Bug#1050592: perl: F_GETLK / F_GETLK64 confusion on ppc64el breaking libfile-fcntllock-perl
- To: Niko Tyni <ntyni@debian.org>
- Cc: 1050592@bugs.debian.org
- Subject: Bug#1050592: perl: F_GETLK / F_GETLK64 confusion on ppc64el breaking libfile-fcntllock-perl
- From: Aurelien Jarno <aurel32@debian.org>
- Date: Sun, 3 Sep 2023 18:04:11 +0200
- Message-id: <[🔎] ZPSue7DC0USkx+iJ@aurel32.net>
- Reply-to: Aurelien Jarno <aurel32@debian.org>, 1050592@bugs.debian.org
- In-reply-to: <ZPScvDhZnaBgp7fk@birgitta.local.invalid>
- References: <ZOo/agH5jjCZ4yE4@birgitta.local.invalid> <ZOs285VXRafcEfZ9@birgitta.local.invalid> <ZOz71cDoyMxBqh2Y@aurel32.net> <ZOo/agH5jjCZ4yE4@birgitta.local.invalid> <ZO0Vpi1BbBp2A4hT@aurel32.net> <ZPScvDhZnaBgp7fk@birgitta.local.invalid> <ZOo/agH5jjCZ4yE4@birgitta.local.invalid>
On 2023-09-03 17:48, Niko Tyni wrote:
> Control: reassign -1 libc6-dev 2.37-2
> Control: found -1 2.36-9+deb12u1
> Control: tag -1 bookworm
>
> On Mon, Aug 28, 2023 at 11:46:14PM +0200, Aurelien Jarno wrote:
>
> > > I think it's an ABI breakage that should be fixed, but just reverting
> > > the patch will break the case without -D_FILE_OFFSET_BITS=64. I'll check
> > > with upstream and try to get the issue fixed in both testing/sid and
> > > stable. I'll keep you updated. In the meantime feel free to reassign the
> > > bug to the glibc.
> >
> > I have opened a bug upstream:
> > https://sourceware.org/bugzilla/show_bug.cgi?id=30804
> >
> > And submitted a possible patch:
> > https://sourceware.org/pipermail/libc-alpha/2023-August/151199.html
>
> Many thanks! I'm reassigning this. Hope I got the versions right.
>
> Looks like the discussion upstream has quieted down now.
I have submitted the v3 of the patch and I am waiting for it to be
rebuilt:
https://sourceware.org/pipermail/libc-alpha/2023-August/151273.html
Unfortunately upstream does not consider it as an ABI breakage on the
glibc side, just an API breakage. But it causes an ABI breakage on the
perl side.
> My understanding is that for sid/trixie we'll just need a binNMU of perl
> on ppc64el afterwards. I'll request that once glibc has the fix.
I believe so (well probably on all architectures for multiarch sync).
> For stable I don't think anything needs to be done on the perl side.
> Both perl and libfile-fcntllock-perl were build with the old constants
> before the regression. So just fixing glibc should be enough.
I agree.
--
Aurelien Jarno GPG: 4096R/1DDD8C9B
aurelien@aurel32.net http://aurel32.net
Reply to: