Add base::test::ScopedAsyncTaskScheduler.

Add base::test::ScopedAsyncTaskScheduler.

base::test::ScopedAsyncTaskScheduler allows usage of the
base/task_scheduler/post_task.h API within its scope. Tasks run
asynchronously, one at a time.

This is required to allow the migration of v8 from WorkerPool to
TaskScheduler. Currently, v8 posts tasks to WorkerPool and waits
until they run using a v8::base::Semaphore. Posting these tasks to
a synchronous TaskScheduler would result in a deadlock (tasks can't
run because the main thread is waiting on a semaphore, the wait
cannot end unless tasks run). Posting these tasks to an
asynchronous TaskScheduler would work as desired.

BUG=553459
Review-Url: https://codereview.chromium.org/2622633004
Cr-Commit-Position: refs/heads/master@{#442364}
Committed: https://chromium.googlesource.com/chromium/src/+/83519fc19e33362d2221fb2767e1eff7649eef84

[fdoray, 2017-01-09 19:46:48]

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

[fdoray, 2017-01-09 19:46:58]

fdoray@chromium.org changed reviewers:
+ gab@chromium.org, robliao@chromium.org

[fdoray, 2017-01-09 19:46:58]

PTAL

[commit-bot: I haz the power, 2017-01-09 19:47:22]

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

[gab, 2017-01-09 19:55:47]

lgtm % comment nit

https://codereview.chromium.org/2622633004/diff/20001/base/test/scoped_async_...
File base/test/scoped_async_task_scheduler.cc (right):

https://codereview.chromium.org/2622633004/diff/20001/base/test/scoped_async_...
base/test/scoped_async_task_scheduler.cc:20: const TimeDelta
kSuggestedReclaimTime = TimeDelta::Max();
constexpr

https://codereview.chromium.org/2622633004/diff/20001/base/test/scoped_async_...
File base/test/scoped_async_task_scheduler.h (right):

https://codereview.chromium.org/2622633004/diff/20001/base/test/scoped_async_...
base/test/scoped_async_task_scheduler.h:24: // deterministic.
Does it really make them more deterministic? If the proper flushes are in place,
the tasks are still posted in order and I don't think it increases determinism.
It does make the tests simpler though because everything all side-effects
(posted tasks) from a task are resolved in a single block (rephrase to highlight
that instead?)

[fdoray, 2017-01-09 20:23:48]

https://codereview.chromium.org/2622633004/diff/20001/base/test/scoped_async_...
File base/test/scoped_async_task_scheduler.cc (right):

https://codereview.chromium.org/2622633004/diff/20001/base/test/scoped_async_...
base/test/scoped_async_task_scheduler.cc:20: const TimeDelta
kSuggestedReclaimTime = TimeDelta::Max();
On 2017/01/09 19:55:46, gab wrote:
> constexpr

TimeDelta::Max is not constexpr (yet) :(

https://codereview.chromium.org/2622633004/diff/20001/base/test/scoped_async_...
File base/test/scoped_async_task_scheduler.h (right):

https://codereview.chromium.org/2622633004/diff/20001/base/test/scoped_async_...
base/test/scoped_async_task_scheduler.h:24: // deterministic.
On 2017/01/09 19:55:46, gab wrote:
> Does it really make them more deterministic? If the proper flushes are in
place,
> the tasks are still posted in order and I don't think it increases
determinism.
> It does make the tests simpler though because everything all side-effects
> (posted tasks) from a task are resolved in a single block (rephrase to
highlight
> that instead?)

Done.

Note: With a ScopedAsyncTaskScheduler, the order in which instructions from the
test body and instructions from asynchronous tasks run is not deterministic. It
is with a ScopedTaskScheduler.

