[Freeciv-Dev] Re: (PR#2370) Path finding

["Raimar Falke" <rf13@xxxxxxxxxxxxxxxxx>]

On Sat, Feb 15, 2003 at 03:00:52PM -0800, Gregory Berkolaiko wrote:
> Quoting Raimar Falke <rf13@xxxxxxxxxxxxxxxxx>:
> 
> > On Thu, Feb 13, 2003 at 09:48:04AM -0800, Gregory Berkolaiko wrote:
> > > Please put it back then ;)
> > 
> > Done. Under the name of TM_CAPPED.
> 
> Good name.
> 
> > I also removed the bool return type of *construct_path*. Added some
> > die()s. Added some comments. Move pqueue to common. Add omniscience
> > handling.
> 
> I rather liked bool return of construct_path, but I can live without it.

Why? The return value is in path->is_valid. If we get too long paths
we have the following solutions:
 - increase MAX_PATH_LEN (easy but increases memory usage)
 - make it dynamic (cleanest but more work)
 - handle it somehow in the caller (I like this least)

> I didn't notice any extra omniscience handling.

Just look for "omniscience".

> There is a strange change here:
> ==========================================================================
> diff -Nurd -X clean/diff_ignore clean/common/shared.h path/common/shared.h
> --- clean/common/shared.h     Fri Feb 14 17:28:53 2003
> +++ path/common/shared.h      Sat Feb 15 11:14:31 2003
> @@ -19,7 +19,7 @@
>  #include <posix/be_prim.h>
>  #define __bool_true_false_are_defined 1
>  #else
> -#ifdef HAVE_STDBOOL_H
> +#ifdef HAVE_STDBOOL_H2
>  #include <stdbool.h>
>  #else /* Implement <stdbool.h> ourselves */
>  #undef bool
> ==========================================================================

Oops. Left by splint checking.

> and a spurious mention of NOTHROUGH instead of DONT_LEAVE in
> pf_danger_iterate_map.

I leave this to you.

> Otherwise, we only have to fix (x,y) <-> xy conversion and it's
> ready imo.

I leave this to you.

        Raimar

-- 
 email: rf13@xxxxxxxxxxxxxxxxx
  "Real Users find the one combination of bizarre
   input values that shuts down the system for days."