home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.std.c
- Path: info.physics.utoronto.ca!utzoo!spenford!henry
- From: Henry Spencer <henry@zoo.toronto.edu>
- Subject: Re: Undefined result vs. int's holding undefined values.
- References: <oZA8wQ9ytpjN084yn@csn.net> <4cs460$d6e@news.informix.com> <821155074snz@genesis.demon.co.uk>
- Sender: henry%spenford@zoo.toronto.edu (Henry Spencer)
- Organization: SP Systems, Toronto
- Date: Tue, 9 Jan 1996 19:50:08 GMT
- Message-ID: <DKxJ3K.4u1%spenford@zoo.toronto.edu>
-
- > Does an appropriate "SAFE TEST" exist
- >for multiple. Has anyone actually seen a real production program where every
- >calculation was protected against overflow/underflow.
-
- Safe tests exist for all the arithmetic operations, although you have to
- be careful and clever. See the paper "On Legality Assertions in Euclid",
- by Dave Wortman in the July 1979 IEEE Transactions on Software Engineering,
- for a fairly complete discussion of the matter.
-
- A program in which overflow/underflow can occur is a buggy program. In
- most cases, the nature of the program is such that the calculations
- can be shown not to overflow/underflow. Where this is not so, checks
- are necessary for a reliable program.
- --
- Look, look, see Windows 95. Buy, lemmings, buy! | Henry Spencer
- Pay no attention to that cliff ahead... | henry@zoo.toronto.edu
-