onetipp - blog.onetop.com - libraryv3 - htmlpurifier - SOURCE CODE

<?php

/**
 * Converts a stream of HTMLPurifier_Token into an HTMLPurifier_Node,
 * and back again.
 *
 * @note This transformation is not an equivalence.  We mutate the input
 * token stream to make it so; see all [MUT] markers in code.
 */
class HTMLPurifier_Arborize
{
    public static function arborize($tokens, $config, $context) {
        $definition = $config->getHTMLDefinition();
        $parent = new HTMLPurifier_Token_Start($definition->info_parent);
        $stack = array($parent->toNode());
        foreach ($tokens as $token) {
            $token->skip = null; // [MUT]
            $token->carryover = null; // [MUT]
            if ($token instanceof HTMLPurifier_Token_End) {
                $token->start = null; // [MUT]
                $r = array_pop($stack);
                assert($r->name === $token->name);
                assert(empty($token->attr));
                $r->endCol = $token->col;
                $r->endLine = $token->line;
                $r->endArmor = $token->armor;
                continue;
            }
            $node = $token->toNode();
            $stack[count($stack)-1]->children[] = $node;
            if ($token instanceof HTMLPurifier_Token_Start) {
                $stack[] = $node;
            }
        }
        assert(count($stack) == 1);
        return $stack[0];
    }

    public static function flatten($node, $config, $context) {
        $level = 0;
        $nodes = array($level => new HTMLPurifier_Queue(array($node)));
        $closingTokens = array();
        $tokens = array();
        do {
            while (!$nodes[$level]->isEmpty()) {
                $node = $nodes[$level]->shift(); // FIFO
                list($start, $end) = $node->toTokenPair();
                if ($level > 0) {
                    $tokens[] = $start;
                }

... [truncated, 22 more lines] ...

<?php

/**
 * Defines common attribute collections that modules reference
 */

class HTMLPurifier_AttrCollections
{

    /**
     * Associative array of attribute collections, indexed by name.
     * @type array
     */
    public $info = array();

    /**
     * Performs all expansions on internal data for use by other inclusions
     * It also collects all attribute collection extensions from
     * modules
     * @param HTMLPurifier_AttrTypes $attr_types HTMLPurifier_AttrTypes instance
     * @param HTMLPurifier_HTMLModule[] $modules Hash array of HTMLPurifier_HTMLModule members
     */
    public function __construct($attr_types, $modules)
    {
        $this->doConstruct($attr_types, $modules);
    }

    public function doConstruct($attr_types, $modules)
    {
        // load extensions from the modules
        foreach ($modules as $module) {
            foreach ($module->attr_collections as $coll_i => $coll) {
                if (!isset($this->info[$coll_i])) {
                    $this->info[$coll_i] = array();
                }
                foreach ($coll as $attr_i => $attr) {
                    if ($attr_i === 0 && isset($this->info[$coll_i][$attr_i])) {
                        // merge in includes
                        $this->info[$coll_i][$attr_i] = array_merge(
                            $this->info[$coll_i][$attr_i],
                            $attr
                        );
                        continue;
                    }
                    $this->info[$coll_i][$attr_i] = $attr;
                }
            }
        }
        // perform internal expansions and inclusions
        foreach ($this->info as $name => $attr) {

... [truncated, 99 more lines] ...

<?php

/**
 * Base class for all validating attribute definitions.
 *
 * This family of classes forms the core for not only HTML attribute validation,
 * but also any sort of string that needs to be validated or cleaned (which
 * means CSS properties and composite definitions are defined here too).
 * Besides defining (through code) what precisely makes the string valid,
 * subclasses are also responsible for cleaning the code if possible.
 */

abstract class HTMLPurifier_AttrDef
{

    /**
     * Tells us whether or not an HTML attribute is minimized.
     * Has no meaning in other contexts.
     * @type bool
     */
    public $minimized = false;

    /**
     * Tells us whether or not an HTML attribute is required.
     * Has no meaning in other contexts
     * @type bool
     */
    public $required = false;

    /**
     * Validates and cleans passed string according to a definition.
     *
     * @param string $string String to be validated and cleaned.
     * @param HTMLPurifier_Config $config Mandatory HTMLPurifier_Config object.
     * @param HTMLPurifier_Context $context Mandatory HTMLPurifier_Context object.
     */
    abstract public function validate($string, $config, $context);

    /**
     * Convenience method that parses a string as if it were CDATA.
     *
     * This method process a string in the manner specified at
     * <http://www.w3.org/TR/html4/types.html#h-6.2> by removing
     * leading and trailing whitespace, ignoring line feeds, and replacing
     * carriage returns and tabs with spaces.  While most useful for HTML
     * attributes specified as CDATA, it can also be applied to most CSS
     * values.
     *
     * @note This method is not entirely standards compliant, as trim() removes
     *       more types of whitespace than specified in the spec. In practice,

... [truncated, 89 more lines] ...

<?php

/**
 * Processes an entire attribute array for corrections needing multiple values.
 *
 * Occasionally, a certain attribute will need to be removed and popped onto
 * another value.  Instead of creating a complex return syntax for
 * HTMLPurifier_AttrDef, we just pass the whole attribute array to a
 * specialized object and have that do the special work.  That is the
 * family of HTMLPurifier_AttrTransform.
 *
 * An attribute transformation can be assigned to run before or after
 * HTMLPurifier_AttrDef validation.  See HTMLPurifier_HTMLDefinition for
 * more details.
 */

abstract class HTMLPurifier_AttrTransform
{

    /**
     * Abstract: makes changes to the attributes dependent on multiple values.
     *
     * @param array $attr Assoc array of attributes, usually from
     *              HTMLPurifier_Token_Tag::$attr
     * @param HTMLPurifier_Config $config Mandatory HTMLPurifier_Config object.
     * @param HTMLPurifier_Context $context Mandatory HTMLPurifier_Context object
     * @return array Processed attribute array.
     */
    abstract public function transform($attr, $config, $context);

    /**
     * Prepends CSS properties to the style attribute, creating the
     * attribute if it doesn't exist.
     * @param array &$attr Attribute array to process (passed by reference)
     * @param string $css CSS to prepend
     */
    public function prependCSS(&$attr, $css)
    {
        $attr['style'] = isset($attr['style']) ? $attr['style'] : '';
        $attr['style'] = $css . $attr['style'];
    }

    /**
     * Retrieves and removes an attribute
     * @param array &$attr Attribute array to process (passed by reference)
     * @param mixed $key Key of attribute to confiscate
     * @return mixed
     */
    public function confiscateAttr(&$attr, $key)
    {

... [truncated, 11 more lines] ...

