Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: $Revision: 1.6.2.16 $; site ima.UUCP
Path: utzoo!watmath!clyde!bonnie!akgua!gatech!seismo!harvard!think!ima!johnl
From: johnl@ima.UUCP
Newsgroups: net.micro.pc
Subject: Re: (Orphan) Re: How does SideKick work?
Message-ID: <131300001@ima.UUCP>
Date: Fri, 22-Nov-85 18:16:00 EST
Article-I.D.: ima.131300001
Posted: Fri Nov 22 18:16:00 1985
Date-Received: Mon, 25-Nov-85 07:30:56 EST
References: <233@well.UUCP.UUCP>
Lines: 23
Nf-ID: #R:well.UUCP:-23300:ima:131300001:000:1264
Nf-From: ima!johnl    Nov 22 18:16:00 1985


/* Written 11:43 pm  Nov 20, 1985 by patrick@ISM780B in ima:net.micro.pc */
> > SideKick installs in the timer interrupt.
> It can't be as simple as that.

You're right, it's a lot more complicated than that.  I was working on a
program that grabbed the keyboard interrupt to peek at the keystrokes coming
in.  (It's Javelin, read all about it in next week's InfoWorld.)  Anyway,
it worked fine normally but hung and barfed under SideKick.  After a lot of
calling back and forth I finally talked to a guy at Borland named Bela, and
here's what he told me:  SideKick grabs all sorts of interrupts including the
timer, the keyboard, the DOS call interrupt, and lots more.  Every time that
the clock ticks, it looks at the keyboard interrupt to see if some other
program grabbed it back, and if so reinstalls its own routine into the keyboard
interrupt on top of the other one, after saving the other one away.  This means
that programs like Epsilon which establish their own keyboard handlers will
still work with SideKick.  Unfortunately, my program grabbed the keyboard
interrupt twice, out of sloppiness, and SideKick couldn't handle that.  When
I fixed it only to grab the keyboard once, it worked fine.

Strange stuff, this PC software.

John Levine, ima!johnl