https://bz.apache.org/SpamAssassin/show_bug.cgi?id=8056
--- Comment #12 from Sidney Markowitz <sid...@sidney.com> --- (In reply to Henrik Krohns from comment #11) > So this is only for individual committers, to run an "extended" make test > manually on many platforms when needed? That's the basic idea. It's something that I could use when I am preparing a release. It is set up for manual trigger only, and not to be run from our Apache GitHub repo because Apache has limited resources for such things, that are already oversubscribed by the early adopter projects. > Or is the intention to automate this in any way? I guess it wouldn't make > sense to automate anything across many committer accounts, since they would > be just running exact same tests wasting Github resources? I don't think it would be reasonable to automate the tests to run with every push, even if it were easy to filter out the rule update pushes. > Or is there some > resource limits on Github side, and we could coordinate running different > Perl versions etc with different committer accounts? I haven't hit a resource limit on my free personal account. There might be some kind of cap on network usage that I hit occasionally when running many jobs at once that is the cause of some network tests failing very sporadically. Actually, your mention of splitting up perl versions between people gave me an idea. Github Actions can have a cron trigger, We could run one perl version on three platforms each day, which would just be three make test runs per day and cycle through all perl versions every two weeks. I could do that on my personal account with no problem. -- You are receiving this mail because: You are the assignee for the bug.
