Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.3 alpha 4/15/85; site lzwi.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!mhuxr!mhuxt!houxm!ihnp4!lzwi!psc
From: psc@lzwi.UUCP (Paul S. R. Chisholm)
Newsgroups: net.unix-wizards,net.unix
Subject: Re: unix terminal i/o
Message-ID: <179@lzwi.UUCP>
Date: Tue, 18-Jun-85 09:17:57 EDT
Article-I.D.: lzwi.179
Posted: Tue Jun 18 09:17:57 1985
Date-Received: Wed, 19-Jun-85 04:20:11 EDT
References: <452@mmintl.UUCP>
Organization: AT&T-IS Enhanced Network Systems
Lines: 28
Xref: watmath net.unix-wizards:13568 net.unix:4827
Summary: VMIN=VTIME=0 or fcntl(O_NDELAY)

< Smokey the Bar says, "Help stamp out software pirates" [squish!] >

In article <452@mmintl.UUCP>, robg@mmintl.UUCP (Robert Goldman) writes:
> 
> I am working on a program that needs to access the terminal in a pretty raw
> state.  In particular, it must have some way to tell whether there are any
> keys in the input buffer waiting to be read.  Is there some portable unix
> call that responds with this information?
> 			Robert Goldman, MultiMate International

In Unix(tm) System III and V (all releases), you can ioctl() the line
to have VMIN and VTIME set to 0; a read() will only get any characters
in the input queue.  (Note that ioctl(fd,TCFLSH,0) will flush the input
queue.)

More portably, you can fcntl() the file descriptor with O_NDELAY, which
avoids blocking on tty's and pipes (and fifos).  If you're always going
to avoid blocking, you can open() the file with O_NDELAY.

Both ioctl() and fcntl() are pretty cheap to call, so it's not too
expensive to go back and forth.
--
Unix is a trademark of some entity within AT&T, but I don't know what
it's called this week.  AT&T's Bell Labs?
-- 
       -Paul S. R. Chisholm       The above opinions are my own,
       {pegasus,vax135}!lzwi!psc  not necessarily those of any
       {mtgzz,ihnp4}!lznv!psc     telecommunications company.