Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!usc!samsung!uunet!stanford.edu!agate!dog.ee.lbl.gov!elf.ee.lbl.gov!torek
From: torek@elf.ee.lbl.gov (Chris Torek)
Newsgroups: comp.unix.questions
Subject: Re: Problem with sleep and signal/alarm.
Message-ID: <12986@dog.ee.lbl.gov>
Date: 9 May 91 07:53:27 GMT
References: <1991May6.153937.28635@msuinfo.cl.msu.edu> <1991May07.221122.18325@chinet.chi.il.us>
Reply-To: torek@elf.ee.lbl.gov (Chris Torek)
Organization: Lawrence Berkeley Laboratory, Berkeley
Lines: 14
X-Local-Date: Thu, 9 May 91 00:53:27 PDT

In article <1991May07.221122.18325@chinet.chi.il.us>
les@chinet.chi.il.us (Leslie Mikesell) writes:
>... First, the alarm granularity on unix is 1 second intervals, so
>an alarm(1) will go off at the next second transition after you do
>it which could be almost instantly.

Actually, this depends on your Unix variant.  There is still some
unavoidable jitter and delay, and adjtime() goofs things up, but in
general, on 4.2BSD systems and systems descended therefrom, the real
interval timer---alarm is a library function that this---really does
deliver SIGALRMs when you asked for them.
-- 
In-Real-Life: Chris Torek, Lawrence Berkeley Lab CSE/EE (+1 415 486 5427)
Berkeley, CA		Domain:	torek@ee.lbl.gov