Hey.

IIRC, that recommends (which should have been a depends anyway) was introduced
because of #1076494.

Looking at:
https://github.com/fail2ban/fail2ban/issues/3787#issuecomment-2213618204
distutils is allegedly no longer used, and indeed a poor-man’s check:
$ grep -R distutils /usr/lib/python3/dist-packages/fail2ban
$
seems to confirm this (for both, sid and trixie).

So AFAICS, but please double check, the Recommends can be dropped.

Cheers,
Chris.

Hi,

Christoph Anton Mitterer <calestyo@scientia.org> (2025-10-10):

After merging my fail2ban integration for a customer's product (based on
Debian 12 initially, merged into our Debian 13 environment), finally
catching up with trixie, I was quite surprised to see the amount of
extra packages being pulled when deploying.

We have a number of Python packages already (50 of them), and the Debian
12 integration only pulled a few packages:
 - whois

Meanwhile, on Debian 13, that's a lot more!
 - python3-autocommand
 - python3-inflect
 - python3-jaraco.context
 - python3-jaraco.functools
 - python3-jaraco.text
 - python3-more-itertools
 - python3-pkg-resources
 - python3-setuptools
 - python3-typeguard
 - python3-typing-extensions
 - python3-zipp
 - whois

I've prepared a modified package and I can confirm the footprint is back
to something that looks more reasonable (and in line with Debian 12):
 - whois

Meanwhile, the default sshd jail is operational (thanks for fixing the
systemd back thing, by the way), alongside a custom jail that leverages
pyinotify (it scans access.log-like files).

I'm really not sure it qualifies for a stable upload, but it seems to me
that keeping the footprint of an important package like fail2ban (that
one might want to deploy on many systems) to a minimum… deserves a
little more than just “normal” severity. That being said, I'm definitely
not pushing for a possible fix in stable, I'm not sure it qualifies in
the first place, and I'm perfectly able to either live with the extra
packages or with a locally-patched package…

Maintainers, you'll find a patch against master attached, and I've
pushed a pu/bug-1117706 branch for consideration. I haven't done any
kind of tests within unstable or forky though.

Thanks for your time, and thanks for maintaining fail2ban!


Cheers,