| commit | author | age | ||
| 7ac944 | 1 | <?php |
| AM | 2 | |
| 7b9245 | 3 | /* |
| 7ac944 | 4 | +-----------------------------------------------------------------------+ |
| AM | 5 | | This file is part of the Roundcube Webmail client | |
| 6 | | Copyright (C) 2008-2012, The Roundcube Dev Team | | |
| 7 | | | | |
| 8 | | Licensed under the GNU General Public License version 3 or | | |
| 9 | | any later version with exceptions for skins & plugins. | | |
| 10 | | See the README file for a full license statement. | | |
| 11 | | | | |
| 12 | | PURPOSE: | | |
| 13 | | Utility class providing HTML sanityzer (based on Washtml class) | | |
| 14 | +-----------------------------------------------------------------------+ | |
| 15 | | Author: Thomas Bruederli <roundcube@gmail.com> | | |
| 16 | | Author: Aleksander Machniak <alec@alec.pl> | | |
| 17 | | Author: Frederic Motte <fmotte@ubixis.com> | | |
| 18 | +-----------------------------------------------------------------------+ | |
| 19 | */ | |
| 20 | ||
| 7b9245 | 21 | /* |
| 7ac944 | 22 | * Washtml, a HTML sanityzer. |
| AM | 23 | * |
| 24 | * Copyright (c) 2007 Frederic Motte <fmotte@ubixis.com> | |
| 25 | * All rights reserved. | |
| 26 | * | |
| 27 | * Redistribution and use in source and binary forms, with or without | |
| 28 | * modification, are permitted provided that the following conditions | |
| 29 | * are met: | |
| 30 | * 1. Redistributions of source code must retain the above copyright | |
| 31 | * notice, this list of conditions and the following disclaimer. | |
| 32 | * 2. Redistributions in binary form must reproduce the above copyright | |
| 33 | * notice, this list of conditions and the following disclaimer in the | |
| 34 | * documentation and/or other materials provided with the distribution. | |
| 35 | * | |
| 36 | * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR | |
| 37 | * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES | |
| 38 | * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. | |
| 39 | * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, | |
| 40 | * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT | |
| 41 | * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, | |
| 42 | * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY | |
| 43 | * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT | |
| 44 | * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF | |
| 45 | * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. | |
| 46 | * | |
| 47 | * OVERVIEW: | |
| 48 | * | |
| 49 | * Wahstml take an untrusted HTML and return a safe html string. | |
| 50 | * | |
| 51 | * SYNOPSIS: | |
| 52 | * | |
| 53 | * $washer = new washtml($config); | |
| 54 | * $washer->wash($html); | |
| 55 | * It return a sanityzed string of the $html parameter without html and head tags. | |
| 56 | * $html is a string containing the html code to wash. | |
| 57 | * $config is an array containing options: | |
| 58 | * $config['allow_remote'] is a boolean to allow link to remote images. | |
| 59 | * $config['blocked_src'] string with image-src to be used for blocked remote images | |
| 60 | * $config['show_washed'] is a boolean to include washed out attributes as x-washed | |
| 61 | * $config['cid_map'] is an array where cid urls index urls to replace them. | |
| 62 | * $config['charset'] is a string containing the charset of the HTML document if it is not defined in it. | |
| 63 | * $washer->extlinks is a reference to a boolean that is set to true if remote images were removed. (FE: show remote images link) | |
| 64 | * | |
| 65 | * INTERNALS: | |
| 66 | * | |
| 67 | * Only tags and attributes in the static lists $html_elements and $html_attributes | |
| 68 | * are kept, inline styles are also filtered: all style identifiers matching | |
| 69 | * /[a-z\-]/i are allowed. Values matching colors, sizes, /[a-z\-]/i and safe | |
| 70 | * urls if allowed and cid urls if mapped are kept. | |
| 71 | * | |
| 72 | * Roundcube Changes: | |
