Re: Soft float support and casting

2 msg	Failed build of AVR cross-compiler with combine...
2 msg	unwind problem when cross compiling
1 msg	voli low cost
1 msg	fluupe

Soft float support and casting
\ Simon Kagstrom (29 Nov 2006)
. \ John Love-Jensen (29 Nov 2006)
. \ Michael Eager (29 Nov 2006)
. . \ Simon Kagstrom (29 Nov 2006)

1 msg	compiling errors with options '-fprofile-arcs -...
8 msg	Unclear documentation on building GCC together ...
8 msg	alloca attribute?
3 msg	changes made to gmon-sol2.c not reflected after...
2 msg	Compiling error when changing to gcc4.1.1
1 msg	enums, switches, and warnings?
1 msg	Exception stack too large
5 msg	prefetching on pentium 4
2 msg	How, i can work with read-rtl.c
5 msg	gcc for ppc
2 msg	Different versions created executable's compati...
2 msg	Easy linker problem
2 msg	A bug with variables and functions of the same ...
1 msg	Tracking operator ==
2 msg	gcc 3.0 and linux kernel

Subject:	Re: Soft float support and casting
Group:	Gcc-help
From:	Simon Kagstrom
Date:	29 Nov 2006

At Wed, 29 Nov 2006 04:43:06 -0800,
Michael Eager wrote:
>
> Simon Kagstrom wrote:
>
> > I've looked in libgcc2.c and am unfortunately a bit unsure about how
> > to get GCC to directly cast an integer value to a float type. For
> > example, I would like to do something like this:
> >
> > float __floatsisf(long i)
> > {
> > return (float)0x40c00000; /* float value 6.0 */
> > }
> >
> > but the cast will of course turn this into some completely different
> > number. Can someone quickly explain how to do this?
>
> Use a union:
>
> union { float f; unsigned int i; } u;
>
> float __floatsisf(long i)
> {
> u.i = 0x40c00000;
> return u.f;
> }

Ah, of course - less magic than I was expecting. :-)

Thanks,
// Simon