Subject: Re: amd64 stable for production ?
To: NetBSD Users's Discussion List <netbsd-users@netbsd.org>
From: Greg A. Woods <woods@weird.com>
List: netbsd-users
Date: 12/11/2006 11:43:36
--pgp-sign-Multipart_Mon_Dec_11_11:43:29_2006-1
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: quoted-printable
At Mon, 11 Dec 2006 14:55:56 +0100,
Christian Biere wrote:
>=20
> Stephen Borrill wrote:
> > And to get back on topic, I've not moved to NetBSD/amd64 as I'm getting=
=20
> > coredumps from joe/ncurses.
>=20
> It's a save bet that software will still suffer from 64-bit bugs in 64
> years. If it's not trivial to fix, I'd dump it and call that natural
> selection or even QA.
Indeed. (GNU Emacs works very well on 64-bit platforms! :-))
While I don't run a huge variety of different applications on Alpha, I
must say I've been pleasantly surprised at how well everything I do run
has been supported on this 64-bit platform.
The worst problems that I've ever seen were of course int/long/pointer
confusion, but most of those were ironed out by others as many as 5
years ago or more.
I think we can, at least in part, credit this to the popularity of Tru64
and the work done by its users to port freeware to their platform.
I suspect many problems that remain might have more to do with those
ports being done a little too naively, i.e. being done with #ifdefs that
are specific to Tru64 or Alpha, instead of sizeof(long).
NetBSD's choice to keep int at 32 bits is probably also a big benefit,
even though it may in some other ways be limiting to applications. As
much as I hate a few aspects of ANSI-C, I can also admit that porting
applications to NetBSD/alpha would have been a _LOT_ more difficult
without function prototypes (and a complier that can treat missing
prototypes as an error).
--=20
Greg A. Woods
H:+1 416 218-0098 W:+1 416 489-5852 x122 VE3TCP RoboHack <woods@robohack.ca>
Planix, Inc. <woods@planix.com> Secrets of the Weird <woods@weird.com>
--pgp-sign-Multipart_Mon_Dec_11_11:43:29_2006-1
Content-Type: application/pgp-signature
Content-Transfer-Encoding: 7bit
-----BEGIN PGP SIGNATURE-----
Version: PGPfreeware 5.0i for non-commercial use
MessageID: GPEz7RYVCaBln104DfY3AM4w7oE0vB5o
iQA/AwUBRX2Kt2J7XxTCWceFEQITfgCgurU2dNOj4FmZJKBTiaR2zOZdM/4AmwZu
CTmcP9vXEnyIyqAwyXJwkdXQ
=toL1
-----END PGP SIGNATURE-----
--pgp-sign-Multipart_Mon_Dec_11_11:43:29_2006-1--