| 73 | * - added $block_elements | |
| 74 | * - changed $ignore_elements behaviour | |
| 75 | * - added RFC2397 support | |
| 76 | * - base URL support | |
| 77 | * - invalid HTML comments removal before parsing | |
| 78 | * - "fixing" unitless CSS values for XHTML output | |
| 79 | * - base url resolving | |
| 80 | */ | |
| 81 | ||
| 82 | /** | |
| 83 | * Utility class providing HTML sanityzer | |
| 84 | * | |
| 85 | * @package Framework | |
| 86 | * @subpackage Utils | |
| 87 | */ | |
| 88 | class rcube_washtml | |
| 89 | { | |
| 90 | /* Allowed HTML elements (default) */ | |
| 91 | static $html_elements = array('a', 'abbr', 'acronym', 'address', 'area', 'b', | |
| 92 | 'basefont', 'bdo', 'big', 'blockquote', 'br', 'caption', 'center', | |
| 93 | 'cite', 'code', 'col', 'colgroup', 'dd', 'del', 'dfn', 'dir', 'div', 'dl', | |
| 94 | 'dt', 'em', 'fieldset', 'font', 'h1', 'h2', 'h3', 'h4', 'h5', 'h6', 'hr', 'i', | |
| 95 | 'ins', 'label', 'legend', 'li', 'map', 'menu', 'nobr', 'ol', 'p', 'pre', 'q', | |
| 96 | 's', 'samp', 'small', 'span', 'strike', 'strong', 'sub', 'sup', 'table', | |
| 97 | 'tbody', 'td', 'tfoot', 'th', 'thead', 'tr', 'tt', 'u', 'ul', 'var', 'wbr', 'img', | |
| c5bfe6 | 98 | 'video', 'source', |
| 7ac944 | 99 | // form elements |
| 3e4b7c | 100 | 'button', 'input', 'textarea', 'select', 'option', 'optgroup', |
| AM | 101 | // SVG |
| 102 | 'svg', 'altglyph', 'altglyphdef', 'altglyphitem', 'animate', 'animatecolor', | |
| 103 | 'animatemotion', 'animatetransform', 'circle', 'clippath', 'defs', 'desc', | |
| 104 | 'ellipse', 'font', 'g', 'glyph', 'glyphref', 'hkern', 'image', 'line', | |
| 105 | 'lineargradient', 'marker', 'mask', 'mpath', 'path', 'pattern', | |
| 106 | 'polygon', 'polyline', 'radialgradient', 'rect', 'set', 'stop', 'switch', 'symbol', | |
| 107 | 'text', 'textpath', 'tref', 'tspan', 'use', 'view', 'vkern', 'filter', | |
| 108 | // SVG Filters | |
| 109 | 'feblend', 'fecolormatrix', 'fecomponenttransfer', 'fecomposite', | |
| 110 | 'feconvolvematrix', 'fediffuselighting', 'fedisplacementmap', | |
| 111 | 'feflood', 'fefunca', 'fefuncb', 'fefuncg', 'fefuncr', 'fegaussianblur', | |
| 112 | 'feimage', 'femerge', 'femergenode', 'femorphology', 'feoffset', | |
| 113 | 'fespecularlighting', 'fetile', 'feturbulence', | |
| 7ac944 | 114 | ); |
| AM | 115 | |
| 116 | /* Ignore these HTML tags and their content */ | |
| 117 | static $ignore_elements = array('script', 'applet', 'embed', 'object', 'style'); | |
| 118 | ||
| 119 | /* Allowed HTML attributes */ | |
| 120 | static $html_attribs = array('name', 'class', 'title', 'alt', 'width', 'height', | |
| 121 | 'align', 'nowrap', 'col', 'row', 'id', 'rowspan', 'colspan', 'cellspacing', | |
| 122 | 'cellpadding', 'valign', 'bgcolor', 'color', 'border', 'bordercolorlight', | |
| 123 | 'bordercolordark', 'face', 'marginwidth', 'marginheight', 'axis', 'border', | |
| 124 | 'abbr', 'char', 'charoff', 'clear', 'compact', 'coords', 'vspace', 'hspace', | |
| 125 | 'cellborder', 'size', 'lang', 'dir', 'usemap', 'shape', 'media', | |
| 847c77 | 126 | 'background', 'src', 'poster', 'href', |
| 7ac944 | 127 | // attributes of form elements |
| 3e4b7c | 128 | 'type', 'rows', 'cols', 'disabled', 'readonly', 'checked', 'multiple', 'value', |
| AM | 129 | // SVG |
| 130 | 'accent-height', 'accumulate', 'additivive', 'alignment-baseline', | |
| 131 | 'ascent', 'attributename', 'attributetype', 'azimuth', 'basefrequency', 'baseprofile', | |
| 132 | 'baseline-shift', 'begin', 'bias', 'by', 'clip', 'clip-path', 'clip-rule', | |
| 133 | 'color', 'color-interpolation', 'color-interpolation-filters', 'color-profile', | |
| 134 | 'color-rendering', 'cx', 'cy', 'd', 'dx', 'dy', 'diffuseconstant', 'direction', | |
| 135 | 'display', 'divisor', 'dur', 'edgemode', 'elevation', 'end', 'fill', 'fill-opacity', | |
| 136 | 'fill-rule', 'filter', 'flood-color', 'flood-opacity', 'font-family', 'font-size', | |
| 847c77 | 137 | 'font-size-adjust', 'font-stretch', 'font-style', 'font-variant', 'font-weight', 'from', |
