ 'nu;',\r
+ 'oacute;',\r
+ 'oacute',\r
+ 'ocirc;',\r
+ 'ocirc',\r
+ 'oelig;',\r
+ 'ograve;',\r
+ 'ograve',\r
+ 'oline;',\r
+ 'omega;',\r
+ 'omicron;',\r
+ 'oplus;',\r
+ 'or;',\r
+ 'ordf;',\r
+ 'ordf',\r
+ 'ordm;',\r
+ 'ordm',\r
+ 'oslash;',\r
+ 'oslash',\r
+ 'otilde;',\r
+ 'otilde',\r
+ 'otimes;',\r
+ 'ouml;',\r
+ 'ouml',\r
+ 'para;',\r
+ 'para',\r
+ 'part;',\r
+ 'permil;',\r
+ 'perp;',\r
+ 'phi;',\r
+ 'pi;',\r
+ 'piv;',\r
+ 'plusmn;',\r
+ 'plusmn',\r
+ 'pound;',\r
+ 'pound',\r
+ 'prime;',\r
+ 'prod;',\r
+ 'prop;',\r
+ 'psi;',\r
+ 'quot;',\r
+ 'quot',\r
+ 'rArr;',\r
+ 'radic;',\r
+ 'rang;',\r
+ 'raquo;',\r
+ 'raquo',\r
+ 'rarr;',\r
+ 'rceil;',\r
+ 'rdquo;',\r
+ 'real;',\r
+ 'reg;',\r
+ 'reg',\r
+ 'rfloor;',\r
+ 'rho;',\r
+ 'rlm;',\r
+ 'rsaquo;',\r
+ 'rsquo;',\r
+ 'sbquo;',\r
+ 'scaron;',\r
+ 'sdot;',\r
+ 'sect;',\r
+ 'sect',\r
+ 'shy;',\r
+ 'shy',\r
+ 'sigma;',\r
+ 'sigmaf;',\r
+ 'sim;',\r
+ 'spades;',\r
+ 'sub;',\r
+ 'sube;',\r
+ 'sum;',\r
+ 'sup1;',\r
+ 'sup1',\r
+ 'sup2;',\r
+ 'sup2',\r
+ 'sup3;',\r
+ 'sup3',\r
+ 'sup;',\r
+ 'supe;',\r
+ 'szlig;',\r
+ 'szlig',\r
+ 'tau;',\r
+ 'there4;',\r
+ 'theta;',\r
+ 'thetasym;',\r
+ 'thinsp;',\r
+ 'thorn;',\r
+ 'thorn',\r
+ 'tilde;',\r
+ 'times;',\r
+ 'times',\r
+ 'trade;',\r
+ 'uArr;',\r
+ 'uacute;',\r
+ 'uacute',\r
+ 'uarr;',\r
+ 'ucirc;',\r
+ 'ucirc',\r
+ 'ugrave;',\r
+ 'ugrave',\r
+ 'uml;',\r
+ 'uml',\r
+ 'upsih;',\r
+ 'upsilon;',\r
+ 'uuml;',\r
+ 'uuml',\r
+ 'weierp;',\r
+ 'xi;',\r
+ 'yacute;',\r
+ 'yacute',\r
+ 'yen;',\r
+ 'yen',\r
+ 'yuml;',\r
+ 'yuml',\r
+ 'zeta;',\r
+ 'zwj;',\r
+ 'zwnj;'\r
+ );\r
+\r
+ const PCDATA = 0;\r
+ const RCDATA = 1;\r
+ const CDATA = 2;\r
+ const PLAINTEXT = 3;\r
+\r
+ const DOCTYPE = 0;\r
+ const STARTTAG = 1;\r
+ const ENDTAG = 2;\r
+ const COMMENT = 3;\r
+ const CHARACTR = 4;\r
+ const EOF = 5;\r
+\r
+ public function __construct($data)\r
+ {\r
+ $this->data = $data;\r
+ $this->char = -1;\r
+ $this->EOF = strlen($data);\r
+ $this->tree = new HTML5TreeConstructer;\r
+ $this->content_model = self::PCDATA;\r
+\r
+ $this->state = 'data';\r
+\r
+ while ($this->state !== null) {\r
+ $this->{$this->state . 'State'}();\r
+ }\r
+ }\r
+\r
+ public function save()\r
+ {\r
+ return $this->tree->save();\r
+ }\r
+\r
+ private function char()\r
+ {\r
+ return ($this->char < $this->EOF)\r
+ ? $this->data[$this->char]\r
+ : false;\r
+ }\r
+\r
+ private function character($s, $l = 0)\r
+ {\r
+ if ($s + $l < $this->EOF) {\r
+ if ($l === 0) {\r
+ return $this->data[$s];\r
+ } else {\r
+ return substr($this->data, $s, $l);\r
+ }\r
+ }\r
+ }\r
+\r
+ private function characters($char_class, $start)\r
+ {\r
+ return preg_replace('#^([' . $char_class . ']+).*#s', '\\1', substr($this->data, $start));\r
+ }\r
+\r
+ private function dataState()\r
+ {\r
+ // Consume the next input character\r
+ $this->char++;\r
+ $char = $this->char();\r
+\r
+ if ($char === '&' && ($this->content_model === self::PCDATA || $this->content_model === self::RCDATA)) {\r
+ /* U+0026 AMPERSAND (&)\r
+ When the content model flag is set to one of the PCDATA or RCDATA\r
+ states: switch to the entity data state. Otherwise: treat it as per\r
+ the "anything else" entry below. */\r
+ $this->state = 'entityData';\r
+\r
+ } elseif ($char === '-') {\r
+ /* If the content model flag is set to either the RCDATA state or\r
+ the CDATA state, and the escape flag is false, and there are at\r
+ least three characters before this one in the input stream, and the\r
+ last four characters in the input stream, including this one, are\r
+ U+003C LESS-THAN SIGN, U+0021 EXCLAMATION MARK, U+002D HYPHEN-MINUS,\r
+ and U+002D HYPHEN-MINUS ("<!--"), then set the escape flag to true. */\r
+ if (($this->content_model === self::RCDATA || $this->content_model ===\r
+ self::CDATA) && $this->escape === false &&\r
+ $this->char >= 3 && $this->character($this->char - 4, 4) === '<!--'\r
+ ) {\r
+ $this->escape = true;\r
+ }\r
+\r
+ /* In any case, emit the input character as a character token. Stay\r
+ in the data state. */\r
+ $this->emitToken(\r
+ array(\r
+ 'type' => self::CHARACTR,\r
+ 'data' => $char\r
+ )\r
+ );\r
+\r
+ /* U+003C LESS-THAN SIGN (<) */\r
+ } elseif ($char === '<' && ($this->content_model === self::PCDATA ||\r
+ (($this->content_model === self::RCDATA ||\r
+ $this->content_model === self::CDATA) && $this->escape === false))\r
+ ) {\r
+ /* When the content model flag is set to the PCDATA state: switch\r
+ to the tag open state.\r
+\r
+ When the content model flag is set to either the RCDATA state or\r
+ the CDATA state and the escape flag is false: switch to the tag\r
+ open state.\r
+\r
+ Otherwise: treat it as per the "anything else" entry below. */\r
+ $this->state = 'tagOpen';\r
+\r
+ /* U+003E GREATER-THAN SIGN (>) */\r
+ } elseif ($char === '>') {\r
+ /* If the content model flag is set to either the RCDATA state or\r
+ the CDATA state, and the escape flag is true, and the last three\r
+ characters in the input stream including this one are U+002D\r
+ HYPHEN-MINUS, U+002D HYPHEN-MINUS, U+003E GREATER-THAN SIGN ("-->"),\r
+ set the escape flag to false. */\r
+ if (($this->content_model === self::RCDATA ||\r
+ $this->content_model === self::CDATA) && $this->escape === true &&\r
+ $this->character($this->char, 3) === '-->'\r
+ ) {\r
+ $this->escape = false;\r
+ }\r
+\r
+ /* In any case, emit the input character as a character token.\r
+ Stay in the data state. */\r
+ $this->emitToken(\r
+ array(\r
+ 'type' => self::CHARACTR,\r
+ 'data' => $char\r
+ )\r
+ );\r
+\r
+ } elseif ($this->char === $this->EOF) {\r
+ /* EOF\r
+ Emit an end-of-file token. */\r
+ $this->EOF();\r
+\r
+ } elseif ($this->content_model === self::PLAINTEXT) {\r
+ /* When the content model flag is set to the PLAINTEXT state\r
+ THIS DIFFERS GREATLY FROM THE SPEC: Get the remaining characters of\r
+ the text and emit it as a character token. */\r
+ $this->emitToken(\r
+ array(\r
+ 'type' => self::CHARACTR,\r
+ 'data' => substr($this->data, $this->char)\r
+ )\r
+ );\r
+\r
+ $this->EOF();\r
+\r
+ } else {\r
+ /* Anything else\r
+ THIS DIFFERS GREATLY FROM THE SPEC: Get as many character that\r
+ otherwise would also be treated as a character token and emit it\r
+ as a single character token. Stay in the data state. */\r
+ $len = strcspn($this->data, '<&', $this->char);\r
+ $char = substr($this->data, $this->char, $len);\r
+ $this->char += $len - 1;\r
+\r
+ $this->emitToken(\r
+ array(\r
+ 'type' => self::CHARACTR,\r
+ 'data' => $char\r
+ )\r
+ );\r
+\r
+ $this->state = 'data';\r
+ }\r
+ }\r
+\r
+ private function entityDataState()\r
+ {\r
+ // Attempt to consume an entity.\r
+ $entity = $this->entity();\r
+\r
+ // If nothing is returned, emit a U+0026 AMPERSAND character token.\r
+ // Otherwise, emit the character token that was returned.\r
+ $char = (!$entity) ? '&' : $entity;\r
+ $this->emitToken(\r
+ array(\r
+ 'type' => self::CHARACTR,\r
+ 'data' => $char\r
+ )\r
+ );\r
+\r
+ // Finally, switch to the data state.\r
+ $this->state = 'data';\r
+ }\r
+\r
+ private function tagOpenState()\r
+ {\r
+ switch ($this->content_model) {\r
+ case self::RCDATA:\r
+ case self::CDATA:\r
+ /* If the next input character is a U+002F SOLIDUS (/) character,\r
+ consume it and switch to the close tag open state. If the next\r
+ input character is not a U+002F SOLIDUS (/) character, emit a\r
+ U+003C LESS-THAN SIGN character token and switch to the data\r
+ state to process the next input character. */\r
+ if ($this->character($this->char + 1) === '/') {\r
+ $this->char++;\r
+ $this->state = 'closeTagOpen';\r
+\r
+ } else {\r
+ $this->emitToken(\r
+ array(\r
+ 'type' => self::CHARACTR,\r
+ 'data' => '<'\r
+ )\r
+ );\r
+\r
+ $this->state = 'data';\r
+ }\r
+ break;\r
+\r
+ case self::PCDATA:\r
+ // If the content model flag is set to the PCDATA state\r
+ // Consume the next input character:\r
+ $this->char++;\r
+ $char = $this->char();\r
+\r
+ if ($char === '!') {\r
+ /* U+0021 EXCLAMATION MARK (!)\r
+ Switch to the markup declaration open state. */\r
+ $this->state = 'markupDeclarationOpen';\r
+\r
+ } elseif ($char === '/') {\r
+ /* U+002F SOLIDUS (/)\r
+ Switch to the close tag open state. */\r
+ $this->state = 'closeTagOpen';\r
+\r
+ } elseif (preg_match('/^[A-Za-z]$/', $char)) {\r
+ /* U+0041 LATIN LETTER A through to U+005A LATIN LETTER Z\r
+ Create a new start tag token, set its tag name to the lowercase\r
+ version of the input character (add 0x0020 to the character's code\r
+ point), then switch to the tag name state. (Don't emit the token\r
+ yet; further details will be filled in before it is emitted.) */\r
+ $this->token = array(\r
+ 'name' => strtolower($char),\r
+ 'type' => self::STARTTAG,\r
+ 'attr' => array()\r
+ );\r
+\r
+ $this->state = 'tagName';\r
+\r
+ } elseif ($char === '>') {\r
+ /* U+003E GREATER-THAN SIGN (>)\r
+ Parse error. Emit a U+003C LESS-THAN SIGN character token and a\r
+ U+003E GREATER-THAN SIGN character token. Switch to the data state. */\r
+ $this->emitToken(\r
+ array(\r
+ 'type' => self::CHARACTR,\r
+ 'data' => '<>'\r
+ )\r
+ );\r
+\r
+ $this->state = 'data';\r
+\r
+ } elseif ($char === '?') {\r
+ /* U+003F QUESTION MARK (?)\r
+ Parse error. Switch to the bogus comment state. */\r
+ $this->state = 'bogusComment';\r
+\r
+ } else {\r
+ /* Anything else\r
+ Parse error. Emit a U+003C LESS-THAN SIGN character token and\r
+ reconsume the current input character in the data state. */\r
+ $this->emitToken(\r
+ array(\r
+ 'type' => self::CHARACTR,\r
+ 'data' => '<'\r
+ )\r
+ );\r
+\r
+ $this->char--;\r
+ $this->state = 'data';\r
+ }\r
+ break;\r
+ }\r
+ }\r
+\r
+ private function closeTagOpenState()\r
+ {\r
+ $next_node = strtolower($this->characters('A-Za-z', $this->char + 1));\r
+ $the_same = count($this->tree->stack) > 0 && $next_node === end($this->tree->stack)->nodeName;\r
+\r
+ if (($this->content_model === self::RCDATA || $this->content_model === self::CDATA) &&\r
+ (!$the_same || ($the_same && (!preg_match(\r
+ '/[\t\n\x0b\x0c >\/]/',\r
+ $this->character($this->char + 1 + strlen($next_node))\r
+ ) || $this->EOF === $this->char)))\r
+ ) {\r
+ /* If the content model flag is set to the RCDATA or CDATA states then\r
+ examine the next few characters. If they do not match the tag name of\r
+ the last start tag token emitted (case insensitively), or if they do but\r
+ they are not immediately followed by one of the following characters:\r
+ * U+0009 CHARACTER TABULATION\r
+ * U+000A LINE FEED (LF)\r
+ * U+000B LINE TABULATION\r
+ * U+000C FORM FEED (FF)\r
+ * U+0020 SPACE\r
+ * U+003E GREATER-THAN SIGN (>)\r
+ * U+002F SOLIDUS (/)\r
+ * EOF\r
+ ...then there is a parse error. Emit a U+003C LESS-THAN SIGN character\r
+ token, a U+002F SOLIDUS character token, and switch to the data state\r
+ to process the next input character. */\r
+ $this->emitToken(\r
+ array(\r
+ 'type' => self::CHARACTR,\r
+ 'data' => '</'\r
+ )\r
+ );\r
+\r
+ $this->state = 'data';\r
+\r
+ } else {\r
+ /* Otherwise, if the content model flag is set to the PCDATA state,\r
+ or if the next few characters do match that tag name, consume the\r
+ next input character: */\r
+ $this->char++;\r
+ $char = $this->char();\r
+\r
+ if (preg_match('/^[A-Za-z]$/', $char)) {\r
+ /* U+0041 LATIN LETTER A through to U+005A LATIN LETTER Z\r
+ Create a new end tag token, set its tag name to the lowercase version\r
+ of the input character (add 0x0020 to the character's code point), then\r
+ switch to the tag name state. (Don't emit the token yet; further details\r
+ will be filled in before it is emitted.) */\r
+ $this->token = array(\r
+ 'name' => strtolower($char),\r
+ 'type' => self::ENDTAG\r
+ );\r
+\r
+ $this->state = 'tagName';\r
+\r
+ } elseif ($char === '>') {\r
+ /* U+003E GREATER-THAN SIGN (>)\r
+ Parse error. Switch to the data state. */\r
+ $this->state = 'data';\r
+\r
+ } elseif ($this->char === $this->EOF) {\r
+ /* EOF\r
+ Parse error. Emit a U+003C LESS-THAN SIGN character token and a U+002F\r
+ SOLIDUS character token. Reconsume the EOF character in the data state. */\r
+ $this->emitToken(\r
+ array(\r
+ 'type' => self::CHARACTR,\r
+ 'data' => '</'\r
+ )\r
+ );\r
+\r
+ $this->char--;\r
+ $this->state = 'data';\r
+\r
+ } else {\r
+ /* Parse error. Switch to the bogus comment state. */\r
+ $this->state = 'bogusComment';\r
+ }\r
+ }\r
+ }\r
+\r
+ private function tagNameState()\r
+ {\r
+ // Consume the next input character:\r
+ $this->char++;\r
+ $char = $this->character($this->char);\r
+\r
+ if (preg_match('/^[\t\n\x0b\x0c ]$/', $char)) {\r
+ /* U+0009 CHARACTER TABULATION\r
+ U+000A LINE FEED (LF)\r
+ U+000B LINE TABULATION\r
+ U+000C FORM FEED (FF)\r
+ U+0020 SPACE\r
+ Switch to the before attribute name state. */\r
+ $this->state = 'beforeAttributeName';\r
+\r
+ } elseif ($char === '>') {\r
+ /* U+003E GREATER-THAN SIGN (>)\r
+ Emit the current tag token. Switch to the data state. */\r
+ $this->emitToken($this->token);\r
+ $this->state = 'data';\r
+\r
+ } elseif ($this->char === $this->EOF) {\r
+ /* EOF\r
+ Parse error. Emit the current tag token. Reconsume the EOF\r
+ character in the data state. */\r
+ $this->emitToken($this->token);\r
+\r
+ $this->char--;\r
+ $this->state = 'data';\r
+\r
+ } elseif ($char === '/') {\r
+ /* U+002F SOLIDUS (/)\r
+ Parse error unless this is a permitted slash. Switch to the before\r
+ attribute name state. */\r
+ $this->state = 'beforeAttributeName';\r
+\r
+ } else {\r
+ /* Anything else\r
+ Append the current input character to the current tag token's tag name.\r
+ Stay in the tag name state. */\r
+ $this->token['name'] .= strtolower($char);\r
+ $this->state = 'tagName';\r
+ }\r
+ }\r
+\r
+ private function beforeAttributeNameState()\r
+ {\r
+ // Consume the next input character:\r
+ $this->char++;\r
+ $char = $this->character($this->char);\r
+\r
+ if (preg_match('/^[\t\n\x0b\x0c ]$/', $char)) {\r
+ /* U+0009 CHARACTER TABULATION\r
+ U+000A LINE FEED (LF)\r
+ U+000B LINE TABULATION\r
+ U+000C FORM FEED (FF)\r
+ U+0020 SPACE\r
+ Stay in the before attribute name state. */\r
+ $this->state = 'beforeAttributeName';\r
+\r
+ } elseif ($char === '>') {\r
+ /* U+003E GREATER-THAN SIGN (>)\r
+ Emit the current tag token. Switch to the data state. */\r
+ $this->emitToken($this->token);\r
+ $this->state = 'data';\r
+\r
+ } elseif ($char === '/') {\r
+ /* U+002F SOLIDUS (/)\r
+ Parse error unless this is a permitted slash. Stay in the before\r
+ attribute name state. */\r
+ $this->state = 'beforeAttributeName';\r
+\r
+ } elseif ($this->char === $this->EOF) {\r
+ /* EOF\r
+ Parse error. Emit the current tag token. Reconsume the EOF\r
+ character in the data state. */\r
+ $this->emitToken($this->token);\r
+\r
+ $this->char--;\r
+ $this->state = 'data';\r
+\r
+ } else {\r
+ /* Anything else\r
+ Start a new attribute in the current tag token. Set that attribute's\r
+ name to the current input character, and its value to the empty string.\r
+ Switch to the attribute name state. */\r
+ $this->token['attr'][] = array(\r
+ 'name' => strtolower($char),\r
+ 'value' => null\r
+ );\r
+\r
+ $this->state = 'attributeName';\r
+ }\r
+ }\r
+\r
+ private function attributeNameState()\r
+ {\r
+ // Consume the next input character:\r
+ $this->char++;\r
+ $char = $this->character($this->char);\r
+\r
+ if (preg_match('/^[\t\n\x0b\x0c ]$/', $char)) {\r
+ /* U+0009 CHARACTER TABULATION\r
+ U+000A LINE FEED (LF)\r
+ U+000B LINE TABULATION\r
+ U+000C FORM FEED (FF)\r
+ U+0020 SPACE\r
+ Stay in the before attribute name state. */\r
+ $this->state = 'afterAttributeName';\r
+\r
+ } elseif ($char === '=') {\r
+ /* U+003D EQUALS SIGN (=)\r
+ Switch to the before attribute value state. */\r
+ $this->state = 'beforeAttributeValue';\r
+\r
+ } elseif ($char === '>') {\r
+ /* U+003E GREATER-THAN SIGN (>)\r
+ Emit the current tag token. Switch to the data state. */\r
+ $this->emitToken($this->token);\r
+ $this->state = 'data';\r
+\r
+ } elseif ($char === '/' && $this->character($this->char + 1) !== '>') {\r
+ /* U+002F SOLIDUS (/)\r
+ Parse error unless this is a permitted slash. Switch to the before\r
+ attribute name state. */\r
+ $this->state = 'beforeAttributeName';\r
+\r
+ } elseif ($this->char === $this->EOF) {\r
+ /* EOF\r
+ Parse error. Emit the current tag token. Reconsume the EOF\r
+ character in the data state. */\r
+ $this->emitToken($this->token);\r
+\r
+ $this->char--;\r
+ $this->state = 'data';\r
+\r
+ } else {\r
+ /* Anything else\r
+ Append the current input character to the current attribute's name.\r
+ Stay in the attribute name state. */\r
+ $last = count($this->token['attr']) - 1;\r
+ $this->token['attr'][$last]['name'] .= strtolower($char);\r
+\r
+ $this->state = 'attributeName';\r
+ }\r
+ }\r
+\r
