> On Tue, 18 Aug 1998, Richard Earnshaw wrote:
> 
> : The compiler cannot compile libc's softfloat.c in the way that it was 
> : written (it exploited a bug in gcc 2.7 to return a structure that was 
> : larger than 32 bits in a register) -- I'm working on a fix for this, but 
> : haven't completed it yet.
> 
> Hm.  The suggestion about "unsigned long long" may be useful; of course, the
> problem there is knowing which 32 bits are the high and low ones (though you
> could make the assumption, since the software is in libc/arch/arm32).
> 

This will work, but my experiments show that the code will be less 
efficient that the code generated by the old 2.7 compiler.  I'm currently 
working on a patch that would give us something like the Norcroft 
compiler's "return_in_regs" attribute, but it won't be in the official 
egcs 1.1 releases since these are now feature-frozen.

Richard.