| 3e4b7c | 138 | 'fx', 'fy', 'g1', 'g2', 'glyph-name', 'glyphref', 'gradientunits', 'gradienttransform', |
| AM | 139 | 'image-rendering', 'in', 'in2', 'k', 'k1', 'k2', 'k3', 'k4', 'kerning', 'keypoints', |
| 140 | 'keysplines', 'keytimes', 'lengthadjust', 'letter-spacing', 'kernelmatrix', | |
| 141 | 'kernelunitlength', 'lighting-color', 'local', 'marker-end', 'marker-mid', | |
| 142 | 'marker-start', 'markerheight', 'markerunits', 'markerwidth', 'maskcontentunits', | |
| 143 | 'maskunits', 'max', 'mask', 'mode', 'min', 'numoctaves', 'offset', 'operator', | |
| 144 | 'opacity', 'order', 'orient', 'orientation', 'origin', 'overflow', 'paint-order', | |
| 145 | 'path', 'pathlength', 'patterncontentunits', 'patterntransform', 'patternunits', | |
| 146 | 'points', 'preservealpha', 'r', 'rx', 'ry', 'radius', 'refx', 'refy', 'repeatcount', | |
| 147 | 'repeatdur', 'restart', 'rotate', 'scale', 'seed', 'shape-rendering', 'specularconstant', | |
| 148 | 'specularexponent', 'spreadmethod', 'stddeviation', 'stitchtiles', 'stop-color', | |
| 149 | 'stop-opacity', 'stroke-dasharray', 'stroke-dashoffset', 'stroke-linecap', | |
| 150 | 'stroke-linejoin', 'stroke-miterlimit', 'stroke-opacity', 'stroke', 'stroke-width', | |
| 151 | 'surfacescale', 'targetx', 'targety', 'transform', 'text-anchor', 'text-decoration', | |
| 847c77 | 152 | 'text-rendering', 'textlength', 'to', 'u1', 'u2', 'unicode', 'values', 'viewbox', |
| 3e4b7c | 153 | 'visibility', 'vert-adv-y', 'version', 'vert-origin-x', 'vert-origin-y', 'word-spacing', |
| AM | 154 | 'wrap', 'writing-mode', 'xchannelselector', 'ychannelselector', 'x', 'x1', 'x2', |
| 155 | 'xmlns', 'y', 'y1', 'y2', 'z', 'zoomandpan', | |
| 156 | // XML | |
| 847c77 | 157 | 'xml:id', 'xlink:title', |
| 7ac944 | 158 | ); |
| AM | 159 | |
| cb3e2f | 160 | /* Elements which could be empty and be returned in short form (<tag />) */ |
| AM | 161 | static $void_elements = array('area', 'base', 'br', 'col', 'command', 'embed', 'hr', |
| 3e4b7c | 162 | 'img', 'input', 'keygen', 'link', 'meta', 'param', 'source', 'track', 'wbr', |
| AM | 163 | // SVG |
| 164 | 'altglyph', 'altglyphdef', 'altglyphitem', 'animate', 'animatecolor', | |
| 165 | 'animatemotion', 'animatetransform', 'circle', 'clippath', 'defs', 'desc', | |
| 166 | 'ellipse', 'font', 'g', 'glyph', 'glyphref', 'hkern', 'image', 'line', | |
| 167 | 'lineargradient', 'marker', 'mask', 'mpath', 'path', 'pattern', | |
| 168 | 'polygon', 'polyline', 'radialgradient', 'rect', 'set', 'stop', 'switch', 'symbol', | |
| 169 | 'text', 'textpath', 'tref', 'tspan', 'use', 'view', 'vkern', 'filter', | |
| 7ac944 | 170 | ); |
| AM | 171 | |
| 172 | /* State for linked objects in HTML */ | |
| 173 | public $extlinks = false; | |
| 174 | ||
| 175 | /* Current settings */ | |
| 176 | private $config = array(); | |
| 177 | ||
| 178 | /* Registered callback functions for tags */ | |
| 179 | private $handlers = array(); | |
| 180 | ||
| 181 | /* Allowed HTML elements */ | |
| 182 | private $_html_elements = array(); | |
| 183 | ||
| 184 | /* Ignore these HTML tags but process their content */ | |
| 185 | private $_ignore_elements = array(); | |
| 186 | ||
| cb3e2f | 187 | /* Elements which could be empty and be returned in short form (<tag />) */ |
| AM | 188 | private $_void_elements = array(); |
| 7ac944 | 189 | |
| AM | 190 | /* Allowed HTML attributes */ |
| 191 | private $_html_attribs = array(); | |
| 192 | ||
| a89940 | 193 | /* Max nesting level */ |
| AM | 194 | private $max_nesting_level; |
| 195 | ||
| 7ac944 | 196 | |
| AM | 197 | /** |
| 198 | * Class constructor | |
| 199 | */ | |
| 200 | public function __construct($p = array()) | |
| 201 | { | |
| 847c77 | 202 | $this->_html_elements = array_flip((array)$p['html_elements']) + array_flip(self::$html_elements); |
| AM | 203 | $this->_html_attribs = array_flip((array)$p['html_attribs']) + array_flip(self::$html_attribs); |
