[Maxima-commits] CVS: maxima/src comm.lisp,1.47,1.48

[Raymond T. <rt...@us...>]

Update of /cvsroot/maxima/maxima/src
In directory sfp-cvsdas-4.v30.ch3.sourceforge.com:/tmp/cvs-serv5154

Modified Files:
	comm.lisp 
Log Message:
Gcl doesn't signal errors on float(log(n)) for large n.  Hence, don't
try to compute that.  Instead, for gcl, just use the more complicated
algorithm(s) for all values of n.


Index: comm.lisp
===================================================================
RCS file: /cvsroot/maxima/maxima/src/comm.lisp,v
retrieving revision 1.47
retrieving revision 1.48
diff -u -d -r1.47 -r1.48
--- comm.lisp	22 May 2010 20:20:03 -0000	1.47
+++ comm.lisp	28 May 2010 02:05:26 -0000	1.48
@@ -1165,11 +1165,17 @@
 	       (list (ncons (caar e)) ($float (cadr e)) (caddr e)))))
 	((and (eq (caar e) '%log)
 	      (complex-number-p (second e) '$ratnump))
+	 ;; Basically we try to compute float(log(x)) as directly as
+	 ;; possible, expecting Lisp to return some error if it can't.
+	 ;; Then we do a more complicated approach to compute the
+	 ;; result.  However, gcl doesn't signal errors in these
+	 ;; cases, so we always use the complicated approach for gcl.
 	 (let ((n (second e)))
 	   (cond ((integerp n)
-		  ;; float(log(int)).  First try to compute (log (float n)).
-		  ;; If that works, we're done.  Otherwise we need to do more.
-		  (to (or (ignore-errors (log (float n)))
+		  ;; float(log(int)).  First try to compute (log
+		  ;; (float n)).  If that works, we're done.
+		  ;; Otherwise we need to do more.  
+		  (to (or #-gcl (ignore-errors (log (float n)))
 			  (let ((m (integer-length n)))
 			    ;; Write n as (n/2^m)*2^m where m is the number of
 			    ;; bits in n.  Then log(n) = log(2^m) + log(n/2^m).
@@ -1180,7 +1186,7 @@
 		 (($ratnump n)
 		  ;; float(log(n/m)) where n and m are integers.  Try computing
 		  ;; it first.  If it fails, compute as log(n) - log(m).
-		  (let ((try (ignore-errors (log (fpcofrat n)))))
+		  (let ((try #-gcl (ignore-errors (log (fpcofrat n)))))
 		    (if try
 			(to try)
 			(sub  ($float `((%log) ,(second n)))
@@ -1189,7 +1195,8 @@
 		  ;; float(log(n+m*%i)).
 		  (let ((re ($realpart n))
 			(im ($imagpart n)))
-		    (to (or (ignore-errors (log (complex (float re)
+		    (to (or #-gcl
+			    (ignore-errors (log (complex (float re)
 							 (float im))))
 			    (let* ((size (max (integer-length re)
 					      (integer-length im)))