[bug#49456] [PATCH] gnu: add environment-modules

[Ivan Gankevich]

> > +++ b/gnu/packages/parallel.scm
>
> How about ‘package-management.scm’ instead?
>
> You probably need to add a copyright line for you too.

Moved to ‘package-management.scm’, added copyright line.


> > +(define-public environment-modules
> > +  (package
> > +    (name "environment-modules")
>
> Should the package name be “modules”, since that’s the name that
> upstream seems to be using?

Renamed to “modules”.


> > +          (add-after 'configure 'patch-/bin/sh-in-tests
> > +            (lambda _
> > +              (for-each
> > +                (lambda (file)
> > +                  (substitute* file
> > +                    (("/bin/sh") (which "bash"))
> > +                    ;; For some reason "kvm" group cannot be resolved for
> > +                    ;; "nixbld" user. We remove "-n" switch here to not
> > +                    ;; resolve the groups at all.
> > +                    (("exec id -G -n -z") "exec id -G -z")
> > +                    (("exec id -G -n") "exec id -G")
>
> Is this change made for tests?  In the build environment, the build user
> is potentially in the “kvm” group if it exists, but indeed, /etc/group
> lacks “kvm” (see nix/libstore/build.cc:1777).
>
> Should a post-check phase reinstate ‘-n’?

This change is needed for tests only, main programme uses different
configuration.

I have updated to the version 4.8.0 and unfortunately these changes no longer
work (developers replaced calls to “id” with Tcl extensions).  Now I disabled
tests that use group information.

Can we add all supplementary groups to /etc/groups? Not adding them to
/etc/group makes some shell commands return an error (“groups”, “id -G -n”).


> > +    (synopsis "Shell environment variables and aliases management")
> > +    (description "A tool that simplify shell initialization and lets users
> > +easily modify their environment during the session with modulefiles.")
>
> Please write full sentences for the description.

Changed description.


> Could you send an updated patch?
>
> Bonus points if you can provide a commit log that follows our
> conventions.  :-)

I’ve sent an updated patch in a separate email. Thank you for the corrections!