Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!csd4.milw.wisc.edu!bionet!apple!netcom!hinton
From: hinton@netcom.UUCP (Greg Hinton)
Newsgroups: comp.os.minix
Subject: Re^3: Future of Minix
Message-ID: <1821@netcom.UUCP>
Date: 17 Jul 89 23:49:51 GMT
References: <19654@louie.udel.EDU> <2885@ast.cs.vu.nl> <465@medusa.informatik.uni-erlangen.de> <2897@ast.cs.vu.nl>
Reply-To: hinton@netcom.UUCP (Greg Hinton)
Organization: NetCom -- San Jose, CA
Lines: 18

In article <2897@ast.cs.vu.nl> ast@cs.vu.nl (Andy Tanenbaum) writes:
>I think assertions have utility as comments even when no code is generated.
>There is already a file assert.h in V1.3.  The only thing wrong with it
>is that instead of #ifdef DEBUG, it says #ifdef NDEBUG.  
>I think NDEBUG is 10x harder to understand.

I agree that NDEBUG is backasswards & causes at least a moment of puzzlement.
However, its use in assert.h is ANSI standard, & we all want to be
standardized now, don't we:-)?
Actually, it does make some sense when you realize that the default mode is
to generate debugging (assert'ing) code.  It's only when you have a polished
product that you
cc -DNDEBUG ....
to eliminate debugging overhead.
-- 
Greg Hinton
INET: hinton@netcom.uucp
UUCP: ...!uunet!apple!netcom!hinton