[Freeciv-Dev] Re: Changing interface for generate_warmap (PR#1108)

[rf13@xxxxxxxxxxxxxxxxxxxxxx]

On Mon, Jan 07, 2002 at 02:09:23PM +0000, Gregory Berkolaiko wrote:
>  --- Raimar Falke <hawk@xxxxxxxxxxxxxxxxxxxxxxx> wrote: 
> > On Thu, Dec 13, 2001 at 05:57:31PM +0200, Gregory Berkolaiko wrote:
> > > Dammit, the attachment is broken again.
> > > Trying from a different address.
> > > 
> > > The patch is updated to todays CVS.
> > 
> > So what is the status of the patch? Do you have an updated version?
> 
> If I remember correctly, you insisted that the change in interface should
> be done together with removal of all old warmap uses.  I felt that this
> project requires not only a lot of work (which is ok) but also a
> long-term commitment (due to average time of a patch acceptance being
> quite long).  I am afraid that I cannot take up the responsibility right
> now, it's a rather hot time for me.  
> 
> However if you consider applying the patch as is, i.e. leaving the
> possibility of global warmap too, I will update the patch and change the
> names of the variables to please everyone ;)

And what about the idea to have the two parallel interfaces?

> > Besides the discussion about the interface I have also other
> > questions:
> > 
> > > +  /* find_the_shortest_path corrupts warmap, cannot be recyled */
> > 
> > Why? Can't this be fixed?
> 
> Warmap gives the approximate distance to all accessible locations.
> When you look for shortest path though, you 
> 1. Take into account more than just geography
> 2. Having found your destination, you stop
> 
> Thus it is two different functions which use same global variables.

So it is more a "find_the_shortest_path regenerate the global warmap"
than a "find_the_shortest_path writes random values into the global
warmap struct"?! However if find_the_shortest_path does regenerate the
global warmap why is this a problem?

        Raimar

-- 
 email: rf13@xxxxxxxxxxxxxxxxx
  The trick is to keep breathing.