Re: Finding cause of test fails on the cfbot site

On Sat, Feb 20, 2021 at 11:33 AM Thomas Munro <thomas(dot)munro(at)gmail(dot)com> wrote:
> On Sat, Feb 20, 2021 at 10:31 AM Peter Smith <smithpb2250(at)gmail(dot)com> wrote:
> > Here is another related question about the cfbot error reporting -
> >
> > The main cfbot "status page" [1] still shows a couple of fails for the
> > 32/2914 (for freebsd & linux). But looking more closely, those fails
> > are not from the latest run. e.g. I also found this execution
> > "history" page [2] for our patch which shows the most recent run was
> > ok for commit a7e715.

> Hmmph. It seems like there is indeed some kind of occasional glitch,
> possible a bug in my code for picking up statuses from Cirrus through
> their GraphQL API (that's a recent thing; we had to change providers
> due to another CI's policy changes in January, and apparently
> something in the new pipeline isn't quite fully baked). Will look
> into that this weekend. Sorry about that, and thanks for letting me
> know.

Should be fixed now. (Embarrassing detail: I used flock(LOCK_EX) to
prevent two copies of a cron job from running at the same time, but
Python helpfully garbage collected and closed the fd, releasing the
lock, Rarely, the script would run for long enough to have some state
clobbered by the next run, so you wouldn't see the latest result.)