unofficial mirror of guix-patches@gnu.org 
 help / color / mirror / code / Atom feed
From: "Ludovic Courtès" <ludo@gnu.org>
To: Hilton Chain <hako@ultrarare.space>
Cc: Josselin Poiret <dev@jpoiret.xyz>,
	Simon Tournier <zimon.toutoune@gmail.com>,
	Mathieu Othacehe <othacehe@gnu.org>,
	Tobias Geerinckx-Rice <me@tobias.gr>,
	Ricardo Wurmus <rekado@elephly.net>,
	65062@debbugs.gnu.org, Christopher Baines <guix@cbaines.net>
Subject: [bug#65062] [PATCH core-updates] packages: Lookup inputs by specification.
Date: Wed, 20 Dec 2023 22:26:51 +0100	[thread overview]
Message-ID: <87msu4ftus.fsf_-_@gnu.org> (raw)
In-Reply-To: <dac85ddb4c5637ab522b4b37f926f00af567cc33.1696323536.git.hako@ultrarare.space> (Hilton Chain's message of "Tue, 3 Oct 2023 17:17:02 +0800")

Hi,

Hilton Chain <hako@ultrarare.space> skribis:

> * guix/packages.scm (specification->inputs): New procedure.
> (lookup-input,replace-input): Use it.
> (delete-input): New procedure.
> (modify-inputs)[delete]: Use it.

I’ve been thinking about this change lately.

The problem we have now is that it looks like input labels are gone, but
they’re not; in particular ‘modify-inputs’ preserves labels, which is a
source of confusion.  For instance, if you do:

  (modify-inputs x
    (replace "openmpi" mpich))

then ‘mpich’ remains associated with the “openmpi” label.  Ugh.

So I sympathize with the goal.  I think we can do something simpler
though:

>  (define (lookup-input inputs name)
>    "Lookup NAME among INPUTS, an input list."
>    ;; Note: Currently INPUTS is assumed to be an input list that contains input
>    ;; labels.  In the future, input labels will be gone and this procedure will
>    ;; check package names.
> -  (match (assoc-ref inputs name)
> -    ((obj) obj)
> -    ((obj _) obj)
> -    (#f #f)))
> +  (let ((candidates (specification->inputs name inputs)))
> +    (and (not (null? candidates))
> +         (second (first candidates)))))

How about:

  (find (match-lambda
          ((_ (? package? package) . _)
           (string=? (package-name package) name))
          (_ #f))
        inputs)

?

That way, ‘lookup-input’ would honor package names and ignore labels.

> +(define (delete-input name inputs)
> +  "Delete input NAME within INPUTS."
> +  (let ((to-delete (specification->inputs name inputs)))
> +    (lset-difference equal? inputs to-delete)))

And we do something similar here.

Thus, no need to fiddle with specifications.

How does that sound?

Now, I think this is the way forward, but I also think it’s going to
break many packages and workflows (‘--with-input’…).  So it should go
hand in hand with an effort to fully remove labels in Guix.

Thanks,
Ludo’.




      reply	other threads:[~2023-12-20 21:28 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-08-05  2:50 [bug#65062] [PATCH core-updates 0/1] Specify output in input label when it's not "out" Hilton Chain via Guix-patches via
2023-08-05  2:53 ` [bug#65062] [PATCH core-updates 1/1] packages: " Hilton Chain via Guix-patches via
2023-08-22 16:00   ` Ludovic Courtès
2023-08-24  3:42     ` Hilton Chain via Guix-patches via
2023-08-25 11:10       ` Josselin Poiret via Guix-patches via
2023-09-08 22:03       ` Ludovic Courtès
2023-10-03  9:13         ` Hilton Chain via Guix-patches via
2023-08-05  3:01 ` [bug#65062] [PATCH core-updates 0/1] " Hilton Chain via Guix-patches via
2023-08-05  3:19   ` Hilton Chain via Guix-patches via
2023-10-03  9:15 ` [bug#65062] [PATCH v2 core-updates 0/2] packages: Lookup inputs by specification Hilton Chain via Guix-patches via
2023-10-03  9:17   ` [bug#65062] [PATCH v2 core-updates 1/2] ui: package-specification->name+version+output: Move to (guix packages) Hilton Chain via Guix-patches via
2023-10-03  9:17   ` [bug#65062] [PATCH v2 core-updates 2/2] packages: Lookup inputs by specification Hilton Chain via Guix-patches via
2023-12-20 21:26     ` Ludovic Courtès [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

  List information: https://guix.gnu.org/

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=87msu4ftus.fsf_-_@gnu.org \
    --to=ludo@gnu.org \
    --cc=65062@debbugs.gnu.org \
    --cc=dev@jpoiret.xyz \
    --cc=guix@cbaines.net \
    --cc=hako@ultrarare.space \
    --cc=me@tobias.gr \
    --cc=othacehe@gnu.org \
    --cc=rekado@elephly.net \
    --cc=zimon.toutoune@gmail.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
Code repositories for project(s) associated with this public inbox

	https://git.savannah.gnu.org/cgit/guix.git

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).