On Tuesday, 14 September 2010 at 11:31, Giel van Schijndel wrote: > On Tue, Sep 14, 2010 at 12:56:53AM -0400, buginator wrote: > > On 9/13/10, Christian Ohm wrote: > >> So how long would this bugfix 2.3 live? Until 2.5 is done, and then > >> 2.4 is fixes only? By what criteria will the jump to 2.5 happen? > > > > I am not really sure what the new criteria will be required for a > > major version bump. > > Is it netcode ? Gfx ? Lobby code ? Other ? Dunno. > > Might I suggest a blasphemous approach? Only *officially* maintaining a > single 2.X branch at a time, i.e. as soon as 2.4 is released support for > 2.3 is dropped. I explicitly emphasised "officially" there to allow for > the case that we have a very simple bugfix which someone is willing to > backport. > > >> How about doing 2.3.5.x releases for bugfixes only instead? So for > >> example we (finally) branch 2.3.5 now (and immediately push > >> CorvusCorax's projectile fixes in 2.3 for 2.3.6), tag (hopefully no > >> more) RCs from it, the release, and the 2.3.5.x fixes. > > > > I can live with this. > > I don't think we have enough manpower to maintain two release branches. > > >> (Well, and there's the question of how well the bugfix only thing > >> will work out, I remember both the 2.2 and 2.3 jumps, and though both > >> times there was some interest in keeping the old line alive in > >> parallel for a while, in the end only the new version was released.) > > > > This is true as well, and since we always lack the manpower to do the > > upkeep that is required, it is possible the same thing will happen > > with the purposed 2.3.5.x branch. > > > > We are going to end up with 3 testing versions, one for trunk (new > > features + bug fixes), one for 2.X (new features + bug fixes), and one > > for 2.3.5.x (bug fixes). > > > > It might not always be possible to just do testing in 2.X, since the > > new feature might skew the results. > > Another radical suggestion: reducing 2.X branches to feature-frozen > branches, i.e. the only time to add new features would be when creating > a new 2.X branch.
How's that different from what I proposed? Sounds about the same to me, except you use 2.x instead of 2.3.x. _______________________________________________ Warzone-dev mailing list [email protected] https://mail.gna.org/listinfo/warzone-dev
