Split <object> tests further to reduce timeout risk

Split <object> tests further to reduce timeout risk

BUG=392640
Committed: https://crrev.com/7f21a74094c93b91b2c47e5f1b9ff79841ddab21
Cr-Commit-Position: refs/heads/master@{#356822}

[davve, 2015-10-29 08:24:20]

The CQ bit was checked by davve@opera.com to run a CQ dry run

[commit-bot: I haz the power, 2015-10-29 08:25:16]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1423093005/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1423093005/1

[davve, 2015-10-29 09:24:08]

davve@opera.com changed reviewers:
+ fs@opera.com

[davve, 2015-10-29 09:24:08]

Down to ~72 tests per file now. Not sure it's low enough, the tests take ~2900
ms on my Nexus 5 now and around ~600 ms on my desktop. I suppose I could split
them further given my mighty python script :)

[commit-bot: I haz the power, 2015-10-29 09:24:44]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2015-10-29 09:24:44]

Dry run: This issue passed the CQ dry run.

[davve, 2015-10-29 09:45:01]

On 2015/10/29 09:24:08, David Vest wrote:
> ... the tests take ~2900 ms on my Nexus 5

To clarify, this was given as a frame of reference not as a indication we should
actually run these test on Android (as part of SmokeTests or anything).

[fs, 2015-10-29 09:52:46]

lgtm

On 2015/10/29 at 09:24:08, davve wrote:
> Down to ~72 tests per file now. Not sure it's low enough, the tests take ~2900
ms on my Nexus 5 now and around ~600 ms on my desktop. I suppose I could split
them further given my mighty python script :)

I think that the various debug bots are the ones to worry (most) about:

http://test-results.appspot.com/dashboards/flakiness_dashboard.html#tests=svg...

(I notice now that the win_chromium_rel_ng bot seems to use 6s pretty
consistently for some of the tests, so a split factor of > 2 is hopefully
enough...)

https://codereview.chromium.org/1423093005/diff/1/third_party/WebKit/LayoutTe...
File third_party/WebKit/LayoutTests/TestExpectations (left):

https://codereview.chromium.org/1423093005/diff/1/third_party/WebKit/LayoutTe...
third_party/WebKit/LayoutTests/TestExpectations:201: crbug.com/392640 [ Release
] svg/as-object/sizing/svg-in-object-placeholder-height-auto.html [ Timeout Pass
]
Could also just keep these and look at the results on the bots (and then remove
if it starts to look ok.) This is probably fine too though.

https://codereview.chromium.org/1423093005/diff/1/third_party/WebKit/LayoutTe...
File third_party/WebKit/LayoutTests/svg/as-object/sizing/svg-in-object.js
(right):

https://codereview.chromium.org/1423093005/diff/1/third_party/WebKit/LayoutTe...
third_party/WebKit/LayoutTests/svg/as-object/sizing/svg-in-object.js:416: 1,
One full (CSS) pixel seems a bit high as an error bound, but I suppose it's not
_that_ high in a relative sense.

[davve, 2015-10-29 10:16:10]

https://codereview.chromium.org/1423093005/diff/1/third_party/WebKit/LayoutTe...
File third_party/WebKit/LayoutTests/TestExpectations (left):

https://codereview.chromium.org/1423093005/diff/1/third_party/WebKit/LayoutTe...
third_party/WebKit/LayoutTests/TestExpectations:201: crbug.com/392640 [ Release
] svg/as-object/sizing/svg-in-object-placeholder-height-auto.html [ Timeout Pass
]
On 2015/10/29 09:52:46, fs wrote:
> Could also just keep these and look at the results on the bots (and then
remove
> if it starts to look ok.) This is probably fine too though.

The tests has been renamed so I would have to expand them into the more
specialized forms.

