Skip to content
GitLab
Switch branch/tag
  1. 04 Jun, 2010 9 commits
  3. 02 Jun, 2010 1 commit
    • eric@webkit.org's avatar
      2010-06-02 Eric Seidel <eric@webkit.org> · 7e86028c
      eric@webkit.org authored 
              Reviewed by Adam Barth.

        HTML5 parser does not track line numbers
        https://bugs.webkit.org/show_bug.cgi?id=39984

        Update our expectations to show the proper line numbers.
        These new line numbers match what the old parser produces.

        * html5lib/runner-expected-html5.txt:
        * html5lib/webkit-resumer-expected-html5.txt:
2010-06-02  Eric Seidel  <eric@webkit.org>

        Reviewed by Adam Barth.

        HTML5 parser does not track line numbers
        https://bugs.webkit.org/show_bug.cgi?id=39984

        This adds rudimentary line-number tracking to the HTML5Lexer.
        We'll still need to handle \r\n cases eventually, but this
        fixes the vast majority of our line-number related failures.

        Fixes 268 layout tests. :)

        * html/HTML5Lexer.cpp:
        (WebCore::HTML5Lexer::reset):
         - Reset m_lineNumber to 0.
        (WebCore::HTML5Lexer::consumeEntity):
         - Pass m_lineNumber to advance().
        (WebCore::HTML5Lexer::nextToken):
         - Pass m_lineNumber to advance().
        * html/HTML5Lexer.h:
        (WebCore::HTML5Lexer::lineNumber):
         - Expose m_lineNumber for HTML5Tokenizer.
        (WebCore::HTML5Lexer::columnNumber):
         - Mirrors the implementation in the old HTMLTokenizer.
           we might some day support column number tracking in
           the Lexer.  It certainly would be possible.
        * html/HTML5ScriptRunner.cpp:
        (WebCore::HTML5ScriptRunner::HTML5ScriptRunner):
         - ASSERT we're passed a host.
        (WebCore::HTML5ScriptRunner::sourceFromPendingScript):
         - Use PendingScript.startingLineNumber when PendingScript is an inline script.
        (WebCore::HTML5ScriptRunner::execute):
         - Now expects a start line number passed from the TreeBuilder.
        (WebCore::HTML5ScriptRunner::runScript):
         - Now expects a start line number.
        * html/HTML5ScriptRunner.h:
        (WebCore::HTML5ScriptRunner::PendingScript::PendingScript):
         - Track startingLineNumber for any inline script.
        * html/HTML5Tokenizer.cpp:
        (WebCore::HTML5Tokenizer::begin):
         - Add a FIXME.  The old HTMLTokenizer resets its "lexer" state here.
        (WebCore::HTML5Tokenizer::pumpLexer):
         - Pass scriptStartLine to the HTML5ScriptRunner
        (WebCore::HTML5Tokenizer::lineNumber):
         - Implementation for Tokenizer.h
        (WebCore::HTML5Tokenizer::columnNumber):
         - Implementation for Tokenizer.h
        * html/HTML5Tokenizer.h:
        * html/HTML5TreeBuilder.cpp:
        (WebCore::HTML5TreeBuilder::HTML5TreeBuilder):
         - Initialize the new line number tracking variables.
        (WebCore::HTML5TreeBuilder::handleScriptEndTag):
         - Save off the line number from the start tag to pass to the ScriptRunner.
        (WebCore::HTML5TreeBuilder::takeScriptToProcess):
         - Return the line number from the start tag.
        (WebCore::HTML5TreeBuilder::passTokenToLegacyParser):
         - Save off the line number of ever script start tag we see.
        * html/HTML5TreeBuilder.h:
         - Keep both the line number of the last start tag, and the
           line number of the script we know we're expecting our caller to
           run before resuming parsing.  These are separate to keep the code
           clean, especially since the last script start tag line number is
           a hack specific to using the old HTMLParser.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@60553 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      7e86028c
  5. 01 Jun, 2010 1 commit
    • eric@webkit.org's avatar
      2010-05-31 Adam Barth <abarth@webkit.org> · 715c149d
      eric@webkit.org authored 
              Reviewed by Eric Seidel.

        Fix default action for EndTagOpenState
        https://bugs.webkit.org/show_bug.cgi?id=39982

        Add a test for a bogus end tag to webkit01.dat and update expected
        results now that we pass some more tests.

