Re: How/where do we submit patches now?

Larry McVoy (
Sun, 07 Jun 1998 00:40:53 -0700

: >>>>> "AC" == Alan Cox <> writes:
: >> Isn't that already the case more or less? I carry some changes in my trees
: >> that aren't ready to be released yet, I'm sure you and Dave do too. Probably
: >> the only one who doesn't do it is Linus :)
: AC> Yes but until now I've tried to make sure people dont get my changes until
: AC> Linus had them. If I understand Larry's armwavings right between pieces of
: AC> chinese food then you can sync to any repository with his system and everyone
: AC> has their own tree. Which is incredibly neat but..
: But this forking already happens with Linux Mama etc.
: Basically with lots of patches that are 99% ready for inclusion and
: just need some final comments.
: Linus maintains the official versions, and only after a change set
: passes through the Linus gate should it be part of ``official'' linux.

Exactly. As I said in an earlier message, I'm happy to work on the
mechanism to make this a formal part of the system. I think it is
a good thing (TM) to maintain history on things, even things that
aren't yet integrated.

One thing I haven't worked out is this: suppose someone makes a new
feature, called "Feature", and the basis is Linux-2.1.89. The world
moves forward and by the time the change is ready for integration, the
tree is at 2.1.97. The change is essentially a branch off of 2.1.89
with a merge delta that sucks the change into 2.1.97. So far, so good.
But Linus isn't ready (or the change isn't ready), so Feature is
rejected for the timing being. The tree moves forward to 2.1.101
and the change is resubmitted, same deal. At this point, the delta[s]
that merged the change into 2.1.97 are uninteresting and should probably
be disgarded. This can help several if Linus takes a while to get
to merging. Ideas?

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to