IDLE vs. the command line?

[Otto T Hyvarinen]

We have this big Python program, developed by another institute, that
uses C extensions a lot. One of those C methods does the segmentation
fault when run from the command line, but _not_ when run from the IDLE.
This is quite strange.

What is different when a program is run in IDLE? Or is there no
difference and our case is just a fluke? Or should I update my Python
knowledge (and how)? If we can tell something about this in our bug
report, the developers would probably be grateful. 

We are using Mandrake 9.2 and Python 2.3. The program is developed in 
Python 2.2.*. 

Regards,

Otto Hyvärinen