cc: Make LayerTreeTests use a DelegatingRenderer and Display.

cc: Make LayerTreeTests use a DelegatingRenderer and Display.

Wherein we rewrite a few thousand cc unit_tests to not use a
DirectRenderer in LayerTreeHostImpl so that we can remove that
possibility from the codebase.

Currently some LayerTreeTests use a DirectRenderer in the
LayerTreeHostImpl, which is the last case of this occuring. Instead
give all LayerTreeTests a TestDelegatingOutputSurface (which means
a DelegatingRenderer in LayerTreeHostImpl). TestDelegatingOutputSurface
delegates drawing to a DirectRenderer via Display, which matches how
things "really work" now.

This means SwapBuffers hooks become async from operations on
LayerTreeHostImpl, so we introduce 3 new hooks for tests:
- DisplayReceivedCompositorFrameOnThread.
- DisplayWillDrawAndSwapOnThread.
- DisplayDidDrawAndSwapOnThread.
None of these receive a LayerTreeHostImpl* since they are async
from it and so using its state from the hook would be racey. These
hooks come from the TestDelegatingOutputSurface instead of from the
LayerTreeHostImpl.

LayerTreeTest gets two methods that can be overridden to control the
output surface:
- CreateDelegatingOutputSurface makes a TestDelegatingOutputSurface
so tests can control the arguments passed to its constructor, or the
ContextProviders given to it.
- CreateDisplayOutputSurface makes an OutputSurface subclass that
will be owned by the Display and used by the DirectRenderer. This
allows tests to control which ContextProvider it uses, and what
subclass of OutputSurface to use, to allow mocking out things as
desired.

Now also with fixes for cc_perftests:

The TestDelegatingOutputSurface was not honoring disable_display_vsync
so they were not saturating the thread, making them take much longer
than 5s to do 5s of thread work, and they'd time out.

Also the TetureMailbox doesn't get released to the main thread
immediately after commit anymore, as you need to wait for the
mailbox to be given to the Display (DisplayReceivedCompositorFrame
is the hook for that fwiw). So we wait for the mailbox to actually
return before ending the test to avoid a crash.

R=enne
BUG=606056
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel

Committed: https://crrev.com/cae1058e62b7960796709512c4e1e650b6389c0f
Committed: https://crrev.com/014316e35eb9a90a024302daa3a6c3ca3983a8bd
Cr-Original-Commit-Position: refs/heads/master@{#409270}
Cr-Commit-Position: refs/heads/master@{#409845}

[danakj, 2016-07-29 22:45:09]

Description was changed from

==========
cc: Make LayerTreeTests use a Display for a DirectRenderer.

Currently some LayerTreeTests use a DirectRenderer in the
LayerTreeHostImpl, which is the last case of this occuring. Instead
give all LayerTreeTests a TestDelegatingOutputSurface (which means
a DelegatingRenderer in LayerTreeHostImpl). TestDelegatingOutputSurface
delegates drawing to a DirectRenderer via Display, which matches how
things "really work" now.

This means SwapBuffers hooks become async from operations on
LayerTreeHostImpl, so we introduce 3 new hooks for tests:
- DisplayReceivedCompositorFrameOnThread.
- DisplayWillDrawAndSwapOnThread.
- DisplayDidDrawAndSwapOnThread.
None of these receive a LayerTreeHostImpl* since they are async
from it and so using its state from the hook would be racey. These
hooks come from the TestDelegatingOutputSurface instead of from the
LayerTreeHostImpl.

LayerTreeTest gets two methods that can be overridden to control the
output surface:
- CreateDelegatingOutputSurface makes a TestDelegatingOutputSurface
so tests can control the arguments passed to its constructor, or the
ContextProviders given to it.
- CreateDisplayOutputSurface makes an OutputSurface subclass that
will be owned by the Display and used by the DirectRenderer. This
allows tests to control which ContextProvider it uses, and what
subclass of OutputSurface to use, to allow mocking out things as
desired.

R=enne
BUG=606056
==========

to

==========
cc: Make LayerTreeTests use a Display for a DirectRenderer.

Currently some LayerTreeTests use a DirectRenderer in the
LayerTreeHostImpl, which is the last case of this occuring. Instead
give all LayerTreeTests a TestDelegatingOutputSurface (which means
a DelegatingRenderer in LayerTreeHostImpl). TestDelegatingOutputSurface
delegates drawing to a DirectRenderer via Display, which matches how
things "really work" now.

This means SwapBuffers hooks become async from operations on
LayerTreeHostImpl, so we introduce 3 new hooks for tests:
- DisplayReceivedCompositorFrameOnThread.
- DisplayWillDrawAndSwapOnThread.
- DisplayDidDrawAndSwapOnThread.
None of these receive a LayerTreeHostImpl* since they are async
from it and so using its state from the hook would be racey. These
hooks come from the TestDelegatingOutputSurface instead of from the
LayerTreeHostImpl.

LayerTreeTest gets two methods that can be overridden to control the
output surface:
- CreateDelegatingOutputSurface makes a TestDelegatingOutputSurface
so tests can control the arguments passed to its constructor, or the
ContextProviders given to it.
- CreateDisplayOutputSurface makes an OutputSurface subclass that
will be owned by the Display and used by the DirectRenderer. This
allows tests to control which ContextProvider it uses, and what
subclass of OutputSurface to use, to allow mocking out things as
desired.

