How to use mouse with mc in console

[Andrew V. Samoilov]

Peter Hugosson-Miller wrote:
> Björn Eriksson wrote:
> 
> 
>>[No CC:s please, Mail-Followup-To: mc-devel at gnome.org]
>>
> 
> OK, for original thread, see
> http://mail.gnome.org/archives/mc/2002-April/msg00009.html
> 
> 
>>Hi Peter, greetings from Göteborg,
>>
> 
> And hi from a fellow Swede in Stockholm!
> 
> 
>> If you have reasons to believe it's a mouse/device/gpm-problem then
>>there are perhaps better places to ask. I'm afraid I don't have any
>>list-suggestions though.
>>
> 
> I was just hoping this might be a situation that someone else might have seen
> before, or there might be some way to get mc to print out debugging
> information, so I can track the problem down. Now that I have it working on
> one installation and not working on another, with no visible differences
> (yet), it ought to be possible to compare them and see what is different.
> 
> 
>> A couple of thoughs: BTW, I assume you mean the *Linux* console.
>>
> 
> Yes. No X is involved in this problem. On the system that doesn't work, X
> isn't even installed - it's a bootable Debian CD-ROM, developed by Timo Benk
> (who can't explain the problem either). Check out the project home page
> http://rescuecd.sourceforge.net/ for more details.
> 
> Timo was very careful to use just standard packages from woody (the up-coming
> Debian release - available Real Soon Now TM). The whole system is loaded into
> RAM, with (he believes) everything writeable that ought to be writeable.
> 
> 
>> [0. Why use a mouse? :) And from the console at that??]
>>
> 
> The mouse is useful for select and paste operation in the console, and I like
> it for mc too (I grew up on the Norton Commander, back in the early eighties,
> so I'm used to using the mouse).
> 
> 
>> 1. Are you running X at the same time? I think you are on the machine
>>where it's working. (As /dev/gpmdata is the repeater pipe.)
>>
> 
> No X installed at all (except on the machine where it's working). As you point
> out, /dev/gpmdata should be a fifo, not a file, and Timo says that this will
> be fixed in the next release. I have deleted the file, and let gpm (or
> gpmconfig) re-create it, but it makes no difference to the situation. My guess
> is that mc uses /dev/gpmctl to get its mouse input (but maybe someone else
> knows better ;-).
> 
> 
>> 2. It's not likely to be a mc problem. More likely a library problem
>>since you've asserted that the mouse 'works' [the cursor moves and you
>>can paste]. Have you tried any other console-apps like lynx, w3m, links?
>>
> 
> Not yet, that will be my next step, then. Any idea which libraries might be
> causing the problem, or a good place to start looking?
> 

This came from Slackware current ChangeLog.txt.

----------------------------
Wed Mar  6 16:24:30 PST 2002
a/gpm-1.19.6-i386-1.tgz:  Removed broken curses support that interferes with
   ncurses, upgraded to gpm-1.19.6.  This should restore application mouse
   support for things like mc.  Thanks to Thomas Dickey (and Google) for
   supplying the clues I needed to solve this mystery. :)  Also added a
   description of the setup.mouse script that's seen when reconfiguring with
   pkgtool, removed the suid bit from disable-paste, and patched for 
<time.h>
   changes in recent glibc.  Fixed rc.gpm to accept "start" and "stop".
   Moved shared library into /lib, since we'll be linking libncurses to 
it and
   that's where libncurses.so.* is installed.
l/ncurses-5.2-i386-2.tgz:  Linked to /lib/libgpm.so.1.  This automagically
   brings console mouse support to lynx and possibly other apps that use
   libncurses.