stylo: Enable reftestts for checking stacking context in layout/reftests/web-animations

RESOLVED FIXED in Firefox 55

Status

()

Core
CSS Parsing and Computation
P2
normal
RESOLVED FIXED
5 months ago
2 months ago

People

(Reporter: hiro, Assigned: boris)

Tracking

(Blocks: 1 bug)

unspecified
mozilla55
Points:
---
Dependency tree / graph

Firefox Tracking Flags

(firefox55 fixed)

Details

MozReview Requests

()

Submitter Diff Changes Open Issues Last Updated
Loading...
Error loading review requests:

Attachments

(1 attachment)

(Reporter)

Description

5 months ago
I don't know the reason yet, but those tests fail intermittently, very high frequent. 
An interesting point is that we have similar tests (checking stacking context) for CSS animations, but they are not intermittent.
Blocks: 1243581
(Reporter)

Comment 1

3 months ago
Boris, could you please check that these tests don't cause intermittent failures with your patch for bug 1334036?
Flags: needinfo?(boris.chiou)
(Assignee)

Comment 2

3 months ago
Sure!
(Assignee)

Comment 3

2 months ago
(In reply to Hiroyuki Ikezoe (:hiro) from comment #0)
> I don't know the reason yet, but those tests fail intermittently, very high
> frequent. 
> An interesting point is that we have similar tests (checking stacking
> context) for CSS animations, but they are not intermittent.

Hi Hiro, does the "very high frequent" happens on local build? Or only happens on try server? (In other words, can you easily reproduce these intermittents?)
Flags: needinfo?(boris.chiou)
(Assignee)

Comment 4

2 months ago
(In reply to Boris Chiou [:boris] from comment #3)
> (In reply to Hiroyuki Ikezoe (:hiro) from comment #0)
> > I don't know the reason yet, but those tests fail intermittently, very high
> > frequent. 
> > An interesting point is that we have similar tests (checking stacking
> > context) for CSS animations, but they are not intermittent.
> 
> Hi Hiro, does the "very high frequent" happens on local build? Or only
> happens on try server? (In other words, can you easily reproduce these
> intermittents?)

any way, I will send to try to check this.
(Reporter)

Comment 5

2 months ago
(In reply to Boris Chiou [:boris] from comment #3)
> (In reply to Hiroyuki Ikezoe (:hiro) from comment #0)
> > I don't know the reason yet, but those tests fail intermittently, very high
> > frequent. 
> > An interesting point is that we have similar tests (checking stacking
> > context) for CSS animations, but they are not intermittent.
> 
> Hi Hiro, does the "very high frequent" happens on local build? Or only
> happens on try server? (In other words, can you easily reproduce these
> intermittents?)

Locally.  Has this been solved?
(Assignee)

Comment 6

2 months ago
(In reply to Hiroyuki Ikezoe (:hiro) from comment #5)
> (In reply to Boris Chiou [:boris] from comment #3)
> > (In reply to Hiroyuki Ikezoe (:hiro) from comment #0)
> > > I don't know the reason yet, but those tests fail intermittently, very high
> > > frequent. 
> > > An interesting point is that we have similar tests (checking stacking
> > > context) for CSS animations, but they are not intermittent.
> > 
> > Hi Hiro, does the "very high frequent" happens on local build? Or only
> > happens on try server? (In other words, can you easily reproduce these
> > intermittents?)
> 
> Locally.  Has this been solved?

