Remove the kDisableGpuProcessPrelaunch flag and disable tests.

Remove the kDisableGpuProcessPrelaunch flag and disable tests.

Tests that depend on this flag actually fail since the UI compositor
creates an OutputSurface which makes the GPU process launch before
the test body gets to run.

Since this flag doesn't really make sense with a UI compositor present,
we should move these tests to not be browser tests. Since they just
pass by early-out right now, remove the early outs and disable them.

R=jbauman@chromium.org, piman@chromium.org
BUG=356876
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=260083

[piman, 2014-03-26 23:48:09]

LGTM I guess, though I'm sad these regressed :(

[danakj, 2014-03-26 23:48:58]

On 2014/03/26 23:48:09, piman wrote:
> LGTM I guess, though I'm sad these regressed :(

Ya :/ I'm not sure how to fix them right now. If you have ideas let me know. I'd
rather have it in the code that theyre disabled than fake passing.

[danakj, 2014-03-26 23:49:02]

The CQ bit was checked by danakj@chromium.org

[commit-bot: I haz the power, 2014-03-26 23:52:47]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/danakj@chromium.org/212603006/1

[no sievers, 2014-03-27 00:00:45]

Where can we fit these tests though? I guess we need multi-process but without
UI compositor.

[danakj, 2014-03-27 00:05:13]

On 2014/03/27 00:00:45, sievers wrote:
> Where can we fit these tests though? I guess we need multi-process but without
> UI compositor.

There are some tests in unit_tests that have a child process FWIW.

[danakj, 2014-03-27 00:57:52]

On 2014/03/27 00:05:13, danakj wrote:
> On 2014/03/27 00:00:45, sievers wrote:
> > Where can we fit these tests though? I guess we need multi-process but
without
> > UI compositor.
> 
> There are some tests in unit_tests that have a child process FWIW.

This is one multi-process test, for posterity: MacSandboxTest.FontLoadingTest

https://code.google.com/p/chromium/codesearch#chromium/src/content/common/san...

I'm not sure what we need to have a functional IPC system though, that test just
executes a child process and runs stuff in it.

Or maybe we could run the Gpu process on a thread in a unit test instead of
another process? I'm not sure if that'd give the same kind of test coverage but
maybe it would. Also not sure if we can simulate GPU crashes in a thread without
destroying the whole process.

[commit-bot: I haz the power, 2014-03-27 07:36:40]

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

[commit-bot: I haz the power, 2014-03-27 07:36:40]

Commit queue rejected this change because the description was changed
between the time the change entered the commit queue and the time it
was ready to commit. You can safely check the commit box again.

[danakj, 2014-03-28 06:02:09]

The CQ bit was checked by danakj@chromium.org

[commit-bot: I haz the power, 2014-03-28 06:03:31]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/danakj@chromium.org/212603006/1

[commit-bot: I haz the power, 2014-03-28 07:22:07]

Change committed as 260083

[no sievers, 2014-03-29 20:50:19]

https://codereview.chromium.org/212603006/diff/1/content/browser/browser_main...
File content/browser/browser_main_loop.cc (right):

https://codereview.chromium.org/212603006/diff/1/content/browser/browser_main...
content/browser/browser_main_loop.cc:961: established_gpu_channel =
Should be initialized to |true| I think.

[danakj, 2014-03-31 14:40:42]

https://codereview.chromium.org/212603006/diff/1/content/browser/browser_main...
File content/browser/browser_main_loop.cc (right):

https://codereview.chromium.org/212603006/diff/1/content/browser/browser_main...
content/browser/browser_main_loop.cc:961: established_gpu_channel =
On 2014/03/29 20:50:19, sievers wrote:
> Should be initialized to |true| I think.

Oh, yah, thanks.