This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Results for 2.97 20010212 (experimental) testsuite on i686-pc -cyg win
- To: DJ Delorie <dj at redhat dot com>
- Subject: Re: Results for 2.97 20010212 (experimental) testsuite on i686-pc -cyg win
- From: Chris Faylor <cgf at redhat dot com>
- Date: Wed, 28 Feb 2001 20:05:59 -0500
- Cc: neil at daikokuya dot demon dot co dot uk, David dot Billinghurst at riotinto dot com, gcc-bugs at gcc dot gnu dot org, gcc-patches at gcc dot gnu dot org
- References: <8D00C32549556B4E977F81DBC24E985DC6E2@crtsmail1.technol_exch.corp.riotinto.org> <20010228074833.C28670@daikokuya.demon.co.uk> <20010228195131.D6418@redhat.com> <200103010057.TAA30711@greed.delorie.com>
On Wed, Feb 28, 2001 at 07:57:18PM -0500, DJ Delorie wrote:
>
>The problem is that gcc supresses certain warnings in system headers,
>but -idirafter doesn't mark things as system headers whereas -isystem
>does. So, if you use (say) "gcc -ansi" it complains about headers in
>/usr/include where it shouldn't.
Hmm. I wondered about that when I read the description of -idirafter.
Then I don't think using an -isystem is correct either. This will cause
the include order to be:
/usr/local/include
/usr/include
/usr/lib/gcc-lib/i686-pc-cygwin/2.97/include
rather than
/usr/local/include
/usr/lib/gcc-lib/i686-pc-cygwin/2.97/include
/usr/include
The intent of most of my recent changes to cygwin.h has been to make
gcc search the right directories when given options like -mno-cygwin
or -mno-win32. In the -mno-cygwin case we never want to search
/usr/include, so it can't just be a STANDARD_INCLUDE_DIR.
cgf