Remove ShapeOutside tests in blink_perf.layout, as the only seem to be measuring overhead.

Remove ShapeOutside tests in blink_perf.layout, as they only seem to be
measuring overhead.

These tests used to report numbers in the 200ms range, but they've all
dropped down to the 1ms range. They no longer seem to be covering
anything but overhead.

BUG=498051
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=196954

[sullivan, 2015-06-10 15:46:41]

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

[Julien - ping for review, 2015-06-10 17:29:14]

jchaffraix@chromium.org changed reviewers:
+ bjonesbe@adobe.com

[Julien - ping for review, 2015-06-10 17:29:15]

What's wrong with removing these tests from Blink? Skipping doesn't give us much
AFAICT as we don't have a path to unskip them.

Also your data point is based on a single test
(https://code.google.com/p/chromium/codesearch#chromium/src/third_party/WebKit...),
not on all the tests. I would like more details if you're going to remove all
our shape-outside perf testing.

+Bem as this is their code and they may have some interest in this effort

[sullivan, 2015-06-10 17:37:01]

On 2015/06/10 17:29:15, Julien Chaffraix - PST wrote:
> What's wrong with removing these tests from Blink? Skipping doesn't give us
much
> AFAICT as we don't have a path to unskip them.

I'd be happy to update this to remove the files.

> Also your data point is based on a single test
>
(https://code.google.com/p/chromium/codesearch#chromium/src/third_party/WebKit...),
> not on all the tests. I would like more details if you're going to remove all
> our shape-outside perf testing.

I think I'm missing something here? It looks like all the shape-outside perf
tests went down to near-zero:
https://chromeperf.appspot.com/report?masters=ChromiumPerf&bots=chromium-rel-...
https://chromeperf.appspot.com/report?masters=ChromiumPerf&bots=chromium-rel-...
https://chromeperf.appspot.com/report?masters=ChromiumPerf&bots=chromium-rel-...
https://chromeperf.appspot.com/report?masters=ChromiumPerf&bots=chromium-rel-...
https://chromeperf.appspot.com/report?masters=ChromiumPerf&bots=chromium-rel-...
https://chromeperf.appspot.com/report?masters=ChromiumPerf&bots=chromium-rel-...
https://chromeperf.appspot.com/report?masters=ChromiumPerf&bots=chromium-rel-...
https://chromeperf.appspot.com/report?masters=ChromiumPerf&bots=chromium-rel-...


> +Bem as this is their code and they may have some interest in this effort

[Julien - ping for review, 2015-06-10 20:41:19]

On 2015/06/10 at 17:37:01, sullivan wrote:
> On 2015/06/10 17:29:15, Julien Chaffraix - PST wrote:
> > What's wrong with removing these tests from Blink? Skipping doesn't give us
much
> > AFAICT as we don't have a path to unskip them.
> 
> I'd be happy to update this to remove the files.

We should just remove the tests in Blink.

> > Also your data point is based on a single test
> >
(https://code.google.com/p/chromium/codesearch#chromium/src/third_party/WebKit...),
> > not on all the tests. I would like more details if you're going to remove
all
> > our shape-outside perf testing.
> 
> I think I'm missing something here? It looks like all the shape-outside perf
tests went down to near-zero:
>
https://chromeperf.appspot.com/report?masters=ChromiumPerf&bots=chromium-rel-...
>
https://chromeperf.appspot.com/report?masters=ChromiumPerf&bots=chromium-rel-...
>
https://chromeperf.appspot.com/report?masters=ChromiumPerf&bots=chromium-rel-...
>
https://chromeperf.appspot.com/report?masters=ChromiumPerf&bots=chromium-rel-...
>
https://chromeperf.appspot.com/report?masters=ChromiumPerf&bots=chromium-rel-...
>
https://chromeperf.appspot.com/report?masters=ChromiumPerf&bots=chromium-rel-...
>
https://chromeperf.appspot.com/report?masters=ChromiumPerf&bots=chromium-rel-...
>
https://chromeperf.appspot.com/report?masters=ChromiumPerf&bots=chromium-rel-...

The linked to bug is pretty ambiguous when it comes to which test(s) it is
talking about. The regression is about a single test but some-but-not-all
comments delve into all the shape-outside tests.