R=enne
BUG=606056
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel
==========

[danakj, 2016-07-30 00:51:16]

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

[commit-bot: I haz the power, 2016-07-30 00:51:34]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[danakj, 2016-07-30 00:54:24]

https://codereview.chromium.org/2193293004/diff/20001/cc/trees/layer_tree_hos...
File cc/trees/layer_tree_host_unittest.cc (right):

https://codereview.chromium.org/2193293004/diff/20001/cc/trees/layer_tree_hos...
cc/trees/layer_tree_host_unittest.cc:2645: class OnDrawOutputSurface : public
TestDelegatingOutputSurface {
This is the singular time we override CreateDelegatingOutputSurface and actually
need to change what the function *returns*.

The other cases all just want to change the params passed to TestDOS
constructor.

I considered maybe replacing CreateDelegatingOutputSurface with
OverrideDelegatingOutputSurfaceArgs(struct T*) where you can change what things
you want to pass to it in tests, but the actual code is all in LayerTreeTest
class always. But then there's this case.

I could add Invalidate to TestDOSClient and TestHooks, then I could do the above
still.

What do you think?

[danakj, 2016-07-30 00:59:06]

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

[danakj, 2016-07-30 00:59:06]

+fsamuel for mus

[danakj, 2016-07-30 00:59:54]

Description was changed from

==========
cc: Make LayerTreeTests use a Display for a DirectRenderer.

Currently some LayerTreeTests use a DirectRenderer in the
LayerTreeHostImpl, which is the last case of this occuring. Instead
give all LayerTreeTests a TestDelegatingOutputSurface (which means
a DelegatingRenderer in LayerTreeHostImpl). TestDelegatingOutputSurface
delegates drawing to a DirectRenderer via Display, which matches how
things "really work" now.

This means SwapBuffers hooks become async from operations on
LayerTreeHostImpl, so we introduce 3 new hooks for tests:
- DisplayReceivedCompositorFrameOnThread.
- DisplayWillDrawAndSwapOnThread.
- DisplayDidDrawAndSwapOnThread.
None of these receive a LayerTreeHostImpl* since they are async
from it and so using its state from the hook would be racey. These
hooks come from the TestDelegatingOutputSurface instead of from the
LayerTreeHostImpl.

LayerTreeTest gets two methods that can be overridden to control the
output surface:
- CreateDelegatingOutputSurface makes a TestDelegatingOutputSurface
so tests can control the arguments passed to its constructor, or the
ContextProviders given to it.
- CreateDisplayOutputSurface makes an OutputSurface subclass that
will be owned by the Display and used by the DirectRenderer. This
allows tests to control which ContextProvider it uses, and what
subclass of OutputSurface to use, to allow mocking out things as
desired.

R=enne
BUG=606056
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel
==========

to

==========
cc: Make LayerTreeTests use a Display for a DirectRenderer.

Wherein we rewrite a few thousand cc unit_tests to not use a
DirectRenderer in LayerTreeHostImpl so that we can remove that
possibility from the codebase.

Currently some LayerTreeTests use a DirectRenderer in the
LayerTreeHostImpl, which is the last case of this occuring. Instead
give all LayerTreeTests a TestDelegatingOutputSurface (which means
a DelegatingRenderer in LayerTreeHostImpl). TestDelegatingOutputSurface
delegates drawing to a DirectRenderer via Display, which matches how
things "really work" now.

This means SwapBuffers hooks become async from operations on
LayerTreeHostImpl, so we introduce 3 new hooks for tests:
- DisplayReceivedCompositorFrameOnThread.
- DisplayWillDrawAndSwapOnThread.
- DisplayDidDrawAndSwapOnThread.
None of these receive a LayerTreeHostImpl* since they are async
from it and so using its state from the hook would be racey. These
hooks come from the TestDelegatingOutputSurface instead of from the
LayerTreeHostImpl.

LayerTreeTest gets two methods that can be overridden to control the
output surface:
- CreateDelegatingOutputSurface makes a TestDelegatingOutputSurface
so tests can control the arguments passed to its constructor, or the
ContextProviders given to it.
- CreateDisplayOutputSurface makes an OutputSurface subclass that
will be owned by the Display and used by the DirectRenderer. This
allows tests to control which ContextProvider it uses, and what
subclass of OutputSurface to use, to allow mocking out things as
desired.

R=enne
BUG=606056
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel
==========

[commit-bot: I haz the power, 2016-07-30 01:00:03]

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

[commit-bot: I haz the power, 2016-07-30 01:00:04]

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

[danakj, 2016-07-30 01:00:22]

Description was changed from

==========
cc: Make LayerTreeTests use a Display for a DirectRenderer.

Wherein we rewrite a few thousand cc unit_tests to not use a
DirectRenderer in LayerTreeHostImpl so that we can remove that
possibility from the codebase.

Currently some LayerTreeTests use a DirectRenderer in the
LayerTreeHostImpl, which is the last case of this occuring. Instead
give all LayerTreeTests a TestDelegatingOutputSurface (which means
a DelegatingRenderer in LayerTreeHostImpl). TestDelegatingOutputSurface
delegates drawing to a DirectRenderer via Display, which matches how
things "really work" now.

This means SwapBuffers hooks become async from operations on
LayerTreeHostImpl, so we introduce 3 new hooks for tests:
- DisplayReceivedCompositorFrameOnThread.
- DisplayWillDrawAndSwapOnThread.
- DisplayDidDrawAndSwapOnThread.
None of these receive a LayerTreeHostImpl* since they are async
from it and so using its state from the hook would be racey. These
hooks come from the TestDelegatingOutputSurface instead of from the
LayerTreeHostImpl.

LayerTreeTest gets two methods that can be overridden to control the
output surface:
- CreateDelegatingOutputSurface makes a TestDelegatingOutputSurface
so tests can control the arguments passed to its constructor, or the
ContextProviders given to it.
- CreateDisplayOutputSurface makes an OutputSurface subclass that
will be owned by the Display and used by the DirectRenderer. This
allows tests to control which ContextProvider it uses, and what
subclass of OutputSurface to use, to allow mocking out things as
desired.

