Date: Mon, 24 Nov 1997 13:05:25 +0200 (IST)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
To: Nate Eldredge <eldredge AT ap DOT net>
cc: Cesar Scarpini Rabak <csrabak AT dce03 DOT ipt DOT br>,
        Thomas Demmer <demmer AT LSTM DOT Ruhr-UNI-Bochum DOT De>, djgpp AT delorie DOT com
Subject: Re: stdout flushing after printf, SIGALRM signal and GNU Make
In-Reply-To: <199711220501.VAA16015@adit.ap.net>
Message-ID: <Pine.SUN.3.91.971124130436.23713D-100000@is>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Precedence: bulk


On Fri, 21 Nov 1997, Nate Eldredge wrote:

> >>Anyway, because this is the most frequently asked frequently
> >>asked question, DJ is planning to change stdout to unbuffered.
> >                 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 
> >If this is being considered, might we have then some sort of flag which
> >allow the programmer to choose between buffered or not?
> IMHO, this already exists: it's called `setvbuf'. :)

No, `setvbuf' won't change anything in the way current v2.02 alpha works. 
`printf' in v2.02 will call "fflush (stdout)" unconditionally before it
returns.  Presumably, programs that care about performance should either
not call `printf' for too small chunks, or use `fprintf' which still
retains the previous buffered behavior. 

If you don't agree with that change, lobby DJ.