Cross compiler problem (xgcc does not find standard includes)

Cross compiler problem (xgcc does not find standard includes)

[Martin Kahlert]

Hi,
i have this problem:
i configured by
../egcs-19990629/configure --target=i686-pc-mingw32

and built on linux x86 with

make CFLAGS='-O3 -fomit-frame-pointer' LIBCFLAGS='-O3 -fomit-frame-pointer' LIBCXXFLAGS='-O3 -fno-implicit-templates -fomit-frame-pointer' cross

and get this:
/home/kahlert/LOCAL/obj/gcc/xgcc -B/home/kahlert/LOCAL/obj/gcc/
                 -B/usr/local/i686-pc-mingw32/bin/ -c -DHAVE_CONFIG_H -O3
		 -fomit-frame-pointer -I. -I../../../egcs-19990629/libiberty/../include 
		 ../../../egcs-19990629/libiberty/choose-temp.c
../../../egcs-19990629/libiberty/choose-temp.c:29: stdio.h: No such file or directory
../../../egcs-19990629/libiberty/choose-temp.c:30: sys/types.h: No such file or directory
../../../egcs-19990629/libiberty/choose-temp.c:32: unistd.h: No such file or directory
../../../egcs-19990629/libiberty/choose-temp.c:35: stdlib.h: No such file or directory
../../../egcs-19990629/libiberty/choose-temp.c:38: sys/file.h: No such file or directory
make[2]: *** [choose-temp.o] Error 1
make[2]: Leaving directory `/sw/local_l/kahlert/obj/i686-pc-mingw32/libiberty'
make[1]: *** [all-target-libiberty] Error 2
make[1]: Leaving directory `/sw/local_l/kahlert/obj'
make: *** [cross] Error 2                                                               

I assume the headers should be the one for the host. So they should be found in the standard 
directories.

What do i make wrong?

Thanks for every hint,
Martin.
>From law@cygnus.com Wed Jun 30 03:56:00 1999
From: Jeffrey A Law <law@cygnus.com>
To: Martin Kahlert <martin.kahlert@mchp.siemens.de>
Cc: egcs-bugs@egcs.cygnus.com
Subject: Re: Cross compiler problem (xgcc does not find standard includes) 
Date: Wed, 30 Jun 1999 03:56:00 -0000
Message-id: <24121.930740082@upchuck.cygnus.com>
References: <199906301033.MAA00950@keksy.mchp.siemens.de>
X-SW-Source: 1999-06/msg00999.html
Content-length: 289

  In message < 199906301033.MAA00950@keksy.mchp.siemens.de >you write:
  > I assume the headers should be the one for the host. So they should be
  > found in the standard directories.
No.  You are trying to build the target libiberty.  Thus you need include files
for your target.

jeff

Re: Cross compiler problem (xgcc does not find standard includes)

[Jeffrey A Law]

  In message < 199906301033.MAA00950@keksy.mchp.siemens.de >you write:
  > I assume the headers should be the one for the host. So they should be
  > found in the standard directories.
No.  You are trying to build the target libiberty.  Thus you need include files
for your target.

jeff

Re: Cross compiler problem (xgcc does not find standard includes)

[Mumit Khan]

Martin,

As Jeff says, you need the target includes (and libraries if you want
actually link!).

See my howto for building cross and canadian cross tools for Cygwin/Mingw
at:
  
  http://www.xraylith.wisc.edu/~khan/software/gnu-win32/
  http://www.delorie.com/howto/

I take a slightly different approach than the crossgcc folks, but I find
it easier and trivially scriptable.

Regards,
Mumit

Re: Cross compiler problem (xgcc does not find standard includes)

[Martin Kahlert]

Quoting Mumit Khan (khan@xraylith.wisc.EDU):
> Martin,
> 
> As Jeff says, you need the target includes (and libraries if you want
> actually link!).
> 
> See my howto for building cross and canadian cross tools for Cygwin/Mingw
> at:
>   
>   http://www.xraylith.wisc.edu/~khan/software/gnu-win32/
>   http://www.delorie.com/howto/
> 
> I take a slightly different approach than the crossgcc folks, but I find
> it easier and trivially scriptable.

Thanks for your reply,
as i wrote Jeff in a private mail, these includes/libs were actually copied to their
position before (as indicated by the cross-FAQ).

I think, my problem was LANGUAGES=... instead of --enable-languages=...
After i changed that, the compilation passed through.
I am no struggling with cross building f77.
(with egcs-1.1.2 this was much simpler i think).

Thanks again,
Martin.
 
-- 
esa$ gcc -Wall -o ariane5 ariane5.c
ariane5.c: 666: warning: long float implicitly truncated to unsigned type
esa$ ariane5
>From law@cygnus.com Thu Jul 01 01:36:00 1999
From: Jeffrey A Law <law@cygnus.com>
To: "Kaveh R. Ghazi" <ghazi@caip.rutgers.edu>
Cc: egcs-bugs@egcs.cygnus.com
Subject: Re: gcc-2.95 extra testsuite failures when built w/-O3 -funroll-all-loops
Date: Thu, 01 Jul 1999 01:36:00 -0000
Message-id: <28426.930818099@upchuck.cygnus.com>
References: <199906281343.JAA03000@caip.rutgers.edu>
X-SW-Source: 1999-07/msg00007.html
Content-length: 568

  In message <199906281343.JAA03000@caip.rutgers.edu>you write:
  > 	When I bootstrap the 2.95 prerelease on solaris2 or OSF4 using
  > -g -O3 -funroll-all-loops, I get a successful three stage bootstrap
  > and compare, but I see ~60 extra g++ failures beyond what a normal
  > bootstrap would produce.  They appear to mostly deal with exception
  > handling.  They can be seen in:
Yes.  We currently miscompile some of the EH code in libgcc.a at -O3.  As far
as I know, nobody's actually analyzed the problem any deeper than that.

Want to take a stab at it?

jeff