R=enne
BUG=606056
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel
==========

to

==========
cc: Make LayerTreeTests use a DelegatingRenderer and Display.

Wherein we rewrite a few thousand cc unit_tests to not use a
DirectRenderer in LayerTreeHostImpl so that we can remove that
possibility from the codebase.

Currently some LayerTreeTests use a DirectRenderer in the
LayerTreeHostImpl, which is the last case of this occuring. Instead
give all LayerTreeTests a TestDelegatingOutputSurface (which means
a DelegatingRenderer in LayerTreeHostImpl). TestDelegatingOutputSurface
delegates drawing to a DirectRenderer via Display, which matches how
things "really work" now.

This means SwapBuffers hooks become async from operations on
LayerTreeHostImpl, so we introduce 3 new hooks for tests:
- DisplayReceivedCompositorFrameOnThread.
- DisplayWillDrawAndSwapOnThread.
- DisplayDidDrawAndSwapOnThread.
None of these receive a LayerTreeHostImpl* since they are async
from it and so using its state from the hook would be racey. These
hooks come from the TestDelegatingOutputSurface instead of from the
LayerTreeHostImpl.

LayerTreeTest gets two methods that can be overridden to control the
output surface:
- CreateDelegatingOutputSurface makes a TestDelegatingOutputSurface
so tests can control the arguments passed to its constructor, or the
ContextProviders given to it.
- CreateDisplayOutputSurface makes an OutputSurface subclass that
will be owned by the Display and used by the DirectRenderer. This
allows tests to control which ContextProvider it uses, and what
subclass of OutputSurface to use, to allow mocking out things as
desired.

R=enne
BUG=606056
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel
==========

[danakj, 2016-07-30 01:05:04]

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

[commit-bot: I haz the power, 2016-07-30 01:05:21]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[danakj, 2016-07-30 01:06:53]

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

[danakj, 2016-07-30 01:06:53]

+piman for trivial change in content/../layouttest_support

[commit-bot: I haz the power, 2016-07-30 01:17:44]

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

[commit-bot: I haz the power, 2016-07-30 01:17:45]

Dry run: Try jobs failed on following builders:
  linux_android_rel_ng on master.tryserver.chromium.android (JOB_FAILED,
https://build.chromium.org/p/tryserver.chromium.android/builders/linux_androi...)
  cast_shell_linux on master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/cast_shell_linu...)

[danakj, 2016-07-30 01:43:31]

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

[commit-bot: I haz the power, 2016-07-30 01:43:57]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[danakj, 2016-07-30 01:44:12]

Patchset #4 (id:60001) has been deleted

[commit-bot: I haz the power, 2016-07-30 01:45:29]

Dry run: Try jobs failed on following builders:
  ios-device on master.tryserver.chromium.mac (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.mac/builders/ios-device/builds...)
  ios-device-gn on master.tryserver.chromium.mac (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.mac/builders/ios-device-gn/bui...)
  ios-simulator on master.tryserver.chromium.mac (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.mac/builders/ios-simulator/bui...)
  ios-simulator-gn on master.tryserver.chromium.mac (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.mac/builders/ios-simulator-gn/...)

[danakj, 2016-07-30 01:47:32]

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

[danakj, 2016-07-30 01:47:56]

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

[danakj, 2016-07-30 01:47:57]

+boliu for the webview bits (just extending the DisplayClient stubs)

[commit-bot: I haz the power, 2016-07-30 01:47:57]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[danakj, 2016-07-30 01:48:38]

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

[danakj, 2016-07-30 01:48:45]

On 2016/07/30 01:47:57, danakj wrote:
> +boliu for the webview bits (just extending the DisplayClient stubs)

[Fady Samuel, 2016-07-30 01:51:14]

lgtm

[commit-bot: I haz the power, 2016-07-30 02:10:14]

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

[commit-bot: I haz the power, 2016-07-30 02:10:15]

Dry run: Try jobs failed on following builders:
  android_compile_dbg on master.tryserver.chromium.android (JOB_FAILED,
https://build.chromium.org/p/tryserver.chromium.android/builders/android_comp...)

[danakj, 2016-07-30 02:13:19]

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

[commit-bot: I haz the power, 2016-07-30 02:13:38]

Dry run: 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, 2016-07-30 03:55:36]

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

[commit-bot: I haz the power, 2016-07-30 03:55:37]

Dry run: Try jobs failed on following builders:
  win_chromium_rel_ng on master.tryserver.chromium.win (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.win/builders/win_chromium_rel_...)

[danakj, 2016-07-30 04:05:28]

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

[commit-bot: I haz the power, 2016-07-30 04:05:38]

Dry run: 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, 2016-07-30 05:08:34]

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

[commit-bot: I haz the power, 2016-07-30 05:08:36]

Dry run: This issue passed the CQ dry run.

[boliu, 2016-08-01 16:18:04]

rs lgtm

[piman, 2016-08-02 02:53:42]

lgtm

[enne (OOO), 2016-08-02 17:23:41]

https://codereview.chromium.org/2193293004/diff/120001/cc/surfaces/surface_di...
File cc/surfaces/surface_display_output_surface.h (right):

https://codereview.chromium.org/2193293004/diff/120001/cc/surfaces/surface_di...
cc/surfaces/surface_display_output_surface.h:60: void DisplayDidDrawAndSwap()
override;
Also in a cut down in client functions sense, why does this need to be here?