<?php

/**
 * Provides lookup array of attribute types to HTMLPurifier_AttrDef objects
 */
class HTMLPurifier_AttrTypes
{
    /**
     * Lookup array of attribute string identifiers to concrete implementations.
     * @type HTMLPurifier_AttrDef[]
     */
    protected $info = array();

    /**
     * Constructs the info array, supplying default implementations for attribute
     * types.
     */
    public function __construct()
    {
        // XXX This is kind of poor, since we don't actually /clone/
        // instances; instead, we use the supplied make() attribute. So,
        // the underlying class must know how to deal with arguments.
        // With the old implementation of Enum, that ignored its
        // arguments when handling a make dispatch, the IAlign
        // definition wouldn't work.

        // pseudo-types, must be instantiated via shorthand
        $this->info['Enum']    = new HTMLPurifier_AttrDef_Enum();
        $this->info['Bool']    = new HTMLPurifier_AttrDef_HTML_Bool();

        $this->info['CDATA']    = new HTMLPurifier_AttrDef_Text();
        $this->info['ID']       = new HTMLPurifier_AttrDef_HTML_ID();
        $this->info['Length']   = new HTMLPurifier_AttrDef_HTML_Length();
        $this->info['MultiLength'] = new HTMLPurifier_AttrDef_HTML_MultiLength();
        $this->info['NMTOKENS'] = new HTMLPurifier_AttrDef_HTML_Nmtokens();
        $this->info['Pixels']   = new HTMLPurifier_AttrDef_HTML_Pixels();
        $this->info['Text']     = new HTMLPurifier_AttrDef_Text();
        $this->info['URI']      = new HTMLPurifier_AttrDef_URI();
        $this->info['LanguageCode'] = new HTMLPurifier_AttrDef_Lang();
        $this->info['Color']    = new HTMLPurifier_AttrDef_HTML_Color();
        $this->info['IAlign']   = self::makeEnum('top,middle,bottom,left,right');
        $this->info['LAlign']   = self::makeEnum('top,bottom,left,right');
        $this->info['FrameTarget'] = new HTMLPurifier_AttrDef_HTML_FrameTarget();

        // unimplemented aliases
        $this->info['ContentType'] = new HTMLPurifier_AttrDef_Text();
        $this->info['ContentTypes'] = new HTMLPurifier_AttrDef_Text();
        $this->info['Charsets'] = new HTMLPurifier_AttrDef_Text();
        $this->info['Character'] = new HTMLPurifier_AttrDef_Text();


... [truncated, 47 more lines] ...

<?php

/**
 * Validates the attributes of a token. Doesn't manage required attributes
 * very well. The only reason we factored this out was because RemoveForeignElements
 * also needed it besides ValidateAttributes.
 */
class HTMLPurifier_AttrValidator
{

    /**
     * Validates the attributes of a token, mutating it as necessary.
     * that has valid tokens
     * @param HTMLPurifier_Token $token Token to validate.
     * @param HTMLPurifier_Config $config Instance of HTMLPurifier_Config
     * @param HTMLPurifier_Context $context Instance of HTMLPurifier_Context
     */
    public function validateToken($token, $config, $context)
    {
        $definition = $config->getHTMLDefinition();
        $e =& $context->get('ErrorCollector', true);

        // initialize IDAccumulator if necessary
        $ok =& $context->get('IDAccumulator', true);
        if (!$ok) {
            $id_accumulator = HTMLPurifier_IDAccumulator::build($config, $context);
            $context->register('IDAccumulator', $id_accumulator);
        }

        // initialize CurrentToken if necessary
        $current_token =& $context->get('CurrentToken', true);
        if (!$current_token) {
            $context->register('CurrentToken', $token);
        }

        if (!$token instanceof HTMLPurifier_Token_Start &&
            !$token instanceof HTMLPurifier_Token_Empty
        ) {
            return;
        }

        // create alias to global definition array, see also $defs
        // DEFINITION CALL
        $d_defs = $definition->info_global_attr;

        // don't update token until the very end, to ensure an atomic update
        $attr = $token->attr;

        // do global transformations (pre)
        // nothing currently utilizes this

... [truncated, 129 more lines] ...

<?php

// constants are slow, so we use as few as possible
if (!defined('HTMLPURIFIER_PREFIX')) {
    define('HTMLPURIFIER_PREFIX', realpath(dirname(__FILE__) . '/..'));
}

// accomodations for versions earlier than 5.0.2
// borrowed from PHP_Compat, LGPL licensed, by Aidan Lister <aidan@php.net>
if (!defined('PHP_EOL')) {
    switch (strtoupper(substr(PHP_OS, 0, 3))) {
        case 'WIN':
            define('PHP_EOL', "\r\n");
            break;
        case 'DAR':
            define('PHP_EOL', "\r");
            break;
        default:
            define('PHP_EOL', "\n");
    }
}

/**
 * Bootstrap class that contains meta-functionality for HTML Purifier such as
 * the autoload function.
 *
 * @note
 *      This class may be used without any other files from HTML Purifier.
 */
class HTMLPurifier_Bootstrap
{

    /**
     * Autoload function for HTML Purifier
     * @param string $class Class to load
     * @return bool
     */
    public static function autoload($class)
    {
        $file = HTMLPurifier_Bootstrap::getPath($class);
        if (!$file) {
            return false;
        }
        // Technically speaking, it should be ok and more efficient to
        // just do 'require', but Antonio Parraga reports that with
        // Zend extensions such as Zend debugger and APC, this invariant
        // may be broken.  Since we have efficient alternatives, pay
        // the cost here and avoid the bug.
        require_once HTMLPURIFIER_PREFIX . '/' . $file;
        return true;

... [truncated, 75 more lines] ...