+ private function afterAttributeNameState()\r
+ {\r
+ // Consume the next input character:\r
+ $this->char++;\r
+ $char = $this->character($this->char);\r
+\r
+ if (preg_match('/^[\t\n\x0b\x0c ]$/', $char)) {\r
+ /* U+0009 CHARACTER TABULATION\r
+ U+000A LINE FEED (LF)\r
+ U+000B LINE TABULATION\r
+ U+000C FORM FEED (FF)\r
+ U+0020 SPACE\r
+ Stay in the after attribute name state. */\r
+ $this->state = 'afterAttributeName';\r
+\r
+ } elseif ($char === '=') {\r
+ /* U+003D EQUALS SIGN (=)\r
+ Switch to the before attribute value state. */\r
+ $this->state = 'beforeAttributeValue';\r
+\r
+ } elseif ($char === '>') {\r
+ /* U+003E GREATER-THAN SIGN (>)\r
+ Emit the current tag token. Switch to the data state. */\r
+ $this->emitToken($this->token);\r
+ $this->state = 'data';\r
+\r
+ } elseif ($char === '/' && $this->character($this->char + 1) !== '>') {\r
+ /* U+002F SOLIDUS (/)\r
+ Parse error unless this is a permitted slash. Switch to the\r
+ before attribute name state. */\r
+ $this->state = 'beforeAttributeName';\r
+\r
+ } elseif ($this->char === $this->EOF) {\r
+ /* EOF\r
+ Parse error. Emit the current tag token. Reconsume the EOF\r
+ character in the data state. */\r
+ $this->emitToken($this->token);\r
+\r
+ $this->char--;\r
+ $this->state = 'data';\r
+\r
+ } else {\r
+ /* Anything else\r
+ Start a new attribute in the current tag token. Set that attribute's\r
+ name to the current input character, and its value to the empty string.\r
+ Switch to the attribute name state. */\r
+ $this->token['attr'][] = array(\r
+ 'name' => strtolower($char),\r
+ 'value' => null\r
+ );\r
+\r
+ $this->state = 'attributeName';\r
+ }\r
+ }\r
+\r
+ private function beforeAttributeValueState()\r
+ {\r
+ // Consume the next input character:\r
+ $this->char++;\r
+ $char = $this->character($this->char);\r
+\r
+ if (preg_match('/^[\t\n\x0b\x0c ]$/', $char)) {\r
+ /* U+0009 CHARACTER TABULATION\r
+ U+000A LINE FEED (LF)\r
+ U+000B LINE TABULATION\r
+ U+000C FORM FEED (FF)\r
+ U+0020 SPACE\r
+ Stay in the before attribute value state. */\r
+ $this->state = 'beforeAttributeValue';\r
+\r
+ } elseif ($char === '"') {\r
+ /* U+0022 QUOTATION MARK (")\r
+ Switch to the attribute value (double-quoted) state. */\r
+ $this->state = 'attributeValueDoubleQuoted';\r
+\r
+ } elseif ($char === '&') {\r
+ /* U+0026 AMPERSAND (&)\r
+ Switch to the attribute value (unquoted) state and reconsume\r
+ this input character. */\r
+ $this->char--;\r
+ $this->state = 'attributeValueUnquoted';\r
+\r
+ } elseif ($char === '\'') {\r
+ /* U+0027 APOSTROPHE (')\r
+ Switch to the attribute value (single-quoted) state. */\r
+ $this->state = 'attributeValueSingleQuoted';\r
+\r
+ } elseif ($char === '>') {\r
+ /* U+003E GREATER-THAN SIGN (>)\r
+ Emit the current tag token. Switch to the data state. */\r
+ $this->emitToken($this->token);\r
+ $this->state = 'data';\r
+\r
+ } else {\r
+ /* Anything else\r
+ Append the current input character to the current attribute's value.\r
+ Switch to the attribute value (unquoted) state. */\r
+ $last = count($this->token['attr']) - 1;\r
+ $this->token['attr'][$last]['value'] .= $char;\r
+\r
+ $this->state = 'attributeValueUnquoted';\r
+ }\r
+ }\r
+\r
+ private function attributeValueDoubleQuotedState()\r
+ {\r
+ // Consume the next input character:\r
+ $this->char++;\r
+ $char = $this->character($this->char);\r
+\r
+ if ($char === '"') {\r
+ /* U+0022 QUOTATION MARK (")\r
+ Switch to the before attribute name state. */\r
+ $this->state = 'beforeAttributeName';\r
+\r
+ } elseif ($char === '&') {\r
+ /* U+0026 AMPERSAND (&)\r
+ Switch to the entity in attribute value state. */\r
+ $this->entityInAttributeValueState('double');\r
+\r
+ } elseif ($this->char === $this->EOF) {\r
+ /* EOF\r
+ Parse error. Emit the current tag token. Reconsume the character\r
+ in the data state. */\r
+ $this->emitToken($this->token);\r
+\r
+ $this->char--;\r
+ $this->state = 'data';\r
+\r
+ } else {\r
+ /* Anything else\r
+ Append the current input character to the current attribute's value.\r
+ Stay in the attribute value (double-quoted) state. */\r
+ $last = count($this->token['attr']) - 1;\r
+ $this->token['attr'][$last]['value'] .= $char;\r
+\r
+ $this->state = 'attributeValueDoubleQuoted';\r
+ }\r
+ }\r
+\r
+ private function attributeValueSingleQuotedState()\r
+ {\r
+ // Consume the next input character:\r
+ $this->char++;\r
+ $char = $this->character($this->char);\r
+\r
+ if ($char === '\'') {\r
+ /* U+0022 QUOTATION MARK (')\r
+ Switch to the before attribute name state. */\r
+ $this->state = 'beforeAttributeName';\r
+\r
+ } elseif ($char === '&') {\r
+ /* U+0026 AMPERSAND (&)\r
+ Switch to the entity in attribute value state. */\r
+ $this->entityInAttributeValueState('single');\r
+\r
+ } elseif ($this->char === $this->EOF) {\r
+ /* EOF\r
+ Parse error. Emit the current tag token. Reconsume the character\r
+ in the data state. */\r
+ $this->emitToken($this->token);\r
+\r
+ $this->char--;\r
+ $this->state = 'data';\r
+\r
+ } else {\r
+ /* Anything else\r
+ Append the current input character to the current attribute's value.\r
+ Stay in the attribute value (single-quoted) state. */\r
+ $last = count($this->token['attr']) - 1;\r
+ $this->token['attr'][$last]['value'] .= $char;\r
+\r
+ $this->state = 'attributeValueSingleQuoted';\r
+ }\r
+ }\r
+\r
+ private function attributeValueUnquotedState()\r
+ {\r
+ // Consume the next input character:\r
+ $this->char++;\r
+ $char = $this->character($this->char);\r
+\r
+ if (preg_match('/^[\t\n\x0b\x0c ]$/', $char)) {\r
+ /* U+0009 CHARACTER TABULATION\r
+ U+000A LINE FEED (LF)\r
+ U+000B LINE TABULATION\r
+ U+000C FORM FEED (FF)\r
+ U+0020 SPACE\r
+ Switch to the before attribute name state. */\r
+ $this->state = 'beforeAttributeName';\r
+\r
+ } elseif ($char === '&') {\r
+ /* U+0026 AMPERSAND (&)\r
+ Switch to the entity in attribute value state. */\r
+ $this->entityInAttributeValueState();\r
+\r
+ } elseif ($char === '>') {\r
+ /* U+003E GREATER-THAN SIGN (>)\r
+ Emit the current tag token. Switch to the data state. */\r
+ $this->emitToken($this->token);\r
+ $this->state = 'data';\r
+\r
+ } else {\r
+ /* Anything else\r
+ Append the current input character to the current attribute's value.\r
+ Stay in the attribute value (unquoted) state. */\r
+ $last = count($this->token['attr']) - 1;\r
+ $this->token['attr'][$last]['value'] .= $char;\r
+\r
+ $this->state = 'attributeValueUnquoted';\r
+ }\r
+ }\r
+\r
+ private function entityInAttributeValueState()\r
+ {\r
+ // Attempt to consume an entity.\r
+ $entity = $this->entity();\r
+\r
+ // If nothing is returned, append a U+0026 AMPERSAND character to the\r
+ // current attribute's value. Otherwise, emit the character token that\r
+ // was returned.\r
+ $char = (!$entity)\r
+ ? '&'\r
+ : $entity;\r
+\r
+ $last = count($this->token['attr']) - 1;\r
+ $this->token['attr'][$last]['value'] .= $char;\r
+ }\r
+\r
+ private function bogusCommentState()\r
+ {\r
+ /* Consume every character up to the first U+003E GREATER-THAN SIGN\r
+ character (>) or the end of the file (EOF), whichever comes first. Emit\r
+ a comment token whose data is the concatenation of all the characters\r
+ starting from and including the character that caused the state machine\r
+ to switch into the bogus comment state, up to and including the last\r
+ consumed character before the U+003E character, if any, or up to the\r
+ end of the file otherwise. (If the comment was started by the end of\r
+ the file (EOF), the token is empty.) */\r
+ $data = $this->characters('^>', $this->char);\r
+ $this->emitToken(\r
+ array(\r
+ 'data' => $data,\r
+ 'type' => self::COMMENT\r
+ )\r
+ );\r
+\r
+ $this->char += strlen($data);\r
+\r
+ /* Switch to the data state. */\r
+ $this->state = 'data';\r
+\r
+ /* If the end of the file was reached, reconsume the EOF character. */\r
+ if ($this->char === $this->EOF) {\r
+ $this->char = $this->EOF - 1;\r
+ }\r
+ }\r
+\r
+ private function markupDeclarationOpenState()\r
+ {\r
+ /* If the next two characters are both U+002D HYPHEN-MINUS (-)\r
+ characters, consume those two characters, create a comment token whose\r
+ data is the empty string, and switch to the comment state. */\r
+ if ($this->character($this->char + 1, 2) === '--') {\r
+ $this->char += 2;\r
+ $this->state = 'comment';\r
+ $this->token = array(\r
+ 'data' => null,\r
+ 'type' => self::COMMENT\r
+ );\r
+\r
+ /* Otherwise if the next seven chacacters are a case-insensitive match\r
+ for the word "DOCTYPE", then consume those characters and switch to the\r
+ DOCTYPE state. */\r
+ } elseif (strtolower($this->character($this->char + 1, 7)) === 'doctype') {\r
+ $this->char += 7;\r
+ $this->state = 'doctype';\r
+\r
+ /* Otherwise, is is a parse error. Switch to the bogus comment state.\r
+ The next character that is consumed, if any, is the first character\r
+ that will be in the comment. */\r
+ } else {\r
+ $this->char++;\r
+ $this->state = 'bogusComment';\r
+ }\r
+ }\r
+\r
+ private function commentState()\r
+ {\r
+ /* Consume the next input character: */\r
+ $this->char++;\r
+ $char = $this->char();\r
+\r
+ /* U+002D HYPHEN-MINUS (-) */\r
+ if ($char === '-') {\r
+ /* Switch to the comment dash state */\r
+ $this->state = 'commentDash';\r
+\r
+ /* EOF */\r
+ } elseif ($this->char === $this->EOF) {\r
+ /* Parse error. Emit the comment token. Reconsume the EOF character\r
+ in the data state. */\r
+ $this->emitToken($this->token);\r
+ $this->char--;\r
+ $this->state = 'data';\r
+\r
+ /* Anything else */\r
+ } else {\r
+ /* Append the input character to the comment token's data. Stay in\r
+ the comment state. */\r
+ $this->token['data'] .= $char;\r
+ }\r
+ }\r
+\r
+ private function commentDashState()\r
+ {\r
+ /* Consume the next input character: */\r
+ $this->char++;\r
+ $char = $this->char();\r
+\r
+ /* U+002D HYPHEN-MINUS (-) */\r
+ if ($char === '-') {\r
+ /* Switch to the comment end state */\r
+ $this->state = 'commentEnd';\r
+\r
+ /* EOF */\r
+ } elseif ($this->char === $this->EOF) {\r
+ /* Parse error. Emit the comment token. Reconsume the EOF character\r
+ in the data state. */\r
+ $this->emitToken($this->token);\r
+ $this->char--;\r
+ $this->state = 'data';\r
+\r
+ /* Anything else */\r
+ } else {\r
+ /* Append a U+002D HYPHEN-MINUS (-) character and the input\r
+ character to the comment token's data. Switch to the comment state. */\r
+ $this->token['data'] .= '-' . $char;\r
+ $this->state = 'comment';\r
+ }\r
+ }\r
+\r
+ private function commentEndState()\r
+ {\r
+ /* Consume the next input character: */\r
+ $this->char++;\r
+ $char = $this->char();\r
+\r
+ if ($char === '>') {\r
+ $this->emitToken($this->token);\r
+ $this->state = 'data';\r
+\r
+ } elseif ($char === '-') {\r
+ $this->token['data'] .= '-';\r
+\r
+ } elseif ($this->char === $this->EOF) {\r
+ $this->emitToken($this->token);\r
+ $this->char--;\r
+ $this->state = 'data';\r
+\r
+ } else {\r
+ $this->token['data'] .= '--' . $char;\r
+ $this->state = 'comment';\r
+ }\r
+ }\r
+\r
+ private function doctypeState()\r
+ {\r
+ /* Consume the next input character: */\r
+ $this->char++;\r
+ $char = $this->char();\r
+\r
+ if (preg_match('/^[\t\n\x0b\x0c ]$/', $char)) {\r
+ $this->state = 'beforeDoctypeName';\r
+\r
+ } else {\r
+ $this->char--;\r
+ $this->state = 'beforeDoctypeName';\r
+ }\r
+ }\r
+\r
+ private function beforeDoctypeNameState()\r
+ {\r
+ /* Consume the next input character: */\r
+ $this->char++;\r
+ $char = $this->char();\r
+\r
+ if (preg_match('/^[\t\n\x0b\x0c ]$/', $char)) {\r
+ // Stay in the before DOCTYPE name state.\r
+\r
+ } elseif (preg_match('/^[a-z]$/', $char)) {\r
+ $this->token = array(\r
+ 'name' => strtoupper($char),\r
+ 'type' => self::DOCTYPE,\r
+ 'error' => true\r
+ );\r
+\r
+ $this->state = 'doctypeName';\r
+\r
+ } elseif ($char === '>') {\r
+ $this->emitToken(\r
+ array(\r
+ 'name' => null,\r
+ 'type' => self::DOCTYPE,\r
+ 'error' => true\r
+ )\r
+ );\r
+\r
+ $this->state = 'data';\r
+\r
+ } elseif ($this->char === $this->EOF) {\r
+ $this->emitToken(\r
+ array(\r
+ 'name' => null,\r
+ 'type' => self::DOCTYPE,\r
+ 'error' => true\r
+ )\r
+ );\r
+\r
+ $this->char--;\r
+ $this->state = 'data';\r
+\r
+ } else {\r
+ $this->token = array(\r
+ 'name' => $char,\r
+ 'type' => self::DOCTYPE,\r
+ 'error' => true\r
+ );\r
+\r
+ $this->state = 'doctypeName';\r
+ }\r
+ }\r
+\r
+ private function doctypeNameState()\r
+ {\r
+ /* Consume the next input character: */\r
+ $this->char++;\r
+ $char = $this->char();\r
+\r
+ if (preg_match('/^[\t\n\x0b\x0c ]$/', $char)) {\r
+ $this->state = 'AfterDoctypeName';\r
+\r
+ } elseif ($char === '>') {\r
+ $this->emitToken($this->token);\r
+ $this->state = 'data';\r
+\r
+ } elseif (preg_match('/^[a-z]$/', $char)) {\r
+ $this->token['name'] .= strtoupper($char);\r
+\r
+ } elseif ($this->char === $this->EOF) {\r
+ $this->emitToken($this->token);\r
+ $this->char--;\r
+ $this->state = 'data';\r
+\r
+ } else {\r
+ $this->token['name'] .= $char;\r
+ }\r
+\r
+ $this->token['error'] = ($this->token['name'] === 'HTML')\r
+ ? false\r
+ : true;\r
+ }\r
+\r
+ private function afterDoctypeNameState()\r
+ {\r
+ /* Consume the next input character: */\r
+ $this->char++;\r
+ $char = $this->char();\r
+\r
+ if (preg_match('/^[\t\n\x0b\x0c ]$/', $char)) {\r
+ // Stay in the DOCTYPE name state.\r
+\r
+ } elseif ($char === '>') {\r
+ $this->emitToken($this->token);\r
+ $this->state = 'data';\r
+\r
+ } elseif ($this->char === $this->EOF) {\r
+ $this->emitToken($this->token);\r
+ $this->char--;\r
+ $this->state = 'data';\r
+\r
+ } else {\r
+ $this->token['error'] = true;\r
+ $this->state = 'bogusDoctype';\r
+ }\r
+ }\r
+\r
+ private function bogusDoctypeState()\r
+ {\r
+ /* Consume the next input character: */\r
+ $this->char++;\r
+ $char = $this->char();\r
+\r
+ if ($char === '>') {\r
+ $this->emitToken($this->token);\r
+ $this->state = 'data';\r
+\r
+ } elseif ($this->char === $this->EOF) {\r
+ $this->emitToken($this->token);\r
+ $this->char--;\r
+ $this->state = 'data';\r
+\r
+ } else {\r
+ // Stay in the bogus DOCTYPE state.\r
+ }\r
+ }\r
+\r
+ private function entity()\r
+ {\r
+ $start = $this->char;\r
+\r
+ // This section defines how to consume an entity. This definition is\r
+ // used when parsing entities in text and in attributes.\r
+\r
+ // The behaviour depends on the identity of the next character (the\r
+ // one immediately after the U+0026 AMPERSAND character):\r
+\r
+ switch ($this->character($this->char + 1)) {\r
+ // U+0023 NUMBER SIGN (#)\r
+ case '#':\r
+\r
+ // The behaviour further depends on the character after the\r
+ // U+0023 NUMBER SIGN:\r
+ switch ($this->character($this->char + 1)) {\r
+ // U+0078 LATIN SMALL LETTER X\r
+ // U+0058 LATIN CAPITAL LETTER X\r
+ case 'x':\r
+ case 'X':\r
+ // Follow the steps below, but using the range of\r
+ // characters U+0030 DIGIT ZERO through to U+0039 DIGIT\r
+ // NINE, U+0061 LATIN SMALL LETTER A through to U+0066\r
+ // LATIN SMALL LETTER F, and U+0041 LATIN CAPITAL LETTER\r
+ // A, through to U+0046 LATIN CAPITAL LETTER F (in other\r
+ // words, 0-9, A-F, a-f).\r
+ $char = 1;\r
+ $char_class = '0-9A-Fa-f';\r
+ break;\r
+\r
+ // Anything else\r
+ default:\r
+ // Follow the steps below, but using the range of\r
+ // characters U+0030 DIGIT ZERO through to U+0039 DIGIT\r
+ // NINE (i.e. just 0-9).\r
+ $char = 0;\r
+ $char_class = '0-9';\r
+ break;\r
+ }\r
+\r
+ // Consume as many characters as match the range of characters\r
+ // given above.\r
+ $this->char++;\r
+ $e_name = $this->characters($char_class, $this->char + $char + 1);\r
+ $entity = $this->character($start, $this->char);\r
+ $cond = strlen($e_name) > 0;\r
+\r
+ // The rest of the parsing happens below.\r
+ break;\r
+\r
+ // Anything else\r
+ default:\r
+ // Consume the maximum number of characters possible, with the\r
+ // consumed characters case-sensitively matching one of the\r
+ // identifiers in the first column of the entities table.\r
+\r
+ $e_name = $this->characters('0-9A-Za-z;', $this->char + 1);\r
+ $len = strlen($e_name);\r
+\r
+ for ($c = 1; $c <= $len; $c++) {\r
+ $id = substr($e_name, 0, $c);\r
+ $this->char++;\r
+\r
+ if (in_array($id, $this->entities)) {\r
+ if ($e_name[$c - 1] !== ';') {\r
+ if ($c < $len && $e_name[$c] == ';') {\r
+ $this->char++; // consume extra semicolon\r
+ }\r
+ }\r
+ $entity = $id;\r
+ break;\r
+ }\r
+ }\r
+\r
+ $cond = isset($entity);\r
+ // The rest of the parsing happens below.\r
+ break;\r
+ }\r
+\r
+ if (!$cond) {\r
+ // If no match can be made, then this is a parse error. No\r
+ // characters are consumed, and nothing is returned.\r
+ $this->char = $start;\r
+ return false;\r
+ }\r
+\r
