Please review the draft for April's report
James Addison
jay at jp-hosting.net
Thu May 7 18:15:44 UTC 2026
Hi Manuel and RB folks,
On Thu, 7 May 2026 at 17:04, Manuel Jacob <me at manueljacob.de> wrote:
> [ ... snip ... ]
> I am a bit confused that
> https://reproducible-builds.org/reports/2026-04/?draft#upstream-patches
> mostly contains what I would call downstream patches to the Debian
> package. For most of the patches, I couldn’t find that they were
> submitted upstream. What about renaming the section to just “Patches”?
> [ ... snip ... ]
I'd be curious whether we could gather any statistics about how often RB
patches are upstreamed; it's usually the objective, and although I don't
know for sure about other distros, Debian does have patch metadata[1] that
can indicate whether patches are suitable for forwarding further upstream.
(Debian itself is used as a base for other operating systems -- so I would
tend to think of there being relative levels of upstreamness; Debian packages
are usually maintained separately from the source repository though, I admit).
NB: Manuel asked me off-list back in March about a binutils patch I'd
submitted[2][3], a case where I _did not_ push for upstreaming (it's
build-time file permissions, and I felt that those and the core utils in
use across different OSes/distros might make it unsuitable).
I don't know for certain whether that's the motivating factor behind
trying to understand the terminology and intent of the patch-upstreaming
process and documentation, but I mention it in case it helps as a case
to refer to.
Personally I lean slightly towards keeping the word upstreaming in place,
because I think it's the ideal objective; we try to get the patches upstream,
and/or obsolete the need for them by encouraging better software development
and toolchain practices. If a patch does have to be retained locally by a
distro/OS for a while... well, reluctantly that's acceptable, but eventually we
want to avoid that.
(this can be true for package bugreports that are about problems unrelated to
build reproducibility, too)
Sorry for a long answer to a short question.
Regards,
James
[1] - https://dep-team.pages.debian.net/deps/dep3/
[2] - https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1090395
[3] - https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1092870
More information about the rb-general
mailing list