If gpm doesn't start properly, like for example if it can't open the mouse device it starts for a moment, forks a process, prints a error mesage and exits with status 0. The problem is, twofold, first it leaves /var/run/gpm.pid with a pid that has exited (gpm's fault here) and 2nd, since it exits with a status of 0, start-stop-daemon thinks it started ok, and says [OK]. Suggestion: Find an appropiate way (maybe via pidof) to determine if it did start and it it's pid matches the in /var/run/gpm.pid
It's relative (I think) with bug ID 3297. There could be some "clever" mechanism for doing with PIDs in rc scripts. if "start" is called a) look in /mnt/.init.d/started if I'm running b) look for pid c) look, if process of this pid is working with combination of this information do the start a) say-> I'm running, if (a && b && c = true) b) start if (a && b && c = false) for b) do cleaning dependant on a || b || c || -> logical OR && -> logical AND Martin Schlemmer is working on it, I hope :-))
I will take a look at gpm code, I've been so busy as of late haven't had much chance to take a look at stuff. We can probably just patch gpm to handle this correctly maybe? Might be the easiest. Naz
new init scripts handle this.. out dated bug
