"Peter J. Braam" wrote:
> Our branching was 4.3.14 to 4.4.  We should at some point start using the
> tags in the files -- true -- it sounds like I can learn some things about
> release engineering from you. 

Mostly the things you've probably encountered as well, it's not easy.
Especially when the release and development versions start to diverge,
(and forgot to tag the released version in the cvs tree).

> Your patch will go into the next release -- we do actually test stuff a
> bit; the head of our tree is quite different at the moment.  If I had put
> it in we would have had to re-test on all platforms (FBSD 2.2.5, NBSD 1.2,
> 1.3 and Linux).  

So it's entirely unlike linus'es `it compiles here, have fun guys'
strategy for kernel development. And I was hoping for the latest,
greatest, and unstablest version with a whopping amount of bugs to
squash.

Oh well, I guess there are others who prefer things to work.

> There are not a lot of new features since 4.2.?.  As mentioned the setup
> situation is much better and a lot of little bugs were fixed.  The kernel
> code for 2.1.?? has changed a lot.  I expect 4.6 to have many more changes.  

You're being a bit modest, as I recall libc6, freebsd, and win32 were
not yet supported in the 4.2 code base.

Jan