Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!uunet!crdgw1!vdsvax!trub
From: perley@trub (Donald P Perley)
Newsgroups: comp.sys.amiga
Subject: Re: The Inevitable 1.4 Rom Problem (was Re: Ferrari Formula One and Kickstart 1.3?)
Message-ID: <10543@vdsvax.crd.ge.com>
Date: 4 Jan 90 14:41:13 GMT
References: <1017@tuminfo1.lan.informatik.tu-muenchen.dbp.de> <7797@cognos.UUCP>
Sender: news@vdsvax.crd.ge.com
Reply-To: perley@trub (Donald P Perley)
Organization: GE Corp. R & D, Schenectady, NY 12345
Lines: 27
In-reply-to: gords@cognos.UUCP (Gord Smith)

In article <7797@cognos.UUCP>, gords@cognos (Gord Smith) writes:
>I have been concerned for some time about how having KickStart in ROM is going
>to be a reeeeeaaaaal problem once 1.4 comes out, in terms of still being able
>to run misbehaved but nevertheless enjoyable software that works under 1.2.
>
>The quoted article below points out that there are even problems with some
>software under KickStart 1.3, whose only supposed change was autoboot.

If you access the routines in the "proper way", that is opening libraries, and 
calling routines at the addresses in the jump table you get from the open, 
than I think things should be ok.

There is some software out there that will call a routine at location X
because thats where they empirically found it was in KS1.2, or for a real
hackers delight, will call a routine at a non-standard entry point to make
it do just what they want it to.

Consider what happens when you write a program with a number of routines.
If you change just one of them and recompile, you could say that all the other
ones were unchanged.  In reality, though, the one you changed will be a little
different in length, so all the other ones are shifted around a little. 
(ignoring relocation adjustment for the moment)
If you call the routines by absolute location you would say they are broken.
If you call them by name, then they are unchanged.

-don perley
perley@trub.crd.ge.com