+ // Return a character token for the character corresponding to the\r
+ // entity name (as given by the second column of the entities table).\r
+ return html_entity_decode('&' . rtrim($entity, ';') . ';', ENT_QUOTES, 'UTF-8');\r
+ }\r
+\r
+ private function emitToken($token)\r
+ {\r
+ $emit = $this->tree->emitToken($token);\r
+\r
+ if (is_int($emit)) {\r
+ $this->content_model = $emit;\r
+\r
+ } elseif ($token['type'] === self::ENDTAG) {\r
+ $this->content_model = self::PCDATA;\r
+ }\r
+ }\r
+\r
+ private function EOF()\r
+ {\r
+ $this->state = null;\r
+ $this->tree->emitToken(\r
+ array(\r
+ 'type' => self::EOF\r
+ )\r
+ );\r
+ }\r
+}\r
+\r
+class HTML5TreeConstructer\r
+{\r
+ public $stack = array();\r
+\r
+ private $phase;\r
+ private $mode;\r
+ private $dom;\r
+ private $foster_parent = null;\r
+ private $a_formatting = array();\r
+\r
+ private $head_pointer = null;\r
+ private $form_pointer = null;\r
+\r
+ private $scoping = array('button', 'caption', 'html', 'marquee', 'object', 'table', 'td', 'th');\r
+ private $formatting = array(\r
+ 'a',\r
+ 'b',\r
+ 'big',\r
+ 'em',\r
+ 'font',\r
+ 'i',\r
+ 'nobr',\r
+ 's',\r
+ 'small',\r
+ 'strike',\r
+ 'strong',\r
+ 'tt',\r
+ 'u'\r
+ );\r
+ private $special = array(\r
+ 'address',\r
+ 'area',\r
+ 'base',\r
+ 'basefont',\r
+ 'bgsound',\r
+ 'blockquote',\r
+ 'body',\r
+ 'br',\r
+ 'center',\r
+ 'col',\r
+ 'colgroup',\r
+ 'dd',\r
+ 'dir',\r
+ 'div',\r
+ 'dl',\r
+ 'dt',\r
+ 'embed',\r
+ 'fieldset',\r
+ 'form',\r
+ 'frame',\r
+ 'frameset',\r
+ 'h1',\r
+ 'h2',\r
+ 'h3',\r
+ 'h4',\r
+ 'h5',\r
+ 'h6',\r
+ 'head',\r
+ 'hr',\r
+ 'iframe',\r
+ 'image',\r
+ 'img',\r
+ 'input',\r
+ 'isindex',\r
+ 'li',\r
+ 'link',\r
+ 'listing',\r
+ 'menu',\r
+ 'meta',\r
+ 'noembed',\r
+ 'noframes',\r
+ 'noscript',\r
+ 'ol',\r
+ 'optgroup',\r
+ 'option',\r
+ 'p',\r
+ 'param',\r
+ 'plaintext',\r
+ 'pre',\r
+ 'script',\r
+ 'select',\r
+ 'spacer',\r
+ 'style',\r
+ 'tbody',\r
+ 'textarea',\r
+ 'tfoot',\r
+ 'thead',\r
+ 'title',\r
+ 'tr',\r
+ 'ul',\r
+ 'wbr'\r
+ );\r
+\r
+ // The different phases.\r
+ const INIT_PHASE = 0;\r
+ const ROOT_PHASE = 1;\r
+ const MAIN_PHASE = 2;\r
+ const END_PHASE = 3;\r
+\r
+ // The different insertion modes for the main phase.\r
+ const BEFOR_HEAD = 0;\r
+ const IN_HEAD = 1;\r
+ const AFTER_HEAD = 2;\r
+ const IN_BODY = 3;\r
+ const IN_TABLE = 4;\r
+ const IN_CAPTION = 5;\r
+ const IN_CGROUP = 6;\r
+ const IN_TBODY = 7;\r
+ const IN_ROW = 8;\r
+ const IN_CELL = 9;\r
+ const IN_SELECT = 10;\r
+ const AFTER_BODY = 11;\r
+ const IN_FRAME = 12;\r
+ const AFTR_FRAME = 13;\r
+\r
+ // The different types of elements.\r
+ const SPECIAL = 0;\r
+ const SCOPING = 1;\r
+ const FORMATTING = 2;\r
+ const PHRASING = 3;\r
+\r
+ const MARKER = 0;\r
+\r
+ public function __construct()\r
+ {\r
+ $this->phase = self::INIT_PHASE;\r
+ $this->mode = self::BEFOR_HEAD;\r
+ $this->dom = new DOMDocument;\r
+\r
+ $this->dom->encoding = 'UTF-8';\r
+ $this->dom->preserveWhiteSpace = true;\r
+ $this->dom->substituteEntities = true;\r
+ $this->dom->strictErrorChecking = false;\r
+ }\r
+\r
+ // Process tag tokens\r
+ public function emitToken($token)\r
+ {\r
+ switch ($this->phase) {\r
+ case self::INIT_PHASE:\r
+ return $this->initPhase($token);\r
+ break;\r
+ case self::ROOT_PHASE:\r
+ return $this->rootElementPhase($token);\r
+ break;\r
+ case self::MAIN_PHASE:\r
+ return $this->mainPhase($token);\r
+ break;\r
+ case self::END_PHASE :\r
+ return $this->trailingEndPhase($token);\r
+ break;\r
+ }\r
+ }\r
+\r
+ private function initPhase($token)\r
+ {\r
+ /* Initially, the tree construction stage must handle each token\r
+ emitted from the tokenisation stage as follows: */\r
+\r
+ /* A DOCTYPE token that is marked as being in error\r
+ A comment token\r
+ A start tag token\r
+ An end tag token\r
+ A character token that is not one of one of U+0009 CHARACTER TABULATION,\r
+ U+000A LINE FEED (LF), U+000B LINE TABULATION, U+000C FORM FEED (FF),\r
+ or U+0020 SPACE\r
+ An end-of-file token */\r
+ if ((isset($token['error']) && $token['error']) ||\r
+ $token['type'] === HTML5::COMMENT ||\r
+ $token['type'] === HTML5::STARTTAG ||\r
+ $token['type'] === HTML5::ENDTAG ||\r
+ $token['type'] === HTML5::EOF ||\r
+ ($token['type'] === HTML5::CHARACTR && isset($token['data']) &&\r
+ !preg_match('/^[\t\n\x0b\x0c ]+$/', $token['data']))\r
+ ) {\r
+ /* This specification does not define how to handle this case. In\r
+ particular, user agents may ignore the entirety of this specification\r
+ altogether for such documents, and instead invoke special parse modes\r
+ with a greater emphasis on backwards compatibility. */\r
+\r
+ $this->phase = self::ROOT_PHASE;\r
+ return $this->rootElementPhase($token);\r
+\r
+ /* A DOCTYPE token marked as being correct */\r
+ } elseif (isset($token['error']) && !$token['error']) {\r
+ /* Append a DocumentType node to the Document node, with the name\r
+ attribute set to the name given in the DOCTYPE token (which will be\r
+ "HTML"), and the other attributes specific to DocumentType objects\r
+ set to null, empty lists, or the empty string as appropriate. */\r
+ $doctype = new DOMDocumentType(null, null, 'HTML');\r
+\r
+ /* Then, switch to the root element phase of the tree construction\r
+ stage. */\r
+ $this->phase = self::ROOT_PHASE;\r
+\r
+ /* A character token that is one of one of U+0009 CHARACTER TABULATION,\r
+ U+000A LINE FEED (LF), U+000B LINE TABULATION, U+000C FORM FEED (FF),\r
+ or U+0020 SPACE */\r
+ } elseif (isset($token['data']) && preg_match(\r
+ '/^[\t\n\x0b\x0c ]+$/',\r
+ $token['data']\r
+ )\r
+ ) {\r
+ /* Append that character to the Document node. */\r
+ $text = $this->dom->createTextNode($token['data']);\r
+ $this->dom->appendChild($text);\r
+ }\r
+ }\r
+\r
+ private function rootElementPhase($token)\r
+ {\r
+ /* After the initial phase, as each token is emitted from the tokenisation\r
+ stage, it must be processed as described in this section. */\r
+\r
+ /* A DOCTYPE token */\r
+ if ($token['type'] === HTML5::DOCTYPE) {\r
+ // Parse error. Ignore the token.\r
+\r
+ /* A comment token */\r
+ } elseif ($token['type'] === HTML5::COMMENT) {\r
+ /* Append a Comment node to the Document object with the data\r
+ attribute set to the data given in the comment token. */\r
+ $comment = $this->dom->createComment($token['data']);\r
+ $this->dom->appendChild($comment);\r
+\r
+ /* A character token that is one of one of U+0009 CHARACTER TABULATION,\r
+ U+000A LINE FEED (LF), U+000B LINE TABULATION, U+000C FORM FEED (FF),\r
+ or U+0020 SPACE */\r
+ } elseif ($token['type'] === HTML5::CHARACTR &&\r
+ preg_match('/^[\t\n\x0b\x0c ]+$/', $token['data'])\r
+ ) {\r
+ /* Append that character to the Document node. */\r
+ $text = $this->dom->createTextNode($token['data']);\r
+ $this->dom->appendChild($text);\r
+\r
+ /* A character token that is not one of U+0009 CHARACTER TABULATION,\r
+ U+000A LINE FEED (LF), U+000B LINE TABULATION, U+000C FORM FEED\r
+ (FF), or U+0020 SPACE\r
+ A start tag token\r
+ An end tag token\r
+ An end-of-file token */\r
+ } elseif (($token['type'] === HTML5::CHARACTR &&\r
+ !preg_match('/^[\t\n\x0b\x0c ]+$/', $token['data'])) ||\r
+ $token['type'] === HTML5::STARTTAG ||\r
+ $token['type'] === HTML5::ENDTAG ||\r
+ $token['type'] === HTML5::EOF\r
+ ) {\r
+ /* Create an HTMLElement node with the tag name html, in the HTML\r
+ namespace. Append it to the Document object. Switch to the main\r
+ phase and reprocess the current token. */\r
+ $html = $this->dom->createElement('html');\r
+ $this->dom->appendChild($html);\r
+ $this->stack[] = $html;\r
+\r
+ $this->phase = self::MAIN_PHASE;\r
+ return $this->mainPhase($token);\r
+ }\r
+ }\r
+\r
+ private function mainPhase($token)\r
+ {\r
+ /* Tokens in the main phase must be handled as follows: */\r
+\r
+ /* A DOCTYPE token */\r
+ if ($token['type'] === HTML5::DOCTYPE) {\r
+ // Parse error. Ignore the token.\r
+\r
+ /* A start tag token with the tag name "html" */\r
+ } elseif ($token['type'] === HTML5::STARTTAG && $token['name'] === 'html') {\r
+ /* If this start tag token was not the first start tag token, then\r
+ it is a parse error. */\r
+\r
+ /* For each attribute on the token, check to see if the attribute\r
+ is already present on the top element of the stack of open elements.\r
+ If it is not, add the attribute and its corresponding value to that\r
+ element. */\r
+ foreach ($token['attr'] as $attr) {\r
+ if (!$this->stack[0]->hasAttribute($attr['name'])) {\r
+ $this->stack[0]->setAttribute($attr['name'], $attr['value']);\r
+ }\r
+ }\r
+\r
+ /* An end-of-file token */\r
+ } elseif ($token['type'] === HTML5::EOF) {\r
+ /* Generate implied end tags. */\r
+ $this->generateImpliedEndTags();\r
+\r
+ /* Anything else. */\r
+ } else {\r
+ /* Depends on the insertion mode: */\r
+ switch ($this->mode) {\r
+ case self::BEFOR_HEAD:\r
+ return $this->beforeHead($token);\r
+ break;\r
+ case self::IN_HEAD:\r
+ return $this->inHead($token);\r
+ break;\r
+ case self::AFTER_HEAD:\r
+ return $this->afterHead($token);\r
+ break;\r
+ case self::IN_BODY:\r
+ return $this->inBody($token);\r
+ break;\r
+ case self::IN_TABLE:\r
+ return $this->inTable($token);\r
+ break;\r
+ case self::IN_CAPTION:\r
+ return $this->inCaption($token);\r
+ break;\r
+ case self::IN_CGROUP:\r
+ return $this->inColumnGroup($token);\r
+ break;\r
+ case self::IN_TBODY:\r
+ return $this->inTableBody($token);\r
+ break;\r
+ case self::IN_ROW:\r
+ return $this->inRow($token);\r
+ break;\r
+ case self::IN_CELL:\r
+ return $this->inCell($token);\r
+ break;\r
+ case self::IN_SELECT:\r
+ return $this->inSelect($token);\r
+ break;\r
+ case self::AFTER_BODY:\r
+ return $this->afterBody($token);\r
+ break;\r
+ case self::IN_FRAME:\r
+ return $this->inFrameset($token);\r
+ break;\r
+ case self::AFTR_FRAME:\r
+ return $this->afterFrameset($token);\r
+ break;\r
+ case self::END_PHASE:\r
+ return $this->trailingEndPhase($token);\r
+ break;\r
+ }\r
+ }\r
+ }\r
+\r
+ private function beforeHead($token)\r
+ {\r
+ /* Handle the token as follows: */\r
+\r
+ /* A character token that is one of one of U+0009 CHARACTER TABULATION,\r
+ U+000A LINE FEED (LF), U+000B LINE TABULATION, U+000C FORM FEED (FF),\r
+ or U+0020 SPACE */\r
+ if ($token['type'] === HTML5::CHARACTR &&\r
+ preg_match('/^[\t\n\x0b\x0c ]+$/', $token['data'])\r
+ ) {\r
+ /* Append the character to the current node. */\r
+ $this->insertText($token['data']);\r
+\r
+ /* A comment token */\r
+ } elseif ($token['type'] === HTML5::COMMENT) {\r
+ /* Append a Comment node to the current node with the data attribute\r
+ set to the data given in the comment token. */\r
+ $this->insertComment($token['data']);\r
+\r
+ /* A start tag token with the tag name "head" */\r
+ } elseif ($token['type'] === HTML5::STARTTAG && $token['name'] === 'head') {\r
+ /* Create an element for the token, append the new element to the\r
+ current node and push it onto the stack of open elements. */\r
+ $element = $this->insertElement($token);\r
+\r
+ /* Set the head element pointer to this new element node. */\r
+ $this->head_pointer = $element;\r
+\r
+ /* Change the insertion mode to "in head". */\r
+ $this->mode = self::IN_HEAD;\r
+\r
+ /* A start tag token whose tag name is one of: "base", "link", "meta",\r
+ "script", "style", "title". Or an end tag with the tag name "html".\r
+ Or a character token that is not one of U+0009 CHARACTER TABULATION,\r
+ U+000A LINE FEED (LF), U+000B LINE TABULATION, U+000C FORM FEED (FF),\r
+ or U+0020 SPACE. Or any other start tag token */\r
+ } elseif ($token['type'] === HTML5::STARTTAG ||\r
+ ($token['type'] === HTML5::ENDTAG && $token['name'] === 'html') ||\r
+ ($token['type'] === HTML5::CHARACTR && !preg_match(\r
+ '/^[\t\n\x0b\x0c ]$/',\r
+ $token['data']\r
+ ))\r
+ ) {\r
+ /* Act as if a start tag token with the tag name "head" and no\r
+ attributes had been seen, then reprocess the current token. */\r
+ $this->beforeHead(\r
+ array(\r
+ 'name' => 'head',\r
+ 'type' => HTML5::STARTTAG,\r
+ 'attr' => array()\r
+ )\r
+ );\r
+\r
+ return $this->inHead($token);\r
+\r
+ /* Any other end tag */\r
+ } elseif ($token['type'] === HTML5::ENDTAG) {\r
+ /* Parse error. Ignore the token. */\r
+ }\r
+ }\r
+\r
+ private function inHead($token)\r
+ {\r
+ /* Handle the token as follows: */\r
+\r
+ /* A character token that is one of one of U+0009 CHARACTER TABULATION,\r
+ U+000A LINE FEED (LF), U+000B LINE TABULATION, U+000C FORM FEED (FF),\r
+ or U+0020 SPACE.\r
+\r
+ THIS DIFFERS FROM THE SPEC: If the current node is either a title, style\r
+ or script element, append the character to the current node regardless\r
+ of its content. */\r
+ if (($token['type'] === HTML5::CHARACTR &&\r
+ preg_match('/^[\t\n\x0b\x0c ]+$/', $token['data'])) || (\r
+ $token['type'] === HTML5::CHARACTR && in_array(\r
+ end($this->stack)->nodeName,\r
+ array('title', 'style', 'script')\r
+ ))\r
+ ) {\r
+ /* Append the character to the current node. */\r
+ $this->insertText($token['data']);\r
+\r
+ /* A comment token */\r
+ } elseif ($token['type'] === HTML5::COMMENT) {\r
+ /* Append a Comment node to the current node with the data attribute\r
+ set to the data given in the comment token. */\r
+ $this->insertComment($token['data']);\r
+\r
+ } elseif ($token['type'] === HTML5::ENDTAG &&\r
+ in_array($token['name'], array('title', 'style', 'script'))\r
+ ) {\r
+ array_pop($this->stack);\r
+ return HTML5::PCDATA;\r
+\r
+ /* A start tag with the tag name "title" */\r
+ } elseif ($token['type'] === HTML5::STARTTAG && $token['name'] === 'title') {\r
+ /* Create an element for the token and append the new element to the\r
+ node pointed to by the head element pointer, or, if that is null\r
+ (innerHTML case), to the current node. */\r
+ if ($this->head_pointer !== null) {\r
+ $element = $this->insertElement($token, false);\r
+ $this->head_pointer->appendChild($element);\r
+\r
+ } else {\r
+ $element = $this->insertElement($token);\r
+ }\r
+\r
+ /* Switch the tokeniser's content model flag to the RCDATA state. */\r
+ return HTML5::RCDATA;\r
+\r
+ /* A start tag with the tag name "style" */\r
+ } elseif ($token['type'] === HTML5::STARTTAG && $token['name'] === 'style') {\r
+ /* Create an element for the token and append the new element to the\r
+ node pointed to by the head element pointer, or, if that is null\r
+ (innerHTML case), to the current node. */\r
+ if ($this->head_pointer !== null) {\r
+ $element = $this->insertElement($token, false);\r
+ $this->head_pointer->appendChild($element);\r
+\r
+ } else {\r
+ $this->insertElement($token);\r
+ }\r
+\r
+ /* Switch the tokeniser's content model flag to the CDATA state. */\r
+ return HTML5::CDATA;\r
+\r
+ /* A start tag with the tag name "script" */\r
+ } elseif ($token['type'] === HTML5::STARTTAG && $token['name'] === 'script') {\r
+ /* Create an element for the token. */\r
+ $element = $this->insertElement($token, false);\r
+ $this->head_pointer->appendChild($element);\r
+\r
+ /* Switch the tokeniser's content model flag to the CDATA state. */\r
+ return HTML5::CDATA;\r
+\r
+ /* A start tag with the tag name "base", "link", or "meta" */\r
+ } elseif ($token['type'] === HTML5::STARTTAG && in_array(\r
+ $token['name'],\r
+ array('base', 'link', 'meta')\r
+ )\r
+ ) {\r
+ /* Create an element for the token and append the new element to the\r
+ node pointed to by the head element pointer, or, if that is null\r
+ (innerHTML case), to the current node. */\r
+ if ($this->head_pointer !== null) {\r
+ $element = $this->insertElement($token, false);\r
+ $this->head_pointer->appendChild($element);\r
+ array_pop($this->stack);\r
+\r
+ } else {\r
+ $this->insertElement($token);\r
+ }\r
+\r
+ /* An end tag with the tag name "head" */\r
+ } elseif ($token['type'] === HTML5::ENDTAG && $token['name'] === 'head') {\r
+ /* If the current node is a head element, pop the current node off\r
+ the stack of open elements. */\r
+ if ($this->head_pointer->isSameNode(end($this->stack))) {\r
+ array_pop($this->stack);\r
+\r
+ /* Otherwise, this is a parse error. */\r
+ } else {\r
+ // k\r
+ }\r
+\r
+ /* Change the insertion mode to "after head". */\r
+ $this->mode = self::AFTER_HEAD;\r
+\r
+ /* A start tag with the tag name "head" or an end tag except "html". */\r
+ } elseif (($token['type'] === HTML5::STARTTAG && $token['name'] === 'head') ||\r
+ ($token['type'] === HTML5::ENDTAG && $token['name'] !== 'html')\r
+ ) {\r
+ // Parse error. Ignore the token.\r
+\r
+ /* Anything else */\r
+ } else {\r
+ /* If the current node is a head element, act as if an end tag\r
+ token with the tag name "head" had been seen. */\r
+ if ($this->head_pointer->isSameNode(end($this->stack))) {\r
+ $this->inHead(\r
+ array(\r
+ 'name' => 'head',\r
