1 NEWS ( CHANGELOG and HISTORY ) HTMLPurifier
2 |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
4 = KEY ====================
10 ==========================
12 4.0.0, unknown release date
13 # APIs for ConfigSchema subsystem have substantially changed. See
14 docs/dev-config-bcbreaks.txt for details; in essence, anything that
15 had both namespace and directive now have a single unified key.
16 # Some configuration directives were renamed, specifically:
17 %AutoFormatParam.PurifierLinkifyDocURL -> %AutoFormat.PurifierLinkify.DocURL
18 %FilterParam.ExtractStyleBlocksEscaping -> %Filter.ExtractStyleBlocks.Escaping
19 %FilterParam.ExtractStyleBlocksScope -> %Filter.ExtractStyleBlocks.Scope
20 %FilterParam.ExtractStyleBlocksTidyImpl -> %Filter.ExtractStyleBlocks.TidyImpl
21 As usual, the old directive names will still work, but will throw E_NOTICE
23 # The allowed values for class have been relaxed to allow all of CDATA for
24 doctypes that are not XHTML 1.1 or XHTML 2.0. For old behavior, set
25 %Attr.ClassUseCDATA to false.
26 ! More robust support for name="" and id=""
27 ! HTMLPurifier_Config::inherit($config) allows you to inherit one
28 configuration, and have changes to that configuration be propagated
29 to all of its children.
30 ! Implement %HTML.Attr.Name.UseCDATA, which relaxes validation rules on
31 the name attribute when set. Use with care. Thanks Ian Cook for
33 ! Implement %AutoFormat.RemoveEmpty.RemoveNbsp, which removes empty
34 tags that contain non-breaking spaces as well other whitespace. You
35 can also modify which tags should have maintained with
36 %AutoFormat.RemoveEmpty.RemoveNbsp.Exceptions.
37 ! Implement %Attr.AllowedClasses, which allows administrators to restrict
38 classes users can use to a specified finite set of classes, and
39 %Attr.ForbiddenClasses, which is the logical inverse.
40 - Fix bug where URIDefinition would not get cleared if it's directives got
42 . Created script maintenance/rename-config.php for renaming a configuration
43 directive while maintaining its alias. This script does not change source code.
44 . Implement namespace locking for definition construction, to prevent
45 bugs where a directive is used for definition construction but is not
46 used to construct the cache hash.
48 3.3.0, released 2009-02-16
49 ! Implement CSS property 'overflow' when %CSS.AllowTricky is true.
50 ! Implement generic property list classess
51 - Fix bug with testEncodingSupportsASCII() algorithm when iconv() implementation
52 does not do the "right thing" with characters not supported in the output
54 - Spellcheck UTF-8: The Secret To Character Encoding
55 - Fix improper removal of the contents of elements with only whitespace. Thanks
56 Eric Wald for reporting.
57 - Fix broken test suite in versions of PHP without spl_autoload_register()
58 - Fix degenerate case with YouTube filter involving double hyphens.
59 Thanks Pierre Attar for reporting.
60 - Fix YouTube rendering problem on certain versions of Firefox.
61 - Fix CSSDefinition Printer problems with decorators
62 - Add text parameter to unit tests, forces text output
63 . Add verbose mode to command line test runner, use (--verbose)
64 . Turn on unit tests for UnitConverter
65 . Fix missing version number in configuration %Attr.DefaultImageAlt (added 3.2.0)
66 . Fix newline errors that caused spurious failures when CRLF HTML Purifier was
68 . Removed trailing whitespace from all text files, see
69 remote-trailing-whitespace.php maintenance script.
70 . Convert configuration to use property list backend.
72 3.2.0, released 2008-10-31
73 # Using %Core.CollectErrors forces line number/column tracking on, whereas
74 previously you could theoretically turn it off.
75 # HTMLPurifier_Injector->notifyEnd() is formally deprecated. Please
76 use handleEnd() instead.
77 ! %Output.AttrSort for when you need your attributes in alphabetical order to
78 deal with a bug in FCKEditor. Requested by frank farmer.
79 ! Enable HTML comments when %HTML.Trusted is on. Requested by Waldo Jaquith.
80 ! Proper support for name attribute. It is now allowed and equivalent to the id
81 attribute in a and img tags, and is only converted to id when %HTML.TidyLevel
82 is heavy (for all doctypes).
83 ! %AutoFormat.RemoveEmpty to remove some empty tags from documents. Please don't
84 use on hand-written HTML.
85 ! Add error-cases for unsupported elements in MakeWellFormed. This enables
86 the strategy to be used, standalone, on untrusted input.
87 ! %Core.AggressivelyFixLt is on by default. This causes more sensible
88 processing of left angled brackets in smileys and other whatnot.
89 ! Test scripts now have a 'type' parameter, which lets you say 'htmlpurifier',
90 'phpt', 'vtest', etc. in order to only execute those tests. This supercedes
91 the --only-phpt parameter, although for backwards-compatibility the flag
93 ! AutoParagraph auto-formatter will now preserve double-newlines upon output.
94 Users who are not performing inbound filtering, this may seem a little
95 useless, but as a bonus, the test suite and handling of edge cases is also
97 ! Experimental implementation of forms for %HTML.Trusted
98 ! Track column numbers when maintain line numbers is on
99 ! Proprietary 'background' attribute on table-related elements converted into
100 corresponding CSS. Thanks Fusemail for sponsoring this feature!
101 ! Add forward(), forwardUntilEndToken(), backward() and current() to Injector
103 ! HTMLPurifier_Injector->handleEnd() permits modification to end tokens. The
104 time of operation varies slightly from notifyEnd() as *all* end tokens are
105 processed by the injector before they are subject to the well-formedness rules.
106 ! %Attr.DefaultImageAlt allows overriding default behavior of setting alt to
107 basename of image when not present.
108 ! %AutoFormat.DisplayLinkURI neuters <a> tags into plain text URLs.