| 204 | $this->_ignore_elements = array_flip((array)$p['ignore_elements']) + array_flip(self::$ignore_elements); | |
| 205 | $this->_void_elements = array_flip((array)$p['void_elements']) + array_flip(self::$void_elements); | |
| 7ac944 | 206 | |
| cb3e2f | 207 | unset($p['html_elements'], $p['html_attribs'], $p['ignore_elements'], $p['void_elements']); |
| 7ac944 | 208 | |
| AM | 209 | $this->config = $p + array('show_washed' => true, 'allow_remote' => false, 'cid_map' => array()); |
| 210 | } | |
| 211 | ||
| 212 | /** | |
| 213 | * Register a callback function for a certain tag | |
| 214 | */ | |
| 215 | public function add_callback($tagName, $callback) | |
| 216 | { | |
| 217 | $this->handlers[$tagName] = $callback; | |
| 218 | } | |
| 219 | ||
| 220 | /** | |
| 221 | * Check CSS style | |
| 222 | */ | |
| 223 | private function wash_style($style) | |
| 224 | { | |
| f96fec | 225 | $result = array(); |
| 7ac944 | 226 | |
| ca7fc7 | 227 | // Remove unwanted white-space characters so regular expressions below work better |
| AM | 228 | $style = preg_replace('/[\n\r\s\t]+/', ' ', $style); |
| 229 | ||
| 7ac944 | 230 | foreach (explode(';', $style) as $declaration) { |
| AM | 231 | if (preg_match('/^\s*([a-z\-]+)\s*:\s*(.*)\s*$/i', $declaration, $match)) { |
| 232 | $cssid = $match[1]; | |
| 233 | $str = $match[2]; | |
| 234 | $value = ''; | |
| 235 | ||
| f96fec | 236 | foreach ($this->explode_style($str) as $val) { |
| AM | 237 | if (preg_match('/^url\(/i', $val)) { |
| 238 | if (preg_match('/^url\(\s*[\'"]?([^\'"\)]*)[\'"]?\s*\)/iu', $val, $match)) { | |
| 239 | $url = $match[1]; | |
| 240 | if (($src = $this->config['cid_map'][$url]) | |
| 241 | || ($src = $this->config['cid_map'][$this->config['base_url'].$url]) | |
| 242 | ) { | |
| 243 | $value .= ' url('.htmlspecialchars($src, ENT_QUOTES) . ')'; | |
| 7ac944 | 244 | } |
| f96fec | 245 | else if (preg_match('!^(https?:)?//[a-z0-9/._+-]+$!i', $url, $m)) { |
| AM | 246 | if ($this->config['allow_remote']) { |
| 247 | $value .= ' url('.htmlspecialchars($m[0], ENT_QUOTES).')'; | |
| 248 | } | |
| 249 | else { | |
| 250 | $this->extlinks = true; | |
| 251 | } | |
| 7ac944 | 252 | } |
| f96fec | 253 | else if (preg_match('/^data:.+/i', $url)) { // RFC2397 |
| AM | 254 | $value .= ' url('.htmlspecialchars($url, ENT_QUOTES).')'; |
| 255 | } | |
| 7ac944 | 256 | } |
| AM | 257 | } |
| f96fec | 258 | else if (!preg_match('/^(behavior|expression)/i', $val)) { |
| 7ac944 | 259 | // whitelist ? |
| f96fec | 260 | $value .= ' ' . $val; |
| 7ac944 | 261 | |
| AM | 262 | // #1488535: Fix size units, so width:800 would be changed to width:800px |
| 5bf83d | 263 | if (preg_match('/^(left|right|top|bottom|width|height)/i', $cssid) |
| f96fec | 264 | && preg_match('/^[0-9]+$/', $val) |
| 7ac944 | 265 | ) { |
| AM | 266 | $value .= 'px'; |
| 267 | } | |
| 268 | } | |
| 269 | } | |
| 270 | ||
| 271 | if (isset($value[0])) { | |
| f96fec | 272 | $result[] = $cssid . ':' . $value; |
| 7ac944 | 273 | } |
| AM | 274 | } |
| 275 | } | |
| 276 | ||
| f96fec | 277 | return implode('; ', $result); |
| 7ac944 | 278 | } |
| AM | 279 | |
| 280 | /** | |
| 281 | * Take a node and return allowed attributes and check values | |
| 282 | */ | |
| 283 | private function wash_attribs($node) | |
| 284 | { | |
| 847c77 | 285 | $result = ''; |
| AM | 286 | $washed = array(); |
| 7ac944 | 287 | |
| 3e4b7c | 288 | foreach ($node->attributes as $name => $attr) { |
| AM | 289 | $key = strtolower($name); |
| 290 | $value = $attr->nodeValue; | |
| 7ac944 | 291 | |
| 847c77 | 292 | if ($key == 'style' && ($style = $this->wash_style($value))) { |
| 786aa0 | 293 | // replace double quotes to prevent syntax error and XSS issues (#1490227) |
| 847c77 | 294 | $result .= ' style="' . str_replace('"', '"', $style) . '"'; |
| 7ac944 | 295 | } |
| 847c77 | 296 | else if (isset($this->_html_attribs[$key])) { |
| AM | 297 | $value = trim($value); |
| 298 | $out = ''; | |
| 299 | ||
| 300 | if ($this->is_image($node->tagName, $key)) { | |
