TestWebKitAPI.IndexedDB.DatabaseProcessKill is a flakey timeout on iOS wk2. HISTORY: https://results.webkit.org/?suite=api-tests&test=TestWebKitAPI.IndexedDB.DatabaseProcessKill Link to stdio file where timeout is noted: https://build.webkit.org/#/builders/28/builds/910/steps/11/logs/stdio It is flakey, but it does timeout more than it passes.
<rdar://problem/76558602>
This appears to be trivially reproducible with a release build, but does not appear to affect debug or production builds. As such, I did reproduce the failure at iOS Release ToT using the following test: run-api-tests --ios-simulator TestWebKitAPI.IndexedDB.DatabaseProcessKill --child-process=50 Note that it is only occurring on release, and will pass on Production and Debug builds. I was also able to verify a regress point. The timeouts occur in r275792, and did not occur in 275791. It appears this changeset caused the timeouts: https://trac.webkit.org/changeset/275792/webkit
Created attachment 425825 [details] Patch
Comment on attachment 425825 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=425825&action=review > Tools/TestWebKitAPI/Tests/WebKitCocoa/IndexedDBDatabaseProcessKill.mm:85 > [webView _test_waitForDidFinishNavigation]; Why not just replace this with TestWebKitAPI::Util::run(&receivedScriptMessage)?
Comment on attachment 425825 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=425825&action=review >> Tools/TestWebKitAPI/Tests/WebKitCocoa/IndexedDBDatabaseProcessKill.mm:85 >> [webView _test_waitForDidFinishNavigation]; > > Why not just replace this with TestWebKitAPI::Util::run(&receivedScriptMessage)? Maybe it could just be removed. I'll test.
https://trac.webkit.org/changeset/275890/webkit