        Adding this test revealed a bug in resuming the bogus comment state.
        I've left these broken expectations, but I'll fix the bug in a future
        patch. (The bug existed prior to this patch, just not the test.)

        * html5lib/resources/webkit01.dat:
        * html5lib/runner-expected-html5.txt:
        * html5lib/webkit-resumer-expected-html5.txt:
2010-05-31  Adam Barth  <abarth@webkit.org>

        Reviewed by Eric Seidel.

        Fix default action for EndTagOpenState
        https://bugs.webkit.org/show_bug.cgi?id=39982

        Add a test for a bogus end tag to webkit01.dat and update expected
        results now that we pass some more tests.

        Adding this test revealed a bug in resuming the bogus comment state.
        I've left these broken expectations, but I'll fix the bug in a future
        patch. (The bug existed prior to this patch, just not the test.)

        * html5lib/resources/webkit01.dat:
        * html5lib/runner-expected-html5.txt:
        * html5lib/webkit-resumer-expected-html5.txt:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@60465 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      715c149d
  7. 31 May, 2010 1 commit
  9. 28 May, 2010 2 commits
    • abarth@webkit.org's avatar
      2010-05-28 Adam Barth <abarth@webkit.org> · 036a420c
      abarth@webkit.org authored 
              Reviewed by Eric Seidel.

        Named entities in attributes aren't parsed correctly by HTML5 parser
        https://bugs.webkit.org/show_bug.cgi?id=39873

        I misplaced this if statement when writing this code originally.  Now
        that we have test coverage for this paragraph in the spec, we can see
        and fix the bug.

        * html/HTML5Lexer.cpp:
        (WebCore::HTML5Lexer::consumeEntity):
2010-05-28  Adam Barth  <abarth@webkit.org>

        Reviewed by Eric Seidel.

        Named entities in attributes aren't parsed correctly by HTML5 parser
        https://bugs.webkit.org/show_bug.cgi?id=39873

        Add a test suite for parsing entities in attributes and update expected results.

        * html5lib/resources/entities02.dat: Added.
        * html5lib/runner-expected-html5.txt:
        * html5lib/runner-expected.txt:
        * html5lib/runner.html:


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@60370 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      036a420c
    • abarth@webkit.org's avatar
      2010-05-28 Adam Barth <abarth@webkit.org> · fc8e393e
      abarth@webkit.org authored 
              Reviewed by Eric Seidel.

        Handle edge cases in HTML5 entity parsing
        https://bugs.webkit.org/show_bug.cgi?id=39823

        The HTML5 specification tells us to handle HTML entities in a somewhat
        complicated way.  This patch attempts to correctly handle numeric
        entities.  Some of this code is duplicated from HTMLTokenizer.

        * html/HTML5Lexer.cpp:
        (WebCore::HTMLNames::):
        (WebCore::HTMLNames::adjustEntity):
        (WebCore::HTMLNames::legalEntityFor):
        (WebCore::HTML5Lexer::consumeEntity):
        (WebCore::HTML5Lexer::processEntity):
        (WebCore::HTML5Lexer::nextToken):
        (WebCore::HTML5Lexer::emitCodePoint):
        * html/HTML5Lexer.h:
2010-05-28  Adam Barth  <abarth@webkit.org>

        Reviewed by Eric Seidel.

        Handle edge cases in HTML5 entity parsing
        https://bugs.webkit.org/show_bug.cgi?id=39823

        Tests a bunch of the edge cases of entity handling in the HTML5
        specification.

        * html5lib/resources/entities01.dat:
        * html5lib/runner-expected.txt:


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@60369 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      fc8e393e
  11. 27 May, 2010 1 commit
    • eric@webkit.org's avatar
      2010-05-27 Eric Seidel <eric@webkit.org> · 3f0d7bf8
      eric@webkit.org authored 
              Reviewed by Adam Barth.

        Add <pre>/<listing> hack to HTML5Lexer to fix the last remaining HTML5 test suite regressions
        https://bugs.webkit.org/show_bug.cgi?id=39818

        * html5lib/runner-expected-html5.txt:
2010-05-27  Eric Seidel  <eric@webkit.org>

        Reviewed by Adam Barth.

        Add <pre>/<listing> hack to HTML5Lexer to fix the last remaining HTML5 test suite regressions
        https://bugs.webkit.org/show_bug.cgi?id=39818

        HTML parsers are supposed to ignore the first \n after a <pre> or <listing> tag
        for authoring convenience.  Our new HTML5Lexer didn't have this hack yet
        so there were 4 HTML5 tests failing.  Fixing this fixed the last of the HTML5
        test suite regressions using the HTML5Lexer vs the old lexer.

