We do not have test cases for mask-position and mask-size animation, so there is no way to prevent regression issues such as bug 1273804.

test_transitions_per_property.html has some tests. Would they not have caught the regression?

(In reply to David Baron :dbaron: ⌚️UTC-7 from comment #1) > test_transitions_per_property.html has some tests. Would they not have > caught the regression? Two reasons that I don't think test_transitions_per_property.html can catch this problem 1. we didn't turn on mask-as-shorthand at that moment, so transition test for mask-position is skipped. 2. Transition value is correct, the root cause of this regression is wrong dirty region computing in display item(fixed in bug 1309646), so reftest is needed.

Example of how to create an animation test for mask-size/mask-position http://searchfox.org/mozilla-central/rev/886d5186f0598ab2f8a9953eb5a4dab9750ef834/testing/web-platform/tests/web-animations/animation-model/animation-types/spacing-keyframes-shapes.html#62-81

Comment on attachment 8823560 [details] Bug 1273807 - mask-position and mask-size animation test cases. https://reviewboard.mozilla.org/r/102094/#review102800 ::: layout/reftests/css-animations/mask-anim-ref.html:8 (Diff revision 4) > +<style> > +#test { > + height: 100px; > + width: 100px; > + background-repeat: no-repeat; > + background-image:url('data:image/svg+xml,<svg xmlns="http://www.w3.org/2000/svg" width="25" height="25"><rect x="0" y="0" width="25" height="25" fill="blue"/></svg>'); Nit: space after ":". (Same in the other files.) ::: layout/reftests/css-animations/mask-position-after-finish-1b.html:23 (Diff revision 4) > + #inner { > + height: 100px; > + width: 100px; > + background-color: blue; > + box-sizing: border-box; > + /* Apply will-change property to force paint mask on mask layer*/ Nit: space after "layer". (And in the other files.) ::: layout/reftests/css-animations/mask-size-in-delay-1a.html:4 (Diff revision 4) > + @keyframes changeMaskSize { > + from { mask-size: 50px 50px; } > + to { mask-size: 100px 100px; } > + } Any reason to have the animation (which takes no effect since we're in the delay phase) different from the mask-position tests, where we use the same from and to values? If not, may as well make them have the same form. ::: layout/reftests/css-animations/mask-size-in-delay-1a.html:19 (Diff revision 4) > + mask-position: center center; > + mask-image:url('data:image/svg+xml,<svg xmlns="http://www.w3.org/2000/svg" width="25" height="25"><rect x="0" y="0" width="25" height="25" fill="blue"/></svg>'); > + animation: changeMaskSize 100s 100s infinite; > + } > +</style> > +<div id="test" class="reftest-wait"></div> Remove the class="reftest-wait"?

Pushed by cku@mozilla.com: https://hg.mozilla.org/integration/autoland/rev/75c24febe680 mask-position and mask-size animation test cases. r=heycam