On Mon, 2026-01-05 at 21:22 +0100, Sebastian Andrzej Siewior wrote: > On 2026-01-05 18:54:57 [+0100], To Adam D. Barratt wrote: > > On 2026-01-05 17:34:32 [+0000], Adam D. Barratt wrote: > > > > Emilio pointed out on IRC that there are also packages that > > > > depend on > > > > other binary packages from src:clamav. Most of those are > > > > arch:all, > > > > with e2guardian being the exception. From the description, I > > > > guess > > > > that probably still has at least some use if the clamav > > > > functionality > > > > is disabled. > > > > > > Do either of you have any time (/ spoons / inclination) to have a > > > look > > > at that? Hopefully that's the last piece of the puzzle for the > > > removal. > > > > Oh. So there is clamsmtp and e2guardian that has a Depends: on > > clamav-daemon. This sounds wrong because the daemon can be accessed > > via > > network so a Recommends/ Suggests would be sufficient since it > > could > > access it on a remote host (unless they can access the unix socket > > directly). Whether is is feasible in a setup is a different story. > > I opened #1124704 for e2guardian where I disabled the clamav bits. It > doesn't support IP sockets.
Thank you! > Please remove clamsmtp for armel and mips*. It only supports unix > sockets and it makes no sense to have it without clamav. That's now #1124705, FTR. > That should be it. Fingers crossed. I need to teach our tooling how to request partial removals on particular arxhitectures (for libc-icap-mod-virus-scan, where the other binary packages continue to exist on the affected architectures), which will be another fun yak. Regards, Adam
