Delete invalidateTreeIfNeeded.

Delete invalidateTreeIfNeeded.

This is just setting and clearing paint invalidation dirty bits.
Also remove a bunch of the paint invalidation dirty bits entirely.
There's plenty more to do.

R=esprehn@chromium.org

Committed: https://chromium.googlesource.com/external/mojo/+/10edd955e8e01e10c0dbefd7521b2238ee768aeb

[esprehn, 2015-01-10 04:54:38]

esprehn@chromium.org changed reviewers:
+ esprehn@chromium.org

[esprehn, 2015-01-10 04:54:38]

lgtm

[esprehn, 2015-01-10 04:56:47]

On second thought, I think we need to keep a bunch of calls to
scheduleVisualUpdate() to make sure new frames get scheduled.

[ojan, 2015-01-10 05:08:22]

On 2015/01/10 at 04:56:47, esprehn wrote:
> On second thought, I think we need to keep a bunch of calls to
scheduleVisualUpdate() to make sure new frames get scheduled.

We chatted in person. I need to add the canvas one back and one to
RenderObject::imageChanged. All the others should be fine because all the
setNeedsLayout, etc style methods already will call scheduleVisualUpdate.

[ojan, 2015-01-12 21:14:21]

I added back in the scheduleVisualUpdate calls. I left them out of the places I
felt confident that we didn't need them.

I also simplified the imageChanged method to just schedule an update if the
renderer is in the tree. The old logic was very complicated and written when the
repaint system was very different (in 2007). Also, it's not clear to me how we
get into the imageChanged callback without the image or it's rect changing.
Maybe that code was just for when the rect changed, but the image didn't. In
that case, we still want to schedule a frame anyways, so we don't need the extra
logic.

I also found a couple places where we don't need the callbacks because the code
is dead in sky. Added FIXMEs appropriately.

https://codereview.chromium.org/791023006/diff/1/sky/engine/core/rendering/Re...
File sky/engine/core/rendering/RenderObject.cpp (left):

https://codereview.chromium.org/791023006/diff/1/sky/engine/core/rendering/Re...
sky/engine/core/rendering/RenderObject.cpp:2250:
frame()->page()->animator().scheduleVisualUpdate(); // In case that this is
called not during FrameView::updateLayoutAndStyleForPainting().
This is the only potential change in behavior. I removed all the
setShouldDoFullPaintInvalidation calls, so now we might not be scheduling a
frame in all the cases where we used to.

Elliott and I talked about this. I'll replace all these calls with
scheduleVisualUpdate calls instead for now and we can clean those up as needed
since it's hard to tell right now which ones are needed for the frame versus for
doing paint invalidation.

[esprehn, 2015-01-12 21:27:32]

lgtm

[ojan, 2015-01-12 21:29:17]

Committed patchset #3 (id:40001) manually as
10edd955e8e01e10c0dbefd7521b2238ee768aeb (presubmit successful).