Initial commit.
This commit is contained in:
@@ -0,0 +1,195 @@
|
||||
<?php
|
||||
|
||||
require_once 'HTMLPurifier/Lexer.php';
|
||||
require_once 'HTMLPurifier/TokenFactory.php';
|
||||
|
||||
/**
|
||||
* Parser that uses PHP 5's DOM extension (part of the core).
|
||||
*
|
||||
* In PHP 5, the DOM XML extension was revamped into DOM and added to the core.
|
||||
* It gives us a forgiving HTML parser, which we use to transform the HTML
|
||||
* into a DOM, and then into the tokens. It is blazingly fast (for large
|
||||
* documents, it performs twenty times faster than
|
||||
* HTMLPurifier_Lexer_DirectLex,and is the default choice for PHP 5.
|
||||
*
|
||||
* @note Any empty elements will have empty tokens associated with them, even if
|
||||
* this is prohibited by the spec. This is cannot be fixed until the spec
|
||||
* comes into play.
|
||||
*
|
||||
* @note PHP's DOM extension does not actually parse any entities, we use
|
||||
* our own function to do that.
|
||||
*
|
||||
* @warning DOM tends to drop whitespace, which may wreak havoc on indenting.
|
||||
* If this is a huge problem, due to the fact that HTML is hand
|
||||
* edited and you are unable to get a parser cache that caches the
|
||||
* the output of HTML Purifier while keeping the original HTML lying
|
||||
* around, you may want to run Tidy on the resulting output or use
|
||||
* HTMLPurifier_DirectLex
|
||||
*/
|
||||
|
||||
class HTMLPurifier_Lexer_DOMLex extends HTMLPurifier_Lexer
|
||||
{
|
||||
|
||||
private $factory;
|
||||
|
||||
public function __construct() {
|
||||
// setup the factory
|
||||
parent::__construct();
|
||||
$this->factory = new HTMLPurifier_TokenFactory();
|
||||
}
|
||||
|
||||
public function tokenizeHTML($html, $config, $context) {
|
||||
|
||||
$html = $this->normalize($html, $config, $context);
|
||||
|
||||
// attempt to armor stray angled brackets that cannot possibly
|
||||
// form tags and thus are probably being used as emoticons
|
||||
if ($config->get('Core', 'AggressivelyFixLt')) {
|
||||
$char = '[^a-z!\/]';
|
||||
$comment = "/<!--(.*?)(-->|\z)/is";
|
||||
$html = preg_replace_callback($comment, array($this, 'callbackArmorCommentEntities'), $html);
|
||||
$html = preg_replace("/<($char)/i", '<\\1', $html);
|
||||
$html = preg_replace_callback($comment, array($this, 'callbackUndoCommentSubst'), $html); // fix comments
|
||||
}
|
||||
|
||||
// preprocess html, essential for UTF-8
|
||||
$html = $this->wrapHTML($html, $config, $context);
|
||||
|
||||
$doc = new DOMDocument();
|
||||
$doc->encoding = 'UTF-8'; // theoretically, the above has this covered
|
||||
|
||||
set_error_handler(array($this, 'muteErrorHandler'));
|
||||
$doc->loadHTML($html);
|
||||
restore_error_handler();
|
||||
|
||||
$tokens = array();
|
||||
$this->tokenizeDOM(
|
||||
$doc->getElementsByTagName('html')->item(0)-> // <html>
|
||||
getElementsByTagName('body')->item(0)-> // <body>
|
||||
getElementsByTagName('div')->item(0) // <div>
|
||||
, $tokens);
|
||||
return $tokens;
|
||||
}
|
||||
|
||||
/**
|
||||
* Recursive function that tokenizes a node, putting it into an accumulator.
|
||||
*
|
||||
* @param $node DOMNode to be tokenized.
|
||||
* @param $tokens Array-list of already tokenized tokens.
|
||||
* @param $collect Says whether or start and close are collected, set to
|
||||
* false at first recursion because it's the implicit DIV
|
||||
* tag you're dealing with.
|
||||
* @returns Tokens of node appended to previously passed tokens.