+ 'type' => HTML5::ENDTAG\r
+ )\r
+ );\r
+\r
+ /* Otherwise, change the insertion mode to "after head". */\r
+ } else {\r
+ $this->mode = self::AFTER_HEAD;\r
+ }\r
+\r
+ /* Then, reprocess the current token. */\r
+ return $this->afterHead($token);\r
+ }\r
+ }\r
+\r
+ private function afterHead($token)\r
+ {\r
+ /* Handle the token as follows: */\r
+\r
+ /* A character token that is one of one of U+0009 CHARACTER TABULATION,\r
+ U+000A LINE FEED (LF), U+000B LINE TABULATION, U+000C FORM FEED (FF),\r
+ or U+0020 SPACE */\r
+ if ($token['type'] === HTML5::CHARACTR &&\r
+ preg_match('/^[\t\n\x0b\x0c ]+$/', $token['data'])\r
+ ) {\r
+ /* Append the character to the current node. */\r
+ $this->insertText($token['data']);\r
+\r
+ /* A comment token */\r
+ } elseif ($token['type'] === HTML5::COMMENT) {\r
+ /* Append a Comment node to the current node with the data attribute\r
+ set to the data given in the comment token. */\r
+ $this->insertComment($token['data']);\r
+\r
+ /* A start tag token with the tag name "body" */\r
+ } elseif ($token['type'] === HTML5::STARTTAG && $token['name'] === 'body') {\r
+ /* Insert a body element for the token. */\r
+ $this->insertElement($token);\r
+\r
+ /* Change the insertion mode to "in body". */\r
+ $this->mode = self::IN_BODY;\r
+\r
+ /* A start tag token with the tag name "frameset" */\r
+ } elseif ($token['type'] === HTML5::STARTTAG && $token['name'] === 'frameset') {\r
+ /* Insert a frameset element for the token. */\r
+ $this->insertElement($token);\r
+\r
+ /* Change the insertion mode to "in frameset". */\r
+ $this->mode = self::IN_FRAME;\r
+\r
+ /* A start tag token whose tag name is one of: "base", "link", "meta",\r
+ "script", "style", "title" */\r
+ } elseif ($token['type'] === HTML5::STARTTAG && in_array(\r
+ $token['name'],\r
+ array('base', 'link', 'meta', 'script', 'style', 'title')\r
+ )\r
+ ) {\r
+ /* Parse error. Switch the insertion mode back to "in head" and\r
+ reprocess the token. */\r
+ $this->mode = self::IN_HEAD;\r
+ return $this->inHead($token);\r
+\r
+ /* Anything else */\r
+ } else {\r
+ /* Act as if a start tag token with the tag name "body" and no\r
+ attributes had been seen, and then reprocess the current token. */\r
+ $this->afterHead(\r
+ array(\r
+ 'name' => 'body',\r
+ 'type' => HTML5::STARTTAG,\r
+ 'attr' => array()\r
+ )\r
+ );\r
+\r
+ return $this->inBody($token);\r
+ }\r
+ }\r
+\r
+ private function inBody($token)\r
+ {\r
+ /* Handle the token as follows: */\r
+\r
+ switch ($token['type']) {\r
+ /* A character token */\r
+ case HTML5::CHARACTR:\r
+ /* Reconstruct the active formatting elements, if any. */\r
+ $this->reconstructActiveFormattingElements();\r
+\r
+ /* Append the token's character to the current node. */\r
+ $this->insertText($token['data']);\r
+ break;\r
+\r
+ /* A comment token */\r
+ case HTML5::COMMENT:\r
+ /* Append a Comment node to the current node with the data\r
+ attribute set to the data given in the comment token. */\r
+ $this->insertComment($token['data']);\r
+ break;\r
+\r
+ case HTML5::STARTTAG:\r
+ switch ($token['name']) {\r
+ /* A start tag token whose tag name is one of: "script",\r
+ "style" */\r
+ case 'script':\r
+ case 'style':\r
+ /* Process the token as if the insertion mode had been "in\r
+ head". */\r
+ return $this->inHead($token);\r
+ break;\r
+\r
+ /* A start tag token whose tag name is one of: "base", "link",\r
+ "meta", "title" */\r
+ case 'base':\r
+ case 'link':\r
+ case 'meta':\r
+ case 'title':\r
+ /* Parse error. Process the token as if the insertion mode\r
+ had been "in head". */\r
+ return $this->inHead($token);\r
+ break;\r
+\r
+ /* A start tag token with the tag name "body" */\r
+ case 'body':\r
+ /* Parse error. If the second element on the stack of open\r
+ elements is not a body element, or, if the stack of open\r
+ elements has only one node on it, then ignore the token.\r
+ (innerHTML case) */\r
+ if (count($this->stack) === 1 || $this->stack[1]->nodeName !== 'body') {\r
+ // Ignore\r
+\r
+ /* Otherwise, for each attribute on the token, check to see\r
+ if the attribute is already present on the body element (the\r
+ second element) on the stack of open elements. If it is not,\r
+ add the attribute and its corresponding value to that\r
+ element. */\r
+ } else {\r
+ foreach ($token['attr'] as $attr) {\r
+ if (!$this->stack[1]->hasAttribute($attr['name'])) {\r
+ $this->stack[1]->setAttribute($attr['name'], $attr['value']);\r
+ }\r
+ }\r
+ }\r
+ break;\r
+\r
+ /* A start tag whose tag name is one of: "address",\r
+ "blockquote", "center", "dir", "div", "dl", "fieldset",\r
+ "listing", "menu", "ol", "p", "ul" */\r
+ case 'address':\r
+ case 'blockquote':\r
+ case 'center':\r
+ case 'dir':\r
+ case 'div':\r
+ case 'dl':\r
+ case 'fieldset':\r
+ case 'listing':\r
+ case 'menu':\r
+ case 'ol':\r
+ case 'p':\r
+ case 'ul':\r
+ /* If the stack of open elements has a p element in scope,\r
+ then act as if an end tag with the tag name p had been\r
+ seen. */\r
+ if ($this->elementInScope('p')) {\r
+ $this->emitToken(\r
+ array(\r
+ 'name' => 'p',\r
+ 'type' => HTML5::ENDTAG\r
+ )\r
+ );\r
+ }\r
+\r
+ /* Insert an HTML element for the token. */\r
+ $this->insertElement($token);\r
+ break;\r
+\r
+ /* A start tag whose tag name is "form" */\r
+ case 'form':\r
+ /* If the form element pointer is not null, ignore the\r
+ token with a parse error. */\r
+ if ($this->form_pointer !== null) {\r
+ // Ignore.\r
+\r
+ /* Otherwise: */\r
+ } else {\r
+ /* If the stack of open elements has a p element in\r
+ scope, then act as if an end tag with the tag name p\r
+ had been seen. */\r
+ if ($this->elementInScope('p')) {\r
+ $this->emitToken(\r
+ array(\r
+ 'name' => 'p',\r
+ 'type' => HTML5::ENDTAG\r
+ )\r
+ );\r
+ }\r
+\r
+ /* Insert an HTML element for the token, and set the\r
+ form element pointer to point to the element created. */\r
+ $element = $this->insertElement($token);\r
+ $this->form_pointer = $element;\r
+ }\r
+ break;\r
+\r
+ /* A start tag whose tag name is "li", "dd" or "dt" */\r
+ case 'li':\r
+ case 'dd':\r
+ case 'dt':\r
+ /* If the stack of open elements has a p element in scope,\r
+ then act as if an end tag with the tag name p had been\r
+ seen. */\r
+ if ($this->elementInScope('p')) {\r
+ $this->emitToken(\r
+ array(\r
+ 'name' => 'p',\r
+ 'type' => HTML5::ENDTAG\r
+ )\r
+ );\r
+ }\r
+\r
+ $stack_length = count($this->stack) - 1;\r
+\r
+ for ($n = $stack_length; 0 <= $n; $n--) {\r
+ /* 1. Initialise node to be the current node (the\r
+ bottommost node of the stack). */\r
+ $stop = false;\r
+ $node = $this->stack[$n];\r
+ $cat = $this->getElementCategory($node->tagName);\r
+\r
+ /* 2. If node is an li, dd or dt element, then pop all\r
+ the nodes from the current node up to node, including\r
+ node, then stop this algorithm. */\r
+ if ($token['name'] === $node->tagName || ($token['name'] !== 'li'\r
+ && ($node->tagName === 'dd' || $node->tagName === 'dt'))\r
+ ) {\r
+ for ($x = $stack_length; $x >= $n; $x--) {\r
+ array_pop($this->stack);\r
+ }\r
+\r
+ break;\r
+ }\r
+\r
+ /* 3. If node is not in the formatting category, and is\r
+ not in the phrasing category, and is not an address or\r
+ div element, then stop this algorithm. */\r
+ if ($cat !== self::FORMATTING && $cat !== self::PHRASING &&\r
+ $node->tagName !== 'address' && $node->tagName !== 'div'\r
+ ) {\r
+ break;\r
+ }\r
+ }\r
+\r
+ /* Finally, insert an HTML element with the same tag\r
+ name as the token's. */\r
+ $this->insertElement($token);\r
+ break;\r
+\r
+ /* A start tag token whose tag name is "plaintext" */\r
+ case 'plaintext':\r
+ /* If the stack of open elements has a p element in scope,\r
+ then act as if an end tag with the tag name p had been\r
+ seen. */\r
+ if ($this->elementInScope('p')) {\r
+ $this->emitToken(\r
+ array(\r
+ 'name' => 'p',\r
+ 'type' => HTML5::ENDTAG\r
+ )\r
+ );\r
+ }\r
+\r
+ /* Insert an HTML element for the token. */\r
+ $this->insertElement($token);\r
+\r
+ return HTML5::PLAINTEXT;\r
+ break;\r
+\r
+ /* A start tag whose tag name is one of: "h1", "h2", "h3", "h4",\r
+ "h5", "h6" */\r
+ case 'h1':\r
+ case 'h2':\r
+ case 'h3':\r
+ case 'h4':\r
+ case 'h5':\r
+ case 'h6':\r
+ /* If the stack of open elements has a p element in scope,\r
+ then act as if an end tag with the tag name p had been seen. */\r
+ if ($this->elementInScope('p')) {\r
+ $this->emitToken(\r
+ array(\r
+ 'name' => 'p',\r
+ 'type' => HTML5::ENDTAG\r
+ )\r
+ );\r
+ }\r
+\r
+ /* If the stack of open elements has in scope an element whose\r
+ tag name is one of "h1", "h2", "h3", "h4", "h5", or "h6", then\r
+ this is a parse error; pop elements from the stack until an\r
+ element with one of those tag names has been popped from the\r
+ stack. */\r
+ while ($this->elementInScope(array('h1', 'h2', 'h3', 'h4', 'h5', 'h6'))) {\r
+ array_pop($this->stack);\r
+ }\r
+\r
+ /* Insert an HTML element for the token. */\r
+ $this->insertElement($token);\r
+ break;\r
+\r
+ /* A start tag whose tag name is "a" */\r
+ case 'a':\r
+ /* If the list of active formatting elements contains\r
+ an element whose tag name is "a" between the end of the\r
+ list and the last marker on the list (or the start of\r
+ the list if there is no marker on the list), then this\r
+ is a parse error; act as if an end tag with the tag name\r
+ "a" had been seen, then remove that element from the list\r
+ of active formatting elements and the stack of open\r
+ elements if the end tag didn't already remove it (it\r
+ might not have if the element is not in table scope). */\r
+ $leng = count($this->a_formatting);\r
+\r
+ for ($n = $leng - 1; $n >= 0; $n--) {\r
+ if ($this->a_formatting[$n] === self::MARKER) {\r
+ break;\r
+\r
+ } elseif ($this->a_formatting[$n]->nodeName === 'a') {\r
+ $this->emitToken(\r
+ array(\r
+ 'name' => 'a',\r
+ 'type' => HTML5::ENDTAG\r
+ )\r
+ );\r
+ break;\r
+ }\r
+ }\r
+\r
+ /* Reconstruct the active formatting elements, if any. */\r
+ $this->reconstructActiveFormattingElements();\r
+\r
+ /* Insert an HTML element for the token. */\r
+ $el = $this->insertElement($token);\r
+\r
+ /* Add that element to the list of active formatting\r
+ elements. */\r
+ $this->a_formatting[] = $el;\r
+ break;\r
+\r
+ /* A start tag whose tag name is one of: "b", "big", "em", "font",\r
+ "i", "nobr", "s", "small", "strike", "strong", "tt", "u" */\r
+ case 'b':\r
+ case 'big':\r
+ case 'em':\r
+ case 'font':\r
+ case 'i':\r
+ case 'nobr':\r
+ case 's':\r
+ case 'small':\r
+ case 'strike':\r
+ case 'strong':\r
+ case 'tt':\r
+ case 'u':\r
+ /* Reconstruct the active formatting elements, if any. */\r
+ $this->reconstructActiveFormattingElements();\r
+\r
+ /* Insert an HTML element for the token. */\r
+ $el = $this->insertElement($token);\r
+\r
+ /* Add that element to the list of active formatting\r
+ elements. */\r
+ $this->a_formatting[] = $el;\r
+ break;\r
+\r
+ /* A start tag token whose tag name is "button" */\r
+ case 'button':\r
+ /* If the stack of open elements has a button element in scope,\r
+ then this is a parse error; act as if an end tag with the tag\r
+ name "button" had been seen, then reprocess the token. (We don't\r
+ do that. Unnecessary.) */\r
+ if ($this->elementInScope('button')) {\r
+ $this->inBody(\r
+ array(\r
+ 'name' => 'button',\r
+ 'type' => HTML5::ENDTAG\r
+ )\r
+ );\r
+ }\r
+\r
+ /* Reconstruct the active formatting elements, if any. */\r
+ $this->reconstructActiveFormattingElements();\r
+\r
+ /* Insert an HTML element for the token. */\r
+ $this->insertElement($token);\r
+\r
+ /* Insert a marker at the end of the list of active\r
+ formatting elements. */\r
+ $this->a_formatting[] = self::MARKER;\r
+ break;\r
+\r
+ /* A start tag token whose tag name is one of: "marquee", "object" */\r
+ case 'marquee':\r
+ case 'object':\r
+ /* Reconstruct the active formatting elements, if any. */\r
+ $this->reconstructActiveFormattingElements();\r
+\r
+ /* Insert an HTML element for the token. */\r
+ $this->insertElement($token);\r
+\r
+ /* Insert a marker at the end of the list of active\r
+ formatting elements. */\r
+ $this->a_formatting[] = self::MARKER;\r
+ break;\r
+\r
+ /* A start tag token whose tag name is "xmp" */\r
+ case 'xmp':\r
+ /* Reconstruct the active formatting elements, if any. */\r
+ $this->reconstructActiveFormattingElements();\r
+\r
+ /* Insert an HTML element for the token. */\r
+ $this->insertElement($token);\r
+\r
+ /* Switch the content model flag to the CDATA state. */\r
+ return HTML5::CDATA;\r
+ break;\r
+\r
+ /* A start tag whose tag name is "table" */\r
+ case 'table':\r
+ /* If the stack of open elements has a p element in scope,\r
+ then act as if an end tag with the tag name p had been seen. */\r
+ if ($this->elementInScope('p')) {\r
+ $this->emitToken(\r
+ array(\r
+ 'name' => 'p',\r
+ 'type' => HTML5::ENDTAG\r
+ )\r
+ );\r
+ }\r
+\r
+ /* Insert an HTML element for the token. */\r
+ $this->insertElement($token);\r
+\r
+ /* Change the insertion mode to "in table". */\r
+ $this->mode = self::IN_TABLE;\r
+ break;\r
+\r
+ /* A start tag whose tag name is one of: "area", "basefont",\r
+ "bgsound", "br", "embed", "img", "param", "spacer", "wbr" */\r
+ case 'area':\r
+ case 'basefont':\r
+ case 'bgsound':\r
+ case 'br':\r
+ case 'embed':\r
+ case 'img':\r
+ case 'param':\r
+ case 'spacer':\r
+ case 'wbr':\r
+ /* Reconstruct the active formatting elements, if any. */\r
+ $this->reconstructActiveFormattingElements();\r
+\r
+ /* Insert an HTML element for the token. */\r
+ $this->insertElement($token);\r
+\r
+ /* Immediately pop the current node off the stack of open elements. */\r
+ array_pop($this->stack);\r
+ break;\r
+\r
+ /* A start tag whose tag name is "hr" */\r
+ case 'hr':\r
+ /* If the stack of open elements has a p element in scope,\r
+ then act as if an end tag with the tag name p had been seen. */\r
+ if ($this->elementInScope('p')) {\r
+ $this->emitToken(\r
+ array(\r
+ 'name' => 'p',\r
+ 'type' => HTML5::ENDTAG\r
+ )\r
+ );\r
+ }\r
+\r
+ /* Insert an HTML element for the token. */\r
+ $this->insertElement($token);\r
+\r
+ /* Immediately pop the current node off the stack of open elements. */\r
+ array_pop($this->stack);\r
+ break;\r
+\r
+ /* A start tag whose tag name is "image" */\r
+ case 'image':\r
+ /* Parse error. Change the token's tag name to "img" and\r
+ reprocess it. (Don't ask.) */\r
+ $token['name'] = 'img';\r
+ return $this->inBody($token);\r
+ break;\r
+\r
+ /* A start tag whose tag name is "input" */\r
+ case 'input':\r
+ /* Reconstruct the active formatting elements, if any. */\r
+ $this->reconstructActiveFormattingElements();\r
+\r
+ /* Insert an input element for the token. */\r
+ $element = $this->insertElement($token, false);\r
+\r
+ /* If the form element pointer is not null, then associate the\r
+ input element with the form element pointed to by the form\r
+ element pointer. */\r
+ $this->form_pointer !== null\r
+ ? $this->form_pointer->appendChild($element)\r
+ : end($this->stack)->appendChild($element);\r
+\r
+ /* Pop that input element off the stack of open elements. */\r
+ array_pop($this->stack);\r
+ break;\r
+\r
+ /* A start tag whose tag name is "isindex" */\r
+ case 'isindex':\r
+ /* Parse error. */\r
+ // w/e\r
+\r
+ /* If the form element pointer is not null,\r
+ then ignore the token. */\r
+ if ($this->form_pointer === null) {\r
+ /* Act as if a start tag token with the tag name "form" had\r
+ been seen. */\r
+ $this->inBody(\r
+ array(\r
+ 'name' => 'body',\r
+ 'type' => HTML5::STARTTAG,\r
+ 'attr' => array()\r
+ )\r
+ );\r
+\r
+ /* Act as if a start tag token with the tag name "hr" had\r
+ been seen. */\r
+ $this->inBody(\r
+ array(\r
+ 'name' => 'hr',\r
+ 'type' => HTML5::STARTTAG,\r
+ 'attr' => array()\r
+ )\r
+ );\r
+\r
+ /* Act as if a start tag token with the tag name "p" had\r
+ been seen. */\r
+ $this->inBody(\r
+ array(\r
+ 'name' => 'p',\r
+ 'type' => HTML5::STARTTAG,\r
+ 'attr' => array()\r
+ )\r
+ );\r
+\r
+ /* Act as if a start tag token with the tag name "label"\r
+ had been seen. */\r
+ $this->inBody(\r
+ array(\r
+ 'name' => 'label',\r
+ 'type' => HTML5::STARTTAG,\r
+ 'attr' => array()\r
+ )\r
+ );\r
+\r
+ /* Act as if a stream of character tokens had been seen. */\r
+ $this->insertText(\r
+ 'This is a searchable index. ' .\r
+ 'Insert your search keywords here: '\r
+ );\r
+\r
+ /* Act as if a start tag token with the tag name "input"\r
+ had been seen, with all the attributes from the "isindex"\r
+ token, except with the "name" attribute set to the value\r
+ "isindex" (ignoring any explicit "name" attribute). */\r
+ $attr = $token['attr'];\r
+ $attr[] = array('name' => 'name', 'value' => 'isindex');\r
