This mega-patch rips out the FixNesting implementation and the related
ChildDef components. The primary algorithmic change is to convert from
use of tokens to tree nodes, which are far more amenable to the style
of processing that FixNesting uses. Additionally, FixNesting has been
changed to go bottom-up rather than top-down, in order to avoid needing
to implement backtracking.
This patch simplifies a good deal of the relevant logic, since we no
longer need to continually recalculate the nesting structure when
processing things. However, the conversion to the alternate format
incurs some overhead, so for small inputs these changes are not a win.
One possibility to greatly reduce the constant factors here is to switch
to entirely using libxml's representation, and never serializing tokens;
this would require one to rewrite injectors, however.
The iterative post-order traversal in FixNesting is a bit subtle, but
we have essentially reified the stack and continuations.
We've removed support for %Core.EscapeInvalidChildren.
Signed-off-by: Edward Z. Yang <ezyang@mit.edu>
- Add FSTools:globr()
- require_once removed from all files
- HTMLPurifier.autoload.php added to register autoload handler
- Removed redundant chdir in maintenance script
- Modified standalone to use HTMLPurifier.includes.php for including stuff
- Added maintenance script remove-require-once.php which we used once and should never use again
git-svn-id: http://htmlpurifier.org/svnroot/htmlpurifier/trunk@1516 48356398-32a2-884e-a903-53898d9a118a
- Update TODO list
- URISchemeRegistry doesn't return a reference for instance anymore, should do the same for other singletons
git-svn-id: http://htmlpurifier.org/svnroot/htmlpurifier/trunk@1477 48356398-32a2-884e-a903-53898d9a118a
- Buggy treatment of end tags of elements that have required attributes fixed (does not manifest on default tag-set)
- Spurious internal content reorganization error suppressed
. Error unit tests can now specify the expectation of no errors. Future iterations of the harness will be extremely strict about what errors are allowed
git-svn-id: http://htmlpurifier.org/svnroot/htmlpurifier/trunk@1424 48356398-32a2-884e-a903-53898d9a118a
- Rework descendants_are_inline to have default value as false, ins/del handling now works top-level when parent element is not block
- Remove CleanUTF8OnGeneration, feature didn't even work
git-svn-id: http://htmlpurifier.org/svnroot/htmlpurifier/trunk@1086 48356398-32a2-884e-a903-53898d9a118a
- Fixed buggy chameleon-support for ins and del
. Removed context variable ParentType, replaced with IsInline, which
is false when you're not inline and an integer of the parent that
caused you to become inline when you are (so possibly zero)
. Removed ElementDef->type in favor of ElementDef->descendants_are_inline
and HTMLDefinition->content_sets
. StrictBlockquote now reports what elements its supposed to allow,
rather than what it does allow
. Removed HTMLDefinition->info_flow_elements in favor of
HTMLDefinition->content_sets['Flow']
. Removed redundant "exclusionary" definitions from DTD roster
. StrictBlockquote now requires a construction parameter as if it
were an Required ChildDef, this is the "real" set of allowed elements
git-svn-id: http://htmlpurifier.org/svnroot/htmlpurifier/trunk@710 48356398-32a2-884e-a903-53898d9a118a
+ Transparently handles inline elements in block context (blockquote)
! Added GET method to demo for easier validation, added 50kb max input size
! New directive %HTML.BlockWrapper, for block-ifying inline elements
! New directive %HTML.Parent, allows you to only allow inline content
- Added missing type to ChildDef_Chameleon
. ChildDef_Required guards against empty tags
. Lookup table HTMLDefinition->info_flow_elements added
. Added peace-of-mind variable initialization to Strategy_FixNesting
git-svn-id: http://htmlpurifier.org/svnroot/htmlpurifier/trunk@560 48356398-32a2-884e-a903-53898d9a118a
- Partially finished migrating to new Context object (done in r485).
- Created HTMLPurifier_Harness to assist with testing, ChildDefTest migrated to that framework.
git-svn-id: http://htmlpurifier.org/svnroot/htmlpurifier/trunk@484 48356398-32a2-884e-a903-53898d9a118a
- Defined new directive %Core.EscapeInvalidChildren, for previously commented out functionality
- Removed convenience configuration generation: you *have* to pass it unless you're interfacing with HTMLPurifier
- Homogenized function parameters even when only a few of them are used
- Rewrote unit tests that expected previous behavior
- Introduced configuration object to ChildDef tests
git-svn-id: http://htmlpurifier.org/svnroot/htmlpurifier/trunk@243 48356398-32a2-884e-a903-53898d9a118a
- Move around child definitions so they make a little more sense (rename to Custom) and also add $allow_empty property to help FixNesting.php determine whether or not to double-check.
git-svn-id: http://htmlpurifier.org/svnroot/htmlpurifier/trunk@136 48356398-32a2-884e-a903-53898d9a118a