|
||||
*/
|
||||
protected function tokenizeDOM($node, &$tokens, $collect = false) {
|
||||
|
||||
// intercept non element nodes. WE MUST catch all of them,
|
||||
// but we're not getting the character reference nodes because
|
||||
// those should have been preprocessed
|
||||
if ($node->nodeType === XML_TEXT_NODE) {
|
||||
$tokens[] = $this->factory->createText($node->data);
|
||||
return;
|
||||
} elseif ($node->nodeType === XML_CDATA_SECTION_NODE) {
|
||||
// undo DOM's special treatment of <script> tags
|
||||
$tokens[] = $this->factory->createText($this->parseData($node->data));
|
||||
return;
|
||||
} elseif ($node->nodeType === XML_COMMENT_NODE) {
|
||||
$tokens[] = $this->factory->createComment($node->data);
|
||||
return;
|
||||
} elseif (
|
||||
// not-well tested: there may be other nodes we have to grab
|
||||
$node->nodeType !== XML_ELEMENT_NODE
|
||||
) {
|
||||
return;
|
||||
}
|
||||
|
||||
$attr = $node->hasAttributes() ?
|
||||
$this->transformAttrToAssoc($node->attributes) :
|
||||
array();
|
||||
|
||||
// We still have to make sure that the element actually IS empty
|
||||
if (!$node->childNodes->length) {
|
||||
if ($collect) {
|
||||
$tokens[] = $this->factory->createEmpty($node->tagName, $attr);
|
||||
}
|
||||
} else {
|
||||
if ($collect) { // don't wrap on first iteration
|
||||
$tokens[] = $this->factory->createStart(
|
||||
$tag_name = $node->tagName, // somehow, it get's dropped
|
||||
$attr
|
||||
);
|
||||
}
|
||||
foreach ($node->childNodes as $node) {
|
||||
// remember, it's an accumulator. Otherwise, we'd have
|
||||
// to use array_merge
|
||||
$this->tokenizeDOM($node, $tokens, true);
|
||||
}
|
||||
if ($collect) {
|
||||
$tokens[] = $this->factory->createEnd($tag_name);
|
||||
}
|
||||
}
|
||||
|
||||
}
|
||||
|
||||
/**
|
||||
* Converts a DOMNamedNodeMap of DOMAttr objects into an assoc array.
|
||||
*
|
||||
* @param $attribute_list DOMNamedNodeMap of DOMAttr objects.
|
||||
* @returns Associative array of attributes.
|
||||
*/
|
||||
protected function transformAttrToAssoc($node_map) {
|
||||
// NamedNodeMap is documented very well, so we're using undocumented
|
||||
// features, namely, the fact that it implements Iterator and
|
||||
// has a ->length attribute
|
||||
if ($node_map->length === 0) return array();
|
||||
$array = array();
|
||||
foreach ($node_map as $attr) {
|
||||
$array[$attr->name] = $attr->value;
|
||||
}
|
||||
return $array;
|
||||
}
|
||||
|
||||
/**
|
||||
* An error handler that mutes all errors
|
||||
*/
|
||||
public function muteErrorHandler($errno, $errstr) {}
|
||||
|
||||
/**
|
||||
* Callback function for undoing escaping of stray angled brackets
|
||||
* in comments
|
||||
*/
|
||||
public function callbackUndoCommentSubst($matches) {
|
||||
return '<!--' . strtr($matches[1], array('&'=>'&','<'=>'<')) . $matches[2];
|
||||
}
|
||||
|
||||
/**
|
||||
* Callback function that entity-izes ampersands in comments so that
|
||||
* callbackUndoCommentSubst doesn't clobber them
|
||||
*/
|
||||
public function callbackArmorCommentEntities($matches) {
|
||||
return '<!--' . str_replace('&', '&', $matches[1]) . $matches[2];
|
||||
}
|
||||
|
||||
/**
|
||||
* Wraps an HTML fragment in the necessary HTML
|
||||
*/
|
||||
protected function wrapHTML($html, $config, $context) {
|
||||
$def = $config->getDefinition('HTML');
|
||||
$ret = '';
|
||||
|
||||
if (!empty($def->doctype->dtdPublic) || !empty($def->doctype->dtdSystem)) {
|
||||
$ret .= '<!DOCTYPE html ';
|
||||
if (!empty($def->doctype->dtdPublic)) $ret .= 'PUBLIC "' . $def->doctype->dtdPublic . '" ';
|
||||
if (!empty($def->doctype->dtdSystem)) $ret .= '"' . $def->doctype->dtdSystem . '" ';
|
||||
$ret .= '>';
|
||||
}
|
||||
|
||||
$ret .= '<html><head>';
|
||||
$ret .= '<meta http-equiv="Content-Type" content="text/html; charset=utf-8" />';
|
||||
// No protection if $html contains a stray </div>!
