>>>>> "Stavros" == Stavros Macrakis <macrakis at alum.mit.edu> writes:
Stavros> In Maxima 5.25.1 GCL 2.6.8 on Windows, 1000! + 0.1 does
Stavros> print??1..#INF00e+000E+10, so it doesn't seem to be a
Stavros> GCL-specific problem. ?But then, 1.#INF00e+000E+10 is a
Stavros> ridiculous way to print the value, and not a first-class
Stavros> Maxima object.
Stavros> But in any case, I would consider any of these results to
Stavros> be bugs in the Maxima implementation, considered as a
Stavros> black box. ?The user doesn't care whether it's Clisp,
Stavros> GCL, SBCL, etc., just that Maxima is not producing a
Stavros> useful result or a meaningful error message (not a Lisp
Stavros> error message).
While I agree with this sentiment, the question is how much effort
should be expended in getting maxima to handle gcl's quirks. I think
we consider it a defect in gcl that there's no way to get gcl to
signal an error on overflow.
On the other hand, I think there are some very recent changes in gcl
that will allow gcl to signal such errors and then we won't have to do
anything too special for this.
Ray