Sender: tih@nhh.no
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: Tom Ivar Helbekkmo <tih@nhh.no>, Paul A Vixie <vixie@vix.com>,
	hackers@postgreSQL.org
Subject: Re: [HACKERS] Open 6.4 items
References: <23787.907968022@sss.pgh.pa.us>
From: Tom Ivar Helbekkmo <tih@nhh.no>
Date: 10 Oct 1998 14:38:25 +0200
In-Reply-To: Tom Lane's message of "Fri, 09 Oct 1998 17:20:22 -0400"
Message-ID: <86iuhsd22m.fsf@athene.nhh.no>
Lines: 30

Tom Lane <tgl@sss.pgh.pa.us> writes:

> Tom Ivar Helbekkmo <tih@nhh.no> writes:
> > As far as I
> > can tell, the only actual error in Paul Vixie's code is that the two
> > lines you quote above should be:
> > 			else if (--size > 0)
> > 				*++dst = 0, dirty = 0;
> 
> No, that's still wrong, because it will error out (jump to emsgsize)
> one byte sooner than it should.  The loop is fundamentally broken
> because it wants to grab and zero a byte before it knows whether there
> are any digits for the byte.

Damn!  You're right!  I swear I tested that change at home, and found
it to do what I wanted -- but now that I've put it into my production
system at work, it doesn't work here.  Heck, I even read the code very
carefully to verify that it was all that was needed!  I think I really
need the holiday I'm taking next week!

To keep this in sync with BIND, maybe Paul could take a look, and fix
the code the way he wants it there?

Whoops.  It just hit me.  When I tested my "quick fix" last night, I
just looked at what happened to the data, and ignored the return
value.  Silly of me...

-tih
-- 
Popularity is the hallmark of mediocrity.  --Niles Crane, "Frasier"