NetNews Usenet Archive 1993 #1/NN_1993_1.iso/spool/comp/lang/c/19378

Text File | 1993-01-07 | 1.7 KB | 37 lines

Newsgroups: comp.lang.c
Path: sparky!uunet!zaphod.mps.ohio-state.edu!darwin.sura.net!convex!constellation!prowler.ecn.uoknor.edu!norlin
From: [e-mail] (Norman Lin)
Subject: Re: Problem with Turbo-C: calculating with pointers
Sender: [e-mail] (Usenet Administrator)
Message-ID: <[e-mail]>
Date: Thu, 7 Jan 1993 16:05:18 GMT
References: <[e-mail]>
Nntp-Posting-Host: prowler.ecn.uoknor.edu
Organization: Engineering Computer Network, University of Oklahoma, Norman, OK, USA
Lines: 24

[e-mail] (Peter De Rijk) writes:

>I am porting a program (working on different platforms such as VAX and DEC) to
>the PC. I am using Turbo C++ v.3.0. I compile using the huge memory model.
>I seem to have tracked down the problem (well, one of the problems.) I don't
>know what to do about it though. Here's a little program to illustrate the
>problem:

>I would expect end - begin to return 71800. However it seems to return 6264.
>This is exactly 64K less.

In the "huge" memory model in Turbo C++, all pointers default to "far"
pointers.  Far pointers, it says in the TC++ manual, should not be used
for pointer arithmetic because they are not segment aligned.  Declare the
pointers to be huge pointers (e.g. char huge *begin, huge *end), and the
math should work, because huge pointers are segment aligned.

In general, if you're doing ANY pointer arithmetic, you should not use far
pointers, and remember, the huge memory model defaults to using far pointers.
You should be aware that using huge pointers incurs slight performance
overhead (due to the necessity of normalization).

Norman Lin
[e-mail]