Xref: utzoo comp.sys.ibm.pc:23036 comp.sys.intel:656
Path: utzoo!attcan!uunet!lll-winken!ames!xanth!ukma!rutgers!elbereth.rutgers.edu!ron.rutgers.edu!ron
From: ron@ron.rutgers.edu (Ron Natalie)
Newsgroups: comp.sys.ibm.pc,comp.sys.intel
Subject: Re: unsafe computing (Was: correct code for pointer subtraction)
Keywords: C pointer math runtime errors
Message-ID: <Jan.10.15.56.46.1989.1362@ron.rutgers.edu>
Date: 10 Jan 89 20:56:48 GMT
References: <597@mks.UUCP> <3845@pt.cs.cmu.edu> <18123@santra.UUCP> <5178@lynx.UUCP> <492@babbage.acc.virginia.edu>
Organization: Rutgers Univ., New Brunswick, N.J.
Lines: 11

> presumaby you expect C to detect arithmetic overflow as well.  > and
subscript errors.  > and invalid pointers > and ...  Nope, I expect it to do
what the spec says.  The spec says that subscripting out of bounds and
overflow is undefined.  It also says that pointer subtraction works.  The
Intel C compiler seems not to conform to the spec, so it isn't C.

> C is not a high level language.
Untrue.
> C is not a language for the average programmer.
We're not talking average programmers.  Were talking about people who know how C is
supposed to work and have been using it on other machines.