[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: XDM loop: Fix report.


On 18-Feb-98 The Thought Assassin wrote:
- -> * find out whether X was killed forcibly, and act accordingly - do we
- ->   restart or not? depends whether the person "logged off" or "shutdown".


- -> So XDM can take care of this for us, and we have much better control of
- -> how, and whether XDM exits - we can make XDM exit to whatever kind of
- -> error/info screen we want.


- -> This seems to be to be both the "right way" and the most powerful way.
- -> (the two often coincide) I hate to use a workaround when we could take
- -> away the cause of the problem by giving XDM the brains not to restart.

I'm not proficient enough to go digging through a program like XDM to add the
ability to do some checking before it restarts.

- -> <Innovative but slightly ugly solution deleted>

Watchit buddy!  It took me all of 10 min's to come up with this.

- -> I really think that the architecture is in place to remove this problem at
- -> the source rather than having to fix the symptoms by starting a process
- -> just to kill it:)
- -> What do the rest of us think?

I think that this type of fix is the best way to do it until a skilled
programmer, perhaps my co-leader, Orn, gets the code fixed if it needs it.

Although this sort of script will be refined for failsafe and such even if the
loop is fixed in the code.

A note:

I was just looking over the man page on XDM and noted it has some connect
timeout and retry args.  I'm going to play with these tomorrow and see if they
might effect this loop.  Might be right in the program already to fix this
problem.  Innovative, eh?  These are in addition to XDMCP timeouts which makes
me think they just might be for the local display.

- ---
E-Mail: Cyberdyn@seul.org       Date: 18-Feb-98        Time: 04:00:41
Playing an unamplified electric guitar is like strumming on a picnic table.
                -- Dave Barry, "The Snake"

Version: 2.6.3a
Charset: noconv