|
||||
$ret .= '</head><body><div>'.$html.'</div></body></html>';
|
||||
return $ret;
|
||||
}
|
||||
|
||||
}
|
||||
|
||||
@@ -0,0 +1,498 @@
|
||||
<?php
|
||||
|
||||
require_once 'HTMLPurifier/Lexer.php';
|
||||
|
||||
HTMLPurifier_ConfigSchema::define(
|
||||
'Core', 'DirectLexLineNumberSyncInterval', 0, 'int', '
|
||||
<p>
|
||||
Specifies the number of tokens the DirectLex line number tracking
|
||||
implementations should process before attempting to resyncronize the
|
||||
current line count by manually counting all previous new-lines. When
|
||||
at 0, this functionality is disabled. Lower values will decrease
|
||||
performance, and this is only strictly necessary if the counting
|
||||
algorithm is buggy (in which case you should report it as a bug).
|
||||
This has no effect when %Core.MaintainLineNumbers is disabled or DirectLex is
|
||||
not being used. This directive has been available since 2.0.0.
|
||||
</p>
|
||||
');
|
||||
|
||||
/**
|
||||
* Our in-house implementation of a parser.
|
||||
*
|
||||
* A pure PHP parser, DirectLex has absolutely no dependencies, making
|
||||
* it a reasonably good default for PHP4. Written with efficiency in mind,
|
||||
* it can be four times faster than HTMLPurifier_Lexer_PEARSax3, although it
|
||||
* pales in comparison to HTMLPurifier_Lexer_DOMLex.
|
||||
*
|
||||
* @todo Reread XML spec and document differences.
|
||||
*/
|
||||
class HTMLPurifier_Lexer_DirectLex extends HTMLPurifier_Lexer
|
||||
{
|
||||
|
||||
/**
|
||||
* Whitespace characters for str(c)spn.
|
||||
*/
|
||||
protected $_whitespace = "\x20\x09\x0D\x0A";
|
||||
|
||||
/**
|
||||
* Callback function for script CDATA fudge
|
||||
* @param $matches, in form of array(opening tag, contents, closing tag)
|
||||
*/
|
||||
protected function scriptCallback($matches) {
|
||||
return $matches[1] . htmlspecialchars($matches[2], ENT_COMPAT, 'UTF-8') . $matches[3];
|
||||
}
|
||||
|
||||
public function tokenizeHTML($html, $config, $context) {
|
||||
|
||||
// special normalization for script tags without any armor
|
||||
// our "armor" heurstic is a < sign any number of whitespaces after
|
||||
// the first script tag
|
||||
if ($config->get('HTML', 'Trusted')) {
|
||||
$html = preg_replace_callback('#(<script[^>]*>)(\s*[^<].+?)(</script>)#si',
|
||||
array($this, 'scriptCallback'), $html);
|
||||
}
|
||||
|
||||
$html = $this->normalize($html, $config, $context);
|
||||
|
||||
$cursor = 0; // our location in the text
|
||||
$inside_tag = false; // whether or not we're parsing the inside of a tag
|
||||
$array = array(); // result array
|
||||
|
||||
$maintain_line_numbers = $config->get('Core', 'MaintainLineNumbers');
|
||||
|
||||
if ($maintain_line_numbers === null) {
|
||||
// automatically determine line numbering by checking
|
||||
// if error collection is on
|
||||
$maintain_line_numbers = $config->get('Core', 'CollectErrors');
|
||||
}
|
||||
|
||||
if ($maintain_line_numbers) $current_line = 1;
|
||||
else $current_line = false;
|
||||
$context->register('CurrentLine', $current_line);
|
||||
$nl = "\n";
|
||||
// how often to manually recalculate. This will ALWAYS be right,
|
||||
// but it's pretty wasteful. Set to 0 to turn off
|
||||
$synchronize_interval = $config->get('Core', 'DirectLexLineNumberSyncInterval');
|
||||
|
||||
$e = false;
|
||||
if ($config->get('Core', 'CollectErrors')) {
|
||||
$e =& $context->get('ErrorCollector');
|
||||
}
|
||||
|
||||
// infinite loop protection
|
||||
// has to be pretty big, since html docs can be big
|
||||
// we're allow two hundred thousand tags... more than enough?
