From: huott AT pinebush DOT com (Ed Huott)
Subject: Re: cat and binary files
10 Apr 1997 19:21:57 -0700
Approved: cygnus DOT gnu-win32 AT cygnus DOT com
Distribution: cygnus
Message-ID: <199704101734.NAA16374.cygnus.gnu-win32@sol.pinebush.com>
Original-To: Tim Iverson <iverson AT cisco DOT com>
Original-cc: marc AT watson DOT ibm DOT com, gnu-win32 AT cygnus DOT com
In-reply-to: Your message of "Wed, 09 Apr 1997 19:38:31 PDT."
             <199704100238 DOT TAA21049 AT rottweiler DOT cisco DOT com> 
Original-Sender: owner-gnu-win32 AT cygnus DOT com

In message <199704100238 DOT TAA21049 AT rottweiler DOT cisco DOT com>, Tim Iverson writes:
>If cat is stopping at Ctl-Z on anything but console input, it is a bug.
>There has been no actual "end-of-file" marker in DOS since version 2.0.
>The Ctl-Z is only EOF for the *console*, just like Ctl-D is EOF for
>/dev/tty under Unix.  And, like /dev/tty, it's only EOF when the console is
>in cooked mode.  If you put the console into raw mode, Ctl-Z does nothing.
>
>Some DOS programs have yet to alter their behavior and still place Ctl-Z at
>the end of disk files.  This is also a bug, though if you want to support
>DOS 1.0, you could call it a feature.
>
>All this aside, IMHO, cat is inherently a binary program -- it shouldn't
>care whether newline is CR, LF, or CR+LF unless one of the line-based
>options is used (bns).
>
>BTW, I wouldn't recommend using binary mode.  I tried that briefly and
>found that rebuilding will fail rather miserably.
>

The relative merits of text vs. binary mode have all been discussed on
this list before in some excruciating detail.  I would suggest
perusing the mailing list archives to anyone new who is interested in
the subject.  I do not wish to rekindle that thread here.

<ED>
-
For help on using this list (especially unsubscribing), send a message to
"gnu-win32-request AT cygnus DOT com" with one line of text: "help".