Small fix for URLRequestJobs when entering suspending mode.

Small fix for URLRequestJobs when entering suspending mode.

When entering suspend mode, there was a window of time between
when the request was canceled, and when the NetworkDelegate was
informed it was canceled.  This didn't exist for other cancellation
paths, and if an embedder invoked certain callbacks in this period,
thinking the request was still live, crashes could result.

BUG=289715
Committed: https://crrev.com/05608da225e05b4a20da3f4ca0cb9a5276fcf3d6
Cr-Commit-Position: refs/heads/master@{#350938}

[mmenke, 2015-09-25 17:02:54]

mmenke@chromium.org changed reviewers:
+ davidben@chromium.org

[mmenke, 2015-09-25 17:02:54]

I'm extremely skeptical that this will fix the crash, but I have verified that
this path did result in calling NetworkDelegate::NotifyCompleted too late,
though the added test doesn't test for that.  Checking that the message loop
isn't run between the OnSuspend() call and the NotifyCompleted call seems too
silly and regression prone, though we could certainly write a test that would
have crashed before this CL landed.

[davidben, 2015-09-25 21:24:55]

lgtm

https://codereview.chromium.org/1366203004/diff/20001/net/url_request/url_req...
File net/url_request/url_request_unittest.cc (right):

https://codereview.chromium.org/1366203004/diff/20001/net/url_request/url_req...
net/url_request/url_request_unittest.cc:3440:
base::MessageLoop::current()->Run();
Nit: base::RunLoop().Run(); ?

[mmenke, 2015-09-25 21:37:52]

Thanks!

https://codereview.chromium.org/1366203004/diff/20001/net/url_request/url_req...
File net/url_request/url_request_unittest.cc (right):

https://codereview.chromium.org/1366203004/diff/20001/net/url_request/url_req...
net/url_request/url_request_unittest.cc:3440:
base::MessageLoop::current()->Run();
On 2015/09/25 21:24:55, David Benjamin wrote:
> Nit: base::RunLoop().Run(); ?

Done (Not really sure...we're waiting on URLRequestDelegate to call
base::MessageLoop::current()->Quit().  Changing that to be more sane will be a
fun CL!).

Also fixed test name and description, and moved down in the file a bit.

[mmenke, 2015-09-25 21:37:56]

The CQ bit was checked by mmenke@chromium.org

[mmenke, 2015-09-25 21:37:56]

The patchset sent to the CQ was uploaded after l-g-t-m from
davidben@chromium.org
Link to the patchset: https://codereview.chromium.org/1366203004/#ps40001
(title: "Response to comments, fix test name/description")

[commit-bot: I haz the power, 2015-09-25 21:38:17]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1366203004/40001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1366203004/40001

[commit-bot: I haz the power, 2015-09-25 22:58:56]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2015-09-25 22:59:39]

Patchset 3 (id:??) landed as
https://crrev.com/05608da225e05b4a20da3f4ca0cb9a5276fcf3d6
Cr-Commit-Position: refs/heads/master@{#350938}

[mmenke, 2015-09-28 22:06:43]

On 2015/09/25 22:59:39, commit-bot: I haz the power wrote:
> Patchset 3 (id:??) landed as
> https://crrev.com/05608da225e05b4a20da3f4ca0cb9a5276fcf3d6
> Cr-Commit-Position: refs/heads/master@{#350938}

Seeing odd failures on android bots that may be due to this CL...Going to
revert.  :(

[mmenke, 2015-09-28 22:07:33]

A revert of this CL (patchset #3 id:40001) has been created in
https://codereview.chromium.org/1372213003/ by mmenke@chromium.org.

The reason for reverting is: Seeing odd failures that make no sense on android
bots that may somehow be due to this CL...Trying a revert.  :(.