* @copyright 2004-2009 Roman Ivanov, Miguel Vazquez Gocobachi
* @license http://www.debian.org/misc/bsd.license BSD License (3 Clause)
* @version SVN: $Id$
* @link http://pear.php.net/package/HTML_Safe
*/
/**
* This package requires HTMLSax3 package
*/
require_once 'XML/HTMLSax3.php';
/**
* HTML_Safe Parser
*
* This parser strips down all potentially dangerous content within HTML:
*
* - opening tag without its closing tag
* - closing tag without its opening tag
* - any of these tags: "base", "basefont", "head", "html", "body", "applet",
* "object", "iframe", "frame", "frameset", "script", "layer", "ilayer", "embed",
* "bgsound", "link", "meta", "style", "title", "blink", "xml" etc.
* - any of these attributes: on*, data*, dynsrc
* - javascript:/vbscript:/about: etc. protocols
* - expression/behavior etc. in styles
* - any other active content
*
* It also tries to convert code to XHTML valid, but htmltidy is far better
* solution for this task.
*
* Example:
*
* $parser = new HTML_Safe;
* $result = $parser->parse($doc);
*
*
* @category HTML
* @package HTML_Safe
* @author Roman Ivanov
* @author Miguel Vazquez Gocobachi
* @copyright 2004-2009 Roman Ivanov, Miguel Vazquez Gocobachi
* @license http://www.debian.org/misc/bsd.license BSD License (3 Clause)
* @version Release: @package_version@
* @link http://pear.php.net/package/HTML_Safe
*/
class HTML_Safe
{
/**
* Storage for resulting HTML output
*
* @var string
*/
protected $xhtml = '';
/**
* Array of counters for each tag
*
* @var array
*/
protected $counter = array();
/**
* Stack of unclosed tags
*
* @var array
*/
protected $stack = array();
/**
* Array of counters for tags that must be deleted with all content
*
* @var array
*/
protected $dcCounter = array();
/**
* Stack of unclosed tags that must be deleted with all content
*
* @var array
*/
protected $dcStack = array();
/**
* Stores level of list (ol/ul) nesting
*
* @var int
*/
protected $listScope = 0;
/**
* Stack of unclosed list tags
*
* @var array
*/
protected $liStack = array();
/**
* Array of prepared regular expressions for protocols (schemas) matching
*
* @var array
*/
protected $protoRegexps = array();
/**
* Array of prepared regular expressions for CSS matching
*
* @var array
*/
protected $cssRegexps = array();
/**
* Allowed tags
*
* @var array
*/
protected $allowTags = array();
/**
* List of single tags ("")
*
* @var array
*/
public $singleTags = array('area', 'br', 'img', 'input', 'hr', 'wbr', );
/**
* List of dangerous tags (such tags will be deleted)
*
* @var array
*/
public $deleteTags = array(
'applet', 'base', 'basefont', 'bgsound', 'blink', 'body',
'embed', 'frame', 'frameset', 'head', 'html', 'ilayer',
'iframe', 'layer', 'link', 'meta', 'object', 'style',
'title', 'script',
);
/**
* List of dangerous tags (such tags will be deleted, and all content
* inside this tags will be also removed)
*
* @var array
*/
public $deleteTagsContent = array('script', 'style', 'title', 'xml', );
/**
* Type of protocols filtering ('white' or 'black')
*
* @var string
*/
public $protocolFiltering = 'white';
/**
* List of "dangerous" protocols (used for blacklist-filtering)
*
* @var array
*/
public $blackProtocols = array(
'about', 'chrome', 'data', 'disk', 'hcp',
'help', 'javascript', 'livescript', 'lynxcgi', 'lynxexec',
'ms-help', 'ms-its', 'mhtml', 'mocha', 'opera',
'res', 'resource', 'shell', 'vbscript', 'view-source',
'vnd.ms.radio', 'wysiwyg',
);
/**
* List of "safe" protocols (used for whitelist-filtering)
*
* @var array
*/
public $whiteProtocols = array(
'ed2k', 'file', 'ftp', 'gopher', 'http', 'https',
'irc', 'mailto', 'news', 'nntp', 'telnet', 'webcal',
'xmpp', 'callto',
);
/**
* List of attributes that can contain protocols
*
* @var array
*/
public $protocolAttributes = array(
'action', 'background', 'codebase', 'dynsrc', 'href', 'lowsrc', 'src',
);
/**
* List of dangerous CSS keywords
*
* Whole style="" attribute will be removed, if parser will find one of
* these keywords
*
* @var array
*/
public $cssKeywords = array(
'absolute', 'behavior', 'behaviour', 'content', 'expression',
'fixed', 'include-source', 'moz-binding',
);
/**
* List of tags that can have no "closing tag"
*
* @var array
* @deprecated XHTML does not allow such tags
*/
public $noClose = array();
/**
* List of block-level tags that terminates paragraph
*
* Paragraph will be closed when this tags opened
