The test bidi-004.html was removed yesterday from the CSS 2.1 test suite, as it was deemed wrong. WebKit should have the new version of the test, so that we don't get more comments like https://bugs.webkit.org/show_bug.cgi?id=57459#c25.
Created attachment 90335 [details] Patch.
Created attachment 90336 [details] Patch. One more try :)
Created attachment 90337 [details] Patch.
Comment on attachment 90337 [details] Patch. Not sure why removing images is causing a problem. Will commit manually after review.
Comment on attachment 90337 [details] Patch. Your local checkout is not correctly attributing pngs as binary.
(At least I believe that's what happened to your diff.)
I'm a bit confused with all this. Can you point me to the decision where bidi-04 was deemed wrong? Can you explain what your new bidi-04 does?
(In reply to comment #7) > I'm a bit confused with all this. Can you point me to the decision where bidi-04 was deemed wrong? Can you explain what your new bidi-04 does? After a discussion on IRC last night, Hixie said he was going to send an e-mail to www-style about bidi-004.xht and bidi-004a.xht being wrong. I do not see that e-mail, so it probably did not happen yet. I created this bug based on Hixie's comment and the test he created in http://www.hixie.ch/tests/adhoc/css/box/inline/bidi/012.html .
(In reply to comment #5) > (From update of attachment 90337 [details]) > Your local checkout is not correctly attributing pngs as binary. I set svn propset svn:mime-type on these files, but that did not help. Anything else I should do? thanks,
(In reply to comment #7) > I'm a bit confused with all this. Can you point me to the decision where bidi-04 was deemed wrong? Can you explain what your new bidi-04 does? Yael and I talked to Hixie on IRC and discussed the change I had made to treat BR's as paragraph separators that don't break out of style/dom directionality, but only unicode control characters. That change prevented us from passing the old bidi-004.html test, and after showing it to him, he concurred that in fact the test was flawed.
Comment on attachment 90337 [details] Patch. It may not be your diff, it may be prettypatch which is causing the strageness.
Committed r84398: <http://trac.webkit.org/changeset/84398>
http://trac.webkit.org/changeset/84398 might have broken Windows 7 Release (Tests) The following tests are not passing: fast/borders/bidi-012.html