[ On Tuesday, February 8, 2000 at 14:56:10 (-0500), Allen Briggs wrote: ]
> Subject: Re: m88k system call convention
>
> I would love to have this information if no one else has claimed it yet.
> I'd be happy to pay for shipping to me.

So far you're it!  Let me know your shipping address and what kind of
service you'd like and I'll get you a price (or if you have a FedEx or
UPS account you might be able to arrange a pickup direct from me).

> DG definitely paid attention to 88Open--at least, they did when I was
> working in RTP in 1989.  I'm pretty sure that 88Open defined COFF,
> though, and not ELF.  At least in earlier versions.

Yes, that's right -- 88Open BCS v1.1 is definitely COFF all the way
through, and of course is SVID-based with extensions necessary to be in
conformance with IEEE 1003.1 (1988).

I think, IIRC, that the ELF standard was never completed.  This version
simply says that binaries in conformance with the 88open BCS v1.1 are
assured compatability with (i.e. will run on) AT&T-based System V
Release 4 environments and that future definitions of the standard will
be in full conformance with the SVR4.0 ABI published by AT&T: including
use of ELF; use of dynamic shared libraries for access to system calls
(instead of direct maps as used in the current revision); and the use of
library interfaces to avoid the naming of, and physical format specs
for, system files.

-- 
							Greg A. Woods

+1 416 218-0098      VE3TCP      <gwoods@acm.org>      <robohack!woods>
Planix, Inc. <woods@planix.com>; Secrets of the Weird <woods@weird.com>