Matt Woodrow (:mattwoodrow) Assignee
	Description • 10 years ago

When we create a mask layer for a ThebesLayer we mark the number of rounded clips that should be ignored on each display item using mCommonClipCount.

We ignore these clips correctly when painting, but invalidation still tracks them.

Fixing this is hard, since we need to know the final common clip count and that hasn't been computed yet when we get to InvalidateForLayerChange.

I think we could leave the actual layer change invalidations here, but move the geometry change invalidation code into PopThebesLayerData.

Unfortunately this means that we have to keep the old geometry around longer, and it also means an extra iteration over all display items in the layer. I hope that isn't going to hurt us performance wise.

	Matt Woodrow (:mattwoodrow) Assignee
	Comment 1 • 10 years ago

This mainly just splits InvalidateForLayerChange into two pieces, and calls the latter when we process display items at the end looking for removed ones.

It also splits DisplayItemData::UpdateContent into two pieces so that the data we need for invalidation is kept around.

	Matt Woodrow (:mattwoodrow) Assignee
	Comment 3 • 10 years ago

https://hg.mozilla.org/integration/mozilla-inbound/rev/575d0363e8e2
https://hg.mozilla.org/integration/mozilla-inbound/rev/912c96edb352

	Carsten Book [:Tomcat]
	Comment 4 • 10 years ago

https://hg.mozilla.org/mozilla-central/rev/575d0363e8e2
https://hg.mozilla.org/mozilla-central/rev/912c96edb352

	Markus Stange [:mstange]
	Comment 5 • 9 years ago

We should write a test for this. I'll put that in my queue.

	Markus Stange [:mstange]
	Comment 6 • 9 years ago

Landed two basic tests: https://hg.mozilla.org/integration/mozilla-inbound/rev/6b528bd3294c

	Carsten Book [:Tomcat]
	Comment 7 • 9 years ago

sorry had to back this out for test failures like https://treeherder.mozilla.org/logviewer.html#?job_id=7736255&repo=mozilla-inbound

	Markus Stange [:mstange]
	Comment 8 • 9 years ago

Sorry about that. The bug is still fixed. I'll push a fixed test the next time I have something substantial to push.

	Markus Stange [:mstange]
	Comment 9 • 9 years ago

https://hg.mozilla.org/integration/mozilla-inbound/rev/259e491ccb21

	Phil Ringnalda (:philor)
	Comment 10 • 9 years ago

https://hg.mozilla.org/mozilla-central/rev/259e491ccb21