Should patch view contain cgit version information?

Should patch view contain cgit version information?

From: Leah Neukirchen
Hi,

for packaging purposes we sometime fetch patches from cgit,
and store their sha hashes to ensure the patches are fetched unmodified.

However, cgit-pink (and cgit) appends its version number at the end of
a patch, which changes the output when cgit-pink is updated.

Other Git hosting sites like Github, Gitlab and Gitea don't append the
version, and I think it doesn't serve much purpose, so I suggest to
drop it.

Thoughts?

cu,
-- 
Leah Neukirchen  <leah@vuxu.org>  https://leahneukirchen.org/

Re: Should patch view contain cgit version information?

From: Klemens Nanni
05.01.2023 22:22, Leah Neukirchen пишет:
> Hi,
> 
> for packaging purposes we sometime fetch patches from cgit,
> and store their sha hashes to ensure the patches are fetched unmodified.

Same goes for tarballs obviously, were certain sites are already a pain
to work with since they either provide tarballs that change on every
download or there simply is no guarantee for autogenerated ones to be
stable due to whatever reason they have.

> However, cgit-pink (and cgit) appends its version number at the end of
> a patch, which changes the output when cgit-pink is updated.

If that is the case, then I'd appreciate such version to disappear.

> Other Git hosting sites like Github, Gitlab and Gitea don't append the
> version, and I think it doesn't serve much purpose, so I suggest to
> drop it.

As ports/package maintainer, I do this all the time with GitHub and
regenerating checksums due to platform changes (while code remains
identical) would break this workflow and/or force me to save them as
local copies.

> 
> Thoughts?
> 
> cu,

Re: Should patch view contain cgit version information?

From: Leah Neukirchen
Klemens Nanni <kn@openbsd.org> writes:

> 05.01.2023 22:22, Leah Neukirchen пишет:
>> Hi,
>> 
>> for packaging purposes we sometime fetch patches from cgit,
>> and store their sha hashes to ensure the patches are fetched unmodified.
>
> Same goes for tarballs obviously, were certain sites are already a pain
> to work with since they either provide tarballs that change on every
> download or there simply is no guarantee for autogenerated ones to be
> stable due to whatever reason they have.

True, but I think everything (including cgit) uses git archive now,
which has been mostly stable (with tar + GNU gzip) for the last years.

>> However, cgit-pink (and cgit) appends its version number at the end of
>> a patch, which changes the output when cgit-pink is updated.
>
> If that is the case, then I'd appreciate such version to disappear.
>
>> Other Git hosting sites like Github, Gitlab and Gitea don't append the
>> version, and I think it doesn't serve much purpose, so I suggest to
>> drop it.
>
> As ports/package maintainer, I do this all the time with GitHub and
> regenerating checksums due to platform changes (while code remains
> identical) would break this workflow and/or force me to save them as
> local copies.

Full ack.

-- 
Leah Neukirchen  <leah@vuxu.org>  https://leahneukirchen.org