19:50:28 There was 1 failure: 19:50:28 19:50:28 1) MediaWiki\CheckUser\Tests\Integration\Maintenance\GenerateStatsAboutClientHintsDataTest::testGenerateCounts 19:50:28 Failed asserting that two arrays are equal. 19:50:28 --- Expected 19:50:28 +++ Actual 19:50:28 @@ @@ 19:50:28 'totalRowCount' => Array (...) 19:50:28 'averageItemsPerNamePerReferenceId' => Array (...) 19:50:28 'mapTableRowCountBreakdown' => Array (...) 19:50:28 - 'averageMapRowsPerReferenceId' => 8.5 19:50:28 + 'averageMapRowsPerReferenceId' => '7.5556' 19:50:28 'invalidMapRowsCount' => 1 19:50:28 ) 19:50:28 19:50:28 /workspace/src/tests/phpunit/MediaWikiTestCaseTrait.php:287 19:50:28 /workspace/src/extensions/CheckUser/tests/phpunit/integration/maintenance/GenerateStatsAboutClientHintsDataTest.php:68 19:50:28 phpvfscomposer:///workspace/src/vendor/phpunit/phpunit/phpunit:106
Description
Description
Details
Details
| Subject | Repo | Branch | Lines +/- | |
|---|---|---|---|---|
| [WIP] Experimenting to find cause of test failure | mediawiki/extensions/CheckUser | master | +8 -0 |
Related Objects
Related Objects
Event Timeline
Comment Actions
Seems unlikely, though, because the test passed just a few hours ago, while the patch in T379195 merged ~10 hours ago.
Comment Actions
Strangely it passes on the gated extensions test run, but fails on the other (e.g. https://integration.wikimedia.org/ci/job/wmf-quibble-vendor-mysql-php74/23728/consoleFull#console-section-0 vs https://integration.wikimedia.org/ci/job/quibble-vendor-mysql-php74-noselenium/52895/consoleFull#console-section-0). If not related to parallel testing, perhaps it's an extension that CheckUser depends on but is not a gated extension that is conflicting here?
The test passes for me locally on the master branch
Comment Actions
Change #1087981 had a related patch set uploaded (by Dreamy Jazz; author: Dreamy Jazz):
[mediawiki/extensions/CheckUser@master] [WIP] Experimenting to find cause of test failure
Comment Actions
Interestingly, the test fails for me locally when I have https://gerrit.wikimedia.org/r/c/mediawiki/extensions/CheckUser/+/1087934/9 checked out. Maybe something changed in between the last test run on that patch and gate-and-submit-wmf in another repository that broke things?
Comment Actions
This isn't actually broken on the master branch. It seems that in https://gerrit.wikimedia.org/r/c/mediawiki/extensions/CheckUser/+/1087934/9 the test was using array_rand, and the behaviour that causes the test failure wasn't seen until a few test runs in.
Comment Actions
Change #1087981 abandoned by Dreamy Jazz:
[mediawiki/extensions/CheckUser@master] [WIP] Experimenting to find cause of test failure
Comment Actions
Ah, right. Thanks for spotting that. I was too quick to jump to conclusions that it had to do with the parallel testing rollout/changes.