Nickolay Olshevsky

@andrey-utkin Could you please add check whether run is made under the GHA, so only local runs would be cleaned up automatically? (this should be helpful: https://docs.github.com/en/actions/learn-github-actions/environment-variables#default-environment-variables ) Basically, current...

> For now I've added a way to preserve the tempdirs by setting an env var `KEEPTEMP`. > > I'd rather make GHA workflows declare explicitly that they want this,...

Additionally, we may check whether document is already signed (includes one-pass signature packets at the beginning), and update their nested field (one-pass signature packet is not hashed).

@kaie Thanks for reporting, I'll take a look.

@kaie I was able to import this key to RNP with PERMISSIVE option enabled. It actually contains a bunch of illformed signatures, GnuPG reports those as well. Or is it...

> Notably, the final state of the codebase is exactly identical to current master, perhaps we could do without duplicating the history? It's visible that you've carefully considered each commit,...

> Well, that was my entire proposal. If the current state of `master` branch is what we are happy with for the release, then I suggest we tag that state...

> It would be nice if fix for #1894 is also included in this release, otherwise looks fine Please see reply in the original issue. If it's okay to ship...

This was changed since GnuPG 2.3+ : https://github.com/gpg/gnupg/blob/master/agent/keyformat.txt

@andrey-utkin Actually, it's what issue heading says - in this particular case correctly would be to say that subkey is expired, not invalid. That's what this issue about :)