Re: Graft hooks

[Ludovic Courtès]

Hi Mark,

Mark H Weaver <address@hidden> skribis:

> address@hidden (Ludovic Courtès) writes:
>
>> Since this is used when grafting Racket, I would suggest moving this
>> graft to the “build side” entirely, similar to what I did in
>> <https://debbugs.gnu.org/cgi/bugreport.cgi?bug=19973#25>.  Probably
>> you’d just add a single procedure to (guix build graft) and add it to
>> %graft-hooks.
>>
>> That procedure could be the same as what you have above, except that
>> it’d run OUT/bin/raco, if it exists, and do nothing if OUT/bin/raco does
>> not exist.
>>
>> WDYT?
>
> I think it would be quite unfortunate if _every_ graft had to run
> _every_ graft hook, or if every graft hook had to be defined in
> (guix grafts) and/or (guix build graft).
>
> It's reasonable to have a few global graft hooks, e.g. for handling
> debugging information, but I would greatly prefer for Guix to also have
> a mechanism allowing individual packages or build systems to introduce
> graft hooks without modifying (guix grafts) or (guix build graft), and
> for such a mechanism to be used for Racket and its libraries.

Sure, like I wrote, a more extensible solution along the lines of what
Timothy and you suggest sounds better long-term.  It just happens to be
harder to implement today (in particular until ‘wip-build-systems-gexp’
is merged), and not entirely clear how this could work, as discussed
with Timothy.

That’s why I’m proposing the simple approach for now.  I think we can
revisit the issue if/when we have graft hooks tied to a particular
package or build system.

Note that the .gnu_debuglink hook in <https://bugs.gnu.org/19973> runs
only when there’s a “debug” output, and the Racket hook that Timothy
proposes would only run when Racket is available.  So you get extra
run-time overhead only if there’s something to do.

HTH,
Ludo’.