WebKit Bugzilla
New
Browse
Log In
×
Sign in with GitHub
or
Remember my login
Create Account
·
Forgot Password
Forgotten password account recovery
RESOLVED FIXED
204707
Add a unit-test to EWS for scenario when there is flakiness with patch and also on clean tree
https://bugs.webkit.org/show_bug.cgi?id=204707
Summary
Add a unit-test to EWS for scenario when there is flakiness with patch and al...
Aakash Jain
Reported
2019-11-29 10:06:01 PST
Add a unit-test to old EWS for scenario when there is flakiness with patch and also on clean tree.
Attachments
Unit-test for old EWS
(1.79 KB, patch)
2019-11-29 10:08 PST
,
Aakash Jain
no flags
Details
Formatted Diff
Diff
Similar unit-test for new EWS
(1.98 KB, patch)
2019-11-29 10:16 PST
,
Aakash Jain
no flags
Details
Formatted Diff
Diff
View All
Add attachment
proposed patch, testcase, etc.
Aakash Jain
Comment 1
2019-11-29 10:08:16 PST
Created
attachment 384518
[details]
Unit-test for old EWS
Aakash Jain
Comment 2
2019-11-29 10:09:51 PST
Committed
r252943
: <
https://trac.webkit.org/changeset/252943
>
Radar WebKit Bug Importer
Comment 3
2019-11-29 10:10:18 PST
<
rdar://problem/57533998
>
Aakash Jain
Comment 4
2019-11-29 10:16:59 PST
Created
attachment 384519
[details]
Similar unit-test for new EWS
Aakash Jain
Comment 5
2019-11-29 10:18:13 PST
.
Aakash Jain
Comment 6
2019-11-29 10:20:48 PST
Committed
r252946
: <
https://trac.webkit.org/changeset/252946
>
Alexey Proskuryakov
Comment 7
2019-11-29 22:06:53 PST
I think that I know why I didn't believe that this was the old EWS behavior. Looking at the "if self._results_failed_different_tests" branch in PatchAnalysisTask._retry_layout_tests, I see that it calls self._should_defer_patch_or_throw, which says that the result should be a PASS. But then we ignore the result, and make it a DEFER anyway! if tests_that_consistently_failed: if self._should_defer_patch_or_throw(tests_that_consistently_failed, first_results_archive, first_script_error, first_failure_status_id): return False # Defer patch # At this point we know that at least one test flaked, but no consistent failures # were introduced. This is a bit of a grey-zone. return False # Defer patch What's the point of the first "return False" here if we are going to defer either way?
Note
You need to
log in
before you can comment on or make changes to this bug.
Top of Page
Format For Printing
XML
Clone This Bug