Add support for sparse histograms to WebKitPlatformSupportImpl

Add support for creating sparse histograms to WebKitPlatformSupportImpl

This is the chromium side of adding support for creating sparse histograms
to Blink. The Blink side of this is http://crrev.com/19490005.

BUG=139612
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=212717

[ajuma, 2013-07-19 17:15:40]

PTAL.

[jar (doing other things), 2013-07-19 17:55:27]

LGTM: But please add the comment below.

https://codereview.chromium.org/19741003/diff/1/webkit/glue/webkitplatformsup...
File webkit/glue/webkitplatformsupport_impl.cc (right):

https://codereview.chromium.org/19741003/diff/1/webkit/glue/webkitplatformsup...
webkit/glue/webkitplatformsupport_impl.cc:448: UMA_HISTOGRAM_SPARSE_SLOWLY(name,
sample);
nit: This works perfectly... but it scared me, since you can't use the UMA
histogram macros (usually) when the name is variable.  Can you add the comment:

// For sparse histograms, we can use the macro, as it does not incorporate a
static.



FWIW: We should not use a macro definition for this, and the sparse_histogram.h
file should instead define a *function* in this case.  I didn't see the review,
or I would have pointed out that we prefer functions, and only use macros when
we are forced to.  Oh well.

[ajuma, 2013-07-19 18:37:20]

+jamesr for webkit/ OWNERS.

https://codereview.chromium.org/19741003/diff/1/webkit/glue/webkitplatformsup...
File webkit/glue/webkitplatformsupport_impl.cc (right):

https://codereview.chromium.org/19741003/diff/1/webkit/glue/webkitplatformsup...
webkit/glue/webkitplatformsupport_impl.cc:448: UMA_HISTOGRAM_SPARSE_SLOWLY(name,
sample);
On 2013/07/19 17:55:28, jar wrote:
> nit: This works perfectly... but it scared me, since you can't use the UMA
> histogram macros (usually) when the name is variable.  Can you add the
comment:
> 
> // For sparse histograms, we can use the macro, as it does not incorporate a
> static.
> 
> 
> 
> FWIW: We should not use a macro definition for this, and the
sparse_histogram.h
> file should instead define a *function* in this case.  I didn't see the
review,
> or I would have pointed out that we prefer functions, and only use macros when
> we are forced to.  Oh well.

Done.

[jamesr, 2013-07-19 19:12:13]

lgtm

[commit-bot: I haz the power, 2013-07-19 19:15:40]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/ajuma@chromium.org/19741003/5001

[commit-bot: I haz the power, 2013-07-20 01:45:21]

Change committed as 212717