Can TestDelegatingOutputSurface::DrawCallback do this and only do it for tests,
or do you also need to wait on swap for real?

https://codereview.chromium.org/2193293004/diff/120001/cc/trees/layer_tree_ho...
File cc/trees/layer_tree_host_unittest_copyrequest.cc (right):

https://codereview.chromium.org/2193293004/diff/120001/cc/trees/layer_tree_ho...
cc/trees/layer_tree_host_unittest_copyrequest.cc:495: void
DisplayWillDrawAndSwapOnThread(
In a "cut down on client functions" thought, it looks like this is the only test
that needs all this plumbing for will draw and swap? Does it really need "will
draw and swap"? If you get a did draw and swap, is that sufficient? Also, can't
the compositor frame function tell you what the render passes are?

[danakj, 2016-08-02 17:28:40]

https://codereview.chromium.org/2193293004/diff/120001/cc/surfaces/surface_di...
File cc/surfaces/surface_display_output_surface.h (right):

https://codereview.chromium.org/2193293004/diff/120001/cc/surfaces/surface_di...
cc/surfaces/surface_display_output_surface.h:60: void DisplayDidDrawAndSwap()
override;
On 2016/08/02 17:23:40, enne wrote:
> Also in a cut down in client functions sense, why does this need to be here?
> 
> Can TestDelegatingOutputSurface::DrawCallback do this and only do it for
tests,
> or do you also need to wait on swap for real?

We need to wait for swap for real to check on things the GLRenderer did, for
copy requests and stuff.

https://codereview.chromium.org/2193293004/diff/120001/cc/trees/layer_tree_ho...
File cc/trees/layer_tree_host_unittest_copyrequest.cc (right):

https://codereview.chromium.org/2193293004/diff/120001/cc/trees/layer_tree_ho...
cc/trees/layer_tree_host_unittest_copyrequest.cc:495: void
DisplayWillDrawAndSwapOnThread(
On 2016/08/02 17:23:40, enne wrote:
> In a "cut down on client functions" thought, it looks like this is the only
test
> that needs all this plumbing for will draw and swap? Does it really need "will
> draw and swap"? If you get a did draw and swap, is that sufficient? Also,
can't
> the compositor frame function tell you what the render passes are?

The compositor frame has different ids, so we can't verify
HasAllocatedResoucesForTesting without knowing the IDs that the
SurfaceAggregator made, which are the RenderPassList.

[enne (OOO), 2016-08-02 17:50:41]

Ok ok! I understand now why you need these.

lgtm

[danakj, 2016-08-02 17:52:50]

The CQ bit was checked by danakj@chromium.org

[danakj, 2016-08-02 17:52:51]

The patchset sent to the CQ was uploaded after l-g-t-m from fsamuel@chromium.org
Link to the patchset: https://codereview.chromium.org/2193293004/#ps120001
(title: "display-layertreetest: windows-pixeltests-default-renderer-settings")

[commit-bot: I haz the power, 2016-08-02 17:53:43]

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, 2016-08-02 19:27:05]

Description was changed from

==========
cc: Make LayerTreeTests use a DelegatingRenderer and Display.

Wherein we rewrite a few thousand cc unit_tests to not use a
DirectRenderer in LayerTreeHostImpl so that we can remove that
possibility from the codebase.

Currently some LayerTreeTests use a DirectRenderer in the
LayerTreeHostImpl, which is the last case of this occuring. Instead
give all LayerTreeTests a TestDelegatingOutputSurface (which means
a DelegatingRenderer in LayerTreeHostImpl). TestDelegatingOutputSurface
delegates drawing to a DirectRenderer via Display, which matches how
things "really work" now.

This means SwapBuffers hooks become async from operations on
LayerTreeHostImpl, so we introduce 3 new hooks for tests:
- DisplayReceivedCompositorFrameOnThread.
- DisplayWillDrawAndSwapOnThread.
- DisplayDidDrawAndSwapOnThread.
None of these receive a LayerTreeHostImpl* since they are async
from it and so using its state from the hook would be racey. These
hooks come from the TestDelegatingOutputSurface instead of from the
LayerTreeHostImpl.

LayerTreeTest gets two methods that can be overridden to control the
output surface:
- CreateDelegatingOutputSurface makes a TestDelegatingOutputSurface
so tests can control the arguments passed to its constructor, or the
ContextProviders given to it.
- CreateDisplayOutputSurface makes an OutputSurface subclass that
will be owned by the Display and used by the DirectRenderer. This
allows tests to control which ContextProvider it uses, and what
subclass of OutputSurface to use, to allow mocking out things as
desired.

R=enne
BUG=606056
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel
==========

to

==========
cc: Make LayerTreeTests use a DelegatingRenderer and Display.

Wherein we rewrite a few thousand cc unit_tests to not use a
DirectRenderer in LayerTreeHostImpl so that we can remove that
possibility from the codebase.

Currently some LayerTreeTests use a DirectRenderer in the
LayerTreeHostImpl, which is the last case of this occuring. Instead
give all LayerTreeTests a TestDelegatingOutputSurface (which means
a DelegatingRenderer in LayerTreeHostImpl). TestDelegatingOutputSurface
delegates drawing to a DirectRenderer via Display, which matches how
things "really work" now.

This means SwapBuffers hooks become async from operations on
LayerTreeHostImpl, so we introduce 3 new hooks for tests:
- DisplayReceivedCompositorFrameOnThread.
- DisplayWillDrawAndSwapOnThread.
- DisplayDidDrawAndSwapOnThread.
None of these receive a LayerTreeHostImpl* since they are async
from it and so using its state from the hook would be racey. These
hooks come from the TestDelegatingOutputSurface instead of from the
LayerTreeHostImpl.

