Re: Lets get branches names right this time (was: Re: Proposed release process/plans)
- From: Gediminas Paulauskas <menesis delfi lt>
- To: Bill Haneman <bill haneman sun com>
- Cc: Chema Celorio <chema ximian com>, Havoc Pennington <hp redhat com>, desktop-devel-list gnome org, gnome-hackers gnome org
- Subject: Re: Lets get branches names right this time (was: Re: Proposed release process/plans)
- Date: Thu, 13 Jun 2002 20:03:11 +0200
Bill Haneman wrote:
On Thu, 2002-06-13 at 18:38, Chema Celorio wrote:
On Thu, 2002-06-13 at 12:27, Bill Haneman wrote:
The problem with this scheme is that 2-0-0 is "two branches removed"
from HEAD, which is a significant problem in that important patches
(i.e. for stoppers) have to be committed to three places, and ordinary
bugfixes to two places (i.e. 2-0 and HEAD). That's a bit burdensome...
which is why I propose to hold new features until 2-0-1 branches, or
else put them on their own branch.
Yes, the problem here is that each module is different. It is hard to
say to hold new features for all modules, they are in different
development cycles and stability (I'm thinking libgnomeprint).
Does anyone have an issue with me using gnome-2-0-0 for
2.0.0-release-only code, and continuing to use HEAD for 2.0.1 until it
enters a "deeper" freeze? That way at least the branch names for the
actual release candidates match up.
Branch at latest time possible. If you do not add new features, only
bugfixes which may go to stable, do not make a branch at all.
Second-level branch (2-0-0) should be used only for short period just
before the release.
Most of the applications will be dveloped only on HEAD I guess, such
deep branching is relevant only for very active modules which have
developers working on future features.
] [Thread Prev