<?php

/**
 * Defines allowed CSS attributes and what their values are.
 * @see HTMLPurifier_HTMLDefinition
 */
class HTMLPurifier_CSSDefinition extends HTMLPurifier_Definition
{

    public $type = 'CSS';

    /**
     * Assoc array of attribute name to definition object.
     * @type HTMLPurifier_AttrDef[]
     */
    public $info = array();

    /**
     * Constructs the info array.  The meat of this class.
     * @param HTMLPurifier_Config $config
     */
    protected function doSetup($config)
    {
        $this->info['text-align'] = new HTMLPurifier_AttrDef_Enum(
            array('left', 'right', 'center', 'justify'),
            false
        );

        $border_style =
            $this->info['border-bottom-style'] =
            $this->info['border-right-style'] =
            $this->info['border-left-style'] =
            $this->info['border-top-style'] = new HTMLPurifier_AttrDef_Enum(
                array(
                    'none',
                    'hidden',
                    'dotted',
                    'dashed',
                    'solid',
                    'double',
                    'groove',
                    'ridge',
                    'inset',
                    'outset'
                ),
                false
            );

        $this->info['border-style'] = new HTMLPurifier_AttrDef_CSS_Multiple($border_style);


... [truncated, 438 more lines] ...

<?php

/**
 * Defines allowed child nodes and validates nodes against it.
 */
abstract class HTMLPurifier_ChildDef
{
    /**
     * Type of child definition, usually right-most part of class name lowercase.
     * Used occasionally in terms of context.
     * @type string
     */
    public $type;

    /**
     * Indicates whether or not an empty array of children is okay.
     *
     * This is necessary for redundant checking when changes affecting
     * a child node may cause a parent node to now be disallowed.
     * @type bool
     */
    public $allow_empty;

    /**
     * Lookup array of all elements that this definition could possibly allow.
     * @type array
     */
    public $elements = array();

    /**
     * Get lookup of tag names that should not close this element automatically.
     * All other elements will do so.
     * @param HTMLPurifier_Config $config HTMLPurifier_Config object
     * @return array
     */
    public function getAllowedElements($config)
    {
        return $this->elements;
    }

    /**
     * Validates nodes according to definition and returns modification.
     *
     * @param HTMLPurifier_Node[] $children Array of HTMLPurifier_Node
     * @param HTMLPurifier_Config $config HTMLPurifier_Config object
     * @param HTMLPurifier_Context $context HTMLPurifier_Context object
     * @return bool|array true to leave nodes as is, false to remove parent node, array of replacement children
     */
    abstract public function validateChildren($children, $config, $context);
}

... [truncated, 3 more lines] ...

<?php

/**
 * Configuration object that triggers customizable behavior.
 *
 * @warning This class is strongly defined: that means that the class
 *          will fail if an undefined directive is retrieved or set.
 *
 * @note Many classes that could (although many times don't) use the
 *       configuration object make it a mandatory parameter.  This is
 *       because a configuration object should always be forwarded,
 *       otherwise, you run the risk of missing a parameter and then
 *       being stumped when a configuration directive doesn't work.
 *
 * @todo Reconsider some of the public member variables
 */
class HTMLPurifier_Config
{

    /**
     * HTML Purifier's version
     * @type string
     */
    public $version = '4.8.0';

    /**
     * Whether or not to automatically finalize
     * the object if a read operation is done.
     * @type bool
     */
    public $autoFinalize = true;

    // protected member variables

    /**
     * Namespace indexed array of serials for specific namespaces.
     * @see getSerial() for more info.
     * @type string[]
     */
    protected $serials = array();

    /**
     * Serial for entire configuration object.
     * @type string
     */
    protected $serial;

    /**
     * Parser for variables.
     * @type HTMLPurifier_VarParser_Flexible

... [truncated, 871 more lines] ...

<?php

/**
 * Configuration definition, defines directives and their defaults.
 */
class HTMLPurifier_ConfigSchema
{
    /**
     * Defaults of the directives and namespaces.
     * @type array
     * @note This shares the exact same structure as HTMLPurifier_Config::$conf
     */
    public $defaults = array();

    /**
     * The default property list. Do not edit this property list.
     * @type array
     */
    public $defaultPlist;

    /**
     * Definition of the directives.
     * The structure of this is:
     *
     *  array(
     *      'Namespace' => array(
     *          'Directive' => new stdclass(),
     *      )
     *  )
     *
     * The stdclass may have the following properties:
     *
     *  - If isAlias isn't set:
     *      - type: Integer type of directive, see HTMLPurifier_VarParser for definitions
     *      - allow_null: If set, this directive allows null values
     *      - aliases: If set, an associative array of value aliases to real values
     *      - allowed: If set, a lookup array of allowed (string) values
     *  - If isAlias is set:
     *      - namespace: Namespace this directive aliases to
     *      - name: Directive name this directive aliases to
     *
     * In certain degenerate cases, stdclass will actually be an integer. In
     * that case, the value is equivalent to an stdclass with the type
     * property set to the integer. If the integer is negative, type is
     * equal to the absolute value of integer, and allow_null is true.
     *
     * This class is friendly with HTMLPurifier_Config. If you need introspection
     * about the schema, you're better of using the ConfigSchema_Interchange,
     * which uses more memory but has much richer information.
     * @type array

... [truncated, 127 more lines] ...

<?php

/**
 * @todo Unit test
 */
class HTMLPurifier_ContentSets
{

    /**
     * List of content set strings (pipe separators) indexed by name.
     * @type array
     */
    public $info = array();

    /**
     * List of content set lookups (element => true) indexed by name.
     * @type array
     * @note This is in HTMLPurifier_HTMLDefinition->info_content_sets
     */
    public $lookup = array();

    /**
     * Synchronized list of defined content sets (keys of info).
     * @type array
     */
    protected $keys = array();
    /**
     * Synchronized list of defined content values (values of info).
     * @type array
     */
    protected $values = array();

    /**
     * Merges in module's content sets, expands identifiers in the content
     * sets and populates the keys, values and lookup member variables.
     * @param HTMLPurifier_HTMLModule[] $modules List of HTMLPurifier_HTMLModule
     */
    public function __construct($modules)
    {
        if (!is_array($modules)) {
            $modules = array($modules);
        }
        // populate content_sets based on module hints
        // sorry, no way of overloading
        foreach ($modules as $module) {
            foreach ($module->content_sets as $key => $value) {
                $temp = $this->convertToLookup($value);
                if (isset($this->lookup[$key])) {
                    // add it into the existing content set
                    $this->lookup[$key] = array_merge($this->lookup[$key], $temp);

... [truncated, 121 more lines] ...

