Bug#247193: kdm: reserve doesn't seem to work
On Wednesday 05 May 2004 07:25, you wrote:
> On Tue, May 04, 2004 at 11:50:54AM +0200, Achim Bohnet wrote:
> > On Monday 03 May 2004 21:18, Ross Boylan wrote:
> > > Package: kdm
> > > Version: 4:3.2.2-1
> > > Severity: normal
> > >
> > > I uncommented the first commented display line in /etc/kde3/kdm/Xservers:
> > > :1 local@tty2 reserve /usr/X11R6/bin/X -nolisten tcp :1 vt8
> > > and did
> > > invoke-rc.d kdm restart
> > >
> > > Despite this /var/run/xdmctl/ only showed xdmctl and xdmctl-:0.
> > Reserve is working fine here (can start other parallel session
> > with k-menu -> 'Start new session'. Nevertheless I only have
> I don't see that option anywhere on my K-menu, either top level or
> further down in any likely spot (system, utilities, control center).
> I think I remember seeing it somewhere (on the kdm greeter?) and
> wondering what it did.
> Oh... maybe this only appears if you have a reserve session possible.
> I may have seen it when I had one, and now it's gone because I removed
> the reserve word since I couldn't make it work.
> > xdmctl and xdmctl-:0 too. When I start the first new session
> > xdmctl-:1 is created. So xdmctl-:? are only created when really
> > used and not when defined in Xservers.
> > >
> > > As I understand it, there should have been an xdmctl-:1 to which I
> > > could direct
> > > echo reserve
> > > to start up the second login screen.
> > From my tests it looks like you miss understood something ;)
> > >
> > > I also tried restarting my X server on vt7. This didn't help.
> > >
> > > I have xfree86-common version 4.3.0-7, and am in a mostly testing
> > > system.
> > >
> > > I may not have understood how this is supposed to work, but this is
> > > what I got from the docs.
> > I had a look at the /usr/share/doc/kdm/README.gz and
> > echo reserve > /var/run/xdmctl/xdmctl-$DISPLAY
> > works fine here (as user log in on $DISPLAY). You should send
> > the reserve command to the xdmctl file responsible for your display.
> But there is no such file. Or are you saying that to start session 1
> I should send this to xdmctl-:0?
Yes. To start another session from a session running on DISPLAY :0
echo reserve to xdmctl-:0.
> > I guess you get confused by the assumption that xdmctl
> > file should exist for every reserve setting in Xservers.
> Correct. Since, as written, it appears you must send the "reserve"
> command to the corresponding FIFO.
Yeah, to the corresponding FIFO of 'your' display, because it's
your communitation line to kdm. Assume that a second session
is already running. echo reserve to xdmtcl-:0 will create a third
> > If you've some suggestion how to enhance kdms README please
> > send a patch. Otherwise I suggest to close the bug report.
> The fact that a bug report does not include a fix does not mean there
> is no bug. Short of a patch, I suggest the document should explain
> how to make reserve work.
okay, but then subject change:'reserve doesn't seem to work' to 'reserve
docu misleading' ;)
> This would involve several elements:
> 1. How to get the "start a session" to appear on your menu. I don't
> know how.
o Uncomment one or more of the 'reserve' line in /etc/kde3/kdm/Xservers.
o logout from kde
o restart kdm (??or is reload enough??)
> 2. Stating that to activate a reserve session you should invoke start
> a session and then send the "reserve" command to the corresponding
> FIFO (I'm assuming that's the correct method).
to the fifo of __your__ display. xdmctl-:0 if you are on DISPLAY=:0
and want to start a second (or third) session on :1 (or :2).
> 3. (Optional) Explain a pure FIFO way to achieve the effect of start a
> sessio without getting the GUI involved. (I suspect there is such a
> method, but I don't know it.)
echo reserve > /var/.../xdmctl-$DISPLAY
> At any rate, the missing piece of the puzzle for me seems to be how to
> invoke the "start session" function, or how to get it on my menu.
> > Achim
> > P.S. The word 'seem' in the subject is an indication that
> > the report is better first send to debian-kde list to make
> > sure it's a bug ;)
> 'seem' referred to the fact there might be some way to make it work,
> as there seems to be. But even if there is, the documentation needs
> to explain it, so it's a documentation bug.
I wanted, discussing this on debian-kde would have given probably
more and additional infos than on the here on the bug list. The
better [patch is best;)] the bug report is the easier is it for
the overloaded kde pkg mainatainers (I'm not a maintainer) to fix
(my) philosophie is: collect as much, precise infos and when
everything is collected, submit a bug report. That's the best
we could do to help debian.
> When I get to a good point to tweak my Xservers and kill all my
> sessions I'll give this all a try.
Have fun ;)
To me vi is Zen. To use vi is to practice zen. Every command is
a koan. Profound to the user, unintelligible to the uninitiated.
You discover truth everytime you use it.