This page lists files in the current directory. You can view content, get download/execute commands for Wget, Curl, or PowerShell, or filter the list using wildcards (e.g., `*.sh`).
wget 'https://sme10.lists2.roe3.org/guppy/inc/htmlpurifier/library/HTMLPurifier/Strategy/Composite.php'<?php
/**
* Composite strategy that runs multiple strategies on tokens.
*/
abstract class HTMLPurifier_Strategy_Composite extends HTMLPurifier_Strategy
{
/**
* List of strategies to run tokens through.
* @type HTMLPurifier_Strategy[]
*/
protected $strategies = array();
/**
* @param HTMLPurifier_Token[] $tokens
* @param HTMLPurifier_Config $config
* @param HTMLPurifier_Context $context
* @return HTMLPurifier_Token[]
*/
public function execute($tokens, $config, $context)
{
foreach ($this->strategies as $strategy) {
$tokens = $strategy->execute($tokens, $config, $context);
}
return $tokens;
}
}
// vim: et sw=4 sts=4
wget 'https://sme10.lists2.roe3.org/guppy/inc/htmlpurifier/library/HTMLPurifier/Strategy/Core.php'<?php
/**
* Core strategy composed of the big four strategies.
*/
class HTMLPurifier_Strategy_Core extends HTMLPurifier_Strategy_Composite
{
public function __construct()
{
$this->strategies[] = new HTMLPurifier_Strategy_RemoveForeignElements();
$this->strategies[] = new HTMLPurifier_Strategy_MakeWellFormed();
$this->strategies[] = new HTMLPurifier_Strategy_FixNesting();
$this->strategies[] = new HTMLPurifier_Strategy_ValidateAttributes();
}
}
// vim: et sw=4 sts=4
wget 'https://sme10.lists2.roe3.org/guppy/inc/htmlpurifier/library/HTMLPurifier/Strategy/FixNesting.php'<?php
/**
* Takes a well formed list of tokens and fixes their nesting.
*
* HTML elements dictate which elements are allowed to be their children,
* for example, you can't have a p tag in a span tag. Other elements have
* much more rigorous definitions: tables, for instance, require a specific
* order for their elements. There are also constraints not expressible by
* document type definitions, such as the chameleon nature of ins/del
* tags and global child exclusions.
*
* The first major objective of this strategy is to iterate through all
* the nodes and determine whether or not their children conform to the
* element's definition. If they do not, the child definition may
* optionally supply an amended list of elements that is valid or
* require that the entire node be deleted (and the previous node
* rescanned).
*
* The second objective is to ensure that explicitly excluded elements of
* an element do not appear in its children. Code that accomplishes this
* task is pervasive through the strategy, though the two are distinct tasks
* and could, theoretically, be seperated (although it's not recommended).
*
* @note Whether or not unrecognized children are silently dropped or
* translated into text depends on the child definitions.
*
* @todo Enable nodes to be bubbled out of the structure. This is
* easier with our new algorithm.
*/
class HTMLPurifier_Strategy_FixNesting extends HTMLPurifier_Strategy
{
/**
* @param HTMLPurifier_Token[] $tokens
* @param HTMLPurifier_Config $config
* @param HTMLPurifier_Context $context
* @return array|HTMLPurifier_Token[]
*/
public function execute($tokens, $config, $context)
{
//####################################################################//
// Pre-processing
// O(n) pass to convert to a tree, so that we can efficiently
// refer to substrings
$top_node = HTMLPurifier_Arborize::arborize($tokens, $config, $context);
// get a copy of the HTML definition
$definition = $config->getHTMLDefinition();
$excludes_enabled = !$config->get('Core.DisableExcludes');
// setup the context variable 'IsInline', for chameleon processing
// is 'false' when we are not inline, 'true' when it must always
// be inline, and an integer when it is inline for a certain
// branch of the document tree
$is_inline = $definition->info_parent_def->descendants_are_inline;
$context->register('IsInline', $is_inline);
// setup error collector
$e =& $context->get('ErrorCollector', true);
//####################################################################//
// Loop initialization
// stack that contains all elements that are excluded
// it is organized by parent elements, similar to $stack,
// but it is only populated when an element with exclusions is
// processed, i.e. there won't be empty exclusions.
