JP Mens

Today in "diagnostic messages I've not seen before", comes this beaut

edit: solved -- see follow-up

#debian13

Jun 7 at 5:48amWeb
Show threadJP MensJun 7

thanks to a question @and1bm asked, I found the problem: it really is an older agent: I forward gpg-agent using SSH, and Debian (where the diagnostic was emitted) has a newer gpg than my workstation has.

For those wanting to know why I do that, I wrote that up a year ago: https://jpmens.net/2025/04/04/forwarding-gnupg-agent-over-ssh/

Jan-Piet Mens :: Forwarding GnuPG agent over SSH

Show threadJP MensJun 7
and, fwiw, if the server side has a higher version, that appears to be acceptable: no diagnostic message is printed to the effect.
Show threadPaul Wouters 🇪🇺🇨🇦Jun 7
@jpmens shouldn’t you avoid gnupg 2.5.x anyway because that’s the internet / OpenPGP hostile version of gnupg ?
Show threadJP MensJun 7
@letoams should I avoid it? I wasn't aware there's a hostile version ...
Show threadBen ZaninJun 7
@jpmens I've experienced the same error message under macOS/Homebrew, where updating the package apparently does not automatically restart the daemon! Surprised and puzzled the heck out of me for about an hour. Thanks for mentioning this other potentially causative case too!
Show threadotto@openbsdJun 7
@jpmens maybe "us" is a library? Writing clear error messages is hard.
Show threadand1bmJun 7
@jpmens an old gpg-agent still running, started before a recent upgrade?
Show threadJP MensJun 7

@and1bm that's what it sounded like to me, but after upgrade (last night) I rebooted.

To be sure, I ran gpgconf --kill gpg-agent but no difference.