>>>>> "Barton" == Barton Willis <willisb at unk.edu> writes:
Barton> No, it is not. You should report this bug to the Maxima bug list.
Barton> Going from 5.9.2 to 5.9.3, we made many changes to the way
Barton> Maxima does floating point evaluation of the trig/log/exponential
Barton> functions. We (I was involved) messed this one up.
This is fixed with this simple patch for simpexpt in src/simp.lisp:
diff -u -r1.21 simp.lisp
--- simp.lisp 2 May 2006 17:42:10 -0000 1.21
+++ simp.lisp 31 May 2006 16:42:53 -0000
@@ -1597,7 +1597,7 @@
(let ((x ($realpart pot))
(y ($imagpart pot)))
(cond ((and ($bfloatp x) (like 0 y))
- ($bfloat `((mexpt simp) $%e ,pot)))
+ (return ($bfloat `((mexpt simp) $%e ,pot))))
((or ($bfloatp x) ($bfloatp y))
(let ((z (add ($bfloat x) (mul '$%i ($bfloat y)))))
(setq z ($rectform `((mexpt simp) $%e ,z)))
We forgot to return the bfloat value the exponent is a real-valued
bigfloat.
I'll check this in shortly, and add a test for this.
Ray