As part of enabling limited EWS bots to check new tests, webkit-patch should have the ability to run new tests and report results. This could be done by adding an option to webkit-patch build-and-test to iterate a number of times on new tests and make a report in case of failures.
Created attachment 249485 [details] Patch
Created attachment 249585 [details] WIP
(In reply to comment #2) > Created attachment 249585 [details] > WIP Another try after 143090. Plan is to add a unit test for the new_or_modified_potential_tests function. Any feedback appreciated before doing that extra step.
Created attachment 249749 [details] WIP2
Created attachment 249811 [details] Patch
Created attachment 268450 [details] Rebasing
(In reply to comment #6) > Created attachment 268450 [details] > Rebasing Introduced option (iterate_on_new_tests) may allow gtk/efl/win EWS bots to check new/modified tests at a hopefully small cost. This option may also be useful in general to check for test flakiness.
Comment on attachment 268450 [details] Rebasing View in context: https://bugs.webkit.org/attachment.cgi?id=268450&action=review > Tools/ChangeLog:10 > + Computationf of new or modified tests is doen by LayoutTestFinder.find_touched_tests Computationf -> Computation doen -> done
Created attachment 268531 [details] Patch for landing
Thanks for the review. > > Tools/ChangeLog:10 > > + Computationf of new or modified tests is doen by LayoutTestFinder.find_touched_tests > > Computationf -> Computation > > doen -> done Doen... hum done ;)
Comment on attachment 268531 [details] Patch for landing Clearing flags on attachment: 268531 Committed r194761: <http://trac.webkit.org/changeset/194761>
All reviewed patches have been landed. Closing bug.