sinl/cosl for ldbl-96

[Jakub Jelinek]

On Fri, Mar 16, 2012 at 10:56:18AM +0100, Andreas Jaeger wrote:
> Wow, thanks a lot! I would like to see Jakub review the ported
> routines as he's the original author - to me it looks fine,

I'm afraid I don't remember the details after 13 years of writing this,
so I'll trust Joseph if he analysed it now.
Unfortunately I didn't take notes on how the indices were computed, usually
those were some short gmp proglets if I remember it at all, and the
implementation is also not very close to .5ulp precision.

	Jakub