        * html/HTML5Lexer.cpp:
        (WebCore::HTML5Lexer::reset):
        (WebCore::HTML5Lexer::nextToken):
        * html/HTML5Lexer.h:
        (WebCore::HTML5Lexer::skipLeading...
      3f0d7bf8
  13. 26 May, 2010 2 commits
    • abarth@webkit.org's avatar
      2010-05-26 Adam Barth <abarth@webkit.org> · 76e97fff
      abarth@webkit.org authored 
              Reviewed by Eric Seidel.

        Fix webkit01.dat resumer tests in HTML5 parser
        https://bugs.webkit.org/show_bug.cgi?id=39796

        Update expectations to show massive passage.

        * html5lib/webkit-resumer-expected-html5.txt:
2010-05-26  Nate Chapin  <japhet@chromium.org>

        Reviewed by Adam Barth.

        Factor PageCache functionality out of FrameLoader and into
        PageCache.

        https://bugs.webkit.org/show_bug.cgi?id=39382

        Refactor only, so no new tests.

        * history/PageCache.cpp:
        (WebCore::pageCacheLogPrefix):
        (WebCore::pageCacheLog):
        (WebCore::logCanCacheFrameDecision):
        (WebCore::logCanCachePageDecision):
        (WebCore::PageCache::canCachePageContainingThisFrame):
        (WebCore::PageCache::canCache):
        (WebCore::PageCache::add):
        (WebCore::PageCache::get):
        * history/PageCache.h:
        * loader/DocumentLoader.cpp:
        * loader/FrameLoader.cpp:
        (WebCore::FrameLoader::commitProvisionalLoad):
        (WebCore::FrameLoader::prepareForCachedPageRestore):
        (WebCore::FrameLoader::continueLoadAfterNavigationPolicy):
        (WebCore::FrameLoader::loadProvisionalItemFromCachedPage):
        (WebCore::FrameLoader::navigateToDifferentDocument):
        * loader/FrameLoader.h:
        (WebCore::FrameLoader::quickRedirectComing):
        * svg/graphics/SVGImage.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@60257 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      76e97fff
    • abarth@webkit.org's avatar
      2010-05-26 Adam Barth <abarth@webkit.org> · 656187c0
      abarth@webkit.org authored 
              Reviewed by Eric Seidel.

        Make HTML5 lexer not ASSERT when resuming partial parses
        https://bugs.webkit.org/show_bug.cgi?id=39755

        Add a test suite for partial parsing.  This test runs all our parsing
        test cases, but stops and starts the parser at every character in the
        parse stream to make sure we resume parsing properly.  Currently, a
        bunch of the test cases are commented out, but I'll comment them back
        in as I get them running.

        * html5lib/webkit-resumer-expected-html5.txt: Added.
        * html5lib/webkit-resumer-expected.txt: Added.
        * html5lib/webkit-resumer.html: Added.
2010-05-26  Adam Barth  <abarth@webkit.org>

        Reviewed by Eric Seidel.

        Make HTML5 lexer not ASSERT when resuming partial parses
        https://bugs.webkit.org/show_bug.cgi?id=39755

        I'm working through a variation of the webkit-runner.html test suite
        that stops the parser at every character to make sure we can resume
        parsing correctly.  This patch fixes some errors caught by ASSERTs,
        which prevent the basic tests from running to completion.  There's a
        bunch more work to do, however.

        Test: html5lib/webkit-resumer.html

        * html/HTML5Lexer.cpp:
        (WebCore::HTMLNames::isEndTagBufferingState):
        (WebCore::HTML5Lexer::nextToken):
        (WebCore::HTML5Lexer::addToPossibleEndTag):
        * html/HTML5Lexer.h:
        * html/HTML5Tokenizer.cpp:
        (WebCore::HTML5Tokenizer::write):
        * html/HTML5Tokenizer.h:
2010-05-26  Adam Barth  <abarth@webkit.org>

        Reviewed by Eric Seidel.

        Make HTML5 lexer not ASSERT when resuming partial parses
        https://bugs.webkit.org/show_bug.cgi?id=39755

        Add webkit-resumer.html to the HTML5 parser test suite.