I enable all the stacking-context-{*} reftests in layout/reftests/web-animations/reftest.list, and
run "./mach reftest --disable-e10s layout/reftests/web-animations/reftest.list".

tried 10 times: always get this result locally (mac os):

REFTEST INFO | Successful: 18 (18 pass, 0 load only)
REFTEST INFO | Unexpected: 0 (0 unexpected fail, 0 unexpected pass, 0 unexpected asserts, 0 failed load, 0 exception)
REFTEST INFO | Known problems: 2 (2 known fail, 0 known asserts, 0 random, 0 skipped, 0 slow)

The 2 known problems are style-updates-on-xxxx.
(Assignee)

Comment 7

2 months ago
(In reply to Boris Chiou [:boris] from comment #6)
> I enable all the stacking-context-{*} reftests in
> layout/reftests/web-animations/reftest.list, and
> run "./mach reftest --disable-e10s
> layout/reftests/web-animations/reftest.list".
> 
> tried 10 times: always get this result locally (mac os):
> 
> REFTEST INFO | Successful: 18 (18 pass, 0 load only)
> REFTEST INFO | Unexpected: 0 (0 unexpected fail, 0 unexpected pass, 0
> unexpected asserts, 0 failed load, 0 exception)
> REFTEST INFO | Known problems: 2 (2 known fail, 0 known asserts, 0 random, 0
> skipped, 0 slow)
> 
> The 2 known problems are style-updates-on-xxxx.

https://treeherder.mozilla.org/#/jobs?repo=try&revision=b0823142defdbd3d7f04f279a66d02c76cf3cfca

We can check Rs16 in the try. Only style-updates-on-{*}s are not passed.
(Reporter)

Comment 8

2 months ago
Thanks for confirming!
Assignee: nobody → boris.chiou
Blocks: 1334036
Status: NEW → ASSIGNED
Comment hidden (mozreview-request)
(Reporter)

Comment 10

2 months ago
Oops, I did accidentally add bug 1334036 into blocks..  And I can't change it to dependency list because recursive dependencies?
(Assignee)

Updated

2 months ago
No longer blocks: 1334036
See Also: → bug 1334036
(Assignee)

Updated

2 months ago
No longer blocks: 1302945
Depends on: 1302945, 1334036
See Also: bug 1334036
(Assignee)

Comment 11

2 months ago
I tested 100 times locally with patches of bug 1334036 (on MacOS, --disable-e10s), but unfortunately, when I ran the reftest.list (in layout/reftests/web-animations) in the first trial, still have one intermittent: "stacking-context-transform-changing-display-property.html", and the rest 99 trials, all tests are passed. (probability: 1/100).

I start to believe maybe it's the problem of the environment of my laptop because I cannot test reftest with e10s (resolution is too low.) and this only happened in the first trial. (/‵Д′)/~ ╧╧
Comment hidden (mozreview-request)
(Reporter)

Comment 13

2 months ago
mozreview-review
Comment on attachment 8867643 [details]
Bug 1346049 - Enable reftests for checking stacking context.

https://reviewboard.mozilla.org/r/139234/#review144446
Attachment #8867643 - Flags: review?(hikezoe) → review+
(Reporter)

Comment 14

2 months ago
(In reply to Boris Chiou [:boris] from comment #11)
> I tested 100 times locally with patches of bug 1334036 (on MacOS,
> --disable-e10s), but unfortunately, when I ran the reftest.list (in
> layout/reftests/web-animations) in the first trial, still have one
> intermittent: "stacking-context-transform-changing-display-property.html",
> and the rest 99 trials, all tests are passed. (probability: 1/100).
> 
> I start to believe maybe it's the problem of the environment of my laptop
> because I cannot test reftest with e10s (resolution is too low.) and this
> only happened in the first trial. (/‵Д′)/~ ╧╧

Yeah, I couldn't run some reftest with e10s either. In Taipei Shing told me that there are some tests that are not able to run with e10s.

Anyway, let's see what happens on our automation servers.
Comment hidden (mozreview-request)

Comment 16

2 months ago
Pushed by bchiou@mozilla.com:
https://hg.mozilla.org/integration/autoland/rev/f251b0f3faee
Enable reftests for checking stacking context. r=hiro

Comment 17

2 months ago
bugherder
https://hg.mozilla.org/mozilla-central/rev/f251b0f3faee
Status: ASSIGNED → RESOLVED
Last Resolved: 2 months ago
status-firefox55: --- → fixed
Resolution: --- → FIXED
Target Milestone: --- → mozilla55
You need to log in before you can comment on or make changes to this bug.