2010-06-15 Adam Barth <abarth@webkit.org>

2010-06-15  Adam Barth  <abarth@webkit.org>

        Reviewed by Eric Seidel.

        Enable HTML5 lexer
        https://bugs.webkit.org/show_bug.cgi?id=40650

        * page/Settings.cpp:
        (WebCore::Settings::Settings):
2010-06-15  Adam Barth  <abarth@webkit.org>

        Reviewed by Eric Seidel.

        Enable HTML5 lexer
        https://bugs.webkit.org/show_bug.cgi?id=40650

        This patch updates the expected results for tests that conflict with
        the HTML5 specification.  In a later patch, I'll update the tests
        themselves to fix pass/fail messages.

        * fast/doctypes/005-case-preserving-expected.txt:
            - HTML5 specifies case normalization for DOCTYPEs.
        * fast/js/missing-title-end-tag-js-expected.txt:
            - HTML5 requires us to swallow the entire document when <title> is
              unterminated.
        * fast/parser/comment-in-iframe-expected.txt:
        * fast/parser/comment-in-script-tricky-expected.txt:
        * fast/parser/comment-in-title-expected.txt:
        * fast/parser/comments-expected.txt:
            - Differences due to HTML5 comment parsing.
        * fast/parser/eightdigithexentity-expected.txt:
            - HTML5 specifies a different behavior for &#x00FFFFFF;
        * fast/parser/entity-end-iframe-tag-expected.txt:
            - HTML5 specifies a different behavior for &lt;/iframe>
        * fast/parser/entity-surrogate-pairs-expected.txt:
            - HTML5 doesn't allow entities to create surrogate pairs.
        * fast/parser/html-whitespace-expected.txt:
            - HTML5 has a different handling of CR in this case.
        * fast/parser/script-tag-with-trailing-slash-expected.txt:
            - HTML5 does not allow self-closing <script/> tags.
        * fast/parser/tag-with-exclamation-point-expected.txt:
            - HTML5 treats bogus DOCTYPEs as comments.
        * fast/parser/xml-directive-in-dom-expected.txt:
            - HTML5 attaches <?xml?> processing directives to the DOM as
              comments.
        * fast/xpath/xpath-functional-test-expected.txt:
            - I don't fully understand this test, but our new behavior matches
              the Firefox nightly builds (which have an HTML5 parser).
        * html5lib/runner-expected.txt:
        * html5lib/webkit-resumer-expected.txt:
            - Massive progressions on HTML5 conformance.
        * http/tests/loading/gmail-assert-on-load-expected.txt:
            - I don't fully understand this behavior change.  It might
              represent a bug.  We're still investigating.
        * http/tests/security/xssAuditor/embed-tag-null-char-expected.txt:
            - Update expected results for change in NULL char handling.
        * http/tests/security/xssAuditor/img-onerror-tricky-expected.txt:
            - This exploit no longer works in the HTML5 parser, so it's not
              blocked.
        * http/tests/security/xssAuditor/javascript-link-HTML-entities-null-char-expected.txt:
            - This a real regression.  We need to update the XSSAuditor to use
              the HTML5 parser's HTML entity decoder.  I'll do that in a
              followup patch because there doesn't appear to be an easy way to
              key the choice off of the WebCore::Setting.
        * http/tests/security/xssAuditor/malformed-HTML-expected.txt:
            - Slightly different handling of malformed HTML.  The exploit is still blocked.
        * http/tests/security/xssAuditor/object-embed-tag-null-char-expected.txt:
            - Update expected result due to different NULL character handling.
        * inspector/audits-panel-functional-expected.txt:
        * inspector/timeline-script-tag-1-expected.txt:
            - Slightly different semantics around document.write.
        * inspector/elements-panel-structure-expected.txt:
            - HTML5 specifies case normalization for DOCTYPEs.
        * platform/mac/css1/box_properties/float_elements_in_series-expected.txt:
            - HTML5 specifies different handling of "<foo<bar"
        * platform/mac/fast/doctypes/003-expected.txt:
            - HTML5 specifies case normalization for DOCTYPEs.
        * platform/mac/fast/dom/stripNullFromTextNodes-expected.txt:
            - HTML5 specifies different handling of NULL characters.
        * platform/mac/fast/invalid/016-expected.txt:
            - HTML5 specifies different handling of "<foo<bar"
        * platform/mac/fast/parser/broken-comments-vs-parsing-mode-expected.txt:
            - HTML5 specifies different handling of HTML comments.
        * platform/mac/fast/parser/comment-in-style-expected.txt:
            - HTML5 specifies different handling of HTML comments.
        * platform/mac/fast/parser/parseCommentsInTitles-expected.txt:
            - HTML5 specifies different handling of HTML comments.
        * platform/mac/fast/text/international/bidi-linebreak-001-expected.txt:
        * platform/mac/fast/text/international/bidi-linebreak-002-expected.txt:
        * platform/mac/fast/text/international/bidi-linebreak-003-expected.txt:
        * platform/mac/fast/text/international/hindi-whitespace-expected.txt:
            - HTML5 specifies different handling CR.  I don't fully understand
              the consequences of this change, but our new behavior matches the
              Firefox nightly.
        * platform/mac/fast/text/stripNullFromText-expected.txt:
            - HTML5 specifies different handling of NULL characters.
        * platform/mac/fast/tokenizer/missing-title-end-tag-1-expected.txt:
        * platform/mac/fast/tokenizer/missing-title-end-tag-2-expected.txt:
            - HTML5 requires us to swallow the entire document when <title> is
              unterminated.
        * platform/mac/fast/tokenizer/script_extra_close-expected.txt:
            - We're coalescing text nodes slightly different in the new parser.
              We're going to get this up to spec when we work on the
              TreeConstructor (which is where the spec handles text node
              coalescing).
        * platform/mac/tables/mozilla/images/adforce_imgis_com-expected.txt:
            - Different handling of crazy unicode characters (replaced with
              FFFD).  Still investigating whether this is a real bug.
        * webarchive/archive-empty-frame-dom-expected.webarchive:
            - HTML5 specifies case normalization for DOCTYPEs.
2010-06-15  Adam Barth  <abarth@webkit.org>

        Reviewed by Eric Seidel.

        Enable HTML5 lexer
        https://bugs.webkit.org/show_bug.cgi?id=40650

        * DumpRenderTree/mac/DumpRenderTree.mm:
        (initializeGlobalsFromCommandLineOptions):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@61234 268f45cc-cd09-0410-ab3c-d52691b4dbfc