Use SaveLayerAlpha bounds when doing Save-Draw-Restore optimization

cc: Use SaveLayerAlpha bounds when doing Save-Draw-Restore optimization

When a recording does SaveLayerAlpha followed by a single drawing op
and a Restore, we can optimize that as drawing the single op with a
modified alpha. When we can't do this optimization, we were sometimes
throwing away the bounds of the SaveLayerAlpha. This would make a
different set of calls for SaveLayerAlpha-Draw-Draw-Restore vs
SaveLayerAlpha-Draw-Restore, where the latter would not use the
bounds but the former would.

Because we were dropping bounds, we've accrued code that gave
incorrect bounds and relied on them being dropped. I've fixed that
in PaintGeneratedImage in this CL to give correct LayoutTest results.
This change is verified by layout tests such as
http/tests/csspaint/background-image-alpha.html
which would fail with this CL if we didn't make the change, as this
CL would cause us to clip things outside the incorrect bounds.

BUG=671433
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_trusty_blink_rel

Review-Url: https://codereview.chromium.org/2914863003
Cr-Commit-Position: refs/heads/master@{#476303}
Committed: https://chromium.googlesource.com/chromium/src/+/b8da9c95f075fcaf5cbe8e646d22a711f1ccac30

[danakj, 2017-05-31 20:47:13]

Description was changed from

==========
savebounds

BUG=
==========

to

==========
savebounds

BUG=
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_trusty_blink_rel
==========

[danakj, 2017-05-31 21:58:07]

Description was changed from

==========
savebounds

BUG=
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_trusty_blink_rel
==========

to

==========
cc: Use SaveLayerAlpha bounds when doing Save-Draw-Restore optimization

When a recording does SaveLayerAlpha followed by a single drawing op
and a Restore, we can optimize that as drawing the single op with a
modified alpha. When we can't do this optimization, we were sometimes
throwing away the bounds of the SaveLayerAlpha. This would make a
different set of calls for SaveLayerAlpha-Draw-Draw-Restore vs
SaveLayerAlpha-Draw-Restore, where the latter would not use the
bounds but the former would.

Because we were dropping bounds, we've accrued code that gave
incorrect bounds and relied on them being dropped. I've fixed that
in PaintGeneratedImage in this CL to give correct LayoutTest results.

BUG=
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_trusty_blink_rel
==========

[danakj, 2017-05-31 21:58:07]

danakj@chromium.org changed reviewers:
+ chrishtr@chromium.org, enne@chromium.org

[danakj, 2017-05-31 21:58:30]

The CQ bit was checked by danakj@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2017-05-31 21:59:46]

Dry run: CQ is trying da patch.

Follow status at:
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[danakj, 2017-05-31 22:02:11]

The CQ bit was checked by danakj@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2017-05-31 22:02:30]

Dry run: CQ is trying da patch.

Follow status at:
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[danakj, 2017-05-31 22:03:47]

The CQ bit was checked by danakj@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2017-05-31 22:04:04]

Dry run: CQ is trying da patch.

Follow status at:
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[danakj, 2017-05-31 22:05:39]

The CQ bit was checked by danakj@chromium.org to run a CQ dry run

[danakj, 2017-05-31 22:05:45]

Patchset #4 (id:60001) has been deleted

[danakj, 2017-05-31 22:05:48]

Patchset #2 (id:20001) has been deleted

[danakj, 2017-05-31 22:05:56]

Patchset #2 (id:40001) has been deleted

[commit-bot: I haz the power, 2017-05-31 22:06:24]

Dry run: CQ is trying da patch.

Follow status at:
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[danakj, 2017-05-31 22:12:14]

Description was changed from

==========
cc: Use SaveLayerAlpha bounds when doing Save-Draw-Restore optimization

When a recording does SaveLayerAlpha followed by a single drawing op
and a Restore, we can optimize that as drawing the single op with a
modified alpha. When we can't do this optimization, we were sometimes
throwing away the bounds of the SaveLayerAlpha. This would make a
different set of calls for SaveLayerAlpha-Draw-Draw-Restore vs
SaveLayerAlpha-Draw-Restore, where the latter would not use the
bounds but the former would.

Because we were dropping bounds, we've accrued code that gave
incorrect bounds and relied on them being dropped. I've fixed that
in PaintGeneratedImage in this CL to give correct LayoutTest results.

