Path: utzoo!mnetor!uunet!amdahl!dlb!ardent!rap
From: rap@ardent.UUCP (Rob Peck)
Newsgroups: comp.sys.amiga
Subject: Re: One silly question...
Message-ID: <318@ardent.UUCP>
Date: 22 Feb 88 20:29:07 GMT
References: <8802200055.AA00454@jade.berkeley.edu>
Organization: Dana Computer, Inc., Sunnyvale, CA
Lines: 23

In article <8802200055.AA00454@jade.berkeley.edu>, FATQW@USU.BITNET writes:
> Why is APTR defined to be a pointer to a pointer (STRPTR) to UBYTEs???

As I remember, Carl Sassenrath wanted to have something that was a
generic pointer variable and this is what he came up with.  The side
effect of this choice was that if:

	APTR p;

then when you later used

	p++;

it would only increment the pointer by the size of one long word.
(if p == 0x10000, then p++ == 0x10004).

Not sure exactly what the "full" reason, but this is part of it.

(It may also have some bearing on one of the macros that converts an
APTR to a BPTR and vice versa.  If APTR was badly defined, the C
compiler could have a field day with the value it returns.)

Rob Peck		(Amiga-history-buff)	...ihnp4!hplabs!ardent!rap