But that is what I do.
Talk to people. When a maintainer sends me patches to an area that he
maintains, I just apply it. I read it through, but I _very_ seldom do
anything but just feed it directly to "patch".
The only requirement I have is that the patches aren't too large, and even
that is slightly relaxed for architecture updates. And that's not so much
because I want to read through it (even if I do), but because I know there
will be bugs, and when the sh*t hits the fan I want me and others to be
able to ask things like "ok, did it work with 2.3.15? and you're sure it's
broken in 2.3.16?" and then be able to get some kind of rational end
result of that question.
> But Alan's patches are usually involved in the "kernel" of the "kernel".
> Or maybe more precisely: the generic parts of it. Drivers are a much
> much different matter.
And that is why Alan actually gets a few rejects every once in a while.
> Don't wonder a firm self assured and competent person doesn't feel
> oblidget for the slavish work of the splitting (which is indeed
> plenty of), where it's really not neccesary at all for any technical
> reason other then maybe someones need for total controll.
Note that you don't have to split the patches up. The way it _should_ be
done is that the patches are sent off often enough that they just get
applied.
The "split things up" is something you need to do only when that is too
late, ie things like the current backlog of network driver patches. It
shouldn't need to happen, and in fact most people do not do that (Alan
does it, but Alan does it mainly because he is one of the few maintainers
that are all OVER the place, and he has his own full tree - he's basically
"Linus with a beard").
> I remember the discussion about the tulip where you where arguing
> over and over again in fact against the competence of Donald.
I was NOT arguing against the competence of Donald.
But with both Tulip and eepro100, people have asked me to use drivers that
simply do not work on my machines. What do you expect me to do?
Should I say
"Oh, I'm sure that this driver is really much better than the old one,
despite the fact that the old one works for me and the new one does
not"?
or should I say
"I don't care if you say that it works for somebody else - if it stopped
working for me, then it's probably going to stop working for a lot of
other people too"
I'll take #2 any day, and I don't think you'll be able to convince me
otherwise.
The other reason patches get ignored even from maintainers is if they
obviously haven't been tracking the mainline at all. For example, I (and
others) do occasional cleanups, where we've removed #ifdef's that test for
linux-2.0, because too much baggage is just too much baggage. You're
better off having two versions, than having one version that tries to do
everything (see my pervious post on this issue from another angle; it's
true of software as well as hardware).
Linus
-
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/