Re: Variable tileset size

[Jens Granseuer <jensgr@xxxxxxx>]

On 01.03.2005 22:21, Matthias Grimm wrote:
> On Tue, 01 Mar 2005 17:25:38 +0000
> Jens Granseuer <jensgr@xxxxxxx> wrote:
>
> > > > Looks more like width / 4 + 1... :-P
> > > 
> > > Ooops. Thats right.
> > 
> > No, looks better, but is still wrong. It should really be
> > 9 * (width / 32) for the current outline. But I'm thinking about
> > defining
> > that in the set as well, so the engine could theoretically support
> > different shapes (still have to be hexagons, though).
> 
> I created three hexagon masks before I know about the shift value and
> the formula and 'width / 4 + 1' match all of them. Your new formula will
> only match your default tileset but none of my bigger masks. I might be
> really the best to define this value in the tileset.

Really? Off the top of my head, it's wrong for 64x56 where it should
be 2x9=18 when scaling, but is 17 using your formula. But it's true
that mine only works for multiples of 32 (the minimum?), not for
intermediate sizes.

> > Might be the best solution indeed, although I think I'd prefer
> > on-the-fly
> > scaling for unavailable sizes to not bloat the sets.
> 
> Agreed. This would also allow smooth zooming not only in two or three
> steps. But how would you handle the mask? If you scale the mask as well
> the risk to get gaps between tiles increases rapidly.

Maybe we should require 32x28 and allow only multiples of that. That
way we can make sure there are no gaps.

Jens