109 - Fix two bugs in %URI.MakeAbsolute; one involving empty paths in base URLs,
110 the other involving an undefined $is_folder error.
111 - Throw error when %Core.Encoding is set to a spurious value. Previously,
112 this errored silently and returned false.
113 - Redirected stderr to stdout for flush error output.
114 - %URI.DisableExternal will now use the host in %URI.Base if %URI.Host is not
116 - Do not re-munge URL if the output URL has the same host as the input URL.
118 - Fix error in documentation regarding %Filter.ExtractStyleBlocks
119 - Prevent <![CDATA[<body></body>]]> from triggering %Core.ConvertDocumentToFragment
120 - Fix bug with inline elements in blockquotes conflicting with strict doctype
121 - Detect if HTML support is disabled for DOM by checking for loadHTML() method.
122 - Fix bug where dots and double-dots in absolute URLs without hostname were
123 not collapsed by URIFilter_MakeAbsolute.
124 - Fix bug with anonymous modules operating on SafeEmbed or SafeObject elements
125 by reordering their addition.
126 - Will now throw exception on many error conditions during lexer creation; also
127 throw an exception when MaintainLineNumbers is true, but a non-tracksLineNumbers
129 - Detect if domxml extension is loaded, and use DirectLEx accordingly.
130 - Improve handling of big numbers with floating point arithmetic in UnitConverter.
131 Reported by David Morton.
132 . Strategy_MakeWellFormed now operates in-place, saving memory and allowing
133 for more interesting filter-backtracking
134 . New HTMLPurifier_Injector->rewind() functionality, allows injectors to rewind
135 index to reprocess tokens.
136 . StringHashParser now allows for multiline sections with "empty" content;
137 previously the section would remain undefined.
138 . Added --quick option to multitest.php, which tests only the most recent
139 release for each series.
140 . Added --distro option to multitest.php, which accepts either 'normal' or
141 'standalone'. This supercedes --exclude-normal and --exclude-standalone
143 3.1.1, released 2008-06-19
144 # %URI.Munge now, by default, does not munge resources (for example, <img src="">)
145 In order to enable this again, please set %URI.MungeResources to true.
146 ! More robust imagecrash protection with height/width CSS with %CSS.MaxImgLength,
147 and height/width HTML with %HTML.MaxImgLength.
148 ! %URI.MungeSecretKey for secure URI munging. Thanks Chris
149 for sponsoring this feature. Check out the corresponding documentation
150 for details. (Att Nightly testers: The API for this feature changed before
151 the general release. Namely, rename your directives %URI.SecureMungeSecretKey =>
152 %URI.MungeSecretKey and and %URI.SecureMunge => %URI.Munge)
153 ! Implemented post URI filtering. Set member variable $post to true to set
155 ! Allow modules to define injectors via $info_injector. Injectors are
156 automatically disabled if injector's needed elements are not found.
157 ! Support for "safe" objects added, use %HTML.SafeObject and %HTML.SafeEmbed.
158 Thanks Chris for sponsoring. If you've been using ad hoc code from the
159 forums, PLEASE use this instead.
160 ! Added substitutions for %e, %n, %a and %p in %URI.Munge (in order,
161 embedded, tag name, attribute name, CSS property name). See %URI.Munge
162 for more details. Requested by Jochem Blok.
163 - Disable percent height/width attributes for img.
164 - AttrValidator operations are now atomic; updates to attributes are not
165 manifest in token until end of operations. This prevents naughty internal
166 code from directly modifying CurrentToken when they're not supposed to.
167 This semantics change was requested by frank farmer.
168 - Percent encoding checks enabled for URI query and fragment
169 - Fix stray backslashes in font-family; CSS Unicode character escapes are
170 now properly resolved (although *only* in font-family). Thanks Takeshi Terada
172 - Improve parseCDATA algorithm to take into account newline normalization
173 - Account for browser confusion between Yen character and backslash in
174 Shift_JIS encoding. This fix generalizes to any other encoding which is not
175 a strict superset of printable ASCII. Thanks Takeshi Terada for reporting.
176 - Fix missing configuration parameter in Generator calls. Thanks vs for the
178 - Improved adherence to Unicode by checking for non-character codepoints.
179 Thanks Geoffrey Sneddon for reporting. This may result in degraded
180 performance for extremely large inputs.
181 - Allow CSS property-value pair ''text-decoration: none''. Thanks Jochem Blok
183 . Added HTMLPurifier_UnitConverter and HTMLPurifier_Length for convenient
184 handling of CSS-style lengths. HTMLPurifier_AttrDef_CSS_Length now uses
186 . API of HTMLPurifier_AttrDef_CSS_Length changed from __construct($disable_negative)
187 to __construct($min, $max). __construct(true) is equivalent to
189 . Added HTMLPurifier_AttrDef_Switch class
190 . Rename HTMLPurifier_HTMLModule_Tidy->construct() to setup() and bubble method
191 up inheritance hierarchy to HTMLPurifier_HTMLModule. All HTMLModules
192 get this called with the configuration object. All modules now
193 use this rather than __construct(), although legacy code using constructors
194 will still work--the new format, however, lets modules access the
195 configuration object for HTML namespace dependant tweaks.
196 . AttrDef_HTML_Pixels now takes a single construction parameter, pixels.
197 . ConfigSchema data-structure heavily optimized; on average it uses a third
198 the memory it did previously. The interface has changed accordingly,
199 consult changes to HTMLPurifier_Config for details.
200 . Variable parsing types now are magic integers instead of strings
201 . Added benchmark for ConfigSchema
202 . HTMLPurifier_Generator requires $config and $context parameters. If you
203 don't know what they should be, use HTMLPurifier_Config::createDefault()
204 and new HTMLPurifier_Context().
205 . Printers now properly distinguish between output configuration, and
206 target configuration. This is not applicable to scripts using
207 the Printers for HTML Purifier related tasks.
