Date: Sat, 11 Nov 2000 01:22:56 +0200
From: "Eli Zaretskii" <eliz AT is DOT elta DOT co DOT il>
Sender: halo1 AT zahav DOT net DOT il
To: richdawe AT bigfoot DOT com
Message-Id: <1659-Sat11Nov2000012256+0200-eliz@is.elta.co.il>
X-Mailer: Emacs 20.6 (via feedmail 8.3.emacs20_6 I) and Blat ver 1.8.6
CC: djgpp-workers AT delorie DOT com
In-reply-to: <3A0C59A8.9DB02164@bigfoot.com> (message from Richard Dawe on
	Fri, 10 Nov 2000 20:25:12 +0000)
Subject: Re: Summary of the snprintf() situation
References: <3A09EAB9 DOT 6924045A AT bigfoot DOT com> <3A0B329F DOT D33F7737 AT bigfoot DOT com> <7458-Fri10Nov2000111820+0200-eliz AT is DOT elta DOT co DOT il> <3A0C59A8 DOT 9DB02164 AT bigfoot DOT com>
Reply-To: djgpp-workers AT delorie DOT com
Errors-To: nobody AT delorie DOT com
X-Mailing-List: djgpp-workers AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk

> Date: Fri, 10 Nov 2000 20:25:12 +0000
> From: Richard Dawe <richdawe AT bigfoot DOT com>
>
> > > I have to admit, reading this again, I am confused by what you say,
> > > Eli. If you don't track the position in the buffer using _cnt, what
> > > do you use?
> > 
> > p->_cnt is not used for tracking the position, it is only used to know
> > how much more free space do we have in p->_buf.  To track the buffer
> > position, we use p->_ptr; see the code above.
> 
> By 'p->_buf', I guess you mean 'p->_base'.

Yes.

> As it stands, the snprintf() code does not set up p->_base. So if we
> do not use p->_cnt, this will need fixing.

It does use p->_ptr, which is the vehicle for tracking the buffer
position.  Isn't that what you were asking about?

> If n is too large, then _cnt (an int) will be negative if you just put n
> (a size_t) into it (actually n - 1 to leave space for nul). Therefore I
> suggest we put a restriction on n, to avoid this (and hence wrap-around &
> returning EOF).

Why not?  I don't see anything wrong with this approach.

> Things are pretty clear in my mind now. Maybe I will get time to produce
> some patches on Sunday. Then the code can do the talking.

Sounds like a good plan ;-)