Re: On elisp running native

[Michael Albinus]

Andrea Corallo <address@hidden> writes:

Hi,

>> Hmm. Maybe we shall run this only when no other job runs in parallel?
>> Resource groups <https://docs.gitlab.com/ee/ci/yaml/#resource_group>
>> seem to do this.
>
> I think this would fix the memory issue but can we afford to keep the CI
> exclusively busy for so long?  I fear speed 2 would take something close
> to one day in this setup.

Understood.

> I propose to keep speed 0 there, is also useful because I always
> bootstrap but only speed 2.  We could revisit the setup when we have a
> more affordable compile time for bootstrapping speed 2 also on small
> machines.  What do you think?

OK. In this case, we even don't need a schedule I believe. Let it run as
normal pipeline. Maybe you just uncomment the schedule for speed 0; for
the other speeds we still might need it in the future.

The timeout is still needed (with a shorter value) I guess.

> Regards
>
>   Andrea
>
> PS I think ideally we should check all speed from 0 too 2 when/if will
> be possible.

Yes. That's what a CI is good for :-)

Best regards, Michael.