Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site aum.UUCP
Path: utzoo!linus!decvax!decwrl!pyramid!hplabs!well!ptsfa!aum!freed
From: freed@aum.UUCP (Erik Freed)
Newsgroups: net.micro.atari16
Subject: Re: --- Megamax Printf bug ---
Message-ID: <449@aum.UUCP>
Date: Sat, 29-Mar-86 15:28:31 EST
Article-I.D.: aum.449
Posted: Sat Mar 29 15:28:31 1986
Date-Received: Wed, 2-Apr-86 20:32:53 EST
References: <8603270054.AA00660@ucbvax.berkeley.edu> <964@megaron.UUCP>
Organization: The Aurora Systems Bunch
Lines: 19

> > Despite Jwahar R. Bammi's rebuttal, the problem described (no printing
> > of statements not  ending in '\n' until explicit buffer flush) is a bug.
> > It is possible to use fflush(stdout) to get around it, as he describes,
> > but that is not how most standard library printf()s require things.
> 
> No, no.  I have yet to see any printf do any automatic buffer flushing
> on micros, minis, or main frames.
> Jwahar was right!

I have only seen this behavior on one out of about 5 machines I have worked
on. Since, I believe, this is not documented in K&R (and in most Unix manuals) 
the nomer of "bug" does seem correct. It is just a widespread and familiar 
one...
-- 
-------------------------------------------------------------------------------
                           Erik James Freed
			   Aurora Systems
			   San Francisco, CA
			   {dual,ptsfa}!aum!freed