Float assignment and conversion to float

[Gilles Rayrat]

Hi Rob, and many thanks for the answer...
Please forget my first question: just explaining better the problem to 
you made me find my answer ... :) sorry for that.
But FYI, you did get the question right and there is no miracle 
solution, just use pow and div

For the second one, I have a concern about performance in down-casting 
the result of get_d() into a float: why doing two operations when you 
can save one ?
But this would lead to changing the gmp lib, and I cannot do that for a 
matter of licensing - and time...

Thanks again
Gilles.