BogusCommentState should come in from the cold
Created attachment 60174 [details] Patch
Comment on attachment 60174 [details] Patch LayoutTests/html5lib/webkit-resumer-expected.txt:159 + resources/webkit01.dat: Geez, we still need to fix this output. WebCore/html/HTMLTokenizer.cpp:1078 + else if (cc == InputStreamPreprocessor::endOfFileMarker) I should have written these as isEndOfFileMarker(cc) Why do we need ContinueBogusCommentState? Why not just check m_token->type() == Comment?
> WebCore/html/HTMLTokenizer.cpp:1078 > + else if (cc == InputStreamPreprocessor::endOfFileMarker) > I should have written these as isEndOfFileMarker(cc) Oh, ok. I just copied another state/ > Why do we need ContinueBogusCommentState? Why not just check m_token->type() == Comment? Because that would be an extra branch per character.
(In reply to comment #3) > > Why do we need ContinueBogusCommentState? Why not just check m_token->type() == Comment? > > Because that would be an extra branch per character. Really? I meant in the BogusCommentState itself. Are we commonly in the BogusCommentState? I agree this is fewer branches.
Comment on attachment 60174 [details] Patch Clearing flags on attachment: 60174 Committed r62233: <http://trac.webkit.org/changeset/62233>
All reviewed patches have been landed. Closing bug.