208 . HTML/CSS Printers must be primed with prepareGenerator($gen_config), otherwise
209 fatal errors will ensue.
210 . URIFilter->prepare can return false in order to abort loading of the filter
211 . Factory for AttrDef_URI implemented, URI#embedded to indicate URI that embeds
212 an external resource.
213 . %URI.Munge functionality factored out into a post-filter class.
214 . Added CurrentCSSProperty context variable during CSS validation
216 3.1.0, released 2008-05-18
217 # Unnecessary references to objects (vestiges of PHP4) removed from method
218 signatures. The following methods do not need references when assigning from
219 them and will result in E_STRICT errors if you try:
220 + HTMLPurifier_Config->get*Definition() [* = HTML, CSS]
221 + HTMLPurifier_ConfigSchema::instance()
222 + HTMLPurifier_DefinitionCacheFactory::instance()
223 + HTMLPurifier_DefinitionCacheFactory->create()
224 + HTMLPurifier_DoctypeRegistry->register()
225 + HTMLPurifier_DoctypeRegistry->get()
226 + HTMLPurifier_HTMLModule->addElement()
227 + HTMLPurifier_HTMLModule->addBlankElement()
228 + HTMLPurifier_LanguageFactory::instance()
229 # Printer_ConfigForm's get*() functions were static-ified
230 # %HTML.ForbiddenAttributes requires attribute declarations to be in the
231 form of tag@attr, NOT tag.attr (which will throw an error and won't do
232 anything). This is for forwards compatibility with XML; you'd do best
233 to migrate an %HTML.AllowedAttributes directives to this syntax too.
234 ! Allow index to be false for config from form creation
235 ! Added HTMLPurifier::VERSION constant
236 ! Commas, not dashes, used for serializer IDs. This change is forwards-compatible
237 and allows for version numbers like "3.1.0-dev".
238 ! %HTML.Allowed deals gracefully with whitespace anywhere, anytime!
239 ! HTML Purifier's URI handling is a lot more robust, with much stricter
240 validation checks and better percent encoding handling. Thanks Gareth Heyes
241 for indicating security vulnerabilities from lax percent encoding.
242 ! Bootstrap autoloader deals more robustly with classes that don't exist,
243 preventing class_exists($class, true) from barfing.
244 - InterchangeBuilder now alphabetizes its lists
245 - Validation error in configdoc output fixed
246 - Iconv and other encoding errors muted even with custom error handlers that
247 do not honor error_reporting
248 - Add protection against imagecrash attack with CSS height/width
249 - HTMLPurifier::instance() created for consistency, is equivalent to getInstance()
250 - Fixed and revamped broken ConfigForm smoketest
251 - Bug with bool/null fields in Printer_ConfigForm fixed
252 - Bug with global forbidden attributes fixed
253 - Improved error messages for allowed and forbidden HTML elements and attributes
254 - Missing (or null) in configdoc documentation restored
255 - If DOM throws and exception during parsing with PH5P (occurs in newer versions
256 of DOM), HTML Purifier punts to DirectLex
257 - Fatal error with unserialization of ScriptRequired
258 - Created directories are now chmod'ed properly
259 - Fixed bug with fallback languages in LanguageFactory
260 - Standalone testing setup properly with autoload
261 . Out-of-date documentation revised
262 . UTF-8 encoding check optimization as suggested by Diego
263 . HTMLPurifier_Error removed in favor of exceptions
264 . More copy() function removed; should use clone instead
265 . More extensive unit tests for HTMLDefinition
266 . assertPurification moved to central harness
267 . HTMLPurifier_Generator accepts $config and $context parameters during
268 instantiation, not runtime
269 . Double-quotes outside of attribute values are now unescaped
271 3.1.0rc1, released 2008-04-22
272 # Autoload support added. Internal require_once's removed in favor of an
273 explicit require list or autoloading. To use HTML Purifier,
274 you must now either use HTMLPurifier.auto.php
275 or HTMLPurifier.includes.php; setting the include path and including
276 HTMLPurifier.php is insufficient--in such cases include HTMLPurifier.autoload.php
277 as well to register our autoload handler (or modify your autoload function
278 to check HTMLPurifier_Bootstrap::getPath($class)). You can also use
279 HTMLPurifier.safe-includes.php for a less performance friendly but more
280 user-friendly library load.
281 # HTMLPurifier_ConfigSchema static functions are officially deprecated. Schema
282 information is stored in the ConfigSchema directory, and the
283 maintenance/generate-schema-cache.php generates the schema.ser file, which
284 is now instantiated. Support for userland schema changes coming soon!
285 # HTMLPurifier_Config will now throw E_USER_NOTICE when you use a directive
286 alias; to get rid of these errors just modify your configuration to use
287 the new directive name.
288 # HTMLPurifier->addFilter is deprecated; built-in filters can now be
289 enabled using %Filter.$filter_name or by setting your own filters using
291 # Directive-level safety properties superceded in favor of module-level
292 safety. Internal method HTMLModule->addElement() has changed, although
293 the externally visible HTMLDefinition->addElement has *not* changed.
294 ! Extra utility classes for testing and non-library operations can
295 be found in extras/. Specifically, these are FSTools and ConfigDoc.
296 You may find a use for these in your own project, but right now they
297 are highly experimental and volatile.
298 ! Integration with PHPT allows for automated smoketests
299 ! Limited support for proprietary HTML elements, namely <marquee>, sponsored
300 by Chris. You can enable them with %HTML.Proprietary if your client
302 ! Support for !important CSS cascade modifier. By default, this will be stripped
303 from CSS, but you can enable it using %CSS.AllowImportant
304 ! Support for display and visibility CSS properties added, set %CSS.AllowTricky
306 ! HTML Purifier now has its own Exception hierarchy under HTMLPurifier_Exception.
