[Freeciv-Dev] Re: patches

[Raimar Falke <hawk@xxxxxxxxxxxxxxxxxxxxxxx>]

On Mon, Sep 10, 2001 at 03:27:34AM -0400, Jason Dorje Short wrote:
> "Ross W. Wetmore" wrote:
> > 
> > Better.
> > 
> > But note that the rotational system allows you to code the
> > DIR_IS_CARDINAL test into the loop parameters so you only
> > iterate over the right ones in the loop code as well.
> > 
> > All the iterate macros are the same tight loop code just with
> > different start/stop/increment values as loop parameters.
> > 
> > And there is a sleazy trick that lets you do this even with
> > the DIR_DX system. Define a mapping array from one to the
> > other, Do the loop over DIR_DX2 indicies, and do a lookup
> > aka transformation whenever you need a "dir" value for DIR_DX
> > i.e. _itr_dir = DX2toDX[_loop_dir].
> 
> Ross,
> 
> I entirely agree that a better system is possible.  But I do think an
> intermediate step such as the patch I proposed is not a bad idea; at a
> minimum it reduces the amount of code we have to deal with in a later
> change.
> 
> 
> Raimar and others,
> 
> Trent proposed such a macro a month or two ago.  I would change it so
> that it is schema-independent.  It would be #defined as something like
> this:
> 
> #define adjc_all_iterate(center_x, center_y, 
>                          itr_x, itr_y, itr_dir, card)
> {
>   int itr_dir, itr_x, itr_y;
>   for (itr_dir=0;
>        itr_dir<8;
>        itr_dir += 1 + card ? DIR_NEXT_CARD(itr_dir+1) : 0) {
>     itr_x = center_x + DIR_DX[itr_dir];
>     itr_y = center_y + DIR_DY[itr_dir];
>     if (!normalize_map_pos(&itr_x, &itr_y))
>       continue;
> 
> DIR_NEXT_CARD returns the difference to the next cardinal direction from
> the one given; in the case of a rotational system it would just be "1"
> whereas for the current system it would be much more complicated.
> 
> A macro like this allows us to write all iterators as wrappers to this
> one, saving dozens of lines of code and making things simpler overall
> (IMO, of course).  It should also be able to retain the efficiency of
> the current code (if we use a rotational system).

I personally see much in unification of the macros. These macros are
only a few and easy in structure. As you pointed you adjc_all_iterate
will only make this easier if we have rotational system. Lets wait
till we have this system.

        Raimar

-- 
 email: rf13@xxxxxxxxxxxxxxxxx
 "Of course, someone who knows more about this will correct me if I'm
  wrong, and someone who knows less will correct me if I'm right."
    -- David Palmer (palmer@xxxxxxxxxxxxxxxxxx)