gEDA-dev: magnetic net mode is so damn cool

[Ales Hvezda]

> I've looked into the code and in my Sedgewick book. Is the tile code 
> based on the ideas of that book?

I don't recall using that book for the algorithm; it's a basic
rectangular area spatial reduction/search.

> 
> I'll play with the tile mechanism and see whether it's much faster.
> 
> I'll add a function to the tile code:
> GList* s_tile_get_object_lists(TOPLEVEL *toplevel, int wx1, int wy1, int 
> wx2, int wy2)
> The returned list will be a list of object lists of the tiles. 
> 
> I guess some objects are in more than one tile list, isn't it?

Yes, an object will be N tiles if it spans those N tiles.  Also, when
doing searching for connection points, you have to make sure that you
have covered enough of the adjacent tiles (since object which are close
to a border will only be in their respective tiles). 

-Ales