[gpm] wheelmouse up event not propagated?

[Piotr Sawuk]

Alessandro Rubini wrote:
> 

> Without -R gpm checks the console mode every pair of seconds, and if
> it is in graphics mode it closes the mouse fd, reopening and
> reinitializing it when the current console is text.

Of course this mechanism does work, but for example mouse-movements
during that initialization does cause

*** err [mice.c(1863)]: imps2: PS/2 mouse failed init

and since I have an optical mouse there happens to be some superfluous
mouse-movement from time to time because of dust-particles or whatever
moving beneath. Maybe some delay might help, but this would require
further experiments since the bug isn't repeatable.

Thanks for this great hint anyway, I'll try setting up a command for
switching to non-repeater mode for the rare cases where those additional
thumb-buttons are needed. This did indeed answer all my questions.
> 

> Actually, I run X to have _more_ text mode windows at the same time
> than what the bare console offers; most of my real computers are
> simply headless, with a serial console and a network connection.

As of now I haven't found any possibility to access my bios' fonts
from within X, nor do I know how to make use of other console-fonts
from within X. Also I haven't yet managed to allow for multiple users
being logged in on the same X-server. GPM still is extremely valuable,
especially as I perform all root-tasks from console in order to
stay aware of the fact that I'm messing with my whole system...

Btw, is it true that mouse-support within X is more advanced than
in GPM? I don't have the newest version, but for me exps2 doesn't
seem to work in gpm, but that's what I'm using in X for my mouse...

P