PasswordAutofillAgent: avoid NULL frames in provisionally_saved_forms_

PasswordAutofillAgent: avoid NULL frames in provisionally_saved_forms_

This should stop the NULL dereference reported in the bug referenced below. It only targets the symptom, not the cause, though. The proper fix will hopefully come with a planned refactoring as referenced in the code comment.

BUG=420519
TBR=isherman@chromium.org

Committed: https://crrev.com/d4a9b6f46e47662f75ebee6d777611329d2f2673
Cr-Commit-Position: refs/heads/master@{#300143}

[vabr (Chromium), 2014-10-17 15:54:19]

vabr@chromium.org changed reviewers:
+ mkwst@chromium.org

[vabr (Chromium), 2014-10-17 15:54:20]

Please have a look.

My biggest concern is that by masking the symptom, I won't be able to find the
cause. But I don't seem able to find it anyway, and this is a stable blocker.

Cheers,
Vaclav

[Mike West, 2014-10-17 15:58:27]

LGTM. Landing this to prevent the null deref is a reasonable short-term
solution.

Another approach to consider would be to convert the DCHECK to a CHECK in the
hopes of getting more data. We certainly wouldn't want to leave that in, of
course, but it might be reasonable in the medium term if the refactoring doesn't
work well.

[vabr (Chromium), 2014-10-17 15:59:32]

The CQ bit was checked by vabr@chromium.org

[commit-bot: I haz the power, 2014-10-17 16:00:19]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/666503002/1

[commit-bot: I haz the power, 2014-10-17 16:07:25]

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

[commit-bot: I haz the power, 2014-10-17 16:07:27]

Try jobs failed on following builders:
  chromium_presubmit on tryserver.chromium.linux
(http://build.chromium.org/p/tryserver.chromium.linux/builders/chromium_presub...)

[vabr (Chromium), 2014-10-17 16:10:41]

vabr@chromium.org changed reviewers:
+ isherman@chromium.org

[vabr (Chromium), 2014-10-17 16:10:41]

After https://crrev.com/26c1c52ff4376cb78f9ffb6098f237d80952aa59, I should be an
OWNER for the changed file.

I'll investigate why this did not work with the CQ, but for the time being,
TBR-ing Ilya.

Cheers,
Vaclav

[vabr (Chromium), 2014-10-17 16:11:21]

The CQ bit was checked by vabr@chromium.org

[commit-bot: I haz the power, 2014-10-17 16:12:18]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/666503002/1

[commit-bot: I haz the power, 2014-10-17 16:29:24]

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

[commit-bot: I haz the power, 2014-10-17 16:29:25]

Try jobs failed on following builders:
  linux_gpu_triggered_tests on tryserver.chromium.gpu
(http://build.chromium.org/p/tryserver.chromium.gpu/builders/linux_gpu_trigger...)

[vabr (Chromium), 2014-10-17 19:32:11]

The CQ bit was checked by vabr@chromium.org

[commit-bot: I haz the power, 2014-10-17 19:33:15]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/666503002/1

[commit-bot: I haz the power, 2014-10-17 19:34:45]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2014-10-17 19:36:15]

Patchset 1 (id:??) landed as
https://crrev.com/d4a9b6f46e47662f75ebee6d777611329d2f2673
Cr-Commit-Position: refs/heads/master@{#300143}

[Ilya Sherman, 2014-10-17 19:37:50]

LGTM.  I'd love to understand why your and Garrett's OWNERS status is being
ignored by the tools...

[vabr (Chromium), 2014-10-17 19:51:17]

On 2014/10/17 19:37:50, Ilya Sherman wrote:
> LGTM.  I'd love to understand why your and Garrett's OWNERS status is being
> ignored by the tools...

Thanks, Ilya!
I'm currently looking in the presubmit script and its dependencies, and
hopefully will be able to find out and fix it. Maybe the per-file directive uses
a different syntax to include subdirectories.

I'll go to bed soon, but will get back to figuring this out on Monday.

Cheers,
Vaclav