[bug#70632] [PATCH 1/2] aux-files: comp-integrity: Adjust for newer emacs.

[Morgan Smith]

Liliana Marie Prikler <liliana.prikler@gmail.com> writes:

> Am Mittwoch, dem 01.05.2024 um 16:06 -0400 schrieb Morgan Smith:
> TL;DR: If we do a new file or a phase, we only rebuild emacs-next.  If
> we modify the file in-place, we rebuild emacs, because it uses it. 
> Between a phase and a new file, a new file is preferable, because we
> can then 'mv' it over the old one.

I apologize.  In my mind we where talking about two different changes at
the same time.  A change to emacs-next right now, and a change that
would rebuild all of emacs but live on a branch for now.

I still think that's the optimal route.  Apply the phase patch right now
to main, and then the file modification to some branch.

On a related note, I do have two changes I'd like to make when we
rebuild all of emacs.  I've attached the changes to this mail despite it
not being quite the right place to send them.  Also I didn't actually
test them because I don't really want to rebuild all of that myself.

>> In guix/build/emacs-utils.scm:emacs-generate-autoloads, there is a
>> condition to support emacs 28.  I don't think we ever use that path
>> anymore but it is nice to have a robust function that "just works".
>> Espiaclly back when we did have emacs 28 and 29 packages in guix.
> This is somewhat legacy code that has grown that way back when Emacs 29
> was emacs-next.  There was no good reason to drop it with the switch,
> but come Emacs 30, 31, and maintainability might be one.

My issue isn't that you're wrong, but rather I would argue "it's not a
big deal".  To maintain clean minimal code we should remove that
condition and so I see your point.  But also, it is such a small
addition of code that adds significant backwards compatibility to the
function.

0001-build-system-emacs-Allow-usage-of-allowed-disallowed.patch
Description: Text Data

0002-guix-emacs-utils-Be-more-verbose-in-build-phase.patch
Description: Text Data