bug in bessel_j.c

[Gerard Jungman]

Thomas Walter wrote:
>
> I think there is a bug in 'specfunc/bessel_j.c' and friends.
> Calling 'bessel_jl_impl (0,0)' should return '1' like 'bessel_j0_impl
> (0)' does.

Oops. I just fixed it in CVS.


> PS:
> What about adding incomplete beta function?

Not likely to happen soon. In the meantime, you
could try using the relation to the confluent
hypergeometric function. That might work for you
if the arguments are not in one of the very
difficult limits.


-- 
G. Jungman