NEW 162815
☔ Implement variation fonts
https://bugs.webkit.org/show_bug.cgi?id=162815
Summary ☔ Implement variation fonts
Myles C. Maxfield
Reported 2016-09-30 17:34:28 PDT
Implement variation fonts
Attachments
Michael Catanzaro
Comment 1 2016-10-09 13:12:40 PDT
Hey Myles! So all variation fonts tests are failing on GTK since r206839 "Put variation fonts work behind a compile-time flag" which added an XCode-specific build setting; the setting doesn't exist for any other WebKit ports, though, so it's off by default everywhere else. Is this desired? Judging from bug #162818 I guess this feature is intended to be enabled on macOS; if it doesn't require any platform-specific implementation then I'm sure other ports would like to match that, right? We should probably add it in WebKitFeatures.cmake at any rate? In the meantime, I'm marking all tests expected to fail on GTK, against this bug (so please don't close this unless we either remove those expectations or open another bug).
Michael Catanzaro
Comment 2 2016-10-09 13:18:27 PDT
(In reply to comment #1) > Hey Myles! So all variation fonts tests are failing on GTK since r206839 > "Put variation fonts work behind a compile-time flag" which added an > XCode-specific build setting Well that's not true, only a few of them are failing: animations/font-variation-settings-unlike.html fast/text/variations/advances.html fast/text/variations/exist.html fast/text/variations/getComputedStyle.html Also we have two timeouts, which are probably bad even if the option is off: animations/font-variation-settings-order.html animations/font-variation-settings.html I'm skipping fast/text/variations, marking animations/font-variation-settings-unlike.html as expected fail, and marking the timeouts as expected timeouts, all against this bug. Feel free to move the expectations to different bug reports or whatever else.
Myles C. Maxfield
Comment 3 2016-10-10 09:58:13 PDT
(In reply to comment #1) > Hey Myles! So all variation fonts tests are failing on GTK since r206839 > "Put variation fonts work behind a compile-time flag" which added an > XCode-specific build setting; the setting doesn't exist for any other WebKit > ports, though, so it's off by default everywhere else. Is this desired? > Judging from bug #162818 I guess this feature is intended to be enabled on > macOS; if it doesn't require any platform-specific implementation then I'm > sure other ports would like to match that, right? We should probably add it > in WebKitFeatures.cmake at any rate? > > In the meantime, I'm marking all tests expected to fail on GTK, against this > bug (so please don't close this unless we either remove those expectations > or open another bug). Variation fonts support requires platform support. It doesn't seem to make any sense to me to turn on the feature in WebKit but not hook it up to the EFL/GTK platforms (which I don't know how to do). Of course, each port an make the decision of turning this feature on or not. The timeouts are particularly worrisome. Do you know why they are happening?
Myles C. Maxfield
Comment 4 2016-10-10 10:02:17 PDT
Clarification: for Cocoa-based ports, the platform support is hooked up, so it makes sense for it to be on by default there.
Michael Catanzaro
Comment 5 2016-10-10 10:30:09 PDT
(In reply to comment #3) > Variation fonts support requires platform support. It doesn't seem to make > any sense to me to turn on the feature in WebKit but not hook it up to the > EFL/GTK platforms (which I don't know how to do). Of course, each port an > make the decision of turning this feature on or not. OK, then the status quo seems correct, though I think that non-Mac-specific build flags should probably always be added to WebKitFeatures.cmake (and just set to OFF), and also FeatureList.pm. > The timeouts are particularly worrisome. Do you know why they are happening? Nope.
Miguel Gomez
Comment 6 2016-10-21 07:29:21 PDT
These two animations/font-variations/font-variation-settings-order.html animations/font-variations/font-variation-settings.html are not timing out anymore since r207642. They are now failing with ImageOnlyFailure. I'll update expectations accordingly.
Note You need to log in before you can comment on or make changes to this bug.