Proposal to revitalize release branches: Make them cheap, and make them work.

[Siddhesh Poyarekar]

On 3 October 2014 02:24, Carlos O'Donell <carlos@redhat.com> wrote:
> * The only per-requisite you need to apply a bug fix to a stable release
>   branch is that the same bug fix was applied to master.
>         - no ack from release maintainer required.
>         - obviously you have to do some testing before committing, that's
>           just expected.
>
> * The release maintainer retains veto for patches on a release branch.

This is not strictly necessary.  Any reverts if necessary could be
consensus based with the release manager/Roland making the final call
if there is a deadlock.  That is, it could just be like master except
that the consensus would be for a revert instead of a commit.

> * Post all release branch commits to new libc-stable mailing list.

This should be automated using commit hooks.  On a marginally related
topic, can we expose the commit hook scripts somehow so that more
maintainers can have a look at it and fix them up?  They could use
some improvements with things like ignoring private namespace branches
when doing IRC notifications, better commit announces in the bugzilla,
etc.

> If nobody disagrees I can adjust the Release wiki to indicate the changes.

+1.

Siddhesh
-- 
http://siddhesh.in