Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!watmath!clyde!rutgers!lll-lcc!seismo!cmcl2!rna!rocky2!reintom
From: reintom@rocky2.UUCP
Newsgroups: comp.sys.ibm.pc
Subject: Re: DOS line editing
Message-ID: <335@rocky2.UUCP>
Date: Sun, 25-Jan-87 21:52:31 EST
Article-I.D.: rocky2.335
Posted: Sun Jan 25 21:52:31 1987
Date-Received: Tue, 27-Jan-87 02:14:56 EST
References: <307@cognos.UUCP> <321@rocky2.UUCP> <324@cognos.UUCP>
Organization: Rockefeller Univ.,N.Y.C 10021
Lines: 126

In article <324@cognos.UUCP>, brianc@cognos.UUCP (Brian Campbell) writes:
> In article <321@rocky2.UUCP> reintom@rocky2.UUCP (Tom Reingold) writes:
> >
> >In article <307@cognos.UUCP>, brianc@cognos.UUCP (Brian Campbell) writes:
> >> 
> >>      In almost every version I used up to and including 3.1, there has
> >> been a Ctrl-U and Ctrl-W keystroke which is recognized.  [For those of
> >> you who didn't know of or didn't have these features, Ctrl-U erases the
> >> entire line (similar to ESC, but on the same line) and Ctrl-W erases
> >> backward to the last non-alphanumeric character]. 
> >> 
> >> 
> >I formatted our hard disk at work and noticed the same thing as you.  It is
> >not the version of DOS that is in question.  It is really a bug in the
> >format program because it happens only on some hard disks.  I was using
> >the original IBM floppy disk that had no write-enable notch.  There were
> >two differences between the system files on the floppy and the hard disk.
> >I patched the one on the hard disk to be the same as on the floppy and
> >the editing keys returned.
> >
> >These editing features were not documented and have been removed from DOS
> >3.2  It seemed like a shame to me but I recently got a copy of CED, a
> >freeware command editor.  Having it is far better than having Ctrl-U and
> >Ctrl-W.  It gives command history and much more.  I suggest you get a copy.
> >It's really good.  You won't miss the editing keys.
> >
> >-Tom
> 
> Ridiculous!!!
> 
> Not only are the editing features available in DOS 3.20 (patch below), but
> I find it *extremely* hard to believe that the format program could just
> occasionally replace four different non-consecutive bytes of one of the system
> files with NOP's (which is what the difference between having Ctrl-U & Ctrl-W
> and not amounts to).
>
> [...]
>
> Brian Campbell

You say ridiculous.  I say yes it is.  Not that it is not true, however.
My statement was not based on conjecture.  It was based on measured
observation.  I am a programmer and it does seem terribly unlikely that a
bug like this would creep into a program but it seems to be the case.

My observation was that using the *same* IBM DOS disk, formatting different
hard disks created varying copies of the file called IBMDOS.COM.  (Either
that or IBMDOS.COM -- I don't remember which.)  I can't tell you why this
is.  I am not an adept disassembler of other people's code.  If I wrote a
format program, I think I would copy that system file over without
modifying it but I don't know what Microsoft or IBM intended to do.  I
am merely observing the end result.

Tom Reingold

This paragraph is necessary  because  my
postnews  program  does not permit me to
write *less* text than I  am  responding
to.   Isn't  that  clever?   Now you are
reading this superfluous paragraph as  a
result.




























































-- 
Tom Reingold;  The Rockefeller University; 1230 York Av; NY 10021
PHONE: (212) 570-7709 [office]; (212) 304-2504 [home]
ARPANET: reintom@rockefeller.arpa BITNET: REINTOM@ROCKVAX
UUCP: {seismo|ihnp4|yale|harvard|philabs|phri}!cmcl2!rna!rocky2!reintom