Open darix opened 11 years ago
a pure publish event may not be enough, since the publisher skips re-publishing if no binary has changed.
I don't understand this issue, as signatures done with a currently expired key are still considered ok, as long as the key was not expired when the signature was made.
but the key file in the repository is still with the old expiration date.
And how is that a problem?
user wants to install package from repos with outdated key: e.g. http://download.opensuse.org/repositories/security:/OpenVAS:/STABLE:/v4/Debian_6.0/Release.key
of course the package manager will and should warn him "key outdated". having it republish the repos with the updated key file shouldnt cause my IO load and would solve that issue.
If only apt behaves that way it is maybe enough to do that only for debian type repos?
The package manager should not warn about an outdated key, an outdated key is perfectly acceptable as long as the signatures were done when the key was not expired. That's how pgp/gpg works, somebody should fix apt.
i guess the problem is related with #12333
The package manager should not warn about an outdated key, an outdated key is perfectly acceptable as long as the signatures were done when the key was not expired. That's how pgp/gpg works, somebody should fix apt.
well the reality is ... it does warn ... so we should republish to make this whole process more userfriendly.
The package manager should not warn about an outdated key, an outdated key is perfectly acceptable as long as the signatures were done when the key was not expired. That's how pgp/gpg works, somebody should fix apt.
As this issue has been lingering for mor than a decade, and is still happening with OBS + apt, I was wondering if you were aware of any bug reports filed against apt to try to get this fixed?
@laf0rge after reading https://www.phoronix.com/news/Linux-Mint-APT-Captain-Aptkit
i wonder if apt doesnt have a general problem with moving forward.
otherwise the extended key might not be published for a very long time. this is especially important for sub projects, as the maintainer might forget to trigger a build in those to force the publish.