Add performance test for WebRTC protocol.

Add performance test for WebRTC protocol.

Added two new tests that allow to measure latency of the new
WebRTC-based protocol.

BUG=547158
Committed: https://crrev.com/d5e1e73225fe9f160d80790ea1c6bda65def8039
Cr-Commit-Position: refs/heads/master@{#372255}

[Sergey Ulanov, 2016-01-27 22:24:14]

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

[Jamie, 2016-01-28 01:32:22]

https://codereview.chromium.org/1642793002/diff/1/remoting/test/cyclic_frame_...
File remoting/test/cyclic_frame_generator.cc (right):

https://codereview.chromium.org/1642793002/diff/1/remoting/test/cyclic_frame_...
remoting/test/cyclic_frame_generator.cc:142: CHECK(value < (2U <<
kBarcodeBits));
s/2/1/?

Eg, with 2 bits, you can express anything less than 4, which is 1 << 2.

https://codereview.chromium.org/1642793002/diff/1/remoting/test/cyclic_frame_...
remoting/test/cyclic_frame_generator.cc:151: if (last_frame_type_ !=
FrameType::EMPTY) {
Even if the frame had no changes before adding the barcode, it does now.
Shouldn't this be testing against FULL instead (no need to update the changed
region if it was already the full screen)?

Alternatively, since it looks like you're not saving the frame info for EMPTY
frames, do you even want to generate the barcode?

[Sergey Ulanov, 2016-01-28 02:05:49]

https://codereview.chromium.org/1642793002/diff/1/remoting/test/cyclic_frame_...
File remoting/test/cyclic_frame_generator.cc (right):

https://codereview.chromium.org/1642793002/diff/1/remoting/test/cyclic_frame_...
remoting/test/cyclic_frame_generator.cc:142: CHECK(value < (2U <<
kBarcodeBits));
On 2016/01/28 01:32:22, Jamie wrote:
> s/2/1/?
> 
> Eg, with 2 bits, you can express anything less than 4, which is 1 << 2.

Done.

https://codereview.chromium.org/1642793002/diff/1/remoting/test/cyclic_frame_...
remoting/test/cyclic_frame_generator.cc:151: if (last_frame_type_ !=
FrameType::EMPTY) {
On 2016/01/28 01:32:22, Jamie wrote:
> Even if the frame had no changes before adding the barcode, it does now.
Barcode changes only for non-empty frames, so EMPTY frames still have empty
updated_region.

> Shouldn't this be testing against FULL instead (no need to update the changed
> region if it was already the full screen)?

It's not necessary, but it doesn't hurt either. DesktopRegion will test the new
rect already overlaps with the rects that already there. AddRect() call for FULL
frames doesn't change anything, but it's still better to call it to make this
code clearer. Changed this condition to frame_id != last_frame_id_.


> 
> Alternatively, since it looks like you're not saving the frame info for EMPTY
> frames, do you even want to generate the barcode?

It still need to be drawn to ensure that the content of the frame is the same
(e.g. webrtc currently ignores updated_region()).

[Sergey Ulanov, 2016-01-28 02:06:30]

Description was changed from

==========
Add performance test for WebRTC protocol.

Added two new tests that allow to measure latency with the new
WebRTC-based protocol.

BUG=547158
==========

to

==========
Add performance test for WebRTC protocol.

Added two new tests that allow to measure latency of the new
WebRTC-based protocol.

BUG=547158
==========

[Jamie, 2016-01-28 18:10:13]

lgtm

[Sergey Ulanov, 2016-01-28 18:28:10]

The CQ bit was checked by sergeyu@chromium.org

[Sergey Ulanov, 2016-01-28 18:28:18]

The CQ bit was unchecked by sergeyu@chromium.org

[Sergey Ulanov, 2016-01-29 02:00:44]

The CQ bit was checked by sergeyu@chromium.org

[commit-bot: I haz the power, 2016-01-29 02:02:51]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1642793002/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1642793002/20001

[commit-bot: I haz the power, 2016-01-29 02:23:57]

Description was changed from

==========
Add performance test for WebRTC protocol.

Added two new tests that allow to measure latency of the new
WebRTC-based protocol.

BUG=547158
==========

to

==========
Add performance test for WebRTC protocol.

Added two new tests that allow to measure latency of the new
WebRTC-based protocol.

BUG=547158
==========

[commit-bot: I haz the power, 2016-01-29 02:23:58]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2016-01-29 02:24:54]

Description was changed from

==========
Add performance test for WebRTC protocol.

Added two new tests that allow to measure latency of the new
WebRTC-based protocol.

BUG=547158
==========

to

==========
Add performance test for WebRTC protocol.

Added two new tests that allow to measure latency of the new
WebRTC-based protocol.

BUG=547158

Committed: https://crrev.com/d5e1e73225fe9f160d80790ea1c6bda65def8039
Cr-Commit-Position: refs/heads/master@{#372255}
==========

[commit-bot: I haz the power, 2016-01-29 02:24:55]

Patchset 2 (id:??) landed as
https://crrev.com/d5e1e73225fe9f160d80790ea1c6bda65def8039
Cr-Commit-Position: refs/heads/master@{#372255}