Join Nostr
2026-01-03 00:34:50 UTC
in reply to

npub177…teh9a on Nostr: Sorry for the late reply -- this is a fair criticism, and I agree with you on the ...

Sorry for the late reply -- this is a fair criticism, and I agree with you on the
core issue: tagfiles stop being first-class citizens the moment the installer
exits. There’s no lifecycle connection to slackpkg -- by design.

That said, I’ve been experimenting with tagfiles as a deliberate project rather
than a convenience feature. I actively use this approach for a jump machine at
work, built on Slackware-current, and I keep the tagfiles updated as -current
evolves.

The cost is exactly what you describe: manually auditing dependencies and keeping
the set coherent over time. There’s no automation here -- just careful review and
explicit decisions.

In other words, tagfiles can scale -- but only if you accept that dependency
resolution becomes a human responsibility, not something Slackware tries to
infer or automate.

In practice, I treat these systems as intentionally “finished”: slackpkg is
used conservatively, often without install-new, and changes are explicit and
reviewed. It’s not elegant, but it’s honest -- and very Slackware.

I’d love to see better tooling around this, but I’m also wary of solving it in a
way that undermines Slackware’s transparency. Your idea of exploring tagfile
support in slackpkg is interesting precisely because it raises that question.

For completeness: to detect dependencies when building and updating these
tagfiles, I rely on an existing Slackware helper script that scans for broken
library links. This script helps *identify* missing dependencies, but it does
not resolve them or make decisions -- the review remains manual.

Script (by mozes@slackware.com, 2010):
https://git.sr.ht/~r1w1s1/slackware-scripts/blob/main/find-whatneedsrecompiling

Tagfiles repo (Slackware-current, used in practice, shared as reference):
https://git.sr.ht/~r1w1s1/slackware-tagfiles