Re: ellisoncurve() bug?

[Karim BELABAS <Karim.Belabas@math.u-psud.fr>]

[Michael Somos:]
> gp> E=ellinit([0,0,1,-1,0]);
> gp> ellisoncurve(E,ellztopoint(E,.1))
> %2 = 0
> 
> Looks like a bug to me, but I could be wrong. Shalom, Michael

This "bug" has a long history... The problem is to give a sensible meaning to
the question `is the point on the curve ?' when the point has inexact
coordinates. ellisoncurve evaluates the defining polynomial on the given
point, using some trivial heuristics to decide how small the result should
be. The patch below modifies the heuristics; it should accept many more
points, perhaps a bit too many...