307 Developer error (not enduser error) can cause these to be triggered.
308 ! Experimental kses() wrapper introduced with HTMLPurifier.kses.php
309 ! Finally %CSS.AllowedProperties for tweaking allowed CSS properties without
310 mucking around with HTMLPurifier_CSSDefinition
311 ! ConfigDoc output has been enhanced with version and deprecation info.
312 ! %HTML.ForbiddenAttributes and %HTML.ForbiddenElements implemented.
313 - Autoclose now operates iteratively, i.e. <span><span><div> now has
314 both span tags closed.
315 - Various HTMLPurifier_Config convenience functions now accept another parameter
316 $schema which defines what HTMLPurifier_ConfigSchema to use besides the
318 - Fix bug with trusted script handling in libxml versions later than 2.6.28.
319 - Fix bug in ExtractStyleBlocks with comments in style tags
320 - Fix bug in comment parsing for DirectLex
321 - Flush output now displayed when in command line mode for unit tester
322 - Fix bug with rgb(0, 1, 2) color syntax with spaces inside shorthand syntax
323 - HTMLPurifier_HTMLDefinition->addAttribute can now be called multiple times
324 on the same element without emitting errors.
325 - Fixed fatal error in PH5P lexer with invalid tag names
326 . Plugins now get their own changelogs according to project conventions.
327 . Convert tokens to use instanceof, reducing memory footprint and
328 improving comparison speed.
329 . Dry runs now supported in SimpleTest; testing facilities improved
330 . Bootstrap class added for handling autoloading functionality
331 . Implemented recursive glob at FSTools->globr
332 . ConfigSchema now has instance methods for all corresponding define*
334 . A couple of new historical maintenance scripts were added.
335 . HTMLPurifier/HTMLModule/Tidy/XHTMLAndHTML4.php split into two files
336 . tests/index.php can now be run from any directory.
337 . HTMLPurifier_Token subclasses split into seperate files
338 . HTMLPURIFIER_PREFIX now is defined in Bootstrap.php, NOT HTMLPurifier.php
339 . HTMLPURIFIER_PREFIX can now be defined outside of HTML Purifier
340 . New --php=php flag added, allows PHP executable to be specified (command
342 . htmlpurifier_add_test() preferred method to translate test files in to
343 classes, because it handles PHPT files too.
344 . Debugger class is deprecated and will be removed soon.
345 . Command line argument parsing for testing scripts revamped, now --opt value
347 . Smoketests now cleanup after magic quotes
348 . Generator now can output comments (however, comments are still stripped
349 from HTML Purifier output)
350 . HTMLPurifier_ConfigSchema->validate() deprecated in favor of
351 HTMLPurifier_VarParser->parse()
352 . Integers auto-cast into float type by VarParser.
353 . HTMLPURIFIER_STRICT removed; no validation is performed on runtime, only
354 during cache generation
355 . Reordered script calls in maintenance/flush.php
356 . Command line scripts now honor exit codes
357 . When --flush fails in unit testers, abort tests and print message
358 . Improved documentation in docs/dev-flush.html about the maintenance scripts
359 . copy() methods removed in favor of clone keyword
361 3.0.0, released 2008-01-06
362 # HTML Purifier is PHP 5 only! The 2.1.x branch will be maintained
363 until PHP 4 is completely deprecated, but no new features will be added
365 + Visibility declarations added
366 + Constructor methods renamed to __construct()
367 + PHP4 reference cruft removed (in progress)
368 ! CSS properties are now case-insensitive
369 ! DefinitionCacheFactory now can register new implementations
370 ! New HTMLPurifier_Filter_ExtractStyleBlocks for extracting <style> from
371 documents and cleaning their contents up. Requires the CSSTidy library
372 <http://csstidy.sourceforge.net/>. You can access the blocks with the
373 'StyleBlocks' Context variable ($purifier->context->get('StyleBlocks')).
374 The output CSS can also be "scoped" for a specific element, use:
375 %Filter.ExtractStyleBlocksScope
376 ! Experimental support for some proprietary CSS attributes allowed:
377 opacity (and all of the browser-specific equivalents) and scrollbar colors.
378 Enable by setting %CSS.Proprietary to true.
379 - Colors missing # but in hex form will be corrected
380 - CSS Number algorithm improved
381 - Unit testing and multi-testing now on steroids: command lines,
382 XML output, and other goodies now added.
383 . Unit tests for Injector improved
385 + HTMLPurifier_AttrDef_CSS_AlphaValue
386 + HTMLPurifier_AttrDef_CSS_Filter
387 . Multitest now has a file docblock
389 2.1.3, released 2007-11-05
390 ! tests/multitest.php allows you to test multiple versions by running
391 tests/index.php through multiple interpreters using `phpv` shell
392 script (you must provide this script!)
393 - Fixed poor include ordering for Email URI AttrDefs, causes fatal errors
395 - Injector algorithm further refined: off-by-one error regarding skip
396 counts for dormant injectors fixed
397 - Corrective blockquote definition now enabled for HTML 4.01 Strict
398 - Fatal error when <img> tag (or any other element with required attributes)
399 has 'id' attribute fixed, thanks NykO18 for reporting
400 - Fix warning emitted when a non-supported URI scheme is passed to the
401 MakeAbsolute URIFilter, thanks NykO18 (again)
402 - Further refine AutoParagraph injector. Behavior inside of elements
403 allowing paragraph tags clarified: only inline content delimeted by
404 double newlines (not block elements) are paragraphed.
