My "select item rtp20 missing from program stack" has returned. Obviously, rebooting the server is only a stop gap to whatever is causing this problem.

I know this occurs on my archive.shell program and always after I have done the archiving of cases – copy cases from one volume to second second volume and then delete off original. After that the first time a select is

performed that has more than one 'with' statement it fails. It doesn't even work if I try 2 select statements always on the second with statement it fails.

I have 5 servers all running Arev databases with the same code same setup same everything and only 1 fails. If the TCL levels were nested too deeply wouldn't it fail on all. I discovered a topic that said Arev may be losing track of the stack pointer or its pointing to a place in the stack that doesn't exist. How do I find out if this is what's happening?

What about placing push.session and pop.session in my program?

The "R" command in the debugger will show the Return/Program Stack.

The "#" command will do a garbage collect and report on memory usage.

It sounds as if you're still running low on memory below 640K.

Are you running the archive program on the server(s) or a from a workstation? Does the tcl WHO command show reasonably close available memory between servers/workstations?

Are symbolics being used for the select criteria? If so, are the symbolics using the {braces} or calculate functions within them? Each of these will consume a entry in the Program Stack.

Did you try adding a FLUSH; GARBAGECOLLECT (or GARBAGECOLLECT; FLUSH - can never remember the proper sequence) before the select statements in your shell program?

Are you doing another select within a readnext loop? If so, you need a push/pop.select.

View this thread on the forum...