Don't DCHECK on aggregating quads with noninvertible transforms.

Don't DCHECK on aggregating quads with noninvertible transforms.

It's possible that the renderer could send these legitimately.

BUG=600886
CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel

Committed: https://crrev.com/2c859fdc9391ee6988071a3cb84c25145cfc9a4c
Cr-Commit-Position: refs/heads/master@{#387495}

[jbauman, 2016-04-13 01:29:46]

Description was changed from

==========
Don't DCHECK on aggregating quads with noninvertible transforms.

It's possible that the renderer could send these legitimately.

BUG=600886
==========

to

==========
Don't DCHECK on aggregating quads with noninvertible transforms.

It's possible that the renderer could send these legitimately.

BUG=600886
CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel
==========

[jbauman, 2016-04-13 22:25:40]

jbauman@chromium.org changed reviewers:
+ danakj@chromium.org

[danakj, 2016-04-13 22:32:47]

https://codereview.chromium.org/1883533008/diff/1/cc/surfaces/surface_aggrega...
File cc/surfaces/surface_aggregator.cc (right):

https://codereview.chromium.org/1883533008/diff/1/cc/surfaces/surface_aggrega...
cc/surfaces/surface_aggregator.cc:381: if (damage_rect_in_quad_space_valid &&
Do you want to draw quads with non-invertible damage rects or skip them? Drawing
them seems.. well idk, what was the example that generates these?

[jbauman, 2016-04-14 00:22:30]

On 2016/04/13 22:32:47, danakj wrote:
>
https://codereview.chromium.org/1883533008/diff/1/cc/surfaces/surface_aggrega...
> File cc/surfaces/surface_aggregator.cc (right):
> 
>
https://codereview.chromium.org/1883533008/diff/1/cc/surfaces/surface_aggrega...
> cc/surfaces/surface_aggregator.cc:381: if (damage_rect_in_quad_space_valid &&
> Do you want to draw quads with non-invertible damage rects or skip them?
Drawing
> them seems.. well idk, what was the example that generates these?

I think drawing them makes sense; this code is only intended as an optimization,
so I'd prefer if it didn't change behavior unnecessarily.

The original example is a flaky test involving various chrome OS accessibility
settings, so I haven't been able to reproduce it and don't exactly why it's
generating non-invertible quads. One possibility is floating-point precision
issues, but there may be other reasons.

[danakj, 2016-04-14 00:54:51]

Can we maybe DCHECK this in the renderer when it sends the IPC? Then we could
track down this bug more easily?

LGTM

[jbauman, 2016-04-14 23:42:17]

The CQ bit was checked by jbauman@chromium.org

[commit-bot: I haz the power, 2016-04-14 23:42:45]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1883533008/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1883533008/1

[commit-bot: I haz the power, 2016-04-15 00:46:08]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2016-04-15 00:47:31]

Description was changed from

==========
Don't DCHECK on aggregating quads with noninvertible transforms.

It's possible that the renderer could send these legitimately.

BUG=600886
CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel
==========

to

==========
Don't DCHECK on aggregating quads with noninvertible transforms.

It's possible that the renderer could send these legitimately.

BUG=600886
CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel

Committed: https://crrev.com/2c859fdc9391ee6988071a3cb84c25145cfc9a4c
Cr-Commit-Position: refs/heads/master@{#387495}
==========

[commit-bot: I haz the power, 2016-04-15 00:47:32]

Patchset 1 (id:??) landed as
https://crrev.com/2c859fdc9391ee6988071a3cb84c25145cfc9a4c
Cr-Commit-Position: refs/heads/master@{#387495}