405 - Buggy treatment of end tags of elements that have required attributes
406 fixed (does not manifest on default tag-set)
407 - Spurious internal content reorganization error suppressed
408 - HTMLDefinition->addElement now returns a reference to the created
409 element object, as implied by the documentation
410 - Phorum mod's HTML Purifier help message expanded (unreleased elsewhere)
411 - Fix a theoretical class of infinite loops from DirectLex reported
413 - Work around unnecessary DOMElement type-cast in PH5P that caused errors
415 - Work around PHP 4 SimpleTest lack-of-error complaining for one-time-only
416 HTMLDefinition errors, this may indicate problems with error-collecting
418 - Make ErrorCollectorEMock work in both PHP 4 and PHP 5
419 - Make PH5P work with PHP 5.0 by removing unnecessary array parameter typedef
420 . %Core.AcceptFullDocuments renamed to %Core.ConvertDocumentToFragment
421 to better communicate its purpose
422 . Error unit tests can now specify the expectation of no errors. Future
423 iterations of the harness will be extremely strict about what errors
425 . Extend Injector hooks to allow for more powerful injector routines
426 . HTMLDefinition->addBlankElement created, as according to the HTMLModule
428 . Doxygen configuration file updated, with minor improvements
429 . Test runner now checks for similarly named files in conf/ directory too.
430 . Minor cosmetic change to flush-definition-cache.php: trailing newline is
432 . Maintenance script for generating PH5P patch added, original PH5P source
433 file also added under version control
434 . Full unit test runner script title made more descriptive with PHP version
435 . Updated INSTALL file to state that 4.3.7 is the earliest version we
438 2.1.2, released 2007-09-03
439 ! Implemented Object module for trusted users
440 ! Implemented experimental HTML5 parsing mode using PH5P. To use, add
442 require_once 'HTMLPurifier/Lexer/PH5P.php';
443 $config->set('Core', 'LexerImpl', 'PH5P');
444 Note that this Lexer introduces some classes not in the HTMLPurifier
445 namespace. Also, this is PHP5 only.
446 ! CSS property border-spacing implemented
447 - Fix non-visible parsing error in DirectLex with empty tags that have
448 slashes inside attribute values.
449 - Fix typo in CSS definition: border-collapse:seperate; was incorrectly
450 accepted as valid CSS. Usually non-visible, because this styling is the
451 default for tables in most browsers. Thanks Brett Zamir for pointing
453 - Fix validation errors in configuration form
454 - Hammer out a bunch of edge-case bugs in the standalone distribution
455 - Inclusion reflection removed from URISchemeRegistry; you must manually
456 include any new schema files you wish to use
457 - Numerous typo fixes in documentation thanks to Brett Zamir
458 . Unit test refactoring for one logical test per test function
459 . Config and context parameters in ComplexHarness deprecated: instead, edit
460 the $config and $context member variables
461 . HTML wrapper in DOMLex now takes DTD identifiers into account; doesn't
462 really make a difference, but is good for completeness sake
463 . merge-library.php script refactored for greater code reusability and
466 2.1.1, released 2007-08-04
467 - Fix show-stopper bug in %URI.MakeAbsolute functionality
468 - Fix PHP4 syntax error in standalone version
469 . Add prefix directory to include path for standalone, this prevents
470 other installations from clobbering the standalone's URI schemes
471 . Single test methods can be invoked by prefixing with __only
473 2.1.0, released 2007-08-02
474 # flush-htmldefinition-cache.php superseded in favor of a generic
475 flush-definition-cache.php script, you can clear a specific cache
476 by passing its name as a parameter to the script
477 ! Phorum mod implemented for HTML Purifier
478 ! With %Core.AggressivelyFixLt, <3 and similar emoticons no longer
479 trigger HTML removal in PHP5 (DOMLex). This directive is not necessary
480 for PHP4 (DirectLex).
481 ! Standalone file now available, which greatly reduces the amount of
482 includes (although there are still a few files that reside in the
484 ! Relative URIs can now be transformed into their absolute equivalents
485 using %URI.Base and %URI.MakeAbsolute
486 ! Ruby implemented for XHTML 1.1
487 ! You can now define custom URI filtering behavior, see enduser-uri-filter.html
489 ! UTF-8 font names now supported in CSS
490 - AutoFormatters emit friendly error messages if tags or attributes they
492 - ConfigForm's compactification of directive names is now configurable
493 - AutoParagraph autoformatter algorithm refined after field-testing
494 - XHTML 1.1 now applies XHTML 1.0 Strict cleanup routines, namely
496 - Contents of <style> tags removed by default when tags are removed
497 . HTMLPurifier_Config->getSerial() implemented, this is extremely useful
498 for output cache invalidation
499 . ConfigForm printer now can retrieve CSS and JS files as strings, in
500 case HTML Purifier's directory is not publically accessible
501 . Introduce new text/itext configuration directive values: these represent
502 longer strings that would be more appropriately edited with a textarea
503 . Allow newlines to act as separators for lists, hashes, lookups and
505 . ConfigForm generates textareas instead of text inputs for lists, hashes,
506 lookups, text and itext fields
507 . Hidden element content removal genericized: %Core.HiddenElements can
508 be used to customize this behavior, by default <script> and <style> are
510 . Added HTMLPURIFIER_PREFIX constant, should be used instead of dirname(__FILE__)
511 . Custom ChildDef added to default include list
512 . URIScheme reflection improved: will not attempt to include file if class
513 already exists. May clobber autoload, so I need to keep an eye on it
514 . ConfigSchema heavily optimized, will only collect information and validate
515 definitions when HTMLPURIFIER_SCHEMA_STRICT is true.
516 . AttrDef_URI unit tests and implementation refactored
517 . benchmarks/ directory now protected from public view with .htaccess file;
518 run the tests via command line
519 . URI scheme is munged off if there is no authority and the scheme is the
521 . All unit tests inherit from HTMLPurifier_Harness, not UnitTestCase
522 . Interface for URIScheme changed
523 . Generic URI object to hold components of URI added, most systems involved
524 in URI validation have been migrated to use it
525 . Custom filtering for URIs factored out to URIDefinition interface for
526 maximum extensibility
528 2.0.1, released 2007-06-27
529 ! Tag auto-closing now based on a ChildDef heuristic rather than a
530 manually set auto_close array; some behavior may change
531 ! Experimental AutoFormat functionality added: auto-paragraph and
532 linkify your HTML input by setting %AutoFormat.AutoParagraph and
533 %AutoFormat.Linkify to true
534 ! Newlines normalized internally, and then converted back to the
535 value of PHP_EOL. If this is not desired, set your newline format
536 using %Output.Newline.
