Support experiment control of profile-management flags.

Support experiment control of profile-management flags.

Three flags can now be controlled via an "experiment".
  EnableWebBasedSignin
  GoogleProfileInfo
  NewProfileManagement

Also, correct name of function to match command-line flag change:
  EnableInlineSignin => !EnableWebBasedSignin

BUG=324046
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=243668

[bcwhite, 2013-12-23 18:41:09]

Merry Christmas!

[Nico, 2013-12-23 18:45:00]

(replacing me with guohui since I'm probably here for the ` EnableInlineSignin
=> !EnableWebBasedSignin` change which is in guohui's code)

[bcwhite, 2013-12-23 19:41:28]

removed unused global

[guohui, 2013-12-23 20:45:14]

lgtm

[bcwhite, 2014-01-06 16:22:50]

give bcwhite OWNER permission for new switches file

[bcwhite, 2014-01-06 16:24:19]

+thakis  (OWNER for chrome/common)

Niko, I need your approval for the files under chrome/common.  I've added a
per-file entry to the OWNERS of that directory so I can approve changes to this
file in the future.

[Roger Tawa OOO till Jul 10th, 2014-01-07 17:01:54]

lgtm, one nit below.

https://codereview.chromium.org/111033005/diff/270001/chrome/common/profile_m...
File chrome/common/profile_management_switches.cc (right):

https://codereview.chromium.org/111033005/diff/270001/chrome/common/profile_m...
chrome/common/profile_management_switches.cc:22: return false ^ negate;
Nit: since:
  true^negate == !negate
  false^negate == negate

I would suggest using the RHS expressions since they are simpler.

[Nico, 2014-01-07 19:58:14]

lgtm

[bcwhite, 2014-01-07 21:21:02]

use ! instead of ^ to return desired state

[bcwhite, 2014-01-07 21:22:45]

On 2014/01/07 21:21:02, bcwhite wrote:
> use ! instead of ^ to return desired state

Done.

[bcwhite, 2014-01-07 21:23:20]

https://codereview.chromium.org/111033005/diff/270001/chrome/common/profile_m...
File chrome/common/profile_management_switches.cc (right):

https://codereview.chromium.org/111033005/diff/270001/chrome/common/profile_m...
chrome/common/profile_management_switches.cc:22: return false ^ negate;
On 2014/01/07 17:01:54, Roger Tawa wrote:
> Nit: since:
>   true^negate == !negate
>   false^negate == negate
> 
> I would suggest using the RHS expressions since they are simpler.

Done.

[commit-bot: I haz the power, 2014-01-08 13:44:49]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/bcwhite@chromium.org/111033005/350001

[commit-bot: I haz the power, 2014-01-08 16:51:13]

Retried try job too often on win_rel for step(s) browser_tests
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=win_rel&nu...

[commit-bot: I haz the power, 2014-01-08 19:14:16]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/bcwhite@chromium.org/111033005/350001

[commit-bot: I haz the power, 2014-01-08 22:12:09]

Change committed as 243668

[acleung1, 2014-01-14 22:20:25]

https://codereview.chromium.org/111033005/diff/350001/chrome/common/profile_m...
File chrome/common/profile_management_switches.cc (right):

https://codereview.chromium.org/111033005/diff/350001/chrome/common/profile_m...
chrome/common/profile_management_switches.cc:24: return
CommandLine::ForCurrentProcess()->HasSwitch(command_switch);
Hi Brian.

Shouldn't this take priority? For people enabling it for testing purpose?

[bcwhite, 2014-01-15 20:28:57]

On 2014/01/14 22:20:25, acleung1 wrote:
>
https://codereview.chromium.org/111033005/diff/350001/chrome/common/profile_m...
> File chrome/common/profile_management_switches.cc (right):
> 
>
https://codereview.chromium.org/111033005/diff/350001/chrome/common/profile_m...
> chrome/common/profile_management_switches.cc:24: return
> CommandLine::ForCurrentProcess()->HasSwitch(command_switch);
> Hi Brian.
> 
> Shouldn't this take priority? For people enabling it for testing purpose?

By that you mean, if somebody specifically adds that flag, we should honor it
even if an experiment is trying to do something to the contrary?

Good question.  We could get odd results due to mixing between experiment
settings and command-line flags.  There's another CL
(https://codereview.chromium.org/136693007/) that allows a single command-line
flag to behave the same as enabling the experiment; perhaps it, at least, should
come before the experiment test.