From owner-freebsd-bugs  Thu Jul 25 16:20:06 1996
Return-Path: owner-bugs
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.5/8.7.3) id QAA09656
          for bugs-outgoing; Thu, 25 Jul 1996 16:20:06 -0700 (PDT)
Received: (from gnats@localhost)
          by freefall.freebsd.org (8.7.5/8.7.3) id QAA09632;
          Thu, 25 Jul 1996 16:20:04 -0700 (PDT)
Date: Thu, 25 Jul 1996 16:20:04 -0700 (PDT)
Message-Id: <199607252320.QAA09632@freefall.freebsd.org>
To: freebsd-bugs
Cc: 
From: "Andrew L. Moore" <alm@woops.slewsys.org>
Subject: Re: ports/1416: cflow(1) doesn't parse GNU C __attribute__ syntax 
Reply-To: "Andrew L. Moore" <alm@woops.slewsys.org>
Sender: owner-bugs@FreeBSD.ORG
X-Loop: FreeBSD.org
Precedence: bulk

The following reply was made to PR ports/1416; it has been noted by GNATS.

From: "Andrew L. Moore" <alm@woops.slewsys.org>
To: Bill Fenner <fenner@parc.xerox.com>
Cc: FreeBSD-gnats-submit@freebsd.org
Subject: Re: ports/1416: cflow(1) doesn't parse GNU C __attribute__ syntax 
Date: Thu, 25 Jul 1996 16:16:41 -0700

 The macro `-D__attribute__(x)=' is still needed in cflow(1), whether the
 `-g' option is specified or not.  Someone added __attribute__ to the
 gnu_keywords list in prcc.c.  This is wrong, since __attribute__ defines
 not just a new C type, but a new C grammar as well.
 
 For portability, I suggest the following.
 >From the end-user's standpoint, it would be nice if `cflow *.c' "just worked."
 Maybe borrow from the GNU autoconf `configure' script:
 
 cat >${TMPFILE}.c <<\EOF
 #ifdef __GNUC__
   yes;
 #endif
 EOF
 if  ${CPP} ${TMPFILE}.c | egrep yes >/dev/null 2>&1; then
 	CPP="${CPP} -D__attribute__(x)="
 fi
 rm -f ${TMPFILE}.c