See bug 932615. Given that we're going to be shipping SkiaGL on B2G, we need to find issues there faster -- let's turn on SkiaGL for Canvas on OSX right now, and deal with the fallout. (Optimistically assigning to gw280, but could go elsewhere.)

Let's let George finish the koi+ bugs first. Let's also prioritize all the items, so that we know where this fits in the list. If it's #1, we'll find somebody to do it.

Seems reasonable to me.

Also I can take this.

Any idea when this will happen now?

https://tbpl.mozilla.org/?tree=Graphics&rev=52733043ba3f Let's see how we look.

I thought bug 996611 was tracking content issues - this is just canvas.

Oh, oops. My bad!

Forgot to change a unit test in the try run mentioned in comment 7, this one is more relevant https://tbpl.mozilla.org/?tree=Try&rev=6e18f132df45

We skip this test with accelerated Skia: test_2d_gradient_radial_inside1

Looks like all the dependent bugs are fixed now. Anything else keeping us from flipping this pref?

We didn't enter bugs for all the failures: https://treeherder.mozilla.org/ui/#/jobs?repo=try&revision=fc75af631b63, there are still problems.

As of https://treeherder.mozilla.org/#/jobs?repo=try&revision=4cfa5824cb47 (SkiaGL): These are the failures: REFTEST TEST-UNEXPECTED-FAIL | file:///builds/slave/talos-slave/test/build/tests/reftest/tests/layout/reftests/w3c-css/submitted/variables/variable-supports-06.html | assertion count 1 is more than expected 0 assertions REFTEST TEST-UNEXPECTED-FAIL | file:///builds/slave/talos-slave/test/build/tests/reftest/tests/layout/reftests/border-radius/curved-stripe-border.html | image comparison (==), max difference: 4, number of differing pixels: 582 REFTEST TEST-UNEXPECTED-PASS | file:///builds/slave/talos-slave/test/build/tests/reftest/tests/layout/reftests/canvas/672646-alpha-radial-gradient.html | image comparison (==) REFTEST TEST-UNEXPECTED-FAIL | file:///builds/slave/talos-slave/test/build/tests/reftest/tests/layout/reftests/css-blending/background-blending-darken.html | assertion count 1 is more than expected 0 assertions REFTEST TEST-UNEXPECTED-FAIL | file:///builds/slave/talos-slave/test/build/tests/reftest/tests/layout/reftests/css-blending/background-blending-hard-light.html | image comparison (==), max difference: 1, number of differing pixels: 1600 REFTEST TEST-UNEXPECTED-FAIL | file:///builds/slave/talos-slave/test/build/tests/reftest/tests/layout/reftests/css-gradients/linear-diagonal-9a.html | image comparison (==), max difference: 4, number of differing pixels: 29437 REFTEST TEST-UNEXPECTED-FAIL | file:///builds/slave/talos-slave/test/build/tests/reftest/tests/layout/reftests/css-gradients/radial-1a.html | image comparison (==), max difference: 6, number of differing pixels: 5606 REFTEST TEST-UNEXPECTED-FAIL | file:///builds/slave/talos-slave/test/build/tests/reftest/tests/layout/reftests/css-gradients/radial-1b.html | image comparison (==), max difference: 6, number of differing pixels: 5606 REFTEST TEST-UNEXPECTED-FAIL | file:///builds/slave/talos-slave/test/build/tests/reftest/tests/layout/reftests/css-gradients/radial-1c.html | image comparison (==), max difference: 6, number of differing pixels: 5606 REFTEST TEST-UNEXPECTED-FAIL | file:///builds/slave/talos-slave/test/build/tests/reftest/tests/layout/reftests/css-gradients/radial-2a.html | image comparison (==), max difference: 3, number of differing pixels: 7056 REFTEST TEST-UNEXPECTED-FAIL | file:///builds/slave/talos-slave/test/build/tests/reftest/tests/layout/reftests/css-gradients/radial-2b.html | image comparison (==), max difference: 3, number of differing pixels: 7056 REFTEST TEST-UNEXPECTED-FAIL | file:///builds/slave/talos-slave/test/build/tests/reftest/tests/layout/reftests/css-gradients/radial-2c.html | image comparison (==), max difference: 3, number of differing pixels: 7056 REFTEST TEST-UNEXPECTED-FAIL | file:///builds/slave/talos-slave/test/build/tests/reftest/tests/layout/reftests/css-gradients/radial-2d.html | image comparison (==), max difference: 3, number of differing pixels: 7056 REFTEST TEST-UNEXPECTED-FAIL | file:///builds/slave/talos-slave/test/build/tests/reftest/tests/layout/reftests/css-gradients/radial-2e.html | image comparison (==), max difference: 3, number of differing pixels: 7056 REFTEST TEST-UNEXPECTED-FAIL | file:///builds/slave/talos-slave/test/build/tests/reftest/tests/layout/reftests/css-gradients/radial-2f.html | image comparison (==), max difference: 3, number of differing pixels: 7056 REFTEST TEST-UNEXPECTED-FAIL | file:///builds/slave/talos-slave/test/build/tests/reftest/tests/layout/reftests/css-gradients/radial-shape-closest-corner-1a.html | image comparison (==), max difference: 4, number of differing pixels: 22301 REFTEST TEST-UNEXPECTED-FAIL | file:///builds/slave/talos-slave/test/build/tests/reftest/tests/layout/reftests/css-gradients/radial-shape-closest-corner-1b.html | image comparison (==), max difference: 4, number of differing pixels: 22565 REFTEST TEST-UNEXPECTED-FAIL | file:///builds/slave/talos-slave/test/build/tests/reftest/tests/layout/reftests/css-gradients/radial-shape-farthest-corner-1b.html | image comparison (==), max difference: 2, number of differing pixels: 41281 REFTEST TEST-UNEXPECTED-FAIL | file:///builds/slave/talos-slave/test/build/tests/reftest/tests/layout/reftests/image-element/gradient-html-07a.html | image comparison (==), max difference: 2, number of differing pixels: 42305 REFTEST TEST-UNEXPECTED-FAIL | file:///builds/slave/talos-slave/test/build/tests/reftest/tests/layout/reftests/svg/as-image/canvas-drawImage-alpha-1.html | image comparison (==), max difference: 1, number of differing pixels: 10000 REFTEST TEST-UNEXPECTED-FAIL | file:///builds/slave/talos-slave/test/build/tests/reftest/tests/layout/reftests/svg/as-image/canvas-drawImage-alpha-2.html | image comparison (==), max difference: 1, number of differing pixels: 676 REFTEST TEST-UNEXPECTED-FAIL | file:///builds/slave/talos-slave/test/build/tests/reftest/tests/layout/reftests/svg/filters/svg-filter-chains/clip-input.svg | assertion count 1 is more than expected 0 assertions REFTEST TEST-UNEXPECTED-FAIL | file:///builds/slave/talos-slave/test/build/tests/reftest/tests/layout/reftests/svg/non-scaling-stroke-01.svg | image comparison (==), max difference: 1, number of differing pixels: 122 REFTEST TEST-UNEXPECTED-FAIL | file:///builds/slave/talos-slave/test/build/tests/reftest/tests/layout/reftests/svg/opacity-and-gradient-02.svg | image comparison (==), max difference: 1, number of differing pixels: 99974 REFTEST TEST-UNEXPECTED-FAIL | file:///builds/slave/talos-slave/test/build/tests/reftest/tests/layout/reftests/svg/radialGradient-basic-03.svg | image comparison (==), max difference: 4, number of differing pixels: 15982 REFTEST TEST-UNEXPECTED-FAIL | file:///builds/slave/talos-slave/test/build/tests/reftest/tests/layout/reftests/text-svgglyphs/svg-glyph-objectgradient-zoom.svg | image comparison (==), max difference: 1, number of differing pixels: 7007

