longjmp(3)

Rick Jaffe rsj at oxtrap.UUCP
Mon Apr 18 08:28:48 AEST 1988


when it receives a signal (SIGUSR1) telling it not to bother: the
answer isn't needed any more.  The program assures that locks and
global variables won't cause any trouble, and executes a longjmp(3)
from the signal handler back up to the main procedure.

Is this "safe"?  Is there a better way, short of setting a global
flag, and then either returning from the several hundred procedure
calls or executing a longjmp() from inside the recursive procedure?



More information about the Comp.unix.questions mailing list