| blob | 20c71e118e62129c4b6dbea8cbc12d62f1915c78 |
4 <head>
7 <meta name="keywords" content="HTMLPurifier, HTML Purifier, HTML, filter, filtering, HTML_Safe, PEAR, comparison, kses, striptags, SafeHTMLChecker" />
12 <!--[if lt IE 7.]><script defer="defer" type="text/javascript" src="./pngfix.js"></script><![endif]-->
14 </head>
15 <body>
20 <div id="header"><a href="./"><span class="html">HTML</span> <span class="purifier">Purifier</span></a></div>
26 gone from passive consumer to active producer of content on the World Wide
30 put the user in control.</p>
32 <p>Give the user too much control, however, and you set yourself up
36 has proven to be both a blessing and a curse, and the software that processes
37 it must strike a fine balance between security and usability. How do
38 we prevent users from injecting JavaScript or inserting malformed
40 a rich syntax of tags, attributes and <acronym
44 about sloppy coding messing up <acronym
47 developer has come across this problem before, and many have tried
48 (albeit unsuccessfully) to solve this problem. We will analyze existing
49 libraries to demonstrate how they are ineffective and, of course,
50 how HTML Purifier solves all our problems and achieves world peace.</p>
52 <p>I will take no quarter and pull no punches: as of the time of writing,
54 for richly formatted documents. However, it's important to note HTML
56 There are cases in which HTML Purifier is overkill, in such situations,
57 use the tool best suited for the job.</p>
61 generated. Please enable JavaScript to see it.</p>
62 </noscript></div>
71 <thead>
72 <tr>
84 </tr>
85 </thead>
87 <tbody>
89 <tr>
101 </tr>
103 <tr>
115 </tr>
117 <tr>
129 </tr>
131 <tr>
143 </tr>
145 <tr>
157 </tr>
159 <tr>
171 </tr>
173 </tbody>
175 </table>
176 </div>
184 </blockquote>
188 they are the number one method of taking user input and processing it into
189 something more than plain old text. These libraries forgo
195 such markup languages (although it should be noted that
196 Wikitext and Markdown can allow
198 The benefits (to those who use it, anyway) are clear: simplicity and
199 security.
200 </p>
203 <thead>
204 <tr>
207 </tr>
208 </thead>
209 <tbody>
210 <tr>
213 </tr>
214 <tr>
217 </tr>
218 <tr>
221 </tr>
222 <tr>
225 </tr>
226 <tr>
228 <td><tt><b>B</b> <i>i</i> <a href="http://www.example.com/">link</a></tt></td>
229 </tr>
230 <tr>
233 </tr>
234 </tbody>
235 </table>
240 There are many variants of Wikitext currently extant.</li>
241 <li>Strictly speaking, the Markdown syntax is not equivalent: bold text
244 however, map those two semantic tags to the associated styling, so
245 many users assume that it really is italics (and use it improperly for,
246 say, book titles.)</li>
247 </ol>
252 source code is often criticized for being difficult to read. For example,
253 compare:</p>
255 <pre>
256 * Item 1
257 * Item 2
258 </pre>
262 <pre>
267 </pre>
269 <p>Which would you prefer to edit? The answer seems obvious, but be careful
270 not to fall into the fallacy of <a
274 editor, which blows earlier choices out of the water in terms
275 of usability.</p>
277 <p>Note that rich text editors and alternate markup syntaxes are not
278 mutually exclusive, but, when push comes to shove, it's easier
279 implement this sort of editor on top of <acronym
281 markup language. And in the cases when it is done, you usually end up with
282 a live preview, not a true rich text editor.</p>
287 editors aren't all that great." There are many good arguments
288 against these editors, and <a
290 people have written essays</a> devoted to
292 In addition to the usual arguments against said editors, the web poses
293 another limitation: no JavaScript means no
294 editor, and no editor means... (gasp) manually typing in code.</p>
296 <p>Even the most dogmatic purist, however, should recognize that for all
298 There are steps you can take to mitigate the associated drawbacks of
299 these editors.</p>
301 <p>It is often asserted that
304 this is the case with any markup language that allows the smallest
307 A good way to reduce this trouble is to simply eliminate the
308 dialogue boxes that allow users to change colors or fonts (which
309 usually have no legitimate use) and adopt a
311 allowing users to select contextually correct formatting styles
312 for segments of text.</p>
313 </blockquote>
315 <p>Simplicity is also a double-edged sword. The moment any remotely
316 complex markup is needed, these lightweight markup languages fail to
317 produce. Sure you can make '''this text bold''' with Wikitext, but that
320 These languages face the same troubles as regular <acronym
322 whitelist is too restrictive (besides the fact that their table markup
323 is extraordinarily complex).</p>
329 the angled brackets with square brackets and omitting the occasional parameter
330 name? How much more un-original can you get? Somehow, I don't think BBCode
331 was meant to readable. <a
334 <blockquote>
335 BBCode was devised and put to use in order to provide a safer, easier
336 and more limited way of allowing users to format their messages.
