Bug 217001

Summary: fast/scrolling/latching/latched-scroll-into-nonfast-region.html is a flaky failure
Product: WebKit Reporter: Karl Rackler <rackler>
Component: ScrollingAssignee: Nobody <webkit-unassigned>
Status: NEW ---    
Severity: Normal CC: ap, simon.fraser, webkit-bot-watchers-bugzilla, webkit-bug-importer
Priority: P2 Keywords: InRadar
Version: WebKit Nightly Build   
Hardware: Unspecified   
OS: Unspecified   

Description Karl Rackler 2020-09-25 16:41:51 PDT
Description:
fast/scrolling/latching/latched-scroll-into-nonfast-region.html

The first failure that I see this on the dashboard was 09/03/20 at r266538 on Catalina debug wk2.  It has had a frequency of approximately every 4 days, but on 09/16/20 increased in frequency to daily occurrences. 

I can reproduce this on r266932, but unable to reproduce on r266931.
Commit r266932 has to do with MediaToolboxSoftLink.cpp and MediaToolboxSoftLink.h, which could have introduced the flakiness.  The change was introduced here https://trac.webkit.org/changeset/266932/webkit.

REPRODUCTION STEPS
Command: 
run-webkit-tests --iterations 2000 --exit-after-n-failures 1 --exit-after-n-crashes-or-timeouts 1 --debug-rwt-logging --no-retry --force --no-build -f fast/scrolling/latching/latched-scroll-into-nonfast-region.html

Result: 
Regressions: Unexpected text-only failures (1)
  fast/scrolling/latching/latched-scroll-into-nonfast-region.html [ Failure ]

History:
https://results.webkit.org/?suite=layout-tests&test=fast%2Fscrolling%2Flatching%2Flatched-scroll-into-nonfast-region.html&limit=50000&version_name=Catalina&flavor=wk2

Diff Log:
--- /Volumes/Data/slave/catalina-release-tests-wk2/build/layout-test-results/fast/scrolling/latching/latched-scroll-into-nonfast-region-expected.txt
+++ /Volumes/Data/slave/catalina-release-tests-wk2/build/layout-test-results/fast/scrolling/latching/latched-scroll-into-nonfast-region-actual.txt
@@ -4,7 +4,7 @@
 
 
 Scrolling the main frame with latching
-PASS wheelEventHandlerCount is 0
+FAIL wheelEventHandlerCount should be 0. Was 3.
 PASS successfullyParsed is true
 
 TEST COMPLETE
Comment 1 Radar WebKit Bug Importer 2020-09-25 16:42:34 PDT
<rdar://problem/69600982>
Comment 2 Karl Rackler 2020-10-16 07:18:48 PDT
This issue is also affecting Mojave.

I have marked this test as failing also on Mojave while this issue is investigated.
https://trac.webkit.org/changeset/268588/webkit
Comment 3 Alexey Proskuryakov 2020-10-17 18:40:25 PDT
Looking at test history (both webkit.org and internal), I think that this was just landed flaky. Frequency mysteriously changes over time, but some bots saw multiple failures a day from the start.