Xref: utzoo comp.mail.uucp:3456 comp.sys.m68k:1223
Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!rutgers!orstcs!ogccse!unmvax!ncar!mailrus!ames!amdahl!nsc!pyramid!csg
From: csg@pyramid.pyramid.com (Carl S. Gutekunst)
Newsgroups: comp.mail.uucp,comp.sys.m68k
Subject: Re: UUCP sending BREAK
Message-ID: <81745@pyramid.pyramid.com>
Date: 23 Aug 89 04:09:28 GMT
References: <109@bdt.UUCP>
Reply-To: csg@pyramid.pyramid.com (Carl S. Gutekunst)
Distribution: usa
Organization: Pyramid Technology Corp., Mountain View, CA
Lines: 19

In article <109@bdt.UUCP> david@bdt.UUCP (David Beckemeyer) writes:
>The problem is the tty driver doesn't seem to know how to send a BREAK.
>
>	if (isatty(fn)) 
>		(void) ioctl(fn, TCSBRK, 0);
>
>Nothing seems to happen when the above code is executed.

Ugly. Assuming you have source -- it appears that you do -- the classic work-
around from Version 7 UNIX should suffice. Simply drop the line baud to some
very small number, like B50, transmit a NULL character, and then restore the
speed. At 50 bps, one NULL gives you a nicely shaped 180ms pulse, which will
do for a BREAK almost anywhere.

On 4.3BSD, which has no timed BREAK, I've actually found this works better
than the "normal" technique of forcing the line to a space state, waiting on
an interval timer, then restoring.

<csg>