+\r
+ $this->inBody(\r
+ array(\r
+ 'name' => 'input',\r
+ 'type' => HTML5::STARTTAG,\r
+ 'attr' => $attr\r
+ )\r
+ );\r
+\r
+ /* Act as if a stream of character tokens had been seen\r
+ (see below for what they should say). */\r
+ $this->insertText(\r
+ 'This is a searchable index. ' .\r
+ 'Insert your search keywords here: '\r
+ );\r
+\r
+ /* Act as if an end tag token with the tag name "label"\r
+ had been seen. */\r
+ $this->inBody(\r
+ array(\r
+ 'name' => 'label',\r
+ 'type' => HTML5::ENDTAG\r
+ )\r
+ );\r
+\r
+ /* Act as if an end tag token with the tag name "p" had\r
+ been seen. */\r
+ $this->inBody(\r
+ array(\r
+ 'name' => 'p',\r
+ 'type' => HTML5::ENDTAG\r
+ )\r
+ );\r
+\r
+ /* Act as if a start tag token with the tag name "hr" had\r
+ been seen. */\r
+ $this->inBody(\r
+ array(\r
+ 'name' => 'hr',\r
+ 'type' => HTML5::ENDTAG\r
+ )\r
+ );\r
+\r
+ /* Act as if an end tag token with the tag name "form" had\r
+ been seen. */\r
+ $this->inBody(\r
+ array(\r
+ 'name' => 'form',\r
+ 'type' => HTML5::ENDTAG\r
+ )\r
+ );\r
+ }\r
+ break;\r
+\r
+ /* A start tag whose tag name is "textarea" */\r
+ case 'textarea':\r
+ $this->insertElement($token);\r
+\r
+ /* Switch the tokeniser's content model flag to the\r
+ RCDATA state. */\r
+ return HTML5::RCDATA;\r
+ break;\r
+\r
+ /* A start tag whose tag name is one of: "iframe", "noembed",\r
+ "noframes" */\r
+ case 'iframe':\r
+ case 'noembed':\r
+ case 'noframes':\r
+ $this->insertElement($token);\r
+\r
+ /* Switch the tokeniser's content model flag to the CDATA state. */\r
+ return HTML5::CDATA;\r
+ break;\r
+\r
+ /* A start tag whose tag name is "select" */\r
+ case 'select':\r
+ /* Reconstruct the active formatting elements, if any. */\r
+ $this->reconstructActiveFormattingElements();\r
+\r
+ /* Insert an HTML element for the token. */\r
+ $this->insertElement($token);\r
+\r
+ /* Change the insertion mode to "in select". */\r
+ $this->mode = self::IN_SELECT;\r
+ break;\r
+\r
+ /* A start or end tag whose tag name is one of: "caption", "col",\r
+ "colgroup", "frame", "frameset", "head", "option", "optgroup",\r
+ "tbody", "td", "tfoot", "th", "thead", "tr". */\r
+ case 'caption':\r
+ case 'col':\r
+ case 'colgroup':\r
+ case 'frame':\r
+ case 'frameset':\r
+ case 'head':\r
+ case 'option':\r
+ case 'optgroup':\r
+ case 'tbody':\r
+ case 'td':\r
+ case 'tfoot':\r
+ case 'th':\r
+ case 'thead':\r
+ case 'tr':\r
+ // Parse error. Ignore the token.\r
+ break;\r
+\r
+ /* A start or end tag whose tag name is one of: "event-source",\r
+ "section", "nav", "article", "aside", "header", "footer",\r
+ "datagrid", "command" */\r
+ case 'event-source':\r
+ case 'section':\r
+ case 'nav':\r
+ case 'article':\r
+ case 'aside':\r
+ case 'header':\r
+ case 'footer':\r
+ case 'datagrid':\r
+ case 'command':\r
+ // Work in progress!\r
+ break;\r
+\r
+ /* A start tag token not covered by the previous entries */\r
+ default:\r
+ /* Reconstruct the active formatting elements, if any. */\r
+ $this->reconstructActiveFormattingElements();\r
+\r
+ $this->insertElement($token, true, true);\r
+ break;\r
+ }\r
+ break;\r
+\r
+ case HTML5::ENDTAG:\r
+ switch ($token['name']) {\r
+ /* An end tag with the tag name "body" */\r
+ case 'body':\r
+ /* If the second element in the stack of open elements is\r
+ not a body element, this is a parse error. Ignore the token.\r
+ (innerHTML case) */\r
+ if (count($this->stack) < 2 || $this->stack[1]->nodeName !== 'body') {\r
+ // Ignore.\r
+\r
+ /* If the current node is not the body element, then this\r
+ is a parse error. */\r
+ } elseif (end($this->stack)->nodeName !== 'body') {\r
+ // Parse error.\r
+ }\r
+\r
+ /* Change the insertion mode to "after body". */\r
+ $this->mode = self::AFTER_BODY;\r
+ break;\r
+\r
+ /* An end tag with the tag name "html" */\r
+ case 'html':\r
+ /* Act as if an end tag with tag name "body" had been seen,\r
+ then, if that token wasn't ignored, reprocess the current\r
+ token. */\r
+ $this->inBody(\r
+ array(\r
+ 'name' => 'body',\r
+ 'type' => HTML5::ENDTAG\r
+ )\r
+ );\r
+\r
+ return $this->afterBody($token);\r
+ break;\r
+\r
+ /* An end tag whose tag name is one of: "address", "blockquote",\r
+ "center", "dir", "div", "dl", "fieldset", "listing", "menu",\r
+ "ol", "pre", "ul" */\r
+ case 'address':\r
+ case 'blockquote':\r
+ case 'center':\r
+ case 'dir':\r
+ case 'div':\r
+ case 'dl':\r
+ case 'fieldset':\r
+ case 'listing':\r
+ case 'menu':\r
+ case 'ol':\r
+ case 'pre':\r
+ case 'ul':\r
+ /* If the stack of open elements has an element in scope\r
+ with the same tag name as that of the token, then generate\r
+ implied end tags. */\r
+ if ($this->elementInScope($token['name'])) {\r
+ $this->generateImpliedEndTags();\r
+\r
+ /* Now, if the current node is not an element with\r
+ the same tag name as that of the token, then this\r
+ is a parse error. */\r
+ // w/e\r
+\r
+ /* If the stack of open elements has an element in\r
+ scope with the same tag name as that of the token,\r
+ then pop elements from this stack until an element\r
+ with that tag name has been popped from the stack. */\r
+ for ($n = count($this->stack) - 1; $n >= 0; $n--) {\r
+ if ($this->stack[$n]->nodeName === $token['name']) {\r
+ $n = -1;\r
+ }\r
+\r
+ array_pop($this->stack);\r
+ }\r
+ }\r
+ break;\r
+\r
+ /* An end tag whose tag name is "form" */\r
+ case 'form':\r
+ /* If the stack of open elements has an element in scope\r
+ with the same tag name as that of the token, then generate\r
+ implied end tags. */\r
+ if ($this->elementInScope($token['name'])) {\r
+ $this->generateImpliedEndTags();\r
+\r
+ }\r
+\r
+ if (end($this->stack)->nodeName !== $token['name']) {\r
+ /* Now, if the current node is not an element with the\r
+ same tag name as that of the token, then this is a parse\r
+ error. */\r
+ // w/e\r
+\r
+ } else {\r
+ /* Otherwise, if the current node is an element with\r
+ the same tag name as that of the token pop that element\r
+ from the stack. */\r
+ array_pop($this->stack);\r
+ }\r
+\r
+ /* In any case, set the form element pointer to null. */\r
+ $this->form_pointer = null;\r
+ break;\r
+\r
+ /* An end tag whose tag name is "p" */\r
+ case 'p':\r
+ /* If the stack of open elements has a p element in scope,\r
+ then generate implied end tags, except for p elements. */\r
+ if ($this->elementInScope('p')) {\r
+ $this->generateImpliedEndTags(array('p'));\r
+\r
+ /* If the current node is not a p element, then this is\r
+ a parse error. */\r
+ // k\r
+\r
+ /* If the stack of open elements has a p element in\r
+ scope, then pop elements from this stack until the stack\r
+ no longer has a p element in scope. */\r
+ for ($n = count($this->stack) - 1; $n >= 0; $n--) {\r
+ if ($this->elementInScope('p')) {\r
+ array_pop($this->stack);\r
+\r
+ } else {\r
+ break;\r
+ }\r
+ }\r
+ }\r
+ break;\r
+\r
+ /* An end tag whose tag name is "dd", "dt", or "li" */\r
+ case 'dd':\r
+ case 'dt':\r
+ case 'li':\r
+ /* If the stack of open elements has an element in scope\r
+ whose tag name matches the tag name of the token, then\r
+ generate implied end tags, except for elements with the\r
+ same tag name as the token. */\r
+ if ($this->elementInScope($token['name'])) {\r
+ $this->generateImpliedEndTags(array($token['name']));\r
+\r
+ /* If the current node is not an element with the same\r
+ tag name as the token, then this is a parse error. */\r
+ // w/e\r
+\r
+ /* If the stack of open elements has an element in scope\r
+ whose tag name matches the tag name of the token, then\r
+ pop elements from this stack until an element with that\r
+ tag name has been popped from the stack. */\r
+ for ($n = count($this->stack) - 1; $n >= 0; $n--) {\r
+ if ($this->stack[$n]->nodeName === $token['name']) {\r
+ $n = -1;\r
+ }\r
+\r
+ array_pop($this->stack);\r
+ }\r
+ }\r
+ break;\r
+\r
+ /* An end tag whose tag name is one of: "h1", "h2", "h3", "h4",\r
+ "h5", "h6" */\r
+ case 'h1':\r
+ case 'h2':\r
+ case 'h3':\r
+ case 'h4':\r
+ case 'h5':\r
+ case 'h6':\r
+ $elements = array('h1', 'h2', 'h3', 'h4', 'h5', 'h6');\r
+\r
+ /* If the stack of open elements has in scope an element whose\r
+ tag name is one of "h1", "h2", "h3", "h4", "h5", or "h6", then\r
+ generate implied end tags. */\r
+ if ($this->elementInScope($elements)) {\r
+ $this->generateImpliedEndTags();\r
+\r
+ /* Now, if the current node is not an element with the same\r
+ tag name as that of the token, then this is a parse error. */\r
+ // w/e\r
+\r
+ /* If the stack of open elements has in scope an element\r
+ whose tag name is one of "h1", "h2", "h3", "h4", "h5", or\r
+ "h6", then pop elements from the stack until an element\r
+ with one of those tag names has been popped from the stack. */\r
+ while ($this->elementInScope($elements)) {\r
+ array_pop($this->stack);\r
+ }\r
+ }\r
+ break;\r
+\r
+ /* An end tag whose tag name is one of: "a", "b", "big", "em",\r
+ "font", "i", "nobr", "s", "small", "strike", "strong", "tt", "u" */\r
+ case 'a':\r
+ case 'b':\r
+ case 'big':\r
+ case 'em':\r
+ case 'font':\r
+ case 'i':\r
+ case 'nobr':\r
+ case 's':\r
+ case 'small':\r
+ case 'strike':\r
+ case 'strong':\r
+ case 'tt':\r
+ case 'u':\r
+ /* 1. Let the formatting element be the last element in\r
+ the list of active formatting elements that:\r
+ * is between the end of the list and the last scope\r
+ marker in the list, if any, or the start of the list\r
+ otherwise, and\r
+ * has the same tag name as the token.\r
+ */\r
+ while (true) {\r
+ for ($a = count($this->a_formatting) - 1; $a >= 0; $a--) {\r
+ if ($this->a_formatting[$a] === self::MARKER) {\r
+ break;\r
+\r
+ } elseif ($this->a_formatting[$a]->tagName === $token['name']) {\r
+ $formatting_element = $this->a_formatting[$a];\r
+ $in_stack = in_array($formatting_element, $this->stack, true);\r
+ $fe_af_pos = $a;\r
+ break;\r
+ }\r
+ }\r
+\r
+ /* If there is no such node, or, if that node is\r
+ also in the stack of open elements but the element\r
+ is not in scope, then this is a parse error. Abort\r
+ these steps. The token is ignored. */\r
+ if (!isset($formatting_element) || ($in_stack &&\r
+ !$this->elementInScope($token['name']))\r
+ ) {\r
+ break;\r
+\r
+ /* Otherwise, if there is such a node, but that node\r
+ is not in the stack of open elements, then this is a\r
+ parse error; remove the element from the list, and\r
+ abort these steps. */\r
+ } elseif (isset($formatting_element) && !$in_stack) {\r
+ unset($this->a_formatting[$fe_af_pos]);\r
+ $this->a_formatting = array_merge($this->a_formatting);\r
+ break;\r
+ }\r
+\r
+ /* 2. Let the furthest block be the topmost node in the\r
+ stack of open elements that is lower in the stack\r
+ than the formatting element, and is not an element in\r
+ the phrasing or formatting categories. There might\r
+ not be one. */\r
+ $fe_s_pos = array_search($formatting_element, $this->stack, true);\r
+ $length = count($this->stack);\r
+\r
+ for ($s = $fe_s_pos + 1; $s < $length; $s++) {\r
+ $category = $this->getElementCategory($this->stack[$s]->nodeName);\r
+\r
+ if ($category !== self::PHRASING && $category !== self::FORMATTING) {\r
+ $furthest_block = $this->stack[$s];\r
+ }\r
+ }\r
+\r
+ /* 3. If there is no furthest block, then the UA must\r
+ skip the subsequent steps and instead just pop all\r
+ the nodes from the bottom of the stack of open\r
+ elements, from the current node up to the formatting\r
+ element, and remove the formatting element from the\r
+ list of active formatting elements. */\r
+ if (!isset($furthest_block)) {\r
+ for ($n = $length - 1; $n >= $fe_s_pos; $n--) {\r
+ array_pop($this->stack);\r
+ }\r
+\r
+ unset($this->a_formatting[$fe_af_pos]);\r
+ $this->a_formatting = array_merge($this->a_formatting);\r
+ break;\r
+ }\r
+\r
+ /* 4. Let the common ancestor be the element\r
+ immediately above the formatting element in the stack\r
+ of open elements. */\r
+ $common_ancestor = $this->stack[$fe_s_pos - 1];\r
+\r
+ /* 5. If the furthest block has a parent node, then\r
+ remove the furthest block from its parent node. */\r
+ if ($furthest_block->parentNode !== null) {\r
+ $furthest_block->parentNode->removeChild($furthest_block);\r
+ }\r
+\r
+ /* 6. Let a bookmark note the position of the\r
+ formatting element in the list of active formatting\r
+ elements relative to the elements on either side\r
+ of it in the list. */\r
+ $bookmark = $fe_af_pos;\r
+\r
+ /* 7. Let node and last node be the furthest block.\r
+ Follow these steps: */\r
+ $node = $furthest_block;\r
+ $last_node = $furthest_block;\r
+\r
+ while (true) {\r
+ for ($n = array_search($node, $this->stack, true) - 1; $n >= 0; $n--) {\r
+ /* 7.1 Let node be the element immediately\r
+ prior to node in the stack of open elements. */\r
+ $node = $this->stack[$n];\r
+\r
+ /* 7.2 If node is not in the list of active\r
+ formatting elements, then remove node from\r
+ the stack of open elements and then go back\r
+ to step 1. */\r
+ if (!in_array($node, $this->a_formatting, true)) {\r
+ unset($this->stack[$n]);\r
+ $this->stack = array_merge($this->stack);\r
+\r
+ } else {\r
+ break;\r
+ }\r
+ }\r
+\r
+ /* 7.3 Otherwise, if node is the formatting\r
+ element, then go to the next step in the overall\r
+ algorithm. */\r
+ if ($node === $formatting_element) {\r
+ break;\r
+\r
+ /* 7.4 Otherwise, if last node is the furthest\r
+ block, then move the aforementioned bookmark to\r
+ be immediately after the node in the list of\r
+ active formatting elements. */\r
+ } elseif ($last_node === $furthest_block) {\r
+ $bookmark = array_search($node, $this->a_formatting, true) + 1;\r
+ }\r
+\r
+ /* 7.5 If node has any children, perform a\r
+ shallow clone of node, replace the entry for\r
+ node in the list of active formatting elements\r
+ with an entry for the clone, replace the entry\r
+ for node in the stack of open elements with an\r
+ entry for the clone, and let node be the clone. */\r
+ if ($node->hasChildNodes()) {\r
+ $clone = $node->cloneNode();\r
+ $s_pos = array_search($node, $this->stack, true);\r
+ $a_pos = array_search($node, $this->a_formatting, true);\r
+\r
+ $this->stack[$s_pos] = $clone;\r
+ $this->a_formatting[$a_pos] = $clone;\r
+ $node = $clone;\r
+ }\r
+\r
+ /* 7.6 Insert last node into node, first removing\r
+ it from its previous parent node if any. */\r
+ if ($last_node->parentNode !== null) {\r
+ $last_node->parentNode->removeChild($last_node);\r
+ }\r
+\r
+ $node->appendChild($last_node);\r
+\r
+ /* 7.7 Let last node be node. */\r
+ $last_node = $node;\r
+ }\r
+\r
+ /* 8. Insert whatever last node ended up being in\r
+ the previous step into the common ancestor node,\r
+ first removing it from its previous parent node if\r
+ any. */\r
+ if ($last_node->parentNode !== null) {\r
+ $last_node->parentNode->removeChild($last_node);\r
+ }\r
+\r
+ $common_ancestor->appendChild($last_node);\r
+\r
+ /* 9. Perform a shallow clone of the formatting\r
+ element. */\r
+ $clone = $formatting_element->cloneNode();\r
+\r
+ /* 10. Take all of the child nodes of the furthest\r
+ block and append them to the clone created in the\r
+ last step. */\r
+ while ($furthest_block->hasChildNodes()) {\r
+ $child = $furthest_block->firstChild;\r
+ $furthest_block->removeChild($child);\r
+ $clone->appendChild($child);\r
+ }\r
+\r
+ /* 11. Append that clone to the furthest block. */\r
+ $furthest_block->appendChild($clone);\r
+\r
+ /* 12. Remove the formatting element from the list\r
+ of active formatting elements, and insert the clone\r
+ into the list of active formatting elements at the\r
+ position of the aforementioned bookmark. */\r
+ $fe_af_pos = array_search($formatting_element, $this->a_formatting, true);\r
+ unset($this->a_formatting[$fe_af_pos]);\r
+ $this->a_formatting = array_merge($this->a_formatting);\r
+\r
+ $af_part1 = array_slice($this->a_formatting, 0, $bookmark - 1);\r
+ $af_part2 = array_slice($this->a_formatting, $bookmark, count($this->a_formatting));\r
+ $this->a_formatting = array_merge($af_part1, array($clone), $af_part2);\r
+\r
+ /* 13. Remove the formatting element from the stack\r
+ of open elements, and insert the clone into the stack\r
+ of open elements immediately after (i.e. in a more\r
+ deeply nested position than) the position of the\r
+ furthest block in that stack. */\r
+ $fe_s_pos = array_search($formatting_element, $this->stack, true);\r
+ $fb_s_pos = array_search($furthest_block, $this->stack, true);\r
+ unset($this->stack[$fe_s_pos]);\r
+\r
+ $s_part1 = array_slice($this->stack, 0, $fb_s_pos);\r
+ $s_part2 = array_slice($this->stack, $fb_s_pos + 1, count($this->stack));\r
+ $this->stack = array_merge($s_part1, array($clone), $s_part2);\r
+\r
+ /* 14. Jump back to step 1 in this series of steps. */\r