$exclude_stack = array($definition->info_parent_def->excludes);
// variable that contains the start token while we are processing
// nodes. This enables error reporting to do its job
$node = $top_node;
// dummy token
list($token, $d) = $node->toTokenPair();
$context->register('CurrentNode', $node);
$context->register('CurrentToken', $token);
//####################################################################//
// Loop
// We need to implement a post-order traversal iteratively, to
// avoid running into stack space limits. This is pretty tricky
// to reason about, so we just manually stack-ify the recursive
// variant:
//
// function f($node) {
// foreach ($node->children as $child) {
// f($child);
// }
// validate($node);
// }
//
// Thus, we will represent a stack frame as array($node,
// $is_inline, stack of children)
// e.g. array_reverse($node->children) - already processed
// children.
$parent_def = $definition->info_parent_def;
$stack = array(
array($top_node,
$parent_def->descendants_are_inline,
$parent_def->excludes, // exclusions
0)
);
while (!empty($stack)) {
list($node, $is_inline, $excludes, $ix) = array_pop($stack);
// recursive call
$go = false;
$def = empty($stack) ? $definition->info_parent_def : $definition->info[$node->name];
while (isset($node->children[$ix])) {
$child = $node->children[$ix++];
if ($child instanceof HTMLPurifier_Node_Element) {
$go = true;
$stack[] = array($node, $is_inline, $excludes, $ix);
$stack[] = array($child,
// ToDo: I don't think it matters if it's def or
// child_def, but double check this...
$is_inline || $def->descendants_are_inline,
empty($def->excludes) ? $excludes
: array_merge($excludes, $def->excludes),
0);
break;
}
};
if ($go) continue;
list($token, $d) = $node->toTokenPair();
// base case
if ($excludes_enabled && isset($excludes[$node->name])) {
$node->dead = true;
if ($e) $e->send(E_ERROR, 'Strategy_FixNesting: Node excluded');
} else {
// XXX I suppose it would be slightly more efficient to
// avoid the allocation here and have children
// strategies handle it
$children = array();
foreach ($node->children as $child) {
if (!$child->dead) $children[] = $child;
}
$result = $def->child->validateChildren($children, $config, $context);
if ($result === true) {
// nop
$node->children = $children;
} elseif ($result === false) {
$node->dead = true;
if ($e) $e->send(E_ERROR, 'Strategy_FixNesting: Node removed');
} else {
$node->children = $result;
if ($e) {
// XXX This will miss mutations of internal nodes. Perhaps defer to the child validators
if (empty($result) && !empty($children)) {
$e->send(E_ERROR, 'Strategy_FixNesting: Node contents removed');
} else if ($result != $children) {
$e->send(E_WARNING, 'Strategy_FixNesting: Node reorganized');
}
}
}
}
}
//####################################################################//
// Post-processing
// remove context variables
$context->destroy('IsInline');
$context->destroy('CurrentNode');
$context->destroy('CurrentToken');
//####################################################################//
// Return
return HTMLPurifier_Arborize::flatten($node, $config, $context);
}
}
// vim: et sw=4 sts=4
wget 'https://sme10.lists2.roe3.org/guppy/inc/htmlpurifier/library/HTMLPurifier/Strategy/MakeWellFormed.php'<?php
/**
* Takes tokens makes them well-formed (balance end tags, etc.)
*
* Specification of the armor attributes this strategy uses:
*
* - MakeWellFormed_TagClosedError: This armor field is used to
* suppress tag closed errors for certain tokens [TagClosedSuppress],
* in particular, if a tag was generated automatically by HTML
* Purifier, we may rely on our infrastructure to close it for us
* and shouldn't report an error to the user [TagClosedAuto].
*/
class HTMLPurifier_Strategy_MakeWellFormed extends HTMLPurifier_Strategy
{
/**
* Array stream of tokens being processed.
* @type HTMLPurifier_Token[]
*/
protected $tokens;
/**
* Current token.
* @type HTMLPurifier_Token
*/
protected $token;
/**
* Zipper managing the true state.
* @type HTMLPurifier_Zipper
*/
protected $zipper;
/**
* Current nesting of elements.
* @type array
*/
protected $stack;
/**
* Injectors active in this stream processing.
* @type HTMLPurifier_Injector[]
*/
protected $injectors;
/**
* Current instance of HTMLPurifier_Config.
* @type HTMLPurifier_Config
*/
protected $config;
/**
* Current instance of HTMLPurifier_Context.