LayerTreeTest gets two methods that can be overridden to control the
output surface:
- CreateDelegatingOutputSurface makes a TestDelegatingOutputSurface
so tests can control the arguments passed to its constructor, or the
ContextProviders given to it.
- CreateDisplayOutputSurface makes an OutputSurface subclass that
will be owned by the Display and used by the DirectRenderer. This
allows tests to control which ContextProvider it uses, and what
subclass of OutputSurface to use, to allow mocking out things as
desired.

R=enne
BUG=606056
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel
==========

[commit-bot: I haz the power, 2016-08-02 19:27:07]

Committed patchset #6 (id:120001)

[commit-bot: I haz the power, 2016-08-02 19:28:50]

Description was changed from

==========
cc: Make LayerTreeTests use a DelegatingRenderer and Display.

Wherein we rewrite a few thousand cc unit_tests to not use a
DirectRenderer in LayerTreeHostImpl so that we can remove that
possibility from the codebase.

Currently some LayerTreeTests use a DirectRenderer in the
LayerTreeHostImpl, which is the last case of this occuring. Instead
give all LayerTreeTests a TestDelegatingOutputSurface (which means
a DelegatingRenderer in LayerTreeHostImpl). TestDelegatingOutputSurface
delegates drawing to a DirectRenderer via Display, which matches how
things "really work" now.

This means SwapBuffers hooks become async from operations on
LayerTreeHostImpl, so we introduce 3 new hooks for tests:
- DisplayReceivedCompositorFrameOnThread.
- DisplayWillDrawAndSwapOnThread.
- DisplayDidDrawAndSwapOnThread.
None of these receive a LayerTreeHostImpl* since they are async
from it and so using its state from the hook would be racey. These
hooks come from the TestDelegatingOutputSurface instead of from the
LayerTreeHostImpl.

LayerTreeTest gets two methods that can be overridden to control the
output surface:
- CreateDelegatingOutputSurface makes a TestDelegatingOutputSurface
so tests can control the arguments passed to its constructor, or the
ContextProviders given to it.
- CreateDisplayOutputSurface makes an OutputSurface subclass that
will be owned by the Display and used by the DirectRenderer. This
allows tests to control which ContextProvider it uses, and what
subclass of OutputSurface to use, to allow mocking out things as
desired.

R=enne
BUG=606056
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel
==========

to

==========
cc: Make LayerTreeTests use a DelegatingRenderer and Display.

Wherein we rewrite a few thousand cc unit_tests to not use a
DirectRenderer in LayerTreeHostImpl so that we can remove that
possibility from the codebase.

Currently some LayerTreeTests use a DirectRenderer in the
LayerTreeHostImpl, which is the last case of this occuring. Instead
give all LayerTreeTests a TestDelegatingOutputSurface (which means
a DelegatingRenderer in LayerTreeHostImpl). TestDelegatingOutputSurface
delegates drawing to a DirectRenderer via Display, which matches how
things "really work" now.

This means SwapBuffers hooks become async from operations on
LayerTreeHostImpl, so we introduce 3 new hooks for tests:
- DisplayReceivedCompositorFrameOnThread.
- DisplayWillDrawAndSwapOnThread.
- DisplayDidDrawAndSwapOnThread.
None of these receive a LayerTreeHostImpl* since they are async
from it and so using its state from the hook would be racey. These
hooks come from the TestDelegatingOutputSurface instead of from the
LayerTreeHostImpl.

LayerTreeTest gets two methods that can be overridden to control the
output surface:
- CreateDelegatingOutputSurface makes a TestDelegatingOutputSurface
so tests can control the arguments passed to its constructor, or the
ContextProviders given to it.
- CreateDisplayOutputSurface makes an OutputSurface subclass that
will be owned by the Display and used by the DirectRenderer. This
allows tests to control which ContextProvider it uses, and what
subclass of OutputSurface to use, to allow mocking out things as
desired.

R=enne
BUG=606056
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel

Committed: https://crrev.com/cae1058e62b7960796709512c4e1e650b6389c0f
Cr-Commit-Position: refs/heads/master@{#409270}
==========

[commit-bot: I haz the power, 2016-08-02 19:28:52]

Patchset 6 (id:??) landed as
https://crrev.com/cae1058e62b7960796709512c4e1e650b6389c0f
Cr-Commit-Position: refs/heads/master@{#409270}

[petrcermak, 2016-08-04 09:33:13]

A revert of this CL (patchset #6 id:120001) has been created in
https://codereview.chromium.org/2208693003/ by petrcermak@chromium.org.

The reason for reverting is: Bisect identified that this patch broke
cc_perftests on Linux and Android perf bots (crbug.com/634013)..

[danakj, 2016-08-04 17:13:32]

Description was changed from

==========
cc: Make LayerTreeTests use a DelegatingRenderer and Display.

Wherein we rewrite a few thousand cc unit_tests to not use a
DirectRenderer in LayerTreeHostImpl so that we can remove that
possibility from the codebase.

Currently some LayerTreeTests use a DirectRenderer in the
LayerTreeHostImpl, which is the last case of this occuring. Instead
give all LayerTreeTests a TestDelegatingOutputSurface (which means
a DelegatingRenderer in LayerTreeHostImpl). TestDelegatingOutputSurface
delegates drawing to a DirectRenderer via Display, which matches how
things "really work" now.

This means SwapBuffers hooks become async from operations on
LayerTreeHostImpl, so we introduce 3 new hooks for tests:
- DisplayReceivedCompositorFrameOnThread.
- DisplayWillDrawAndSwapOnThread.
- DisplayDidDrawAndSwapOnThread.
None of these receive a LayerTreeHostImpl* since they are async
from it and so using its state from the hook would be racey. These
hooks come from the TestDelegatingOutputSurface instead of from the
LayerTreeHostImpl.

