Re: Coredumps and 4GL
Posted in 1998
Topics: Stored Procedures & SPL, Connectivity: ESQL/C, 4GL & Embedded SQL
If you have not INITIALIZEd your variables before using them, they could contain anything at the time you try to use them, this is often a reason for core dumps. Thomas J. Girsch wrote: > John - > > I've seen this behavior in the past, but not with 4GL directly. It was > with C programs. I believe 4GL compiles to C, so this might be relevant. > > Basically, in C, this behavior usually meant that you were overflowing a > string somewhere. By inserting display (printf) statements, you were > affecting the programs stack space, and thereby affecting when the overflow > would actually get you into trouble. > > Just as a SWAG, I'd suggest that you comb through the code and make sure you > don't have any strings (CHARs) that might be too small. > > Otherwise, if your UNIX system has sdb or adb, you could try stepping > through with one of those cryptic debuggers. It's crude, but it's > effective. > > Carlson@WHSmith wrote in message <36388F94.5AE@bellsouth.net>... > >4GL: 7.20.UD2 > >Engine: 7.23.UC11 > >OS: HPUX 10.20 > > > >When I add a few displays (yeah, we're IFMX-debugger-deprived), I'm able > >to see where it is in the program when it dumps. When I add a few other > >displays, the location changes. > > > >When I clean up the code ( close cursor / free cursor ), it runs a bit > >longer and then still coredumps. > > > >I remember the 'old' (4.1) days where a display or a sleep would "take > >care" of the problem, but this is ridiculous. Anyone out there have any > >ideas?