[bug#41790] [PATCH] Update emacs-direnv

[Katherine Cox-Buday <cox.katherine.e@gmail.com>]

Nicolas Goaziou <mail@nicolasgoaziou.fr> writes:

> Katherine Cox-Buday <cox.katherine.e@gmail.com> writes:
>
>> I disagree. If propagated inputs are not for this -- making the package
>> even functional -- what are they for?
>
> IIUC, they are to be used as a last resort, e.g., when the package
> cannot possibly build without them.

I checked the Guix manual for the intended use of propagated inputs
since I didn't completely understand how what is in the runtime
environment would affect the build environment. I found something which
maybe is open to interpretation:

  "To ensure that libraries written in those languages can find library
  code they depend on at run time, run-time dependencies must be listed
  in propagated-inputs rather than inputs."

Maybe a binary required to operate is no different than "library code
they depend on at run time"?

> Reason is propagated inputs pollute user's profile. E.g., someone may
> want to use a different direnv, and this propagated input would
> conflict with their package.

That's another good point I hadn't considered: what if for some reason a
user wants a different version of the tool (presumably provided by Guix
as well)?

> I'll let maintainers decide about this, and hopefully clarify what can
> and cannot be propagated, at least in Emacs packages. This will be
> useful feedback for future reviews.

That's a good idea. I have an opinion, but it is not fully informed. I
appreciate the review, conversation, and appeal to authority!

> Meanwhile, you can still provide a patch only bumping emacs-direnv. It
> is also fine in you prefer to wait.

I'll let this one sit. The version bump was a side-effect of me
believing the package should also install the tool.

-- 
Katherine