Re: Should we delete pkg-go-tools, provisioning, migrate-pkg-go-to-salsa etc?
Hi!
I didn't get any responses to this email last month. I guess people
active on this list don't have knowledge/opinion about these?
Martina, as you previously expressed you represent the decision makers
in the Go team, could you state your preference/decisions on the
questions below?
Having a bunch of outdated cruft adds unnecessary toil and distraction
to people trying to maintain current packages, so cleaning away old
stuff can help make current contributors more active. For example
Santiago spent today time debugging the Go CI job test_the_archive,
which in fact seems useless and should be deleted and replaced with
Salsa CI.
On Mon, 28 Apr 2025 at 10:50, Otto Kekäläinen <otto@debian.org> wrote:
>
> Hi,
>
> What is the current maintenance plan for these?
>
> https://salsa.debian.org/go-team/infra/pkg-go-tools
> https://salsa.debian.org/go-team/infra/provisioning
> https://salsa.debian.org/go-team/migrate-pkg-go-to-salsa
> https://salsa.debian.org/go-team/stapelbergtest
>
> Should we plan to stop using them and gracefully deprecate them, mark
> them archived or delete them?
>
> I have not seen the CI job test_the_archive be useful in any of the
> packages I worked on. I am not fully sure if I even understand what
> the intent of that CI job, as it is doing only a source package build
> and otherwise testing other packages than what actually triggered the
> CI run.
>
> See e.g. https://salsa.debian.org/go-team/packages/usql/-/jobs/7283517
> failing on various go commands but job passes anyway:
> go: errors parsing go.mod:
> /builds/go-team/packages/usql/go.mod:5: unknown directive: toolchain
>
> The only way to get the test_the_archive job to fail is to give inputs
> that don't even produce a source package, e.g. vs
> https://salsa.debian.org/go-team/packages/usql/-/jobs/7283480
>
>
> I am sure these Go team infra pieces were useful 2-3 years ago, but
> going forward, does it makes sense to spend energy on maintaining
> these?
Reply to: