Use NTP time for video frame timestamp. Add tracing for timestamps.

Use NTP time for video frame timestamp.
Add tracing for capture timestamps.

BUG=225209
TEST=Check that AppRTC still works.

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=192670

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=194268

[holmer, 2013-04-05 07:33:44]

lgtm

[no longer working on chromium, 2013-04-05 08:56:34]

lgtm %nit

https://codereview.chromium.org/13554003/diff/5001/content/renderer/media/rtc...
File content/renderer/media/rtc_video_capturer.cc (right):

https://codereview.chromium.org/13554003/diff/5001/content/renderer/media/rtc...
content/renderer/media/rtc_video_capturer.cc:19: base::Time::Exploded exploded;
nit, initialize the value to {}

[justinlin, 2013-04-05 19:03:42]

thanks

https://codereview.chromium.org/13554003/diff/5001/content/renderer/media/rtc...
File content/renderer/media/rtc_video_capturer.cc (right):

https://codereview.chromium.org/13554003/diff/5001/content/renderer/media/rtc...
content/renderer/media/rtc_video_capturer.cc:19: base::Time::Exploded exploded;
On 2013/04/05 08:56:34, xians1 wrote:
> nit, initialize the value to {}

Done.

[commit-bot: I haz the power, 2013-04-05 19:19:14]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/justinlin@chromium.org/13554003/11001

[commit-bot: I haz the power, 2013-04-06 00:33:31]

Change committed as 192670

[justinlin, 2013-04-15 06:33:51]

Fixed the issue that caused the test breakage: the
webrtc_call.WebrtcCallTest.testLocalPreview test bypasses the normal path
through libjingle (which discards the frame's timestamp). The test somehow sets
the frame's timestamp to the one we assign in DeliverFrame.

There was a problem with the originally submitted CL because RenderFrame was
still expecting the timestamp in nanoseconds. ptal.

[holmer, 2013-04-15 08:14:18]

On 2013/04/15 06:33:51, justinlin wrote:
> Fixed the issue that caused the test breakage: the
> webrtc_call.WebrtcCallTest.testLocalPreview test bypasses the normal path
> through libjingle (which discards the frame's timestamp). The test somehow
sets
> the frame's timestamp to the one we assign in DeliverFrame.
> 
> There was a problem with the originally submitted CL because RenderFrame was
> still expecting the timestamp in nanoseconds. ptal.

lgtm

[justinlin, 2013-04-16 01:12:23]

Committed patchset #3 manually as r194268 (presubmit successful).