There are also the ones under css-transforms/animation/ I expect that at least some of them are due to spec changes that occurred after CJ's patches were written (e.g. making `scale: 2` affect both x and y).

(In reply to Brian Birtles (:birtles) from comment #1) > There are also the ones under css-transforms/animation/ > > I expect that at least some of them are due to spec changes that occurred > after CJ's patches were written (e.g. making `scale: 2` affect both x and y). Yes. After changing "scale: 2;" to "scale: 2 2;", we could pass this wpt reftest.

The current spec says: "If only the X value is given, the Y value defaults to the same value.", so we should update the behavior.

try: https://treeherder.mozilla.org/#/jobs?repo=try&revision=bc24953c420e72fe2633d462de3e74dc3d249111

I filed a spec issue if we want to serialize the scale property with only one <number> [1]. If `scale` is set as "2", Chrome serializes this as "2 2", but Firefox serializes this as "2". However, I think the shorter one is preferred. Therefore, I will keep our serialization the same as before. [1] https://github.com/w3c/fxtf-drafts/issues/313.

We should then probably use Scale(Number, Number) instead of Scale(Number, Option<Number>), and implement ToCss. Otherwise we'd serialize "scale: 2 2" as "scale: 2 2" instead of "scale: 2".

(In reply to Emilio Cobos Álvarez (:emilio) from comment #6) > We should then probably use Scale(Number, Number) instead of Scale(Number, > Option<Number>), and implement ToCss. > > Otherwise we'd serialize "scale: 2 2" as "scale: 2 2" instead of "scale: 2". Thanks. In conclusion: "scale: 2" => "scale: 2" "scale: 2 2" => "scale: 2" "scale: 2 2 1" => "scale: 2 2 1" And I think I need to update "Animate for ComputedScale" to get the correct enum value, instead of always `Scale3D`. BTW, tabatkins filed the chrome bug for the bad serialization: https://bugs.chromium.org/p/chromium/issues/detail?id=901509

(In reply to Boris Chiou [:boris] (UTC-7) from comment #5) > I filed a spec issue if we want to serialize the scale property with only > one <number> [1]. > If `scale` is set as "2", Chrome serializes this as "2 2", but Firefox > serializes this as "2". However, I think the shorter one is preferred. > Therefore, I will keep our serialization the same as before. > > [1] https://github.com/w3c/fxtf-drafts/issues/313. This may also fix Bug 1464791.

(In reply to Boris Chiou [:boris] (UTC-7) from comment #9) > This may also fix Bug 1464791. I decide not to fix Bug 1464791 here.

try: https://treeherder.mozilla.org/#/jobs?repo=try&revision=05af6785f5e1b8a8fdaa97a8a08e8b8a40728124

About the serialization: https://drafts.csswg.org/css-transforms-2/#individual-transform-serialization

OOps, there are so many UNEXPECTED_PASS in css-transform/animation/ [1]. Will update this patch later. (Good news.) [1] https://treeherder.mozilla.org/#/jobs?repo=try&revision=bb922ecbf75d7ef155a61c53af0148e007134f2f&selectedJob=209958673

(In reply to Boris Chiou [:boris] (UTC-7) from comment #13) > OOps, there are so many UNEXPECTED_PASS in css-transform/animation/ [1]. > Will update this patch later. (Good news.) > > [1] > https://treeherder.mozilla.org/#/ > jobs?repo=try&revision=bb922ecbf75d7ef155a61c53af0148e007134f2f&selectedJob=2 > 09958673 Which means those tests should be passed because we didn't enable the preference of individual-transform for all the css-transform in wpt before applying my patch.

Some testcases confuse me, especially for animation [1]. The spec says we should convert the `scale: none` value into `scale: 1`, for animations and transitions, but the wpt doesn't match? I filed an spec issue [2] for this. [1] https://searchfox.org/mozilla-central/rev/b096dcf0ea226af628fe03f7e7acb56a25853533/testing/web-platform/tests/css/css-transforms/animation/scale-interpolation.html#41-52 [2] https://github.com/w3c/fxtf-drafts/issues/315

Pushed by boris.chiou@gmail.com: https://hg.mozilla.org/integration/autoland/rev/ffb8562ce45d Fix the default behavior of scale:<number>{1}. r=birtles

Created web-platform-tests PR https://github.com/web-platform-tests/wpt/pull/13961 for changes under testing/web-platform/tests

Upstream web-platform-tests status checks passed, PR will merge once commit reaches central.

web-platform-tests PR at https://github.com/web-platform-tests/wpt/pull/13974

Upstream PR merged