Hello Ludo,

I hadn't found that post, but what I'm trying to get across holds.
Barring distribution of binaries since I don't think that is critical
for improving and incremental adoption of Guix to support
infrastructure. A `manifest.scm` could look like the following,
keeping a close eye on the line containing "(remote-repo-dep URI
GIT-REF)".

```
(use-modules (guix)
             (guix build-system gnu)
             ((guix licenses) #:prefix license:)
             (gnu packages autotools)
             (gnu packages ...)
             (gnu remote-repo-dep-mod)
             (gnu packages ...)
             (gnu packages version-control))

(package
  (name "your-project-name")
  (version "???")
  (source #t)
  (build-system ???)
  (native-inputs
   (append (list autoconf
                 ...
                 (remote-repo-dep URI GIT-REF)
                 ...
                 pkg-config)
  (inputs
   (list libffi bash-minimal))
  (propagated-inputs
   (list libunistring libgc))

  (native-search-paths
   (list (search-path-specification
          (variable "GUILE_LOAD_PATH")
          (files '("share/guile/site/3.0")))
         (search-path-specification
          (variable "GUILE_LOAD_COMPILED_PATH")
          (files '("lib/guile/3.0/site-ccache")))))
  (synopsis "...")
  (description "...")
  (home-page "...")
  (license license:lgpl3+))
```

Now, having that single line as a way to include code as opposed to
what can be accomplished similarly with similar expressiveness for the
purpose of the target user group of researchers (think mechanical
engineers, biologists, electrical labs, chemists) and general
developers would lower the barrier of entry to use Guix and improve
the quality of their projects to the point where it just wouldn't make
sense not to.  Now, since there is such a well developed and
documented workflow for a similar process, would I be wrong in
assuming that such functionality could be relatively easy to
implement?

Let me know if gmail is messing up formatting.  I think something is off.

On Wed, Nov 20, 2024 at 4:45 PM Ludovic Courtès <l...@gnu.org> wrote:
>
> Josh Marshall <joshua.r.marshall.1...@gmail.com> skribis:
>
> > This is a technical workaround, not direct support.  If such a feature
> > is to be aimed at general developers and researchers, it can't be so
> > circuitous. The Invocation from any form of Guix installation needs to
> > be `guix shell` and just work.
>
> To make sure we’re talking about the same thing, Efraim is referring to
> the conventions documented at:
>
>   https://guix.gnu.org/cookbook/en/html_node/Software-Development.html
>
> Ludo’.

Reply via email to