Hi Barton.
Thanks for your follow-up test with "cl::real".
Looks like GCL 2.6.5 (the stable release I tried yesterday) has problems
with PCL and ANSI, as following on from your result I took CVS Maxima
(using the Maxima-local batch file) built with CVS GCL and tried out
Rays' original code (without the package manipulation I suggested) ie:
1. to_lisp()
2. Cut and paste Ray's original Lisp CLOS code
3. (to-maxima)
4. Cut and paste Ray's origional Maxima test code and I'm pleased to say
that modulo some compiler noise, it worked as below.
I may not be subscribed to the Maxima list from this email address so it
may take a while for this message to get through there.
Cheers
Mike Thomas.
JACOBI-SN-BIGFLOAT
MAXIMA> (to-maxima)
Returning to Maxima
(%o1) true
(%i2) fpprec:50;
(%o2) 50
(%i3) jacobi_sn(x,1);
(%o3) tanh(x)
(%i4) tanh(0.5b0);
(%o4) 4.6211715726000975850231848364367254873028928033011b-1
(%i5) jacobi_sn(0.5b0,0.99999999999999999999999999999b0),bfloat;
Compiling ./gazonk0.lsp.
End of Pass 1.
End of Pass 2.
OPTIMIZE levels: Safety=2, Space=3, Speed=2
Finished compiling ./gazonk0.lsp.
(%o5) 4.6211715726000975850231848364384478195723189546368b-1
(%i6)