Re: Linux Stability & cold.system

Hans-Joachim Widmaier (hjwidmai@foxboro.com)
Fri, 21 Nov 1997 11:30:01 +0100


> > That's why the implementors are usually no good testers for their own
> > stuff. They don't want to see it break.

Alan Cox:
> The real reason is much more fundamental than that. Implementors generally
> do test code but they tend to have included in the code all the checks for
> things that they thought might be unusual and break it. Those they didnt think
> of in the code they wont normally think of in the testing

Certainly. But whatever the exact reason is, there's really a need for
a ingeniously devilish test. And I use it as kind of an excuse for me
that the affs is so lousy in many respects. ;-)

-- 
Hans-Joachim Widmaier