On Wed, 29 Oct 2025 18:27:53 +0000 "Niedermayr, BENEDIKT" <[email protected]> wrote: > On 29.10.25 17:38, Jochen Sprickerhof wrote: > > * Benedikt Niedermayr <[email protected]> [2025-10-29 15:27]: > >> Package: sbuild > >> Version: 0.91.1 > > > > ^^ > > > > This version seems to be false, I can't reproduce it there. > > sorry, the version in question is that one from the trixie release, > during my tests I switched between different Debian Distros and > accidentally created the bug report from within sid (a docker container). > > The version where this BUG actually appears is of sbuild=0.89.3+deb13u1. > > > > > > > * Helmut Grohne <[email protected]> [2025-10-29 17:27]: > >> If I change Sbuild/AptResolver.pm to arch qualify the $dummy_pkg_name > >> (in the line defining @apt_args in sub install_deps), then building for > >> bookworm works again. Is there any downside to always including the host > >> architecture there? > > > > This is already done in: > > > > https://salsa.debian.org/debian/sbuild/-/ > > commit/0da1ccda97e6a4134c2aadf85913680fd8274e2b > > I've applied that patch ontop of 0.89.3+deb13u1 and this indeed fixes > this error! > > > We should maybe put it in the next stable upload. > > Does that mean this would be shipped into trixie? >
Today, we were also able to avoid the issue by using unpatched sbuild with --build-dep-resolver=aptitude. But now we are unsure if (and how broadly) to use that workaround or better wait for a stable fix in trixie. Comments welcome. Thanks, Jan
