Dennis Lee Bieber wrote: > And that was a direct cut&paste from a command window; showing it > had slept for some 90 seconds before I killed it.
Interesting. Just tried a 2.7.2 on a 32-bit MS Windows with following results: 1. sleep(5 - 2**32) sleeps for a few seconds 2. sleep(-1) sleeps much longer So this seems to confirm that it's a 32-bit underflow while preparing the argument for win32's Sleep() function. That said, an "IOError" is a bit better but still leaves room for improvement. I'll take this to the developers mailinglist and see if they consider the behaviour a bug. At the very least the docs are bad, I would say. Cheers! Uli -- Domino Laser GmbH Geschäftsführer: Thorsten Föcking, Amtsgericht Hamburg HR B62 932 -- http://mail.python.org/mailman/listinfo/python-list