Refactor declaration order for 2 A/B tests....

Refactor declaration order for 2 A/B tests.

Rearranged the declaration order for probability groups contained in 2 A/B
tests in a way that preserves the remaining probability space to the default
value, after the other groups receive an equal share of the sample space.

BUG=50451
r=jar



Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=53894

[jar (doing other things), 2010-07-28 00:39:42]

LGTM (please adjust the comment nit below)

You should add a BUG= to the description, with a bug indicating that you're
reordering the experiment to make it easier to change values on a beta or stable
branch (with reduced risk of errors).

http://codereview.chromium.org/3080004/diff/1/2
File chrome/browser/browser_main.cc (right):

http://codereview.chromium.org/3080004/diff/1/2#newcode245
chrome/browser/browser_main.cc:245: // declaration order as the
"ConnCountImpact" A/B test defined above.
IMO, you shouldn't bother to refer to the other A/B test, as that really only
confuses the issue.  I think you'd be better off repeating the gist of the
comment.

[ziadh, 2010-07-28 01:22:10]

On 2010/07/28 00:39:42, jar wrote:
> LGTM (please adjust the comment nit below)
> 
> You should add a BUG= to the description, with a bug indicating that you're
> reordering the experiment to make it easier to change values on a beta or
stable
> branch (with reduced risk of errors).
> 
> http://codereview.chromium.org/3080004/diff/1/2
> File chrome/browser/browser_main.cc (right):
> 
> http://codereview.chromium.org/3080004/diff/1/2#newcode245
> chrome/browser/browser_main.cc:245: // declaration order as the
> "ConnCountImpact" A/B test defined above.
> IMO, you shouldn't bother to refer to the other A/B test, as that really only
> confuses the issue.  I think you'd be better off repeating the gist of the
> comment.

Done.

[ziadh, 2010-07-28 01:29:55]

http://codereview.chromium.org/3080004/diff/1/2
File chrome/browser/browser_main.cc (right):

http://codereview.chromium.org/3080004/diff/1/2#newcode245
chrome/browser/browser_main.cc:245: // declaration order as the
"ConnCountImpact" A/B test defined above.
On 2010/07/28 00:39:42, jar wrote:
> IMO, you shouldn't bother to refer to the other A/B test, as that really only
> confuses the issue.  I think you'd be better off repeating the gist of the
> comment.

Done.