[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[bug#74979] [PATCH 0/4] scripts: style: Sort more kinds of package
From: |
Ludovic Courtès |
Subject: |
[bug#74979] [PATCH 0/4] scripts: style: Sort more kinds of package |
Date: |
Mon, 23 Dec 2024 18:28:39 +0100 |
User-agent: |
Gnus/5.13 (Gnus v5.13) |
Hi Herman,
Herman Rimm <herman@rimm.ee> skribis:
> The warnings added in [PATCH 4/4] are emitted multiple times. How
> should I prevent that? Or should I put them behind a --verbose option?
Not sure, do you have an example on how to trigger it?
> +++ b/guix/scripts/style.scm
> @@ -508,9 +508,15 @@ (define (order-packages lst)
> (((or 'package 'package/inherit) fields ...)
> (let ((name (and=> (assoc-ref fields 'name) first))
> (version (and=> (assoc-ref fields 'version) first)))
> + (if (and name version)
> + (unless (and (string? name) (string? version))
> + (warning (G_ "non-string name/version for ~a~%") pkg))
> + (warning (G_ "package fields not found for ~a~%") pkg))
> (values name version)))
> - (_ (and (values #f #f)))))
> - (_ (and (values #f #f)))))
> + (_ (and (warning (G_ "package record missing for ~a~%") pkg)
> + (values #f #f)))))
> + (_ (and (info (G_ "not sorting top-level S-exp.: ~a~%") pkg)
> + (values #f #f)))))
You shouldn’t rely on the return value of ‘info’, ‘warning’, etc.:
they’re not specified (that’s generally the case for procedures called
for side effects only).
So I’d recommend:
(begin
(info …)
(values #f #f))
(You don’t even need ‘begin’ in this context.)
The other patches LGTM, though perhaps ‘tests/guix-style.sh’ could be
augmented a bit to cover the new cases?
Thanks,
Ludo’.