* @type HTMLPurifier_Context
*/
protected $context;
/**
* @param HTMLPurifier_Token[] $tokens
* @param HTMLPurifier_Config $config
* @param HTMLPurifier_Context $context
* @return HTMLPurifier_Token[]
* @throws HTMLPurifier_Exception
*/
public function execute($tokens, $config, $context)
{
$definition = $config->getHTMLDefinition();
// local variables
$generator = new HTMLPurifier_Generator($config, $context);
$escape_invalid_tags = $config->get('Core.EscapeInvalidTags');
// used for autoclose early abortion
$global_parent_allowed_elements = $definition->info_parent_def->child->getAllowedElements($config);
$e = $context->get('ErrorCollector', true);
$i = false; // injector index
list($zipper, $token) = HTMLPurifier_Zipper::fromArray($tokens);
if ($token === NULL) {
return array();
}
$reprocess = false; // whether or not to reprocess the same token
$stack = array();
// member variables
$this->stack =& $stack;
$this->tokens =& $tokens;
$this->token =& $token;
$this->zipper =& $zipper;
$this->config = $config;
$this->context = $context;
// context variables
$context->register('CurrentNesting', $stack);
$context->register('InputZipper', $zipper);
$context->register('CurrentToken', $token);
// -- begin INJECTOR --
$this->injectors = array();
$injectors = $config->getBatch('AutoFormat');
$def_injectors = $definition->info_injector;
$custom_injectors = $injectors['Custom'];
unset($injectors['Custom']); // special case
foreach ($injectors as $injector => $b) {
// XXX: Fix with a legitimate lookup table of enabled filters
if (strpos($injector, '.') !== false) {
continue;
}
$injector = "HTMLPurifier_Injector_$injector";
if (!$b) {
continue;
}
$this->injectors[] = new $injector;
}
foreach ($def_injectors as $injector) {
// assumed to be objects
$this->injectors[] = $injector;
}
foreach ($custom_injectors as $injector) {
if (!$injector) {
continue;
}
if (is_string($injector)) {
$injector = "HTMLPurifier_Injector_$injector";
$injector = new $injector;
}
$this->injectors[] = $injector;
}
// give the injectors references to the definition and context
// variables for performance reasons
foreach ($this->injectors as $ix => $injector) {
$error = $injector->prepare($config, $context);
if (!$error) {
continue;
}
array_splice($this->injectors, $ix, 1); // rm the injector
trigger_error("Cannot enable {$injector->name} injector because $error is not allowed", E_USER_WARNING);
}
// -- end INJECTOR --
// a note on reprocessing:
// In order to reduce code duplication, whenever some code needs
// to make HTML changes in order to make things "correct", the
// new HTML gets sent through the purifier, regardless of its
// status. This means that if we add a start token, because it
// was totally necessary, we don't have to update nesting; we just
// punt ($reprocess = true; continue;) and it does that for us.
// isset is in loop because $tokens size changes during loop exec
for (;;
// only increment if we don't need to reprocess
$reprocess ? $reprocess = false : $token = $zipper->next($token)) {
// check for a rewind
if (is_int($i)) {
// possibility: disable rewinding if the current token has a
// rewind set on it already. This would offer protection from
// infinite loop, but might hinder some advanced rewinding.
$rewind_offset = $this->injectors[$i]->getRewindOffset();
if (is_int($rewind_offset)) {
for ($j = 0; $j < $rewind_offset; $j++) {
if (empty($zipper->front)) break;
$token = $zipper->prev($token);
// indicate that other injectors should not process this token,
// but we need to reprocess it. See Note [Injector skips]
unset($token->skip[$i]);
$token->rewind = $i;
if ($token instanceof HTMLPurifier_Token_Start) {
array_pop($this->stack);
} elseif ($token instanceof HTMLPurifier_Token_End) {
$this->stack[] = $token->start;
}
}
}
$i = false;
}
// handle case of document end
if ($token === NULL) {
// kill processing if stack is empty
if (empty($this->stack)) {
break;
}
// peek
$top_nesting = array_pop($this->stack);
$this->stack[] = $top_nesting;
// send error [TagClosedSuppress]
if ($e && !isset($top_nesting->armor['MakeWellFormed_TagClosedError'])) {
$e->send(E_NOTICE, 'Strategy_MakeWellFormed: Tag closed by document end', $top_nesting);
}
// append, don't splice, since this is the end
$token = new HTMLPurifier_Token_End($top_nesting->name);
// punt!
$reprocess = true;
continue;
}
//echo '<br>'; printZipper($zipper, $token);//printTokens($this->stack);
//flush();
// quick-check: if it's not a tag, no need to process
if (empty($token->is_tag)) {
if ($token instanceof HTMLPurifier_Token_Text) {
foreach ($this->injectors as $i => $injector) {
if (isset($token->skip[$i])) {
// See Note [Injector skips]
continue;
}
if ($token->rewind !== null && $token->rewind !== $i) {
continue;
}
// XXX fuckup
$r = $token;
$injector->handleText($r);
$token = $this->processToken($r, $i);
$reprocess = true;
break;
}
}
// another possibility is a comment
continue;
}
if (isset($definition->info[$token->name])) {
$type = $definition->info[$token->name]->child->type;
} else {
$type = false; // Type is unknown, treat accordingly
}
// quick tag checks: anything that's *not* an end tag
$ok = false;
if ($type === 'empty' && $token instanceof HTMLPurifier_Token_Start) {
// claims to be a start tag but is empty
$token = new HTMLPurifier_Token_Empty(
$token->name,
$token->attr,
$token->line,
$token->col,
$token->armor
);
$ok = true;
} elseif ($type && $type !== 'empty' && $token instanceof HTMLPurifier_Token_Empty) {
// claims to be empty but really is a start tag
// NB: this assignment is required
$old_token = $token;
$token = new HTMLPurifier_Token_End($token->name);
$token = $this->insertBefore(
new HTMLPurifier_Token_Start($old_token->name, $old_token->attr, $old_token->line, $old_token->col, $old_token->armor)
);
// punt (since we had to modify the input stream in a non-trivial way)
$reprocess = true;
continue;
} elseif ($token instanceof HTMLPurifier_Token_Empty) {
// real empty token
$ok = true;
} elseif ($token instanceof HTMLPurifier_Token_Start) {
// start tag
// ...unless they also have to close their parent
if (!empty($this->stack)) {
// Performance note: you might think that it's rather
// inefficient, recalculating the autoclose information
// for every tag that a token closes (since when we
// do an autoclose, we push a new token into the
// stream and then /process/ that, before
// re-processing this token.) But this is
// necessary, because an injector can make an
// arbitrary transformations to the autoclosing
// tokens we introduce, so things may have changed
// in the meantime. Also, doing the inefficient thing is
// "easy" to reason about (for certain perverse definitions
// of "easy")
$parent = array_pop($this->stack);
$this->stack[] = $parent;
$parent_def = null;
$parent_elements = null;
$autoclose = false;
if (isset($definition->info[$parent->name])) {
$parent_def = $definition->info[$parent->name];
$parent_elements = $parent_def->child->getAllowedElements($config);
$autoclose = !isset($parent_elements[$token->name]);
}
if ($autoclose && $definition->info[$token->name]->wrap) {
// Check if an element can be wrapped by another
// element to make it valid in a context (for
// example, <ul><ul> needs a <li> in between)
$wrapname = $definition->info[$token->name]->wrap;
$wrapdef = $definition->info[$wrapname];
$elements = $wrapdef->child->getAllowedElements($config);
if (isset($elements[$token->name]) && isset($parent_elements[$wrapname])) {
$newtoken = new HTMLPurifier_Token_Start($wrapname);
$token = $this->insertBefore($newtoken);
$reprocess = true;
continue;
}
}
$carryover = false;
if ($autoclose && $parent_def->formatting) {
$carryover = true;
}
if ($autoclose) {
// check if this autoclose is doomed to fail
// (this rechecks $parent, which his harmless)
$autoclose_ok = isset($global_parent_allowed_elements[$token->name]);
if (!$autoclose_ok) {
foreach ($this->stack as $ancestor) {
$elements = $definition->info[$ancestor->name]->child->getAllowedElements($config);
if (isset($elements[$token->name])) {
$autoclose_ok = true;
break;
}
if ($definition->info[$token->name]->wrap) {
$wrapname = $definition->info[$token->name]->wrap;
$wrapdef = $definition->info[$wrapname];
$wrap_elements = $wrapdef->child->getAllowedElements($config);
if (isset($wrap_elements[$token->name]) && isset($elements[$wrapname])) {
$autoclose_ok = true;
break;
}
}
}
}
if ($autoclose_ok) {
// errors need to be updated
$new_token = new HTMLPurifier_Token_End($parent->name);
$new_token->start = $parent;
// [TagClosedSuppress]
if ($e && !isset($parent->armor['MakeWellFormed_TagClosedError'])) {
if (!$carryover) {
$e->send(E_NOTICE, 'Strategy_MakeWellFormed: Tag auto closed', $parent);
} else {
$e->send(E_NOTICE, 'Strategy_MakeWellFormed: Tag carryover', $parent);
}
}
if ($carryover) {
$element = clone $parent;
