Re: 2.6 not cat proof

[Tim Connors <tconnors+linuxkernel1069985926@astro.swin.edu.au>]

bert hubert <ahu@ds9a.nl> said on Wed, 26 Nov 2003 21:10:52 +0100:
> This bug has been seen here over eight years ago and it is back.. linux
> 2.6.0-test4 is still not cat proof :-)
> 
> I found my cat asleep on the warm laptop, it is winter here, and the
> keyboard was dead. Mouse still works, but I had to reboot before I could use
> the keyboard again. Restarting X, which I could do with the mouse, did not
> help.

Are you sure he didn't press alt-sysreq-r or something? Mine has at
various times, pressed ctrl-alt-backspace, ctrl-alt-delete,
alt-sysrq-b (without s and u first, bastard), and the power button
directly. This is all on a laptop, so sysrq requires you pressing fn
first.

One of the cats have also turned my heater on by the touch sensetive
buttons, in the middle of summer while I was out for the day. The
heater, and the power button now have molly gaurds, but the key
combinations are hard to gaurd against.

Read the fvwm webpage - most of the developers of fvwm have cats, and
have, IIRC, documented various ways around their key typing
habits. IIRC, there is a package that at least detects when a cat is
walking across a keyboard, and forces you to click the mouse in a
window under X. Won't stop the kernel crashing, but will do something.

In all seriousness, on my laptop, (do to this and other reasons, I
have stopped using 2.6 for the time being), when I pressed h and j (or
something like that) at the same time, when I realeased one of the
keys, I got some kernel warning like "invalid keyscan" or
something. Never wrote it down, sorry.

-- 
TimC -- http://astronomy.swin.edu.au/staff/tconnors/
When cryptography is outlawed, bayl bhgynjf jvyy unir cevinpl.