Remove old perf format conversion functions.

Remove old perf format conversion functions.

In WebRTC r4714 (https://code.google.com/p/webrtc/source/detail?r=4714)
the format of the measured numbers were converted into Chrome perf format.

This means we can get rid of this old parsing code and use the stdout
without modifications.

BUG=none
TEST=
Modify src/DEPS to have webrtc_revision > 4714
gclient sync
ninja -C out/Release
out/Release/browser_tests --run-manual --gtest_filter=WebrtcVideoQualityBrowserTest.* --ui-test-action-max-timeout=1000000
and inspecting the output.

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=223577

[kjellander_chromium, 2013-09-11 15:30:58]

phoglund: main review.
tommi: rubberstamp.

https://codereview.chromium.org/23946005/diff/2001/chrome/browser/media/chrom...
File chrome/browser/media/chrome_webrtc_video_quality_browsertest.cc (right):

https://codereview.chromium.org/23946005/diff/2001/chrome/browser/media/chrom...
chrome/browser/media/chrome_webrtc_video_quality_browsertest.cc:296:
compare_command.AppendArg("--label=VGA");
This is a new argument for the script, preparing for supporting both VGA and HD
(needs to separated in the output).

[tommi (sloooow) - chröme, 2013-09-11 20:01:45]

rs lgtm

[phoglund_chromium, 2013-09-12 07:32:10]

https://codereview.chromium.org/23946005/diff/2001/chrome/browser/media/chrom...
File chrome/browser/media/chrome_webrtc_video_quality_browsertest.cc (right):

https://codereview.chromium.org/23946005/diff/2001/chrome/browser/media/chrom...
chrome/browser/media/chrome_webrtc_video_quality_browsertest.cc:267: // |width|
x |height|.
Add a comment which describes how the output gets printed, since that's not
really readily apparent. Additionally, perhaps we should rename this method
CompareVideosAndPrintResult?

https://codereview.chromium.org/23946005/diff/2001/chrome/browser/media/chrom...
chrome/browser/media/chrome_webrtc_video_quality_browsertest.cc:313: LOG(INFO)
<< "Output was:\n\n" << output;
Ok, so this LOG(INFO) is what really prints the output. I don't like putting a
program's formal output into a log statement, so I would rather have this
printed on stdout if possible. I can also imagine other potential problems, for
instance if we print

LOG(INFO) << "RESULT psnr...\nRESULT ssim\n...";

The printout will be something like

[29185:29185:0912/093019:INFO:...cc] RESULT psnr...
RESULT ssim...
...

I'm not sure if the perf tool will pick up the first line. One can solve this by
making sure the compare scripts prints nonimportant information on the first
line, but that seems hackish.

[kjellander_chromium, 2013-09-12 09:02:35]

PTAL

https://codereview.chromium.org/23946005/diff/2001/chrome/browser/media/chrom...
File chrome/browser/media/chrome_webrtc_video_quality_browsertest.cc (right):

https://codereview.chromium.org/23946005/diff/2001/chrome/browser/media/chrom...
chrome/browser/media/chrome_webrtc_video_quality_browsertest.cc:267: // |width|
x |height|.
On 2013/09/12 07:32:10, phoglund wrote:
> Add a comment which describes how the output gets printed, since that's not
> really readily apparent. Additionally, perhaps we should rename this method
> CompareVideosAndPrintResult?

I agree with both, updated.

https://codereview.chromium.org/23946005/diff/2001/chrome/browser/media/chrom...
chrome/browser/media/chrome_webrtc_video_quality_browsertest.cc:313: LOG(INFO)
<< "Output was:\n\n" << output;
On 2013/09/12 07:32:10, phoglund wrote:
> Ok, so this LOG(INFO) is what really prints the output. I don't like putting a
> program's formal output into a log statement, so I would rather have this
> printed on stdout if possible. I can also imagine other potential problems,
for
> instance if we print
> 
> LOG(INFO) << "RESULT psnr...\nRESULT ssim\n...";
> 
> The printout will be something like
> 
> [29185:29185:0912/093019:INFO:...cc] RESULT psnr...
> RESULT ssim...
> ...
> 
> I'm not sure if the perf tool will pick up the first line. One can solve this
by
> making sure the compare scripts prints nonimportant information on the first
> line, but that seems hackish.

Right, there's a ton of other stuff printed before the RESULT lines are printed,
but I changed it to std:cout just to be sure for the future.

[phoglund_chromium, 2013-09-12 09:17:59]

https://codereview.chromium.org/23946005/diff/14001/chrome/browser/media/chro...
File chrome/browser/media/chrome_webrtc_video_quality_browsertest.cc (right):

https://codereview.chromium.org/23946005/diff/14001/chrome/browser/media/chro...
chrome/browser/media/chrome_webrtc_video_quality_browsertest.cc:315: // Print to
stdout to ensure the perf numbers are parsed properly.
... by the buildbot step.

https://codereview.chromium.org/23946005/diff/14001/chrome/browser/media/chro...
chrome/browser/media/chrome_webrtc_video_quality_browsertest.cc:316: std::cout
<< "Output was:\n\n" << output;
Style guide says: use fprintf or printf here.

[kjellander_chromium, 2013-09-12 11:10:52]

and again...

https://codereview.chromium.org/23946005/diff/14001/chrome/browser/media/chro...
File chrome/browser/media/chrome_webrtc_video_quality_browsertest.cc (right):

https://codereview.chromium.org/23946005/diff/14001/chrome/browser/media/chro...
chrome/browser/media/chrome_webrtc_video_quality_browsertest.cc:315: // Print to
stdout to ensure the perf numbers are parsed properly.
On 2013/09/12 09:17:59, phoglund wrote:
> ... by the buildbot step.

Done.

https://codereview.chromium.org/23946005/diff/14001/chrome/browser/media/chro...
chrome/browser/media/chrome_webrtc_video_quality_browsertest.cc:316: std::cout
<< "Output was:\n\n" << output;
On 2013/09/12 09:17:59, phoglund wrote:
> Style guide says: use fprintf or printf here.

Done.

[phoglund1, 2013-09-12 11:19:11]

lgtm

[commit-bot: I haz the power, 2013-09-17 07:11:58]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/kjellander@chromium.org/23946005/18001

[commit-bot: I haz the power, 2013-09-17 10:03:28]

Change committed as 223577