Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!allegra!mit-eddie!genrad!panda!talcott!harvard!seismo!ll-xn!cit-vax!jon
From: jon@cit-vax.Caltech.Edu (Jonathan P. Leech)
Newsgroups: net.lang.c++
Subject: System V preprocessor problem with <generic.h>
Message-ID: <410@cit-vax.Caltech.Edu>
Date: Fri, 25-Apr-86 00:56:35 EDT
Article-I.D.: cit-vax.410
Posted: Fri Apr 25 00:56:35 1986
Date-Received: Sun, 27-Apr-86 07:24:22 EDT
Reply-To: jon@cit-vax.UUCP (Jonathan P. Leech)
Distribution: net
Organization: California Institute of Technology
Lines: 35
Summary: Ryserisms aren't in System V /lib/cpp

Expires:
Sender:
Followup-To:

Organization : California Institute of Technology
Keywords:

    The <generic.h> header file defines constructs of the form

	#define name2(a,b) a\
	b
	#define name3(a,b,c) a\
	b\
	c

    with the intent of producing

	name2(a,b) => ab    etc.

    unfortunately the  System  V  cpp  (at  least  the	HP  and  ELXSI
versions) instead produces

	name2(a,b) => a b

    this is a serious problem in  every  case  I  can  find  of  these
constructs  being  used  -  specifically  <vector.h>.  I  don't  think
there's any way to produce this behavior with the System V cpp.

    Combined with the problem of '//' comments	not  being  elided  by
cpp, I think C++ should come with its own  preprocessor  (or  have  it
integrated into cfront).

    Additional	amusement:  try  running  cpp  with  input  from  your
keyboard and #include <generic.h>. Note that it now takes > 1  EOF  to
get out of cpp!
-- 
    Jon Leech (jon@csvax.caltech.edu || ...seismo!cit-vax!jon)
    Caltech Computer Science Graphics Group
    __@/