Thanks for the explanation, Ross. Sounds like it should be mostly smooth-sailing so I'm happy to give it a try and see how it goes.
On Wed, Nov 11, 2015 at 2:53 PM, 'Ross McIlroy' via v8-dev < [email protected]> wrote: > On 11 November 2015 at 18:23, Adam Klein <[email protected]> wrote: > >> On Wed, Nov 11, 2015 at 9:03 AM, Ross McIlroy <[email protected]> >> wrote: >> >>> Recently we have started running a subset of the mjsunit and test262 >>> tests in the Ignition interpreter on the V8 test bots. In order to keep the >>> waterfall green, we are going to start running these tests on the commit >>> queue (CQ) as well (once https://codereview.chromium.org/1440673002/ >>> lands). >>> >>> Since Ignition doesn't support all JavaScript features yet, you may see >>> CQ failures on these tests for your own CLs if you add a new mjsunit test >>> or modify V8 such that existing test runs differently. If you see a >>> failure for a test in MjsunitIgnition or Test262Ignition, just add a line >>> to the ignition section of test/mjsunit/mjsunit.status or >>> test/test262/test262.status to skip the test on ignition, and let me know. >>> >>> Please let me know if you have any problems with these tests or the >>> process of if adding these skips becomes too burdensome. >>> >> >> If the only action when the CQ bots fail is going to be to add lines to >> the status file, why is this approach more useful than just letting the bot >> go red on failures (and leave it out of the CQ)? >> > > The problem is it's not an independent bot - for resource reasons it is an > additional step on all the existing bots, therefore if it goes red, the > main waterfall goes red (and LKGRs and Chromium rolls stop). > > >> It seems like in either strategy, manual intervention from the Ignition >> team will be necessary to further-triage the failure, and the plan proposed >> by Ross is strictly more work. >> > > This shouldn't be much work - I've tried to make sure all ES6 feature > tests are wildcard blacklisted, so I'm hoping there should be very few > (hopefully no) situations where any additional test skipping is necessary - > the tests have been running on the waterfall for a couple of days and, > other than a few initial flakes which we've now suppressed, have been > stable. This PSA is meant as a heads up - I'm not thinking this should > require any changes to day-to-day business, if it does let me know and we > can try something else. > > > - Adam >> >> >>> >>> Cheers, >>> Ross >>> >>> -- >>> -- >>> v8-dev mailing list >>> [email protected] >>> http://groups.google.com/group/v8-dev >>> --- >>> You received this message because you are subscribed to the Google >>> Groups "v8-dev" group. >>> To unsubscribe from this group and stop receiving emails from it, send >>> an email to [email protected]. >>> For more options, visit https://groups.google.com/d/optout. >>> >> >> -- >> -- >> v8-dev mailing list >> [email protected] >> http://groups.google.com/group/v8-dev >> --- >> You received this message because you are subscribed to the Google Groups >> "v8-dev" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to [email protected]. >> For more options, visit https://groups.google.com/d/optout. >> > > -- > -- > v8-dev mailing list > [email protected] > http://groups.google.com/group/v8-dev > --- > You received this message because you are subscribed to the Google Groups > "v8-dev" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > For more options, visit https://groups.google.com/d/optout. > -- -- v8-dev mailing list [email protected] http://groups.google.com/group/v8-dev --- You received this message because you are subscribed to the Google Groups "v8-dev" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
