That AC proposal neglets to mention just how exactly this behavior is
triggered, which is rather critical piece of information. Are new APIs and cli
switches being added - which ones?
Stuff like "move new instance to proper place" is an uninteresting
implementation detail, what we're really interested is the semantics that are
also testable. For example, "add all keys" sounds like all keys get moved no
matter what, but that's not what should happen, we should weed out any no
longer legit keys in the process - in other words, those that fail to import.
So it's more like "attempt to reimport all previously imported keys" than just
"add all". And as a part of that reimport, any short keyids get converted to
fingerprints - which is what we want. And on the semantic level I dont' think
we should or need to differentiate between rpmdb and other backends, from
rpmkeys perspective they should all behave the same.
So I think the requirements are more like
- a way to invoke keystore rebuild from the command line and API (and describe
how it'll look like)
- supports moving keys between all keystores, including itself (eg rpmdb ->
rpmdb)
- old invalid (expired, weak crypto, whatnot) keys are discarded in the process
with a warning message
- old short keyid based keys are converted to fingerprint based ones
--
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/issues/3347#issuecomment-2472590239
You are receiving this because you are subscribed to this thread.
Message ID: <rpm-software-management/rpm/issues/3347/[email protected]>
_______________________________________________
Rpm-maint mailing list
[email protected]
http://lists.rpm.org/mailman/listinfo/rpm-maint
