Currently the CSS parser calls to the Document and StyleSheetInternal objects to get settings, base URL etc. The required information should be passed in on parser construction instead. The parser should not need to know about Document at all.
Created attachment 137237 [details] patch
Created attachment 137238 [details] rebased
Attachment 137238 [details] did not pass style-queue: Failed to run "['Tools/Scripts/check-webkit-style', '--diff-files', u'Source/WebCore/ChangeLog', u'Source/WebCor..." exit_code: 1 Source/WebCore/css/CSSParser.h:329: The parameter name "context" adds no information, so it should be removed. [readability/parameter_name] [5] Total errors found: 1 in 23 files If any of these errors are false positives, please file a bug against check-webkit-style.
Comment on attachment 137238 [details] rebased Attachment 137238 [details] did not pass chromium-ews (chromium-xvfb): Output: http://queues.webkit.org/results/12405802
Created attachment 137239 [details] try to fix chromium build
Comment on attachment 137239 [details] try to fix chromium build Attachment 137239 [details] did not pass chromium-ews (chromium-xvfb): Output: http://queues.webkit.org/results/12406713 New failing tests: fast/regions/css-regions-disabled.html css3/filters/custom/custom-filter-property-parsing.html
Created attachment 137240 [details] Archive of layout-test-results from ec2-cr-linux-03 The attached test failures were seen while running run-webkit-tests on the chromium-ews. Bot: ec2-cr-linux-03 Port: <class 'webkitpy.common.config.ports.ChromiumXVFBPort'> Platform: Linux-2.6.35-28-virtual-x86_64-with-Ubuntu-10.10-maverick
Created attachment 137242 [details] modify the failing tests to not expect settings to change the behavior of the existing stylesheets
Are there any cases where we keep a CSSParser object around for a while? Are we sure the baseURL of the Document can never change in the lifetime of the parser?
(In reply to comment #9) > Are there any cases where we keep a CSSParser object around for a while? No. It is always stack allocated and thrown away after parsing finishes. > Are we sure the baseURL of the Document can never change in the lifetime of the parser? Pretty sure.
Comment on attachment 137242 [details] modify the failing tests to not expect settings to change the behavior of the existing stylesheets r=me.
http://trac.webkit.org/changeset/114217
Is this in preparation for moving CSSParsing off into another thread? Just curious, as we've discussed such for the HTML parser.
(In reply to comment #13) > Is this in preparation for moving CSSParsing off into another thread? Just curious, as we've discussed such for the HTML parser. This is prep work for sharable style sheets (see the meta bug I just added as a dependent.) Moving CSS parsing to a secondary thread is certainly made easier by this effort, but not currently a goal in itself.
(In reply to comment #12) > http://trac.webkit.org/changeset/114217 It seems after the commit fast/repaint/line-flow-with-floats-in-regions.html layout test fails on Qt Release bots http://build.webkit.org/results/Qt%20Linux%20Release/r114217%20%2845824%29/fast/repaint/line-flow-with-floats-in-regions-pretty-diff.html