|
||||
// NOTE: this is also used for synchronization, so watch out
|
||||
$loops = 0;
|
||||
|
||||
while(true) {
|
||||
|
||||
// infinite loop protection
|
||||
if (++$loops > 200000) return array();
|
||||
|
||||
// recalculate lines
|
||||
if (
|
||||
$maintain_line_numbers && // line number tracking is on
|
||||
$synchronize_interval && // synchronization is on
|
||||
$cursor > 0 && // cursor is further than zero
|
||||
$loops % $synchronize_interval === 0 // time to synchronize!
|
||||
) {
|
||||
$current_line = 1 + $this->substrCount($html, $nl, 0, $cursor);
|
||||
}
|
||||
|
||||
$position_next_lt = strpos($html, '<', $cursor);
|
||||
$position_next_gt = strpos($html, '>', $cursor);
|
||||
|
||||
// triggers on "<b>asdf</b>" but not "asdf <b></b>"
|
||||
// special case to set up context
|
||||
if ($position_next_lt === $cursor) {
|
||||
$inside_tag = true;
|
||||
$cursor++;
|
||||
}
|
||||
|
||||
if (!$inside_tag && $position_next_lt !== false) {
|
||||
// We are not inside tag and there still is another tag to parse
|
||||
$token = new
|
||||
HTMLPurifier_Token_Text(
|
||||
$this->parseData(
|
||||
substr(
|
||||
$html, $cursor, $position_next_lt - $cursor
|
||||
)
|
||||
)
|
||||
);
|
||||
if ($maintain_line_numbers) {
|
||||
$token->line = $current_line;
|
||||
$current_line += $this->substrCount($html, $nl, $cursor, $position_next_lt - $cursor);
|
||||
}
|
||||
$array[] = $token;
|
||||
$cursor = $position_next_lt + 1;
|
||||
$inside_tag = true;
|
||||
continue;
|
||||
} elseif (!$inside_tag) {
|
||||
// We are not inside tag but there are no more tags
|
||||
// If we're already at the end, break
|
||||
if ($cursor === strlen($html)) break;
|
||||
// Create Text of rest of string
|
||||
$token = new
|
||||
HTMLPurifier_Token_Text(
|
||||
$this->parseData(
|
||||
substr(
|
||||
$html, $cursor
|
||||
)
|
||||
)
|
||||
);
|
||||
if ($maintain_line_numbers) $token->line = $current_line;
|
||||
$array[] = $token;
|
||||
break;
|
||||
} elseif ($inside_tag && $position_next_gt !== false) {
|
||||
// We are in tag and it is well formed
|
||||
// Grab the internals of the tag
|
||||
$strlen_segment = $position_next_gt - $cursor;
|
||||
|
||||
if ($strlen_segment < 1) {
|
||||
// there's nothing to process!
