Re: Guix days guix home discussion

["Ludovic Courtès" <ludo@gnu.org>]

Hi!

Gábor Boskovits <gboskovits@gmail.com> skribis:

> 1. It is rare that guix home import does the right thing (it is usually
> removing some startup file customizations, does not seem to arrange to pick
> up profiles, not even its own). Either we should improve it, or document
> that it only gives a skeleton configuration and add some guidance on the
> steps needed to get a working one.

‘guix home import’ does exactly that: generate a first configuration
that you may find necessary to tweak.  Perhaps the manual should clarify
that?

I’m not sure what you mean though when you say it rarely does the right
thing.  Did you discuss examples?

> 2. The user default profile and the home package profile being separate is
> causing some issues. It might be enough to document all the special
> profiles somewhere (which as of now include at least system profile, user
> profile, pull profile and home profile), but we can also think about a bit
> more general solution, along the lines of a home service that ensures that
> a given profile matches the supplied manifest, and have variables for the
> special profiles. (These could then provide extensions to the shell
> services which could arrange to pick them up)

Likewise, would be nice to see if there are specific examples of
problems that people stumble upon.

> 3. Sometime on home reconfigure the shell prompt customizations seem to get
> lost. Sourcing the shell startup file fixes it. I will have to look into
> this more to file a proper bug report.
> 4. Creating a guix development environment service would be beneficial, to
> showcase the possibilities and to simplify onboarding. On top of there
> could be an additional service that adds emacs integration to this
> development environment.

Interesting, I like that.

(A colleague of mine has been working on a pre-configured Emacs
targeting developers using Guix and writing C/C++/Fortran and LaTeX:
<https://elementaryx.gitlabpages.inria.fr/>.)

> 5. There was a recommendation to relax the expectations on the home
> services merged. Right now a lot of people are just writing services for
> private use. Most probably such a single usecase service would already be
> beneficial to multiple people. The idea is the following: make it easy for
> an initial home service to be merged. (Example: do not ask to implement
> options that the submitter is not using). Then take care that if there is
> an addition to the service that it really gets merged with what we already
> have. This needs a bit of up front design, we have to make sure that the
> services can be extended while maintaining backwards compatibility.

I personally try to lower the barrier for Home services, but I think few
people (if any) beyond me review Home services.

We should expand the ‘home’ team; who’s in?

Thanks for the feedback,
Ludo’.