// [TagClosedAuto]
$element->armor['MakeWellFormed_TagClosedError'] = true;
$element->carryover = true;
$token = $this->processToken(array($new_token, $token, $element));
} else {
$token = $this->insertBefore($new_token);
}
} else {
$token = $this->remove();
}
$reprocess = true;
continue;
}
}
$ok = true;
}
if ($ok) {
foreach ($this->injectors as $i => $injector) {
if (isset($token->skip[$i])) {
// See Note [Injector skips]
continue;
}
if ($token->rewind !== null && $token->rewind !== $i) {
continue;
}
$r = $token;
$injector->handleElement($r);
$token = $this->processToken($r, $i);
$reprocess = true;
break;
}
if (!$reprocess) {
// ah, nothing interesting happened; do normal processing
if ($token instanceof HTMLPurifier_Token_Start) {
$this->stack[] = $token;
} elseif ($token instanceof HTMLPurifier_Token_End) {
throw new HTMLPurifier_Exception(
'Improper handling of end tag in start code; possible error in MakeWellFormed'
);
}
}
continue;
}
// sanity check: we should be dealing with a closing tag
if (!$token instanceof HTMLPurifier_Token_End) {
throw new HTMLPurifier_Exception('Unaccounted for tag token in input stream, bug in HTML Purifier');
}
// make sure that we have something open
if (empty($this->stack)) {
if ($escape_invalid_tags) {
if ($e) {
$e->send(E_WARNING, 'Strategy_MakeWellFormed: Unnecessary end tag to text');
}
$token = new HTMLPurifier_Token_Text($generator->generateFromToken($token));
} else {
if ($e) {
$e->send(E_WARNING, 'Strategy_MakeWellFormed: Unnecessary end tag removed');
}
$token = $this->remove();
}
$reprocess = true;
continue;
}
// first, check for the simplest case: everything closes neatly.
// Eventually, everything passes through here; if there are problems
// we modify the input stream accordingly and then punt, so that
// the tokens get processed again.
$current_parent = array_pop($this->stack);
if ($current_parent->name == $token->name) {
$token->start = $current_parent;
foreach ($this->injectors as $i => $injector) {
if (isset($token->skip[$i])) {
// See Note [Injector skips]
continue;
}
if ($token->rewind !== null && $token->rewind !== $i) {
continue;
}
$r = $token;
$injector->handleEnd($r);
$token = $this->processToken($r, $i);
$this->stack[] = $current_parent;
$reprocess = true;
break;
}
continue;
}
// okay, so we're trying to close the wrong tag
// undo the pop previous pop
$this->stack[] = $current_parent;
// scroll back the entire nest, trying to find our tag.
// (feature could be to specify how far you'd like to go)
$size = count($this->stack);
// -2 because -1 is the last element, but we already checked that
$skipped_tags = false;
for ($j = $size - 2; $j >= 0; $j--) {
if ($this->stack[$j]->name == $token->name) {
$skipped_tags = array_slice($this->stack, $j);
break;
}
}
// we didn't find the tag, so remove
if ($skipped_tags === false) {
if ($escape_invalid_tags) {
if ($e) {
$e->send(E_WARNING, 'Strategy_MakeWellFormed: Stray end tag to text');
}
$token = new HTMLPurifier_Token_Text($generator->generateFromToken($token));
} else {
if ($e) {
$e->send(E_WARNING, 'Strategy_MakeWellFormed: Stray end tag removed');
}
$token = $this->remove();
}
$reprocess = true;
continue;
}
// do errors, in REVERSE $j order: a,b,c with </a></b></c>
$c = count($skipped_tags);
if ($e) {
for ($j = $c - 1; $j > 0; $j--) {
// notice we exclude $j == 0, i.e. the current ending tag, from
// the errors... [TagClosedSuppress]
if (!isset($skipped_tags[$j]->armor['MakeWellFormed_TagClosedError'])) {
$e->send(E_NOTICE, 'Strategy_MakeWellFormed: Tag closed by element end', $skipped_tags[$j]);
}
}
}
// insert tags, in FORWARD $j order: c,b,a with </a></b></c>
$replace = array($token);
for ($j = 1; $j < $c; $j++) {
// ...as well as from the insertions
$new_token = new HTMLPurifier_Token_End($skipped_tags[$j]->name);
$new_token->start = $skipped_tags[$j];
array_unshift($replace, $new_token);
if (isset($definition->info[$new_token->name]) && $definition->info[$new_token->name]->formatting) {
// [TagClosedAuto]
$element = clone $skipped_tags[$j];
$element->carryover = true;
$element->armor['MakeWellFormed_TagClosedError'] = true;
$replace[] = $element;
}
}
$token = $this->processToken($replace);
$reprocess = true;
continue;
}
$context->destroy('CurrentToken');
$context->destroy('CurrentNesting');
$context->destroy('InputZipper');
unset($this->injectors, $this->stack, $this->tokens);
return $zipper->toArray($token);
}
/**
* Processes arbitrary token values for complicated substitution patterns.
