Bug in polinterpolate? (A12-140)

[Bill Daly <bill@wmdaly.com>]

I uploaded a test file to http://wmdaly.com/math/err.gp , because it is 
a bit too big to include in a post. Essentially, I have a vector (qv) of 
25 polynomials, each of which is of degree 120 with fairly large 
coefficients in Z, and when I invoke

 q1 = polinterpolate(qv,,a,&err)

the result is exact (naturally), specifically a polynomial of degree 120 
in x whose coefficients are rational functions of a, but the error 
returned is not zero, as I would have expected. Am I misunderstanding 
something? (I get the same result if I call

 q1 = polinterpolate(vector(25,j,j),qv,a,&err)

instead.) (I ran it with a stacksize of 250M and the default precision 
of 28, but I don't know if that is relevant.)


Regards, Bill