<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso-8859-1">
<META NAME="Generator" CONTENT="MS Exchange Server version 6.5.7638.1">
<TITLE>RE: [Spread-users] Hi, here's a bug, and a patch.</TITLE>
</HEAD>
<BODY>
<!-- Converted from text/plain format -->
<P><FONT SIZE=2>Yeah, you're probably right. I was thinking of that. Spread won't compile on a pre-ANSI compiler anyway, or probably on a freestanding implementation.<BR>
<BR>
Since I don't know the code at all, I was trying to be conservative. I assumed there must be some reason someone would explicitly define NULL, and I just didn't know it.<BR>
<BR>
But even if that's true, there's no reason the uses of NULL couldn't be eliminated, or changed to a local define that wouldn't screw up the rest of the world. (I only noticed one inessential place.)<BR>
<BR>
-----Original Message-----<BR>
From: Neil Conway [<A HREF="mailto:neilc@samurai.com">mailto:neilc@samurai.com</A>]<BR>
Sent: Thu 5/4/2006 2:25 AM<BR>
To: Andrew Barnert<BR>
Cc: spread-users@lists.spread.org<BR>
Subject: Re: [Spread-users] Hi, here's a bug, and a patch.<BR>
<BR>
Andrew Barnert said:<BR>
> Anyway, I found a bug in spread-3.17.3. Spread's C/C++ interface defines<BR>
> NULL as (void *)0, which is perfectly fine for C, but not for C++.<BR>
<BR>
Why does Spread need to provide its own definition of NULL in any case? It<BR>
should be provided by any reasonably modern platform -- IMHO the right fix<BR>
would be to just remove the #ifndef NULL block (and #include <stddef.h> as<BR>
needed).<BR>
<BR>
-Neil<BR>
<BR>
<BR>
<BR>
<BR>
</FONT>
</P>
</BODY>
</HTML>