[clug] [OT] With Coders like these, how far can Apple go?

Craig Small csmall at enc.com.au
Tue Jan 15 03:41:59 MST 2013


On Tue, Jan 15, 2013 at 12:29:31PM +1100, steve jenkin wrote:
> Thank goodness these ideas don't propagate in the FOSS world, especially
> the Linux kernel.
> Though Linus will give you a well deserved serve for being a fool and
> attempting to submit code beyond your understanding/capability.
I think there needs to be a clear distinction between hacking on MY
kernel as opposed to having my code put into THE kernel.  If I'm
fiddling around and my computer crashes because I did something wrong,
I've learnt something.  I think that sort of thing should be encouraged.

A lot of the kernel programming is incremental not greenfield (not all
of course).  Creating a new device driver or protocol stack is often a
matter of looking up (what will the protocol/shim want from me, what
ioctls do I have to do) and down (what bit pattern to make the chip do
X). Copy some code across and put way to many printks to figure out
where you are at, then move in small increments.

It's a great feeling when you push that string of bytes into the top
stack and see the hardware do the right thing with it (or vice versa).
It was 18 years ago now for me but I still remember it.

It's hard to tell what Apple are saying (MY kernel or THEIR kernel), but
I hope it doesn't discourage people from poking around and breaking
things; just make sure its not on anything important or unrecovable.

I'd also agree it is way way easier to write stuff in userland than in
the kernel and probably not ideal your second decent program is in the
kernel; but if it scratches your itch why not (just don't submit it in
for the main kernel yet perhaps :)

 - Craig
-- 
Craig Small VK2XLZ   http://enc.com.au/          csmall at : enc.com.au
Debian GNU/Linux     http://www.debian.org/      csmall at : debian.org
GPG fingerprint:     5D2F B320 B825 D939 04D2  0519 3938 F96B DF50 FEA5


More information about the linux mailing list