        * Scripts/test-html5-parser:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@60253 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      656187c0
  15. 25 May, 2010 2 commits
    • abarth@webkit.org's avatar
      2010-05-25 Adam Barth <abarth@webkit.org> · 7ad367da
      abarth@webkit.org authored 
              Unreviewed.  Fix typo pointed out by Maciej.

        * html/HTML5Lexer.cpp:
        (WebCore::HTMLNames::unconsumeCharacters):
        (WebCore::HTML5Lexer::consumeEntity):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@60139 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      7ad367da
    • abarth@webkit.org's avatar
      2010-05-24 Adam Barth <abarth@webkit.org> · ed1f7c11
      abarth@webkit.org authored 
              Reviewed by Eric Seidel.

        First cut at HTML5 entities
        https://bugs.webkit.org/show_bug.cgi?id=39649

        Update the expected results to show the new passing tests.  Yay!

        * html5lib/webkit-runner-expected-html5.txt:
2010-05-24  Adam Barth  <abarth@webkit.org>

        Reviewed by Eric Seidel.

        First cut at HTML5 entities
        https://bugs.webkit.org/show_bug.cgi?id=39649

        There's still a bunch more work to do to get our entity parsing fully
        up to spec.  This patch contains the bulk of the implementation
        however.

        The basics are covered by the existing html5lib tests.  I'll add more
        detailed tests in a followup patch.

        * html/HTML5Lexer.cpp:
        (WebCore::HTMLNames::legalEntityFor):
        (WebCore::HTMLNames::isHexDigit):
        (WebCore::HTMLNames::isAlphaNumeric):
        (WebCore::HTMLNames::uncomsumeCharacters):
        (WebCore::HTML5Lexer::consumeEntity):
        (WebCore::HTML5Lexer::nextToken):
        (WebCore::HTML5Lexer::haveBufferedCharacterToken):
        * html/HTML5Lexer.h:
        (WebCore::HTML5Lexer::):
        * html/HTML5Tokenizer.cpp:
        (WebCore::HTML5Tokenizer::write):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@60130 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      ed1f7c11
  17. 24 May, 2010 3 commits
    • abarth@webkit.org's avatar
      2010-05-24 Eric Seidel <eric@webkit.org> · 7279ee2f
      abarth@webkit.org authored 
              Reviewed by Adam Barth.

        Add RCDATA and RAWTEXT suport to the HTML5 parser
        https://bugs.webkit.org/show_bug.cgi?id=39642

        Adam Barth wrote half of this patch.

        * html/HTML5TreeBuilder.cpp:
        (WebCore::HTML5TreeBuilder::passTokenToLegacyParser):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@60116 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      7279ee2f
    • abarth@webkit.org's avatar
      2010-05-24 Adam Barth <abarth@webkit.org> · 91463047
      abarth@webkit.org authored 
              Reviewed by Eric Seidel.

        Teach the HTML5 parser how to lex escaped script data
        https://bugs.webkit.org/show_bug.cgi?id=39630

        Add a bunch of tests for parsing script data.  I think these tests
        drive the lexer into all the script data states.

        * html5lib/resources/scriptdata01.dat: Added.
        * html5lib/webkit-runner-expected-html5.txt:
        * html5lib/webkit-runner-expected.txt:
        * html5lib/webkit-runner.html:
2010-05-24  Adam Barth  <abarth@webkit.org>

        Reviewed by Eric Seidel.

        Teach the HTML5 parser how to lex escaped script data
        https://bugs.webkit.org/show_bug.cgi?id=39630

        * html/HTML5Lexer.cpp:
        (WebCore::HTMLNames::vectorEqualsString):
        (WebCore::HTML5Lexer::nextToken):
        (WebCore::HTML5Lexer::temporaryBufferIs):
        (WebCore::HTML5Lexer::isAppropriateEndTag):
        (WebCore::HTML5Lexer::maybeFlushBufferedEndTag):
        (WebCore::HTML5Lexer::flushBufferedEndTag):
        * html/HTML5Lexer.h:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@60109 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      91463047
    • abarth@webkit.org's avatar
      2010-05-24 Adam Barth <abarth@webkit.org> · 01ea231a
      abarth@webkit.org authored 
              Reviewed by Eric Seidel.