<?php

/**
 * Registry object that contains information about the current context.
 * @warning Is a bit buggy when variables are set to null: it thinks
 *          they don't exist! So use false instead, please.
 * @note Since the variables Context deals with may not be objects,
 *       references are very important here! Do not remove!
 */
class HTMLPurifier_Context
{

    /**
     * Private array that stores the references.
     * @type array
     */
    private $_storage = array();

    /**
     * Registers a variable into the context.
     * @param string $name String name
     * @param mixed $ref Reference to variable to be registered
     */
    public function register($name, &$ref)
    {
        if (array_key_exists($name, $this->_storage)) {
            trigger_error(
                "Name $name produces collision, cannot re-register",
                E_USER_ERROR
            );
            return;
        }
        $this->_storage[$name] =& $ref;
    }

    /**
     * Retrieves a variable reference from the context.
     * @param string $name String name
     * @param bool $ignore_error Boolean whether or not to ignore error
     * @return mixed
     */
    public function &get($name, $ignore_error = false)
    {
        if (!array_key_exists($name, $this->_storage)) {
            if (!$ignore_error) {
                trigger_error(
                    "Attempted to retrieve non-existent variable $name",
                    E_USER_ERROR
                );
            }

... [truncated, 46 more lines] ...

<?php

/**
 * Super-class for definition datatype objects, implements serialization
 * functions for the class.
 */
abstract class HTMLPurifier_Definition
{

    /**
     * Has setup() been called yet?
     * @type bool
     */
    public $setup = false;

    /**
     * If true, write out the final definition object to the cache after
     * setup.  This will be true only if all invocations to get a raw
     * definition object are also optimized.  This does not cause file
     * system thrashing because on subsequent calls the cached object
     * is used and any writes to the raw definition object are short
     * circuited.  See enduser-customize.html for the high-level
     * picture.
     * @type bool
     */
    public $optimized = null;

    /**
     * What type of definition is it?
     * @type string
     */
    public $type;

    /**
     * Sets up the definition object into the final form, something
     * not done by the constructor
     * @param HTMLPurifier_Config $config
     */
    abstract protected function doSetup($config);

    /**
     * Setup function that aborts if already setup
     * @param HTMLPurifier_Config $config
     */
    public function setup($config)
    {
        if ($this->setup) {
            return;
        }
        $this->setup = true;

... [truncated, 6 more lines] ...

<?php

/**
 * Abstract class representing Definition cache managers that implements
 * useful common methods and is a factory.
 * @todo Create a separate maintenance file advanced users can use to
 *       cache their custom HTMLDefinition, which can be loaded
 *       via a configuration directive
 * @todo Implement memcached
 */
abstract class HTMLPurifier_DefinitionCache
{
    /**
     * @type string
     */
    public $type;

    /**
     * @param string $type Type of definition objects this instance of the
     *      cache will handle.
     */
    public function __construct($type)
    {
        $this->type = $type;
    }

    /**
     * Generates a unique identifier for a particular configuration
     * @param HTMLPurifier_Config $config Instance of HTMLPurifier_Config
     * @return string
     */
    public function generateKey($config)
    {
        return $config->version . ',' . // possibly replace with function calls
               $config->getBatchSerial($this->type) . ',' .
               $config->get($this->type . '.DefinitionRev');
    }

    /**
     * Tests whether or not a key is old with respect to the configuration's
     * version and revision number.
     * @param string $key Key to test
     * @param HTMLPurifier_Config $config Instance of HTMLPurifier_Config to test against
     * @return bool
     */
    public function isOld($key, $config)
    {
        if (substr_count($key, ',') < 2) {
            return true;
        }

... [truncated, 80 more lines] ...

<?php

/**
 * Responsible for creating definition caches.
 */
class HTMLPurifier_DefinitionCacheFactory
{
    /**
     * @type array
     */
    protected $caches = array('Serializer' => array());

    /**
     * @type array
     */
    protected $implementations = array();

    /**
     * @type HTMLPurifier_DefinitionCache_Decorator[]
     */
    protected $decorators = array();

    /**
     * Initialize default decorators
     */
    public function setup()
    {
        $this->addDecorator('Cleanup');
    }

    /**
     * Retrieves an instance of global definition cache factory.
     * @param HTMLPurifier_DefinitionCacheFactory $prototype
     * @return HTMLPurifier_DefinitionCacheFactory
     */
    public static function instance($prototype = null)
    {
        static $instance;
        if ($prototype !== null) {
            $instance = $prototype;
        } elseif ($instance === null || $prototype === true) {
            $instance = new HTMLPurifier_DefinitionCacheFactory();
            $instance->setup();
        }
        return $instance;
    }

    /**
     * Registers a new definition cache object
     * @param string $short Short name of cache object, for reference

... [truncated, 57 more lines] ...

<?php

/**
 * Represents a document type, contains information on which modules
 * need to be loaded.
 * @note This class is inspected by Printer_HTMLDefinition->renderDoctype.
 *       If structure changes, please update that function.
 */
class HTMLPurifier_Doctype
{
    /**
     * Full name of doctype
     * @type string
     */
    public $name;

    /**
     * List of standard modules (string identifiers or literal objects)
     * that this doctype uses
     * @type array
     */
    public $modules = array();

    /**
     * List of modules to use for tidying up code
     * @type array
     */
    public $tidyModules = array();

    /**
     * Is the language derived from XML (i.e. XHTML)?
     * @type bool
     */
    public $xml = true;

    /**
     * List of aliases for this doctype
     * @type array
     */
    public $aliases = array();

    /**
     * Public DTD identifier
     * @type string
     */
    public $dtdPublic;

    /**
     * System DTD identifier
     * @type string

... [truncated, 24 more lines] ...

<?php

class HTMLPurifier_DoctypeRegistry
{

    /**
     * Hash of doctype names to doctype objects.
     * @type array
     */
    protected $doctypes;

    /**
     * Lookup table of aliases to real doctype names.
     * @type array
     */
    protected $aliases;

    /**
     * Registers a doctype to the registry
     * @note Accepts a fully-formed doctype object, or the
     *       parameters for constructing a doctype object
     * @param string $doctype Name of doctype or literal doctype object
     * @param bool $xml
     * @param array $modules Modules doctype will load
     * @param array $tidy_modules Modules doctype will load for certain modes
     * @param array $aliases Alias names for doctype
     * @param string $dtd_public
     * @param string $dtd_system
     * @return HTMLPurifier_Doctype Editable registered doctype
     */
    public function register(
        $doctype,
        $xml = true,
        $modules = array(),
        $tidy_modules = array(),
        $aliases = array(),
        $dtd_public = null,
        $dtd_system = null
    ) {
        if (!is_array($modules)) {
            $modules = array($modules);
        }
        if (!is_array($tidy_modules)) {
            $tidy_modules = array($tidy_modules);
        }
        if (!is_array($aliases)) {
            $aliases = array($aliases);
        }
        if (!is_object($doctype)) {
            $doctype = new HTMLPurifier_Doctype(

... [truncated, 93 more lines] ...

