Replacing SSLIdentity* with scoped_refptr<RTCCertificate> in TransportChannel layer

Replacing SSLIdentity* with scoped_refptr<RTCCertificate> in TransportChannel layer.

BUG=webrtc:4927
R=tommi@webrtc.org, torbjorng@webrtc.org

Committed: https://chromium.googlesource.com/external/webrtc/+/f3ecdb981c172cdfafbe92c939eb25ddcc1ae96e

[hbos, 2015-09-03 13:56:45]

hbos@webrtc.org changed reviewers:
+ tommi@webrtc.org, torbjorng@webrtc.org

[hbos, 2015-09-03 13:56:45]

PTAL torbjorng, tommi (when you have time, not high priority, no functional
change)

Should be the last of the "replace SSLIdentity with RTCCertificate" CLs.

[tommi, 2015-09-03 20:06:04]

lgtm with the below addressed

https://codereview.webrtc.org/1304043008/diff/1/webrtc/p2p/base/dtlstransport...
File webrtc/p2p/base/dtlstransportchannel.cc (right):

https://codereview.webrtc.org/1304043008/diff/1/webrtc/p2p/base/dtlstransport...
webrtc/p2p/base/dtlstransportchannel.cc:98: local_certificate_(nullptr),
this is a scoped_refptr now, so no need for including in initializer list

https://codereview.webrtc.org/1304043008/diff/1/webrtc/p2p/base/fakesession.h
File webrtc/p2p/base/fakesession.h (right):

https://codereview.webrtc.org/1304043008/diff/1/webrtc/p2p/base/fakesession.h...
webrtc/p2p/base/fakesession.h:53: certificate_(nullptr),
no need to include in initializer list

https://codereview.webrtc.org/1304043008/diff/1/webrtc/p2p/base/fakesession.h...
webrtc/p2p/base/fakesession.h:253: rtc::scoped_refptr<rtc::RTCCertificate>*
certificate) const override {
I wonder if we should just change this method to return const scoped_refptr<>& ?

https://codereview.webrtc.org/1304043008/diff/1/webrtc/p2p/base/transportchan...
File webrtc/p2p/base/transportchannelimpl.h (right):

https://codereview.webrtc.org/1304043008/diff/1/webrtc/p2p/base/transportchan...
webrtc/p2p/base/transportchannelimpl.h:88: // TODO(bemasc): Fix the ownership
semantics of this method.
maybe you can remove this todo now?

[hbos, 2015-09-07 08:11:01]

PTAL torbjorng

https://codereview.webrtc.org/1304043008/diff/1/webrtc/p2p/base/dtlstransport...
File webrtc/p2p/base/dtlstransportchannel.cc (right):

https://codereview.webrtc.org/1304043008/diff/1/webrtc/p2p/base/dtlstransport...
webrtc/p2p/base/dtlstransportchannel.cc:98: local_certificate_(nullptr),
On 2015/09/03 20:06:04, tommi (webrtc) wrote:
> this is a scoped_refptr now, so no need for including in initializer list

Done.

https://codereview.webrtc.org/1304043008/diff/1/webrtc/p2p/base/fakesession.h
File webrtc/p2p/base/fakesession.h (right):

https://codereview.webrtc.org/1304043008/diff/1/webrtc/p2p/base/fakesession.h...
webrtc/p2p/base/fakesession.h:53: certificate_(nullptr),
On 2015/09/03 20:06:04, tommi (webrtc) wrote:
> no need to include in initializer list

Done.

https://codereview.webrtc.org/1304043008/diff/1/webrtc/p2p/base/fakesession.h...
webrtc/p2p/base/fakesession.h:253: rtc::scoped_refptr<rtc::RTCCertificate>*
certificate) const override {
On 2015/09/03 20:06:04, tommi (webrtc) wrote:
> I wonder if we should just change this method to return const scoped_refptr<>&
?

Done, except I made it (non-const non-ref) scoped_refptr because some
implementations want to return nullptr and I don't want to have a member
variable dedicated for the value null.

https://codereview.webrtc.org/1304043008/diff/1/webrtc/p2p/base/transportchan...
File webrtc/p2p/base/transportchannelimpl.h (right):

https://codereview.webrtc.org/1304043008/diff/1/webrtc/p2p/base/transportchan...
webrtc/p2p/base/transportchannelimpl.h:88: // TODO(bemasc): Fix the ownership
semantics of this method.
On 2015/09/03 20:06:04, tommi (webrtc) wrote:
> maybe you can remove this todo now?

Done.

[hbos, 2015-09-07 11:50:29]

Last time I landed a CL making use of the override keyword in places I had
touched, but not updating usage of other virtual methods in the same file, the
webrtc trybots were all green but the chrome bots turned red. And I had to add a
CL that made other stuff in the same file use the override keyword.

Trying to prevent this I added PS4 which does virtual -> override updates of
unrelated nearby code. Hope this is OK in the same CL but not sure it is
landable otherwise even if it's all green...

[torbjorng (webrtc), 2015-09-07 12:48:42]

lgtm (but a trybot is red...)

[hbos, 2015-09-08 10:12:14]

Committed patchset #4 (id:60001) manually as
f3ecdb981c172cdfafbe92c939eb25ddcc1ae96e (presubmit successful).