cc: Fix thread race on writing commit_waits_for_activation_.

cc: Fix thread race on writing commit_waits_for_activation_.

Under non-impl-side painting, the thread proxy ends up releasing the
main thread then setting commit_waits_for_activation_ to false, which
is the wrong way around. This puts the order right.

R=enne
BUG=334611
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=245269

[enne (OOO), 2014-01-15 18:01:06]

lgtm

One random thought which you can take or leave--what about wrapping that
variable (and inside_commit) in a setter/getter that verifies that it's on the
main thread or the main thread is blocked?

[danakj, 2014-01-15 18:10:47]

On 2014/01/15 18:01:06, enne wrote:
> lgtm
> 
> One random thought which you can take or leave--what about wrapping that
> variable (and inside_commit) in a setter/getter that verifies that it's on the
> main thread or the main thread is blocked?

I've thought about breaking all the variables in ThreadProxy into two structs,
and I guess those structs could do those things. I want to do this, but I'm also
in the middle of the browsertest swamp.. so maybe :)

[commit-bot: I haz the power, 2014-01-15 18:30:03]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/danakj@chromium.org/139553003/1

[commit-bot: I haz the power, 2014-01-16 18:44:57]

Change committed as 245269