But I like the idea so I'll give that a shot.

https://codereview.chromium.org/1423093005/diff/1/third_party/WebKit/LayoutTe...
File third_party/WebKit/LayoutTests/svg/as-object/sizing/svg-in-object.js
(right):

https://codereview.chromium.org/1423093005/diff/1/third_party/WebKit/LayoutTe...
third_party/WebKit/LayoutTests/svg/as-object/sizing/svg-in-object.js:416: 1,
On 2015/10/29 09:52:46, fs wrote:
> One full (CSS) pixel seems a bit high as an error bound, but I suppose it's
not
> _that_ high in a relative sense.

Yes, I had roughly the same thought but it felt slightly silly writing a smaller
number. Does 0.5 feel more balanced to you?

[fs, 2015-10-29 10:19:41]

https://codereview.chromium.org/1423093005/diff/1/third_party/WebKit/LayoutTe...
File third_party/WebKit/LayoutTests/svg/as-object/sizing/svg-in-object.js
(right):

https://codereview.chromium.org/1423093005/diff/1/third_party/WebKit/LayoutTe...
third_party/WebKit/LayoutTests/svg/as-object/sizing/svg-in-object.js:416: 1,
On 2015/10/29 at 10:16:10, David Vest wrote:
> On 2015/10/29 09:52:46, fs wrote:
> > One full (CSS) pixel seems a bit high as an error bound, but I suppose it's
not
> > _that_ high in a relative sense.
> 
> Yes, I had roughly the same thought but it felt slightly silly writing a
smaller number. Does 0.5 feel more balanced to you?

0.5 would be "within one pixel" (rather than "within two pixels"), so that seems
better.

[fs, 2015-10-29 10:20:48]

https://codereview.chromium.org/1423093005/diff/1/third_party/WebKit/LayoutTe...
File third_party/WebKit/LayoutTests/TestExpectations (left):

https://codereview.chromium.org/1423093005/diff/1/third_party/WebKit/LayoutTe...
third_party/WebKit/LayoutTests/TestExpectations:201: crbug.com/392640 [ Release
] svg/as-object/sizing/svg-in-object-placeholder-height-auto.html [ Timeout Pass
]
On 2015/10/29 at 10:16:10, David Vest wrote:
> On 2015/10/29 09:52:46, fs wrote:
> > Could also just keep these and look at the results on the bots (and then
remove
> > if it starts to look ok.) This is probably fine too though.
> 
> The tests has been renamed so I would have to expand them into the more
specialized forms.

Right, I didn't think about that bit...

[davve, 2015-10-29 10:30:33]

https://codereview.chromium.org/1423093005/diff/1/third_party/WebKit/LayoutTe...
File third_party/WebKit/LayoutTests/svg/as-object/sizing/svg-in-object.js
(right):

https://codereview.chromium.org/1423093005/diff/1/third_party/WebKit/LayoutTe...
third_party/WebKit/LayoutTests/svg/as-object/sizing/svg-in-object.js:416: 1,
On 2015/10/29 10:19:41, fs wrote:
> On 2015/10/29 at 10:16:10, David Vest wrote:
> > On 2015/10/29 09:52:46, fs wrote:
> > > One full (CSS) pixel seems a bit high as an error bound, but I suppose
it's
> not
> > > _that_ high in a relative sense.
> > 
> > Yes, I had roughly the same thought but it felt slightly silly writing a
> smaller number. Does 0.5 feel more balanced to you?
> 
> 0.5 would be "within one pixel" (rather than "within two pixels"), so that
seems
> better.

Done.

[davve, 2015-10-29 10:32:11]

The CQ bit was checked by davve@opera.com

[davve, 2015-10-29 10:32:11]

The patchset sent to the CQ was uploaded after l-g-t-m from fs@opera.com
Link to the patchset: https://codereview.chromium.org/1423093005/#ps40001
(title: "Also include assert_approx_equals adjustment")

[commit-bot: I haz the power, 2015-10-29 10:32:25]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1423093005/40001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1423093005/40001

[commit-bot: I haz the power, 2015-10-29 12:36:10]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2015-10-29 12:36:54]

Patchset 3 (id:??) landed as
https://crrev.com/7f21a74094c93b91b2c47e5f1b9ff79841ddab21
Cr-Commit-Position: refs/heads/master@{#356822}