ppapi: fix crash/infinite loop in VideoEncoder

ppapi: fix crash/infinite loop in VideoEncoder

Depending on the libc used with your NaCl plugins, if the VideoEncoder
proxy receives a EncodeReply message after it has been closed, the
plugin will crash (newlib) or enter an infinite loop (glibc) because
it tries to remove the end() iterator from a std::map.

BUG=455409
TEST=run ppapi/examples/video_encode multiple times and verify the test keeps recording new videos

Committed: https://crrev.com/86bd0448bb11826977519f852522267e403daeb0
Cr-Commit-Position: refs/heads/master@{#328997}

[llandwerlin-old, 2015-05-08 14:10:17]

Hi Bill, I just run into a bug in the video encoder. Please review this fix,
thanks.

[llandwerlin-old, 2015-05-08 14:10:45]

lionel.g.landwerlin@intel.com changed reviewers:
+ bbudge@chromium.org

[llandwerlin-old, 2015-05-08 14:10:45]

Hi Bill, I just run into a bug in the video encoder. Please review this fix,
thanks.

[llandwerlin-old, 2015-05-08 14:30:13]

I'm wondering whether I should prevent the renderer to emit any message after a
Close() message rather than this solution.

[bbudge, 2015-05-08 17:25:04]

On 2015/05/08 14:30:13, llandwerlin wrote:
> I'm wondering whether I should prevent the renderer to emit any message after
a
> Close() message rather than this solution.

I don't think you can prevent a message. When you call Close on the resource, it
sends an async message to the renderer, which may have already sent some
messages.

LGTM with a comment explaining how this can happen.

[bbudge, 2015-05-08 17:25:19]

https://codereview.chromium.org/1138483002/diff/40001/ppapi/proxy/video_encod...
File ppapi/proxy/video_encoder_resource.cc (right):

https://codereview.chromium.org/1138483002/diff/40001/ppapi/proxy/video_encod...
ppapi/proxy/video_encoder_resource.cc:364: if (encode_callbacks_.empty())
We should comment on how this can happen.

[llandwerlin-old, 2015-05-08 17:35:25]

The CQ bit was checked by lionel.g.landwerlin@intel.com

[llandwerlin-old, 2015-05-08 17:35:25]

The patchset sent to the CQ was uploaded after l-g-t-m from bbudge@chromium.org
Link to the patchset: https://codereview.chromium.org/1138483002/#ps60001
(title: "Add comment")

[commit-bot: I haz the power, 2015-05-08 17:36:38]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1138483002/60001

[commit-bot: I haz the power, 2015-05-08 18:51:00]

Committed patchset #4 (id:60001)

[commit-bot: I haz the power, 2015-05-08 18:51:55]

Patchset 4 (id:??) landed as
https://crrev.com/86bd0448bb11826977519f852522267e403daeb0
Cr-Commit-Position: refs/heads/master@{#328997}