From glondu at debian.org Mon Nov 16 08:23:32 2026 From: glondu at debian.org (=?UTF-8?Q?St=C3=A9phane_Glondu?=) Date: Mon, 16 Nov 2026 09:23:32 +0100 Subject: Reproducibility and testing migration In-Reply-To: References: <48562f8e-88a8-48aa-8e74-9f5fc4cc9f48@debian.org> <3cb3e614-7450-4f09-9202-76d4d526432a@debian.org> Message-ID: Hi Holger, Le 14/11/2026 ? 13:22, Holger Levsen a ?crit?: >> [1] https://reproduce.debian.net/excuses.html?source_name=eliom >> >> Wouldn't it make more sense for reproduce.debian.net to report build >> failures as "UNKNOWN" rather than "BAD"? > > not really, because reproduce.debian.net is rebuilding (or is aiming to rebuilding) > packages which have build before. (and it's using the exact same dependencies > as the build on buildd.d.o used), so it should not happen that such a build does > not build. Well, it does happen. Meanwhile, eliom managed to be reproduced on loong64, but not on amd64. And it seems that it doesn't always fail on the same test. And I cannot reproduce the failure on my own machine; it doesn't happen on buildds either. There must be something not obvious specific to the reproduce.debian.net setup. I tried with nice, cpulimit, systemd-run with a CPU limit... to no avail. Meanwhile, packages cannot migrate to testing... Cheers, -- St?phane From holger at layer-acht.org Mon Nov 16 08:57:53 2026 From: holger at layer-acht.org (Holger Levsen) Date: Mon, 16 Nov 2026 08:57:53 +0000 Subject: Reproducibility and testing migration In-Reply-To: References: <48562f8e-88a8-48aa-8e74-9f5fc4cc9f48@debian.org> <3cb3e614-7450-4f09-9202-76d4d526432a@debian.org> Message-ID: On Mon, Nov 16, 2026 at 09:23:32AM +0100, St?phane Glondu wrote: > Well, it does happen. yes, indeed... :/ > Meanwhile, packages cannot migrate to testing... ...this is why you can still just file an unblock bug so it can migrate. -- cheers, Holger ??????? ??????? holger@(debian|reproducible-builds|layer-acht).org ??????? OpenPGP: B8BF54137B09D35CF026FE9D 091AB856069AAA1C ??? "All opinions are not equal. Some are a great deal more robust, sophicated and logical than others." - DouglasAdams -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 833 bytes Desc: not available URL: From vagrant at reproducible-builds.org Mon Nov 16 09:14:14 2026 From: vagrant at reproducible-builds.org (Vagrant Cascadian) Date: Mon, 16 Nov 2026 01:14:14 -0800 Subject: Reproducibility and testing migration In-Reply-To: References: <48562f8e-88a8-48aa-8e74-9f5fc4cc9f48@debian.org> <3cb3e614-7450-4f09-9202-76d4d526432a@debian.org> Message-ID: <87v75xuoih.fsf@wireframe> On 2026-11-16, St?phane Glondu wrote: > Le 14/11/2026 ? 13:22, Holger Levsen a ?crit?: >>> [1] https://reproduce.debian.net/excuses.html?source_name=eliom >>> >>> Wouldn't it make more sense for reproduce.debian.net to report build >>> failures as "UNKNOWN" rather than "BAD"? >> >> not really, because reproduce.debian.net is rebuilding (or is aiming to rebuilding) >> packages which have build before. (and it's using the exact same dependencies >> as the build on buildd.d.o used), so it should not happen that such a build does >> not build. > Well, it does happen. > > Meanwhile, eliom managed to be reproduced on loong64, but not on amd64. > And it seems that it doesn't always fail on the same test. And I cannot > reproduce the failure on my own machine; it doesn't happen on buildds > either. > > There must be something not obvious specific to the reproduce.debian.net > setup. I tried with nice, cpulimit, systemd-run with a CPU limit... to > no avail. Not sure off the top of my head what it might be, but a list of expected and known differences: https://release.debian.org/reproducible-builds-host-requirements.html live well, vagrant