IdentityLaunchWebAuthFlowFunction shouldn't be able to send response twice.

IdentityLaunchWebAuthFlowFunction shouldn't be able to send response twice.

It is possible before this change, because when response it sent, WebAuthFlow
isn't stopped or destroyed and it still can call its delegate methods.

BUG=

Committed: https://crrev.com/ee37f1b7c6da834dec9056283cf83d88b0f2f53c
Cr-Commit-Position: refs/heads/master@{#311028}

[Lukasz Jagielski, 2015-01-08 14:19:33]

ljagielski@opera.com changed reviewers:
+ rogerta@chromium.org

[Lukasz Jagielski, 2015-01-08 14:20:37]

Hi,
PTAL at my CL proposition.
Regards,
Łukasz

[Roger Tawa OOO till Jul 10th, 2015-01-08 22:05:53]

rogerta@chromium.org changed reviewers:
+ courage@chromium.org

[Roger Tawa OOO till Jul 10th, 2015-01-08 22:05:53]

I'm not familiar with this code, adding Michael as reviewer who owns this code.

[Michael Courage, 2015-01-08 22:45:14]

Hi Łukasz, out of curiosity have there been cases where multiple responses were
sent, or was the bug found through code review? This seems unlikely to happen in
practice since I would expect the IdentityLaunchWebAuthFlowFunction to be
destroyed quickly, though it is better to explicitly clean up rather than
relying on this implementation detail.

[Michael Courage, 2015-01-08 22:45:27]

https://codereview.chromium.org/829853006/diff/1/chrome/browser/extensions/ap...
File chrome/browser/extensions/api/identity/identity_api.cc (right):

https://codereview.chromium.org/829853006/diff/1/chrome/browser/extensions/ap...
chrome/browser/extensions/api/identity/identity_api.cc:946: Release();  //
Balanced in RunAsync.
Would we not also want to detach here as well?

[Lukasz Jagielski, 2015-01-09 23:00:50]

Hi,
Test LaunchWebAuthFlowFunctionTest.NonInteractiveSuccess was flaky for me in
certain circumstances and hw setup. IdentityLaunchWebAuthFlowFunction sends
response when the resource load starts. The handling of response quits the
message loop but there can be a resource load fail pending in the queue and the
response will be sent for the second time.
Regards,
Łukasz

[Michael Courage, 2015-01-09 23:24:07]

lgtm, thanks!

[Lukasz Jagielski, 2015-01-12 08:32:23]

The CQ bit was checked by ljagielski@opera.com

[commit-bot: I haz the power, 2015-01-12 08:32:56]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/829853006/20001

[commit-bot: I haz the power, 2015-01-12 09:30:29]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2015-01-12 09:31:18]

Patchset 2 (id:??) landed as
https://crrev.com/ee37f1b7c6da834dec9056283cf83d88b0f2f53c
Cr-Commit-Position: refs/heads/master@{#311028}