537 ! Beta error collection, messages are implemented for the most generic
538 cases involving Lexing or Strategies
539 - Clean up special case code for <script> tags
540 - Reorder includes for DefinitionCache decorators, fixes a possible
542 - Fixed bug where manually modified definitions were not saved via cache
543 (mostly harmless, except for the fact that it would be a little slower)
544 - Configuration objects with different serials do not clobber each
545 others when revision numbers are unequal
546 - Improve Serializer DefinitionCache directory permissions checks
547 - DefinitionCache no longer throws errors when it encounters old
548 serial files that do not conform to the current style
549 - Stray xmlns attributes removed from configuration documentation
550 - configForm.php smoketest no longer has XSS vulnerability due to
551 unescaped print_r output
552 - Printer adheres to configuration's directives on output format
553 - Fix improperly named form field in ConfigForm printer
554 . Rewire some test-cases to swallow errors rather than expect them
555 . HTMLDefinition printer updated with some of the new attributes
556 . DefinitionCache keys reordered to reflect precedence: version number,
557 hash, then revision number
558 . %Core.DefinitionCache renamed to %Cache.DefinitionImpl
559 . Interlinking in configuration documentation added using
560 Injector_PurifierLinkify
561 . Directives now keep track of aliases to themselves
562 . Error collector now requires a severity to be passed, use PHP's internal
563 error constants for this
564 . HTMLPurifier_Config::getAllowedDirectivesForForm implemented, allows
565 much easier selective embedding of configuration values
566 . Doctype objects now accept public and system DTD identifiers
567 . %HTML.Doctype is now constrained by specific values, to specify a custom
568 doctype use new %HTML.CustomDoctype
569 . ConfigForm truncates long directives to keep the form small, and does
570 not re-output namespaces
572 2.0.0, released 2007-06-20
573 # Completely refactored HTMLModuleManager, decentralizing safety
575 # Transform modules changed to Tidy modules, which offer more flexibility
576 and better modularization
577 # Configuration object now finalizes itself when a read operation is
578 performed on it, ensuring that its internal state stays consistent.
579 To revert this behavior, you can set the $autoFinalize member variable
580 off, but it's not recommended.
581 # New compact syntax for AttrDef objects that can be used to instantiate
582 new objects via make()
583 # Definitions (esp. HTMLDefinition) are now cached for a significant
584 performance boost. You can disable caching by setting %Core.DefinitionCache
585 to null. You CANNOT edit raw definitions without setting the corresponding
586 DefinitionID directive (%HTML.DefinitionID for HTMLDefinition).
587 # Contents between <script> tags are now completely removed if <script>
589 # Prototype-declarations for Lexer removed in favor of configuration
590 determination of Lexer implementations.
591 ! HTML Purifier now works in PHP 4.3.2.
592 ! Configuration form-editing API makes tweaking HTMLPurifier_Config a
594 ! Configuration directives that accept hashes now allow new string
595 format: key1:value1,key2:value2
596 ! ConfigDoc now factored into OOP design
597 ! All deprecated elements now natively supported
598 ! Implement TinyMCE styled whitelist specification format in
600 ! Config object gives more friendly error messages when things go wrong
601 ! Advanced API implemented: easy functions for creating elements (addElement)
602 and attributes (addAttribute) on HTMLDefinition
603 ! Add native support for required attributes
604 - Deprecated and removed EnableRedundantUTF8Cleaning. It didn't even work!
605 - DOMLex will not emit errors when a custom error handler that does not
606 honor error_reporting is used
607 - StrictBlockquote child definition refrains from wrapping whitespace
609 - Bug resulting from tag transforms to non-allowed elements fixed
610 - ChildDef_Custom's regex generation has been improved, removing several
612 . Unit test for ElementDef created, ElementDef behavior modified to
614 . Added convenience functions for HTMLModule constructors
615 . AttrTypes now has accessor functions that should be used instead
616 of directly manipulating info
617 . TagTransform_Center deprecated in favor of generic TagTransform_Simple
618 . Add extra protection in AttrDef_URI against phantom Schemes
619 . Doctype object added to HTMLDefinition which describes certain aspects
620 of the operational document type
621 . Lexer is now pre-emptively included, with a conditional include for the
623 . HTMLDefinition and CSSDefinition have a common parent class: Definition.
624 . DirectLex can now track line-numbers
625 . Preliminary error collector is in place, although no code actually reports
627 . Factor out most of ValidateAttributes to new AttrValidator class
629 1.6.1, released 2007-05-05
630 ! Support for more deprecated attributes via transformations:
631 + hspace and vspace in img
632 + size and noshade in hr
635 + align in caption, table, img and hr
636 + type in ul, ol and li
637 ! DirectLex now preserves text in which a < bracket is followed by
638 a non-alphanumeric character. This means that certain emoticons
640 ! %Core.RemoveInvalidImg is now operational, when set to false invalid
641 images will hang around with an empty src
642 ! target attribute in a tag supported, use %Attr.AllowedFrameTargets
644 ! CSS property white-space now allows nowrap (supported in all modern
645 browsers) but not others (which have spotty browser implementations)
646 ! XHTML 1.1 mode now sort-of works without any fatal errors, and
647 lang is now moved over to xml:lang.
648 ! Attribute transformation smoketest available at smoketests/attrTransform.php
649 ! Transformation of font's size attribute now handles super-large numbers
650 - Possibly fatal bug with __autoload() fixed in module manager
651 - Invert HTMLModuleManager->addModule() processing order to check
652 prefixes first and then the literal module
653 - Empty strings get converted to empty arrays instead of arrays with
654 an empty string in them.
