On Sat, Jun 11, 2005 at 08:56:06PM +0100, Jonathan Buzzard wrote: > > [EMAIL PROTECTED] said: > > The problem is that the KILL(2) call is used to check to see if a process > > exists at the specified PID. The default action for SIGUSR1 is to kill the > > targeted process. Thus the original process dies, and the new process exits. > > > > The propper thing to do is to use kill(pid,0), which just checks if a > > process exists with that pid without actually signaling it. This is standard > > behavoir, for this type of thing. > > > > The bug is not that the second hotkey fails to start, but rather that the > > attempt to start the seccond hotkey kills the first and then the second > > exits. > > Hum, I could have sworn that it did not used do this, and there should be a > > signal(SIGUSR1, SIG_IGN); > > line in there to stop the first hotkey dieing. If you look in wmtuxtime you > will see such a line. > > Stupid thing is if I now role up an update to toshutils, it won't be > acceptable for stable because it is new upstream version, despite being > just bug fixes :( > If it doesn't make any difference to you, I keep Sarge backports of my packages on my website. People can always get them there. Personally, I think an update to the utils (as you pointed out in your other mail) is overdue. Let me know what I can to help.
-Roberto -- Roberto C. Sanchez http://familiasanchez.net/~sanchezr
pgpEvL9ZzkuTs.pgp
Description: PGP signature
