Re: too much untested code in new kernels

Philip Blundell (
Fri, 3 Jan 1997 14:55:28 +0000 (GMT)

On Fri, 3 Jan 1997, Patricia Cross wrote:

> Roy P. Turner wrote:
> > I may be wrong, but I thought the purpose of the dev kernels, is so that
> > we, the willing, can test them and report bugs to the maintainers.
> Yes, the development kernels are there so we can test them, but I really
> don't think it's too much to ask that the person submitting a change at
> least _compile_ it before turning it in. Why would anyone turn it a
> change that they don't even know works on one single system?

These things happen. Firstly, Linus often doesn't put patches in verbatim
- he does his own tweaks, or forgets to merge it all, and that can break
code. Secondly, it's quite easy to produce code that compiles for the
person who wrote it but fails for everybody else. Thirdly, it's not
unheard of for people to merge in the wrong version of a file by mistake.
And fourthly, if somebody is in a hurry, it can be quite tempting to
submit a change that "looks right" before testing it at all.