Path: utzoo!attcan!utgpu!news-server.csri.toronto.edu!clyde.concordia.ca!uunet!cs.utexas.edu!usc!apple!portal!motcad!jtc
From: jtc@motcad.portal.com (J.T. Conklin)
Newsgroups: comp.unix.xenix
Subject: Re: XENIX 2.3.2 include files...
Message-ID: <1990Aug19.191911.3317@motcad.portal.com>
Date: 19 Aug 90 19:19:11 GMT
References: <1990Aug18.091631.19939@dhw68k.cts.com> <1536@sixhub.UUCP>
Distribution: na
Organization: CADnet Inc., San Ramon CA
Lines: 43

In article <1536@sixhub.UUCP> davidsen@sixhub.UUCP (bill davidsen) writes:
>The gcc compiler compiles gcc, not C.

Just as the Microsoft compiler compiles "Microsoft C", and the pcc
derived compilers compile "PCC C".  I am unaware of any C compiler
which everyone will agree is the "One True C Compiler".

In fact, it could be argued that gcc will eclipse pcc as the de facto
standard C compiler as vendors abandon pcc due to poor performance, etc.

>It is useful for compiling some
>net programs which make heavy use of the "extensions" to the language.

I haven't seen any "net programs" that really use the gcc extensions.
The only extension I use myself is the function inlining directive, but
I always disable it with a:

	#ifndef __GNUC__
	#define inline
	#endif

if I'm not compiling with gcc.

>The Xenix C compiler often won't compile some vaild code, and gcc is
>useful there, too. In UNIX you have rcc (real cc) and can compile most
>things with that.

I use gcc on SCO Xenix/UNIX for the following reasons:

	* It generates correct code.

	* It understands ANSI C function prototypes.

Niether compiler from SCO meets both criteria.


	--jtc

-- 
J.T. Conklin	UniFax Communications, Inc.

		CADnet Inc, San Ramon California
		jtc@motcad.portal.com, ...!portal!motcad!jtc