Re: Are there kernel testing suites out there? We need them.

Dominik Kubla (dominik.kubla@uni-mainz.de)
Sun, 4 Jul 1999 07:02:57 +0200


On Thu, Jul 01, 1999 at 11:14:05AM -0400, Ramana Juvvadi wrote:
>
> A testing suite assumes that the features we want from the kernel
> are known in advance. The problem is that linux is constantly
> changing. For the test suite to have any meaning, it has to
> at least evolve as fast as linux. That is heck of a task.

That is incorrect: Test suites exist to ensure that the implementation
follows the design. All kernel development aims to achieve a certain
goal. Once you have that defined you can write test cases. From then
on test cases will prevent that "fixes" or "enhancements" break the
documented design.

> If you have any ideas how to separate out ther rapidly changing
> features of kernel from the slowly changing ones, you might
> interest some people.

You claim that Linux has rapidly changing features: prove it.
I think you can not do this, because the kernel features are not
changing, they are merely expanding. So once a test for the behaviour
of a feature exists, this test will be useful until that feature
is dropped from the kernel. How many kernel features do you know that
have been dropped from the kernel? Without checking the kernel history,
i know of just three: callout-devices, extfs and kerneld.

Yours,
Dominik

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/