| 301 | if (($src = $this->config['cid_map'][$value]) | |
| 302 | || ($src = $this->config['cid_map'][$this->config['base_url'].$value]) | |
| 303 | ) { | |
| 304 | $out = $src; | |
| 7ac944 | 305 | } |
| 847c77 | 306 | else if (preg_match('/^(http|https|ftp):.+/i', $value)) { |
| AM | 307 | if ($this->config['allow_remote']) { |
| 308 | $out = $value; | |
| 309 | } | |
| 310 | else { | |
| 311 | $this->extlinks = true; | |
| 312 | if ($this->config['blocked_src']) { | |
| 313 | $out = $this->config['blocked_src']; | |
| 314 | } | |
| 7ac944 | 315 | } |
| AM | 316 | } |
| 847c77 | 317 | else if (preg_match('/^data:image.+/i', $value)) { // RFC2397 |
| AM | 318 | $out = $value; |
| 319 | } | |
| 7ac944 | 320 | } |
| 847c77 | 321 | else if ($this->is_link($node->tagName, $key)) { |
| AM | 322 | if (!preg_match('!^(javascript|vbscript|data:text)!i', $value) |
| 323 | && preg_match('!^([a-z][a-z0-9.+-]+:|//|#).+!i', $value) | |
| 324 | ) { | |
| 325 | $out = $value; | |
| 326 | } | |
| 327 | } | |
| 328 | else { | |
| 329 | $out = $value; | |
| 330 | } | |
| 331 | ||
| 332 | if ($out) { | |
| 333 | $result .= ' ' . $key . '="' . htmlspecialchars($out, ENT_QUOTES) . '"'; | |
| 334 | } | |
| 335 | else if ($value) { | |
| 336 | $washed[] = htmlspecialchars($attr->nodeName, ENT_QUOTES); | |
| 7ac944 | 337 | } |
| AM | 338 | } |
| 339 | else { | |
| 847c77 | 340 | $washed[] = htmlspecialchars($attr->nodeName, ENT_QUOTES); |
| 7ac944 | 341 | } |
| AM | 342 | } |
| 343 | ||
| 847c77 | 344 | if (!empty($washed) && $this->config['show_washed']) { |
| AM | 345 | $result .= ' x-washed="' . implode(' ', $washed) . '"'; |
| 346 | } | |
| 347 | ||
| 348 | return $result; | |
| 349 | } | |
| 350 | ||
| 351 | /** | |
| 352 | * Check it the tag/attribute may contain an URI | |
| 353 | */ | |
| 354 | private function is_link($tag, $attr) | |
| 355 | { | |
| 356 | return $tag == 'a' && $attr == 'href'; | |
| 357 | } | |
| 358 | ||
| 359 | /** | |
| 360 | * Check it the tag/attribute may contain an image URI | |
| 361 | */ | |
| 362 | private function is_image($tag, $attr) | |
| 363 | { | |
| 364 | return $attr == 'background' | |
| 365 | || ($attr == 'poster' && $tag == 'video') | |
| 366 | || ($attr == 'src' && preg_match('/^(img|source)$/i', $tag)); | |
| 7ac944 | 367 | } |
| AM | 368 | |
| 369 | /** | |
| 370 | * The main loop that recurse on a node tree. | |
| c77a84 | 371 | * It output only allowed tags with allowed attributes and allowed inline styles |
| AM | 372 | * |
| 373 | * @param DOMNode $node HTML element | |
| 374 | * @param int $level Recurrence level (safe initial value found empirically) | |
| 7ac944 | 375 | */ |
| c77a84 | 376 | private function dumpHtml($node, $level = 20) |
| 7ac944 | 377 | { |
| AM | 378 | if (!$node->hasChildNodes()) { |
| 379 | return ''; | |
| a89940 | 380 | } |
| AM | 381 | |
| 382 | $level++; | |
| 383 | ||
| 384 | if ($this->max_nesting_level > 0 && $level == $this->max_nesting_level - 1) { | |
| 385 | // log error message once | |
| 386 | if (!$this->max_nesting_level_error) { | |
| 387 | $this->max_nesting_level_error = true; | |
| 388 | rcube::raise_error(array('code' => 500, 'type' => 'php', | |
| 389 | 'line' => __LINE__, 'file' => __FILE__, | |
| 390 | 'message' => "Maximum nesting level exceeded (xdebug.max_nesting_level={$this->max_nesting_level})"), | |
| 391 | true, false); | |
| 392 | } | |
| 393 | return '<!-- ignored -->'; | |
| 7ac944 | 394 | } |
| AM | 395 | |
| 396 | $node = $node->firstChild; | |
| 397 | $dump = ''; | |
| 398 | ||
| 399 | do { | |
| c7c09f | 400 | switch ($node->nodeType) { |
| 7ac944 | 401 | case XML_ELEMENT_NODE: //Check element |
| AM | 402 | $tagName = strtolower($node->tagName); |
| 403 | if ($callback = $this->handlers[$tagName]) { | |
| 404 | $dump .= call_user_func($callback, $tagName, | |
| a89940 | 405 | $this->wash_attribs($node), $this->dumpHtml($node, $level), $this); |
| 7ac944 | 406 | } |
| AM | 407 | else if (isset($this->_html_elements[$tagName])) { |
