Move screen capturers from remoting/capturer to media/video/capturer/screen

Move screen capturers from remoting/capturer to media/video/capturer/screen

Screen capturers will be used in content, and so they have to be moved to
avoid dependency on remoting/capturer in content.

Beside moving the files this CL also renames classes as follows:
  remoting::VideoFrameCapturer -> media::ScreenCapturer
  remoting::VideoFrame -> media::ScreenCaptureFrame
  remoting::CaptureData -> media::ScreenCaptureData

BUG=134249
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=179218

[Sergey Ulanov, 2013-01-25 00:05:48]

Practically all changes are mechanical renames, except SharedBuffer class where
I added new method to avoid dependency on src/ipc.

tsepez@ - Please approve changes in remoting/host/chromoting_messages.h
alexeypa@ - Please take a look at my changes in SharedBuffer.
wez@, scherkus@ - please review the rest.

[scherkus (not reviewing), 2013-01-25 00:30:36]

holy smokes that's a lot of files!

two μ-nits

https://codereview.chromium.org/12047101/diff/2001/media/video/capture/screen...
File media/video/capture/screen/differ_block.h (right):

https://codereview.chromium.org/12047101/diff/2001/media/video/capture/screen...
media/video/capture/screen/differ_block.h:15: static const int kBlockSize = 32;
static ints declared in the .h?

doesn't this end up getting declared in every compilation unit?

https://codereview.chromium.org/12047101/diff/2001/media/video/capture/screen...
File media/video/capture/screen/screen_capturer_mock_objects.cc (right):

https://codereview.chromium.org/12047101/diff/2001/media/video/capture/screen...
media/video/capture/screen/screen_capturer_mock_objects.cc:16:
MockScreenCapturerDelegate::MockScreenCapturerDelegate() {
nit: you're inconsistent on {} vs {\n}

FWIW I prefer the former and the clang-based C++ formatter agrees with me ;)

[Sergey Ulanov, 2013-01-25 00:49:50]

https://codereview.chromium.org/12047101/diff/2001/media/video/capture/screen...
File media/video/capture/screen/differ_block.h (right):

https://codereview.chromium.org/12047101/diff/2001/media/video/capture/screen...
media/video/capture/screen/differ_block.h:15: static const int kBlockSize = 32;
On 2013/01/25 00:30:36, scherkus wrote:
> static ints declared in the .h?
> 
> doesn't this end up getting declared in every compilation unit?

Fixed.

https://codereview.chromium.org/12047101/diff/2001/media/video/capture/screen...
File media/video/capture/screen/screen_capturer_mock_objects.cc (right):

https://codereview.chromium.org/12047101/diff/2001/media/video/capture/screen...
media/video/capture/screen/screen_capturer_mock_objects.cc:16:
MockScreenCapturerDelegate::MockScreenCapturerDelegate() {
On 2013/01/25 00:30:36, scherkus wrote:
> nit: you're inconsistent on {} vs {\n}

Fixed.

> 
> FWIW I prefer the former and

I agree, but several times I was asked to use {\n} when getting my changes
reviewed in various parts of chrome. I think it's more consistent with the rest
of chromium code. Also {} becomes less readable for constructors with member
initializers.

> the clang-based C++ formatter agrees with me ;)

How can I try it?

[Wez, 2013-01-25 00:51:20]

remoting/ LGTM w/ nits, but please let alexeypa@ review the IPC-related stuff.

The naming of some of the components (e.g. Differ, SharedBuffer) in content/ is
a little too general, IMHO, so let's be sure to rename them to something more
appropriate in a follow-up CL.

https://codereview.chromium.org/12047101/diff/2001/remoting/codec/video_encod...
File remoting/codec/video_encoder.h (right):

https://codereview.chromium.org/12047101/diff/2001/remoting/codec/video_encod...
remoting/codec/video_encoder.h:13: class DataBuffer;
nit: I don't think we need this here?

https://codereview.chromium.org/12047101/diff/2001/remoting/host/DEPS
File remoting/host/DEPS (right):

https://codereview.chromium.org/12047101/diff/2001/remoting/host/DEPS#newcode13
remoting/host/DEPS:13: "+media/video/capture/screen",
This belongs before the net/ dependencies.

https://codereview.chromium.org/12047101/diff/2001/remoting/host/chromoting_h...
File remoting/host/chromoting_host.h (right):

https://codereview.chromium.org/12047101/diff/2001/remoting/host/chromoting_h...
remoting/host/chromoting_host.h:49: //    A ScreenRecorder is created with an
Encoder and a media::ScreenCapturer.
nit: ScreenRecorder -> VideoScheduler here and below.

