The current set of layout tests for ruby rendering in fast/ruby uses Japanese for the example sentences, since Japanese Furigana (reading annotations) are a main use case for <ruby>. However, this apparently causes problems on non-Mac platforms, as evidenced by feedback on the mailing list and the fact that most (all?) other layout tests that use non-Latin characters are listed in platform/win/Skipped (for the Windows platform). Since there is no fundamental reason to use Japanese characters for the ruby layout tests, re-doing them to use Latin characters seems prudent.
Created attachment 43832 [details] patch - re-do ruby layout tests New versions of the ruby layout tests that use English text rather than Japanese, includes the pixel test results for Mac.
I suspect that getting the right fonts on windows is probably a pain.
Comment on attachment 43832 [details] patch - re-do ruby layout tests Looks OK to me.
Note: still need to create Windows baselines before committing, in order to avoid having a mixed state with old and new result expectations (my Windows machine for that must arrive Real Soon Now... :p)
I see, so this bug is ready to land, just waiting on additional windows baselines from you? Thus I assume it cannot be commit-queue'd?
(In reply to comment #5) Yes - I received my XP machine yesterday, but I'm not sure I get to do the Windows baselines before my vacation.
Created attachment 45032 [details] patch - base-lines for Windows
style-queue ran check-webkit-style on attachment 45032 [details] without any errors.
So should these now be landed?
Yes, but I'm on vacation ATM, and have no easy source access. Could you perhaps land them for me? (Or I can also land them after I come back - these are hardly time-critical patches.)
Comment on attachment 45032 [details] patch - base-lines for Windows The tab in the ChangeLog will prevent this from being cq'd.
Ping? This bug has been in the pending-commit list a long time, and blocks bug 28420 which has been in the pending-commit list even longer. It's time to take action on these patches or mark them r-.
Committed as rev. 53316
First patch landed as rev. 53605 (was under the wrong impression that this had already landed before)