Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site lsuc.UUCP
Path: utzoo!lsuc!dave
From: dave@lsuc.UUCP (David Sherman)
Newsgroups: net.news,net.news.b,net.unix
Subject: Re: massive rn speedup code posted to net.sources
Message-ID: <527@lsuc.UUCP>
Date: Tue, 19-Mar-85 21:22:34 EST
Article-I.D.: lsuc.527
Posted: Tue Mar 19 21:22:34 1985
Date-Received: Tue, 19-Mar-85 21:28:31 EST
References: <481@lsuc.UUCP> <516@lsuc.UUCP> <1382@utah-gr.UUCP>
Reply-To: dave@lsuc.UUCP (David Sherman)
Organization: Law Society of Upper Canada, Toronto
Lines: 18
Summary: I don't understand how 4.1 could buffer stdout "for free"

In article <1382@utah-gr.UUCP> thomas@utah-gr.UUCP (Spencer W. Thomas) writes:
||In article <516@lsuc.UUCP> dave@lsuc.UUCP (David Sherman) writes:
||>I mentioned in a previous posting that I had implemented buffered
||>stdout in rn, and it made a big difference in output at high speed.
||
||I might point out that, unless Larry deliberately programmed around it,
||you get buffered stdout "for free" in 4.1bsd and above.

Well, mebbe so, but the basic output call in art.c, when spitting
out the article, is putchar, and there were no fflush's in the
source before I put some in. If 4.1 buffered stdout (as opposed
to buffering during printfs, which perhaps it does), I
don't think rn would work under it.

Dave Sherman
-- 
{utzoo pesnta nrcaero utcs hcr}!lsuc!dave
{allegra decvax ihnp4 linus}!utcsri!lsuc!dave