roots

[Kevin Ryde]

Jason Moxham <J.L.Moxham@maths.soton.ac.uk> writes:
>
> There is quite a bit of code

Are you sure that level of complexity is needed?  It's probably not
necessary to squeeze every last bit out of the precisions used and
stuff, since the real work is done in limb sized chunks anyway.

> perfect_power_testing (definitely)

That stuff could do with a revision.  Some residue tests in
particular, in a fashion similar to what I recently added to
mpn/generic/perfsqr.c would be a start.