Re: too much untested code in new kernels

Michael Elizabeth Chastain (
Fri, 3 Jan 1997 05:04:26 -0600

Hello Hackers,

Ok, a couple of people have said that it's unrealistic for Linus
to even *compile* all the code that goes into a 2.1.XX release.

I will try to make my points more clearly:

Point 1: Many patches in the 2.1.XX release are being integrated with
zero testing by the person submitting the patch.

As evidence for point (1), I offer:

2.1.18 modules patch:
prematurely integrated into kernel.

2.1.15 drivers/char/pcwd.c:
looks like it was crunched in a round of patch integration
and never even compiled by anyone

2.1.12 sr_ioctl patch
person who wrote the patch mis-understood the code, broke it
through his mis-understanding, and never executed his code
successfully before distributing it

Point 2: This is a problem.

This is a matter of opinion and it's my opinion that this is a problem.
Specifically, it's a problem for me because: I would like to run current
kernels so that I can pursue my own kernel development on them; and I
have lost faith that all new code in the kernel is getting tested by the
people who wrote it before I see it.

I have my own ideas about solutions for this problem but as the
discussion appears to be centering on discussing whether this is a
problem, I'll keep these ideas to myself for now.

Finally I would like to address the argument 'go run 2.0.XX if you want
a stable kernel'. I don't need a stable kernel. But I do want a kernel
where ALL code gets tested at least ONCE by the people who write it
BEFORE they hand it to Linus.

Is that asking too much?

Michael Chastain
"love without fear"