On 9 February 2010 10:02, Jan Djärv<jan.h.d@swipnet.se> wrote:
Anyway, you have session suport compiled in and Emacs talks to
the session manager.
Do you have a directory ~/.emacs.d and is it writable?
Yes.
If you start emacs,
log out and then in again, do you have any session.* files in emacs.d?
Or any ~/.emacs.session.* files?
OK, the problem seems to be a bit more complex than I thought.
If I close down all my programs, log out, log in, start Emacs, then
log out and back in, everything works, and my session is restored.
If however I try to log out from a long-running session, GNOME says "A
program is still running: Emacs", and offers me the choice to log out
anyway, or to cancel the log-out action. If I ignore the warning and
log out anyway, then indeed the Emacs session is restored when I log
back in.
So in fact, my problem is really: why sometimes does Emacs refuse to
be killed? I tried for example starting an emacs client session, but
that didn't seem to trigger this problem.
In summary, it's a minor annoyance now that I understand what is going
on better, but it would be nice to work out why Emacs sometimes
refuses to die at first, althuogh it doesn't seem to matter if I just
log out anyway.
(I did come across another minor bug in the process which I'll now
file separately.)