Re: USB Mouse Problem

[Adam C Powell IV <hazelsct@mit.edu>]

Bastien Nocera wrote:

> Quoting Ethan Benson <erbenson@alaska.net>:
>
> > I wonder if its worth attempting to get makedev fixed and installed
> > into proposed-updates for potato r3, or has everyone already upgraded
> > to r2 --> broke system --> fixed manually ?
> >
> > fixing makedev would at least make it simpler for people to get the
> > nodes (instead of trying to install ruinmysystem.rpm) but it X and gpm
> > are still going to break since makedev can either repoint the
> > /dev/mouse symlink at input/mice which breaks on 2.2.17 kernels, or it
> > can leave it alone and gpm/X break when the kernel is changed to
> > 2.2.18.  fiddling with config files has the same problem (and is a
> > policy no-no anyway)
>
> I think I have an idea for a quite easy fix. Add to the MAKEDEV package an
> initscript that checks the kernel version and makes /dev/mouse point to the
> correct location.
> In pseudo-code:
>
> if kernel_version <= 2.2.17
>   ln -sf /dev/usbmouse /dev/mouse
> else
>   ln -sf /dev/input/mice /dev/mouse
> fi
>
> I don't know if the policy would have anything against it... Comments ?

I don't know about policy, but this would break for people building their own
2.2.18 with old .configs (updated by make oldconfig of course).

A nitpick, I know, but if there's a way to detect HID, it would be better.

Here's one: grep HID /boot/config-`uname --release` and leaving the default
(usbmouse or adbmouse) if that isn't found (i.e. if someone rolled and installed
their own kernel not as a .deb).  Then again, that would only work for the
currently installed kernel; if someone's upgraded to R2 already they likely
fixed manually already, as Ethan said.  For new installs it would work though...

Just a couple of ideas...

-Adam P.

              Welcome to the best software in the world today cafe!