Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site mcnc.mcnc.UUCP
Path: utzoo!watmath!clyde!bonnie!akgua!mcnc!jnw
From: jnw@mcnc.UUCP (John White)
Newsgroups: net.micro.pc
Subject: Re: (Orphan) Re: How does SideKick work?
Message-ID: <1016@mcnc.mcnc.UUCP>
Date: Mon, 25-Nov-85 22:56:43 EST
Article-I.D.: mcnc.1016
Posted: Mon Nov 25 22:56:43 1985
Date-Received: Wed, 27-Nov-85 05:41:58 EST
References: <233@well.UUCP.UUCP> <26700005@ISM780B.UUCP>
Organization: Microelectronics Center of NC; RTP, NC
Lines: 12

> I suspect that this is connected to Borland`s insistance
> that Sidekick be the _last_ memory-resident program loaded.  Does anyone
> know why this is so?

My guess is that sidekick wants to know if the CS:IP location before
the interupt was in an application program. (The CS:IP values can be
obtained from the stack.) If the CS:IP values from the stack are greater
than its own, then sidekick can assume that an application program 
was actually being executed before the interupt. If an interupt
handler is loaded after sidekick, then this assumption can fail if that
interupt handler had control just before the interupt that gave
sidekick control.