Porting troubles

Jan Bottorff janb at pmatrix.com
Sat Jul 18 00:39:08 UTC 1998


At 03:35 PM 7/14/98 +0200, Georg Gollmann wrote:
>So...shouldn´t all variables that hold addresses (OOPs) be declared as
>unsigned instead of int (or am I barking up the wrong tree) ?

Assuming addresses have a bitvalue that matches positive "int" is a real
bad move. I can quickly think of a couple of OS's where this will not
allways be true. The Enterprise version of Windows NT 4.0 uses a 3 GByte
user mode address space, so 1/3 of the addresses are negative int's. Any
process shared memory on Win9x I believe will also be above 2 GB.

Future porting to 64-bit processors may also be a big issue. I believe
Microsoft has said the Intel/HP Merced version of WinNT will use a 32-bit
int, but 64-bit pointers. Welcome back to FAR pointers. For maximum
portability,the VM really should treat pointers and int's as potentially
very different.

A comment was made here that Squeak Smalltalk should be easily portable for
the next 20 years. My guess is in 20 years we will no longer have 32-bit
processors, so how true is this in the context of 64-bit processors? (with
64-bit pointers and some size of int's)? Is there a 64-bit DEC (or are we
supposed to say Compaq now) Alpha version of Squeak right now?

Jan Bottorff, President
Paradigm Matrix, Inc.
___________________________________________________________________
            Paradigm Matrix Inc., San Ramon California
   "video products and development services for Win32 platforms"
Internet: Jan Bottorff janb at pmatrix.com
          WWW          http://www.pmatrix.com
Phone: voice  (925) 803-9318
       fax    (925) 803-9397
PGP: public key  <http://www-swiss.ai.mit.edu/~bal/pks-toplev.html>
     fingerprint  52 CB FF 60 91 25 F9 44  6F 87 23 C9 AB 5D 05 F6
___________________________________________________________________





More information about the Squeak-dev mailing list