GMP & MPFR think I'm cross compiling!

Gilles Espinasse g.esp at
Sun Nov 23 15:35:16 PST 2008

----- Original Message ----- 
From: "Matthew Burgess" <matthew at>
To: <lfs-dev at>
Sent: Sunday, November 23, 2008 7:16 PM
Subject: GMP & MPFR think I'm cross compiling!

> Hi all,
> I'm trying to compile LFS trunk + local package update patches under a
Windows host, Kubuntu-8.10 guest Virtualbox environment using jhalfs trunk.
> I see the following output in chapter 5's GCC build logs:
> checking build system type... i686-pc-linux-gnu
> checking host system type... none-pc-linux-gnu
> ...
> checking whether we are cross compiling... yes
> ...
> checking for vsnprintf... yes
> checking whether vsnprintf works... probably
> configure: WARNING: cannot check for properly working vsnprintf when cross
compiling, will assume it's ok
> This occurs under the configure runs for all 3 stages of GMP & MPFR, but
not for any of the other configure runs.
> Whilst I'm not overly concerned by this, as it's going to assume that
vsnprintf works correctly, I wonder whether this mis-detection of
cross-compiling will impact on anything else in the build.
> My 'configure'/autoconf foo has gone as far as figuring out that it thinks
I'm cross compiling because the build system type and host system type are
different.  I couldn't, though, figure out how it works out what my host
system type is.  If anyone could point me in the right direction, I'd
appreciate it.  I'd rather fix up the configure script rather than override
its detection using --host, as to me this is an obvious bug given that the
other areas in the tree correctly detect the host type.
Just look in configure.log, you will see at wich line of configure the test
is made.
configure:1654: checking build system type
configure:1672: result: athlon-pc-linux-gnu
configure:1680: checking host system type
configure:1694: result: athlon-pc-linux-gnu

Then you could go looking in those configure lines.
In short, build should come from config.guess and host should come from


More information about the lfs-dev mailing list