+ unset($formatting_element, $fe_af_pos, $fe_s_pos, $furthest_block);\r
+ }\r
+ break;\r
+\r
+ /* An end tag token whose tag name is one of: "button",\r
+ "marquee", "object" */\r
+ case 'button':\r
+ case 'marquee':\r
+ case 'object':\r
+ /* If the stack of open elements has an element in scope whose\r
+ tag name matches the tag name of the token, then generate implied\r
+ tags. */\r
+ if ($this->elementInScope($token['name'])) {\r
+ $this->generateImpliedEndTags();\r
+\r
+ /* Now, if the current node is not an element with the same\r
+ tag name as the token, then this is a parse error. */\r
+ // k\r
+\r
+ /* Now, if the stack of open elements has an element in scope\r
+ whose tag name matches the tag name of the token, then pop\r
+ elements from the stack until that element has been popped from\r
+ the stack, and clear the list of active formatting elements up\r
+ to the last marker. */\r
+ for ($n = count($this->stack) - 1; $n >= 0; $n--) {\r
+ if ($this->stack[$n]->nodeName === $token['name']) {\r
+ $n = -1;\r
+ }\r
+\r
+ array_pop($this->stack);\r
+ }\r
+\r
+ $marker = end(array_keys($this->a_formatting, self::MARKER, true));\r
+\r
+ for ($n = count($this->a_formatting) - 1; $n > $marker; $n--) {\r
+ array_pop($this->a_formatting);\r
+ }\r
+ }\r
+ break;\r
+\r
+ /* Or an end tag whose tag name is one of: "area", "basefont",\r
+ "bgsound", "br", "embed", "hr", "iframe", "image", "img",\r
+ "input", "isindex", "noembed", "noframes", "param", "select",\r
+ "spacer", "table", "textarea", "wbr" */\r
+ case 'area':\r
+ case 'basefont':\r
+ case 'bgsound':\r
+ case 'br':\r
+ case 'embed':\r
+ case 'hr':\r
+ case 'iframe':\r
+ case 'image':\r
+ case 'img':\r
+ case 'input':\r
+ case 'isindex':\r
+ case 'noembed':\r
+ case 'noframes':\r
+ case 'param':\r
+ case 'select':\r
+ case 'spacer':\r
+ case 'table':\r
+ case 'textarea':\r
+ case 'wbr':\r
+ // Parse error. Ignore the token.\r
+ break;\r
+\r
+ /* An end tag token not covered by the previous entries */\r
+ default:\r
+ for ($n = count($this->stack) - 1; $n >= 0; $n--) {\r
+ /* Initialise node to be the current node (the bottommost\r
+ node of the stack). */\r
+ $node = end($this->stack);\r
+\r
+ /* If node has the same tag name as the end tag token,\r
+ then: */\r
+ if ($token['name'] === $node->nodeName) {\r
+ /* Generate implied end tags. */\r
+ $this->generateImpliedEndTags();\r
+\r
+ /* If the tag name of the end tag token does not\r
+ match the tag name of the current node, this is a\r
+ parse error. */\r
+ // k\r
+\r
+ /* Pop all the nodes from the current node up to\r
+ node, including node, then stop this algorithm. */\r
+ for ($x = count($this->stack) - $n; $x >= $n; $x--) {\r
+ array_pop($this->stack);\r
+ }\r
+\r
+ } else {\r
+ $category = $this->getElementCategory($node);\r
+\r
+ if ($category !== self::SPECIAL && $category !== self::SCOPING) {\r
+ /* Otherwise, if node is in neither the formatting\r
+ category nor the phrasing category, then this is a\r
+ parse error. Stop this algorithm. The end tag token\r
+ is ignored. */\r
+ return false;\r
+ }\r
+ }\r
+ }\r
+ break;\r
+ }\r
+ break;\r
+ }\r
+ }\r
+\r
+ private function inTable($token)\r
+ {\r
+ $clear = array('html', 'table');\r
+\r
+ /* A character token that is one of one of U+0009 CHARACTER TABULATION,\r
+ U+000A LINE FEED (LF), U+000B LINE TABULATION, U+000C FORM FEED (FF),\r
+ or U+0020 SPACE */\r
+ if ($token['type'] === HTML5::CHARACTR &&\r
+ preg_match('/^[\t\n\x0b\x0c ]+$/', $token['data'])\r
+ ) {\r
+ /* Append the character to the current node. */\r
+ $text = $this->dom->createTextNode($token['data']);\r
+ end($this->stack)->appendChild($text);\r
+\r
+ /* A comment token */\r
+ } elseif ($token['type'] === HTML5::COMMENT) {\r
+ /* Append a Comment node to the current node with the data\r
+ attribute set to the data given in the comment token. */\r
+ $comment = $this->dom->createComment($token['data']);\r
+ end($this->stack)->appendChild($comment);\r
+\r
+ /* A start tag whose tag name is "caption" */\r
+ } elseif ($token['type'] === HTML5::STARTTAG &&\r
+ $token['name'] === 'caption'\r
+ ) {\r
+ /* Clear the stack back to a table context. */\r
+ $this->clearStackToTableContext($clear);\r
+\r
+ /* Insert a marker at the end of the list of active\r
+ formatting elements. */\r
+ $this->a_formatting[] = self::MARKER;\r
+\r
+ /* Insert an HTML element for the token, then switch the\r
+ insertion mode to "in caption". */\r
+ $this->insertElement($token);\r
+ $this->mode = self::IN_CAPTION;\r
+\r
+ /* A start tag whose tag name is "colgroup" */\r
+ } elseif ($token['type'] === HTML5::STARTTAG &&\r
+ $token['name'] === 'colgroup'\r
+ ) {\r
+ /* Clear the stack back to a table context. */\r
+ $this->clearStackToTableContext($clear);\r
+\r
+ /* Insert an HTML element for the token, then switch the\r
+ insertion mode to "in column group". */\r
+ $this->insertElement($token);\r
+ $this->mode = self::IN_CGROUP;\r
+\r
+ /* A start tag whose tag name is "col" */\r
+ } elseif ($token['type'] === HTML5::STARTTAG &&\r
+ $token['name'] === 'col'\r
+ ) {\r
+ $this->inTable(\r
+ array(\r
+ 'name' => 'colgroup',\r
+ 'type' => HTML5::STARTTAG,\r
+ 'attr' => array()\r
+ )\r
+ );\r
+\r
+ $this->inColumnGroup($token);\r
+\r
+ /* A start tag whose tag name is one of: "tbody", "tfoot", "thead" */\r
+ } elseif ($token['type'] === HTML5::STARTTAG && in_array(\r
+ $token['name'],\r
+ array('tbody', 'tfoot', 'thead')\r
+ )\r
+ ) {\r
+ /* Clear the stack back to a table context. */\r
+ $this->clearStackToTableContext($clear);\r
+\r
+ /* Insert an HTML element for the token, then switch the insertion\r
+ mode to "in table body". */\r
+ $this->insertElement($token);\r
+ $this->mode = self::IN_TBODY;\r
+\r
+ /* A start tag whose tag name is one of: "td", "th", "tr" */\r
+ } elseif ($token['type'] === HTML5::STARTTAG &&\r
+ in_array($token['name'], array('td', 'th', 'tr'))\r
+ ) {\r
+ /* Act as if a start tag token with the tag name "tbody" had been\r
+ seen, then reprocess the current token. */\r
+ $this->inTable(\r
+ array(\r
+ 'name' => 'tbody',\r
+ 'type' => HTML5::STARTTAG,\r
+ 'attr' => array()\r
+ )\r
+ );\r
+\r
+ return $this->inTableBody($token);\r
+\r
+ /* A start tag whose tag name is "table" */\r
+ } elseif ($token['type'] === HTML5::STARTTAG &&\r
+ $token['name'] === 'table'\r
+ ) {\r
+ /* Parse error. Act as if an end tag token with the tag name "table"\r
+ had been seen, then, if that token wasn't ignored, reprocess the\r
+ current token. */\r
+ $this->inTable(\r
+ array(\r
+ 'name' => 'table',\r
+ 'type' => HTML5::ENDTAG\r
+ )\r
+ );\r
+\r
+ return $this->mainPhase($token);\r
+\r
+ /* An end tag whose tag name is "table" */\r
+ } elseif ($token['type'] === HTML5::ENDTAG &&\r
+ $token['name'] === 'table'\r
+ ) {\r
+ /* If the stack of open elements does not have an element in table\r
+ scope with the same tag name as the token, this is a parse error.\r
+ Ignore the token. (innerHTML case) */\r
+ if (!$this->elementInScope($token['name'], true)) {\r
+ return false;\r
+\r
+ /* Otherwise: */\r
+ } else {\r
+ /* Generate implied end tags. */\r
+ $this->generateImpliedEndTags();\r
+\r
+ /* Now, if the current node is not a table element, then this\r
+ is a parse error. */\r
+ // w/e\r
+\r
+ /* Pop elements from this stack until a table element has been\r
+ popped from the stack. */\r
+ while (true) {\r
+ $current = end($this->stack)->nodeName;\r
+ array_pop($this->stack);\r
+\r
+ if ($current === 'table') {\r
+ break;\r
+ }\r
+ }\r
+\r
+ /* Reset the insertion mode appropriately. */\r
+ $this->resetInsertionMode();\r
+ }\r
+\r
+ /* An end tag whose tag name is one of: "body", "caption", "col",\r
+ "colgroup", "html", "tbody", "td", "tfoot", "th", "thead", "tr" */\r
+ } elseif ($token['type'] === HTML5::ENDTAG && in_array(\r
+ $token['name'],\r
+ array(\r
+ 'body',\r
+ 'caption',\r
+ 'col',\r
+ 'colgroup',\r
+ 'html',\r
+ 'tbody',\r
+ 'td',\r
+ 'tfoot',\r
+ 'th',\r
+ 'thead',\r
+ 'tr'\r
+ )\r
+ )\r
+ ) {\r
+ // Parse error. Ignore the token.\r
+\r
+ /* Anything else */\r
+ } else {\r
+ /* Parse error. Process the token as if the insertion mode was "in\r
+ body", with the following exception: */\r
+\r
+ /* If the current node is a table, tbody, tfoot, thead, or tr\r
+ element, then, whenever a node would be inserted into the current\r
+ node, it must instead be inserted into the foster parent element. */\r
+ if (in_array(\r
+ end($this->stack)->nodeName,\r
+ array('table', 'tbody', 'tfoot', 'thead', 'tr')\r
+ )\r
+ ) {\r
+ /* The foster parent element is the parent element of the last\r
+ table element in the stack of open elements, if there is a\r
+ table element and it has such a parent element. If there is no\r
+ table element in the stack of open elements (innerHTML case),\r
+ then the foster parent element is the first element in the\r
+ stack of open elements (the html element). Otherwise, if there\r
+ is a table element in the stack of open elements, but the last\r
+ table element in the stack of open elements has no parent, or\r
+ its parent node is not an element, then the foster parent\r
+ element is the element before the last table element in the\r
+ stack of open elements. */\r
+ for ($n = count($this->stack) - 1; $n >= 0; $n--) {\r
+ if ($this->stack[$n]->nodeName === 'table') {\r
+ $table = $this->stack[$n];\r
+ break;\r
+ }\r
+ }\r
+\r
+ if (isset($table) && $table->parentNode !== null) {\r
+ $this->foster_parent = $table->parentNode;\r
+\r
+ } elseif (!isset($table)) {\r
+ $this->foster_parent = $this->stack[0];\r
+\r
+ } elseif (isset($table) && ($table->parentNode === null ||\r
+ $table->parentNode->nodeType !== XML_ELEMENT_NODE)\r
+ ) {\r
+ $this->foster_parent = $this->stack[$n - 1];\r
+ }\r
+ }\r
+\r
+ $this->inBody($token);\r
+ }\r
+ }\r
+\r
+ private function inCaption($token)\r
+ {\r
+ /* An end tag whose tag name is "caption" */\r
+ if ($token['type'] === HTML5::ENDTAG && $token['name'] === 'caption') {\r
+ /* If the stack of open elements does not have an element in table\r
+ scope with the same tag name as the token, this is a parse error.\r
+ Ignore the token. (innerHTML case) */\r
+ if (!$this->elementInScope($token['name'], true)) {\r
+ // Ignore\r
+\r
+ /* Otherwise: */\r
+ } else {\r
+ /* Generate implied end tags. */\r
+ $this->generateImpliedEndTags();\r
+\r
+ /* Now, if the current node is not a caption element, then this\r
+ is a parse error. */\r
+ // w/e\r
+\r
+ /* Pop elements from this stack until a caption element has\r
+ been popped from the stack. */\r
+ while (true) {\r
+ $node = end($this->stack)->nodeName;\r
+ array_pop($this->stack);\r
+\r
+ if ($node === 'caption') {\r
+ break;\r
+ }\r
+ }\r
+\r
+ /* Clear the list of active formatting elements up to the last\r
+ marker. */\r
+ $this->clearTheActiveFormattingElementsUpToTheLastMarker();\r
+\r
+ /* Switch the insertion mode to "in table". */\r
+ $this->mode = self::IN_TABLE;\r
+ }\r
+\r
+ /* A start tag whose tag name is one of: "caption", "col", "colgroup",\r
+ "tbody", "td", "tfoot", "th", "thead", "tr", or an end tag whose tag\r
+ name is "table" */\r
+ } elseif (($token['type'] === HTML5::STARTTAG && in_array(\r
+ $token['name'],\r
+ array(\r
+ 'caption',\r
+ 'col',\r
+ 'colgroup',\r
+ 'tbody',\r
+ 'td',\r
+ 'tfoot',\r
+ 'th',\r
+ 'thead',\r
+ 'tr'\r
+ )\r
+ )) || ($token['type'] === HTML5::ENDTAG &&\r
+ $token['name'] === 'table')\r
+ ) {\r
+ /* Parse error. Act as if an end tag with the tag name "caption"\r
+ had been seen, then, if that token wasn't ignored, reprocess the\r
+ current token. */\r
+ $this->inCaption(\r
+ array(\r
+ 'name' => 'caption',\r
+ 'type' => HTML5::ENDTAG\r
+ )\r
+ );\r
+\r
+ return $this->inTable($token);\r
+\r
+ /* An end tag whose tag name is one of: "body", "col", "colgroup",\r
+ "html", "tbody", "td", "tfoot", "th", "thead", "tr" */\r
+ } elseif ($token['type'] === HTML5::ENDTAG && in_array(\r
+ $token['name'],\r
+ array(\r
+ 'body',\r
+ 'col',\r
+ 'colgroup',\r
+ 'html',\r
+ 'tbody',\r
+ 'tfoot',\r
+ 'th',\r
+ 'thead',\r
+ 'tr'\r
+ )\r
+ )\r
+ ) {\r
+ // Parse error. Ignore the token.\r
+\r
+ /* Anything else */\r
+ } else {\r
+ /* Process the token as if the insertion mode was "in body". */\r
+ $this->inBody($token);\r
+ }\r
+ }\r
+\r
+ private function inColumnGroup($token)\r
+ {\r
+ /* A character token that is one of one of U+0009 CHARACTER TABULATION,\r
+ U+000A LINE FEED (LF), U+000B LINE TABULATION, U+000C FORM FEED (FF),\r
+ or U+0020 SPACE */\r
+ if ($token['type'] === HTML5::CHARACTR &&\r
+ preg_match('/^[\t\n\x0b\x0c ]+$/', $token['data'])\r
+ ) {\r
+ /* Append the character to the current node. */\r
+ $text = $this->dom->createTextNode($token['data']);\r
+ end($this->stack)->appendChild($text);\r
+\r
+ /* A comment token */\r
+ } elseif ($token['type'] === HTML5::COMMENT) {\r
+ /* Append a Comment node to the current node with the data\r
+ attribute set to the data given in the comment token. */\r
+ $comment = $this->dom->createComment($token['data']);\r
+ end($this->stack)->appendChild($comment);\r
+\r
+ /* A start tag whose tag name is "col" */\r
+ } elseif ($token['type'] === HTML5::STARTTAG && $token['name'] === 'col') {\r
+ /* Insert a col element for the token. Immediately pop the current\r
+ node off the stack of open elements. */\r
+ $this->insertElement($token);\r
+ array_pop($this->stack);\r
+\r
+ /* An end tag whose tag name is "colgroup" */\r
+ } elseif ($token['type'] === HTML5::ENDTAG &&\r
+ $token['name'] === 'colgroup'\r
+ ) {\r
+ /* If the current node is the root html element, then this is a\r
+ parse error, ignore the token. (innerHTML case) */\r
+ if (end($this->stack)->nodeName === 'html') {\r
+ // Ignore\r
+\r
+ /* Otherwise, pop the current node (which will be a colgroup\r
+ element) from the stack of open elements. Switch the insertion\r
+ mode to "in table". */\r
+ } else {\r
+ array_pop($this->stack);\r
+ $this->mode = self::IN_TABLE;\r
+ }\r
+\r
+ /* An end tag whose tag name is "col" */\r
+ } elseif ($token['type'] === HTML5::ENDTAG && $token['name'] === 'col') {\r
+ /* Parse error. Ignore the token. */\r
+\r
+ /* Anything else */\r
+ } else {\r
+ /* Act as if an end tag with the tag name "colgroup" had been seen,\r
+ and then, if that token wasn't ignored, reprocess the current token. */\r
+ $this->inColumnGroup(\r
+ array(\r
+ 'name' => 'colgroup',\r
+ 'type' => HTML5::ENDTAG\r
+ )\r
+ );\r
+\r
+ return $this->inTable($token);\r
+ }\r
+ }\r
+\r
+ private function inTableBody($token)\r
+ {\r
+ $clear = array('tbody', 'tfoot', 'thead', 'html');\r
+\r
+ /* A start tag whose tag name is "tr" */\r
+ if ($token['type'] === HTML5::STARTTAG && $token['name'] === 'tr') {\r
+ /* Clear the stack back to a table body context. */\r
+ $this->clearStackToTableContext($clear);\r
+\r
+ /* Insert a tr element for the token, then switch the insertion\r
+ mode to "in row". */\r
+ $this->insertElement($token);\r
+ $this->mode = self::IN_ROW;\r
+\r
+ /* A start tag whose tag name is one of: "th", "td" */\r
+ } elseif ($token['type'] === HTML5::STARTTAG &&\r
+ ($token['name'] === 'th' || $token['name'] === 'td')\r
+ ) {\r
+ /* Parse error. Act as if a start tag with the tag name "tr" had\r
+ been seen, then reprocess the current token. */\r
+ $this->inTableBody(\r
+ array(\r
+ 'name' => 'tr',\r
+ 'type' => HTML5::STARTTAG,\r
+ 'attr' => array()\r
+ )\r
+ );\r
+\r
+ return $this->inRow($token);\r
+\r
+ /* An end tag whose tag name is one of: "tbody", "tfoot", "thead" */\r
+ } elseif ($token['type'] === HTML5::ENDTAG &&\r
+ in_array($token['name'], array('tbody', 'tfoot', 'thead'))\r
+ ) {\r
+ /* If the stack of open elements does not have an element in table\r
+ scope with the same tag name as the token, this is a parse error.\r
+ Ignore the token. */\r
+ if (!$this->elementInScope($token['name'], true)) {\r
+ // Ignore\r
+\r
+ /* Otherwise: */\r
+ } else {\r
+ /* Clear the stack back to a table body context. */\r
+ $this->clearStackToTableContext($clear);\r
+\r
+ /* Pop the current node from the stack of open elements. Switch\r
+ the insertion mode to "in table". */\r
+ array_pop($this->stack);\r
+ $this->mode = self::IN_TABLE;\r
+ }\r
+\r
+ /* A start tag whose tag name is one of: "caption", "col", "colgroup",\r
+ "tbody", "tfoot", "thead", or an end tag whose tag name is "table" */\r
+ } elseif (($token['type'] === HTML5::STARTTAG && in_array(\r
+ $token['name'],\r
+ array('caption', 'col', 'colgroup', 'tbody', 'tfoor', 'thead')\r
+ )) ||\r
+ ($token['type'] === HTML5::STARTTAG && $token['name'] === 'table')\r
+ ) {\r
+ /* If the stack of open elements does not have a tbody, thead, or\r
