Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!watmath!clyde!rutgers!seismo!nbires!hao!noao!mcdsun!fnf
From: fnf@mcdsun.UUCP
Newsgroups: comp.sys.amiga
Subject: Re: include files
Message-ID: <245@mcdsun.UUCP>
Date: Sat, 21-Feb-87 15:42:09 EST
Article-I.D.: mcdsun.245
Posted: Sat Feb 21 15:42:09 1987
Date-Received: Mon, 23-Feb-87 01:55:49 EST
References: <1627@amiga.amiga.UUCP> <1494@tekigm2.TEK.COM>
Reply-To: fnf@mcdsun.UUCP (Fred Fish)
Organization: Motorola Microcomputer Division
Lines: 29

In article <1494@tekigm2.TEK.COM> phils@tekigm2.UUCP (Philip E Staub) writes:
>Structures declared in the include files as having elements which are
>declared as type "int" will have to be changed to "long" for the Manx
>compiler. I would have to assume that this is included in what you refer to 
>as "editing" the include files, but it is essential nevertheless.

Actually, given that these declarations represent fixed length integral
quantities that are essentially "cast in concrete" and are part of the
fixed hardware (ROM) environment, they should be declared as int32 or int16
as appropriate, with a single compiler typedef taking care of the mapping
to a type supported by whatever random C compiler.

In general, my experiences over the last 16 months with existing code,
lint, and two compilers (Manx and Lattice) have convinced me that the
entire object typing situation on the Amiga is a total disaster.  It
is definately not helped by the split personality of part of the system
software being done in BCPL.  If someone with extensive experience in
portability and use of lint had been working on the project from its
earliest days, I suspect we would not be suffering the current problems.

Sigh.

-Fred

-- 
===========================================================================
Fred Fish  Motorola Computer Division, 3013 S 52nd St, Tempe, Az 85282  USA
{seismo!noao!mcdsun,hplabs!well}!fnf    (602) 438-5976
===========================================================================