New tests will come, bugfixes will happen, we should stay up-to-date with them. Given we're just rerunning a script, I don't think there's much value in getting a review every single time here -- at least, when rerunning is all that's happening, and we don't have to add new skip entries and such. How about we have a blanket rubber-stamp for a JS engine hacker to land the results of running the script against the canonical test262 repository, whenever doing so doesn't change the skipped-test list? Seems like the cases where the skip-set does change could maybe use an eye, but the rest of the time it seems wasteful.
I thought we already had a blanket rs= for non-breaking test additions?
Not sure. I don't remember one now, but I have a vague recollection about tests being different somehow, at one time. Maybe it was tests not needing approval when tree checkins are restricted? Yeah, I think it might have been that, and that's hardly relevant any more. From IRC: <terrence> Waldo: well, either way, for these tests I think automatic inclusion is the right call So I'll give the script a run every so often to keep things reasonably up-to-date, then, and only poke for reviews when there's an actual non-automated change to make.
Whiteboard: [leave open]
Summary: Periodically rerun update-test262.sh to keep our imported test262 copy up-to-date → Periodically rerun test262-update.py to keep our imported test262 copy up-to-date
You need to log in before you can comment on or make changes to this bug.