<?php

/**
 * Structure that stores an HTML element definition. Used by
 * HTMLPurifier_HTMLDefinition and HTMLPurifier_HTMLModule.
 * @note This class is inspected by HTMLPurifier_Printer_HTMLDefinition.
 *       Please update that class too.
 * @warning If you add new properties to this class, you MUST update
 *          the mergeIn() method.
 */
class HTMLPurifier_ElementDef
{
    /**
     * Does the definition work by itself, or is it created solely
     * for the purpose of merging into another definition?
     * @type bool
     */
    public $standalone = true;

    /**
     * Associative array of attribute name to HTMLPurifier_AttrDef.
     * @type array
     * @note Before being processed by HTMLPurifier_AttrCollections
     *       when modules are finalized during
     *       HTMLPurifier_HTMLDefinition->setup(), this array may also
     *       contain an array at index 0 that indicates which attribute
     *       collections to load into the full array. It may also
     *       contain string indentifiers in lieu of HTMLPurifier_AttrDef,
     *       see HTMLPurifier_AttrTypes on how they are expanded during
     *       HTMLPurifier_HTMLDefinition->setup() processing.
     */
    public $attr = array();

    // XXX: Design note: currently, it's not possible to override
    // previously defined AttrTransforms without messing around with
    // the final generated config. This is by design; a previous version
    // used an associated list of attr_transform, but it was extremely
    // easy to accidentally override other attribute transforms by
    // forgetting to specify an index (and just using 0.)  While we
    // could check this by checking the index number and complaining,
    // there is a second problem which is that it is not at all easy to
    // tell when something is getting overridden. Combine this with a
    // codebase where this isn't really being used, and it's perfect for
    // nuking.

    /**
     * List of tags HTMLPurifier_AttrTransform to be done before validation.
     * @type array
     */
    public $attr_transform_pre = array();

... [truncated, 167 more lines] ...

<?php

/**
 * A UTF-8 specific character encoder that handles cleaning and transforming.
 * @note All functions in this class should be static.
 */
class HTMLPurifier_Encoder
{

    /**
     * Constructor throws fatal error if you attempt to instantiate class
     */
    private function __construct()
    {
        trigger_error('Cannot instantiate encoder, call methods statically', E_USER_ERROR);
    }

    /**
     * Error-handler that mutes errors, alternative to shut-up operator.
     */
    public static function muteErrorHandler()
    {
    }

    /**
     * iconv wrapper which mutes errors, but doesn't work around bugs.
     * @param string $in Input encoding
     * @param string $out Output encoding
     * @param string $text The text to convert
     * @return string
     */
    public static function unsafeIconv($in, $out, $text)
    {
        set_error_handler(array('HTMLPurifier_Encoder', 'muteErrorHandler'));
        $r = iconv($in, $out, $text);
        restore_error_handler();
        return $r;
    }

    /**
     * iconv wrapper which mutes errors and works around bugs.
     * @param string $in Input encoding
     * @param string $out Output encoding
     * @param string $text The text to convert
     * @param int $max_chunk_size
     * @return string
     */
    public static function iconv($in, $out, $text, $max_chunk_size = 8000)
    {
        $code = self::testIconvTruncateBug();

... [truncated, 562 more lines] ...

<?php

/**
 * Object that provides entity lookup table from entity name to character
 */
class HTMLPurifier_EntityLookup
{
    /**
     * Assoc array of entity name to character represented.
     * @type array
     */
    public $table;

    /**
     * Sets up the entity lookup table from the serialized file contents.
     * @param bool $file
     * @note The serialized contents are versioned, but were generated
     *       using the maintenance script generate_entity_file.php
     * @warning This is not in constructor to help enforce the Singleton
     */
    public function setup($file = false)
    {
        if (!$file) {
            $file = HTMLPURIFIER_PREFIX . '/HTMLPurifier/EntityLookup/entities.ser';
        }
        $this->table = unserialize(file_get_contents($file));
    }

    /**
     * Retrieves sole instance of the object.
     * @param bool|HTMLPurifier_EntityLookup $prototype Optional prototype of custom lookup table to overload with.
     * @return HTMLPurifier_EntityLookup
     */
    public static function instance($prototype = false)
    {
        // no references, since PHP doesn't copy unless modified
        static $instance = null;
        if ($prototype) {
            $instance = $prototype;
        } elseif (!$instance) {
            $instance = new HTMLPurifier_EntityLookup();
            $instance->setup();
        }
        return $instance;
    }
}

// vim: et sw=4 sts=4

<?php

// if want to implement error collecting here, we'll need to use some sort
// of global data (probably trigger_error) because it's impossible to pass
// $config or $context to the callback functions.

/**
 * Handles referencing and derefencing character entities
 */
class HTMLPurifier_EntityParser
{

    /**
     * Reference to entity lookup table.
     * @type HTMLPurifier_EntityLookup
     */
    protected $_entity_lookup;

    /**
     * Callback regex string for parsing entities.
     * @type string
     */
    protected $_substituteEntitiesRegex =
        '/&(?:[#]x([a-fA-F0-9]+)|[#]0*(\d+)|([A-Za-z_:][A-Za-z0-9.\-_:]*));?/';
        //     1. hex             2. dec      3. string (XML style)

    /**
     * Decimal to parsed string conversion table for special entities.
     * @type array
     */
    protected $_special_dec2str =
            array(
                    34 => '"',
                    38 => '&',
                    39 => "'",
                    60 => '<',
                    62 => '>'
            );

    /**
     * Stripped entity names to decimal conversion table for special entities.
     * @type array
     */
    protected $_special_ent2dec =
            array(
                    'quot' => 34,
                    'amp'  => 38,
                    'lt'   => 60,
                    'gt'   => 62
            );

... [truncated, 104 more lines] ...

