Waterfall to use run-time flag WebRTC-H264WithOpenH264FFmpeg

Waterfall to use run-time flag WebRTC-H264WithOpenH264FFmpeg.

Updated fieldtrial_testing_config file for chromeos, linux,
mac, win. (Feature not supported on android/ios.)

Note: The run-time flag was defined in:
https://codereview.chromium.org/1718883002/.

BUG=chromium:500605, chromium:468365
Committed: https://crrev.com/533bae44c5d341149e4f14b118c34750a5bd9d1a
Cr-Commit-Position: refs/heads/master@{#378424}

[hbos_chromium, 2016-02-18 13:00:06]

hbos@chromium.org changed reviewers:
+ rkaplow@chromium.org

[hbos_chromium, 2016-02-18 13:00:07]

Please take a look rkaplow.

You said in email that usually we land the client-side diversion (i.e. finch
flag and use of it?), then field trial testing like this CL, and then google3
config.

Just curious, doesn't it seem a bit backwards to do the field trial before
google3 config? Since these field trials are referencing the study and groups
defined in the google3 config.

[hbos_chromium, 2016-02-18 13:01:11]

Description was changed from

==========
Waterfall to use Finch experiment WebRTC-H264WithOpenH264FFmpeg/Enabled.

Updated fieldtrial_testing_config file for chromeos, linux, mac, win.
(Feature not supported on android/ios.)

BUG=chromium:500605, chromium:468365
==========

to

==========
Waterfall to use Finch experiment WebRTC-H264WithOpenH264FFmpeg/Enabled.

Updated fieldtrial_testing_config file for chromeos, linux, mac, win.
(Feature not supported on android/ios.)

Note: The finch flag was defined in:
https://codereview.chromium.org/1673183002/ (not landed as of writing this).

COMMIT=False
BUG=chromium:500605, chromium:468365
==========

[rkaplow, 2016-02-18 16:15:00]

On 2016/02/18 13:00:07, hbos_chromium wrote:
> Please take a look rkaplow.
> 
> You said in email that usually we land the client-side diversion (i.e. finch
> flag and use of it?), then field trial testing like this CL, and then google3
> config.
> 
> Just curious, doesn't it seem a bit backwards to do the field trial before
> google3 config? Since these field trials are referencing the study and groups
> defined in the google3 config.

The fields trials here are referencing the ones defined in the client-side
diversion. This is forcing the test bots to run under these experimental groups
- that can work independent of the server side configs.

Does that make more sense?

[rkaplow, 2016-02-18 16:25:14]

lgtm

lg once client side logic is in

[hbos_chromium, 2016-02-19 12:11:33]

On 2016/02/18 16:15:00, rkaplow wrote:
> On 2016/02/18 13:00:07, hbos_chromium wrote:
> > Please take a look rkaplow.
> > 
> > You said in email that usually we land the client-side diversion (i.e. finch
> > flag and use of it?), then field trial testing like this CL, and then
google3
> > config.
> > 
> > Just curious, doesn't it seem a bit backwards to do the field trial before
> > google3 config? Since these field trials are referencing the study and
groups
> > defined in the google3 config.
> 
> The fields trials here are referencing the ones defined in the client-side
> diversion. This is forcing the test bots to run under these experimental
groups
> - that can work independent of the server side configs.
> 
> Does that make more sense?

Hmm. Am I missing something or using the wrong name for the flag?
The CL that introduces the flag (https://codereview.chromium.org/1673183002/)
uses a base::Feature like so:

const base::Feature kWebRtcH264WithOpenH264FFmpeg {
  "WebRtcH264WithOpenH264FFmpeg", base::FEATURE_DISABLED_BY_DEFAULT
};

Without defining any study or group. And this CL adds to fieldtrial:

"WebRTC-H264WithOpenH264FFmpeg": [
    {
        "group_name": "Enabled"
    }
],

Studies/groups are only defined in the google3 CL (where name
"WebRTC-H264WithOpenH264FFmpeg" is used for study and "Enabled" used for group),
which shouldn't be necessary to run with the feature enabled on the bots. So
should the flag be renamed from "WebRtcH264WithOpenH264FFmpeg" to
"WebRTC-H264WithOpenH264FFmpeg" or is something else required?