655 - Merging in attribute lists now works.
656 . Demo script removed: it has been added to the website's repository
657 . Basic.php script modified to work out of the box
658 . Refactor AttrTransform classes to reduce duplication
659 . AttrTransform_TextAlign axed in favor of a more general
660 AttrTransform_EnumToCSS, refer to HTMLModule/TransformToStrict.php to
661 see how the new equivalent is implemented
662 . Unit tests now use exclusively assertIdentical
664 1.6.0, released 2007-04-01
665 ! Support for most common deprecated attributes via transformations:
666 + bgcolor in td, th, tr and table
669 + width in td, th and hr
671 ! Support for CSS attribute 'height' added
672 ! Support for rel and rev attributes in a tags added, use %Attr.AllowedRel
673 and %Attr.AllowedRev to activate
674 - You can define ID blacklists using regular expressions via
675 %Attr.IDBlacklistRegexp
676 - Error messages are emitted when you attempt to "allow" elements or
677 attributes that HTML Purifier does not support
678 - Fix segfault in unit test. The problem is not very reproduceable and
679 I don't know what causes it, but a six line patch fixed it.
681 1.5.0, released 2007-03-23
682 ! Added a rudimentary I18N and L10N system modeled off MediaWiki. It
683 doesn't actually do anything yet, but keep your eyes peeled.
684 ! docs/enduser-utf8.html explains how to use UTF-8 and HTML Purifier
685 ! Newly structured HTMLDefinition modeled off of XHTML 1.1 modules.
686 I am loathe to release beta quality APIs, but this is exactly that;
687 don't use the internal interfaces if you're not willing to do migration
689 - Allow 'x' subtag in language codes
690 - Fixed buggy chameleon-support for ins and del
691 . Added support for IDREF attributes (i.e. for)
692 . Renamed HTMLPurifier_AttrDef_Class to HTMLPurifier_AttrDef_Nmtokens
693 . Removed context variable ParentType, replaced with IsInline, which
694 is false when you're not inline and an integer of the parent that
695 caused you to become inline when you are (so possibly zero)
696 . Removed ElementDef->type in favor of ElementDef->descendants_are_inline
697 and HTMLDefinition->content_sets
698 . StrictBlockquote now reports what elements its supposed to allow,
699 rather than what it does allow
700 . Removed HTMLDefinition->info_flow_elements in favor of
701 HTMLDefinition->content_sets['Flow']
702 . Removed redundant "exclusionary" definitions from DTD roster
703 . StrictBlockquote now requires a construction parameter as if it
704 were an Required ChildDef, this is the "real" set of allowed elements
705 . AttrDef partitioned into HTML, CSS and URI segments
706 . Modify Youtube filter regexp to be multiline
707 . Require both PHP5 and DOM extension in order to use DOMLex, fixes
708 some edge cases where a DOMDocument class exists in a PHP4 environment
709 due to DOM XML extension.
711 1.4.1, released 2007-01-21
712 ! docs/enduser-youtube.html updated according to new functionality
713 - YouTube IDs can have underscores and dashes
715 1.4.0, released 2007-01-21
716 ! Implemented list-style-image, URIs now allowed in list-style
717 ! Implemented background-image, background-repeat, background-attachment
718 and background-position CSS properties. Shorthand property background
719 supports all of these properties.
720 ! Configuration documentation looks nicer
721 ! Added %Core.EscapeNonASCIICharacters to workaround loss of Unicode
722 characters while %Core.Encoding is set to a non-UTF-8 encoding.
723 ! Support for configuration directive aliases added
724 ! Config object can now be instantiated from ini files
725 ! YouTube preservation code added to the core, with two lines of code
726 you can add it as a filter to your code. See smoketests/preserveYouTube.php
728 ! Moved SLOW to docs/enduser-slow.html and added code examples
729 - Replaced version check with functionality check for DOM (thanks Stephen
731 . Added smoketest 'all.php', which loads all other smoketests via frames
732 . Implemented AttrDef_CSSURI for url(http://google.com) style declarations
733 . Added convenient single test selector form on test runner
735 1.3.2, released 2006-12-25
736 ! HTMLPurifier object now accepts configuration arrays, no need to manually
737 instantiate a configuration object
738 ! Context object now accessible to outside
739 ! Added enduser-youtube.html, explains how to embed YouTube videos. See
740 also corresponding smoketest preserveYouTube.php.
741 ! Added purifyArray(), which takes a list of HTML and purifies it all
742 ! Added static member variable $version to HTML Purifier with PHP-compatible
743 version number string.
744 - Fixed fatal error thrown by upper-cased language attributes
745 - printDefinition.php: added labels, added better clarification
746 . HTMLPurifier_Config::create() added, takes mixed variable and converts into
747 a HTMLPurifier_Config object.
749 1.3.1, released 2006-12-06
750 ! Added HTMLPurifier.func.php stub for a convenient function to call the library
751 - Fixed bug in RemoveInvalidImg code that caused all images to be dropped
752 (thanks to .mario for reporting this)
753 . Standardized all attribute handling variables to attr, made it plural
755 1.3.0, released 2006-11-26
756 # Invalid images are now removed, rather than replaced with a dud
757 <img src="" alt="Invalid image" />. Previous behavior can be restored
758 with new directive %Core.RemoveInvalidImg set to false.
759 ! (X)HTML Strict now supported
760 + Transparently handles inline elements in block context (blockquote)
761 ! Added GET method to demo for easier validation, added 50kb max input size
762 ! New directive %HTML.BlockWrapper, for block-ifying inline elements
763 ! New directive %HTML.Parent, allows you to only allow inline content
764 ! New directives %HTML.AllowedElements and %HTML.AllowedAttributes to let
765 users narrow the set of allowed tags
766 ! <li value="4"> and <ul start="2"> now allowed in loose mode
767 ! New directives %URI.DisableExternalResources and %URI.DisableResources
768 ! New directive %Attr.DisableURI, which eliminates all hyperlinking
769 ! New directive %URI.Munge, munges URI so you can use some sort of redirector
770 service to avoid PageRank leaks or warn users that they are exiting your site.
