| | |
| | | |
| | | /* Block elements which could be empty but cannot be returned in short form (<tag />) */ |
| | | static $block_elements = array('div', 'p', 'pre', 'blockquote', 'a', 'font', 'center', |
| | | 'table', 'ul', 'h1', 'h2', 'h3', 'h4', 'h5', 'h6', 'ol', 'dl', 'strong', 'i', 'b'); |
| | | 'table', 'ul', 'h1', 'h2', 'h3', 'h4', 'h5', 'h6', 'ol', 'dl', 'strong', 'i', 'b', 'u'); |
| | | |
| | | /* State for linked objects in HTML */ |
| | | public $extlinks = false; |
| | |
| | | unset($p['html_elements'], $p['html_attribs'], $p['ignore_elements'], $p['block_elements']); |
| | | $this->config = $p + array('show_washed'=>true, 'allow_remote'=>false, 'cid_map'=>array()); |
| | | } |
| | | |
| | | |
| | | /* Register a callback function for a certain tag */ |
| | | public function add_callback($tagName, $callback) |
| | | { |
| | | $this->handlers[$tagName] = $callback; |
| | | } |
| | | |
| | | |
| | | /* Check CSS style */ |
| | | private function wash_style($style) { |
| | | $s = ''; |
| | |
| | | preg_match('/^(url\(\s*[\'"]?([^\'"\)]*)[\'"]?\s*\)'./*1,2*/ |
| | | '|rgb\(\s*[0-9]+\s*,\s*[0-9]+\s*,\s*[0-9]+\s*\)'. |
| | | '|-?[0-9.]+\s*(em|ex|px|cm|mm|in|pt|pc|deg|rad|grad|ms|s|hz|khz|%)?'. |
| | | '|#[0-9a-f]{3,6}|[a-z0-9\-]+'. |
| | | '|#[0-9a-f]{3,6}|[a-z0-9", -]+'. |
| | | ')\s*/i', $str, $match)) { |
| | | if ($match[2]) { |
| | | if (($src = $this->config['cid_map'][$match[2]]) |
| | | || ($src = $this->config['cid_map'][$this->config['base_url'].$match[2]])) { |
| | | $value .= ' url('.htmlspecialchars($src, ENT_QUOTES) . ')'; |
| | | } |
| | | else if (preg_match('/^(http|https|ftp):.*$/i', $match[2], $url)) { |
| | | else if (preg_match('!^(https?:)?//[a-z0-9/._+-]+$!i', $match[2], $url)) { |
| | | if ($this->config['allow_remote']) |
| | | $value .= ' url('.htmlspecialchars($url[0], ENT_QUOTES).')'; |
| | | else |
| | |
| | | $value .= ' url('.htmlspecialchars($match[2], ENT_QUOTES).')'; |
| | | } |
| | | } |
| | | else if ($match[0] != 'url' && $match[0] != 'rbg') //whitelist ? |
| | | else if ($match[0] != 'url' && $match[0] != 'rgb') //whitelist ? |
| | | $value .= ' ' . $match[0]; |
| | | |
| | | $str = substr($str, strlen($match[0])); |
| | | } |
| | | if ($value) |
| | |
| | | $key = strtolower($key); |
| | | $value = $node->getAttribute($key); |
| | | if (isset($this->_html_attribs[$key]) || |
| | | ($key == 'href' && preg_match('/^(http:|https:|ftp:|mailto:|#).+/i', $value))) |
| | | ($key == 'href' && preg_match('!^(http:|https:|ftp:|mailto:|//|#).+!i', $value))) |
| | | $t .= ' ' . $key . '="' . htmlspecialchars($value, ENT_QUOTES) . '"'; |
| | | else if ($key == 'style' && ($style = $this->wash_style($value))) |
| | | $t .= ' style="' . $style . '"'; |
| | | else if ($key == 'style' && ($style = $this->wash_style($value))) { |
| | | $quot = strpos($style, '"') !== false ? "'" : '"'; |
| | | $t .= ' style=' . $quot . $style . $quot; |
| | | } |
| | | else if ($key == 'background' || ($key == 'src' && strtolower($node->tagName) == 'img')) { //check tagName anyway |
| | | if (($src = $this->config['cid_map'][$value]) |
| | | || ($src = $this->config['cid_map'][$this->config['base_url'].$value])) { |
| | |
| | | case XML_ELEMENT_NODE: //Check element |
| | | $tagName = strtolower($node->tagName); |
| | | if ($callback = $this->handlers[$tagName]) { |
| | | $dump .= call_user_func($callback, $tagName, $this->wash_attribs($node), $this->dumpHtml($node)); |
| | | $dump .= call_user_func($callback, $tagName, $this->wash_attribs($node), $this->dumpHtml($node), $this); |
| | | } |
| | | else if (isset($this->_html_elements[$tagName])) { |
| | | $content = $this->dumpHtml($node); |
| | |
| | | return $this->dumpHtml($node); |
| | | } |
| | | |
| | | /** |
| | | * Getter for config parameters |
| | | */ |
| | | public function get_config($prop) |
| | | { |
| | | return $this->config[$prop]; |
| | | } |
| | | |
| | | } |
| | | |
| | | ?> |