* In general:
*
* If $token is an array, it is a list of tokens to substitute for the
* current token. These tokens then get individually processed. If there
* is a leading integer in the list, that integer determines how many
* tokens from the stream should be removed.
*
* If $token is a regular token, it is swapped with the current token.
*
* If $token is false, the current token is deleted.
*
* If $token is an integer, that number of tokens (with the first token
* being the current one) will be deleted.
*
* @param HTMLPurifier_Token|array|int|bool $token Token substitution value
* @param HTMLPurifier_Injector|int $injector Injector that performed the substitution; default is if
* this is not an injector related operation.
* @throws HTMLPurifier_Exception
*/
protected function processToken($token, $injector = -1)
{
// Zend OpCache miscompiles $token = array($token), so
// avoid this pattern. See: https://github.com/ezyang/htmlpurifier/issues/108
// normalize forms of token
if (is_object($token)) {
$tmp = $token;
$token = array(1, $tmp);
}
if (is_int($token)) {
$tmp = $token;
$token = array($tmp);
}
if ($token === false) {
$token = array(1);
}
if (!is_array($token)) {
throw new HTMLPurifier_Exception('Invalid token type from injector');
}
if (!is_int($token[0])) {
array_unshift($token, 1);
}
if ($token[0] === 0) {
throw new HTMLPurifier_Exception('Deleting zero tokens is not valid');
}
// $token is now an array with the following form:
// array(number nodes to delete, new node 1, new node 2, ...)
$delete = array_shift($token);
list($old, $r) = $this->zipper->splice($this->token, $delete, $token);
if ($injector > -1) {
// See Note [Injector skips]
// Determine appropriate skips. Here's what the code does:
// *If* we deleted one or more tokens, copy the skips
// of those tokens into the skips of the new tokens (in $token).
// Also, mark the newly inserted tokens as having come from
// $injector.
$oldskip = isset($old[0]) ? $old[0]->skip : array();
foreach ($token as $object) {
$object->skip = $oldskip;
$object->skip[$injector] = true;
}
}
return $r;
}
/**
* Inserts a token before the current token. Cursor now points to
* this token. You must reprocess after this.
* @param HTMLPurifier_Token $token
*/
private function insertBefore($token)
{
// NB not $this->zipper->insertBefore(), due to positioning
// differences
$splice = $this->zipper->splice($this->token, 0, array($token));
return $splice[1];
}
/**
* Removes current token. Cursor now points to new token occupying previously
* occupied space. You must reprocess after this.
*/
private function remove()
{
return $this->zipper->delete();
}
}
// Note [Injector skips]
// ~~~~~~~~~~~~~~~~~~~~~
// When I originally designed this class, the idea behind the 'skip'
// property of HTMLPurifier_Token was to help avoid infinite loops
// in injector processing. For example, suppose you wrote an injector
// that bolded swear words. Naively, you might write it so that
// whenever you saw ****, you replaced it with <strong>****</strong>.
//
// When this happens, we will reprocess all of the tokens with the
// other injectors. Now there is an opportunity for infinite loop:
// if we rerun the swear-word injector on these tokens, we might
// see **** and then reprocess again to get
// <strong><strong>****</strong></strong> ad infinitum.
//
// Thus, the idea of a skip is that once we process a token with
// an injector, we mark all of those tokens as having "come from"
// the injector, and we never run the injector again on these
// tokens.
