Sender: crough45 AT amc DOT de
Message-Id: <97Jul19.150632gmt+0100.17025@internet01.amc.de>
Date: Sat, 19 Jul 1997 14:11:41 +0100
From: Chris Croughton <crough45 AT amc DOT de>
Mime-Version: 1.0
To: MCLSSAA2 AT fs2 DOT mt DOT umist DOT ac DOT uk
Cc: djgpp AT delorie DOT com
Subject: Re: sizeof() in #if
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Precedence: bulk

Anthony Appleyard wrote:

>  I imagine that, if you simply wrote e.g. if(sizeof(int)==4)
> and the like, that if you compile with optimization, the
> compiler will detect conditions like this which can be worked
> out at compile time, and not bother to compile parts of the
> program which as a result are never obeyed.

Indeed it will.  However, that's no use for the wanted
purpose, which was doing typedefs for the appropriate
size.

(It also depends on te optimiser not making horrible mistakes:
I once lost half a function to an optimisation bug where it 
thought that it knew what the value in an if was but got it
wrong...)

Certainly doing run-time tests for sizes can be useful because
in almost all cases they are optimised out, and they are often
clearer than puting #if around things.

>  They may well come, if you boot up with a floppy in the 
> drive or the like.  There is a computer virus called DaBoys.

I didn't know that.  But does it do its job with sax and 
violins?

(It doesn't matter if I boot up with a floppy in the drive
- a nice thing with a modern BIOS is being able to disable
booting from floppy, and my machine is configured like that
except for the rare occasions when I need to use a real boot
floppy for system recovery.  I recommend anyone with a
modern BIOS to do that, it saves a lot of hassle.)

Chris C