Date: Tue, 22 Sep 1998 21:36:29 -0700 From: "Jordan K. Hubbard" <jkh@time.cdrom.com> To: Mark Tinguely <tinguely@plains.NoDak.edu> Cc: freebsd-chat@FreeBSD.ORG Subject: Re: FYI: function call failures Message-ID: <1293.906525389@time.cdrom.com> In-Reply-To: Your message of "Tue, 22 Sep 1998 14:07:14 CDT." <199809221907.OAA07181@plains.NoDak.edu>
next in thread | previous in thread | raw e-mail | index | archive | help
> from comp.arch a reference (http://www.ices.cmu.edu/ballista/ftcs98/) to > a test of function call failures. FreeBSD did not fair too well compared > to other OSes. These aren't failures and we've already had a flame-fest over a pre-release of this paper. :-) A failure is an [un]expected condition which produces an undesired result. When you pass NULL to strlen(), it is a desirable result for it to dump core just as surely as if you'd gone *((int *)0) = 10; You WANT those errors to crop up early so you can figure out what else went wrong to get an unchecked string into a function which is now blithely passing it around to other routines not written to take NULL as a valid argument. Yes? I think most people here can agree with that. This guy, however, calls it a failure when his test suite passes all manner of bogus arguments to various library functions, functions which are not documented as accepting arbitrarily bogus pointers, and one of them dumps core. He's arguing from a fundamentally different viewpoint here and, like the wackos in aviation circles who keep wanting to do things like reintroduce the bat-wing design to civil aviation, I don't find it useful to pay it much attention. :-) - Jordan To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-chat" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1293.906525389>