"Tim Golden" <[EMAIL PROTECTED]> writes: > Well, one answer would be: because it's really quite hard to > work out if there are, in fact, no changes which would break > existing code. Obviously, if everything's nicely tied up with > tests etc. it should be plain-sailing. But if it's not...
Even with tests... They can show you that what you test is not broken, but they can't make you 100% sure that everything will work. -- Jorge Godoy <[EMAIL PROTECTED]> -- http://mail.python.org/mailman/listinfo/python-list
