Long ago we taught the commit-queue how to be tolerant of flaky tests (just running the tests again and again until it gets consistent results). Recently we taught it how to report to bugzilla when it finds flaky tests.
We're still working on improving the information it reports (including crash reports, failure type, etc.) But even in the limited data we have, I'm seeing a lot of non-text-diff failures from inspector tests: https://bugs.webkit.org/buglist.cgi?bug_id=50923%2C51014%2C51153%2C50987%2C50855%2C51019%2C51131%2C50958%2C50868%2C51182%2C51020%2C51053%2C51470%2C51111%2C51320%2C51138%2C50822%2C51102%2C51039%2C50889%2C51384%2C50880%2C51442%2C51026%2C50863%2C51337%2C50875%2C51314%2C51095%2C51098%2C51056%2C51383%2C51028%2C51304%2C51301%2C50995%2C51009%2C51002%2C50823%2C51024%2C51130%2C50959%2C51361%2C51294%2C51154&field-1-0-0=bug_id&query_format=advanced&remaction=&type-1-0-0=anyexact&value-1-0-0=50923%2C51014%2C51153%2C50987%2C50855%2C51019%2C51131%2C50958%2C50868%2C51182%2C51020%2C51053%2C51470%2C51111%2C51320%2C51138%2C50822%2C51102%2C51039%2C50889%2C51384%2C50880%2C51442%2C51026%2C50863%2C51337%2C50875%2C51314%2C51095%2C51098%2C51056%2C51383%2C51028%2C51304%2C51301%2C50995%2C51009%2C51002%2C50823%2C51024%2C51130%2C50959%2C51361%2C51294%2C51154&order=bugs.short_desc%2Cbugs.bug_id&query_based_on= And I suspect that there may be a marauding crasher in the inspector code. https://bugs.webkit.org/show_bug.cgi?id=50868 seems to support this hypothesis. Do any of the inspector folks have any ideas? -eric p.s. I'm working on teaching the queue how to upload crash reports. I hope to have the patch posted before holiday break. _______________________________________________ webkit-dev mailing list [email protected] http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev
