From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 21463 invoked by alias); 15 Mar 2003 23:56:01 -0000 Mailing-List: contact gcc-prs-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Archive: List-Post: List-Help: Sender: gcc-prs-owner@gcc.gnu.org Received: (qmail 21447 invoked by uid 71); 15 Mar 2003 23:56:00 -0000 Date: Sat, 15 Mar 2003 23:56:00 -0000 Message-ID: <20030315235600.21445.qmail@sources.redhat.com> To: nobody@gcc.gnu.org Cc: gcc-prs@gcc.gnu.org, From: Sean McNeil Subject: Re: c++/10069: -include switch is improperly handled Reply-To: Sean McNeil X-SW-Source: 2003-03/txt/msg01044.txt.bz2 List-Id: The following reply was made to PR c++/10069; it has been noted by GNATS. From: Sean McNeil To: Neil Booth Cc: Steven Bosscher , gcc-gnats@gcc.gnu.org, gcc-bugs@gcc.gnu.org, nobody@gcc.gnu.org, gcc-prs@gcc.gnu.org Subject: Re: c++/10069: -include switch is improperly handled Date: 15 Mar 2003 15:54:54 -0800 It would appear that this issue has been addressed somewhere else. It was a bug in how the cc1plus executable handled switches. That is why I gave the patch to cp/g++spec.c that recognized the switch -include as taking an argument. I'm not exactly sure how it was addressed. I'm suspicious that this problem may come back later. It is peculiar that c++spec.c has to know about switches that take arguments at all. I'm talking about the check: else if (((argv[i][2] == '\0' && (char *)strchr ("bBVDUoeTuIYmLiA", argv[i][1]) != NULL) || strcmp (argv[i], "-Xlinker") == 0 || strcmp (argv[i], "-Tdata") == 0)) Perhaps the cpp initialization now pulls the -include arguments off the list? Sean On Sat, 2003-03-15 at 13:52, Neil Booth wrote: > Sean McNeil wrote:- > > > No, I am not using PCH. I'm trying to use the --include switch but it > > isn't recognized by C++ as a switch with 2 arguments and so it gets > > improperly handled. It ends up reordering my switches and passes -O2 as > > the file to include and tries to compile the include file as an input > > file. > > By "C++" do you mean the g++ driver? That may well be a bug. If so, > the place to look is cp/lang-specs.h. > > > So I observed that > > > > gcc -include hack.h -O2 test.cc > > > > would end up as > > > > switches: > > -include -O2 > > > > input files: > > hack.h test.cc > > > > thus it would fail because it cannot find the file "-O2" and because it > > would precompile hack.h into hack.pch. > > LOL. Though if you're not using PCH I don't understand your comment. > > Neil.