|
||||
$token = new HTMLPurifier_Token_Text('<');
|
||||
$cursor++;
|
||||
continue;
|
||||
}
|
||||
|
||||
$segment = substr($html, $cursor, $strlen_segment);
|
||||
|
||||
if ($segment === false) {
|
||||
// somehow, we attempted to access beyond the end of
|
||||
// the string, defense-in-depth, reported by Nate Abele
|
||||
break;
|
||||
}
|
||||
|
||||
// Check if it's a comment
|
||||
if (
|
||||
substr($segment, 0, 3) === '!--'
|
||||
) {
|
||||
// re-determine segment length, looking for -->
|
||||
$position_comment_end = strpos($html, '-->', $cursor);
|
||||
if ($position_comment_end === false) {
|
||||
// uh oh, we have a comment that extends to
|
||||
// infinity. Can't be helped: set comment
|
||||
// end position to end of string
|
||||
if ($e) $e->send(E_WARNING, 'Lexer: Unclosed comment');
|
||||
$position_comment_end = strlen($html);
|
||||
$end = true;
|
||||
} else {
|
||||
$end = false;
|
||||
}
|
||||
$strlen_segment = $position_comment_end - $cursor;
|
||||
$segment = substr($html, $cursor, $strlen_segment);
|
||||
$token = new
|
||||
HTMLPurifier_Token_Comment(
|
||||
substr(
|
||||
$segment, 3, $strlen_segment - 3
|
||||
)
|
||||
);
|
||||
if ($maintain_line_numbers) {
|
||||
$token->line = $current_line;
|
||||
$current_line += $this->substrCount($html, $nl, $cursor, $strlen_segment);
|
||||
}
|
||||
$array[] = $token;
|
||||
$cursor = $end ? $position_comment_end : $position_comment_end + 3;
|
||||
$inside_tag = false;
|
||||
continue;
|
||||
}
|
||||
|
||||
// Check if it's an end tag
|
||||
$is_end_tag = (strpos($segment,'/') === 0);
|
||||
if ($is_end_tag) {
|
||||
$type = substr($segment, 1);
|
||||
$token = new HTMLPurifier_Token_End($type);
|
||||
if ($maintain_line_numbers) {
|
||||
$token->line = $current_line;
|
||||
$current_line += $this->substrCount($html, $nl, $cursor, $position_next_gt - $cursor);
|
||||
}
|
||||
$array[] = $token;
|
||||
$inside_tag = false;
|
||||
$cursor = $position_next_gt + 1;
|
||||
continue;
|
||||
}
|
||||
|
||||
// Check leading character is alnum, if not, we may
|
||||
// have accidently grabbed an emoticon. Translate into
|
||||
// text and go our merry way
|
||||
if (!ctype_alpha($segment[0])) {
|
||||
// XML: $segment[0] !== '_' && $segment[0] !== ':'
|
||||
if ($e) $e->send(E_NOTICE, 'Lexer: Unescaped lt');
|
||||
$token = new
|
||||
HTMLPurifier_Token_Text(
|
||||
'<' .
|
||||
$this->parseData(
|
||||
$segment
|
||||
) .
|
||||
'>'
|
||||
);
|
||||
if ($maintain_line_numbers) {
|
||||
$token->line = $current_line;
|
||||
$current_line += $this->substrCount($html, $nl, $cursor, $position_next_gt - $cursor);
|
||||
}
|
||||
$array[] = $token;
|
||||
$cursor = $position_next_gt + 1;
|
||||
$inside_tag = false;
|
||||
continue;
|
||||
}
|
||||
|
||||
// Check if it is explicitly self closing, if so, remove
|
||||
// trailing slash. Remember, we could have a tag like <br>, so
|
||||
// any later token processing scripts must convert improperly
|
||||
// classified EmptyTags from StartTags.
|
||||
$is_self_closing = (strrpos($segment,'/') === $strlen_segment-1);
|
||||
if ($is_self_closing) {
|
||||
$strlen_segment--;
|
||||
$segment = substr($segment, 0, $strlen_segment);
|
||||
}
|
||||
|
||||
// Check if there are any attributes
|
||||
$position_first_space = strcspn($segment, $this->_whitespace);
|
||||
|
||||
if ($position_first_space >= $strlen_segment) {
|
||||
if ($is_self_closing) {
|
||||
$token = new HTMLPurifier_Token_Empty($segment);
|
||||
} else {
|
||||
$token = new HTMLPurifier_Token_Start($segment);
|
||||
}
|
||||
if ($maintain_line_numbers) {
|
||||
$token->line = $current_line;
|
||||
$current_line += $this->substrCount($html, $nl, $cursor, $position_next_gt - $cursor);
|
||||
}
|
||||
$array[] = $token;
|
||||
$inside_tag = false;
|
||||
$cursor = $position_next_gt + 1;
|
||||
continue;
|
||||
}
|
||||
|
||||
// Grab out all the data
|
||||
$type = substr($segment, 0, $position_first_space);
|
||||
$attribute_string =
|
||||
trim(
|
||||
substr(
|
||||
$segment, $position_first_space
|
||||
)
|
||||
);
|
||||
if ($attribute_string) {
|
||||
$attr = $this->parseAttributeString(
|
||||
$attribute_string
|
||||
, $config, $context
|
||||
);
|
||||
} else {
|
||||
$attr = array();
|
||||
}
|
||||
|
||||
if ($is_self_closing) {
|
||||
$token = new HTMLPurifier_Token_Empty($type, $attr);
|
||||
} else {
|
||||
$token = new HTMLPurifier_Token_Start($type, $attr);
|
||||
}
|
||||
if ($maintain_line_numbers) {
|
||||
$token->line = $current_line;
|
||||
$current_line += $this->substrCount($html, $nl, $cursor, $position_next_gt - $cursor);
|
||||
}
|
||||
$array[] = $token;
|
||||
$cursor = $position_next_gt + 1;
|
||||
$inside_tag = false;
|
||||
continue;
|
||||
} else {
|
||||
// inside tag, but there's no ending > sign
|
||||
if ($e) $e->send(E_WARNING, 'Lexer: Missing gt');
|
||||
$token = new
|
||||
HTMLPurifier_Token_Text(
|
||||
'<' .
