From xemacs-m  Mon May 12 08:24:39 1997
Received: from frege.math.ethz.ch (root@frege-d-math-north-g-west.math.ethz.ch [129.132.145.3])
	by xemacs.org (8.8.5/8.8.5) with SMTP id IAA06976
	for <xemacs-beta@xemacs.org>; Mon, 12 May 1997 08:24:38 -0500 (CDT)
Received: from fresnel (vroonhof@fresnel [129.132.145.6]) by frege.math.ethz.ch (8.6.12/Main-STAT-mailer) with ESMTP id PAA26018 for <xemacs-beta@xemacs.org>; Mon, 12 May 1997 15:23:21 +0200
Received: (vroonhof@localhost) by fresnel (SMI-8.6/D-MATH-client) id PAA08315; Mon, 12 May 1997 15:24:15 +0200
Sender: vroonhof@math.ethz.ch
To: xemacs-beta@xemacs.org
Subject: Re: Gnuattach
References: <kigiv0pymrs.fsf@jagor.srce.hr>
Mime-Version: 1.0 (generated by tm-edit 7.106)
Content-Type: text/plain; charset=US-ASCII
From: Jan Vroonhof <vroonhof@math.ethz.ch>
Date: 12 May 1997 15:24:15 +0200
In-Reply-To: Hrvoje Niksic's message of 12 May 1997 13:04:23 +0200
Message-ID: <byvi4ox1q8.fsf@fresnel.math.ethz.ch>
Lines: 58
X-Mailer: Gnus v5.4.37/XEmacs 19.15

Hrvoje Niksic <hniksic@srce.hr> writes:

> One aspect of gnuattach that annoys me horribly is that its function
> is totally confused with the function of gnuclient.  For instance, the
> message "press C-x # when done" and the whole /concept/ of pressing
> `C-x #' when done doesn't apply to gnuattach.

The problem is, it currently is two things at once.

1. gnuclient equivalent of xemacs -nw
2. tty-equivalent of make-frame-on-other-display (needed because it
   cannot be done properly without the helper program).

For 1. C-x # makes sense. For 2 absolutely not.

> 1) no more `C-x #' silliness -- what is good for gnuclient isn't
>    necessarily good for gnuattach;

Make this an option, maybe default to on when gnuattach is called with
a filename argument. May be it would be nice to write the
functionality 1. into gnuclient and let if it have a -nw option. It would
be really nice of did the same magic based on $DISPLAY as xemacs does.

 
> 2) when you delete the last frame on a gnuattached TTY device via
>    `C-x 5 0', it should tell the gnuattach process to die;

Put a function in delete-frame-hook that checks whether this is the
last frame on a gnuattach device. If yes tell gnuattach to die, may be
query to save buffer first if functionality is 1). This will require
small changes in the book keeping because gnuattach is now no longer a
client of buffer, but of a display

> 
> 3) suspending gnuattach and resuming still doesn't work well for me, I
>    get
> 
>    (1) (error/warning) Error in process filter: (wrong-type-argument number-char-or-marker-p nil) 
>

If you only get it on suspending and resuming that is really strange
because it basically does an '(server-eval (resume-pid-console <pid>))'

Can you play around with gnudoit and see if you get any problems there?

Jan


P.S. I would be very happy if all the #ifdef GNUATTACH in gnuclient.c
and the seperate gnudoit.c could be dropped in favor of a unified
program doing some magic based on switched and it's name.


P.S. If you make a lot of changes, please make them against the bare
19.15 version of the files, or ask me for a snapshot of the last
version. This features a cleaner gnuserv.el and some extentions to the
gnu* programs. 