LayerTreeTest gets two methods that can be overridden to control the
output surface:
- CreateDelegatingOutputSurface makes a TestDelegatingOutputSurface
so tests can control the arguments passed to its constructor, or the
ContextProviders given to it.
- CreateDisplayOutputSurface makes an OutputSurface subclass that
will be owned by the Display and used by the DirectRenderer. This
allows tests to control which ContextProvider it uses, and what
subclass of OutputSurface to use, to allow mocking out things as
desired.

R=enne
BUG=606056
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel

Committed: https://crrev.com/cae1058e62b7960796709512c4e1e650b6389c0f
Cr-Commit-Position: refs/heads/master@{#409270}
==========

to

==========
cc: Make LayerTreeTests use a DelegatingRenderer and Display.

Wherein we rewrite a few thousand cc unit_tests to not use a
DirectRenderer in LayerTreeHostImpl so that we can remove that
possibility from the codebase.

Currently some LayerTreeTests use a DirectRenderer in the
LayerTreeHostImpl, which is the last case of this occuring. Instead
give all LayerTreeTests a TestDelegatingOutputSurface (which means
a DelegatingRenderer in LayerTreeHostImpl). TestDelegatingOutputSurface
delegates drawing to a DirectRenderer via Display, which matches how
things "really work" now.

This means SwapBuffers hooks become async from operations on
LayerTreeHostImpl, so we introduce 3 new hooks for tests:
- DisplayReceivedCompositorFrameOnThread.
- DisplayWillDrawAndSwapOnThread.
- DisplayDidDrawAndSwapOnThread.
None of these receive a LayerTreeHostImpl* since they are async
from it and so using its state from the hook would be racey. These
hooks come from the TestDelegatingOutputSurface instead of from the
LayerTreeHostImpl.

LayerTreeTest gets two methods that can be overridden to control the
output surface:
- CreateDelegatingOutputSurface makes a TestDelegatingOutputSurface
so tests can control the arguments passed to its constructor, or the
ContextProviders given to it.
- CreateDisplayOutputSurface makes an OutputSurface subclass that
will be owned by the Display and used by the DirectRenderer. This
allows tests to control which ContextProvider it uses, and what
subclass of OutputSurface to use, to allow mocking out things as
desired.

R=enne
BUG=606056
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel

Committed: https://crrev.com/cae1058e62b7960796709512c4e1e650b6389c0f
Cr-Commit-Position: refs/heads/master@{#409270}
==========

[danakj, 2016-08-04 17:13:54]

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

[commit-bot: I haz the power, 2016-08-04 17:14:14]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[danakj, 2016-08-04 17:33:12]

The CQ bit was checked by danakj@chromium.org

[danakj, 2016-08-04 17:33:13]

The patchset sent to the CQ was uploaded after l-g-t-m from
fsamuel@chromium.org, enne@chromium.org, boliu@chromium.org, piman@chromium.org
Link to the patchset: https://codereview.chromium.org/2193293004/#ps160001
(title: "display-layertreetest: withperftestsfix")

[danakj, 2016-08-04 17:33:37]

The CQ bit was unchecked by danakj@chromium.org

[commit-bot: I haz the power, 2016-08-04 17:33:41]

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

[danakj, 2016-08-04 17:34:07]

Description was changed from

==========
cc: Make LayerTreeTests use a DelegatingRenderer and Display.

Wherein we rewrite a few thousand cc unit_tests to not use a
DirectRenderer in LayerTreeHostImpl so that we can remove that
possibility from the codebase.

Currently some LayerTreeTests use a DirectRenderer in the
LayerTreeHostImpl, which is the last case of this occuring. Instead
give all LayerTreeTests a TestDelegatingOutputSurface (which means
a DelegatingRenderer in LayerTreeHostImpl). TestDelegatingOutputSurface
delegates drawing to a DirectRenderer via Display, which matches how
things "really work" now.

This means SwapBuffers hooks become async from operations on
LayerTreeHostImpl, so we introduce 3 new hooks for tests:
- DisplayReceivedCompositorFrameOnThread.
- DisplayWillDrawAndSwapOnThread.
- DisplayDidDrawAndSwapOnThread.
None of these receive a LayerTreeHostImpl* since they are async
from it and so using its state from the hook would be racey. These
hooks come from the TestDelegatingOutputSurface instead of from the
LayerTreeHostImpl.

LayerTreeTest gets two methods that can be overridden to control the
output surface:
- CreateDelegatingOutputSurface makes a TestDelegatingOutputSurface
so tests can control the arguments passed to its constructor, or the
ContextProviders given to it.
- CreateDisplayOutputSurface makes an OutputSurface subclass that
will be owned by the Display and used by the DirectRenderer. This
allows tests to control which ContextProvider it uses, and what
subclass of OutputSurface to use, to allow mocking out things as
desired.

R=enne
BUG=606056
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel

Committed: https://crrev.com/cae1058e62b7960796709512c4e1e650b6389c0f
Cr-Commit-Position: refs/heads/master@{#409270}
==========

to

==========
cc: Make LayerTreeTests use a DelegatingRenderer and Display.

Wherein we rewrite a few thousand cc unit_tests to not use a
DirectRenderer in LayerTreeHostImpl so that we can remove that
possibility from the codebase.

Currently some LayerTreeTests use a DirectRenderer in the
LayerTreeHostImpl, which is the last case of this occuring. Instead
give all LayerTreeTests a TestDelegatingOutputSurface (which means
a DelegatingRenderer in LayerTreeHostImpl). TestDelegatingOutputSurface
delegates drawing to a DirectRenderer via Display, which matches how
things "really work" now.

