The fix for the bug 41175 is innocuous; however, the layout test added is failing in debug-mode WebKit on 10.5. Add an exclusion for that layout test, until a proper fix is developed.
Created attachment 62255 [details] Add an exclusion for the layout test for 41175 under OS X 10.5, until the real problem is fixed
Comment on attachment 62255 [details] Add an exclusion for the layout test for 41175 under OS X 10.5, until the real problem is fixed Every change requires a ChangeLog entry. :(
Created attachment 62325 [details] Excluding the layout test for 41175; including the ChangeLog entry
Comment on attachment 62325 [details] Excluding the layout test for 41175; including the ChangeLog entry https://bugs.webkit.org/show_bug.cgi?id=41808 does not explain the bug at all. It talks about Chromium, yet this is disabling a test on Mac. I do not understand. Please file a bug about the ASSERT and link to it from your bug and skipped list. Or re-purpose https://bugs.webkit.org/show_bug.cgi?id=41808 to be about the ASSERT itself (with information about what assert and a reduced test case, etc.) This patch is fine, but the course forward is not yet.
Comment on attachment 62325 [details] Excluding the layout test for 41175; including the ChangeLog entry This will disable the test for both Leopard and Tiger. But I suppose it’s better than having the bot fail for days.
Comment on attachment 62325 [details] Excluding the layout test for 41175; including the ChangeLog entry Clearing flags on attachment: 62325 Committed r63935: <http://trac.webkit.org/changeset/63935>
All reviewed patches have been landed. Closing bug.
The bug 43069 has been opened. That will address the issue of filtering zero-sized <g> elements at a higher, platform-independent level, as opposed to using platform-specific fixes and workarounds. That should hopefully resolve this issue as well.