It's a moderately large application, and so impractical to post it all. I'll have a go at isolating circumstances where this fault occurs then post the preceeding code - but not before this evening (London time) when I might have some free time. Peter
-- http://mail.python.org/mailman/listinfo/python-list
