summaryrefslogtreecommitdiff
path: root/src/tokeniser
Commit message (Collapse)AuthorAgeFilesLines
* Fix entity consumption. This gets us to the second set of html5lib tests.Andrew Sidwell2008-06-171-35/+54
| | | | svn path=/trunk/hubbub/; revision=4366
* Prevent an infinite loop; get entity matcher a tiny bit further towards ↵Andrew Sidwell2008-06-161-0/+3
| | | | | | compliance. svn path=/trunk/hubbub/; revision=4364
* Remove dead code.Andrew Sidwell2008-06-161-5/+0
| | | | svn path=/trunk/hubbub/; revision=4362
* Was a little too hasty in the previous commit-- the same thing was done ↵Andrew Sidwell2008-06-161-2/+0
| | | | | | earlier on. svn path=/trunk/hubbub/; revision=4361
* Ensure the "additional allowed character" in the "consume a character ↵Andrew Sidwell2008-06-161-0/+1
| | | | | | reference" steps gets reset. svn path=/trunk/hubbub/; revision=4360
* - Move away from using inputstream_push_back() and instead use _rewind() ↵Andrew Sidwell2008-06-161-42/+25
| | | | | | | | for both clarity and efficiency. - Fix a bug where the CDATA-matching code was looking at the doctype-matching count. svn path=/trunk/hubbub/; revision=4359
* Update the tokeniser to the 2008-06-15 draft.Andrew Sidwell2008-06-161-19/+18
| | | | svn path=/trunk/hubbub/; revision=4358
* Update the tokeniser to the 2008-06-09 version of the spec. There are still ↵Andrew Sidwell2008-06-161-95/+1072
| | | | | | some bugs to iron out, mainly around entities. svn path=/trunk/hubbub/; revision=4355
* Add the current entity list from the HTML5 spec.Andrew Sidwell2008-06-161-259/+2137
| | | | svn path=/trunk/hubbub/; revision=4354
* Rework buildsystem so that it no longer calls make recursively and rebuilds ↵John Mark Bell2008-04-071-43/+36
| | | | | | the testcases when the library changes. svn path=/trunk/hubbub/; revision=4077
* hubbub_strings may now be either an offset into the data buffer or a pointer ↵John Mark Bell2008-03-211-49/+64
| | | | | | | | | | | | | to constant data. Fix up tokeniser and treebuilder to deal with this. Fix up testcases, too. The tokeniser will only ever emit strings of type HUBBUB_STRING_OFF. Anything else is a bug which should be fixed. The treebuilder may emit strings of either type. svn path=/trunk/hubbub/; revision=4014
* Import hubbub -- an HTML parsing library.John Mark Bell2007-06-235-0/+2794
Plenty of work still to do (like tree generation ;) svn path=/trunk/hubbub/; revision=3359