I also like this idea. I particularly like the side effect of making it clear when a patch is modifying a test result that is itself not necessarily up to date.
I suggest a policy that says you must rebaseline a test before committing a patch that would change that test (if the test is marked as needing rebaselining). Otherwise I see the potential for situations in which the status of a test gets confused due to overlapping patches or roll-outs. That also means we need an easy way to determine if a patch affects tests that are not reporting errors due to rebaselining. Can we have new-run-webkit-tests show, in the test results summary, the tests that you failed that need rebaselining? Cheers, Stephen. On Thu, Oct 4, 2012 at 6:18 PM, Emil A Eklund <[email protected]> wrote: > That sounds like a great idea! > It is too hard to do this right today and having an easy way to > indicate that a test needs to be rebaselined across some or a subset > of platforms would be great. > > -- > Emil > _______________________________________________ > webkit-dev mailing list > [email protected] > http://lists.webkit.org/mailman/listinfo/webkit-dev >
_______________________________________________ webkit-dev mailing list [email protected] http://lists.webkit.org/mailman/listinfo/webkit-dev
