Re: A patch for ia64

[Jakub Jelinek <jakub at redhat dot com>]

On Tue, Feb 13, 2001 at 10:26:40PM -0800, H . J . Lu wrote:
> On Tue, Feb 13, 2001 at 09:27:53PM -0800, Ulrich Drepper wrote:
> > "H . J . Lu" <hjl@valinux.com> writes:
> > 
> > > We need to check j0l () with 2.0 to verify if it is implemented on
> > > ia64.
> > 
> > It is implemented and therefore it shouldn't matter.  What is more of
> > a concern are the results I see.  But this is with the old GnuPro
> > compiler.
> > 
> 
> It is not implemented:
> 
> # gcc m.c /home/work/build/glibc/math/libm.a 
> /home/work/build/glibc/math/libm.a(s_cosl.o): In function `cosl':
> /work/gnu/src/glibc/math/../sysdeps/ieee754/ldbl-96/s_cosl.c:78: warning: __ieee754_rem_pio2l is not implemented and will always fail
> /work/gnu/src/glibc/math/../sysdeps/ieee754/ldbl-96/s_cosl.c:71: warning: __kernel_cosl is not implemented and will always fail
> /work/gnu/src/glibc/math/../sysdeps/ieee754/ldbl-96/s_cosl.c:81: warning: __kernel_sinl is not implemented and will always fail
> 
> j0l is implemented with something which is not implemented on ia64.
> j0l (0.0) is a special case so that it won't return an error.

Duh, it looks like ia64 long double libm support is even worse than on
sparc64. I might try to look into at least XFmode sinl/cosl/expl in the next
week or two (since I already did that for TFmode) if I find my gmp programs
I was doing it with.

	Jakub