For example, if a scroll handler updates the transform property of an element, we should keep that element layerized until the scrolled scroll frame becomes inactive.

This helps on many sites with parallax effects. Most of them modify background-position from a scroll handler. Keeping the background position layer activity alive between scrolls means there's no lag when a new scroll is started after the page has rested for a while.

Bug 1201330 - Refactor LayerActivity property management. r?roc

Bug 1201330 - Keep scroll handler induced layer activity active until the scroll frame becomes inactive. r?roc

Comment on attachment 8679530 [details] MozReview Request: Bug 1201330 - Refactor LayerActivity property management. r?roc https://reviewboard.mozilla.org/r/23461/#review20949

Comment on attachment 8679531 [details] MozReview Request: Bug 1201330 - Keep scroll handler induced layer activity active until the scroll frame becomes inactive. r?roc https://reviewboard.mozilla.org/r/23463/#review20963

https://hg.mozilla.org/integration/mozilla-inbound/rev/72da2ed9b1e7db209601769d5fc4461c0490963f Bug 1201330 - Refactor LayerActivity property management. r=roc https://hg.mozilla.org/integration/mozilla-inbound/rev/d5d68c0258f3d225bce32140e13a7d40f0f29a2f Bug 1201330 - Keep scroll handler induced layer activity active until the scroll frame becomes inactive. r=roc

sorry had to back this out for test failures like https://treeherder.mozilla.org/logviewer.html#?job_id=16547879&repo=mozilla-inbound

Backout: https://hg.mozilla.org/integration/mozilla-inbound/rev/0fe806b72d85 https://hg.mozilla.org/integration/mozilla-inbound/rev/edf4e01443dc

https://hg.mozilla.org/integration/mozilla-inbound/rev/61217d41be19074ced225fcb09c281168ac7a6cd Bug 1201330 - Refactor LayerActivity property management. r=roc https://hg.mozilla.org/integration/mozilla-inbound/rev/1f3ea4c34877706356d43a1834228b1a7f91f18d Bug 1201330 - Keep scroll handler induced layer activity active until the scroll frame becomes inactive. r=roc

I had to add another nsWeakFrame, for the current scroll frame, because a scroll frame can be destroyed while it's firing a scroll event. The test that failed did that because it was listening for the scroll event in a window it had opened, and during the scroll event handler it closed that window.

https://hg.mozilla.org/mozilla-central/rev/61217d41be19 https://hg.mozilla.org/mozilla-central/rev/1f3ea4c34877

https://hg.mozilla.org/releases/mozilla-b2g44_v2_5/rev/61217d41be19 https://hg.mozilla.org/releases/mozilla-b2g44_v2_5/rev/1f3ea4c34877

removing the b2g 2.5 flag since this commit has been reverted due to an incorrect merge, sorry for the confusion