+ tfoot element in table scope, this is a parse error. Ignore the\r
+ token. (innerHTML case) */\r
+ if (!$this->elementInScope(array('tbody', 'thead', 'tfoot'), true)) {\r
+ // Ignore.\r
+\r
+ /* Otherwise: */\r
+ } else {\r
+ /* Clear the stack back to a table body context. */\r
+ $this->clearStackToTableContext($clear);\r
+\r
+ /* Act as if an end tag with the same tag name as the current\r
+ node ("tbody", "tfoot", or "thead") had been seen, then\r
+ reprocess the current token. */\r
+ $this->inTableBody(\r
+ array(\r
+ 'name' => end($this->stack)->nodeName,\r
+ 'type' => HTML5::ENDTAG\r
+ )\r
+ );\r
+\r
+ return $this->mainPhase($token);\r
+ }\r
+\r
+ /* An end tag whose tag name is one of: "body", "caption", "col",\r
+ "colgroup", "html", "td", "th", "tr" */\r
+ } elseif ($token['type'] === HTML5::ENDTAG && in_array(\r
+ $token['name'],\r
+ array('body', 'caption', 'col', 'colgroup', 'html', 'td', 'th', 'tr')\r
+ )\r
+ ) {\r
+ /* Parse error. Ignore the token. */\r
+\r
+ /* Anything else */\r
+ } else {\r
+ /* Process the token as if the insertion mode was "in table". */\r
+ $this->inTable($token);\r
+ }\r
+ }\r
+\r
+ private function inRow($token)\r
+ {\r
+ $clear = array('tr', 'html');\r
+\r
+ /* A start tag whose tag name is one of: "th", "td" */\r
+ if ($token['type'] === HTML5::STARTTAG &&\r
+ ($token['name'] === 'th' || $token['name'] === 'td')\r
+ ) {\r
+ /* Clear the stack back to a table row context. */\r
+ $this->clearStackToTableContext($clear);\r
+\r
+ /* Insert an HTML element for the token, then switch the insertion\r
+ mode to "in cell". */\r
+ $this->insertElement($token);\r
+ $this->mode = self::IN_CELL;\r
+\r
+ /* Insert a marker at the end of the list of active formatting\r
+ elements. */\r
+ $this->a_formatting[] = self::MARKER;\r
+\r
+ /* An end tag whose tag name is "tr" */\r
+ } elseif ($token['type'] === HTML5::ENDTAG && $token['name'] === 'tr') {\r
+ /* If the stack of open elements does not have an element in table\r
+ scope with the same tag name as the token, this is a parse error.\r
+ Ignore the token. (innerHTML case) */\r
+ if (!$this->elementInScope($token['name'], true)) {\r
+ // Ignore.\r
+\r
+ /* Otherwise: */\r
+ } else {\r
+ /* Clear the stack back to a table row context. */\r
+ $this->clearStackToTableContext($clear);\r
+\r
+ /* Pop the current node (which will be a tr element) from the\r
+ stack of open elements. Switch the insertion mode to "in table\r
+ body". */\r
+ array_pop($this->stack);\r
+ $this->mode = self::IN_TBODY;\r
+ }\r
+\r
+ /* A start tag whose tag name is one of: "caption", "col", "colgroup",\r
+ "tbody", "tfoot", "thead", "tr" or an end tag whose tag name is "table" */\r
+ } elseif ($token['type'] === HTML5::STARTTAG && in_array(\r
+ $token['name'],\r
+ array('caption', 'col', 'colgroup', 'tbody', 'tfoot', 'thead', 'tr')\r
+ )\r
+ ) {\r
+ /* Act as if an end tag with the tag name "tr" had been seen, then,\r
+ if that token wasn't ignored, reprocess the current token. */\r
+ $this->inRow(\r
+ array(\r
+ 'name' => 'tr',\r
+ 'type' => HTML5::ENDTAG\r
+ )\r
+ );\r
+\r
+ return $this->inCell($token);\r
+\r
+ /* An end tag whose tag name is one of: "tbody", "tfoot", "thead" */\r
+ } elseif ($token['type'] === HTML5::ENDTAG &&\r
+ in_array($token['name'], array('tbody', 'tfoot', 'thead'))\r
+ ) {\r
+ /* If the stack of open elements does not have an element in table\r
+ scope with the same tag name as the token, this is a parse error.\r
+ Ignore the token. */\r
+ if (!$this->elementInScope($token['name'], true)) {\r
+ // Ignore.\r
+\r
+ /* Otherwise: */\r
+ } else {\r
+ /* Otherwise, act as if an end tag with the tag name "tr" had\r
+ been seen, then reprocess the current token. */\r
+ $this->inRow(\r
+ array(\r
+ 'name' => 'tr',\r
+ 'type' => HTML5::ENDTAG\r
+ )\r
+ );\r
+\r
+ return $this->inCell($token);\r
+ }\r
+\r
+ /* An end tag whose tag name is one of: "body", "caption", "col",\r
+ "colgroup", "html", "td", "th" */\r
+ } elseif ($token['type'] === HTML5::ENDTAG && in_array(\r
+ $token['name'],\r
+ array('body', 'caption', 'col', 'colgroup', 'html', 'td', 'th', 'tr')\r
+ )\r
+ ) {\r
+ /* Parse error. Ignore the token. */\r
+\r
+ /* Anything else */\r
+ } else {\r
+ /* Process the token as if the insertion mode was "in table". */\r
+ $this->inTable($token);\r
+ }\r
+ }\r
+\r
+ private function inCell($token)\r
+ {\r
+ /* An end tag whose tag name is one of: "td", "th" */\r
+ if ($token['type'] === HTML5::ENDTAG &&\r
+ ($token['name'] === 'td' || $token['name'] === 'th')\r
+ ) {\r
+ /* If the stack of open elements does not have an element in table\r
+ scope with the same tag name as that of the token, then this is a\r
+ parse error and the token must be ignored. */\r
+ if (!$this->elementInScope($token['name'], true)) {\r
+ // Ignore.\r
+\r
+ /* Otherwise: */\r
+ } else {\r
+ /* Generate implied end tags, except for elements with the same\r
+ tag name as the token. */\r
+ $this->generateImpliedEndTags(array($token['name']));\r
+\r
+ /* Now, if the current node is not an element with the same tag\r
+ name as the token, then this is a parse error. */\r
+ // k\r
+\r
+ /* Pop elements from this stack until an element with the same\r
+ tag name as the token has been popped from the stack. */\r
+ while (true) {\r
+ $node = end($this->stack)->nodeName;\r
+ array_pop($this->stack);\r
+\r
+ if ($node === $token['name']) {\r
+ break;\r
+ }\r
+ }\r
+\r
+ /* Clear the list of active formatting elements up to the last\r
+ marker. */\r
+ $this->clearTheActiveFormattingElementsUpToTheLastMarker();\r
+\r
+ /* Switch the insertion mode to "in row". (The current node\r
+ will be a tr element at this point.) */\r
+ $this->mode = self::IN_ROW;\r
+ }\r
+\r
+ /* A start tag whose tag name is one of: "caption", "col", "colgroup",\r
+ "tbody", "td", "tfoot", "th", "thead", "tr" */\r
+ } elseif ($token['type'] === HTML5::STARTTAG && in_array(\r
+ $token['name'],\r
+ array(\r
+ 'caption',\r
+ 'col',\r
+ 'colgroup',\r
+ 'tbody',\r
+ 'td',\r
+ 'tfoot',\r
+ 'th',\r
+ 'thead',\r
+ 'tr'\r
+ )\r
+ )\r
+ ) {\r
+ /* If the stack of open elements does not have a td or th element\r
+ in table scope, then this is a parse error; ignore the token.\r
+ (innerHTML case) */\r
+ if (!$this->elementInScope(array('td', 'th'), true)) {\r
+ // Ignore.\r
+\r
+ /* Otherwise, close the cell (see below) and reprocess the current\r
+ token. */\r
+ } else {\r
+ $this->closeCell();\r
+ return $this->inRow($token);\r
+ }\r
+\r
+ /* A start tag whose tag name is one of: "caption", "col", "colgroup",\r
+ "tbody", "td", "tfoot", "th", "thead", "tr" */\r
+ } elseif ($token['type'] === HTML5::STARTTAG && in_array(\r
+ $token['name'],\r
+ array(\r
+ 'caption',\r
+ 'col',\r
+ 'colgroup',\r
+ 'tbody',\r
+ 'td',\r
+ 'tfoot',\r
+ 'th',\r
+ 'thead',\r
+ 'tr'\r
+ )\r
+ )\r
+ ) {\r
+ /* If the stack of open elements does not have a td or th element\r
+ in table scope, then this is a parse error; ignore the token.\r
+ (innerHTML case) */\r
+ if (!$this->elementInScope(array('td', 'th'), true)) {\r
+ // Ignore.\r
+\r
+ /* Otherwise, close the cell (see below) and reprocess the current\r
+ token. */\r
+ } else {\r
+ $this->closeCell();\r
+ return $this->inRow($token);\r
+ }\r
+\r
+ /* An end tag whose tag name is one of: "body", "caption", "col",\r
+ "colgroup", "html" */\r
+ } elseif ($token['type'] === HTML5::ENDTAG && in_array(\r
+ $token['name'],\r
+ array('body', 'caption', 'col', 'colgroup', 'html')\r
+ )\r
+ ) {\r
+ /* Parse error. Ignore the token. */\r
+\r
+ /* An end tag whose tag name is one of: "table", "tbody", "tfoot",\r
+ "thead", "tr" */\r
+ } elseif ($token['type'] === HTML5::ENDTAG && in_array(\r
+ $token['name'],\r
+ array('table', 'tbody', 'tfoot', 'thead', 'tr')\r
+ )\r
+ ) {\r
+ /* If the stack of open elements does not have an element in table\r
+ scope with the same tag name as that of the token (which can only\r
+ happen for "tbody", "tfoot" and "thead", or, in the innerHTML case),\r
+ then this is a parse error and the token must be ignored. */\r
+ if (!$this->elementInScope($token['name'], true)) {\r
+ // Ignore.\r
+\r
+ /* Otherwise, close the cell (see below) and reprocess the current\r
+ token. */\r
+ } else {\r
+ $this->closeCell();\r
+ return $this->inRow($token);\r
+ }\r
+\r
+ /* Anything else */\r
+ } else {\r
+ /* Process the token as if the insertion mode was "in body". */\r
+ $this->inBody($token);\r
+ }\r
+ }\r
+\r
+ private function inSelect($token)\r
+ {\r
+ /* Handle the token as follows: */\r
+\r
+ /* A character token */\r
+ if ($token['type'] === HTML5::CHARACTR) {\r
+ /* Append the token's character to the current node. */\r
+ $this->insertText($token['data']);\r
+\r
+ /* A comment token */\r
+ } elseif ($token['type'] === HTML5::COMMENT) {\r
+ /* Append a Comment node to the current node with the data\r
+ attribute set to the data given in the comment token. */\r
+ $this->insertComment($token['data']);\r
+\r
+ /* A start tag token whose tag name is "option" */\r
+ } elseif ($token['type'] === HTML5::STARTTAG &&\r
+ $token['name'] === 'option'\r
+ ) {\r
+ /* If the current node is an option element, act as if an end tag\r
+ with the tag name "option" had been seen. */\r
+ if (end($this->stack)->nodeName === 'option') {\r
+ $this->inSelect(\r
+ array(\r
+ 'name' => 'option',\r
+ 'type' => HTML5::ENDTAG\r
+ )\r
+ );\r
+ }\r
+\r
+ /* Insert an HTML element for the token. */\r
+ $this->insertElement($token);\r
+\r
+ /* A start tag token whose tag name is "optgroup" */\r
+ } elseif ($token['type'] === HTML5::STARTTAG &&\r
+ $token['name'] === 'optgroup'\r
+ ) {\r
+ /* If the current node is an option element, act as if an end tag\r
+ with the tag name "option" had been seen. */\r
+ if (end($this->stack)->nodeName === 'option') {\r
+ $this->inSelect(\r
+ array(\r
+ 'name' => 'option',\r
+ 'type' => HTML5::ENDTAG\r
+ )\r
+ );\r
+ }\r
+\r
+ /* If the current node is an optgroup element, act as if an end tag\r
+ with the tag name "optgroup" had been seen. */\r
+ if (end($this->stack)->nodeName === 'optgroup') {\r
+ $this->inSelect(\r
+ array(\r
+ 'name' => 'optgroup',\r
+ 'type' => HTML5::ENDTAG\r
+ )\r
+ );\r
+ }\r
+\r
+ /* Insert an HTML element for the token. */\r
+ $this->insertElement($token);\r
+\r
+ /* An end tag token whose tag name is "optgroup" */\r
+ } elseif ($token['type'] === HTML5::ENDTAG &&\r
+ $token['name'] === 'optgroup'\r
+ ) {\r
+ /* First, if the current node is an option element, and the node\r
+ immediately before it in the stack of open elements is an optgroup\r
+ element, then act as if an end tag with the tag name "option" had\r
+ been seen. */\r
+ $elements_in_stack = count($this->stack);\r
+\r
+ if ($this->stack[$elements_in_stack - 1]->nodeName === 'option' &&\r
+ $this->stack[$elements_in_stack - 2]->nodeName === 'optgroup'\r
+ ) {\r
+ $this->inSelect(\r
+ array(\r
+ 'name' => 'option',\r
+ 'type' => HTML5::ENDTAG\r
+ )\r
+ );\r
+ }\r
+\r
+ /* If the current node is an optgroup element, then pop that node\r
+ from the stack of open elements. Otherwise, this is a parse error,\r
+ ignore the token. */\r
+ if ($this->stack[$elements_in_stack - 1] === 'optgroup') {\r
+ array_pop($this->stack);\r
+ }\r
+\r
+ /* An end tag token whose tag name is "option" */\r
+ } elseif ($token['type'] === HTML5::ENDTAG &&\r
+ $token['name'] === 'option'\r
+ ) {\r
+ /* If the current node is an option element, then pop that node\r
+ from the stack of open elements. Otherwise, this is a parse error,\r
+ ignore the token. */\r
+ if (end($this->stack)->nodeName === 'option') {\r
+ array_pop($this->stack);\r
+ }\r
+\r
+ /* An end tag whose tag name is "select" */\r
+ } elseif ($token['type'] === HTML5::ENDTAG &&\r
+ $token['name'] === 'select'\r
+ ) {\r
+ /* If the stack of open elements does not have an element in table\r
+ scope with the same tag name as the token, this is a parse error.\r
+ Ignore the token. (innerHTML case) */\r
+ if (!$this->elementInScope($token['name'], true)) {\r
+ // w/e\r
+\r
+ /* Otherwise: */\r
+ } else {\r
+ /* Pop elements from the stack of open elements until a select\r
+ element has been popped from the stack. */\r
+ while (true) {\r
+ $current = end($this->stack)->nodeName;\r
+ array_pop($this->stack);\r
+\r
+ if ($current === 'select') {\r
+ break;\r
+ }\r
+ }\r
+\r
+ /* Reset the insertion mode appropriately. */\r
+ $this->resetInsertionMode();\r
+ }\r
+\r
+ /* A start tag whose tag name is "select" */\r
+ } elseif ($token['name'] === 'select' &&\r
+ $token['type'] === HTML5::STARTTAG\r
+ ) {\r
+ /* Parse error. Act as if the token had been an end tag with the\r
+ tag name "select" instead. */\r
+ $this->inSelect(\r
+ array(\r
+ 'name' => 'select',\r
+ 'type' => HTML5::ENDTAG\r
+ )\r
+ );\r
+\r
+ /* An end tag whose tag name is one of: "caption", "table", "tbody",\r
+ "tfoot", "thead", "tr", "td", "th" */\r
+ } elseif (in_array(\r
+ $token['name'],\r
+ array(\r
+ 'caption',\r
+ 'table',\r
+ 'tbody',\r
+ 'tfoot',\r
+ 'thead',\r
+ 'tr',\r
+ 'td',\r
+ 'th'\r
+ )\r
+ ) && $token['type'] === HTML5::ENDTAG\r
+ ) {\r
+ /* Parse error. */\r
+ // w/e\r
+\r
+ /* If the stack of open elements has an element in table scope with\r
+ the same tag name as that of the token, then act as if an end tag\r
+ with the tag name "select" had been seen, and reprocess the token.\r
+ Otherwise, ignore the token. */\r
+ if ($this->elementInScope($token['name'], true)) {\r
+ $this->inSelect(\r
+ array(\r
+ 'name' => 'select',\r
+ 'type' => HTML5::ENDTAG\r
+ )\r
+ );\r
+\r
+ $this->mainPhase($token);\r
+ }\r
+\r
+ /* Anything else */\r
+ } else {\r
+ /* Parse error. Ignore the token. */\r
+ }\r
+ }\r
+\r
+ private function afterBody($token)\r
+ {\r
+ /* Handle the token as follows: */\r
+\r
+ /* A character token that is one of one of U+0009 CHARACTER TABULATION,\r
+ U+000A LINE FEED (LF), U+000B LINE TABULATION, U+000C FORM FEED (FF),\r
+ or U+0020 SPACE */\r
+ if ($token['type'] === HTML5::CHARACTR &&\r
+ preg_match('/^[\t\n\x0b\x0c ]+$/', $token['data'])\r
+ ) {\r
+ /* Process the token as it would be processed if the insertion mode\r
+ was "in body". */\r
+ $this->inBody($token);\r
+\r
+ /* A comment token */\r
+ } elseif ($token['type'] === HTML5::COMMENT) {\r
+ /* Append a Comment node to the first element in the stack of open\r
+ elements (the html element), with the data attribute set to the\r
+ data given in the comment token. */\r
+ $comment = $this->dom->createComment($token['data']);\r
+ $this->stack[0]->appendChild($comment);\r
+\r
+ /* An end tag with the tag name "html" */\r
+ } elseif ($token['type'] === HTML5::ENDTAG && $token['name'] === 'html') {\r
+ /* If the parser was originally created in order to handle the\r
+ setting of an element's innerHTML attribute, this is a parse error;\r
+ ignore the token. (The element will be an html element in this\r
+ case.) (innerHTML case) */\r
+\r
+ /* Otherwise, switch to the trailing end phase. */\r
+ $this->phase = self::END_PHASE;\r
+\r
+ /* Anything else */\r
+ } else {\r
+ /* Parse error. Set the insertion mode to "in body" and reprocess\r
+ the token. */\r
+ $this->mode = self::IN_BODY;\r
+ return $this->inBody($token);\r
+ }\r
+ }\r
+\r
+ private function inFrameset($token)\r
+ {\r
+ /* Handle the token as follows: */\r
+\r
+ /* A character token that is one of one of U+0009 CHARACTER TABULATION,\r
+ U+000A LINE FEED (LF), U+000B LINE TABULATION, U+000C FORM FEED (FF),\r
+ U+000D CARRIAGE RETURN (CR), or U+0020 SPACE */\r
+ if ($token['type'] === HTML5::CHARACTR &&\r
+ preg_match('/^[\t\n\x0b\x0c ]+$/', $token['data'])\r
+ ) {\r
+ /* Append the character to the current node. */\r
+ $this->insertText($token['data']);\r
+\r
+ /* A comment token */\r
+ } elseif ($token['type'] === HTML5::COMMENT) {\r
+ /* Append a Comment node to the current node with the data\r
+ attribute set to the data given in the comment token. */\r
+ $this->insertComment($token['data']);\r
+\r
+ /* A start tag with the tag name "frameset" */\r
+ } elseif ($token['name'] === 'frameset' &&\r
+ $token['type'] === HTML5::STARTTAG\r
+ ) {\r
+ $this->insertElement($token);\r
+\r
+ /* An end tag with the tag name "frameset" */\r
+ } elseif ($token['name'] === 'frameset' &&\r
+ $token['type'] === HTML5::ENDTAG\r
+ ) {\r
+ /* If the current node is the root html element, then this is a\r
+ parse error; ignore the token. (innerHTML case) */\r
+ if (end($this->stack)->nodeName === 'html') {\r
+ // Ignore\r
+\r
+ } else {\r
+ /* Otherwise, pop the current node from the stack of open\r
+ elements. */\r
+ array_pop($this->stack);\r
+\r
+ /* If the parser was not originally created in order to handle\r
