Re: Fw: gpsd | Pipeline #174282730 has failed for release-3.21 | 87055e73

[James Browning]

On Tue, Aug 4, 2020 at 4:05 PM Bernd Zeimetz <bernd@bzed.de> wrote:
>
> On 8/4/20 10:14 PM, James Browning wrote:
> > On Tue, Aug 4, 2020, at 12:55 PM Gary E. Miller <gem@rellim.com> wrote:
> >>
> >> Yo All!
> >>
> >> Yesterday the rpm CI worked.  Today it is broken.
> >>
> >> Anyone know how to fix what RedHat broke last night?
> >
> > see merge request 173. It has been broken for a while. It does not
> > run on every commit. Instead, I _think_ it runs twice a week. Bernd
> > set up the scheduling. I set up the CI job. I have been useless.
>
> Uh, that wasn't me:
>
> 16f3ad091c (James Browning  2020-01-28 10:09:49 -0800 325)   only:
> 16f3ad091c (James Browning  2020-01-28 10:09:49 -0800 326)     - schedules
> 16f3ad091c (James Browning  2020-01-28 10:09:49 -0800 327)     - tags
> 16f3ad091c (James Browning  2020-01-28 10:09:49 -0800 328)     - web
>
>
> Its broken since weeks.
> Basically its telling you, that on Fedora /usr/bin/python might either
> still point to python2 or to python3. So you are forced to write the
> appropriate python version into the shebangs. Scontruct has an option
> for that, it should be used.

Right the other guy named Bernd set up the kicker @
https://gitlab.com/gpsd/gpsd/-/pipeline_schedules
and *I* set up all the broken (s)crap in .gitlab-ci.yml

Today, I guess it got triggered by tagging 3.21 as the
schedule probably does not align, and AFAICT no one
knows how to use the web trigger. There were other
whack-a-mole issues that I _think_ I have cleared up.