Luca Boccassi, le ven. 30 sept. 2022 23:33:39 +0100, a ecrit: > On Thu, 29 Sep 2022 09:06:58 +0200 Samuel Thibault > <sthiba...@debian.org> wrote: > > Marco d'Itri, le jeu. 29 sept. 2022 04:53:32 +0200, a ecrit: > > > On Sep 22, Samuel Thibault <samuel.thiba...@ens-lyon.org> wrote: > > > > > > > On usr merge attempt on a hurd-i386 system, things go really bad: > > > I am quite disappointed that it took 7 years to find out that the > > > package does not work at all on Hurd. > > > > I have not had known usrmerge for 7 years. I have not seen a call for > > testing until recently. So, no, testing it wasn't a priority so far. > > What if we had a fixed version uploaded only to experimental, and leave > it out of unstable until bookworm shipped? That way we don't introduce > risks for bookworm, and hurd-i386 users can test it?
Well, I have already uploaded a +hurd.1 fixed version to debian-ports' unreleased, to save hurd boxes from the mayhem. experimental wouldn't have saved them since upgrades would rather pick up unstable by default (while unrelease does take over unstable). But as soon as a newer version is uploaded to unstable, I have to upload another +hurd.1 version. Or else I could upload a 1: epoch version on debian-ports' unreleased, so it'd override the unstable version anyway. Samuel