//
// There were two more complications, however:
//
// - With HTMLPurifier_Injector_RemoveEmpty, we noticed that if
// you had <b><i></i></b>, after you removed the <i></i>, you
// really would like this injector to go back and reprocess
// the <b> tag, discovering that it is now empty and can be
// removed. So we reintroduced the possibility of infinite looping
// by adding a "rewind" function, which let you go back to an
// earlier point in the token stream and reprocess it with injectors.
// Needless to say, we need to UN-skip the token so it gets
// reprocessed.
//
// - Suppose that you successfuly process a token, replace it with
// one with your skip mark, but now another injector wants to
// process the skipped token with another token. Should you continue
// to skip that new token, or reprocess it? If you reprocess,
// you can end up with an infinite loop where one injector converts
// <a> to <b>, and then another injector converts it back. So
// we inherit the skips, but for some reason, I thought that we
// should inherit the skip from the first token of the token
// that we deleted. Why? Well, it seems to work OK.
//
// If I were to redesign this functionality, I would absolutely not
// go about doing it this way: the semantics are just not very well
// defined, and in any case you probably wanted to operate on trees,
// not token streams.
// vim: et sw=4 sts=4
wget 'https://sme10.lists2.roe3.org/guppy/inc/htmlpurifier/library/HTMLPurifier/Strategy/RemoveForeignElements.php'<?php
/**
* Removes all unrecognized tags from the list of tokens.
*
* This strategy iterates through all the tokens and removes unrecognized
* tokens. If a token is not recognized but a TagTransform is defined for
* that element, the element will be transformed accordingly.
*/
class HTMLPurifier_Strategy_RemoveForeignElements extends HTMLPurifier_Strategy
{
/**
* @param HTMLPurifier_Token[] $tokens
* @param HTMLPurifier_Config $config
* @param HTMLPurifier_Context $context
* @return array|HTMLPurifier_Token[]
*/
public function execute($tokens, $config, $context)
{
$definition = $config->getHTMLDefinition();
$generator = new HTMLPurifier_Generator($config, $context);
$result = array();
$escape_invalid_tags = $config->get('Core.EscapeInvalidTags');
$remove_invalid_img = $config->get('Core.RemoveInvalidImg');
// currently only used to determine if comments should be kept
$trusted = $config->get('HTML.Trusted');
$comment_lookup = $config->get('HTML.AllowedComments');
$comment_regexp = $config->get('HTML.AllowedCommentsRegexp');
$check_comments = $comment_lookup !== array() || $comment_regexp !== null;
$remove_script_contents = $config->get('Core.RemoveScriptContents');
$hidden_elements = $config->get('Core.HiddenElements');
// remove script contents compatibility
if ($remove_script_contents === true) {
$hidden_elements['script'] = true;
} elseif ($remove_script_contents === false && isset($hidden_elements['script'])) {
unset($hidden_elements['script']);
}
$attr_validator = new HTMLPurifier_AttrValidator();
// removes tokens until it reaches a closing tag with its value
$remove_until = false;
// converts comments into text tokens when this is equal to a tag name
$textify_comments = false;
$token = false;
$context->register('CurrentToken', $token);
$e = false;
if ($config->get('Core.CollectErrors')) {
$e =& $context->get('ErrorCollector');
}
foreach ($tokens as $token) {
if ($remove_until) {
if (empty($token->is_tag) || $token->name !== $remove_until) {
continue;
}
}
if (!empty($token->is_tag)) {
// DEFINITION CALL
// before any processing, try to transform the element
if (isset($definition->info_tag_transform[$token->name])) {
$original_name = $token->name;
// there is a transformation for this tag
// DEFINITION CALL
$token = $definition->
info_tag_transform[$token->name]->transform($token, $config, $context);
if ($e) {
$e->send(E_NOTICE, 'Strategy_RemoveForeignElements: Tag transform', $original_name);
}
}
if (isset($definition->info[$token->name])) {
// mostly everything's good, but
// we need to make sure required attributes are in order
if (($token instanceof HTMLPurifier_Token_Start || $token instanceof HTMLPurifier_Token_Empty) &&
$definition->info[$token->name]->required_attr &&
