Nice. Thank you!

On Mon, Dec 19, 2022, 19:42 Guillaume Le Vaillant <glv@posteo.net> wrote:

> Roman Scherer <roman.scherer@burningswell.com> skribis:
>
> >
> > Hi Guillaume,
> >
> > thanks for looking at this patch series and sorry for the trouble. I
> > just rebuild the patch series on top of the current master 3 times with
> > the following command:
> >
> > ```
> > ./pre-inst-env guix build --check --no-grafts --no-substitutes
> --rounds=3 cdo
> > ```
> >
> > I see some tests are skipped, but the whole thing passed every time. I
> > also don't remember seeing them failing a single time.
> >
> > I'm running this on the following system:
> >
> > ```
> > Linux precision 5.15.0-56-generic #62-Ubuntu SMP Tue Nov 22 19:54:14 UTC
> 2022 x86_64 x86_64 x86_64 GNU/Linux
> > ```
> >
> > And here are the logs of one of the test runs:
> >
> > [...]
> >
> > On what kind of system is this failing for you? What can we do?
>
> On my main machine, with an AMD Ryzen 9 CPU with 32 threads, it failed
> every time. On another machine, with an Intel I9 CPU with 16 threads, it
> succeeded every time. It turns out that some tests can fail if the
> machine has many threads, and adding '#:parallel-tests? #f' to the
> 'arguments' field fixes the issue.
>
> I also changed the 'source' field of the libaec package to use
> 'git-fetch' instead of 'url-fetch', because archive files created
> automatically by forges like gitlab can sometimes change in place and
> have a different hash.
>
> Patches pushed as d03b6fb0e3f0b81b35a9b35b89c213c144c59fe6 and
> following. Thanks.
>