Revert Renderer.LayoutMs.

Revert Renderer.LayoutMs.

We created this histogram in the hope that some sort of useful pattern
might emerge. It turned out that there's a huge spike at 0 due to
incremental layouts, then it tapers off smoothly all the way through 1
minute, apparently an exponential distribution. The mean is about 1ms.
If performance optimizations improve benchmarked cases but regress
un-benchmarked cases, we would not be able to tell from these histograms
-- a fraction of samples would be smeared one way and another fraction
would be smeared the other way, likely resulting in no visible
difference. So this histogram is useless and a waste of space.

Instead of counting all layouts together, it might make more sense to
count specific kinds of layout separately, like a
micro-benchmark-in-the-wild. E.g.  one histogram for when a layout
touches >100 floats/table cells/lots of complex text. If we identify
kinds of layout like that, then maybe we'll add new histograms here in
the hope that they would present a log-normal with a mean value that
could visibly move towards zero due to targetted performance
optimizations, or shrink as sites stop abusing layout.


No BUG, but these histograms were created in these CLs:
https://codereview.chromium.org/882293002
https://codereview.chromium.org/877293003

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=191379

[benjhayden, 2015-03-02 23:33:46]

benjhayden@chromium.org changed reviewers:
+ jchaffraix@chromium.org

[benjhayden, 2015-03-02 23:33:47]

PTAL

[Julien - ping for review, 2015-03-02 23:52:25]

jchaffraix@chromium.org changed reviewers:
+ dsinclair@chromium.org

[Julien - ping for review, 2015-03-02 23:52:26]

lgtm.

The description should be improved before landing as it's not useful or clear
for anyone who hasn't seen said histogram. Here is a suggestion for the wording:

Remove Render.LayoutMs

This histogram tried to capture *every* possible
layout in the wild and no real pattern emerged,
thus it's better to eliminate this noisy histogram.

(extra point if we have a BUG discussion about this but I assume we didn't file
one).

[benjhayden, 2015-03-04 19:31:33]

On 2015/03/02 at 23:52:26, jchaffraix wrote:
> lgtm.
> 
> The description should be improved before landing as it's not useful or clear
for anyone who hasn't seen said histogram. Here is a suggestion for the wording:
> 
> Remove Render.LayoutMs
> 
> This histogram tried to capture *every* possible
> layout in the wild and no real pattern emerged,
> thus it's better to eliminate this noisy histogram.
> 
> (extra point if we have a BUG discussion about this but I assume we didn't
file one).

PTAL

[Julien - ping for review, 2015-03-04 21:31:12]

lgtm, the description should be wrapped at 80 columns (see
https://groups.google.com/a/chromium.org/forum/#!topic/blink-dev/rqGhY7PfkYY).
Dana advise 72 and that's what I usually advise as a good guidance.

Ideally the text should be balanced across lines but that's my design OCD
talking. :-)

[benjhayden, 2015-03-05 17:23:29]

On 2015/03/04 at 21:31:12, jchaffraix wrote:
> lgtm, the description should be wrapped at 80 columns (see
https://groups.google.com/a/chromium.org/forum/#!topic/blink-dev/rqGhY7PfkYY).
Dana advise 72 and that's what I usually advise as a good guidance.

What year is it?

> 
> Ideally the text should be balanced across lines but that's my design OCD
talking. :-)

I used vim's gq.

[benjhayden, 2015-03-05 17:23:39]

The CQ bit was checked by benjhayden@chromium.org

[commit-bot: I haz the power, 2015-03-05 17:24:18]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/969973002/1

[Julien - ping for review, 2015-03-05 17:57:56]

On 2015/03/05 at 17:23:29, benjhayden wrote:
> On 2015/03/04 at 21:31:12, jchaffraix wrote:
> > lgtm, the description should be wrapped at 80 columns (see
https://groups.google.com/a/chromium.org/forum/#!topic/blink-dev/rqGhY7PfkYY).
Dana advise 72 and that's what I usually advise as a good guidance.
> 
> What year is it?

Not my decision, Chrome / Google's:
http://google-styleguide.googlecode.com/svn/trunk/cppguide.html#Line_Length

[commit-bot: I haz the power, 2015-03-05 20:08:09]

Committed patchset #1 (id:1) as
https://src.chromium.org/viewvc/blink?view=rev&revision=191379