Re: Mouse support bugs and a couple of feature requests (fwd)

[Pavel Tsekov <ptsekov gmx net>]

On Mon, 4 Sep 2006, Pavel Tsekov wrote:

> ---------- Forwarded message ----------
> Date: Mon, 4 Sep 2006 06:03:52 -0400
> From: Stephan Sokolow
> To: Pavel Tsekov
> Subject: Re: Mouse support bugs and a couple of feature requests
>
> On Monday September 4, 2006 05:15, you wrote:
> > On Sat, 2 Sep 2006, Stephan Sokolow wrote:
> > > First, the bug:
> > > 1. Running mc inside screen at the Linux console prevents gpm mouse
> > > support from working.
> >
> > Seems to be a gpm problem. See what I've got in my syslog:
> >
> > *** warning [client.c(183)]: Failed gpm connect attempt by uid 500 for vc 0
> >
> > I ran MC under debugger in linux console both with screen and in plain
> > console. When I am not running screen everything is normal. If I start
> > screen I get the above message both in syslog and printed on the console
> > ... but the Gpm_Open() call doesn't report error. I'll try to research why
> > this happens but it is obvious that this problem is not MC related.

Screen uses a pseudo-terminal pair to communicate with the shells it 
starts. As such when you start MC it is running on the pseudo terminal
slave device - /dev/pts/1 for example. It is screen that is running on
the virtual console device (the real terminal) and not MC. gpm cannot
handle mouse events on anything other than Linux virtual console. So
requests from MC for mouse events on /dev/pts/1 should fail ... but
there is a patch in the RedHat version of gpm that removes a check which
is supposed to catch that condition. I guess you are using Fedora, right ?