One of my clients has been having a problem with Zope locking up every time she changes something. It appears that Zope goes into a loop somewhere as there are no tracebacks or anything else to indicate an error except that the browser reports that the page is unavailable and any subsequent attempt to go to the management screens or any other part of the site fails. The setup is as follows : Zope 1.10.3 (modified slightly to include the _on_delete_object() protocol and some timezones added to DateTime) running ZopeHTTPServer as localhost Windows NT4 (SP3) IE5 I can perform the same changes against a copy of her database without any problems on my test setup which is running an identical 1.10.3 installation on a NT4 (SP4)/IE5 machine. Has anyone else seen this behaviour or anything like it ? Are there any issues with Zope/Python on an NT box and service pack 3 ? Right now after spending most of this afternoon attempting to isolate the problem I am open to any and all suggestions as to how to track down the problem. The major difficulty is finding out what Zope/Python is doing when it goes off into the ether. Is there some way of breaking into looping Zope/Python code to find out what it is up to ? Any and all help much appreciated Robert Leftwich