+ the setting of an element's innerHTML attribute (innerHTML case),\r
+ and the current node is no longer a frameset element, then change\r
+ the insertion mode to "after frameset". */\r
+ $this->mode = self::AFTR_FRAME;\r
+ }\r
+\r
+ /* A start tag with the tag name "frame" */\r
+ } elseif ($token['name'] === 'frame' &&\r
+ $token['type'] === HTML5::STARTTAG\r
+ ) {\r
+ /* Insert an HTML element for the token. */\r
+ $this->insertElement($token);\r
+\r
+ /* Immediately pop the current node off the stack of open elements. */\r
+ array_pop($this->stack);\r
+\r
+ /* A start tag with the tag name "noframes" */\r
+ } elseif ($token['name'] === 'noframes' &&\r
+ $token['type'] === HTML5::STARTTAG\r
+ ) {\r
+ /* Process the token as if the insertion mode had been "in body". */\r
+ $this->inBody($token);\r
+\r
+ /* Anything else */\r
+ } else {\r
+ /* Parse error. Ignore the token. */\r
+ }\r
+ }\r
+\r
+ private function afterFrameset($token)\r
+ {\r
+ /* Handle the token as follows: */\r
+\r
+ /* A character token that is one of one of U+0009 CHARACTER TABULATION,\r
+ U+000A LINE FEED (LF), U+000B LINE TABULATION, U+000C FORM FEED (FF),\r
+ U+000D CARRIAGE RETURN (CR), or U+0020 SPACE */\r
+ if ($token['type'] === HTML5::CHARACTR &&\r
+ preg_match('/^[\t\n\x0b\x0c ]+$/', $token['data'])\r
+ ) {\r
+ /* Append the character to the current node. */\r
+ $this->insertText($token['data']);\r
+\r
+ /* A comment token */\r
+ } elseif ($token['type'] === HTML5::COMMENT) {\r
+ /* Append a Comment node to the current node with the data\r
+ attribute set to the data given in the comment token. */\r
+ $this->insertComment($token['data']);\r
+\r
+ /* An end tag with the tag name "html" */\r
+ } elseif ($token['name'] === 'html' &&\r
+ $token['type'] === HTML5::ENDTAG\r
+ ) {\r
+ /* Switch to the trailing end phase. */\r
+ $this->phase = self::END_PHASE;\r
+\r
+ /* A start tag with the tag name "noframes" */\r
+ } elseif ($token['name'] === 'noframes' &&\r
+ $token['type'] === HTML5::STARTTAG\r
+ ) {\r
+ /* Process the token as if the insertion mode had been "in body". */\r
+ $this->inBody($token);\r
+\r
+ /* Anything else */\r
+ } else {\r
+ /* Parse error. Ignore the token. */\r
+ }\r
+ }\r
+\r
+ private function trailingEndPhase($token)\r
+ {\r
+ /* After the main phase, as each token is emitted from the tokenisation\r
+ stage, it must be processed as described in this section. */\r
+\r
+ /* A DOCTYPE token */\r
+ if ($token['type'] === HTML5::DOCTYPE) {\r
+ // Parse error. Ignore the token.\r
+\r
+ /* A comment token */\r
+ } elseif ($token['type'] === HTML5::COMMENT) {\r
+ /* Append a Comment node to the Document object with the data\r
+ attribute set to the data given in the comment token. */\r
+ $comment = $this->dom->createComment($token['data']);\r
+ $this->dom->appendChild($comment);\r
+\r
+ /* A character token that is one of one of U+0009 CHARACTER TABULATION,\r
+ U+000A LINE FEED (LF), U+000B LINE TABULATION, U+000C FORM FEED (FF),\r
+ or U+0020 SPACE */\r
+ } elseif ($token['type'] === HTML5::CHARACTR &&\r
+ preg_match('/^[\t\n\x0b\x0c ]+$/', $token['data'])\r
+ ) {\r
+ /* Process the token as it would be processed in the main phase. */\r
+ $this->mainPhase($token);\r
+\r
+ /* A character token that is not one of U+0009 CHARACTER TABULATION,\r
+ U+000A LINE FEED (LF), U+000B LINE TABULATION, U+000C FORM FEED (FF),\r
+ or U+0020 SPACE. Or a start tag token. Or an end tag token. */\r
+ } elseif (($token['type'] === HTML5::CHARACTR &&\r
+ preg_match('/^[\t\n\x0b\x0c ]+$/', $token['data'])) ||\r
+ $token['type'] === HTML5::STARTTAG || $token['type'] === HTML5::ENDTAG\r
+ ) {\r
+ /* Parse error. Switch back to the main phase and reprocess the\r
+ token. */\r
+ $this->phase = self::MAIN_PHASE;\r
+ return $this->mainPhase($token);\r
+\r
+ /* An end-of-file token */\r
+ } elseif ($token['type'] === HTML5::EOF) {\r
+ /* OMG DONE!! */\r
+ }\r
+ }\r
+\r
+ private function insertElement($token, $append = true, $check = false)\r
+ {\r
+ // Proprietary workaround for libxml2's limitations with tag names\r
+ if ($check) {\r
+ // Slightly modified HTML5 tag-name modification,\r
+ // removing anything that's not an ASCII letter, digit, or hyphen\r
+ $token['name'] = preg_replace('/[^a-z0-9-]/i', '', $token['name']);\r
+ // Remove leading hyphens and numbers\r
+ $token['name'] = ltrim($token['name'], '-0..9');\r
+ // In theory, this should ever be needed, but just in case\r
+ if ($token['name'] === '') {\r
+ $token['name'] = 'span';\r
+ } // arbitrary generic choice\r
+ }\r
+\r
+ $el = $this->dom->createElement($token['name']);\r
+\r
+ foreach ($token['attr'] as $attr) {\r
+ if (!$el->hasAttribute($attr['name'])) {\r
+ $el->setAttribute($attr['name'], $attr['value']);\r
+ }\r
+ }\r
+\r
+ $this->appendToRealParent($el);\r
+ $this->stack[] = $el;\r
+\r
+ return $el;\r
+ }\r
+\r
+ private function insertText($data)\r
+ {\r
+ $text = $this->dom->createTextNode($data);\r
+ $this->appendToRealParent($text);\r
+ }\r
+\r
+ private function insertComment($data)\r
+ {\r
+ $comment = $this->dom->createComment($data);\r
+ $this->appendToRealParent($comment);\r
+ }\r
+\r
+ private function appendToRealParent($node)\r
+ {\r
+ if ($this->foster_parent === null) {\r
+ end($this->stack)->appendChild($node);\r
+\r
+ } elseif ($this->foster_parent !== null) {\r
+ /* If the foster parent element is the parent element of the\r
+ last table element in the stack of open elements, then the new\r
+ node must be inserted immediately before the last table element\r
+ in the stack of open elements in the foster parent element;\r
+ otherwise, the new node must be appended to the foster parent\r
+ element. */\r
+ for ($n = count($this->stack) - 1; $n >= 0; $n--) {\r
+ if ($this->stack[$n]->nodeName === 'table' &&\r
+ $this->stack[$n]->parentNode !== null\r
+ ) {\r
+ $table = $this->stack[$n];\r
+ break;\r
+ }\r
+ }\r
+\r
+ if (isset($table) && $this->foster_parent->isSameNode($table->parentNode)) {\r
+ $this->foster_parent->insertBefore($node, $table);\r
+ } else {\r
+ $this->foster_parent->appendChild($node);\r
+ }\r
+\r
+ $this->foster_parent = null;\r
+ }\r
+ }\r
+\r
+ private function elementInScope($el, $table = false)\r
+ {\r
+ if (is_array($el)) {\r
+ foreach ($el as $element) {\r
+ if ($this->elementInScope($element, $table)) {\r
+ return true;\r
+ }\r
+ }\r
+\r
+ return false;\r
+ }\r
+\r
+ $leng = count($this->stack);\r
+\r
+ for ($n = 0; $n < $leng; $n++) {\r
+ /* 1. Initialise node to be the current node (the bottommost node of\r
+ the stack). */\r
+ $node = $this->stack[$leng - 1 - $n];\r
+\r
+ if ($node->tagName === $el) {\r
+ /* 2. If node is the target node, terminate in a match state. */\r
+ return true;\r
+\r
+ } elseif ($node->tagName === 'table') {\r
+ /* 3. Otherwise, if node is a table element, terminate in a failure\r
+ state. */\r
+ return false;\r
+\r
+ } elseif ($table === true && in_array(\r
+ $node->tagName,\r
+ array(\r
+ 'caption',\r
+ 'td',\r
+ 'th',\r
+ 'button',\r
+ 'marquee',\r
+ 'object'\r
+ )\r
+ )\r
+ ) {\r
+ /* 4. Otherwise, if the algorithm is the "has an element in scope"\r
+ variant (rather than the "has an element in table scope" variant),\r
+ and node is one of the following, terminate in a failure state. */\r
+ return false;\r
+\r
+ } elseif ($node === $node->ownerDocument->documentElement) {\r
+ /* 5. Otherwise, if node is an html element (root element), terminate\r
+ in a failure state. (This can only happen if the node is the topmost\r
+ node of the stack of open elements, and prevents the next step from\r
+ being invoked if there are no more elements in the stack.) */\r
+ return false;\r
+ }\r
+\r
+ /* Otherwise, set node to the previous entry in the stack of open\r
+ elements and return to step 2. (This will never fail, since the loop\r
+ will always terminate in the previous step if the top of the stack\r
+ is reached.) */\r
+ }\r
+ }\r
+\r
+ private function reconstructActiveFormattingElements()\r
+ {\r
+ /* 1. If there are no entries in the list of active formatting elements,\r
+ then there is nothing to reconstruct; stop this algorithm. */\r
+ $formatting_elements = count($this->a_formatting);\r
+\r
+ if ($formatting_elements === 0) {\r
+ return false;\r
+ }\r
+\r
+ /* 3. Let entry be the last (most recently added) element in the list\r
+ of active formatting elements. */\r
+ $entry = end($this->a_formatting);\r
+\r
+ /* 2. If the last (most recently added) entry in the list of active\r
+ formatting elements is a marker, or if it is an element that is in the\r
+ stack of open elements, then there is nothing to reconstruct; stop this\r
+ algorithm. */\r
+ if ($entry === self::MARKER || in_array($entry, $this->stack, true)) {\r
+ return false;\r
+ }\r
+\r
+ for ($a = $formatting_elements - 1; $a >= 0; true) {\r
+ /* 4. If there are no entries before entry in the list of active\r
+ formatting elements, then jump to step 8. */\r
+ if ($a === 0) {\r
+ $step_seven = false;\r
+ break;\r
+ }\r
+\r
+ /* 5. Let entry be the entry one earlier than entry in the list of\r
+ active formatting elements. */\r
+ $a--;\r
+ $entry = $this->a_formatting[$a];\r
+\r
+ /* 6. If entry is neither a marker nor an element that is also in\r
+ thetack of open elements, go to step 4. */\r
+ if ($entry === self::MARKER || in_array($entry, $this->stack, true)) {\r
+ break;\r
+ }\r
+ }\r
+\r
+ while (true) {\r
+ /* 7. Let entry be the element one later than entry in the list of\r
+ active formatting elements. */\r
+ if (isset($step_seven) && $step_seven === true) {\r
+ $a++;\r
+ $entry = $this->a_formatting[$a];\r
+ }\r
+\r
+ /* 8. Perform a shallow clone of the element entry to obtain clone. */\r
+ $clone = $entry->cloneNode();\r
+\r
+ /* 9. Append clone to the current node and push it onto the stack\r
+ of open elements so that it is the new current node. */\r
+ end($this->stack)->appendChild($clone);\r
+ $this->stack[] = $clone;\r
+\r
+ /* 10. Replace the entry for entry in the list with an entry for\r
+ clone. */\r
+ $this->a_formatting[$a] = $clone;\r
+\r
+ /* 11. If the entry for clone in the list of active formatting\r
+ elements is not the last entry in the list, return to step 7. */\r
+ if (end($this->a_formatting) !== $clone) {\r
+ $step_seven = true;\r
+ } else {\r
+ break;\r
+ }\r
+ }\r
+ }\r
+\r
+ private function clearTheActiveFormattingElementsUpToTheLastMarker()\r
+ {\r
+ /* When the steps below require the UA to clear the list of active\r
+ formatting elements up to the last marker, the UA must perform the\r
+ following steps: */\r
+\r
+ while (true) {\r
+ /* 1. Let entry be the last (most recently added) entry in the list\r
+ of active formatting elements. */\r
+ $entry = end($this->a_formatting);\r
+\r
+ /* 2. Remove entry from the list of active formatting elements. */\r
+ array_pop($this->a_formatting);\r
+\r
+ /* 3. If entry was a marker, then stop the algorithm at this point.\r
+ The list has been cleared up to the last marker. */\r
+ if ($entry === self::MARKER) {\r
+ break;\r
+ }\r
+ }\r
+ }\r
+\r
+ private function generateImpliedEndTags($exclude = array())\r
+ {\r
+ /* When the steps below require the UA to generate implied end tags,\r
+ then, if the current node is a dd element, a dt element, an li element,\r
+ a p element, a td element, a th element, or a tr element, the UA must\r
+ act as if an end tag with the respective tag name had been seen and\r
+ then generate implied end tags again. */\r
+ $node = end($this->stack);\r
+ $elements = array_diff(array('dd', 'dt', 'li', 'p', 'td', 'th', 'tr'), $exclude);\r
+\r
+ while (in_array(end($this->stack)->nodeName, $elements)) {\r
+ array_pop($this->stack);\r
+ }\r
+ }\r
+\r
+ private function getElementCategory($node)\r
+ {\r
+ $name = $node->tagName;\r
+ if (in_array($name, $this->special)) {\r
+ return self::SPECIAL;\r
+ } elseif (in_array($name, $this->scoping)) {\r
+ return self::SCOPING;\r
+ } elseif (in_array($name, $this->formatting)) {\r
+ return self::FORMATTING;\r
+ } else {\r
+ return self::PHRASING;\r
+ }\r
+ }\r
+\r
+ private function clearStackToTableContext($elements)\r
+ {\r
+ /* When the steps above require the UA to clear the stack back to a\r
+ table context, it means that the UA must, while the current node is not\r
+ a table element or an html element, pop elements from the stack of open\r
+ elements. If this causes any elements to be popped from the stack, then\r
+ this is a parse error. */\r
+ while (true) {\r
+ $node = end($this->stack)->nodeName;\r
+\r
+ if (in_array($node, $elements)) {\r
+ break;\r
+ } else {\r
+ array_pop($this->stack);\r
+ }\r
+ }\r
+ }\r
+\r
+ private function resetInsertionMode()\r
+ {\r
+ /* 1. Let last be false. */\r
+ $last = false;\r
+ $leng = count($this->stack);\r
+\r
+ for ($n = $leng - 1; $n >= 0; $n--) {\r
+ /* 2. Let node be the last node in the stack of open elements. */\r
+ $node = $this->stack[$n];\r
+\r
+ /* 3. If node is the first node in the stack of open elements, then\r
+ set last to true. If the element whose innerHTML attribute is being\r
+ set is neither a td element nor a th element, then set node to the\r
+ element whose innerHTML attribute is being set. (innerHTML case) */\r
+ if ($this->stack[0]->isSameNode($node)) {\r
+ $last = true;\r
+ }\r
+\r
+ /* 4. If node is a select element, then switch the insertion mode to\r
+ "in select" and abort these steps. (innerHTML case) */\r
+ if ($node->nodeName === 'select') {\r
+ $this->mode = self::IN_SELECT;\r
+ break;\r
+\r
+ /* 5. If node is a td or th element, then switch the insertion mode\r
+ to "in cell" and abort these steps. */\r
+ } elseif ($node->nodeName === 'td' || $node->nodeName === 'th') {\r
+ $this->mode = self::IN_CELL;\r
+ break;\r
+\r
+ /* 6. If node is a tr element, then switch the insertion mode to\r
+ "in row" and abort these steps. */\r
+ } elseif ($node->nodeName === 'tr') {\r
+ $this->mode = self::IN_ROW;\r
+ break;\r
+\r
+ /* 7. If node is a tbody, thead, or tfoot element, then switch the\r
+ insertion mode to "in table body" and abort these steps. */\r
+ } elseif (in_array($node->nodeName, array('tbody', 'thead', 'tfoot'))) {\r
+ $this->mode = self::IN_TBODY;\r
+ break;\r
+\r
+ /* 8. If node is a caption element, then switch the insertion mode\r
+ to "in caption" and abort these steps. */\r
+ } elseif ($node->nodeName === 'caption') {\r
+ $this->mode = self::IN_CAPTION;\r
+ break;\r
+\r
+ /* 9. If node is a colgroup element, then switch the insertion mode\r
+ to "in column group" and abort these steps. (innerHTML case) */\r
+ } elseif ($node->nodeName === 'colgroup') {\r
+ $this->mode = self::IN_CGROUP;\r
+ break;\r
+\r
+ /* 10. If node is a table element, then switch the insertion mode\r
+ to "in table" and abort these steps. */\r
+ } elseif ($node->nodeName === 'table') {\r
+ $this->mode = self::IN_TABLE;\r
+ break;\r
+\r
+ /* 11. If node is a head element, then switch the insertion mode\r
+ to "in body" ("in body"! not "in head"!) and abort these steps.\r
+ (innerHTML case) */\r
+ } elseif ($node->nodeName === 'head') {\r
+ $this->mode = self::IN_BODY;\r
+ break;\r
+\r
+ /* 12. If node is a body element, then switch the insertion mode to\r
+ "in body" and abort these steps. */\r
+ } elseif ($node->nodeName === 'body') {\r
+ $this->mode = self::IN_BODY;\r
+ break;\r
+\r
+ /* 13. If node is a frameset element, then switch the insertion\r
+ mode to "in frameset" and abort these steps. (innerHTML case) */\r
+ } elseif ($node->nodeName === 'frameset') {\r
+ $this->mode = self::IN_FRAME;\r
+ break;\r
+\r
+ /* 14. If node is an html element, then: if the head element\r
+ pointer is null, switch the insertion mode to "before head",\r
+ otherwise, switch the insertion mode to "after head". In either\r
+ case, abort these steps. (innerHTML case) */\r
+ } elseif ($node->nodeName === 'html') {\r
+ $this->mode = ($this->head_pointer === null)\r
+ ? self::BEFOR_HEAD\r
+ : self::AFTER_HEAD;\r
+\r
+ break;\r
+\r
+ /* 15. If last is true, then set the insertion mode to "in body"\r
+ and abort these steps. (innerHTML case) */\r
+ } elseif ($last) {\r
+ $this->mode = self::IN_BODY;\r
+ break;\r
+ }\r
+ }\r
+ }\r
+\r
+ private function closeCell()\r
+ {\r
+ /* If the stack of open elements has a td or th element in table scope,\r
+ then act as if an end tag token with that tag name had been seen. */\r
+ foreach (array('td', 'th') as $cell) {\r
+ if ($this->elementInScope($cell, true)) {\r
+ $this->inCell(\r
+ array(\r
+ 'name' => $cell,\r
+ 'type' => HTML5::ENDTAG\r
+ )\r
+ );\r
+\r
+ break;\r
+ }\r
+ }\r
+ }\r
+\r
+ public function save()\r
+ {\r
+ return $this->dom;\r
+ }\r
+}\r
git://git.planwatch.world / planworld_public.git / commitdiff