From: Peter Broadbery <peterb>
Date: Thu, 2 May 96 10:51:50 BST
Received: from co.uk (nags8) by nags2.nag.co.uk (4.1/UK-2.1)
	id AA05483; Thu, 2 May 96 10:51:50 BST
To: ax-bugs

Subject: fixbug
By: PAB
Fixed: bug1076.as
--* Received: from mailer.scri.fsu.edu by nags2.nag.co.uk (4.1/UK-2.1)
--* 	id AA02276; Wed, 1 May 96 05:30:35 BST
--* Received: from ibm4.scri.fsu.edu (ibm4.scri.fsu.edu [144.174.131.4]) by mailer.scri.fsu.edu (8.6.12/8.6.12) with SMTP id AAA03129; Wed, 1 May 1996 00:26:55 -0400
--* From: Tony Kennedy <adk@scri.fsu.edu>
--* Received: by ibm4.scri.fsu.edu (5.67b) id AA38306; Wed, 1 May 1996 00:25:09 -0400
--* Date: Wed, 1 May 1996 00:25:09 -0400
--* Message-Id: <199605010425.AA38306@ibm4.scri.fsu.edu>
--* To: adk@scri.fsu.edu, ax-bugs, edwards@scri.fsu.edu
--* Subject: [3] bug (and fix) in power$BinaryPowering

--@ Bug Number:  bug1076.as 
--@ Fixed  by:  PAB   
--@ Tested by:  pow.as 
--@ Summary:    Fixed as you suggested (thanks). 

-- Command line: none
-- Version: 1.1.5
-- Original bug file name: /dev/null

--+ The following code fragment occurs in `bpower.as' in `libaxllib' for
--+ AXIOM-XL version 1.1.5:
--+ 
--+ 	power(u: T, x: T, n: E): T == {
--+ 		(n < 0) =>
--+ 			if T has with { inv: % -> % } then
--+ 				inv power(u, x, -n)
--+ 			else
--+ 				error "power:  negative exponent";
--+ 
--+ Now,  u * x^n = u * inv(x)^(-n) ~= inv(u * x^(-n)), so the fourth line
--+ should be
--+ 
--+ 				power(u, inv x, -n)
--+ 
--+ rather than
--+ 
--+ 				inv power(u, x, -n)