[chromium] hebrew vowel marks incorrectly positioned
Created attachment 47578 [details] patch
agl: the meat of this change is the HarfbuzzSkia change -- it turns out, confusingly, that metrics->x is the offset and metrics->xOffset is the advance.
LGTM
Comment on attachment 47578 [details] patch Note this is lacking a Mac baseline. I will make one today, but if someone could review the code I'd appreciate it.
Comment on attachment 47578 [details] patch Please add results before landing.
Evan isn't a committer (or is he?) I assume he's supposed to post a new patch?
This landed in r54020. Not sure why the bug didn't close.
Checked in new expected results for Windows w/ r54209. Is this expected (font issue)? Or should I file a new bug about needing different results?
Thanks for checking, and sorry for the additional work. I would expect this to differ slightly between Mac and Windows due to different fonts being available. The real value of the test is as a pixel test; I tried to make the test text make testing pass/fail possible for people who can't read Hebrew. I added this test because the code change that accompanied it fixed the test; I believe it already passed on the other WebKit ports.
No problem, just wanted to make sure I wasn't masking a bug.(In reply to comment #9) > Thanks for checking, and sorry for the additional work. I would expect this to > differ slightly between Mac and Windows due to different fonts being available. > The real value of the test is as a pixel test; I tried to make the test text > make testing pass/fail possible for people who can't read Hebrew. > > I added this test because the code change that accompanied it fixed the test; I > believe it already passed on the other WebKit ports. No problem, just wanted to make sure I wasn't masking over a bug and needing to file a new one. I think there is one about different fonts on Mac and Windows, which is why there are a good number of tests on the Windows Skipped list.