[clug] Causing a kernel oops

[Darren Freeman]

On Tue, 2003-09-02 at 15:01, Simon Fowler wrote:
> On Tue, Sep 02, 2003 at 01:40:26PM +1000, Andrew Pollock wrote:
> > On Tue, Sep 02, 2003 at 01:34:16PM +1000, Martin Pool wrote:
> > > 
> > > It looks just like this:
> > 
> > I remember seeing something involving ASCII art of something pulling a 
> > face, did that get done away with at some point?
> > 
> That's still there, in the SPARC oops handler (at least in 2.4 - I haven't run 2.6 on a SPARC, so I can't say if it is).
> 
> DaveM has a strange sense of humour ;-)

When you've just lost your last hour's work, you really do need a
cartoon to cheer you up. Think of all the suicides this has prevented!

Everybody hates the "blue screen of death", not *just* because it
happens so often, but also because it's boring.

We should compile-in 100 ASCII-art cartoons (in an obfuscated form of
course) so that the only straightforward way to see them is to get an
OOPS(). Then developers would have something to get excited about after
they claim to have seen *all* of them ;) You could have a counter in
NVRAM so that you only get them in order. Maybe they could be like a
comic book, so that a story unfolds as you mess up your system more and
more.

Or you could code in a game of ASCII-tetris that can only be played from
an OOPS()... A panic would have ASCII-asteroids of course, complete with
PC-speaker sound effects. Who wouldn't want that hard-coded into their
high-reliability server?

I'm kinda tempted to go away and implement that now. After all,
monolithic kernels are supposed to have lots of this stuff, aren't they?
;)

> Simon

Darren