I realized that message has been forwarded to *all* buildd maintainers, and I answered there. So let me forward the answer to the bug report.
----- Message transféré de Aurelien Jarno <aure...@debian.org> ----- From: Aurelien Jarno <aure...@debian.org> To: Thorsten Glaser <t.gla...@tarent.de> Cc: buildd-maintain...@buildd.debian.org User-Agent: Mutt/2.2.9 (2022-11-12) Date: Sun, 2 Jul 2023 23:02:38 +0200 Subject: Re: Fwd: (buildd chroot bug) Re: Bug#1040167: openjdk-8-jre-headless: version 8u382~b04-1 depends on libjpeg8 which is not in Debian Message-ID: <zkhl7ie6+4q4r...@aurel32.net> [ tl;dr for buildd-maintainers: nothing to do. ] Hi, On 2023-07-02 21:44, Thorsten Glaser wrote: > Hi, > > please check the footnote below. Thanks! > > ---------- Forwarded message ---------- > Message-ID: <2d85d8b1-45e2-d6a8-5088-7fd838b7...@tarent.de> > Date: Sun, 2 Jul 2023 20:20:16 +0200 (CEST) > > Dixi quod… > > >Indeed. Weird. > > > >Thanks for reporting, I’ll have two or three looks at it… fixing that > >is going to be… fun. Not. > > OK so first analysis is showing the cause of the problem: > > • the buildd chroots for sid/unstable do not identify themselves as > sid/unstable but instead as trixie/testing, which is a bug onto > itself¹ No the chroots do not identify themselves. I guess you check the output of "lsb_release -a". If you believe the output is not correct, please report the bug to the lsb-release package. > however if the buildd chroot bug could be fixed, I’d be glad. There is nothing to fix on the buildd chroot side, please fix your debian/rules script instead. > ① sid buildd chroots should save the following content… > PRETTY_NAME="Debian GNU/Linux sid" > NAME="Debian GNU/Linux" > ID=debian > HOME_URL="https://www.debian.org/" > SUPPORT_URL="https://www.debian.org/support" > BUG_REPORT_URL="https://bugs.debian.org/" > VERSION_ID=unstable > VERSION_CODENAME=sid > … as /usr/lib/os-release.sid (in the chroot) and run… > # dpkg-divert --rename --divert /usr/lib/os-release.dpkg-dist \ > --add /usr/lib/os-release > # ln -sfT os-release.sid /usr/lib/os-release > … in the chroot, so the reported lsb_release is correct. They used > to have this in /etc/lsb-release, but the lsb-release program no > longer uses that. That's plainly wrong, that's not the job of a buildd to configure that. The chroots are just created using debootstrap --variant=buildd, which is basically a minimal chroot with build-essential. There shouldn't be any further visible configuration to the chroot, as your package is supposed to build the same way on a normal installation outside of a chroot. The functionality you describe should just come by installing additional packages. Regards Aurelien -- Aurelien Jarno GPG: 4096R/1DDD8C9B aurel...@aurel32.net http://aurel32.net ----- Fin du message transféré ----- -- Aurelien Jarno GPG: 4096R/1DDD8C9B aurel...@aurel32.net http://aurel32.net