Fix the flaky MediaSourcePlayerTest

Fix the flaky MediaSourcePlayerTest

If seek happens while requesting data for the inactive chunk, the ack for the data request will call OnDecodeCompleted().
However, the decoder job is waiting for the OnDecodeCompleted() from the active chunk.
Tests:We already have seveal tests covering requesting inactive chunk during seek. This CL should fix the flakiness in there.

BUG=361541
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=262934

[qinmin, 2014-04-09 23:59:01]

PTAL

[wolenetz, 2014-04-10 00:15:35]

https://codereview.chromium.org/227523005/diff/1/media/base/android/media_dec...
File media/base/android/media_decoder_job.cc (right):

https://codereview.chromium.org/227523005/diff/1/media/base/android/media_dec...
media/base/android/media_decoder_job.cc:76: // If this data request is for the
inactive chunk, do nothing.
I'm confused. If it's a request for the inactive chunk, on_data_received_cb_
should be just a base::Closure. However, if this data ack was received after
StopDecode() or ClearData(), then done_cb would be null here. Is it just the new
comment that is incorrect here?

[qinmin, 2014-04-10 00:19:36]

https://codereview.chromium.org/227523005/diff/1/media/base/android/media_dec...
File media/base/android/media_decoder_job.cc (right):

https://codereview.chromium.org/227523005/diff/1/media/base/android/media_dec...
media/base/android/media_decoder_job.cc:76: // If this data request is for the
inactive chunk, do nothing.
So if done_cb.is_null() is true, this is the data request for the inactive
chunk. And in this case, we should do nothing, just early return.

On 2014/04/10 00:15:35, wolenetz wrote:
> I'm confused. If it's a request for the inactive chunk, on_data_received_cb_
> should be just a base::Closure. However, if this data ack was received after
> StopDecode() or ClearData(), then done_cb would be null here. Is it just the
new
> comment that is incorrect here?

[wolenetz, 2014-04-10 00:57:45]

lgtm. Also, thanks for clarifying base::Closure().is_null() is true :)

[wolenetz, 2014-04-10 00:58:20]

On 2014/04/10 00:57:45, wolenetz wrote:
> lgtm. Also, thanks for clarifying base::Closure().is_null() is true :)

With this patch, I no longer get repro locally (~4 attempts).

[qinmin, 2014-04-10 00:59:11]

The CQ bit was checked by qinmin@chromium.org

[commit-bot: I haz the power, 2014-04-10 01:00:11]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/qinmin@chromium.org/227523005/20001

[commit-bot: I haz the power, 2014-04-10 05:32:15]

Change committed as 262934