NEW 89368
Mark failing tests as failing
https://bugs.webkit.org/show_bug.cgi?id=89368
Summary Mark failing tests as failing
Tony Payne
Reported 2012-06-18 11:48:40 PDT
Mark failing tests as failing
Attachments
Patch (1.69 KB, patch)
2012-06-18 11:49 PDT, Tony Payne
no flags
Patch (1.70 KB, patch)
2012-06-18 11:52 PDT, Tony Payne
tony: review-
tony: commit-queue-
Tony Payne
Comment 1 2012-06-18 11:49:38 PDT
Tony Payne
Comment 2 2012-06-18 11:52:05 PDT
Adam Barth
Comment 3 2012-06-18 12:05:30 PDT
Tony might have some thoughts about this patch.
Tony Chang
Comment 4 2012-06-18 13:45:32 PDT
Comment on attachment 148146 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=148146&action=review > LayoutTests/platform/chromium/TestExpectations:1448 > +BUGWK89367 LINUX : platform/chromium-linux/fast/text/international/complex-joining-using-gpos.html = IMAGE > +BUGWK89367 LINUX : fast/text/international/thai-line-breaks.html = IMAGE They pass on the waterfall, right? IIRC, the EWS bots run a different version of Ubuntu than the waterfall bots, so these tests only fail on the EWS bots. I think these are always going to fail until we switch to Precise in a year or so.
Tony Payne
Comment 5 2012-06-18 13:47:54 PDT
(In reply to comment #4) > (From update of attachment 148146 [details]) > View in context: https://bugs.webkit.org/attachment.cgi?id=148146&action=review > > > LayoutTests/platform/chromium/TestExpectations:1448 > > +BUGWK89367 LINUX : platform/chromium-linux/fast/text/international/complex-joining-using-gpos.html = IMAGE > > +BUGWK89367 LINUX : fast/text/international/thai-line-breaks.html = IMAGE > > They pass on the waterfall, right? IIRC, the EWS bots run a different version of Ubuntu than the waterfall bots, so these tests only fail on the EWS bots. I think these are always going to fail until we switch to Precise in a year or so. These failures really hurt EWS performance since they guarantee the tests will be run 3 times, and they make it impossible to use the CQ when there are red tests, even if the CL is fixing those red tests. I don't think this is a situation that's tenable for a year. It impacts developer productivity significantly.
Adam Barth
Comment 6 2012-06-18 13:49:40 PDT
That's not really accurate. The EWS remembers the set of failing tests from run to run and tends to ignore these tests anyway.
Tony Payne
Comment 7 2012-06-18 13:50:34 PDT
(In reply to comment #6) > That's not really accurate. The EWS remembers the set of failing tests from run to run and tends to ignore these tests anyway. That doesn't match my experience at all.
Note You need to log in before you can comment on or make changes to this bug.