<?php

/**
 * Error collection class that enables HTML Purifier to report HTML
 * problems back to the user
 */
class HTMLPurifier_ErrorCollector
{

    /**
     * Identifiers for the returned error array. These are purposely numeric
     * so list() can be used.
     */
    const LINENO   = 0;
    const SEVERITY = 1;
    const MESSAGE  = 2;
    const CHILDREN = 3;

    /**
     * @type array
     */
    protected $errors;

    /**
     * @type array
     */
    protected $_current;

    /**
     * @type array
     */
    protected $_stacks = array(array());

    /**
     * @type HTMLPurifier_Language
     */
    protected $locale;

    /**
     * @type HTMLPurifier_Generator
     */
    protected $generator;

    /**
     * @type HTMLPurifier_Context
     */
    protected $context;

    /**
     * @type array

... [truncated, 195 more lines] ...

<?php

/**
 * Records errors for particular segments of an HTML document such as tokens,
 * attributes or CSS properties. They can contain error structs (which apply
 * to components of what they represent), but their main purpose is to hold
 * errors applying to whatever struct is being used.
 */
class HTMLPurifier_ErrorStruct
{

    /**
     * Possible values for $children first-key. Note that top-level structures
     * are automatically token-level.
     */
    const TOKEN     = 0;
    const ATTR      = 1;
    const CSSPROP   = 2;

    /**
     * Type of this struct.
     * @type string
     */
    public $type;

    /**
     * Value of the struct we are recording errors for. There are various
     * values for this:
     *  - TOKEN: Instance of HTMLPurifier_Token
     *  - ATTR: array('attr-name', 'value')
     *  - CSSPROP: array('prop-name', 'value')
     * @type mixed
     */
    public $value;

    /**
     * Errors registered for this structure.
     * @type array
     */
    public $errors = array();

    /**
     * Child ErrorStructs that are from this structure. For example, a TOKEN
     * ErrorStruct would contain ATTR ErrorStructs. This is a multi-dimensional
     * array in structure: [TYPE]['identifier']
     * @type array
     */
    public $children = array();

    /**

... [truncated, 25 more lines] ...

<?php

/**
 * Global exception class for HTML Purifier; any exceptions we throw
 * are from here.
 */
class HTMLPurifier_Exception extends Exception
{

}

// vim: et sw=4 sts=4

<?php

/**
 * Represents a pre or post processing filter on HTML Purifier's output
 *
 * Sometimes, a little ad-hoc fixing of HTML has to be done before
 * it gets sent through HTML Purifier: you can use filters to acheive
 * this effect. For instance, YouTube videos can be preserved using
 * this manner. You could have used a decorator for this task, but
 * PHP's support for them is not terribly robust, so we're going
 * to just loop through the filters.
 *
 * Filters should be exited first in, last out. If there are three filters,
 * named 1, 2 and 3, the order of execution should go 1->preFilter,
 * 2->preFilter, 3->preFilter, purify, 3->postFilter, 2->postFilter,
 * 1->postFilter.
 *
 * @note Methods are not declared abstract as it is perfectly legitimate
 *       for an implementation not to want anything to happen on a step
 */

class HTMLPurifier_Filter
{

    /**
     * Name of the filter for identification purposes.
     * @type string
     */
    public $name;

    /**
     * Pre-processor function, handles HTML before HTML Purifier
     * @param string $html
     * @param HTMLPurifier_Config $config
     * @param HTMLPurifier_Context $context
     * @return string
     */
    public function preFilter($html, $config, $context)
    {
        return $html;
    }

    /**
     * Post-processor function, handles HTML after HTML Purifier
     * @param string $html
     * @param HTMLPurifier_Config $config
     * @param HTMLPurifier_Context $context
     * @return string
     */
    public function postFilter($html, $config, $context)

... [truncated, 7 more lines] ...

<?php

/**
 * Generates HTML from tokens.
 * @todo Refactor interface so that configuration/context is determined
 *       upon instantiation, no need for messy generateFromTokens() calls
 * @todo Make some of the more internal functions protected, and have
 *       unit tests work around that
 */
class HTMLPurifier_Generator
{

    /**
     * Whether or not generator should produce XML output.
     * @type bool
     */
    private $_xhtml = true;

    /**
     * :HACK: Whether or not generator should comment the insides of <script> tags.
     * @type bool
     */
    private $_scriptFix = false;

    /**
     * Cache of HTMLDefinition during HTML output to determine whether or
     * not attributes should be minimized.
     * @type HTMLPurifier_HTMLDefinition
     */
    private $_def;

    /**
     * Cache of %Output.SortAttr.
     * @type bool
     */
    private $_sortAttr;

    /**
     * Cache of %Output.FlashCompat.
     * @type bool
     */
    private $_flashCompat;

    /**
     * Cache of %Output.FixInnerHTML.
     * @type bool
     */
    private $_innerHTMLFix;

    /**

... [truncated, 237 more lines] ...

<?php

/**
 * Definition of the purified HTML that describes allowed children,
 * attributes, and many other things.
 *
 * Conventions:
 *
 * All member variables that are prefixed with info
 * (including the main $info array) are used by HTML Purifier internals
 * and should not be directly edited when customizing the HTMLDefinition.
 * They can usually be set via configuration directives or custom
 * modules.
 *
 * On the other hand, member variables without the info prefix are used
 * internally by the HTMLDefinition and MUST NOT be used by other HTML
 * Purifier internals. Many of them, however, are public, and may be
 * edited by userspace code to tweak the behavior of HTMLDefinition.
 *
 * @note This class is inspected by Printer_HTMLDefinition; please
 *       update that class if things here change.
 *
 * @warning Directives that change this object's structure must be in
 *          the HTML or Attr namespace!
 */
class HTMLPurifier_HTMLDefinition extends HTMLPurifier_Definition
{

    // FULLY-PUBLIC VARIABLES ---------------------------------------------

    /**
     * Associative array of element names to HTMLPurifier_ElementDef.
     * @type HTMLPurifier_ElementDef[]
     */
    public $info = array();

    /**
     * Associative array of global attribute name to attribute definition.
     * @type array
     */
    public $info_global_attr = array();

    /**
     * String name of parent element HTML will be going into.
     * @type string
     */
    public $info_parent = 'div';

    /**
     * Definition for parent element, allows parent element to be a

... [truncated, 444 more lines] ...

<?php

/**
 * Represents an XHTML 1.1 module, with information on elements, tags
 * and attributes.
 * @note Even though this is technically XHTML 1.1, it is also used for
 *       regular HTML parsing. We are using modulization as a convenient
 *       way to represent the internals of HTMLDefinition, and our
 *       implementation is by no means conforming and does not directly
 *       use the normative DTDs or XML schemas.
 * @note The public variables in a module should almost directly
 *       correspond to the variables in HTMLPurifier_HTMLDefinition.
 *       However, the prefix info carries no special meaning in these
 *       objects (include it anyway if that's the correspondence though).
 * @todo Consider making some member functions protected
 */

class HTMLPurifier_HTMLModule
{

    // -- Overloadable ----------------------------------------------------

    /**
     * Short unique string identifier of the module.
     * @type string
     */
    public $name;

    /**
     * Informally, a list of elements this module changes.
     * Not used in any significant way.
     * @type array
     */
    public $elements = array();

    /**
     * Associative array of element names to element definitions.
     * Some definitions may be incomplete, to be merged in later
     * with the full definition.
     * @type array
     */
    public $info = array();

    /**
     * Associative array of content set names to content set additions.
     * This is commonly used to, say, add an A element to the Inline
     * content set. This corresponds to an internal variable $content_sets
     * and NOT info_content_sets member variable of HTMLDefinition.
     * @type array
     */

... [truncated, 235 more lines] ...

<?php

class HTMLPurifier_HTMLModuleManager
{

    /**
     * @type HTMLPurifier_DoctypeRegistry
     */
    public $doctypes;

    /**
     * Instance of current doctype.
     * @type string
     */
    public $doctype;

    /**
     * @type HTMLPurifier_AttrTypes
     */
    public $attrTypes;

    /**
     * Active instances of modules for the specified doctype are
     * indexed, by name, in this array.
     * @type HTMLPurifier_HTMLModule[]
     */
    public $modules = array();

    /**
     * Array of recognized HTMLPurifier_HTMLModule instances,
     * indexed by module's class name. This array is usually lazy loaded, but a
     * user can overload a module by pre-emptively registering it.
     * @type HTMLPurifier_HTMLModule[]
     */
    public $registeredModules = array();

    /**
     * List of extra modules that were added by the user
     * using addModule(). These get unconditionally merged into the current doctype, whatever
     * it may be.
     * @type HTMLPurifier_HTMLModule[]
     */
    public $userModules = array();

    /**
     * Associative array of element name to list of modules that have
     * definitions for the element; this array is dynamically filled.
     * @type array
     */
    public $elementLookup = array();

... [truncated, 415 more lines] ...

<?php

/**
 * Component of HTMLPurifier_AttrContext that accumulates IDs to prevent dupes
 * @note In Slashdot-speak, dupe means duplicate.
 * @note The default constructor does not accept $config or $context objects:
 *       use must use the static build() factory method to perform initialization.
 */
class HTMLPurifier_IDAccumulator
{

    /**
     * Lookup table of IDs we've accumulated.
     * @public
     */
    public $ids = array();

    /**
     * Builds an IDAccumulator, also initializing the default blacklist
     * @param HTMLPurifier_Config $config Instance of HTMLPurifier_Config
     * @param HTMLPurifier_Context $context Instance of HTMLPurifier_Context
     * @return HTMLPurifier_IDAccumulator Fully initialized HTMLPurifier_IDAccumulator
     */
    public static function build($config, $context)
    {
        $id_accumulator = new HTMLPurifier_IDAccumulator();
        $id_accumulator->load($config->get('Attr.IDBlacklist'));
        return $id_accumulator;
    }

    /**
     * Add an ID to the lookup table.
     * @param string $id ID to be added.
     * @return bool status, true if success, false if there's a dupe
     */
    public function add($id)
    {
        if (isset($this->ids[$id])) {
            return false;
        }
        return $this->ids[$id] = true;
    }

    /**
     * Load a list of IDs into the lookup table
     * @param $array_of_ids Array of IDs to load
     * @note This function doesn't care about duplicates
     */
    public function load($array_of_ids)
    {

... [truncated, 8 more lines] ...

<?php

/**
 * Injects tokens into the document while parsing for well-formedness.
 * This enables "formatter-like" functionality such as auto-paragraphing,
 * smiley-ification and linkification to take place.
 *
 * A note on how handlers create changes; this is done by assigning a new
 * value to the $token reference. These values can take a variety of forms and
 * are best described HTMLPurifier_Strategy_MakeWellFormed->processToken()
 * documentation.
 *
 * @todo Allow injectors to request a re-run on their output. This
 *       would help if an operation is recursive.
 */
abstract class HTMLPurifier_Injector
{

    /**
     * Advisory name of injector, this is for friendly error messages.
     * @type string
     */
    public $name;

    /**
     * @type HTMLPurifier_HTMLDefinition
     */
    protected $htmlDefinition;

    /**
     * Reference to CurrentNesting variable in Context. This is an array
     * list of tokens that we are currently "inside"
     * @type array
     */
    protected $currentNesting;

    /**
     * Reference to current token.
     * @type HTMLPurifier_Token
     */
    protected $currentToken;

    /**
     * Reference to InputZipper variable in Context.
     * @type HTMLPurifier_Zipper
     */
    protected $inputZipper;

    /**
     * Array of elements and attributes this injector creates and therefore

... [truncated, 232 more lines] ...

<?php

/**
 * Represents a language and defines localizable string formatting and
 * other functions, as well as the localized messages for HTML Purifier.
 */
class HTMLPurifier_Language
{

    /**
     * ISO 639 language code of language. Prefers shortest possible version.
     * @type string
     */
    public $code = 'en';

    /**
     * Fallback language code.
     * @type bool|string
     */
    public $fallback = false;

    /**
     * Array of localizable messages.
     * @type array
     */
    public $messages = array();

    /**
     * Array of localizable error codes.
     * @type array
     */
    public $errorNames = array();

    /**
     * True if no message file was found for this language, so English
     * is being used instead. Check this if you'd like to notify the
     * user that they've used a non-supported language.
     * @type bool
     */
    public $error = false;

    /**
     * Has the language object been loaded yet?
     * @type bool
     * @todo Make it private, fix usage in HTMLPurifier_LanguageTest
     */
    public $_loaded = false;

    /**
     * @type HTMLPurifier_Config

... [truncated, 155 more lines] ...