337 Previously, many message boards allowed the users to include HTML,
338 which could be used to break/imitate parts of the layout, or run
339 JavaScript. Some implementations of BBCode have suffered problems related
340 to the way they translate the BBCode into HTML, which could negate the
341 security that was intended to be given by BBCode.
342 </blockquote>
346 <blockquote>
347 BBCode came to life when developers where too lazy to parse HTML correctly
348 and decided to invent their own markup language. As with all products of
349 laziness, the result is completely inconsistent, unstandardized, and
350 widely adopted.
351 </blockquote>
353 <p>Well, developers, the whole point of HTML Purifier is that I do the
354 work so you can just execute the ridiculously simple
360 <p>These alternative markup languages have their shiny points, and HTML
361 Purifier is not meant to replace them. However, a major reason for
363 using these languages?</p>
369 neat enough, at least, to make it into PHP as a
371 The premise is simple, the execution effective. Tidy is, in short, a great
374 <p>It is not, however, a filter. I am often surprised when people ask
379 Tidy reads HTML, XHTML and XML files and writes cleaned up markup. For
380 HTML variants, it detects and corrects many common coding errors and
381 strives to produce visually equivalent markup that is both W3C compliant
382 and works on most browsers. A common use of Tidy is to convert plain HTML
383 to XHTML.
384 </blockquote>
393 <p>This is not to say that I haven't seen Tidy be used in this sort of
394 fashion. MediaWiki, for instance, uses Tidy to cleanup the final HTML
395 output before shuttling it off to the browser. The developers, nevertheless,
396 agree that this is only a band-aid solution, and that the real way
397 to fix it is to fix the parser. Tidy's great, but in terms of security,
398 it's unsuited for untrusted sources.</p>
402 <p>I've ordered my analyses according to how bad a library is. The worst
403 is first, and then we move up the spectrum. I will point out the most
404 flagrant problems with the libraries, but note that I will omit more
406 attribute, I really shouldn't reprimand you for letting non-SGML code
407 points through. The ideal solution, however, must do all these things.</p>
417 </table>
421 the classic solution for attempting to clean up
424 The fact that it doesn't validate attributes at all means that anyone can
426 this can be bandaided with a series of regular expressions that strip out
427 on[event] (you're still vulnerable to <acronym
429 quirky browser behavior), striptags() is fundamentally flawed and should not be
430 used.
431 </p>
437 is a souped up version of striptags() with the ability to inspect
438 attributes. (Don't mind the hastily tacked on query escaping function).</p>
451 </table>
453 <p>PHP Input Filter implements an
455 performs very basic checks on whether or not tags and attributes have
456 been defined in the whitelist as well as some smarter <acronym
458 the user to define what they'll permit.</p>
460 <p>With absolutely no checking of well-formedness, it is trivially easy
461 to trick the filter into leaving unclosed tags lying around. While
463 basic sanity checks like this must be implemented.</p>
465 <p>More troubles: Woe to
467 just let <acronym
469 layout not to be badly mutilated. To top things off,
470 the filter doesn't even preserve data properly: attributes have all
471 spaces stripped out of them. Stay away, stay away!</p>
478 filtering library.
