[Freeciv-Dev] Re: (PR#4684) client-side air goto

["Gregory Berkolaiko" <Gregory.Berkolaiko@xxxxxxxxxxxxx>]

<URL: http://rt.freeciv.org/Ticket/Display.html?id=4684 >

On Fri, 23 Jan 2004, Raimar Falke wrote:

> 
> <URL: http://rt.freeciv.org/Ticket/Display.html?id=4684 >
> 
> On Fri, Jan 23, 2004 at 04:01:34PM -0800, Gregory Berkolaiko wrote:
> > 
> > <URL: http://rt.freeciv.org/Ticket/Display.html?id=4684 >
> > 
> > On Fri, 23 Jan 2004, Jason Short wrote:
> > 
> > > > I mean the server-side air PF which is already in.  It wasn't my sole
> > > > effort, of course.
> > > 
> > > I thought server-side air goto still used warmap?  Does the PF have
> > > support for fuel?
> > 
> > No, it never did.  It uses some of the global variables that belong to 
> > warmap, but it's a standalone thing.
> > 
> > It does have support for fuel.  It's build around the assumption that
> > the possible refuel points are "sparse" on the map, so it's better to have 
> > a list of them, rater than "crawl" the map like the big PF does.
> > 
> > I want to reiterate that I never did any timing comparsions (because there 
> > was nothing to compare against).
> 
> The air goto is really a well-known find-shortest-path from vertex to
> vertex but the vertexes aren't tiles but refuel/safe points plus the
> starting position plus the destination position. The reason for this
> is that the finding of a path on a tile level from on of the above
> vertex to another is easy. So you can abstract the tile level away and
> be a lot faster.
> 
> While Jason's patch is a first working version which allows showing
> the path in the client nothing stops us from special casing air-goto
> inside PF. The idea is that you have the same interface (PF) but a
> faster implementation.

That is we basically move current server-side air-PF into common and 
modernize it?  This is not a bad idea.  But who will do it?

G.