<?php

/**
 * Class responsible for generating HTMLPurifier_Language objects, managing
 * caching and fallbacks.
 * @note Thanks to MediaWiki for the general logic, although this version
 *       has been entirely rewritten
 * @todo Serialized cache for languages
 */
class HTMLPurifier_LanguageFactory
{

    /**
     * Cache of language code information used to load HTMLPurifier_Language objects.
     * Structure is: $factory->cache[$language_code][$key] = $value
     * @type array
     */
    public $cache;

    /**
     * Valid keys in the HTMLPurifier_Language object. Designates which
     * variables to slurp out of a message file.
     * @type array
     */
    public $keys = array('fallback', 'messages', 'errorNames');

    /**
     * Instance to validate language codes.
     * @type HTMLPurifier_AttrDef_Lang
     *
     */
    protected $validator;

    /**
     * Cached copy of dirname(__FILE__), directory of current file without
     * trailing slash.
     * @type string
     */
    protected $dir;

    /**
     * Keys whose contents are a hash map and can be merged.
     * @type array
     */
    protected $mergeable_keys_map = array('messages' => true, 'errorNames' => true);

    /**
     * Keys whose contents are a list and can be merged.
     * @value array lookup
     */

... [truncated, 160 more lines] ...

<?php

/**
 * Represents a measurable length, with a string numeric magnitude
 * and a unit. This object is immutable.
 */
class HTMLPurifier_Length
{

    /**
     * String numeric magnitude.
     * @type string
     */
    protected $n;

    /**
     * String unit. False is permitted if $n = 0.
     * @type string|bool
     */
    protected $unit;

    /**
     * Whether or not this length is valid. Null if not calculated yet.
     * @type bool
     */
    protected $isValid;

    /**
     * Array Lookup array of units recognized by CSS 2.1
     * @type array
     */
    protected static $allowedUnits = array(
        'em' => true, 'ex' => true, 'px' => true, 'in' => true,
        'cm' => true, 'mm' => true, 'pt' => true, 'pc' => true
    );

    /**
     * @param string $n Magnitude
     * @param bool|string $u Unit
     */
    public function __construct($n = '0', $u = false)
    {
        $this->n = (string) $n;
        $this->unit = $u !== false ? (string) $u : false;
    }

    /**
     * @param string $s Unit string, like '2em' or '3.4in'
     * @return HTMLPurifier_Length
     * @warning Does not perform validation.

... [truncated, 111 more lines] ...

<?php

/**
 * Forgivingly lexes HTML (SGML-style) markup into tokens.
 *
 * A lexer parses a string of SGML-style markup and converts them into
 * corresponding tokens.  It doesn't check for well-formedness, although its
 * internal mechanism may make this automatic (such as the case of
 * HTMLPurifier_Lexer_DOMLex).  There are several implementations to choose
 * from.
 *
 * A lexer is HTML-oriented: it might work with XML, but it's not
 * recommended, as we adhere to a subset of the specification for optimization
 * reasons. This might change in the future. Also, most tokenizers are not
 * expected to handle DTDs or PIs.
 *
 * This class should not be directly instantiated, but you may use create() to
 * retrieve a default copy of the lexer.  Being a supertype, this class
 * does not actually define any implementation, but offers commonly used
 * convenience functions for subclasses.
 *
 * @note The unit tests will instantiate this class for testing purposes, as
 *       many of the utility functions require a class to be instantiated.
 *       This means that, even though this class is not runnable, it will
 *       not be declared abstract.
 *
 * @par
 *
 * @note
 * We use tokens rather than create a DOM representation because DOM would:
 *
 * @par
 *  -# Require more processing and memory to create,
 *  -# Is not streamable, and
 *  -# Has the entire document structure (html and body not needed).
 *
 * @par
 * However, DOM is helpful in that it makes it easy to move around nodes
 * without a lot of lookaheads to see when a tag is closed. This is a
 * limitation of the token system and some workarounds would be nice.
 */
class HTMLPurifier_Lexer
{

    /**
     * Whether or not this lexer implements line-number/column-number tracking.
     * If it does, set to true.
     */
    public $tracksLineNumbers = false;


... [truncated, 313 more lines] ...

<?php

/**
 * Abstract base node class that all others inherit from.
 *
 * Why do we not use the DOM extension?  (1) It is not always available,
 * (2) it has funny constraints on the data it can represent,
 * whereas we want a maximally flexible representation, and (3) its
 * interface is a bit cumbersome.
 */
abstract class HTMLPurifier_Node
{
    /**
     * Line number of the start token in the source document
     * @type int
     */
    public $line;

    /**
     * Column number of the start token in the source document. Null if unknown.
     * @type int
     */
    public $col;

    /**
     * Lookup array of processing that this token is exempt from.
     * Currently, valid values are "ValidateAttributes".
     * @type array
     */
    public $armor = array();

    /**
     * When true, this node should be ignored as non-existent.
     *
     * Who is responsible for ignoring dead nodes?  FixNesting is
     * responsible for removing them before passing on to child
     * validators.
     */
    public $dead = false;

    /**
     * Returns a pair of start and end tokens, where the end token
     * is null if it is not necessary. Does not include children.
     * @type array
     */
    abstract public function toTokenPair();
}

// vim: et sw=4 sts=4

<?php

/**
 * Class that handles operations involving percent-encoding in URIs.
 *
 * @warning
 *      Be careful when reusing instances of PercentEncoder. The object
 *      you use for normalize() SHOULD NOT be used for encode(), or
 *      vice-versa.
 */
class HTMLPurifier_PercentEncoder
{

    /**
     * Reserved characters to preserve when using encode().
     * @type array
     */
    protected $preserve = array();

    /**
     * String of characters that should be preserved while using encode().
     * @param bool $preserve
     */
    public function __construct($preserve = false)
    {
        // unreserved letters, ought to const-ify
        for ($i = 48; $i <= 57; $i++) { // digits
            $this->preserve[$i] = true;
        }
        for ($i = 65; $i <= 90; $i++) { // upper-case
            $this->preserve[$i] = true;
        }
        for ($i = 97; $i <= 122; $i++) { // lower-case
            $this->preserve[$i] = true;
        }
        $this->preserve[45] = true; // Dash         -
        $this->preserve[46] = true; // Period       .
        $this->preserve[95] = true; // Underscore   _
        $this->preserve[126]= true; // Tilde        ~

        // extra letters not to escape
        if ($preserve !== false) {
            for ($i = 0, $c = strlen($preserve); $i < $c; $i++) {
                $this->preserve[ord($preserve[$i])] = true;
            }
        }
    }

    /**
     * Our replacement for urlencode, it encodes all non-reserved characters,

... [truncated, 62 more lines] ...