Extend VideoCaptureDevice::Client::OnError() to have a tracked_objects::Location param.

Extend VideoCaptureDevice::Client::OnError() to have a tracked_objects::Location param.

Every now and then a bug report would include WebRtc
logs with an error string: this is generated in
VideoCaptureDevice::Client::OnError() and has no
origin information. This CL adds that in a way that
clients add the classical FROM_HERE, so that the error
log includes the originating file and line.

Committed: https://crrev.com/71929f7a5802e2991e7a2e90c6d01f8cd363423b
Cr-Commit-Position: refs/heads/master@{#357154}

[mcasas, 2015-10-29 22:12:01]

Description was changed from

==========
Extend VideoCaptureDevice::Client::OnError() to have a Location param.
==========

to

==========
Extend VideoCaptureDevice::Client::OnError() to have a Location param.

Every now and then a bug report would include WebRtc
logs with an error string: this is generated in 
VideoCaptureDevice::Client::OnError() and has no 
origin information. This CL adds that in a way that
clients add the classical FROM_HERE, so that the error
log includes the originating file and line.
==========

[mcasas, 2015-10-29 22:12:26]

Description was changed from

==========
Extend VideoCaptureDevice::Client::OnError() to have a Location param.

Every now and then a bug report would include WebRtc
logs with an error string: this is generated in 
VideoCaptureDevice::Client::OnError() and has no 
origin information. This CL adds that in a way that
clients add the classical FROM_HERE, so that the error
log includes the originating file and line.
==========

to

==========
Extend VideoCaptureDevice::Client::OnError() to have a tracked_objects::Location
param.

Every now and then a bug report would include WebRtc
logs with an error string: this is generated in 
VideoCaptureDevice::Client::OnError() and has no 
origin information. This CL adds that in a way that
clients add the classical FROM_HERE, so that the error
log includes the originating file and line.
==========

[mcasas, 2015-10-29 23:04:03]

Patchset #1 (id:1) has been deleted

[mcasas, 2015-10-29 23:06:07]

mcasas@chromium.org changed reviewers:
+ miu@chromium.org, tommi@chromium.org

[mcasas, 2015-10-29 23:06:08]

tommi@: PTAL

miu@: tiny thing in 
media/capture/content/screen_capture_device_core.{cc,h}

https://chromiumcodereview.appspot.com/1418263006/diff/20001/media/capture/co...
File media/capture/content/screen_capture_device_core.cc (right):

https://chromiumcodereview.appspot.com/1418263006/diff/20001/media/capture/co...
media/capture/content/screen_capture_device_core.cc:107: static const char*
kStateNames[] = {"Idle", "Capturing", "Error"};
|State| had only 3 entries, so I guess l.111 was stale. To prevent 
this from happening again, added the static_assert() below.

[tommi (sloooow) - chröme, 2015-10-30 15:38:14]

Is the error string that we report today, not unique?

[mcasas, 2015-10-30 15:51:06]

On 2015/10/30 15:38:14, tommi wrote:
> Is the error string that we report today, not unique?

Probably yes.

[tommi (sloooow) - chröme, 2015-10-30 15:56:01]

On 2015/10/30 15:51:06, mcasas wrote:
> On 2015/10/30 15:38:14, tommi wrote:
> > Is the error string that we report today, not unique?
> 
> Probably yes.

haha, ok... I don't suppose you know of some examples?  If you can, then this
approach makes sense to me but it would be good to make sure.

[mcasas, 2015-10-30 16:08:30]

On 2015/10/30 15:56:01, tommi wrote:
> On 2015/10/30 15:51:06, mcasas wrote:
> > On 2015/10/30 15:38:14, tommi wrote:
> > > Is the error string that we report today, not unique?
> > 
> > Probably yes.
> 
> haha, ok... I don't suppose you know of some examples?  If you can, then this
> approach makes sense to me but it would be good to make sure.

I meant: "I think reported errors are unique". So, some code
search would reveal the place where it is generated, if that's
what you are after :)

However the advantage of this way is to add the file:line
 to both DLOG() and WebRtc log, which would help triaging
errors etc, e.g. compare:

ERROR: "Could not open capture device."

with

ERROR: video_capture_device_mac.mm:384 "Coud not open capture device"

or perhaps better (to be done in another CL)

ERROR: video_capture_device_mac.mm:AllocateAndStart():384 "Coud not open capture
device"

These type of errors are possibly read by power users or
testers that might not be familiar with all the capture classes,
embedded in a stream of WebRtc logs... (we've all been 
there, haven't we?).

[tommi (sloooow) - chröme, 2015-10-30 16:09:51]

On 2015/10/30 16:08:30, mcasas wrote:
> On 2015/10/30 15:56:01, tommi wrote:
> > On 2015/10/30 15:51:06, mcasas wrote:
> > > On 2015/10/30 15:38:14, tommi wrote:
> > > > Is the error string that we report today, not unique?
> > > 
> > > Probably yes.
> > 
> > haha, ok... I don't suppose you know of some examples?  If you can, then
this
> > approach makes sense to me but it would be good to make sure.
> 
> I meant: "I think reported errors are unique". So, some code
> search would reveal the place where it is generated, if that's
> what you are after :)
> 
> However the advantage of this way is to add the file:line
>  to both DLOG() and WebRtc log, which would help triaging
> errors etc, e.g. compare:
> 
> ERROR: "Could not open capture device."
> 
> with
> 
> ERROR: video_capture_device_mac.mm:384 "Coud not open capture device"
> 
> or perhaps better (to be done in another CL)
> 
> ERROR: video_capture_device_mac.mm:AllocateAndStart():384 "Coud not open
capture
> device"
> 
> These type of errors are possibly read by power users or
> testers that might not be familiar with all the capture classes,
> embedded in a stream of WebRtc logs... (we've all been 
> there, haven't we?).

Yes.  Yeah, I think this is going to be useful.  Lgtm.

[miu, 2015-10-30 17:55:00]

lgtm

[mcasas, 2015-10-30 18:12:27]

The CQ bit was checked by mcasas@chromium.org

[commit-bot: I haz the power, 2015-10-30 18:14:18]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1418263006/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1418263006/20001

[commit-bot: I haz the power, 2015-10-30 18:47:29]

Committed patchset #1 (id:20001)

[commit-bot: I haz the power, 2015-10-30 18:48:14]

Patchset 1 (id:??) landed as
https://crrev.com/71929f7a5802e2991e7a2e90c6d01f8cd363423b
Cr-Commit-Position: refs/heads/master@{#357154}