On 2 Oct 2006, at 18:14, Roy Schestowitz wrote:
___/ On Mon 02 Oct 2006 12:46:54 BST, [ Matt Mullenweg ] wrote : \___
Elliotte Harold wrote:
It's better to set up separate branches in the source repository
so the crazy fun wild development can run in parallel with the
testing and polishing.
We would continue to create branches for major releases to release
bugfix and security releases as needed.
Unless I misinterpreted something, I think Elliotte was
referring to the need for constant vigorous development
where testing milestones are reached and future extension
carries on simultaneously (not only bugfixes for older
supported releases). A bit like stable and unstable Debian,
or even the Kernel Mm tree that is maintained by Andrew
Morton. I suppose that Fedora and RHEL would be another
example... that in itself signals the need for multiple
patchmasters or maintainers, I think.
That's what I think Elliotte meant as well. I'd throw myself behind
breaking it up into three branches (like Debian), unstable, testing
and stable; however, implementing slightly differently to Debian: so,
we move unstable to testing once we have the feature freeze (then
starting a new unstable branch), with testing replacing stable at the
end of the 4 month release cycle.
- Geoffrey Sneddon.
wp-hackers mailing list