Remove CHECK(!g_webrtc_logging_delegate) for --single-process model

Remove CHECK(!g_webrtc_logging_delegate) for --single-process model

--single-process mode may happen crash in file of webrtc_logging.cc where code
is CHECK(!g_webrtc_logging_delegate). The orginal code is to expect there to
be only one g_webrtc_logging_delegate instance in a single render process.
g_webrtc_logging_delegate is created when a RenderThreadImpl is created.
But in --single-process model all RenderThreadImpl are created in browser
process and so g_webrtc_logging_delegate may be created many times in a single
process.

So we may not do CHECK(!g_webrtc_logging_delegate) or it will fail under
--single-process model.

BUG=455573

[jinlong.zhai, 2015-02-26 08:38:24]

jinlong.zhai@samsung.com changed reviewers:
+ tommi@chromium.org

[jinlong.zhai, 2015-02-26 08:38:25]

hi, tommi:
  Since we can not change from process-level to RenderThread-level (patchset #1
id:1 of https://codereview.chromium.org/931693002),
and also CHECK(!g_webrtc_logging_delegate) will actually fail under
--single-process model, I suggest to remove CHECK(!g_webrtc_logging_delegate).
WDYT?
thanks,
Jinlong Zhai

[jinlong.zhai, 2015-02-28 02:06:03]

jinlong.zhai@samsung.com changed reviewers:
+ grunell@google.com

[tommi (sloooow) - chröme, 2015-02-28 17:25:52]

On 2015/02/28 at 02:06:03, jinlong.zhai wrote:
> 

Sorry, I don't like this. We can't remove a valid and useful check from
production code because of an inconvenience with single-process mode.  I'd
rather have a proper fix for single process mode and don't change how the
browser actually runs.

[jinlong.zhai, 2015-03-04 07:43:34]

jinlong.zhai@samsung.com changed reviewers:
- grunell@google.com

[jinlong.zhai, 2015-03-04 07:43:35]

hi,tommi

 I have an another proper fix for single process mode,
 please to have another look at it.

thanks,
Jinlong

[tommi (sloooow) - chröme, 2015-03-04 13:23:20]

https://codereview.chromium.org/956263003/diff/20001/content/renderer/media/w...
File content/renderer/media/webrtc_logging.cc (right):

https://codereview.chromium.org/956263003/diff/20001/content/renderer/media/w...
content/renderer/media/webrtc_logging.cc:25: !g_webrtc_logging_delegate);
In the case where g_webrtc_logging_delegate is not null, what value will this be
set to?  Will it overwrite the previous value or will it be assigned the same
value it is already set to?

[jinlong.zhai, 2015-03-05 01:50:17]

https://codereview.chromium.org/956263003/diff/20001/content/renderer/media/w...
File content/renderer/media/webrtc_logging.cc (right):

https://codereview.chromium.org/956263003/diff/20001/content/renderer/media/w...
content/renderer/media/webrtc_logging.cc:25: !g_webrtc_logging_delegate);
On 2015/03/04 13:23:20, tommi wrote:
> In the case where g_webrtc_logging_delegate is not null, what value will this
be
> set to?  Will it overwrite the previous value or will it be assigned the same
> value it is already set to?

In this case, it will overwrite the provious value. And I think this feature is
just to notify logging message, it is OK to do this no matter by new value or
previous value in single process.

I also consider another way to re-landing this: create g_webrtc_logging_delegate
as a tls on IO thread, and also we must make sure every place where
WebRtcLogMessage be called should do on IO thread or it will not work by add
CHECK(g_webrtc_logging_delegate_tls->poniter()->Get()) << "WebRtcLogMessage
should be called on IO thread".

WDYT?