|
||||
$this->parseData(
|
||||
substr($html, $cursor)
|
||||
)
|
||||
);
|
||||
if ($maintain_line_numbers) $token->line = $current_line;
|
||||
// no cursor scroll? Hmm...
|
||||
$array[] = $token;
|
||||
break;
|
||||
}
|
||||
break;
|
||||
}
|
||||
|
||||
$context->destroy('CurrentLine');
|
||||
return $array;
|
||||
}
|
||||
|
||||
/**
|
||||
* PHP 4 compatible substr_count that implements offset and length
|
||||
*/
|
||||
protected function substrCount($haystack, $needle, $offset, $length) {
|
||||
static $oldVersion;
|
||||
if ($oldVersion === null) {
|
||||
$oldVersion = version_compare(PHP_VERSION, '5.1', '<');
|
||||
}
|
||||
if ($oldVersion) {
|
||||
$haystack = substr($haystack, $offset, $length);
|
||||
return substr_count($haystack, $needle);
|
||||
} else {
|
||||
return substr_count($haystack, $needle, $offset, $length);
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* Takes the inside of an HTML tag and makes an assoc array of attributes.
|
||||
*
|
||||
* @param $string Inside of tag excluding name.
|
||||
* @returns Assoc array of attributes.
|
||||
*/
|
||||
public function parseAttributeString($string, $config, $context) {
|
||||
$string = (string) $string; // quick typecast
|
||||
|
||||
if ($string == '') return array(); // no attributes
|
||||
|
||||
$e = false;
|
||||
if ($config->get('Core', 'CollectErrors')) {
|
||||
$e =& $context->get('ErrorCollector');
|
||||
}
|
||||
|
||||
// let's see if we can abort as quickly as possible
|
||||
// one equal sign, no spaces => one attribute
|
||||
$num_equal = substr_count($string, '=');
|
||||
$has_space = strpos($string, ' ');
|
||||
if ($num_equal === 0 && !$has_space) {
|
||||
// bool attribute
|
||||
return array($string => $string);
|
||||
} elseif ($num_equal === 1 && !$has_space) {
|
||||
// only one attribute
|
||||
list($key, $quoted_value) = explode('=', $string);
|
||||
$quoted_value = trim($quoted_value);
|
||||
if (!$key) {
|
||||
if ($e) $e->send(E_ERROR, 'Lexer: Missing attribute key');
|
||||
return array();
|
||||
}
|
||||
if (!$quoted_value) return array($key => '');
|
||||
$first_char = @$quoted_value[0];
|
||||
$last_char = @$quoted_value[strlen($quoted_value)-1];
|
||||
|
||||
$same_quote = ($first_char == $last_char);
|
||||
$open_quote = ($first_char == '"' || $first_char == "'");
|
||||
|
||||
if ( $same_quote && $open_quote) {
|
||||
// well behaved
|
||||
$value = substr($quoted_value, 1, strlen($quoted_value) - 2);
|
||||
} else {
|
||||
// not well behaved
|
||||
if ($open_quote) {
|
||||
if ($e) $e->send(E_ERROR, 'Lexer: Missing end quote');
|
||||
$value = substr($quoted_value, 1);
|
||||
} else {
|
||||
$value = $quoted_value;
|
||||
}
|
||||
}
|
||||
if ($value === false) $value = '';
|
||||
return array($key => $value);
|
||||
}
|
||||
|
||||
// setup loop environment
|
||||
$array = array(); // return assoc array of attributes
|
||||
$cursor = 0; // current position in string (moves forward)
|
||||
$size = strlen($string); // size of the string (stays the same)
|
||||
|
||||
// if we have unquoted attributes, the parser expects a terminating
|
||||
// space, so let's guarantee that there's always a terminating space.
