Release process

C. Scott Ananian cscott at
Fri May 23 00:45:26 EDT 2008

On 5/22/08, Michael Stone <michael at> wrote:
> Marco and I (as well as Dennis and Bernie) had some long chats at the
>  beginning of this week about how to work together to pull of the next
>  release. At Marco's request, I've posted one important chunk of this
>  conversation at
>  Please comment freely.

I think that your desire to always provide a stable build is
admirable, but you need to be much more careful not to let it
interfere with development.

A broken build is not a bad thing, as long as there is still a stable
version for people to use.  Linus' philosophy of development is valid
here: the way to get bugs fixed is to get the code distributed.

In my opinion, if the remaining bugs in the FC9 branch are small
enough not to interfere with the sugar developer's work, there is no
reason why they shouldn't start working with it.  The more people
looking at that branch, the more people will be able to help Dennis
pummel the bugs out of it.

I'm all for not landing broken code into joyride (which makes joyride
the 'testing' branch of the debian triumvirate), as long as people
have unstable and experimental branches to land their works in
progress for broad review.  (As I proposed elsewhere, I actually
encourage the creation of *lots* of long-lived unstable branches with
different maintainers and focuses (sugar, cerebro, X, etc) and
well-defined merge windows.

                         ( )

More information about the Devel mailing list