On Fri, Jul 04, 2025 at 12:27:10AM +0800, Hilton Chain wrote:
> Hi Guix,
> 
> In our manual, we link to the "ChangeLog" style[1], but in practice a 
> different
> convention is used.
> 
> Considering the following change (91bbed89b52eb64ee2388bf58be44eb5ae6a9dbb,
> found this when searching ‘if’ in the guix package):
> --8<---------------cut here---------------start------------->8---
> modified   gnu/packages/package-management.scm
> @@ -239,6 +239,12 @@ (define-public guix
>  $(prefix)/etc/openrc\n")))
>  
>                          (invoke "sh" "bootstrap")))
> +                    ,@(if (target-riscv64?)
> +                        `((add-after 'unpack 
> 'use-correct-guile-version-for-tests
> +                            (lambda _
> +                              (substitute* "tests/gexp.scm"
> +                                (("2\\.0") "3.0")))))
> +                        '())
>                      (add-before 'build 'use-host-compressors
>                        (lambda* (#:key inputs target #:allow-other-keys)
>                          (when target
> --8<---------------cut here---------------end--------------->8---
> 
> This is what we use:
> --8<---------------cut here---------------start------------->8---
> * gnu/packages/package-management.scm (guix)[arguments]: Add phase when
> building for riscv64-linux to adjust the test suite.
> --8<---------------cut here---------------end--------------->8---
> 
> 
> But according to GNU Coding Standards, the following might be used instead:
> --8<---------------cut here---------------start------------->8---
> * gnu/packages/package-management.scm (guix) <#:phases> [(target-riscv64?)]: 
> Use
>   correct Guile version for tests.   
> --8<---------------cut here---------------end--------------->8---
> 
> convention:
>   - *  changed file
>   - () changed function or variable
>   - [] conditional change
>   - <> indicating the part changed
> 
> [] is added after <> because the condition happens within that part.
> 
> Should this documented convention be followed instead, or we documenting the 
> one
> currently used?
> 
> Thanks
> 
> [1]: https://www.gnu.org/prep/standards/html_node/Change-Logs.html

Hey, that commit looks familiar! :)

I admit that I have never read the GNU Coding Standards closely, and for
commit messages I generally try to follow what others in a project are
using.  Mentally my model is:

section: (package|specific something): Very short description.

* location/to/file: (base function/package)[subheading]: Use your words
to describe the change, being specific and factual but not too verbose.

(and if you feel a need to add a reason for a change, it should be a
code comment).

who: git author(, Co-Authored-by:)
what: Add a phase ... adjust the test suite
where: location/to/file: (base function/package)[subheading]:
when: when building for riscv64-linux
why: (If it's not clear, this should be a code comment)
how: as seen in the code/diff


-- 
Efraim Flashner   <efr...@flashner.co.il>   אפרים פלשנר
GPG key = A28B F40C 3E55 1372 662D  14F7 41AA E7DC CA3D 8351
Confidentiality cannot be guaranteed on emails sent or received unencrypted

Attachment: signature.asc
Description: PGP signature

Reply via email to