|
||||
$string .= ' ';
|
||||
|
||||
// infinite loop protection
|
||||
$loops = 0;
|
||||
while(true) {
|
||||
|
||||
// infinite loop protection
|
||||
if (++$loops > 1000) {
|
||||
trigger_error('Infinite loop detected in attribute parsing', E_USER_WARNING);
|
||||
return array();
|
||||
}
|
||||
|
||||
if ($cursor >= $size) {
|
||||
break;
|
||||
}
|
||||
|
||||
$cursor += ($value = strspn($string, $this->_whitespace, $cursor));
|
||||
// grab the key
|
||||
|
||||
$key_begin = $cursor; //we're currently at the start of the key
|
||||
|
||||
// scroll past all characters that are the key (not whitespace or =)
|
||||
$cursor += strcspn($string, $this->_whitespace . '=', $cursor);
|
||||
|
||||
$key_end = $cursor; // now at the end of the key
|
||||
|
||||
$key = substr($string, $key_begin, $key_end - $key_begin);
|
||||
|
||||
if (!$key) {
|
||||
if ($e) $e->send(E_ERROR, 'Lexer: Missing attribute key');
|
||||
$cursor += strcspn($string, $this->_whitespace, $cursor + 1); // prevent infinite loop
|
||||
continue; // empty key
|
||||
}
|
||||
|
||||
// scroll past all whitespace
|
||||
$cursor += strspn($string, $this->_whitespace, $cursor);
|
||||
|
||||
if ($cursor >= $size) {
|
||||
$array[$key] = $key;
|
||||
break;
|
||||
}
|
||||
|
||||
// if the next character is an equal sign, we've got a regular
|
||||
// pair, otherwise, it's a bool attribute
|
||||
$first_char = @$string[$cursor];
|
||||
|
||||
if ($first_char == '=') {
|
||||
// key="value"
|
||||
|
||||
$cursor++;
|
||||
$cursor += strspn($string, $this->_whitespace, $cursor);
|
||||
|
||||
if ($cursor === false) {
|
||||
$array[$key] = '';
|
||||
break;
|
||||
}
|
||||
|
||||
// we might be in front of a quote right now
|
||||
|
||||
$char = @$string[$cursor];
|
||||
|
||||
if ($char == '"' || $char == "'") {
|
||||
// it's quoted, end bound is $char
|
||||
$cursor++;
|
||||
$value_begin = $cursor;
|
||||
$cursor = strpos($string, $char, $cursor);
|
||||
$value_end = $cursor;
|
||||
} else {
|
||||
// it's not quoted, end bound is whitespace
|
||||
$value_begin = $cursor;
|
||||
$cursor += strcspn($string, $this->_whitespace, $cursor);
|
||||
$value_end = $cursor;
|
||||
}
|
||||
|
||||
// we reached a premature end
|
||||
if ($cursor === false) {
|
||||
$cursor = $size;
|
||||
$value_end = $cursor;
|
||||
}
|
||||
|
||||
$value = substr($string, $value_begin, $value_end - $value_begin);
|
||||
if ($value === false) $value = '';
|
||||
$array[$key] = $this->parseData($value);
|
||||
$cursor++;
|
||||
|
||||
} else {
|
||||
// boolattr
|
||||
if ($key !== '') {
|
||||
$array[$key] = $key;
|
||||
} else {
|
||||
// purely theoretical
|
||||
if ($e) $e->send(E_ERROR, 'Lexer: Missing attribute key');
|
||||
}
|
||||
|
||||
}
|
||||
}
|
||||
return $array;
|
||||
}
|
||||
|
||||
}
|
||||
|
||||
@@ -0,0 +1,108 @@
|
||||
<?php
|
||||
|
||||
require_once 'XML/HTMLSax3.php'; // PEAR
|
||||
require_once 'HTMLPurifier/Lexer.php';
|
||||
|
||||
/**
|
||||
* Proof-of-concept lexer that uses the PEAR package XML_HTMLSax3 to parse HTML.
