Bug 213517

Summary: [ macOS ] fullscreen/video-inside-flex-item.html is a flaky failure
Product: WebKit Reporter: Karl Rackler <rackler>
Component: New BugsAssignee: Sergio Villar Senin <svillar>
Status: RESOLVED FIXED    
Severity: Normal CC: eric.carlson, ews-watchlist, glenn, jer.noble, peng.liu6, philipj, rackler, sergio, simon.fraser, svillar, tsavell, webkit-bot-watchers-bugzilla, webkit-bug-importer
Priority: P2 Keywords: InRadar
Version: WebKit Nightly Build   
Hardware: Mac   
OS: macOS 10.15   
See Also: https://bugs.webkit.org/show_bug.cgi?id=213110
Bug Depends on: 213549    
Bug Blocks:    
Attachments:
Description Flags
Patch
none
Patch none

Description Karl Rackler 2020-06-23 07:37:02 PDT 
fullscreen/video-inside-flex-item.html

Description: 
This test is flaky failing on macOS.  It does seem to be passing on wk2 release.  It has been flaky failing since introduced.
Introduced here: https://trac.webkit.org/changeset/263389/webkit

History: 
https://results.webkit.org/?suite=layout-tests&test=fullscreen%2Fvideo-inside-flex-item.html&platform=ios&platform=mac

Diff: 
--- /Volumes/Data/slave/catalina-debug-tests-wk2/build/layout-test-results/fullscreen/video-inside-flex-item-expected.txt
+++ /Volumes/Data/slave/catalina-debug-tests-wk2/build/layout-test-results/fullscreen/video-inside-flex-item-actual.txt
@@ -2,7 +2,7 @@
 EVENT(webkitfullscreenchange)
 EXPECTED (800 == '800') OK
 EVENT(webkitfullscreenchange)
-EXPECTED (300 == '300') OK
-EXPECTED (150 == '150') OK
+EXPECTED (320 == '300'), OBSERVED '320' FAIL
+EXPECTED (240 == '150'), OBSERVED '240' FAIL
 END OF TEST
Comment 1 Radar WebKit Bug Importer 2020-06-23 07:38:48 PDT 
<rdar://problem/64643289>
Comment 2 Karl Rackler 2020-06-23 08:46:19 PDT 
Created attachment 402564 [details]
Patch
Comment 3 Sergio Villar Senin 2020-06-23 08:49:59 PDT 
I think in this case you should create specific expectations for the macos/ios failing configurations.
Comment 4 Truitt Savell 2020-06-23 08:56:53 PDT 
(In reply to Sergio Villar Senin from comment #3)
> I think in this case you should create specific expectations for the
> macos/ios failing configurations.

In this case the test would be marked to pass on Mac WK2 Release only. This doesn't seem like an effective test right now if it only functions for one specific configuration.
Comment 5 Simon Fraser (smfr) 2020-06-23 09:21:59 PDT 
Full results:

https://results.webkit.org/?suite=layout-tests&test=fullscreen%2Fvideo-inside-flex-item.html

So the test flakes on macOS WK1 Release, always fails on WK1 Debug, always passes on W2 Release, always fails on WK2 Debug.

Sergio, there's a timing issue with this test, please investigate.
Comment 6 Sergio Villar Senin 2020-06-23 09:38:12 PDT 
(In reply to Simon Fraser (smfr) from comment #5)
> Full results:
> 
> https://results.webkit.org/?suite=layout-tests&test=fullscreen%2Fvideo-
> inside-flex-item.html
> 
> So the test flakes on macOS WK1 Release, always fails on WK1 Debug, always
> passes on W2 Release, always fails on WK2 Debug.
> 
> Sergio, there's a timing issue with this test, please investigate.

Sure, I'll take a look.
Comment 7 Sergio Villar Senin 2020-06-23 12:19:44 PDT 
Created attachment 402581 [details]
Patch
Comment 8 EWS 2020-06-23 19:20:54 PDT 
Committed r263436: <https://trac.webkit.org/changeset/263436>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 402581 [details].
Comment 9 Ryan Haddad 2020-06-23 21:06:12 PDT 
The test is still failing:
--- /Volumes/Data/slave/mojave-release-tests-wk1/build/layout-test-results/fullscreen/video-inside-flex-item-expected.txt
+++ /Volumes/Data/slave/mojave-release-tests-wk1/build/layout-test-results/fullscreen/video-inside-flex-item-actual.txt
@@ -2,7 +2,7 @@
 EVENT(webkitfullscreenchange)
 TEST(video.clientWidth == document.body.clientWidth) OK
 EVENT(webkitfullscreenchange)
-TEST(video.clientWidth == oldVideoWidth) OK
-TEST(video.clientHeight == oldVideoHeight) OK
+TEST(video.clientWidth == oldVideoWidth) FAIL
+TEST(video.clientHeight == oldVideoHeight) FAIL
 END OF TEST
Comment 10 Ryan Haddad 2020-06-23 21:11:29 PDT 
Reopening since this isn't fixed.
Comment 11 Sergio Villar Senin 2020-06-24 02:15:19 PDT 
(In reply to Ryan Haddad from comment #10)
> Reopening since this isn't fixed.

It's indeed failing on Debug, I'll come up with something soon.
Comment 12 Sergio Villar Senin 2020-06-24 03:50:49 PDT 
(In reply to Ryan Haddad from comment #10)
> Reopening since this isn't fixed.

I've submitted a patch to bug 213549 to fix this issue. It's working fine in Mac debug for me locally.
Comment 13 Sergio Villar Senin 2020-06-25 03:18:16 PDT 
Fixed in bug 213549