On Thu 2016-11-03 01:44:44 -0400, Vincent Lefevre wrote: > On 2016-11-02 23:52:43 -0400, Daniel Kahn Gillmor wrote: >> I think we ought to take the "systemd" meme out of discussion here. I >> think it's being used by conflation with "d-bus", and that confusion is >> unlikely to be helpful in resolving these issues. > [...] > > Thanks for the explanations. So, the problem seems to be related > to d-bus (which itself may be related to systemd). When I ssh > to my machine with sysvinit (for which there are no problems), > DBUS_SESSION_BUS_ADDRESS is unset, while when I ssh to my > machine with systemd, DBUS_SESSION_BUS_ADDRESS is set to > "unix:path=/run/user/1000/bus" (i.e. the same as in the physical > X session, so I'm wondering whether this is correct!).
this still isn't necessarily systemd -- you can have a d-bus session connected to an ssh session on sysvinit system. But yes, libpam-systemd does make it easier to have a d-bus session get initialized upon ssh login. > When I ssh to my machine where DBUS_SESSION_BUS_ADDRESS is set, > if I run "gpg -d file.gpg", it doesn't work. If I unset this > environment variable and try again, it still doesn't work, but > that's probably because the gpg-agent that is running (due to > the previous test) has DBUS_SESSION_BUS_ADDRESS set in its > environment. So, if I try again after killing the running > gpg-agent, it works, i.e. I can type my passphrase from the > curses UI. That sounds like the right analysis to me. Hopefully we can get this resolved. --dkg