Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!watmath!clyde!rutgers!brl-adm!adm!bzs@bu-cs.bu.EDU
From: bzs@bu-cs.bu.EDU
Newsgroups: comp.unix.wizards
Subject: Time for 64-bit longs?
Message-ID: <4529@brl-adm.ARPA>
Date: Wed, 18-Feb-87 11:47:56 EST
Article-I.D.: brl-adm.4529
Posted: Wed Feb 18 11:47:56 1987
Date-Received: Thu, 19-Feb-87 20:04:08 EST
Sender: news@brl-adm.ARPA
Lines: 21


From: "Wilson H. Bent" <whb@vax135.uucp>
>This is not really a 128-bit pointer, this is a structure with many parts.

You missed my point (so to speak), I was using the struct to describe
a 128 bit pointer to an audience of C programmers.

For example, A char * in C on a DEC20 has (had?) a similar segmented
structure and could be described by a struct, but you really had to
carry it around in a single char * to get the correct machine code
generated, there were special machine instructions for dealing with
such objects such as ILDB and the only correct way to deal with them
was as a char *, not a struct.  (well, 'correct' meaning other than
excessive bit hacking behavior.) If you didn't have 36-bit pointers
the compiler would be crippled even tho it might appear that 18-bits
was sufficient to address every word in memory (let's ignore extended
addressing on the KL for the moment.)

In sum: I really meant a pointer.

	-Barry Shein, Boston University