Re: [PATCH] Making double precision constants work with -fsingle-precision-constant [was: Re: current subversion (rev 115032) compiling issues]

[Marco Trudel <mtrudel at gmx dot ch>]

Problem solved by updating the gcc I use to build the current subversion 
version to 4.2.0.

regards
Marco


Carlos O'Donell wrote:
> On Wed, Jun 28, 2006 at 11:02:00AM +0200, Marco Trudel wrote:
> > Your patch http://gcc.gnu.org/ml/gcc-patches/2005-12/msg00790.html seems 
> > to break cross-compiling for windows, see attached forwarded message.
>
> > > Ok, next Problem:
> > > $GCC_BUILD_DIR/i686-pc-mingw32/libstdc++-v3/include/limits:1011: error: 
> > > '__FLT_HAS_DENORM__' was not declared in this scope
> > > $GCC_BUILD_DIR/i686-pc-mingw32/libstdc++-v3/include/limits:1068: error: 
> > > '__DBL_HAS_DENORM__' was not declared in this scope
> > > $GCC_BUILD_DIR/i686-pc-mingw32/libstdc++-v3/include/limits:1125: error: 
> > > '__LDBL_HAS_DENORM__' was not declared in this scope
> > >
> > > Here I'm stuck because I'm unable to find the defining header or source 
> > > file for these constants... Any ideas?
>
> The values of these definitions are computed by the compiler in gcc/c-cppbuiltins.c
> (c_cpp_builtins).
>
> Can you provide preprocessed source without macro expansion? I'm curious
> why a Windows hosted toolchain doesn't call c_cpp_builtins.
>
> Cheers,
> Carlos.