479 It should be noted that this is the same library as
481 branding (and a different version number).</p>
494 </table>
496 <p>HTML_Safe's mechanism of action involves parsing
499 validation and filtering as the handlers are called. HTML_Safe does a lot
502 is fundamentally flawed: blacklists.</p>
504 <p>This library maintains arrays of dangerous tags, attributes and
506 has a blacklist of dangerous <acronym
508 intelligently disabled by default in favor of a protocol whitelist.)
509 What this means is that HTML_Safe has no qualms of accepting input
511 the tags in those arrays. Scratch standards-compliance (and that was
512 without even considering proper nesting).</p>
514 <p>For now, HTML_Safe might be safe from
516 In the future, however, one of the infinitely many tags that HTML_Safe lets
517 through might just possibly be given special functionality by browser vendors.
519 solution puts you at a perpetual arms race against crackers who are constantly
520 discovering new and inventive ways to abuse tags and attributes that you
521 didn't blacklist.</p>
526 be the de-facto solution for cleaning
542 </table>
544 <p>To be truthful, I didn't do as comprehensive a code survey for kses
545 as I did for some of the other libraries. Out of
546 all the classes I've reviewed so far, kses was definitely the hardest to
547 understand.</p>
549 <p>kses's modus operandi is splitting up html with a monster regexp
551 suffers from the same problems as Input Filter: no well-formedness
552 checks leading to rampant runaway tags (and no standards-compliance).</p>
554 <p>Its whitelist syntax, however, is the most complex of all these libraries,
555 so I'm going to take some time to argue why this particular implementation
556 is bad. The author of this library was thoughtful enough to provide some
557 basic constraint checks on attributes like maxlen and maxval. Now, barring
558 the fact that there simply aren't enough checks, and the fact that they are
559 all lumped together in one function, we now must wonder whether or not
560 the user will go through the trouble of specifying the maximum length
561 of a title attribute.</p>
563 <p>I have my opinions about inherent human laziness, but perhaps WordPress's
564 default filterset is the most telling example:</p>
566 <pre>
567 $allowedposttags = array (
568 /* formatted and trimmed */
569 'hr' => array (
570 'align' => array (),
571 'noshade' => array (),
572 'size' => array (),
573 'width' => array ()
574 )
575 );
576 </pre>
578 <p>Hmm... do I see a blatant lack of attribute constraints? Conclusion:
579 if the user can get away with not doing work, they will! The biggest
581 the whitelist. The whitelist is just as important as the code that uses
582 the whitelist to filter
589 HTML Checker</a> is (to my knowledge) the first attempt to make a filter
590 that also outputs standards-compliant XHTML. It wasn't even released or
592 search result must have done something right.</p>
605 </table>
607 <p>Indeed, it is quite a well-written piece of code. It demonstrates
608 knowledge of inline versus block elements, thus almost nearly getting
609 nesting correct (the only exception is an unimplemented omitted SGML
612 <p>Unfortunately, part of the reason why it works so well is that it's
613 extremely restrictive. No styling, no tables, very few attributes.
614 Perfectly appropriate for blog comments, but then again, there's always
615 BBCode. This probably means that Safe HTML Checker has a different
616 goal than HTML Purifier.</p>
619 is also quite strict. Accidentally missed a < sign? The parser will
620 complain with the cryptic message:
622 is not well-formed".
623 The solution is not as simple as just switching to a more permissive
624 parser: Safe HTML Checker relies on the fact that the parser will have
625 matched up the tags for them.</p>
640 </table>
650 </table>
652 <p>This is not to say that HTML Purifier doesn't have problems of its own.
653 It's a fairly nascent library (so there's bound to bugs), it's big
654 (while the others usually fit in one file, this one requires a huge
656 features.</a> A big thing I would like to see added to this library is
657 multiple levels of filtering: from a super-permissive lint mode
658 to a super-restrictive blog comment mode. But even in its current state,
659 HTML Purifier is far better than the other libraries.</p>
663 </div>
664 </body>
665 </html>