seedfiles now available in experimental

Martin Steigerwald martin at lichtvoll.de
Sat May 23 16:09:57 BST 2026


Greetings Matthias.

Matthias Geiger - 23.05.26, 02:45:10 CEST:
> seedfiles ( a lightweight systemd-tmpfiles reimplemention) is now
> available in experimental. While it certainly is not well-tested yet, it
> already work for me with daily use. If you have a system to test on,
> I'd appreciate more early adopters. Its main benefit is the really
> small size and portability.

I installed it on my main laptop which is using Runit. However I found I 
did not have systemd-tmpfiles installed on my laptop. Maybe I removed
it at some time. I think I remember there has been some crazy
systemd-tmpfiles let's remove everything failure being reported in the
news and that may be why I removed it.

Ah, yes, that was related to the "--purge" option of the original. Glad 
you did not implement that one. So yes, please avoid implementing stuff 
that can delete your /home or so. :) I appreciate your approach not to
implement everything. I hope you implemented the --remove option in
a very careful way and avoided recursive removals.

I misread your readme and initially tried to run seedfiles with
--cat-config option. I wonder whether this one might be a good one. Or at 
least document which paths it uses to look for config files. Maybe in some 
minimal man page?

By the way, README.md is missing from the package /usr/share/doc 
directory. If you mention it in the package description I think it is 
better to include it.

I found a typo in the readme. Not implemented option --gaceful should be 
--graceful I bet.

> After some more testing I will likely upload it to unstable.

What is there to look out for during the test?

What I tested so far is:

seedfiles --create --dry-run

Then I found a directory, /run/tpm2-tss/eventlog, that had not been
created before. So I ran

seedfiles --create

and it has been created.

So far so good.

> PS: yes, I know, in an ideal world we wouldn't need such
> reimplementations, but here we are ...

Yeah.

Best,
-- 
Martin





More information about the Debian-init-diversity mailing list