[hbos_chromium, 2016-02-19 12:17:00]

Description was changed from

==========
Waterfall to use Finch experiment WebRTC-H264WithOpenH264FFmpeg/Enabled.

Updated fieldtrial_testing_config file for chromeos, linux, mac, win.
(Feature not supported on android/ios.)

Note: The finch flag was defined in:
https://codereview.chromium.org/1673183002/ (not landed as of writing this).

COMMIT=False
BUG=chromium:500605, chromium:468365
==========

to

==========
Waterfall to use run-time flag
WebRTC-H264WithOpenH264FFmpeg/Enabled.

Updated fieldtrial_testing_config file for chromeos, linux,
mac, win. (Feature not supported on android/ios.)

Note: The finch flag was defined in:
https://codereview.chromium.org/1673183002/ (not landed as of writing this).

COMMIT=False
BUG=chromium:500605, chromium:468365
==========

[hbos_chromium, 2016-02-19 17:57:17]

On 2016/02/19 12:11:33, hbos_chromium wrote:
> On 2016/02/18 16:15:00, rkaplow wrote:
> > On 2016/02/18 13:00:07, hbos_chromium wrote:
> > > Please take a look rkaplow.
> > > 
> > > You said in email that usually we land the client-side diversion (i.e.
finch
> > > flag and use of it?), then field trial testing like this CL, and then
> google3
> > > config.
> > > 
> > > Just curious, doesn't it seem a bit backwards to do the field trial before
> > > google3 config? Since these field trials are referencing the study and
> groups
> > > defined in the google3 config.
> > 
> > The fields trials here are referencing the ones defined in the client-side
> > diversion. This is forcing the test bots to run under these experimental
> groups
> > - that can work independent of the server side configs.
> > 
> > Does that make more sense?
> 
> Hmm. Am I missing something or using the wrong name for the flag?
> The CL that introduces the flag (https://codereview.chromium.org/1673183002/)
> uses a base::Feature like so:
> 
> const base::Feature kWebRtcH264WithOpenH264FFmpeg {
>   "WebRtcH264WithOpenH264FFmpeg", base::FEATURE_DISABLED_BY_DEFAULT
> };
> 
> Without defining any study or group. And this CL adds to fieldtrial:
> 
> "WebRTC-H264WithOpenH264FFmpeg": [
>     {
>         "group_name": "Enabled"
>     }
> ],
> 
> Studies/groups are only defined in the google3 CL (where name
> "WebRTC-H264WithOpenH264FFmpeg" is used for study and "Enabled" used for
group),
> which shouldn't be necessary to run with the feature enabled on the bots. So
> should the flag be renamed from "WebRtcH264WithOpenH264FFmpeg" to
> "WebRTC-H264WithOpenH264FFmpeg" or is something else required?

I've seen examples that make sense using base::FieldTrialList::FindFullName
where its just a string->string map between feature name and group name and you
just check what group name you get, that being the value of the flag so to
speak. But because I've been using base::Feature I'm a bit confused where the
group comes in to the picture, since base::Feature is just something that is on
or off.

[rkaplow, 2016-02-19 18:16:24]

rkaplow@chromium.org changed reviewers:
+ asvitkine@google.com

[rkaplow, 2016-02-19 18:17:14]

Looping in alexei since I'm actually not quite sure how the field trial testing
framework works with the feature API.

[Alexei Svitkine (slow), 2016-02-19 18:18:13]

asvitkine@chromium.org changed reviewers:
+ asvitkine@chromium.org

[Alexei Svitkine (slow), 2016-02-19 18:18:13]

You need to specify "enable_features", similar to the server side configs.

[rkaplow, 2016-02-19 18:20:49]

OK, that makes sense. You can see an example here:
https://code.google.com/p/chromium/codesearch#chromium/src/testing/variations...

So this is still fine to (and, should) submit before the server-side config.

I'd also suggest making the group names match but technically it doesn';t matter

[hbos_chromium, 2016-02-22 12:25:45]

Patchset #2 (id:20001) has been deleted

[hbos_chromium, 2016-02-22 12:29:31]

Description was changed from

==========
Waterfall to use run-time flag
WebRTC-H264WithOpenH264FFmpeg/Enabled.

