From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 22154 invoked by alias); 5 May 2005 17:52:04 -0000 Mailing-List: contact binutils-help@sources.redhat.com; run by ezmlm Precedence: bulk List-Subscribe: List-Archive: List-Post: List-Help: , Sender: binutils-owner@sources.redhat.com Received: (qmail 21836 invoked from network); 5 May 2005 17:51:54 -0000 Received: from unknown (HELO palrel10.hp.com) (156.153.255.245) by sourceware.org with SMTP; 5 May 2005 17:51:54 -0000 Received: from smtp1.ptp.hp.com (smtp1.ptp.hp.com [15.1.28.250]) by palrel10.hp.com (Postfix) with ESMTP id D68A8B5F; Thu, 5 May 2005 10:51:53 -0700 (PDT) Received: from hpsje.cup.hp.com (hpsje.cup.hp.com [15.244.96.221]) by smtp1.ptp.hp.com (Postfix) with ESMTP id AA81E1508BF; Thu, 5 May 2005 17:51:53 +0000 (UTC) Received: (from sje@localhost) by hpsje.cup.hp.com (8.9.3 (PHNE_24419+JAGae58098)/8.7.3 TIS Messaging 5.0) id KAA17543; Thu, 5 May 2005 10:51:53 -0700 (PDT) Date: Thu, 05 May 2005 17:55:00 -0000 From: Steve Ellcey Message-Id: <200505051751.KAA17543@hpsje.cup.hp.com> To: drow@false.org Cc: binutils@sources.redhat.com Subject: Re: Another HP-UX IA64 Build patch In-Reply-To: <20050505173236.GA17480@nevyn.them.org> X-SW-Source: 2005-05/txt/msg00189.txt.bz2 > This is wrong. If your system headers have a conflicting prototype, > you will lose, regardless of what compiler you're using. > > Do you really not have a prototype for basename? > > -- > Daniel Jacobowitz > CodeSourcery, LLC Yes, I have a prototype for basename on IA64 HP-UX. It is in /usr/include/libgen.h: extern char *basename(char *); I see it doesn't have the const modifier that libiberty has. I didn't get a warning or error presumbably because the build didn't include libgen.h. The problem I was trying to avoid is with PA HP-UX. It has the same prototype but if someone is trying to build binutils with the bundled HP-UX compiler, that compiler doesn't handle prototypes. It is an old K&R compiler that is really just intended to be used to rebuild the kernel. The real HP compiler that people can buy is, of course, a full ANSI C compiler. But it looks like we require a full ANSI compiler to build binutils now, I see all the uses of PARAM are gone now so I guess I don't need to worry about that scenario anymore. So should I just add __hpux__ to the list of OS's or is the difference between 'char *' and 'const char *' going to bite me somewhere that I haven't seen yet? Steve Ellcey sje@cup.hp.com