number printing problems in 5.9.3 xmaxima onwindows/GCL



> Greetings, and thanks Mike!  Just to catch up on the thread, 
> you need to (setq si::*print-nans* t) to print these numbers. 
>  Apparently, the result is libc version dependent (comes from 
> sprintf), and specifically differs on Linux from Windows.

Aha!!!  Yes, Stavros' result is correct for 2.6.8pre and I suppose you
are correct:

===============================
>(setq si::*print-nans* t)

T

>(* 1e300 1e300)

#<1.#INF00e+000>
===============================

MinGW32 C runtime uses an underlying Microsoft library (the Visual C++
runtime library) for these kinds of things.  WHile it may be possible to
alter the result by fiddling with compiler switches related to floating
point on gcc, I think you are unlikely ever to replicate the Linux libc
results.