*
* @var array
*/
public $closeParagraph = array(
'address', 'blockquote', 'center', 'dd', 'dir', 'div',
'dl', 'dt', 'h1', 'h2', 'h3', 'h4',
'h5', 'h6', 'hr', 'isindex', 'listing', 'marquee',
'menu', 'multicol', 'ol', 'p', 'plaintext', 'pre',
'table', 'ul', 'xmp',
);
/**
* List of table tags, all table tags outside a table will be removed
*
* @var array
*/
public $tableTags = array(
'caption', 'col', 'colgroup', 'tbody', 'td', 'tfoot', 'th',
'thead', 'tr',
);
/**
* List of list tags
*
* @var array
*/
public $listTags = array('dir', 'menu', 'ol', 'ul', 'dl', );
/**
* List of dangerous attributes
*
* @var array
*/
public $attributes = array('dynsrc', 'id', 'name', );
/**
* List of allowed "namespaced" attributes
*
* @var array
*/
public $attributesNS = array('xml:lang', );
/**
* Constructs class
*
* @access public
*/
public function __construct()
{
//making regular expressions based on Proto & CSS arrays
foreach ($this->blackProtocols as $proto) {
$preg = "/[\s\x01-\x1F]*";
for ($i=0; $iprotoRegexps[] = $preg;
}
foreach ($this->cssKeywords as $css) {
$this->cssRegexps[] = '/' . $css . '/i';
}
return true;
}
/**
* Handles the writing of attributes - called from $this->openHandler()
*
* @param array $attrs array of attributes $name => $value
*
* @return boolean
*/
protected function writeAttrs($attrs)
{
if (is_array($attrs)) {
foreach ($attrs as $name => $value) {
$name = strtolower($name);
if (strpos($name, 'on') === 0) {
continue;
}
if (strpos($name, 'data') === 0) {
continue;
}
if (in_array($name, $this->attributes)) {
continue;
}
if (!preg_match('/^[a-z0-9]+$/i', $name)) {
if (!in_array($name, $this->attributesNS)) {
continue;
}
}
if (($value === true) || (is_null($value))) {
$value = $name;
}
if ($name == 'style') {
// removes insignificant backslahes
$value = str_replace("\\", '', $value);
// removes CSS comments
while (1) {
$_value = preg_replace('!/\*.*?\*/!s', '', $value);
if ($_value == $value) {
break;
}
$value = $_value;
}
// replace all & to &
$value = str_replace('&', '&', $value);
$value = str_replace('&', '&', $value);
foreach ($this->cssRegexps as $css) {
if (preg_match($css, $value)) {
continue 2;
}
}
foreach ($this->protoRegexps as $proto) {
if (preg_match($proto, $value)) {
continue 2;
}
}
}
$tempval = preg_replace('/&#(\d+);?/me', "chr('\\1')", $value); //"'
$tempval = preg_replace(
'/&#x([0-9a-f]+);?/mei',
"chr(hexdec('\\1'))",
$tempval
);
if ((in_array($name, $this->protocolAttributes))
&& (strpos($tempval, ':') !== false)
) {
if ($this->protocolFiltering == 'black') {
foreach ($this->protoRegexps as $proto) {
if (preg_match($proto, $tempval)) {
continue 2;
}
}
} else {
$_tempval = explode(':', $tempval);
$proto = $_tempval[0];
if (!in_array($proto, $this->whiteProtocols)) {
continue;
}
}
}
$value = str_replace("\"", '"', $value);
$this->xhtml .= ' ' . $name . '="' . $value . '"';
}
}
return true;
}
/**
* Opening tag handler - called from HTMLSax
*
* @param object &$parser HTML Parser
* @param string $name tag name
* @param array $attrs tag attributes
*
* @return boolean
*/
public function openHandler(&$parser, $name, $attrs)
{
$name = strtolower($name);
if (in_array($name, $this->deleteTagsContent)) {
array_push($this->dcStack, $name);
$this->dcCounter[$name] = isset($this->dcCounter[$name])
? $this->dcCounter[$name]+1 : 1;
}
if (count($this->dcStack) != 0) {
return true;
}
if (in_array($name, $this->deleteTags)
&& !in_array($name, $this->allowTags)
) {
return true;
}
if (!preg_match("/^[a-z0-9]+$/i", $name)) {
if (preg_match("!(?:\@|://)!i", $name)) {
$this->xhtml .= '<' . $name . '>';
}
return true;
}
if (in_array($name, $this->singleTags)) {
$this->xhtml .= '<' . $name;
$this->writeAttrs($attrs);
$this->xhtml .= ' />';
return true;
}
// TABLES: cannot open table elements when we are not inside table
if ((isset($this->counter['table']))
&& ($this->counter['table'] <= 0)
&& (in_array($name, $this->tableTags))
) {
return true;
}
// PARAGRAPHS: close paragraph when closeParagraph tags opening
if ((in_array($name, $this->closeParagraph))
&& (in_array('p', $this->stack))
) {
$this->closeHandler($parser, 'p');
}
// LISTS: we should close if of the same level opening
if (($name == 'li') && count($this->liStack)
&& ($this->listScope == $this->liStack[count($this->liStack) - 1])
) {
$this->closeHandler($parser, 'li');
}
// LISTS: we want to know on what nesting level of lists we are