771 ! Added spiffy new smoketest printDefinition.php, which lets you twiddle with
772 the configuration settings and see how the internal rules are affected.
773 ! New directive %URI.HostBlacklist for blocking links to bad hosts.
774 xssAttacks.php smoketest updated accordingly.
775 - Added missing type to ChildDef_Chameleon
776 - Remove Tidy option from demo if there is not Tidy available
777 . ChildDef_Required guards against empty tags
778 . Lookup table HTMLDefinition->info_flow_elements added
779 . Added peace-of-mind variable initialization to Strategy_FixNesting
780 . Added HTMLPurifier->info_parent_def, parent child processing made special
781 . Added internal documents briefly summarizing future progression of HTML
782 . HTMLPurifier_Config->getBatch($namespace) added
783 . More lenient casting to bool from string in HTMLPurifier_ConfigSchema
784 . Refactored ChildDef classes into their own files
786 1.2.0, released 2006-11-19
787 # ID attributes now disabled by default. New directives:
788 + %HTML.EnableAttrID - restores old behavior by allowing IDs
789 + %Attr.IDPrefix - %Attr.IDBlacklist alternative that munges all user IDs
790 so that they don't collide with your IDs
791 + %Attr.IDPrefixLocal - Same as above, but for when there are multiple
792 instances of user content on the page
793 + Profuse documentation on how to use these available in docs/enduser-id.txt
794 ! Added MODx plugin <http://modxcms.com/forums/index.php/topic,6604.0.html>
795 ! Added percent encoding normalization
796 ! XSS attacks smoketest given facelift
797 ! Configuration documentation now has table of contents
798 ! Added %URI.DisableExternal, which prevents links to external websites. You
799 can also use %URI.Host to permit absolute linking to subdomains
800 ! Non-accessible resources (ex. mailto) blocked from embedded URIs (img src)
801 - Type variable in HTMLDefinition was not being set properly, fixed
802 - Documentation updated
803 + TODO added request Phalanger
804 + TODO added request Native compression
805 + TODO added request Remove redundant tags
806 + TODO added possible plaintext formatter for HTML Purifier documentation
807 + Updated ConfigDoc TODO
808 + Improved inline comments in AttrDef/Class.php, AttrDef/CSS.php
810 + Revamped documentation into HTML, along with misc updates
811 - HTMLPurifier_Context doesn't throw a variable reference error if you attempt
812 to retrieve a non-existent variable
813 . Switched to purify()-wide Context object registry
814 . Refactored unit tests to minimize duplication
815 . XSS attack sheet updated
816 . configdoc.xml now has xml:space attached to default value nodes
817 . Allow configuration directives to permit null values
818 . Cleaned up test-cases to remove unnecessary swallowErrors()
820 1.1.2, released 2006-09-30
821 ! Add HTMLPurifier.auto.php stub file that configures include_path
822 - Documentation updated
823 + INSTALL document rewritten
824 + TODO added semi-lossy conversion
825 + API Doxygen docs' file exclusions updated
826 + Added notes on HTML versus XML attribute whitespace handling
827 + Noted that HTMLPurifier_ChildDef_Custom isn't being used
828 + Noted that config object's definitions are cached versions
829 - Fixed lack of attribute parsing in HTMLPurifier_Lexer_PEARSax3
830 - ftp:// URIs now have their typecodes checked
831 - Hooked up HTMLPurifier_ChildDef_Custom's unit tests (they weren't being run)
832 . Line endings standardized throughout project (svn:eol-style standardized)
833 . Refactored parseData() to general Lexer class
834 . Tester named "HTML Purifier" not "HTMLPurifier"
836 1.1.1, released 2006-09-24
837 ! Configuration option to optionally Tidy up output for indentation to make up
838 for dropped whitespace by DOMLex (pretty-printing for the entire application
839 should be done by a page-wide Tidy)
840 - Various documentation updates
841 - Fixed parse error in configuration documentation script
842 - Fixed fatal error in benchmark scripts, slightly augmented
843 - As far as possible, whitespace is preserved in-between table children
844 - Sample test-settings.php file included
846 1.1.0, released 2006-09-16
847 ! Directive documentation generation using XSLT
848 ! XHTML can now be turned off, output becomes <br>
849 - Made URI validator more forgiving: will ignore leading and trailing
850 quotes, apostrophes and less than or greater than signs.
851 - Enforce alphanumeric namespace and directive names for configuration.
852 - Table child definition made more flexible, will fix up poorly ordered elements
853 . Renamed ConfigDef to ConfigSchema
855 1.0.1, released 2006-09-04
856 - Fixed slight bug in DOMLex attribute parsing
857 - Fixed rejection of case-insensitive configuration values when there is a
858 set of allowed values. This manifested in %Core.Encoding.
859 - Fixed rejection of inline style declarations that had lots of extra
860 space in them. This manifested in TinyMCE.
862 1.0.0, released 2006-09-01
863 ! Shorthand CSS properties implemented: font, border, background, list-style
864 ! Basic color keywords translated into hexadecimal values
865 ! Table CSS properties implemented
866 ! Support for charsets other than UTF-8 (defined by iconv)
867 ! Malformed UTF-8 and non-SGML character detection and cleaning implemented
868 - Fixed broken numeric entity conversion
869 - API documentation completed
870 . (HTML|CSS)Definition de-singleton-ized
872 1.0.0beta, released 2006-08-16
873 ! First public release, most functionality implemented. Notable omissions are:
874 + Shorthand CSS properties
875 + Table CSS properties
876 + Deprecated attribute transformations