Stop using rtc_make_scoped_ptr

Stop using rtc_make_scoped_ptr

It'll go away very soon, when
https://codereview.webrtc.org/1797463002/ lands in WebRTC.

BUG=webrtc:5520
Committed: https://crrev.com/14ff8c364ead6ba03cc52d87e5d20f6f920ca5b7
Cr-Commit-Position: refs/heads/master@{#389646}

[kwiberg-chromium, 2016-04-25 10:19:06]

kwiberg@chromium.org changed reviewers:
+ sergeyu@chromium.org

[tommi, 2016-04-25 10:24:22]

tommi@webrtc.org changed reviewers:
+ tommi@webrtc.org

[tommi, 2016-04-25 10:24:22]

one teeny request

https://codereview.webrtc.org/1921553002/diff/1/remoting/host/desktop_session...
File remoting/host/desktop_session_agent.cc (right):

https://codereview.webrtc.org/1921553002/diff/1/remoting/host/desktop_session...
remoting/host/desktop_session_agent.cc:145: return
rtc::scoped_ptr<webrtc::SharedMemory>(buffer.release());
looks like CreateSharedMemory is only being called from two places and in both
of those places, there's a workaround to translate the return value from
scoped_ptr to std::unique_ptr.  Can we just change the return type and return
buffer here?

[kwiberg-chromium, 2016-04-25 10:41:59]

https://codereview.webrtc.org/1921553002/diff/1/remoting/host/desktop_session...
File remoting/host/desktop_session_agent.cc (right):

https://codereview.webrtc.org/1921553002/diff/1/remoting/host/desktop_session...
remoting/host/desktop_session_agent.cc:145: return
rtc::scoped_ptr<webrtc::SharedMemory>(buffer.release());
On 2016/04/25 10:24:22, tommi-webrtc wrote:
> looks like CreateSharedMemory is only being called from two places and in both
> of those places, there's a workaround to translate the return value from
> scoped_ptr to std::unique_ptr.  Can we just change the return type and return
> buffer here?

We could, but changing the return value of an interface function that's used by
WebRTC clients is a multi-step process if we want to give the clients the option
to avoid going through broken states (which we do).

But if we just do what I suggest in this patch set, the problem will go away
naturally, since the WebRTC CL I'm planning to land this week will turn
rtc::scoped pointer into a type alias for std::unique_ptr. After that, what you
asked for (webrtc::SharedMemoryFactory::CreateSharedMemory returning a
std::unique_ptr) will have effectively happened, and call sites such as this one
will be free to just return their unique_ptrs directly. (A simplification that's
actually very likely to happen, given that two weeks later, I'll remove the
rtc::scoped_ptr name entirely, so I'll need to fix all lines that mention it.)

[tommi, 2016-04-25 12:30:27]

lgtm

https://codereview.webrtc.org/1921553002/diff/1/remoting/host/desktop_session...
File remoting/host/desktop_session_agent.cc (right):

https://codereview.webrtc.org/1921553002/diff/1/remoting/host/desktop_session...
remoting/host/desktop_session_agent.cc:145: return
rtc::scoped_ptr<webrtc::SharedMemory>(buffer.release());
On 2016/04/25 10:41:59, kwiberg-chromium wrote:
> On 2016/04/25 10:24:22, tommi-webrtc wrote:
> > looks like CreateSharedMemory is only being called from two places and in
both
> > of those places, there's a workaround to translate the return value from
> > scoped_ptr to std::unique_ptr.  Can we just change the return type and
return
> > buffer here?
> 
> We could, but changing the return value of an interface function that's used
by
> WebRTC clients is a multi-step process if we want to give the clients the
option
> to avoid going through broken states (which we do).
> 
> But if we just do what I suggest in this patch set, the problem will go away
> naturally, since the WebRTC CL I'm planning to land this week will turn
> rtc::scoped pointer into a type alias for std::unique_ptr. After that, what
you
> asked for (webrtc::SharedMemoryFactory::CreateSharedMemory returning a
> std::unique_ptr) will have effectively happened, and call sites such as this
one
> will be free to just return their unique_ptrs directly. (A simplification
that's
> actually very likely to happen, given that two weeks later, I'll remove the
> rtc::scoped_ptr name entirely, so I'll need to fix all lines that mention it.)

ok, sgtm

[kwiberg-chromium, 2016-04-26 00:05:53]

kwiberg@chromium.org changed reviewers:
+ lambroslambrou@chromium.org

[kwiberg-chromium, 2016-04-26 00:05:54]

Trying a new reviewer since sergeyu@ may be unavailable.

[Lambros, 2016-04-26 00:41:01]

lgtm

[kwiberg-chromium, 2016-04-26 00:53:29]

The CQ bit was checked by kwiberg@chromium.org

[commit-bot: I haz the power, 2016-04-26 00:54:13]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1921553002/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1921553002/1

[commit-bot: I haz the power, 2016-04-26 00:59:42]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2016-04-26 01:00:54]

Description was changed from

==========
Stop using rtc_make_scoped_ptr

It'll go away very soon, when
https://codereview.webrtc.org/1797463002/ lands in WebRTC.

BUG=webrtc:5520
==========

to

==========
Stop using rtc_make_scoped_ptr

It'll go away very soon, when
https://codereview.webrtc.org/1797463002/ lands in WebRTC.

BUG=webrtc:5520

Committed: https://crrev.com/14ff8c364ead6ba03cc52d87e5d20f6f920ca5b7
Cr-Commit-Position: refs/heads/master@{#389646}
==========

[commit-bot: I haz the power, 2016-04-26 01:00:55]

Patchset 1 (id:??) landed as
https://crrev.com/14ff8c364ead6ba03cc52d87e5d20f6f920ca5b7
Cr-Commit-Position: refs/heads/master@{#389646}