On Wed, Feb 26, 2025 at 02:59:54PM +0100, Chris Hofstaedtler wrote: > * Simon McVittie <s...@debian.org> [250226 14:45]: > > Do we know that this is only a practical problem for /usr/games, or are > > there also collisions between /usr/bin/foo and /usr/sbin/foo in practice? > > /usr/bin/foo and /usr/sbin/foo is where this started, and most > packages got fixed. The remaining ones have open bugs.
I finally usertagged and pinged the remaining bin<>sbin bugs: https://udd.debian.org/cgi-bin/bts-usertags.cgi?user=z...@debian.org&tag=bin-sbin-conflict > So really, this is now about the accidential expansion to > /usr/games. I agree it would be good to fix the name conflicts there > too. Just not in time for trixie. Unfortunately it turns out, we _already_ have problems in the archive. I was previously under the impression that it /probably/ didn't really matter. Simon however pointed out the pacman-package-manager <> pacman conflict, which is actually *bad*: pacman installs a pacman.desktop file with "Exec=pacman". It *relies* on /usr/games/pacman a) being in PATH and b) being _the only_ "pacman" program in PATH. At least this conflict is IMO truly a serious bug, _regardless_ of what policy says. Having found this example (thanks to Michael Biebl for some discussion on this), I no longer belief we should really have an exception for trixie. Also, if its ~22 bugs, then it's not the end of the world either. Chris
