-----maxima-bounces at math.utexas.edu wrote: -----
>When designing possible overflow detection
>prior to calling quadpack in a wrapper package,
>I have the problem that bessel functions don't
>have the same overflow return as core math
>functions like exp.
What happens with floating point overflow depends on the version of CL used to compile Maxima.
Example: with Clozure CL
(%i1) exp(800.0);
Maxima encountered a Lisp error: FLOATING-POINT-OVERFLOW detected
Every Lisp error reported by Maxima is at least a venial bug (I think). Likely, making floating
point overflow work the same with all CLs would require a great deal work:
$ wx86cl64
Welcome to Clozure Common Lisp Version 1.7-r14925M (WindowsX8664)!
? (ignore-errors (exp 800.0))
NIL
#<FLOATING-POINT-OVERFLOW #x210054B8CD>
?
But GCL gives
(%i52) :lisp(exp 800.0)
#<1.#INF00e+000>
--Barton