if (in_array($name, $this->listTags)) {
++$this->listScope;
}
if ($name == 'li') {
array_push($this->liStack, $this->listScope);
}
$this->xhtml .= '<' . $name;
$this->writeAttrs($attrs);
$this->xhtml .= '>';
array_push($this->stack, $name);
$this->counter[$name] = isset($this->counter[$name])
? ($this->counter[$name] + 1) : 1;
return true;
}
/**
* Closing tag handler - called from HTMLSax
*
* @param object &$parser HTML parser
* @param string $name tag name
*
* @return boolean
*/
public function closeHandler(&$parser, $name)
{
$name = strtolower($name);
if (isset($this->dcCounter[$name])
&& ($this->dcCounter[$name] > 0)
&& (in_array($name, $this->deleteTagsContent))
) {
while ($name != ($tag = array_pop($this->dcStack))) {
--$this->dcCounter[$tag];
}
--$this->dcCounter[$name];
}
if (count($this->dcStack) != 0) {
return true;
}
if ((isset($this->counter[$name])) && ($this->counter[$name] > 0)) {
while ($name != ($tag = array_pop($this->stack))) {
$this->closeTag($tag);
}
$this->closeTag($name);
}
return true;
}
/**
* Closes tag
*
* @param string $tag tag name
*
* @return boolean
*/
protected function closeTag($tag)
{
if (!in_array($tag, $this->noClose)) {
$this->xhtml .= '';
}
--$this->counter[$tag];
if (in_array($tag, $this->listTags)) {
--$this->listScope;
}
if ($tag == 'li') {
array_pop($this->liStack);
}
return true;
}
/**
* Character data handler - called from HTMLSax
*
* @param object &$parser HTML parser
* @param string $data textual data
*
* @return boolean
*/
public function dataHandler(&$parser, $data)
{
if (count($this->dcStack) == 0) {
$this->xhtml .= $data;
}
return true;
}
/**
* Escape handler - called from HTMLSax
*
* @param object &$parser HTML parser
* @param string $data comments or other type of data
*
* @return boolean
*/
public function escapeHandler(&$parser, $data)
{
return true;
}
/**
* Allow tags
*
* Example:
*
* $safe = new HTML_Safe;
* $safe->setAllowTags(array('body'));
*
*
* @param array $tags Tags to allow
*
* @return void
*/
public function setAllowTags($tags = array())
{
if (is_array($tags)) {
$this->allowTags = $tags;
}
}
/**
* Returns the allowed tags
*
* @return array
*/
public function getAllowTags()
{
return $this->allowTags;
}
/**
* Reset the allowed tags
*
* @return void
*/
public function resetAllowTags()
{
$this->allowTags = array();
}
/**
* Returns the XHTML document
*
* @return string Processed (X)HTML document
*/
public function getXHTML()
{
while ($tag = array_pop($this->stack)) {
$this->closeTag($tag);
}
return $this->xhtml;
}
/**
* Clears current document data
*
* @return boolean
*/
public function clear()
{
$this->xhtml = '';
return true;
}
/**
* Main parsing fuction
*
* @param string $doc HTML document for processing
*
* @return string Processed (X)HTML document
*/
public function parse($doc)
{
$result = '';
// Save all '<' symbols
$doc = preg_replace("/<(?=[^a-zA-Z\/\!\?\%])/", '<', $doc);
// UTF7 pack
$doc = $this->repackUTF7($doc);
// Instantiate the parser
$parser = new XML_HTMLSax3;
// Set up the parser
$parser->set_object($this);
$parser->set_element_handler('openHandler', 'closeHandler');
$parser->set_data_handler('dataHandler');
$parser->set_escape_handler('escapeHandler');
$parser->parse($doc);
$result = $this->getXHTML();
$this->clear();
return $result;
}
/**
* UTF-7 decoding fuction
*
* @param string $str HTML document for recode ASCII part of UTF-7 back to ASCII
* @return string Decoded document
* @access private
*/
function repackUTF7($str)
{
return preg_replace_callback('!\+([0-9a-zA-Z/]+)\-!', array($this, 'repackUTF7Callback'), $str);
}
/**
* Additional UTF-7 decoding fuction
*
* @param string $str String for recode ASCII part of UTF-7 back to ASCII
* @return string Recoded string
* @access private
*/
function repackUTF7Callback($str)
{
$str = base64_decode($str[1]);
$str = preg_replace_callback('/^((?:\x00.)*)((?:[^\x00].)+)/', array($this, 'repackUTF7Back'), $str);
return preg_replace('/\x00(.)/', '$1', $str);
}
/**
* Additional UTF-7 encoding fuction
*
* @param string $str String for recode ASCII part of UTF-7 back to ASCII
* @return string Recoded string
* @access private
*/
function repackUTF7Back($str)
{
return $str[1].'+'.rtrim(base64_encode($str[2]), '=').'-';
}
}
HTML_Safe-0.10.1/tests/testHTML_Safe.php 0000644 0000765 0000024 00000001502 11425670544 016725 0 ustar demrit staff my text