Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!rutgers!sri-unix!hplabs!hao!nbires!isis!dragheb
From: dragheb@isis.UUCP (Darius "OPRDRT" Ragheb)
Newsgroups: comp.sys.ibm.pc
Subject: Re: Memory Res. Prgs.
Message-ID: <1568@isis.UUCP>
Date: Mon, 19-Jan-87 23:14:07 EST
Article-I.D.: isis.1568
Posted: Mon Jan 19 23:14:07 1987
Date-Received: Wed, 21-Jan-87 00:30:59 EST
References: <1541@isis.UUCP> <1610003@hpcvlo.HP.COM> <966@nbires.UUCP>
Reply-To: dragheb@isis.UUCP (Darius "OPRDRT" Ragheb)
Organization: Math/CS, University of Denver
Lines: 19

In article <966@nbires.UUCP> mmm@nbires.UUCP (lloyd w. tabb) writes:
>the proposed method of comparing interrupt handlers to determine
>if a TSR program is already loaded is not sound.  it
>is possible that some other TSR program has stolen your vector (has been
>loaded after you and chained to that interrupt).  in this case your
>handlers would NOT match but the program would be already loaded.  while
>this would work most of the time it is not bullet proof.

This is the exact problem (I think) that occurs when you load sidekick.
As Borland "recommends" loading SK last, it seems to "jump" in front
of all the other interrupt handlers.  Another method was proposed (and
has not yet been tried) is: place a unique sequence of bytes in the data
segment, and when loaded scan memory for it (not too bad timewise, i 
suppose) BUT also have a flag right after the sequence, and when determined
that it is the first time the program is run, change the flag.
-- 
Functionality, Efficiency, Luxury.

isis!dragheb  |  dragheb@isis.cs.du.edu