[Freeciv-Dev] Re: freeciv 1.12.1 spec?

[Raimar Falke <hawk@xxxxxxxxxxxxxxxxxxxxxxx>]

On Thu, Dec 06, 2001 at 05:48:23PM -0500, Jason Short wrote:
> Mike Kaufman wrote:
> 
> > All this talk about 2.0 has made me think that we ought to release a
> > 1.12.1.
> > 
> > I propose that we set some specific goals that then release a new
> > version when we've met them. My preliminary proposal is:
> 
> My recommendation would be, instead of (or in addition to) trying to 
> openly influence the maintainers to push in specific directions, you 
> (and others) just try to help out patches that you think move in the 
> direction of 1.12.1.  Reviews are very helpful.
> 
> Once there's a feeling that we're about ready for 1.12.1, I'd recommend 
> forking off a "stable" CVS branch, releasing a beta version, and going 
> from there.  It is very important IMO to allow development work to continue.
> 
> 
> > map topology: I should be able to play wrap_x, wrap_y, flat, or
> > bicylindrically (Jason,Ross,Gaute?)
> 
> Gaute has a plan for hacking support for these topologies into server 
> and client.  I don't know if he'll complete it, though.
> 
> General-topologies (which will allow all sorts of wrapping, plus 
> iso-rectangular and potentially many other cool topology shapes) is not 
> all that far off now; most cleanups have been done and what remains is 
> mostly the implementation of the actual topology system.  It would 
> therefore be a shame IMO to not complete it at this point.  Things 
> remaining to do include:
> 
> - Put through the currently pending (small) patch to clean up 
> isometric-tile math.
> - Provide another (larger) cleanup patch to make more of the overview 
> code GUI-independent (Gaute needs to do this too, I think, so there may 
> be some help there).
> - Provide one final small cleanup patch (or it could be several 
> micro-patches) that fixes a collection of small problems through the code.
> - Implement and use the function find_representative_map_pos (or 
> unnormalize_map_pos).  Ross and I still have to bash heads on the 
> implementation (see the earlier discussion on "candy-cane wrapping"), 
> but we're agreed on the basic idea.
> - Add the infrastructure for the topology system.  This includes the 
> topology structure itself, information in the savegame about it, and new 
> packets containing the data.  If there was ever a time to break 
> backwards-compatibility for the protocol (which I believe Raimar has 
> been waiting for), this is it.
> - Implement different topologies by changing the backend topology code.

You forgot mapgen. At least I don't see this as a micro problem.

        Raimar

-- 
 email: rf13@xxxxxxxxxxxxxxxxx
  Living on earth may be expensive, but it includes an annual free trip
  around the sun.