Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!ames!lll-winken!ubvax!ardent!rap
From: rap@ardent.UUCP (Rob Peck)
Newsgroups: comp.sys.amiga.tech
Subject: Re: ARP with 1.3
Summary: wish I coud remember...
Message-ID: <5236@ardent.UUCP>
Date: 29 Mar 89 23:30:35 GMT
References: <6472@dayton.UUCP>
Organization: Dana Computer, Inc., Sunnyvale, CA
Lines: 39

In article <6472@dayton.UUCP>, joe@dayton.UUCP (Joseph P. Larson) writes:
> 
> This has probably been asked, but I didn't see it float by.  Can I replace
> my new 1.3 commands with my old ARP commands in my C directory?  When
> will the 1.3 extensions for ARP be coming out?

Thad Floryan has tested all of 'em and knows which ones are incompatible.
Was it RESIDENT and ASSIGN?  I forget.  Thad?

> Finally -- can someone In The Know explain the difference between using
> Jim Goodnow's "rez" program and the Amiga "Resident" command?
> 
Resident (1.3) works if a program is 'pure', meaning no global variables
and fully re-entrant.  Thus it permits you to have only one copy of the
code loaded at one time, with multiple programs using it.  Kinda like
the shared libraries.   Each invocation of the code has to allocate
its own work areas (and if it executes scripts, the <$$> facility
now in 1.3, which gives the current CLI number (ASSIGN HERE<$$>: ""
for example) is a big help in this area).

'rez', which works only with code compiled with Manx 3.6 and above,
recognizes certain explicit code sequences in the loaded code, and
adds a patch to create a separate data-segment (initialized as well as
uninitialized data) so that the code is loaded only once, but the
data segment is created afresh for each incarnation of the program.
He actually patches into the segment-list of the loaded program,
(really neat.. he explained it once at BADGE about a year ago).
I forget the other limitations, if any.  He stopped at version '0.4',
perhaps because of the known progress of 1.3's RESIDENT command.

MAYBE when Manx 5.x comes out, we'll see more (no knowledge in this
area, just speculation).

I guess someone just oughta write a "Complete Guide To Making A Program
Pure and Residentable", if not already available.  But I STILL like
Jim Goodnow's approach... viva la hacker.


Rob Peck