Leaking code on non-compile machine

[Chuck Lackey]

Here's one for you gurus out there. I can't get too involved with my limited knowledge of programming, but if you need more info (assumed), I can get our Development Manager involved here.

We are a software company with an odd problem. We've been building code on HP-UX for years and this is a new one on us.

Our app is C++ based and we build code in-house on an L3000 (PA-8600). Suddenly, on a customer's rp7410, out code is growing by leaps and bounds. We can run the same code in-house on our L3000 with no problem. However, we run the same code on a 700 class machine (J2240) and it leaks like the rp7410. We have matched the system software and patches on the L3000 as close as possible to the rp7410 to no avail.

It's beginning to look as if our code will leak on everything but the same class as the build machine, an L3000 (or PA-8600?). We are in process of rebuilding the code on the J2240 to start the proving process on this theory.

Just wanted to see if anyone has any "smarts" to add to the process while we're waiting on the build to complete.

Much Thanks,

Chuck Lackey

[Tim Sanko]

Is it possible to link the code on the leaking machine?

That does not mean compiling the code, but it does mean using the ld command to complete the linkage for the executable on each machine.

Tim