Re: [Gcl-devel] Re: [Axiom-developer] hascategory bug

[Camm Maguire]

Greetings again!  OK here it is:

--- ../../../../../axiom2/new/new/src/interp/category.boot.pamphlet     
2003-05-05 03:14:25.000000000 +0000
+++ category.boot.pamphlet      2003-07-26 05:43:49.000000000 +0000
@@ -471,7 +471,7 @@
         n:= SIZE $NewCatVec
         FundamentalAncestors:= [[b.(0),condition,n],:FundamentalAncestors]
         $NewCatVec:= LENGTHENVEC($NewCatVec,n+1)
-        copied:= true
+        copied:= false
         originalvector:= false
         $NewCatVec.n:= b.(0)
   if not copied then $NewCatVec:= COPY_-SEQ $NewCatVec


At least in GCL, the code for lengthenvec need not copy the vec to a
new location:

(macros.lisp)

(defun lengthenvec (v n)
  (if (adjustable-array-p v) (adjust-array v n)
    (replace (make-array n) v)))

In this case, the array is adjustable, and again at least in GCL, the
adjust-array need not and in this case does not do a copy.

I think I just compiled xpoly ok.  Trying now a full algebra build.

Take care,

root <address@hidden> writes:

> Man, I'm impressed. I've been chasing this bug for a while with 
> no real success. I'm going to look at the same code and see what
> I can find.
> 
> Tim
> 
> 
> _______________________________________________
> Gcl-devel mailing list
> address@hidden
> http://mail.gnu.org/mailman/listinfo/gcl-devel
> 
> 
> 

-- 
Camm Maguire                                            address@hidden
==========================================================================
"The earth is but one country, and mankind its citizens."  --  Baha'u'llah