Bug 1223658 - Part 4: Consider fillMode in compositor thread as well. r?birtles We need to consider fill mode in compositor thread as well as other properties because pulling the animation back from the compositor thread is sometimes delayed due to the main thread busyness. In such situations, if there is another animation running on the main thread on the same element, users can easily notice a gap between both of animations. MozReview-Commit-ID: 1i7YTWboira

Bug 1223658 - Part 3: Pass delay property to compositor. r?birtles The check of negative elapsedDuration is no longer valid since animation delay is not factored into start time any more. Also, we don't need to call AnimationTimeToTimeStamp for deviding delay by playbackRate since the time passed to AnimationTimeToTimeStamp does not contain delay any more. We can use DocumentTimeline::ToTimeStamp() directly instead. As a result of these changes, we can drop Animation::GetCurrentOrPendingStartTime. MozReview-Commit-ID: IVE2IFfNgm0

Bug 1075025 - part 2: implement AndroidVsyncSource. r?snorp MozReview-Commit-ID: LEsm2SMbjww

Bug 1075025 - part 1: receive vsync events through View.postOnAnimation(). r?snorp MozReview-Commit-ID: 6cB9wdYLwFj

Bug 1223658 - Part 7: Remove Animation::HasInPlayEffect and KeyframeEffectReadOnly::IsInPlay. r?birtles MozReview-Commit-ID: JQ167ol4Ty9

Bug 1223658 - Part 6: Send animations to compositor even though it's in delay phase. r?birtles To send transform animations in before phase we need a transform style if the target element has no transform style in the before phase. In case of opacity animations we don't need such styles because nsDisplayOpacity is built if EffectCompositor::HasAnimationsForCompositor just returns true in the before phase. MozReview-Commit-ID: ILYKig3c08d

Bug 1223658 - Part 5: Add a function to check active duration is zero. r?birtles We are going to pass animations to the compositor in before phase but not to pass when active duration is zero. To distinguish this state we need a new function to check that the active duration is zero. MozReview-Commit-ID: 9fbSgfBwfpw

Bug 1223658 - Part 4: Consider fillMode in compositor thread as well. r?birtles We need to consider fill mode in compositor thread as well as other properties because pulling the animation back from the compositor thread is sometimes delayed due to the main thread busyness. In such situations, if there is another animation running on the main thread on the same element, users can easily notice a gap between both of animations. MozReview-Commit-ID: 1i7YTWboira

Bug 1223658 - Part 3: Pass delay property to compositor. r?birtles The check of negative elapsedDuration is no longer valid since animation delay is not factored into start time any more. Also, we don't need to call AnimationTimeToTimeStamp for deviding delay by playbackRate since the time passed to AnimationTimeToTimeStamp does not contain delay any more. We can use DocumentTimeline::ToTimeStamp() directly instead. As a result of these changes, we can drop Animation::GetCurrentOrPendingStartTime. MozReview-Commit-ID: IVE2IFfNgm0

Bug 1223658 - Part 7: Remove Animation::HasInPlayEffect and KeyframeEffectReadOnly::IsInPlay. r?birtles MozReview-Commit-ID: JQ167ol4Ty9

Bug 1223658 - Part 6: Send animations to compositor even though it's in delay phase. r?birtles To send transform animations in before phase we need a transform style if the target element has no transform style in the before phase. In case of opacity animations we don't need such styles because nsDisplayOpacity is built if EffectCompositor::HasAnimationsForCompositor just returns true in the before phase. MozReview-Commit-ID: ILYKig3c08d

Bug 1223658 - Part 5: Add a function to check active duration is zero. r?birtles We are going to pass animations to the compositor in before phase but not to pass when active duration is zero. To distinguish this state we need a new function to check that the active duration is zero. MozReview-Commit-ID: A1eWb3PFKMq

Bug 1223658 - Part 4: Consider fillMode in compositor thread as well. r?birtles We need to consider fill mode in compositor thread as well as other properties because pulling the animation back from the compositor thread is sometimes delayed due to the main thread busyness. In such situations, if there is another animation running on the main thread on the same element, users can easily notice a gap between both of animations. MozReview-Commit-ID: 1i7YTWboira

Bug 1223658 - Part 3: Pass delay property to compositor. r?birtles The check of negative elapsedDuration is no longer valid since animation delay is not factored into start time any more. Also, we don't need to call AnimationTimeToTimeStamp for deviding delay by playbackRate since the time passed to AnimationTimeToTimeStamp does not contain delay any more. We can use DocumentTimeline::ToTimeStamp() directly instead. As a result of these changes, we can drop Animation::GetCurrentOrPendingStartTime. MozReview-Commit-ID: IVE2IFfNgm0 * * * [mq]: P MozReview-Commit-ID: FjkTI9qoDIa

Bug 1223658 - Part 2: Add SetIdentityTransform to be used for sending animations in before phase. r?dholbert To send transform animations in before phase to compositor we need any transforms in the before phase even though there is no transform on the target element. SetIdentityTransform is for that purporse. MozReview-Commit-ID: 9rpx15pFvNN

Bug 1223658 - Part 1: Move ExtractComputedValueForTransition into StyleAnimationValue. r?dholbert CommonAnimationManager::ExtractComputedValueForTransition does not involve any members of CommonAnimationManager. MozReview-Commit-ID: I4SMwcv30YL

Bug 1265593 - Forward app server keys to Autopush on Android. r?nalexander MozReview-Commit-ID: 3J4mM1k0pcY

Bug 1247685 - Expose `PushSubscriptionOptions` in interface tests. r?baku MozReview-Commit-ID: A6x2K8Phmr7

Bug 1247685 - Send subscription keys to the Push server. r=mt MozReview-Commit-ID: 2i3UqgNGlEt

Bug 1247685 - Validate and store app server keys in the Push service. r=mt MozReview-Commit-ID: KLm6mP22y2E