($token->name != 'img' || $remove_invalid_img) // ensure config option still works
) {
$attr_validator->validateToken($token, $config, $context);
$ok = true;
foreach ($definition->info[$token->name]->required_attr as $name) {
if (!isset($token->attr[$name])) {
$ok = false;
break;
}
}
if (!$ok) {
if ($e) {
$e->send(
E_ERROR,
'Strategy_RemoveForeignElements: Missing required attribute',
$name
);
}
continue;
}
$token->armor['ValidateAttributes'] = true;
}
if (isset($hidden_elements[$token->name]) && $token instanceof HTMLPurifier_Token_Start) {
$textify_comments = $token->name;
} elseif ($token->name === $textify_comments && $token instanceof HTMLPurifier_Token_End) {
$textify_comments = false;
}
} elseif ($escape_invalid_tags) {
// invalid tag, generate HTML representation and insert in
if ($e) {
$e->send(E_WARNING, 'Strategy_RemoveForeignElements: Foreign element to text');
}
$token = new HTMLPurifier_Token_Text(
$generator->generateFromToken($token)
);
} else {
// check if we need to destroy all of the tag's children
// CAN BE GENERICIZED
if (isset($hidden_elements[$token->name])) {
if ($token instanceof HTMLPurifier_Token_Start) {
$remove_until = $token->name;
} elseif ($token instanceof HTMLPurifier_Token_Empty) {
// do nothing: we're still looking
} else {
$remove_until = false;
}
if ($e) {
$e->send(E_ERROR, 'Strategy_RemoveForeignElements: Foreign meta element removed');
}
} else {
if ($e) {
$e->send(E_ERROR, 'Strategy_RemoveForeignElements: Foreign element removed');
}
}
continue;
}
} elseif ($token instanceof HTMLPurifier_Token_Comment) {
// textify comments in script tags when they are allowed
if ($textify_comments !== false) {
$data = $token->data;
$token = new HTMLPurifier_Token_Text($data);
} elseif ($trusted || $check_comments) {
// always cleanup comments
$trailing_hyphen = false;
if ($e) {
// perform check whether or not there's a trailing hyphen
if (substr($token->data, -1) == '-') {
$trailing_hyphen = true;
}
}
$token->data = rtrim($token->data, '-');
$found_double_hyphen = false;
while (strpos($token->data, '--') !== false) {
$found_double_hyphen = true;
$token->data = str_replace('--', '-', $token->data);
}
if ($trusted || !empty($comment_lookup[trim($token->data)]) ||
($comment_regexp !== null && preg_match($comment_regexp, trim($token->data)))) {
// OK good
if ($e) {
if ($trailing_hyphen) {
$e->send(
E_NOTICE,
'Strategy_RemoveForeignElements: Trailing hyphen in comment removed'
);
}
if ($found_double_hyphen) {
$e->send(E_NOTICE, 'Strategy_RemoveForeignElements: Hyphens in comment collapsed');
}
}
} else {
if ($e) {
$e->send(E_NOTICE, 'Strategy_RemoveForeignElements: Comment removed');
}
continue;
}
} else {
// strip comments
if ($e) {
$e->send(E_NOTICE, 'Strategy_RemoveForeignElements: Comment removed');
}
continue;
}
} elseif ($token instanceof HTMLPurifier_Token_Text) {
} else {
continue;
}
$result[] = $token;
}
if ($remove_until && $e) {
// we removed tokens until the end, throw error
$e->send(E_ERROR, 'Strategy_RemoveForeignElements: Token removed to end', $remove_until);
}
$context->destroy('CurrentToken');
return $result;
}
}
// vim: et sw=4 sts=4
wget 'https://sme10.lists2.roe3.org/guppy/inc/htmlpurifier/library/HTMLPurifier/Strategy/ValidateAttributes.php'<?php
/**
* Validate all attributes in the tokens.
*/
class HTMLPurifier_Strategy_ValidateAttributes extends HTMLPurifier_Strategy
{
/**
* @param HTMLPurifier_Token[] $tokens
* @param HTMLPurifier_Config $config
* @param HTMLPurifier_Context $context
* @return HTMLPurifier_Token[]
*/
public function execute($tokens, $config, $context)
{
// setup validator
$validator = new HTMLPurifier_AttrValidator();
$token = false;
$context->register('CurrentToken', $token);
foreach ($tokens as $key => $token) {
// only process tokens that have attributes,
// namely start and empty tags
if (!$token instanceof HTMLPurifier_Token_Start && !$token instanceof HTMLPurifier_Token_Empty) {
continue;
}
// skip tokens that are armored
if (!empty($token->armor['ValidateAttributes'])) {
continue;
}
// note that we have no facilities here for removing tokens
$validator->validateToken($token, $config, $context);
}
$context->destroy('CurrentToken');
return $tokens;
}
}
// vim: et sw=4 sts=4