| a89940 | 408 | $content = $this->dumpHtml($node, $level); |
| 3e4b7c | 409 | $dump .= '<' . $tagName; |
| AM | 410 | |
| 411 | if ($tagName == 'svg') { | |
| 412 | $xpath = new DOMXPath($node->ownerDocument); | |
| 413 | foreach ($xpath->query('namespace::*') as $ns) { | |
| 414 | if ($ns->nodeName != 'xmlns:xml') { | |
| 415 | $dump .= ' ' . $ns->nodeName . '="' . $ns->nodeValue . '"'; | |
| 416 | } | |
| 417 | } | |
| 418 | } | |
| 419 | ||
| 420 | $dump .= $this->wash_attribs($node); | |
| 421 | ||
| 422 | if ($content === '' && isset($this->_void_elements[$tagName])) { | |
| 423 | $dump .= ' />'; | |
| 424 | } | |
| 425 | else { | |
| 426 | $dump .= ">$content</$tagName>"; | |
| 427 | } | |
| 7ac944 | 428 | } |
| AM | 429 | else if (isset($this->_ignore_elements[$tagName])) { |
| 430 | $dump .= '<!-- ' . htmlspecialchars($tagName, ENT_QUOTES) . ' not allowed -->'; | |
| 431 | } | |
| 432 | else { | |
| 433 | $dump .= '<!-- ' . htmlspecialchars($tagName, ENT_QUOTES) . ' ignored -->'; | |
| a89940 | 434 | $dump .= $this->dumpHtml($node, $level); // ignore tags not its content |
| 7ac944 | 435 | } |
| AM | 436 | break; |
| 437 | ||
| 438 | case XML_CDATA_SECTION_NODE: | |
| 439 | $dump .= $node->nodeValue; | |
| 440 | break; | |
| 441 | ||
| 442 | case XML_TEXT_NODE: | |
| 443 | $dump .= htmlspecialchars($node->nodeValue); | |
| 444 | break; | |
| 445 | ||
| 446 | case XML_HTML_DOCUMENT_NODE: | |
| a89940 | 447 | $dump .= $this->dumpHtml($node, $level); |
| 7ac944 | 448 | break; |
| AM | 449 | } |
| c7c09f | 450 | } |
| AM | 451 | while($node = $node->nextSibling); |
| 7ac944 | 452 | |
| AM | 453 | return $dump; |
| 454 | } | |
| 455 | ||
| 456 | /** | |
| 457 | * Main function, give it untrusted HTML, tell it if you allow loading | |
| 458 | * remote images and give it a map to convert "cid:" urls. | |
| 459 | */ | |
| 460 | public function wash($html) | |
| 461 | { | |
| 462 | // Charset seems to be ignored (probably if defined in the HTML document) | |
| 463 | $node = new DOMDocument('1.0', $this->config['charset']); | |
| 464 | $this->extlinks = false; | |
| 465 | ||
| 466 | $html = $this->cleanup($html); | |
| 467 | ||
| 468 | // Find base URL for images | |
| 469 | if (preg_match('/<base\s+href=[\'"]*([^\'"]+)/is', $html, $matches)) { | |
| 470 | $this->config['base_url'] = $matches[1]; | |
| 471 | } | |
| 472 | else { | |
| 473 | $this->config['base_url'] = ''; | |
| 474 | } | |
| a89940 | 475 | |
| AM | 476 | // Detect max nesting level (for dumpHTML) (#1489110) |
| 477 | $this->max_nesting_level = (int) @ini_get('xdebug.max_nesting_level'); | |
| 7ac944 | 478 | |
| 3e4b7c | 479 | // SVG need to be parsed as XML |
| AM | 480 | $xml = stripos($html, '<svg') !== false || stripos($html, '<?xml') !== false; |
| 481 | $method = $xml ? 'loadXML' : 'loadHTML'; | |
| 482 | $options = 0; | |
| 483 | ||
| bfd24f | 484 | // Use optimizations if supported |
| 75bbad | 485 | if (PHP_VERSION_ID >= 50400) { |
| 3e4b7c | 486 | $options = LIBXML_PARSEHUGE | LIBXML_COMPACT | LIBXML_NONET; |
| 847c77 | 487 | @$node->{$method}($html, $options); |
| bfd24f | 488 | } |
| 847c77 | 489 | else { |
| AM | 490 | @$node->{$method}($html); |
| 491 | } | |
| bfd24f | 492 | |
| 7ac944 | 493 | return $this->dumpHtml($node); |
| AM | 494 | } |
| 495 | ||
| 496 | /** | |
| 497 | * Getter for config parameters | |
| 498 | */ | |
| 499 | public function get_config($prop) | |
| 500 | { | |
| 501 | return $this->config[$prop]; | |
| 502 | } | |
| 503 | ||
| 504 | /** | |
| 505 | * Clean HTML input | |
| 506 | */ | |
| 507 | private function cleanup($html) | |
| 508 | { | |
| 3e4b7c | 509 | $html = trim($html); |
| AM | 510 | |
| 7ac944 | 511 | // special replacements (not properly handled by washtml class) |
| AM | 512 | $html_search = array( |
| b7048d | 513 | // space(s) between <NOBR> |
| AM | 514 | '/(<\/nobr>)(\s+)(<nobr>)/i', |
| 515 | // PHP bug #32547 workaround: remove title tag | |
| 516 | '/<title[^>]*>[^<]*<\/title>/i', | |
