I have a bizarre problem that manifests on some but not all RHEL 4/5 systems. Steps to reproduce are:
- run tclsh8.4
- at tcl prompt: package require Oratcl 4.4
- at tcl prompt: parray env
- at tcl prompt: exit

Has anybody else seen this / can anybody reproduce this?

I am using redhat which uname -a reports 2.6.9-67.0.22.ELlargesmp
I am using tclsh 8.4.19 (actually, ActiveTcl 8.4.19).
Before I can do the above, I first have to set my ORACLE_HOME and ORACLE_LIBRARY environment variables.
After I do that, then I can perform the steps you mention above and do NOT get a core dump.
Without setting the ORACLE_ variables, the package require fails.

Thank you. Sorry, I wasn't clear. Oracle environment is configured, and the package loads and works fine. It's just in one script where combined with examining env, it core dumps, but only on exit after successfully completing. It is most bizarre:

prompt> env | grep ORA
ORACLE_SID=ora102
ORACLE_LIBRARY=/site/gngapps/Linux/oracle/lib32/libclntsh.so
ORACLE_HOME=/site/gngapps/Linux/oracle
prompt> tclsh8.4
% package require Oratcl 4.4
4.4
% parray env
env(BASH_ENV)        = /home/fred/.bashrc
<snip>
% exit
Segmentation fault

Having just looked at the Bug list, I think this might be the same as ID 2817940:
http://sourceforge.net/tracker/?func=detail&aid=2817940&group_id=12880&atid=112880

the way to reproduce the problem is to start a new interpreter from teh current interpreter:
$ tclsh
% set interp [interp create]
interp0
% $interp eval {package require Oratcl}
4.4
%
% interp delete $interp
%
% exit
Segmentation fault
$

For details check the bug report i posted - i have not yet got any solution for this.

In my case that core dumps prior to exit:
bash-3.2$ tclsh8.4
% set interp [interp create]
interp0
% $interp eval {package require Oratcl}
4.4
% interp delete $interp
Segmentation fault