Updated fieldtrial_testing_config file for chromeos, linux,
mac, win. (Feature not supported on android/ios.)

Note: The finch flag was defined in:
https://codereview.chromium.org/1673183002/ (not landed as of writing this).

COMMIT=False
BUG=chromium:500605, chromium:468365
==========

to

==========
Waterfall to use run-time flag WebRTC-H264WithOpenH264FFmpeg.

Updated fieldtrial_testing_config file for chromeos, linux,
mac, win. (Feature not supported on android/ios.)

Note: The run-time flag was defined in:
https://codereview.chromium.org/1718883002/.

BUG=chromium:500605, chromium:468365
==========

[hbos_chromium, 2016-02-23 11:50:24]

I'll call the feature flag the same as this and the study is called
(WebRTC-H264WithOpenH264FFmpeg, with hyphen) for less confusion. And I made the
group names match (even though it doesn't matter).

[hbos_chromium, 2016-02-26 09:17:00]

PTAL rkaplow, asvitkine. Still LGTM and OK to land this now that the feature has
landed behind the flag?

Which bots is it that use these flags btw?

[Alexei Svitkine (slow), 2016-02-26 16:33:27]

lgtm

lgtm

this will affect telemetry perf tests & browser tests

[hbos_chromium, 2016-02-29 15:08:54]

Bump rkaplow. (I assume your LGTM is still valid but bump in case.)
I'll probably land this tomorrow morning (some time UTC+1) so I can keep an eye
on the waterfall during the day.

[hbos_chromium, 2016-03-01 10:47:47]

Alright, no word, I'll trust my judgement that the LGTM still holds and that I
can land this.

[hbos_chromium, 2016-03-01 10:47:54]

The CQ bit was checked by hbos@chromium.org

[hbos_chromium, 2016-03-01 10:47:55]

The patchset sent to the CQ was uploaded after l-g-t-m from rkaplow@chromium.org
Link to the patchset: https://codereview.chromium.org/1696383002/#ps40001
(title: ""enable_features": [ "WebRTC-H264WithOpenH264FFmpeg" ]")

[commit-bot: I haz the power, 2016-03-01 10:47:58]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1696383002/40001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1696383002/40001

[commit-bot: I haz the power, 2016-03-01 11:26:03]

Description was changed from

==========
Waterfall to use run-time flag WebRTC-H264WithOpenH264FFmpeg.

Updated fieldtrial_testing_config file for chromeos, linux,
mac, win. (Feature not supported on android/ios.)

Note: The run-time flag was defined in:
https://codereview.chromium.org/1718883002/.

BUG=chromium:500605, chromium:468365
==========

to

==========
Waterfall to use run-time flag WebRTC-H264WithOpenH264FFmpeg.

Updated fieldtrial_testing_config file for chromeos, linux,
mac, win. (Feature not supported on android/ios.)

Note: The run-time flag was defined in:
https://codereview.chromium.org/1718883002/.

BUG=chromium:500605, chromium:468365
==========

[commit-bot: I haz the power, 2016-03-01 11:26:04]

Committed patchset #2 (id:40001)

[commit-bot: I haz the power, 2016-03-01 11:27:41]

Description was changed from

==========
Waterfall to use run-time flag WebRTC-H264WithOpenH264FFmpeg.

Updated fieldtrial_testing_config file for chromeos, linux,
mac, win. (Feature not supported on android/ios.)

Note: The run-time flag was defined in:
https://codereview.chromium.org/1718883002/.

BUG=chromium:500605, chromium:468365
==========

to

==========
Waterfall to use run-time flag WebRTC-H264WithOpenH264FFmpeg.

Updated fieldtrial_testing_config file for chromeos, linux,
mac, win. (Feature not supported on android/ios.)

Note: The run-time flag was defined in:
https://codereview.chromium.org/1718883002/.

BUG=chromium:500605, chromium:468365

Committed: https://crrev.com/533bae44c5d341149e4f14b118c34750a5bd9d1a
Cr-Commit-Position: refs/heads/master@{#378424}
==========

[commit-bot: I haz the power, 2016-03-01 11:27:42]

Patchset 2 (id:??) landed as
https://crrev.com/533bae44c5d341149e4f14b118c34750a5bd9d1a
Cr-Commit-Position: refs/heads/master@{#378424}