You can add that the terse description doesn't give any justification to the
mass-removal whatsoever which didn't clear my confusion :(

[sullivan, 2015-06-10 21:43:53]

On 2015/06/10 20:41:19, Julien Chaffraix - PST wrote:
> On 2015/06/10 at 17:37:01, sullivan wrote:
> > On 2015/06/10 17:29:15, Julien Chaffraix - PST wrote:
> > > What's wrong with removing these tests from Blink? Skipping doesn't give
us
> much
> > > AFAICT as we don't have a path to unskip them.
> > 
> > I'd be happy to update this to remove the files.
> 
> We should just remove the tests in Blink.

Done.

> > > Also your data point is based on a single test
> > >
>
(https://code.google.com/p/chromium/codesearch#chromium/src/third_party/WebKit...),
> > > not on all the tests. I would like more details if you're going to remove
> all
> > > our shape-outside perf testing.
> > 
> > I think I'm missing something here? It looks like all the shape-outside perf
> tests went down to near-zero:
> >
>
https://chromeperf.appspot.com/report?masters=ChromiumPerf&bots=chromium-rel-...
> >
>
https://chromeperf.appspot.com/report?masters=ChromiumPerf&bots=chromium-rel-...
> >
>
https://chromeperf.appspot.com/report?masters=ChromiumPerf&bots=chromium-rel-...
> >
>
https://chromeperf.appspot.com/report?masters=ChromiumPerf&bots=chromium-rel-...
> >
>
https://chromeperf.appspot.com/report?masters=ChromiumPerf&bots=chromium-rel-...
> >
>
https://chromeperf.appspot.com/report?masters=ChromiumPerf&bots=chromium-rel-...
> >
>
https://chromeperf.appspot.com/report?masters=ChromiumPerf&bots=chromium-rel-...
> >
>
https://chromeperf.appspot.com/report?masters=ChromiumPerf&bots=chromium-rel-...
> 
> The linked to bug is pretty ambiguous when it comes to which test(s) it is
> talking about. The regression is about a single test but some-but-not-all
> comments delve into all the shape-outside tests.

Sorry about that! It's autogenerated by the perf dashboard. Automatically
getting the titles right turns out to be hard since new regressions can be added
to the bug after it gets filed. You can see the full list of charts the bug is
referring to in the "All graphs for this bug" link in comment 1 of perf
dashboard bugs. Link for this bug:
https://chromeperf.appspot.com/group_report?bug_id=498051

> You can add that the terse description doesn't give any justification to the
> mass-removal whatsoever which didn't clear my confusion :(

Sorry, I updated the CL description.

[Julien - ping for review, 2015-06-10 22:21:28]

lgtm with the description wrapping at <80 characters (I follow a 72 characters
limit per
https://groups.google.com/a/chromium.org/forum/#!topic/blink-dev/rqGhY7PfkYY).

[sullivan, 2015-06-11 13:57:41]

On 2015/06/10 22:21:28, Julien Chaffraix - PST wrote:
> lgtm with the description wrapping at <80 characters (I follow a 72 characters
> limit per
> https://groups.google.com/a/chromium.org/forum/#!topic/blink-dev/rqGhY7PfkYY).

Done.

[sullivan, 2015-06-11 13:57:45]

The CQ bit was checked by sullivan@chromium.org

[commit-bot: I haz the power, 2015-06-11 13:58:08]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1175733004/20001

[commit-bot: I haz the power, 2015-06-11 15:12:58]

Committed patchset #2 (id:20001) as
https://src.chromium.org/viewvc/blink?view=rev&revision=196954

[Bem Jones-Bey (adobe), 2015-06-11 21:47:51]

bjonesbe@adobe.com changed reviewers:
+ zoltan@adobe.com

[Bem Jones-Bey (adobe), 2015-06-11 21:47:52]

I'm a bit late to the party, but I'm adding Zoltan as a FYI since he wrote the
tests in question.

[Zoltan, 2015-06-11 21:58:33]

On 2015/06/11 21:47:52, Bem Jones-Bey (adobe) wrote:
> I'm a bit late to the party, but I'm adding Zoltan as a FYI since he wrote the
> tests in question.

The change from https://codereview.chromium.org/675983004 didn't include these
files, so they didn't get the new relayout behavior. That's why they stopped
working. We should restore the updated version of them. I can do that later this
week.