Guido van Rossum wrote: > Ah, that's why sundry fails. I think it shouldn't depend on bsddb; > I'll look into that. test_sundry passes with no leaks now. Neal