[Sergey Ulanov, 2013-01-25 02:33:45]

Opened crbug.com/172183 to rename Differ and SharedBuffer

https://codereview.chromium.org/12047101/diff/2001/remoting/codec/video_encod...
File remoting/codec/video_encoder.h (right):

https://codereview.chromium.org/12047101/diff/2001/remoting/codec/video_encod...
remoting/codec/video_encoder.h:13: class DataBuffer;
On 2013/01/25 00:51:21, Wez wrote:
> nit: I don't think we need this here?

Done.

https://codereview.chromium.org/12047101/diff/2001/remoting/host/DEPS
File remoting/host/DEPS (right):

https://codereview.chromium.org/12047101/diff/2001/remoting/host/DEPS#newcode13
remoting/host/DEPS:13: "+media/video/capture/screen",
On 2013/01/25 00:51:21, Wez wrote:
> This belongs before the net/ dependencies.

Done.

https://codereview.chromium.org/12047101/diff/2001/remoting/host/chromoting_h...
File remoting/host/chromoting_host.h (right):

https://codereview.chromium.org/12047101/diff/2001/remoting/host/chromoting_h...
remoting/host/chromoting_host.h:49: //    A ScreenRecorder is created with an
Encoder and a media::ScreenCapturer.
On 2013/01/25 00:51:21, Wez wrote:
> nit: ScreenRecorder -> VideoScheduler here and below.

Done.

[scherkus (not reviewing), 2013-01-25 05:03:40]

LGTM w/ nits

https://codereview.chromium.org/12047101/diff/2001/media/video/capture/screen...
File media/video/capture/screen/screen_capturer_mock_objects.cc (right):

https://codereview.chromium.org/12047101/diff/2001/media/video/capture/screen...
media/video/capture/screen/screen_capturer_mock_objects.cc:16:
MockScreenCapturerDelegate::MockScreenCapturerDelegate() {
On 2013/01/25 00:49:50, sergeyu wrote:
> On 2013/01/25 00:30:36, scherkus wrote:
> > nit: you're inconsistent on {} vs {\n}
> 
> Fixed.
> 
> > 
> > FWIW I prefer the former and
> 
> I agree, but several times I was asked to use {\n} when getting my changes
> reviewed in various parts of chrome. I think it's more consistent with the
rest
> of chromium code. Also {} becomes less readable for constructors with member
> initializers.

Yeah media/ code attempts to use {} for single line functions but {\n} if it
it's more than one line for reasons you describe (i.e., readability w/
initializers)... but same-file consistency is good enough for me

> > the clang-based C++ formatter agrees with me ;)
> 
> How can I try it?

Sent you an email.

https://codereview.chromium.org/12047101/diff/9001/media/video/capture/screen...
File media/video/capture/screen/differ_block.h (right):

https://codereview.chromium.org/12047101/diff/9001/media/video/capture/screen...
media/video/capture/screen/differ_block.h:5: #ifndef
REMOTING_HOST_DIFFER_BLOCK_H_
wrong header guard

https://codereview.chromium.org/12047101/diff/9001/media/video/capture/screen...
File media/video/capture/screen/differ_block_sse2.h (right):

https://codereview.chromium.org/12047101/diff/9001/media/video/capture/screen...
media/video/capture/screen/differ_block_sse2.h:8: #ifndef
REMOTING_HOST_DIFFER_BLOCK_SSE2_H_
wrong header guard

[Tom Sepez, 2013-01-25 18:07:38]

Rubberstamp LGTM on changing names in messages.

[alexeypa (please no reviews), 2013-01-28 18:24:41]

SharedBuffer changes - LGTM