[fdoray, 2017-01-09 20:23:52]

The CQ bit was checked by fdoray@chromium.org

[fdoray, 2017-01-09 20:23:53]

The patchset sent to the CQ was uploaded after l-g-t-m from gab@chromium.org
Link to the patchset: https://codereview.chromium.org/2622633004/#ps40001
(title: "Cr gab #5")

[commit-bot: I haz the power, 2017-01-09 20:24:27]

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

[gab, 2017-01-09 20:27:04]

The CQ bit was unchecked by gab@chromium.org

[gab, 2017-01-09 20:27:11]

lgtm++

https://codereview.chromium.org/2622633004/diff/20001/base/test/scoped_async_...
File base/test/scoped_async_task_scheduler.cc (right):

https://codereview.chromium.org/2622633004/diff/20001/base/test/scoped_async_...
base/test/scoped_async_task_scheduler.cc:20: const TimeDelta
kSuggestedReclaimTime = TimeDelta::Max();
On 2017/01/09 20:23:48, fdoray wrote:
> On 2017/01/09 19:55:46, gab wrote:
> > constexpr
> 
> TimeDelta::Max is not constexpr (yet) :(

Ah weird, the rest of TimeDeltas are...

https://codereview.chromium.org/2622633004/diff/40001/base/test/scoped_async_...
File base/test/scoped_async_task_scheduler.h (right):

https://codereview.chromium.org/2622633004/diff/40001/base/test/scoped_async_...
base/test/scoped_async_task_scheduler.h:23: // the scope of a
ScopedTaskScheduler run synchronously, which make tests easier
nit: s/make/makes/

[fdoray, 2017-01-09 20:28:34]

https://codereview.chromium.org/2622633004/diff/40001/base/test/scoped_async_...
File base/test/scoped_async_task_scheduler.h (right):

https://codereview.chromium.org/2622633004/diff/40001/base/test/scoped_async_...
base/test/scoped_async_task_scheduler.h:23: // the scope of a
ScopedTaskScheduler run synchronously, which make tests easier
On 2017/01/09 20:27:11, gab wrote:
> nit: s/make/makes/

Done.

[fdoray, 2017-01-09 20:28:38]

The CQ bit was checked by fdoray@chromium.org

[fdoray, 2017-01-09 20:28:39]

The patchset sent to the CQ was uploaded after l-g-t-m from gab@chromium.org
Link to the patchset: https://codereview.chromium.org/2622633004/#ps60001
(title: "CR gab #11 (grammar)")

[commit-bot: I haz the power, 2017-01-09 20:29:07]

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, 2017-01-09 21:48:23]

CQ is committing da patch.

Bot data: {"patchset_id": 60001, "attempt_start_ts": 1483993718837840,
"parent_rev": "5829c341bc47c0591d14414eadf4e4635347e69b", "commit_rev":
"83519fc19e33362d2221fb2767e1eff7649eef84"}

[commit-bot: I haz the power, 2017-01-09 21:48:48]

Description was changed from

==========
Add base::test::ScopedAsyncTaskScheduler.

base::test::ScopedAsyncTaskScheduler allows usage of the
base/task_scheduler/post_task.h API within its scope. Tasks run
asynchronously, one at a time.

This is required to allow the migration of v8 from WorkerPool to
TaskScheduler. Currently, v8 posts tasks to WorkerPool and waits
until they run using a v8::base::Semaphore. Posting these tasks to
a synchronous TaskScheduler would result in a deadlock (tasks can't
run because the main thread is waiting on a semaphore, the wait
cannot end unless tasks run). Posting these tasks to an
asynchronous TaskScheduler would work as desired.

BUG=553459
==========

to

==========
Add base::test::ScopedAsyncTaskScheduler.

base::test::ScopedAsyncTaskScheduler allows usage of the
base/task_scheduler/post_task.h API within its scope. Tasks run
asynchronously, one at a time.

This is required to allow the migration of v8 from WorkerPool to
TaskScheduler. Currently, v8 posts tasks to WorkerPool and waits
until they run using a v8::base::Semaphore. Posting these tasks to
a synchronous TaskScheduler would result in a deadlock (tasks can't
run because the main thread is waiting on a semaphore, the wait
cannot end unless tasks run). Posting these tasks to an
asynchronous TaskScheduler would work as desired.

BUG=553459

Review-Url: https://codereview.chromium.org/2622633004
Cr-Commit-Position: refs/heads/master@{#442364}
Committed:
https://chromium.googlesource.com/chromium/src/+/83519fc19e33362d2221fb2767e1...
==========

[commit-bot: I haz the power, 2017-01-09 21:48:49]

Committed patchset #4 (id:60001) as
https://chromium.googlesource.com/chromium/src/+/83519fc19e33362d2221fb2767e1...