Two TaskManager tests are flaky on debug due to a DCHECK in SharedSampler

Two TaskManager tests are flaky on debug due to a DCHECK in SharedSampler.

This is happening due to a Refresh called from a task group
which process_id is zero.

I am not sure if calling Refresh when process_id is still
zero should be supported. Perhaps that is an issue in
TaskManagerImpl / TaskGroup code, I think
https://crbug.com/521197 might be related. For now the
quick fix is move the DCHECK a bit further in the code so
it doesn't get triggered in that case.

BUG=642482
Committed: https://crrev.com/15ee7f73c882bb282d80225c50dd50c678438696
Cr-Commit-Position: refs/heads/master@{#415663}

[stanisc, 2016-08-31 01:09:23]

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

[commit-bot: I haz the power, 2016-08-31 01:09:46]

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

[stanisc, 2016-08-31 01:12:36]

Description was changed from

==========
TaskManagerBrowserTest.DevToolsNewUndockedWindow and
AppBackgroundPageApiTest.NoJsBackgroundPage flaky on debug

BUG=642482
==========

to

==========
TaskManagerBrowserTest.DevToolsNewUndockedWindow and
AppBackgroundPageApiTest.NoJsBackgroundPage flaky on debug

This is happenning due to a Refresh called from a task group
which process_id is zero. I am not sure if it is a problem
with TaskManagerImpl / TaskGroup implementation, but for
now I am fixing it locally by moving a DCHECK to a
different line, after the condition for zero process_id
is checked.

BUG=642482
==========

[stanisc, 2016-08-31 01:12:50]

Description was changed from

==========
TaskManagerBrowserTest.DevToolsNewUndockedWindow and
AppBackgroundPageApiTest.NoJsBackgroundPage flaky on debug

This is happenning due to a Refresh called from a task group
which process_id is zero. I am not sure if it is a problem
with TaskManagerImpl / TaskGroup implementation, but for
now I am fixing it locally by moving a DCHECK to a
different line, after the condition for zero process_id
is checked.

BUG=642482
==========

to

==========
TaskManagerBrowserTest.DevToolsNewUndockedWindow and
AppBackgroundPageApiTest.NoJsBackgroundPage flaky on debug

This is happening due to a Refresh called from a task group
which process_id is zero. I am not sure if it is a problem
with TaskManagerImpl / TaskGroup implementation, but for
now I am fixing it locally by moving a DCHECK to a
different line, after the condition for zero process_id
is checked.

BUG=642482
==========

[stanisc, 2016-08-31 01:20:28]

Description was changed from

==========
TaskManagerBrowserTest.DevToolsNewUndockedWindow and
AppBackgroundPageApiTest.NoJsBackgroundPage flaky on debug

This is happening due to a Refresh called from a task group
which process_id is zero. I am not sure if it is a problem
with TaskManagerImpl / TaskGroup implementation, but for
now I am fixing it locally by moving a DCHECK to a
different line, after the condition for zero process_id
is checked.

BUG=642482
==========

to

==========
Two TaskManager tests are flaky on debug due to a DCHECK in SharedSampler.

This is happening due to a Refresh called from a task group
which process_id is zero. SharedSampler::Refresh is already
checking for that but there is a DCHECK ealier in the code
that fails in that case.

I am not sure if calling Refresh when process_id is still
zero should be supported. Perhaps that is an issue in
TaskManagerImpl / TaskGroup code, I think
https://crbug.com/521197
might be related. For now the quick fix is
move the DCHECK a bit further in the code so it doesn't
get triggered in that case.

BUG=642482
==========

[stanisc, 2016-08-31 01:21:27]

Description was changed from

==========
Two TaskManager tests are flaky on debug due to a DCHECK in SharedSampler.

This is happening due to a Refresh called from a task group
which process_id is zero. SharedSampler::Refresh is already
checking for that but there is a DCHECK ealier in the code
that fails in that case.

I am not sure if calling Refresh when process_id is still
zero should be supported. Perhaps that is an issue in
TaskManagerImpl / TaskGroup code, I think
https://crbug.com/521197
might be related. For now the quick fix is
move the DCHECK a bit further in the code so it doesn't
get triggered in that case.