Tracking SkiaGL for Shumway canvas perf.

Skia software: https://treeherder.mozilla.org/#/jobs?repo=try&revision=3674bb1395fa SkiaGL: https://treeherder.mozilla.org/#/jobs?repo=try&revision=57ba4ba9eaa5

New try push with Skia software and updated reftest annotations: https://treeherder.mozilla.org/#/jobs?repo=try&revision=5c2af84f226f Wherever there was a fuzzy-if(azureQuartz,*,*) annotation that was needed also with Skia, I replaced azureQuartz with cocoaWidget. It's possible there are still failures in test_canvas.html; it doesn't run during reftests so it wasn't included in the last push.

Looks like I missed updating the annotation on non-scaling-stroke-01.svg in the previous push. This patch fixes that.

I filed bug 1150944 for skiagl. Let's keep this one for the skia landing.

/r/6663 - Bug 932958 - Use the Skia Software backend by default for canvas on OS X. r=jrmuizel Pull down this commit: hg pull -r 503c3c2d430f7a0812b99b9950dd822c3b93d184 https://reviewboard-hg.mozilla.org/gecko/

Comment on attachment 8588730 [details] MozReview Request: bz://932958/mstange https://reviewboard.mozilla.org/r/6661/#review5547 Ship It!

Backed out in https://hg.mozilla.org/integration/mozilla-inbound/rev/76df12cb260f because we apparently enjoy setting you up for failure. You missed a few 10.10 things from your try push that did include 10.8 and 10.10, https://treeherder.mozilla.org/logviewer.html#?job_id=8510877&repo=mozilla-inbound You also, unsurprisingly, missed realizing that we now spell reftest "reftest[10.6,10.8,10.10],web-platform-tests-reftests[10.10]" because we both run reftests (some of them just copy-pasted) in a separate suite with a different name, and we also do not run that suite on 10.6, so -a or any of its variants won't run it for you on try, only explicitly saying that you want 10.10 tests will.

Oops, silly of me, it was actually web-platform-tests-1 rather than web-platform-tests-reftests where you had some more bustage (https://treeherder.mozilla.org/logviewer.html#?job_id=8515689&repo=mozilla-inbound). Still a 10.10-only suite, but maybe in a way less painful than it being in the other reftest suite with a different name.

New try push: https://treeherder.mozilla.org/#/jobs?repo=try&revision=fecd344e59d4

That push is looking good. The 10.6 failure was because I accidentally included a debugging change to the failing test (which I'm going to revert before landing), and the 10.8 failure is an existing failure from a test that was added a week ago in bug 1146754 and was only fuzzed for 10.10, because we don't run 10.8 tests on inbound / mozilla-central.