| 517 | // remove <!doctype> before BOM (#1490291) | |
| 518 | '/<\!doctype[^>]+>[^<]*/im', | |
| 519 | // byte-order mark (only outlook?) | |
| 520 | '/^(\0\0\xFE\xFF|\xFF\xFE\0\0|\xFE\xFF|\xFF\xFE|\xEF\xBB\xBF)/', | |
| 521 | // washtml/DOMDocument cannot handle xml namespaces | |
| 522 | '/<html\s[^>]+>/i', | |
| 7ac944 | 523 | ); |
| AM | 524 | |
| 525 | $html_replace = array( | |
| 526 | '\\1'.' '.'\\3', | |
| 527 | '', | |
| 528 | '', | |
| b7048d | 529 | '', |
| 7ac944 | 530 | '<html>', |
| AM | 531 | ); |
| 3e4b7c | 532 | |
| 7ac944 | 533 | $html = preg_replace($html_search, $html_replace, trim($html)); |
| AM | 534 | |
| 3e4b7c | 535 | // Replace all of those weird MS Word quotes and other high characters |
| c72507 | 536 | $badwordchars = array( |
| b6a640 | 537 | "\xe2\x80\x98", // left single quote |
| R | 538 | "\xe2\x80\x99", // right single quote |
| 539 | "\xe2\x80\x9c", // left double quote | |
| 540 | "\xe2\x80\x9d", // right double quote | |
| 541 | "\xe2\x80\x94", // em dash | |
| 3e4b7c | 542 | "\xe2\x80\xa6" // elipses |
| b6a640 | 543 | ); |
| 3e4b7c | 544 | |
| c72507 | 545 | $fixedwordchars = array( |
| b6a640 | 546 | "'", |
| R | 547 | "'", |
| 548 | '"', | |
| 549 | '"', | |
| 550 | '—', | |
| 551 | '...' | |
| 552 | ); | |
| 3e4b7c | 553 | |
| c72507 | 554 | $html = str_replace($badwordchars, $fixedwordchars, $html); |
| b6a640 | 555 | |
| 7ac944 | 556 | // PCRE errors handling (#1486856), should we use something like for every preg_* use? |
| AM | 557 | if ($html === null && ($preg_error = preg_last_error()) != PREG_NO_ERROR) { |
| 558 | $errstr = "Could not clean up HTML message! PCRE Error: $preg_error."; | |
| 559 | ||
| 560 | if ($preg_error == PREG_BACKTRACK_LIMIT_ERROR) { | |
| 561 | $errstr .= " Consider raising pcre.backtrack_limit!"; | |
| 562 | } | |
| 563 | if ($preg_error == PREG_RECURSION_LIMIT_ERROR) { | |
| 564 | $errstr .= " Consider raising pcre.recursion_limit!"; | |
| 565 | } | |
| 566 | ||
| 567 | rcube::raise_error(array('code' => 620, 'type' => 'php', | |
| 568 | 'line' => __LINE__, 'file' => __FILE__, | |
| 569 | 'message' => $errstr), true, false); | |
| a89940 | 570 | |
| 7ac944 | 571 | return ''; |
| AM | 572 | } |
| 573 | ||
| 574 | // fix (unknown/malformed) HTML tags before "wash" | |
| ffec85 | 575 | $html = preg_replace_callback('/(<(?!\!)[\/]*)([^\s>]+)([^>]*)/', array($this, 'html_tag_callback'), $html); |
| 7ac944 | 576 | |
| AM | 577 | // Remove invalid HTML comments (#1487759) |
| 578 | // Don't remove valid conditional comments | |
| 1bce14 | 579 | // Don't remove MSOutlook (<!-->) conditional comments (#1489004) |
| 82ed25 | 580 | $html = preg_replace('/<!--[^-<>\[\n]+>/', '', $html); |
| c72507 | 581 | |
| AM | 582 | // fix broken nested lists |
| 583 | self::fix_broken_lists($html); | |
| 7ac944 | 584 | |
| AM | 585 | // turn relative into absolute urls |
| 586 | $html = self::resolve_base($html); | |
| 587 | ||
| 588 | return $html; | |
| 589 | } | |
| 590 | ||
| 591 | /** | |
| 592 | * Callback function for HTML tags fixing | |
| 593 | */ | |
| 594 | public static function html_tag_callback($matches) | |
| 595 | { | |
| 596 | $tagname = $matches[2]; | |
| 597 | $tagname = preg_replace(array( | |
| 598 | '/:.*$/', // Microsoft's Smart Tags <st1:xxxx> | |
| 3e4b7c | 599 | '/[^a-z0-9_\[\]\!?-]/i', // forbidden characters |
| 7ac944 | 600 | ), '', $tagname); |
| AM | 601 | |
| ffec85 | 602 | // fix invalid closing tags - remove any attributes (#1489446) |
| AM | 603 | if ($matches[1] == '</') { |
| 604 | $matches[3] = ''; | |
| 605 | } | |
| 606 | ||
| 607 | return $matches[1] . $tagname . $matches[3]; | |
| 7ac944 | 608 | } |
| AM | 609 | |
| 610 | /** | |
| 611 | * Convert all relative URLs according to a <base> in HTML | |
| 612 | */ | |
| 613 | public static function resolve_base($body) | |
| 614 | { | |
| 615 | // check for <base href=...> | |
| 616 | if (preg_match('!(<base.*href=["\']?)([hftps]{3,5}://[a-z0-9/.%-]+)!i', $body, $regs)) { | |
