The following tests are still failing on Mac Chromium. [ Mac ]fast/lists/003-vertical.html [Failure] [ Mac ]fast/text/international/vertical-text-glyph-test.html [Failure] [ Mac ]fast/text/shaping/shaping-selection-rect.html [Failure] even after disabling the optimization for HarfBuzz. Updated TextExpectations.
*** Bug 98867 has been marked as a duplicate of this bug. ***
Enrica and I think that this is related to Chromium’s use of HarfBuzz on Mac. It’s also possible, though seems unlikely, that it’s related to sub-pixel layout.
If the tests are still failing, should we re-enable the optimization for harfbuzz-ng platforms? Maybe bashi has time to look into these failures.
This isn't a sub-pixel issue. The differences here are in the float values in the line box tree. My money is on HarfBuzz as well :) Thanks for helping us track this.
I think this is a good plan. To re-enable the optimization, all that needs to be done is remove the #if USE(HARFBUZZ_NG) and the remove the UNUSED_PARAMETER at the beginning of setLogicalWidthForTextRun
Created attachment 168109 [details] Patch
Comment on attachment 168109 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=168109&action=review > Source/WebCore/rendering/RenderBlockLineLayout.cpp:715 > +#if !(PLATFORM(CHROMIUM) && OS(DARWIN)) I think you want this to be #if (PLATFORM(CHROMIUM) && OS(DARWIN)) otherwise you get a compilation error
Comment on attachment 168109 [details] Patch Clearing flags on attachment: 168109 Committed r131000: <http://trac.webkit.org/changeset/131000>
All reviewed patches have been landed. Closing bug.