        Teach the HTML5 parser to lex DOCTYPEs
        https://bugs.webkit.org/show_bug.cgi?id=39571

        * html/HTML5Lexer.cpp:
        (WebCore::HTML5Lexer::nextToken):
        * html/HTML5Lexer.h:
        * html/HTML5Token.h:
        (WebCore::HTML5Token::beginDOCTYPE):
        (WebCore::HTML5Token::publicIdentifier):
        (WebCore::HTML5Token::systemIdentifier):
        (WebCore::HTML5Token::setPublicIdentifierToEmptyString):
        (WebCore::HTML5Token::setSystemIdentifierToEmptyString):
        (WebCore::HTML5Token::appendToPublicIdentifier):
        (WebCore::HTML5Token::appendToSystemIdentifier):
        (WebCore::HTML5Token::DoctypeData::DoctypeData):
        * html/HTML5TreeBuilder.cpp:
        (WebCore::convertToOldStyle):
        (WebCore::HTML5TreeBuilder::passTokenToLegacyParser):
        * platform/text/SegmentedString.h:
        (WebCore::SegmentedString::lookAheadIgnoringCase):
        (WebCore::SegmentedString::advanceAndASSERTIgnoringCase):

2010-05-24  Adam Barth  <abarth@webkit.org>

        Reviewed by Eric Seidel.

        Teach the HTML5 parser to lex DOCTYPEs
        https://bugs.webkit.org/show_bug.cgi?id=39571

        Add tests for DOCTYPEs.

        * html5lib/resources/doctype01.dat: Added.
        * html5lib/webkit-runner.html:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@60087 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      01ea231a
  19. 23 May, 2010 3 commits
    • eric@webkit.org's avatar
      2010-05-23 Adam Barth <abarth@webkit.org> · 1b4078bd
      eric@webkit.org authored 
              Reviewed by Eric Seidel.

        Teach HTML5 parser how to lex comments correctly
        https://bugs.webkit.org/show_bug.cgi?id=39537

        Added a new data file for testing comment parsing.  There are some more
        things we can test here once we implement the bogus comment state.

        * html5lib/resources/comments01.dat: Added.
        * html5lib/webkit-runner-expected.txt:
            - Some of these tests fail in our current parser because our
              current parser doesn't comply with HTML5.  I've set these
              expectations to failing so the tests will "pass" in the current
              parser.  I've double-checked the results in comments01.dat with a
              nightly build of Firefox to make sure we agree on our
              implementation of the spec.
        * html5lib/webkit-runner.html:
2010-05-23  Adam Barth  <abarth@webkit.org>

        Reviewed by Eric Seidel.

        Teach HTML5 parser how to lex commen...
      1b4078bd
    • eric@webkit.org's avatar
      2010-05-23 Adam Barth <abarth@webkit.org> · 82b4d03a
      eric@webkit.org authored 
              Reviewed by Eric Seidel.

        Make the html5lib parser tests run using the HTML5 parser
        https://bugs.webkit.org/show_bug.cgi?id=39533

        This patch adds a version of runner.html that's easier to parse (e.g.,
        no doctype or HTML comments) as well as a new test data file.  The
        exciting thing is that this test passes in the new HTML5 parser using
        this runner.

        Once the HTML5 parser is strong enough, we'll remove webkit-runner.html
        and have runner.html run the new webkit*.dat files.

        * html5lib/resources/webkit01.dat: Added.
        * html5lib/webkit-runner-expected.txt: Added.
        * html5lib/webkit-runner.html: Added.
2010-05-23  Adam Barth  <abarth@webkit.org>

        Reviewed by Eric Seidel.

        Make the html5lib parser tests run using the HTML5 parser
        https://bugs.webkit.org/show_bug.cgi?id=39533

        Turns out we're just one small tweak away from running the html5lib
        tests.  After this patch lands, we should be able to introduce tests
        for every change to the HTML5 parser.

        Test: html5lib/webkit-runner.html

        * html/HTML5Lexer.cpp:
        (WebCore::HTML5Lexer::nextToken):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@60048 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      82b4d03a
    • eric@webkit.org's avatar
      2010-05-23 Adam Barth <abarth@webkit.org> · 38adf9ce
      eric@webkit.org authored 
              Reviewed by Eric Seidel.

        Make HTML5 parser parse <script>alert("<hi>");</script>
        https://bugs.webkit.org/show_bug.cgi?id=39525

        More progress towards getting the html5lib test harness running.  We
        need to parse script data in the ScriptDataState.  The spec says the
        parser should set the lexer's state to ScriptDataState, but we want to
        keep using the existing parser for a while so I've added some temporary
        logic to HTML5Tokenizer to set the state in the lexer.

