On Sat, Apr 30, 2011 at 7:29 PM, Dylan Evans <dy...@contentfree.info> wrote: > I think i see what you are trying to do but it depends on the environment > and your goals. > Generally i think you need to separate your code by forking (or perhaps you > have already done that?), > then you can run a check to see if the process died as expected. I don't > know though, this not much > information to go on, but if you are running untrusted code then you need to > be able to isolate and kill it.
Well, I've gone with a slight variant on this. If the Python script doesn't terminate in a timely manner, the process will be killed with a longjmp straight from the signal handler (the setjmp having been done long long ago back when the process initialized itself). So Py_Finalize() will be called, but no other Python-related functions _at all_, and the process promptly terminates. I'm assuming that Python will flush out all its state on process termination (that is, it doesn't hang onto any system-global resources). Thanks for the advice! Chris Angelico -- http://mail.python.org/mailman/listinfo/python-list
