[gpm]Re: report

[Petr Baudis]

Dear diary, on Thu, Nov 07, 2002 at 04:31:20PM CET, I got a letter, where
Alessandro Rubini <rubini@gnu.org> told me, that...
> > The Linux Kernel Team seems to make great progress in their Input Devi= ces
> > section, perhaps they'll develop a much smaller gpm replacement as soon as
> > = all mice protocols are supported in the kernel.
> 
> Mice protocols should _not_ be supported in the kernel. I don't expect gpm to
> be implemented by the kernel people at all. They ought to _remove_ selection
> stuff from TIOCLINUX instead, and gpm should become a user-space tool with no
> kernel support at all (in addition to /dev/vcs stuff, obviously).

BTW, I'm already working on this - first I submitted a patch to make usable
interface to the kernel selection buffer, but no interest has shown up at all;
and my mind changed as well, I think that the selection support in kernel
should be made deprecated and it shouldn't be used by gpm at all - I'm working
on a patch which aims to do this right in these days. After this will be
finished, I also want to add additional interface to gpm for its internal
selection buffer (kernel's won't be used anymore), thus it will solve the
insane situation of applications wanting to mess with the selections but having
no way to do it and it will also finally make it possible to share one
selection buffer between X and console.

It would be nice to have some usable version of gpm around already, though -
the thing is that gpm 1.20.0+ (and maybe a while before as well) generates way
too many debugging messages, endlessly polluting the console and syslog with
them etc - it's only cosmetical problem, but it scares most users and they'll
go back to the old gpm, I think.

-- 
 
				Petr "Pasky" Baudis
.
This host is a black hole at HTTP wavelengths. GETs go in, and nothing
comes out, not even Hawking radiation.
                -- Graaagh the Mighty on rec.games.roguelike.angband
.
Public PGP key && geekcode && homepage: http://pasky.ji.cz/~pasky/