BUG=642482
==========

to

==========
Two TaskManager tests are flaky on debug due to a DCHECK in SharedSampler.

This is happening due to a Refresh called from a task group
which process_id is zero. 

I am not sure if calling Refresh when process_id is still
zero should be supported. Perhaps that is an issue in
TaskManagerImpl / TaskGroup code, I think
https://crbug.com/521197 might be related. For now the
quick fix is move the DCHECK a bit further in the code so 
it doesn't get triggered in that case.

BUG=642482
==========

[stanisc, 2016-08-31 01:22:02]

stanisc@chromium.org changed reviewers:
+ nick@chromium.org

[stanisc, 2016-08-31 01:22:03]

PTAL!

[commit-bot: I haz the power, 2016-08-31 03:14:10]

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

[commit-bot: I haz the power, 2016-08-31 03:14:10]

Dry run: This issue passed the CQ dry run.

[ncarter (slow), 2016-08-31 16:35:32]

lgtm to fix the bug.

I was mucking with this a few weeks ago; the pid==0 case can happen in the
period of time between (a) when a renderer process is successfully launched and
(b) we hear back its pid on the ui thread. in this period of time, the process
exists and we may actually be communicating with it, but we can't measure its
stats. It's gross.

[stanisc, 2016-08-31 16:54:08]

The CQ bit was checked by stanisc@chromium.org

[commit-bot: I haz the power, 2016-08-31 16:54:36]

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, 2016-08-31 16:58:17]

Description was changed from

==========
Two TaskManager tests are flaky on debug due to a DCHECK in SharedSampler.

This is happening due to a Refresh called from a task group
which process_id is zero. 

I am not sure if calling Refresh when process_id is still
zero should be supported. Perhaps that is an issue in
TaskManagerImpl / TaskGroup code, I think
https://crbug.com/521197 might be related. For now the
quick fix is move the DCHECK a bit further in the code so 
it doesn't get triggered in that case.

BUG=642482
==========

to

==========
Two TaskManager tests are flaky on debug due to a DCHECK in SharedSampler.

This is happening due to a Refresh called from a task group
which process_id is zero. 

I am not sure if calling Refresh when process_id is still
zero should be supported. Perhaps that is an issue in
TaskManagerImpl / TaskGroup code, I think
https://crbug.com/521197 might be related. For now the
quick fix is move the DCHECK a bit further in the code so 
it doesn't get triggered in that case.

BUG=642482
==========

[commit-bot: I haz the power, 2016-08-31 16:58:22]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2016-08-31 17:01:37]

Description was changed from

==========
Two TaskManager tests are flaky on debug due to a DCHECK in SharedSampler.

This is happening due to a Refresh called from a task group
which process_id is zero. 

I am not sure if calling Refresh when process_id is still
zero should be supported. Perhaps that is an issue in
TaskManagerImpl / TaskGroup code, I think
https://crbug.com/521197 might be related. For now the
quick fix is move the DCHECK a bit further in the code so 
it doesn't get triggered in that case.

BUG=642482
==========

to

==========
Two TaskManager tests are flaky on debug due to a DCHECK in SharedSampler.

This is happening due to a Refresh called from a task group
which process_id is zero.

I am not sure if calling Refresh when process_id is still
zero should be supported. Perhaps that is an issue in
TaskManagerImpl / TaskGroup code, I think
https://crbug.com/521197 might be related. For now the
quick fix is move the DCHECK a bit further in the code so
it doesn't get triggered in that case.

BUG=642482

Committed: https://crrev.com/15ee7f73c882bb282d80225c50dd50c678438696
Cr-Commit-Position: refs/heads/master@{#415663}
==========

[commit-bot: I haz the power, 2016-08-31 17:01:38]

Patchset 1 (id:??) landed as
https://crrev.com/15ee7f73c882bb282d80225c50dd50c678438696
Cr-Commit-Position: refs/heads/master@{#415663}