This means SwapBuffers hooks become async from operations on
LayerTreeHostImpl, so we introduce 3 new hooks for tests:
- DisplayReceivedCompositorFrameOnThread.
- DisplayWillDrawAndSwapOnThread.
- DisplayDidDrawAndSwapOnThread.
None of these receive a LayerTreeHostImpl* since they are async
from it and so using its state from the hook would be racey. These
hooks come from the TestDelegatingOutputSurface instead of from the
LayerTreeHostImpl.

LayerTreeTest gets two methods that can be overridden to control the
output surface:
- CreateDelegatingOutputSurface makes a TestDelegatingOutputSurface
so tests can control the arguments passed to its constructor, or the
ContextProviders given to it.
- CreateDisplayOutputSurface makes an OutputSurface subclass that
will be owned by the Display and used by the DirectRenderer. This
allows tests to control which ContextProvider it uses, and what
subclass of OutputSurface to use, to allow mocking out things as
desired.

Now also with fixes for cc_perftests:

The TestDelegatingOutputSurface was not honoring disable_display_vsync
so they were not saturating the thread, making them take much longer
than 5s to do 5s of thread work, and they'd time out.

Also the TetureMailbox doesn't get released to the main thread
immediately after commit anymore, as you need to wait for the
mailbox to be given to the Display (DisplayReceivedCompositorFrame
is the hook for that fwiw). So we wait for the mailbox to actually
return before ending the test to avoid a crash.

R=enne
BUG=606056
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel

Committed: https://crrev.com/cae1058e62b7960796709512c4e1e650b6389c0f
Cr-Commit-Position: refs/heads/master@{#409270}
==========

[danakj, 2016-08-04 17:34:16]

The CQ bit was checked by danakj@chromium.org

[commit-bot: I haz the power, 2016-08-04 17:35:03]

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, 2016-08-04 18:40:42]

Description was changed from

==========
cc: Make LayerTreeTests use a DelegatingRenderer and Display.

Wherein we rewrite a few thousand cc unit_tests to not use a
DirectRenderer in LayerTreeHostImpl so that we can remove that
possibility from the codebase.

Currently some LayerTreeTests use a DirectRenderer in the
LayerTreeHostImpl, which is the last case of this occuring. Instead
give all LayerTreeTests a TestDelegatingOutputSurface (which means
a DelegatingRenderer in LayerTreeHostImpl). TestDelegatingOutputSurface
delegates drawing to a DirectRenderer via Display, which matches how
things "really work" now.

This means SwapBuffers hooks become async from operations on
LayerTreeHostImpl, so we introduce 3 new hooks for tests:
- DisplayReceivedCompositorFrameOnThread.
- DisplayWillDrawAndSwapOnThread.
- DisplayDidDrawAndSwapOnThread.
None of these receive a LayerTreeHostImpl* since they are async
from it and so using its state from the hook would be racey. These
hooks come from the TestDelegatingOutputSurface instead of from the
LayerTreeHostImpl.

LayerTreeTest gets two methods that can be overridden to control the
output surface:
- CreateDelegatingOutputSurface makes a TestDelegatingOutputSurface
so tests can control the arguments passed to its constructor, or the
ContextProviders given to it.
- CreateDisplayOutputSurface makes an OutputSurface subclass that
will be owned by the Display and used by the DirectRenderer. This
allows tests to control which ContextProvider it uses, and what
subclass of OutputSurface to use, to allow mocking out things as
desired.

Now also with fixes for cc_perftests:

The TestDelegatingOutputSurface was not honoring disable_display_vsync
so they were not saturating the thread, making them take much longer
than 5s to do 5s of thread work, and they'd time out.

Also the TetureMailbox doesn't get released to the main thread
immediately after commit anymore, as you need to wait for the
mailbox to be given to the Display (DisplayReceivedCompositorFrame
is the hook for that fwiw). So we wait for the mailbox to actually
return before ending the test to avoid a crash.

R=enne
BUG=606056
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel

Committed: https://crrev.com/cae1058e62b7960796709512c4e1e650b6389c0f
Cr-Commit-Position: refs/heads/master@{#409270}
==========

to

==========
cc: Make LayerTreeTests use a DelegatingRenderer and Display.

Wherein we rewrite a few thousand cc unit_tests to not use a
DirectRenderer in LayerTreeHostImpl so that we can remove that
possibility from the codebase.

Currently some LayerTreeTests use a DirectRenderer in the
LayerTreeHostImpl, which is the last case of this occuring. Instead
give all LayerTreeTests a TestDelegatingOutputSurface (which means
a DelegatingRenderer in LayerTreeHostImpl). TestDelegatingOutputSurface
delegates drawing to a DirectRenderer via Display, which matches how
things "really work" now.

This means SwapBuffers hooks become async from operations on
LayerTreeHostImpl, so we introduce 3 new hooks for tests:
- DisplayReceivedCompositorFrameOnThread.
- DisplayWillDrawAndSwapOnThread.
- DisplayDidDrawAndSwapOnThread.
None of these receive a LayerTreeHostImpl* since they are async
from it and so using its state from the hook would be racey. These
hooks come from the TestDelegatingOutputSurface instead of from the
LayerTreeHostImpl.

LayerTreeTest gets two methods that can be overridden to control the
output surface:
- CreateDelegatingOutputSurface makes a TestDelegatingOutputSurface
so tests can control the arguments passed to its constructor, or the
ContextProviders given to it.
- CreateDisplayOutputSurface makes an OutputSurface subclass that
will be owned by the Display and used by the DirectRenderer. This
allows tests to control which ContextProvider it uses, and what
subclass of OutputSurface to use, to allow mocking out things as
desired.

Now also with fixes for cc_perftests:

The TestDelegatingOutputSurface was not honoring disable_display_vsync
so they were not saturating the thread, making them take much longer
than 5s to do 5s of thread work, and they'd time out.

