GNOME Bugzilla – Bug 644547
must connect to the session manager
Last modified: 2011-03-27 16:50:49 UTC
gsm used to daemonize, which gnome-session interpreted as the sign that gsm is ready. It stopped daemonizing a while ago, so now gnome-session times out the session phase in which gsd is running, slowing down login. Everything that runs in a session phase other than 'applications' _must_ properly notify the session manager that its done starting up, either by daemonizing/exiting, or via xsmp, or by registering as a session client using the dbus api.
It already does that (and has for the past couple of months) in gnome-settings-daemon/main.c in on_client_registered() (and related). What doesn't work there exactly?
Not sure, I've talked with ray about this today and just filed this bug to not forget. Ray wanted to look at it.
Running gnome-session with --debug shows that gnome-settings-daemon is registering. And I don't get a really long delay like I used to either.
at some point login was taking a long time for several people accross the internet. It happened to me, too. I looked in .xsession-errors and noticed a message from gnome-session saying that gnome-settings-daemon didn't register. Matthias mentioned it was slow for him the next day. I told him about the message I saw the day before and guessed the problem was it wasn't registering. It's fast for me now, too, so I say we should just close this one and let it get reopened if anyone is still seeing it. Could be gremlins, could be a g-s-d plugin making a blocking dbus call that times out, could be cosmic rays. Lacking more info, hard to say.
mccann hit this today on a live image, reopening.
So it does connect, but sometimes a timeout is triggered, should this one be closed in favour of bug 645650 ?
(In reply to comment #6) > So it does connect, but sometimes a timeout is triggered, should this one be > closed in favour of bug 645650 ? Fine by me, I'll let Ray handle it though.
*** This bug has been marked as a duplicate of bug 645650 ***