RenderViewTests: Implement proper shutdown.

RenderViewTests: Implement proper shutdown.

Interleave gc cycles with message loop pumping.
Sequentially calling GC didn't work so we had memory leaks.

Depends on blink part: https://codereview.chromium.org/1360723003/

BUG=506433
BUG=46571
BUG=484760
Committed: https://crrev.com/9540f3e597d917dc3c33eaf7e61f5a9feb50fc1e
Cr-Commit-Position: refs/heads/master@{#350298}

[loyso (OOO), 2015-09-22 04:52:32]

loyso@chromium.org changed reviewers:
+ glider@chromium.org, piman@chromium.org, thakis@chromium.org

[loyso (OOO), 2015-09-22 04:53:45]

loyso@chromium.org changed reviewers:
+ haraken@chromium.org

[haraken, 2015-09-22 06:20:38]

LGTM

[loyso (OOO), 2015-09-22 06:34:47]

loyso@chromium.org changed reviewers:
+ jochen@chromium.org, phajdan.jr@chromium.org

[loyso (OOO), 2015-09-22 06:35:31]

render_view_test owners, PTAL anyone?

[Alexander Potapenko, 2015-09-22 09:25:31]

On 2015/09/22 06:35:31, loyso wrote:
> render_view_test owners, PTAL anyone?

Can you please elaborate what are you going to shut down properly?
I'm particularly interested in whether content_browsertests start respecting the
exit codes of child processes upon your change.

[Paweł Hajdan Jr., 2015-09-22 10:32:57]

Could you change the commit message to say RenderViewTest as opposed to Content
Browser Tests?

I can stamp afterwards.

[piman, 2015-09-22 17:30:40]

rs lgtm

[loyso (OOO), 2015-09-23 01:00:06]

On 2015/09/22 10:32:57, Paweł Hajdan Jr. wrote:
> Could you change the commit message to say RenderViewTest as opposed to
Content
> Browser Tests?
> I can stamp afterwards.
Done!

[loyso (OOO), 2015-09-23 01:47:35]

On 2015/09/22 09:25:31, Alexander Potapenko wrote:
> On 2015/09/22 06:35:31, loyso wrote:
> > render_view_test owners, PTAL anyone?
> 
> Can you please elaborate what are you going to shut down properly?
> I'm particularly interested in whether content_browsertests start respecting
the
> exit codes of child processes upon your change.

I'm fixing this:
// Try very hard to collect garbage before shutting down.
// "5" was chosen following http://crbug.com/46571#c9

Also, LSAN build was partially useless with such a broad suppressions.

[loyso (OOO), 2015-09-23 01:58:53]

On 2015/09/22 09:25:31, Alexander Potapenko wrote:
> Can you please elaborate what are you going to shut down properly?
> I'm particularly interested in whether content_browsertests start respecting
the
> exit codes of child processes upon your change.
Also, you can read the story here:
https://code.google.com/p/chromium/issues/detail?id=497658#c18

[loyso (OOO), 2015-09-23 02:25:02]

RenderViewTests non-rs LGTM needed! Somebody? :)

[loyso (OOO), 2015-09-23 02:49:01]

Ooops - just realized that DCHECKs don't work in LSAN build. Looking for
replacement how to raise an ASAN/LSAN error. Any advices are welcome.

[loyso (OOO), 2015-09-23 02:59:43]

On 2015/09/23 02:49:01, loyso wrote:
> Ooops - just realized that DCHECKs don't work in LSAN build. Looking for
> replacement how to raise an ASAN/LSAN error. Any advices are welcome.
That's EXPECT_EQ. PTAL!

[piman, 2015-09-23 03:00:03]

lgtm

[loyso (OOO), 2015-09-23 04:22:13]

The CQ bit was checked by loyso@chromium.org

[loyso (OOO), 2015-09-23 04:22:13]

The patchset sent to the CQ was uploaded after l-g-t-m from haraken@chromium.org
Link to the patchset: https://codereview.chromium.org/1363523002/#ps40001
(title: "Use EXPECT_EQ instead of DCHECK_EQ.")

[commit-bot: I haz the power, 2015-09-23 04:23:20]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1363523002/40001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1363523002/40001

[commit-bot: I haz the power, 2015-09-23 04:29:55]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2015-09-23 04:30:53]

Patchset 3 (id:??) landed as
https://crrev.com/9540f3e597d917dc3c33eaf7e61f5a9feb50fc1e
Cr-Commit-Position: refs/heads/master@{#350298}

[loyso (OOO), 2015-09-23 07:37:27]

To be continued: https://codereview.chromium.org/1364673002/

[loyso (OOO), 2015-09-23 07:51:13]

haraken isn't an owner for lsan_suppressions.cc but chromium_presubmit treats it
ok :(