FreeBSD has serious problems with focus, longevity, and lifecycle

[Freddie Cash]

2012/1/20 Gleb Smirnoff <glebius at freebsd.org>:
> On Fri, Jan 20, 2012 at 02:43:55PM +0100, Damien Fleuriot wrote:
> D> Don't be mistaken, I greatly appreciate the work you put into this and
> D> the time you devoted to fixing this issue which was *a real annoyance*
> D> in our case.
> D>
> D> I'm not saying you didn't merge it Gleb, I'm saying for a loooooooong
> D> time I had to manually patch the 8.2-RELEASE boxes, because for some
> D> reason that I don't know/understand, the patch couldn't (and still
> D> hasn't been, I guess) be merged with 8.2-RELEASE.
> D>
> D> Actually, on topic, what prevents patches from being merged with
> D> -RELEASE, as opposed to waiting for a new -RELEASE bump ?
>
> It cannot be merged into RELEASE! RELEASE is a point on a branch,
> as soon as RELEASE had been released, you can't push anything into
> it, unless you have a time machine.

I think he's asking "what's the criteria to push a patch to
RELENG_8_2, the security/fixes branch of -RELEASE?"  IOW, how does one
increase the patch level (8.2-RELEASE-p5 -> 8.2-RELEASE-p6, etc) of a
release?

I believe, and RE/security team can correct me on this, that the
criteria is along the lines of "security issues and major bug fixes
only".  Perhaps that policy should be looked at and loosened slightly
to also include "driver updates"?  Or something along those lines?
Perhaps look at releasing point releases (8.2.1) like DoubB suggested?

-- 
Freddie Cash
fjwcash at gmail.com