dsg@mbunix.mitre.org (Goldberg) (05/30/90)
An application I wrote originally under Ultrix version 2.3 no longer times out properly. I set an alarm(2) for 4 minutes, do some processing, then fork a subprocess to do some more processing and user interaction. The parent waits for the subprogram to finish, using the return status to determine it's next step. The problem: under 2.3, if the user waited too long, the SIGALARM would be generated, and the whole program would time out, as expected. Under 3.1, the SIGALARM is apparently not being generated (I tested this by sending one myself with kill(1) after waiting over 10 minutes). After I sent the SIGALARM, the program caught it, and timed out as I expected it to. I also found that the SIGALARM is properly generated if there is no subprocess/wait going on. Is this a known bug (Hopefully fixed under 4.X)? Or was the behavior under 2.3 the bug? Thanks in advance for any help. -- Dave Goldberg UNIX Systems Programmer/Administrator The Mitre Corporation MS B020 Bedford, MA 01730 617-271-2460 Domain: dsg@mbunix.mitre.org UUCP: linus!mbunix!dsg