        To get this case running, I needed to flesh out a bunch of the
        ScriptDataState logic.  There's a bunch more to do, but this is a
        start.

        I'm not 100% happy with this patch, but it lets us execute a simplified
        version of the test harness up to the point where it starts running the
        first test.

        * html/HTML5Lexer.cpp:
        (WebCore::HTML5Lexer::nextToken):
        (WebCore::HTML5Lexer::isAppropriateEndTag):
        (WebCore::HTML5Lexer::emitCurrentTagToken):
        * html/HTML5Lexer.h:
        (WebCore::HTML5Lexer::setState):
        * html/HTML5Token.h:
        (WebCore::HTML5Token::beginEndTag):
        (WebCore::HTML5Token::appendToCharacter):
        (WebCore::HTML5Token::name):
        * html/HTML5Tokenizer.cpp:
        (WebCore::HTML5Tokenizer::write):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@60047 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      38adf9ce
  21. 22 May, 2010 1 commit
    • abarth@webkit.org's avatar
      2010-05-21 Adam Barth <abarth@webkit.org> · 3454dd30
      abarth@webkit.org authored 
              Reviewed by Eric Seidel.

        Parse attributes with HTML5 lexer
        https://bugs.webkit.org/show_bug.cgi?id=39520

        This patch lets the HTML5 lexer parse attributes, as in
        <div attr="foo"></div>.  This patch is on the critical path to running
        the html5lib tests.

        * html/HTML5Lexer.cpp:
        (WebCore::HTML5Lexer::nextToken):
        * html/HTML5Lexer.h:
        * html/HTML5Token.h:
        (WebCore::HTML5Token::beginStartTag):
        (WebCore::HTML5Token::beginEndTag):
        (WebCore::HTML5Token::addNewAttribute):
        (WebCore::HTML5Token::appendToAttributeName):
        (WebCore::HTML5Token::appendToAttributeValue):
        (WebCore::HTML5Token::attributes):
        * html/HTML5Tokenizer.cpp:
        (WebCore::convertToOldStyle):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@60001 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      3454dd30
  23. 19 May, 2010 2 commits
    • abarth@webkit.org's avatar
      2010-05-19 Adam Barth <abarth@webkit.org> · c0b247de
      abarth@webkit.org authored 
              Reviewed by David Hyatt.

        Change the API of nextToken to distinguish between emitting a token and just running out of input
        https://bugs.webkit.org/show_bug.cgi?id=39349

        I'm not sure there's a functional difference with this patch yet, but
        as discussed with Eric, we need this to handle attributes.

        * html/HTML5Lexer.cpp:
        (WebCore::HTML5Lexer::reset):
        (WebCore::HTML5Lexer::nextToken):
        * html/HTML5Lexer.h:
        * html/HTML5Token.h:
        (WebCore::HTML5Token::HTML5Token):
        (WebCore::HTML5Token::clear):
        (WebCore::HTML5Token::beginStartTag):
        (WebCore::HTML5Token::beginEndTag):
        (WebCore::HTML5Token::beginCharacter):
        * html/HTML5Tokenizer.cpp:
        (WebCore::HTML5Tokenizer::write):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@59750 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      c0b247de
    • abarth@webkit.org's avatar
      2010-05-18 Adam Barth <abarth@webkit.org> · c2f2b026
      abarth@webkit.org authored 
              Reviewed by Eric Seidel.

        Make the HTML5 parser correctly parse <div>Hello</div>
        https://bugs.webkit.org/show_bug.cgi?id=39345

        Again, this patch is covered by a large number of LayoutTests.

        * html/HTML5Lexer.cpp:
        (WebCore::HTML5Lexer::HTML5Lexer):
        (WebCore::HTML5Lexer::reset):
        (WebCore::HTML5Lexer::nextToken):
        (WebCore::HTML5Lexer::emitCommentToken):
        (WebCore::HTML5Lexer::emitCharacter):
        (WebCore::HTML5Lexer::emitCurrentTagToken):
        * html/HTML5Lexer.h:
        * html/HTML5Token.h:
        (WebCore::HTML5Token::beginCharacter):
        (WebCore::HTML5Token::appendToCharacter):
        (WebCore::HTML5Token::characters):
        * html/HTML5Tokenizer.cpp:
        (WebCore::convertToOldStyle):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@59747 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      c2f2b026
  25. 18 May, 2010 9 commits
  27. 17 May, 2010 3 commits