Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!usc!snorkelwacker.mit.edu!bloom-picayune.mit.edu!athena.mit.edu!jfc
From: jfc@athena.mit.edu (John F Carr)
Newsgroups: comp.std.c
Subject: Re: Definition of "conforming"
Message-ID: <1991Apr4.011628.1983@athena.mit.edu>
Date: 4 Apr 91 01:16:28 GMT
References: <18160@crdgw1.crd.ge.com> <15680@smoke.brl.mil>
Sender: news@athena.mit.edu (News system)
Organization: Massachusetts Institute of Technology
Lines: 15

In article <15680@smoke.brl.mil> gwyn@smoke.brl.mil (Doug Gwyn) writes:

>??  That's a strictly conforming program, so every conforming
>implementation will accept it.

The only requirement I remember from the standard is that a single program
with certain properties be compiled correctly.  I think there is a comment
somewhere in the book (possibly not in the standard itself) saying that ANSI
did not want to write an ANSI C verification package.  If a conforming
implementation must do the right thing with "hello world" but is allowed to
fail on a complex program, that implies a restriction on implementation
limits that I don't remember reading.

--
    John Carr (jfc@athena.mit.edu)