`xroot` should return integer results when possible

[c3d]

trafficstars

While analysing the HP50G xroot behaviour, I noticed that xroot(3;8) reutrns 2, with type 28 (i.e. integer), whereas DB48X returns a decimal value in that case.

[c3d]

Interestingly, this has limits.

xroot(3;8) then EVAL shows 2.

xroot(3;216) then EVAL still shows xroot(3;216), and you need ->NUM to get a numerical value, which is decimal 6.. So it looks like there is some kind of factorization happening for small values up to 5, but that it does not work after that.

The idea that it is a factorization algorithm is reinforced by the observation that xroot(3;64) gives 2^2 :-)