Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/5/84; site reed.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!gamma!epsilon!zeta!sabre!petrus!bellcore!decvax!tektronix!reed!kamath
From: kamath@reed.UUCP (Sean Kamath)
Newsgroups: net.micro.apple
Subject: Re: Help with DOS 3.3 write end of file
Message-ID: <2455@reed.UUCP>
Date: Wed, 5-Feb-86 19:36:16 EST
Article-I.D.: reed.2455
Posted: Wed Feb  5 19:36:16 1986
Date-Received: Fri, 7-Feb-86 20:44:21 EST
References: <189@noscvax.UUCP>
Reply-To: kamath@reed.UUCP (Sean Kamath)
Organization: Reed College, Portland, Oregon
Lines: 27

Here's an interesting tidbit.  If anyone has dossource (Complete
disassembly of dos 3.3) or is in the mood, take a look at where dos
tries to read or write a file.  I forget exactly where it is, but notice
that it checks to see if the file is open.  If it isn't then it opens
it.  Therefore, there is no real need to open a file before reading or
writing it. try it sometime.  Also, you may have noticed that
Applewriter won't create a new file if you try to read one that isn't
there.  But when you try to open one that doesn't exist from your own
programs, you're left with a one sectore file by that name.  How does
Applewriter do it?  Take a look at Beneath Apple Dos.  You'll notice
that there is a marker in dos that tells it if a command can creat a new
file.  They are "open" "save" etc.  Ones that cannot are "load" "delete"
etc.  In the case of open, if you stick a hex 22 in the right spot, no
nasty on sector files.  On the other hand, if you try to open a new file
to write to, you must replace that 22 with a 23.  If anyone's going
crazy because they can't live with out the exact location(s), write me.

Sean Kamath


-- 
________________________________________________________________________________

UUCP {ihnp4,decvax,ucbcad}!tektronix!reed!kamath

And I looked again
And the monster was me...