Also the TetureMailbox doesn't get released to the main thread
immediately after commit anymore, as you need to wait for the
mailbox to be given to the Display (DisplayReceivedCompositorFrame
is the hook for that fwiw). So we wait for the mailbox to actually
return before ending the test to avoid a crash.

R=enne
BUG=606056
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel

Committed: https://crrev.com/cae1058e62b7960796709512c4e1e650b6389c0f
Cr-Commit-Position: refs/heads/master@{#409270}
==========

[commit-bot: I haz the power, 2016-08-04 18:40:45]

Committed patchset #8 (id:160001)

[commit-bot: I haz the power, 2016-08-04 18:42:34]

Description was changed from

==========
cc: Make LayerTreeTests use a DelegatingRenderer and Display.

Wherein we rewrite a few thousand cc unit_tests to not use a
DirectRenderer in LayerTreeHostImpl so that we can remove that
possibility from the codebase.

Currently some LayerTreeTests use a DirectRenderer in the
LayerTreeHostImpl, which is the last case of this occuring. Instead
give all LayerTreeTests a TestDelegatingOutputSurface (which means
a DelegatingRenderer in LayerTreeHostImpl). TestDelegatingOutputSurface
delegates drawing to a DirectRenderer via Display, which matches how
things "really work" now.

This means SwapBuffers hooks become async from operations on
LayerTreeHostImpl, so we introduce 3 new hooks for tests:
- DisplayReceivedCompositorFrameOnThread.
- DisplayWillDrawAndSwapOnThread.
- DisplayDidDrawAndSwapOnThread.
None of these receive a LayerTreeHostImpl* since they are async
from it and so using its state from the hook would be racey. These
hooks come from the TestDelegatingOutputSurface instead of from the
LayerTreeHostImpl.

LayerTreeTest gets two methods that can be overridden to control the
output surface:
- CreateDelegatingOutputSurface makes a TestDelegatingOutputSurface
so tests can control the arguments passed to its constructor, or the
ContextProviders given to it.
- CreateDisplayOutputSurface makes an OutputSurface subclass that
will be owned by the Display and used by the DirectRenderer. This
allows tests to control which ContextProvider it uses, and what
subclass of OutputSurface to use, to allow mocking out things as
desired.

Now also with fixes for cc_perftests:

The TestDelegatingOutputSurface was not honoring disable_display_vsync
so they were not saturating the thread, making them take much longer
than 5s to do 5s of thread work, and they'd time out.

Also the TetureMailbox doesn't get released to the main thread
immediately after commit anymore, as you need to wait for the
mailbox to be given to the Display (DisplayReceivedCompositorFrame
is the hook for that fwiw). So we wait for the mailbox to actually
return before ending the test to avoid a crash.

R=enne
BUG=606056
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel

Committed: https://crrev.com/cae1058e62b7960796709512c4e1e650b6389c0f
Cr-Commit-Position: refs/heads/master@{#409270}
==========

to

==========
cc: Make LayerTreeTests use a DelegatingRenderer and Display.

Wherein we rewrite a few thousand cc unit_tests to not use a
DirectRenderer in LayerTreeHostImpl so that we can remove that
possibility from the codebase.

Currently some LayerTreeTests use a DirectRenderer in the
LayerTreeHostImpl, which is the last case of this occuring. Instead
give all LayerTreeTests a TestDelegatingOutputSurface (which means
a DelegatingRenderer in LayerTreeHostImpl). TestDelegatingOutputSurface
delegates drawing to a DirectRenderer via Display, which matches how
things "really work" now.

This means SwapBuffers hooks become async from operations on
LayerTreeHostImpl, so we introduce 3 new hooks for tests:
- DisplayReceivedCompositorFrameOnThread.
- DisplayWillDrawAndSwapOnThread.
- DisplayDidDrawAndSwapOnThread.
None of these receive a LayerTreeHostImpl* since they are async
from it and so using its state from the hook would be racey. These
hooks come from the TestDelegatingOutputSurface instead of from the
LayerTreeHostImpl.

LayerTreeTest gets two methods that can be overridden to control the
output surface:
- CreateDelegatingOutputSurface makes a TestDelegatingOutputSurface
so tests can control the arguments passed to its constructor, or the
ContextProviders given to it.
- CreateDisplayOutputSurface makes an OutputSurface subclass that
will be owned by the Display and used by the DirectRenderer. This
allows tests to control which ContextProvider it uses, and what
subclass of OutputSurface to use, to allow mocking out things as
desired.

Now also with fixes for cc_perftests:

The TestDelegatingOutputSurface was not honoring disable_display_vsync
so they were not saturating the thread, making them take much longer
than 5s to do 5s of thread work, and they'd time out.

Also the TetureMailbox doesn't get released to the main thread
immediately after commit anymore, as you need to wait for the
mailbox to be given to the Display (DisplayReceivedCompositorFrame
is the hook for that fwiw). So we wait for the mailbox to actually
return before ending the test to avoid a crash.

R=enne
BUG=606056
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel

Committed: https://crrev.com/cae1058e62b7960796709512c4e1e650b6389c0f
Committed: https://crrev.com/014316e35eb9a90a024302daa3a6c3ca3983a8bd
Cr-Original-Commit-Position: refs/heads/master@{#409270}
Cr-Commit-Position: refs/heads/master@{#409845}
==========

[commit-bot: I haz the power, 2016-08-04 18:42:36]

Patchset 8 (id:??) landed as
https://crrev.com/014316e35eb9a90a024302daa3a6c3ca3983a8bd
Cr-Commit-Position: refs/heads/master@{#409845}