*Part* of the problem appears to be that I was not aware of what had been shipped to the customer. It turns out that it was a debug build that contained assert()s which are effectively abort()s.
Charles -----Original Message----- From: IBM Mainframe Discussion List [mailto:[email protected]] On Behalf Of Charles Mills Sent: Thursday, June 16, 2016 10:06 PM To: [email protected] Subject: What causes a Signal 3? I have a POSIX C++ application running as a started task. At a customer the code reported a Signal 3. Looking at signal.h I see that Signal 3 is SIGABRT, and looking at the C library reference I see that SIGABRT is triggered by a call to abort(), which in turn may be caused by a call to terminate(). But there are no explicit calls to either abort() or terminate() anywhere in the source code. My code is very good about putting out messages for abnormal conditions. There were no preceding messages. I'm an MVS guy. I know what a S0C1, S222 or SB37 is but I confess I am not fully up to speed on signals. What the heck would cause an apparently spontaneous Signal 3? ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
