[systemd-devel] (no subject)

[Colin Guthrie]

'Twas brillig, and Colin Guthrie at 03/09/12 23:46 did gyre and gimble:
> Hi,
> 
> OK, so these are the changes I'd propose to fix the issues mentioned
> in this thread. I've not tested them so this is more for general
> feedback as to whether this approach is generally a good one or not.
> 
> The initial patch is just extra debugging I added to try and work out
> why my gdm session is not marked with type=greeter so seems like a
> generally useful change when in debug mode.

It seems that while I still think the principle behind these changes
makes sense, that the various env var files are simply not updated
correctly when various changes occur:

e.g. compare the output and note the discrepancy in the [Ss]tate= values:

[colin at jimmy systemd (master)]$ loginctl show-session 1
Id=1
Timestamp=Tue, 04 Sep 2012 00:27:37 +0100
TimestampMonotonic=12876139
DefaultControlGroup=name=systemd:/user/gdm/1
VTNr=1
Display=:0
Remote=no
Service=gdm-launch-environment
Leader=3075
Audit=1
Type=x11
Class=user
Active=no
State=closing
KillProcesses=no
IdleHint=no
IdleSinceHint=1346714860433194
IdleSinceHintMonotonic=16273801
Name=gdm

[colin at jimmy systemd (master)]$ cat /run/systemd/sessions/1
# This is private data. Do not parse.
UID=492
USER=gdm
ACTIVE=0
STATE=online
REMOTE=0
KILL_PROCESSES=0
TYPE=x11
CLASS=user
CGROUP=/user/gdm/1
FIFO=/run/systemd/sessions/1.ref
SEAT=seat0
DISPLAY=:0
SERVICE=gdm-launch-environment
VTNR=1
LEADER=3075
AUDIT=1




I do have a patch that adds various calls to seat_save(),
session_save(), user_save() at appropriate junctures, but I've not
tested this yet and it would appear to be quite easy to get tied in
knots and add infinite loops!

Col

-- 

Colin Guthrie
gmane(at)colin.guthr.ie
http://colin.guthr.ie/

Day Job:
  Tribalogic Limited http://www.tribalogic.net/
Open Source:
  Mageia Contributor http://www.mageia.org/
  PulseAudio Hacker http://www.pulseaudio.org/
  Trac Hacker http://trac.edgewall.org/