BUG=
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_trusty_blink_rel
==========

to

==========
cc: Use SaveLayerAlpha bounds when doing Save-Draw-Restore optimization

When a recording does SaveLayerAlpha followed by a single drawing op
and a Restore, we can optimize that as drawing the single op with a
modified alpha. When we can't do this optimization, we were sometimes
throwing away the bounds of the SaveLayerAlpha. This would make a
different set of calls for SaveLayerAlpha-Draw-Draw-Restore vs
SaveLayerAlpha-Draw-Restore, where the latter would not use the
bounds but the former would.

Because we were dropping bounds, we've accrued code that gave
incorrect bounds and relied on them being dropped. I've fixed that
in PaintGeneratedImage in this CL to give correct LayoutTest results.
This change is verified by layout tests such as
http/tests/csspaint/background-image-alpha.html
which would fail with this CL if we didn't make the change, as this
CL would cause us to clip things outside the incorrect bounds.

BUG=671433
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_trusty_blink_rel
==========

[enne (OOO), 2017-05-31 22:15:36]

lgtm

[commit-bot: I haz the power, 2017-06-01 00:30:06]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2017-06-01 00:30:07]

Dry run: Try jobs failed on following builders:
  linux_chromium_rel_ng on master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)

[chrishtr, 2017-06-01 00:41:51]

lgtm

[danakj, 2017-06-01 14:05:52]

The CQ bit was checked by danakj@chromium.org

[commit-bot: I haz the power, 2017-06-01 14:06:08]

CQ is trying da patch.

Follow status at:
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2017-06-01 15:51:53]

CQ is committing da patch.

Bot data: {"patchset_id": 80001, "attempt_start_ts": 1496325952674380,
"parent_rev": "907652b824e283a6109261f5e760bf879dc1537a", "commit_rev":
"a247ef3a757817a0e9a1fa38aac00a566f6515a9"}

[commit-bot: I haz the power, 2017-06-01 15:52:05]

CQ is committing da patch.

Bot data: {"patchset_id": 80001, "attempt_start_ts": 1496325952674380,
"parent_rev": "95f5e23f33ef467b01c54a5b769d7d03f6f90b73", "commit_rev":
"b8da9c95f075fcaf5cbe8e646d22a711f1ccac30"}

[commit-bot: I haz the power, 2017-06-01 15:52:17]

Description was changed from

==========
cc: Use SaveLayerAlpha bounds when doing Save-Draw-Restore optimization

When a recording does SaveLayerAlpha followed by a single drawing op
and a Restore, we can optimize that as drawing the single op with a
modified alpha. When we can't do this optimization, we were sometimes
throwing away the bounds of the SaveLayerAlpha. This would make a
different set of calls for SaveLayerAlpha-Draw-Draw-Restore vs
SaveLayerAlpha-Draw-Restore, where the latter would not use the
bounds but the former would.

Because we were dropping bounds, we've accrued code that gave
incorrect bounds and relied on them being dropped. I've fixed that
in PaintGeneratedImage in this CL to give correct LayoutTest results.
This change is verified by layout tests such as
http/tests/csspaint/background-image-alpha.html
which would fail with this CL if we didn't make the change, as this
CL would cause us to clip things outside the incorrect bounds.

BUG=671433
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_trusty_blink_rel
==========

to

==========
cc: Use SaveLayerAlpha bounds when doing Save-Draw-Restore optimization

When a recording does SaveLayerAlpha followed by a single drawing op
and a Restore, we can optimize that as drawing the single op with a
modified alpha. When we can't do this optimization, we were sometimes
throwing away the bounds of the SaveLayerAlpha. This would make a
different set of calls for SaveLayerAlpha-Draw-Draw-Restore vs
SaveLayerAlpha-Draw-Restore, where the latter would not use the
bounds but the former would.

Because we were dropping bounds, we've accrued code that gave
incorrect bounds and relied on them being dropped. I've fixed that
in PaintGeneratedImage in this CL to give correct LayoutTest results.
This change is verified by layout tests such as
http/tests/csspaint/background-image-alpha.html
which would fail with this CL if we didn't make the change, as this
CL would cause us to clip things outside the incorrect bounds.

BUG=671433
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_trusty_blink_rel

Review-Url: https://codereview.chromium.org/2914863003
Cr-Commit-Position: refs/heads/master@{#476303}
Committed:
https://chromium.googlesource.com/chromium/src/+/b8da9c95f075fcaf5cbe8e646d22...
==========

[commit-bot: I haz the power, 2017-06-01 15:52:18]

Committed patchset #2 (id:80001) as
https://chromium.googlesource.com/chromium/src/+/b8da9c95f075fcaf5cbe8e646d22...