|
||||
*
|
||||
* PEAR, not suprisingly, also has a SAX parser for HTML. I don't know
|
||||
* very much about implementation, but it's fairly well written. However, that
|
||||
* abstraction comes at a price: performance. You need to have it installed,
|
||||
* and if the API changes, it might break our adapter. Not sure whether or not
|
||||
* it's UTF-8 aware, but it has some entity parsing trouble (in all areas,
|
||||
* text and attributes).
|
||||
*
|
||||
* Quite personally, I don't recommend using the PEAR class, and the defaults
|
||||
* don't use it. The unit tests do perform the tests on the SAX parser too, but
|
||||
* whatever it does for poorly formed HTML is up to it.
|
||||
*
|
||||
* @todo Generalize so that XML_HTMLSax is also supported.
|
||||
*
|
||||
* @warning Entity-resolution inside attributes is broken.
|
||||
*/
|
||||
|
||||
class HTMLPurifier_Lexer_PEARSax3 extends HTMLPurifier_Lexer
|
||||
{
|
||||
|
||||
/**
|
||||
* Internal accumulator array for SAX parsers.
|
||||
*/
|
||||
protected $tokens = array();
|
||||
|
||||
public function tokenizeHTML($string, $config, $context) {
|
||||
|
||||
$this->tokens = array();
|
||||
|
||||
$string = $this->normalize($string, $config, $context);
|
||||
|
||||
$parser = new XML_HTMLSax3();
|
||||
$parser->set_object($this);
|
||||
$parser->set_element_handler('openHandler','closeHandler');
|
||||
$parser->set_data_handler('dataHandler');
|
||||
$parser->set_escape_handler('escapeHandler');
|
||||
|
||||
// doesn't seem to work correctly for attributes
|
||||
$parser->set_option('XML_OPTION_ENTITIES_PARSED', 1);
|
||||
|
||||
$parser->parse($string);
|
||||
|
||||
return $this->tokens;
|
||||
|
||||
}
|
||||
|
||||
/**
|
||||
* Open tag event handler, interface is defined by PEAR package.
|
||||
*/
|
||||
public function openHandler(&$parser, $name, $attrs, $closed) {
|
||||
// entities are not resolved in attrs
|
||||
foreach ($attrs as $key => $attr) {
|
||||
$attrs[$key] = $this->parseData($attr);
|
||||
}
|
||||
if ($closed) {
|
||||
$this->tokens[] = new HTMLPurifier_Token_Empty($name, $attrs);
|
||||
} else {
|
||||
$this->tokens[] = new HTMLPurifier_Token_Start($name, $attrs);
|
||||
}
|
||||
return true;
|
||||
}
|
||||
|
||||
/**
|
||||
* Close tag event handler, interface is defined by PEAR package.
|
||||
*/
|
||||
public function closeHandler(&$parser, $name) {
|
||||
// HTMLSax3 seems to always send empty tags an extra close tag
|
||||
// check and ignore if you see it:
|
||||
// [TESTME] to make sure it doesn't overreach
|
||||
if ($this->tokens[count($this->tokens)-1]->type == 'empty') {
|
||||
return true;
|
||||
}
|
||||
$this->tokens[] = new HTMLPurifier_Token_End($name);
|
||||
return true;
|
||||
}
|
||||
|
||||
/**
|
||||
* Data event handler, interface is defined by PEAR package.
|
||||
*/
|
||||
public function dataHandler(&$parser, $data) {
|
||||
$this->tokens[] = new HTMLPurifier_Token_Text($data);
|
||||
return true;
|
||||
}
|
||||
|
||||
/**
|
||||
* Escaped text handler, interface is defined by PEAR package.
|
||||
*/
|
||||
public function escapeHandler(&$parser, $data) {
|
||||
if (strpos($data, '--') === 0) {
|
||||
$this->tokens[] = new HTMLPurifier_Token_Comment($data);
|
||||
}
|
||||
// CDATA is handled elsewhere, but if it was handled here:
|
||||
//if (strpos($data, '[CDATA[') === 0) {
|
||||
// $this->tokens[] = new HTMLPurifier_Token_Text(
|
||||
// substr($data, 7, strlen($data) - 9) );
|
||||
//}
|
||||
return true;
|
||||
}
|
||||
|
||||
}
|
||||
|
||||
File diff suppressed because it is too large
Load Diff
Reference in New Issue
Block a user