Fire overflowchanged events at raf timing

Fire overflowchanged events at raf timing

Running script inside layout leads to nasty security bugs and crashes,
instead we should defer overflowchanged events until raf time. This still
lets the author take an action before the paint preventing blinking and
jumpiness which was the reason we ran it inside layout, while avoiding the
pitfalls of synchronous script.

Unfortunately this patch makes us start firing overflowchanged even when a
node has been removed from the tree, but the old protection against it was
bad since it only checked inDocument() so removing a node and putting it in
another document would still let the event fire. Instead I plan to fix the
detach problem in a future patch since scroll events shouldn't fire for
detached nodes either.

This patch also lets us remove the paused-event-dispatch.html test which
was testing the suspend/resume logic I removed and for crashes that happen
with synchronous script inside layout which doesn't apply after this patch.

BUG=293534, 323283
TEST=fast/events/overflowchanged-event-raf-timing.html

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=162655

[ojan, 2013-11-25 22:21:31]

lgtm

Your change description has some missing words.

Would be nice to add a test. Can you have a test that runs a rAF loop, then
forced a layout inside the rAF callback and ensures that we get the
overflowchanged callbacks before the next rAF callback fires? Or something to
that effect.

[esprehn, 2013-11-25 22:53:30]

On 2013/11/25 22:21:31, ojan wrote:
> lgtm
> 
> Your change description has some missing words.
> 
> Would be nice to add a test. Can you have a test that runs a rAF loop, then
> forced a layout inside the rAF callback and ensures that we get the
> overflowchanged callbacks before the next rAF callback fires? Or something to
> that effect.

I added a test.

[commit-bot: I haz the power, 2013-11-25 22:55:48]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/esprehn@chromium.org/82843003/40001

[eseidel, 2013-11-26 00:03:56]

The security team probably owes you a hug. :)

Fantastic!

[commit-bot: I haz the power, 2013-11-26 00:25:39]

Retried try job too often on linux_blink_rel for step(s) webkit_tests
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=linux_blin...

[commit-bot: I haz the power, 2013-11-26 01:26:40]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/esprehn@chromium.org/82843003/40001

[commit-bot: I haz the power, 2013-11-26 02:06:02]

Change committed as 162655