| 617 | $replacer = new rcube_base_replacer($regs[2]); | |
| 618 | $body = $replacer->replace($body); | |
| 619 | } | |
| 620 | ||
| 621 | return $body; | |
| 622 | } | |
| 623 | ||
| c72507 | 624 | /** |
| AM | 625 | * Fix broken nested lists, they are not handled properly by DOMDocument (#1488768) |
| 626 | */ | |
| 627 | public static function fix_broken_lists(&$html) | |
| 628 | { | |
| 629 | // do two rounds, one for <ol>, one for <ul> | |
| 630 | foreach (array('ol', 'ul') as $tag) { | |
| 631 | $pos = 0; | |
| 632 | while (($pos = stripos($html, '<' . $tag, $pos)) !== false) { | |
| 633 | $pos++; | |
| 634 | ||
| 635 | // make sure this is an ol/ul tag | |
| 636 | if (!in_array($html[$pos+2], array(' ', '>'))) { | |
| 637 | continue; | |
| 638 | } | |
| 639 | ||
| 640 | $p = $pos; | |
| 641 | $in_li = false; | |
| 642 | $li_pos = 0; | |
| 643 | ||
| 644 | while (($p = strpos($html, '<', $p)) !== false) { | |
| 645 | $tt = strtolower(substr($html, $p, 4)); | |
| 646 | ||
| 647 | // li open tag | |
| 648 | if ($tt == '<li>' || $tt == '<li ') { | |
| 649 | $in_li = true; | |
| 650 | $p += 4; | |
| 651 | } | |
| 652 | // li close tag | |
| 653 | else if ($tt == '</li' && in_array($html[$p+4], array(' ', '>'))) { | |
| 654 | $li_pos = $p; | |
| 655 | $p += 4; | |
| 656 | $in_li = false; | |
| 657 | } | |
| 658 | // ul/ol closing tag | |
| 659 | else if ($tt == '</' . $tag && in_array($html[$p+4], array(' ', '>'))) { | |
| 660 | break; | |
| 661 | } | |
| 662 | // nested ol/ul element out of li | |
| 663 | else if (!$in_li && $li_pos && ($tt == '<ol>' || $tt == '<ol ' || $tt == '<ul>' || $tt == '<ul ')) { | |
| 664 | // find closing tag of this ul/ol element | |
| 665 | $element = substr($tt, 1, 2); | |
| 666 | $cpos = $p; | |
| 667 | do { | |
| 668 | $tpos = stripos($html, '<' . $element, $cpos+1); | |
| 669 | $cpos = stripos($html, '</' . $element, $cpos+1); | |
| 670 | } | |
| 671 | while ($tpos !== false && $cpos !== false && $cpos > $tpos); | |
| 672 | ||
| 673 | // not found, this is invalid HTML, skip it | |
| 674 | if ($cpos === false) { | |
| 675 | break; | |
| 676 | } | |
| 677 | ||
| 678 | // get element content | |
| 679 | $end = strpos($html, '>', $cpos); | |
| 680 | $len = $end - $p + 1; | |
| 681 | $element = substr($html, $p, $len); | |
| 682 | ||
| 683 | // move element to the end of the last li | |
| 684 | $html = substr_replace($html, '', $p, $len); | |
| 685 | $html = substr_replace($html, $element, $li_pos, 0); | |
| 686 | ||
| 687 | $p = $end; | |
| 688 | } | |
| 689 | else { | |
| 690 | $p++; | |
| 691 | } | |
| 692 | } | |
| 693 | } | |
| 694 | } | |
| 695 | } | |
| f96fec | 696 | |
| AM | 697 | /** |
| 698 | * Explode css style value | |
| 699 | */ | |
| 700 | protected function explode_style($style) | |
| 701 | { | |
| 702 | $style = trim($style); | |
| 703 | ||
| 704 | // first remove comments | |
| 705 | $pos = 0; | |
| 706 | while (($pos = strpos($style, '/*', $pos)) !== false) { | |
| 707 | $end = strpos($style, '*/', $pos+2); | |
| 708 | ||
| 709 | if ($end === false) { | |
| 710 | $style = substr($style, 0, $pos); | |
| 711 | } | |
| 712 | else { | |
| 713 | $style = substr_replace($style, '', $pos, $end - $pos + 2); | |
| 714 | } | |
| 715 | } | |
| 716 | ||
| 717 | $strlen = strlen($style); | |
| 718 | $result = array(); | |
| 719 | ||
| 720 | // explode value | |
| 721 | for ($p=$i=0; $i < $strlen; $i++) { | |
| 722 | if (($style[$i] == "\"" || $style[$i] == "'") && $style[$i-1] != "\\") { | |
| 723 | if ($q == $style[$i]) { | |
| 724 | $q = false; | |
| 725 | } | |
| 726 | else if (!$q) { | |
| 727 | $q = $style[$i]; | |
| 728 | } | |
| 729 | } | |
| 730 | ||
| 731 | if (!$q && $style[$i] == ' ' && !preg_match('/[,\(]/', $style[$i-1])) { | |
| 732 | $result[] = substr($style, $p, $i - $p); | |
| 733 | $p = $i + 1; | |
| 734 | } | |
| 735 | } | |
| 736 | ||
| 737 | $result[] = (string) substr($style, $p); | |
| 738 | ||
| 739 | return $result; | |
| 740 | } | |
| c72507 | 741 | } |
