Ondřej Žára

Hmm.. why closed? The feature itself makes sense, although a typical Bressenham line algo is indeed NOT compatible with the used shadowcasting approach.

@ardcore that would be nice, thanks. Bressenham as an alternative FOV might be computationally sub-optimal, but useful for LOS situations.

@jokeofweek depends on what you need. The current shadowcasting (both algos) operate on O(R^2) for a radius of R; computing a LOS (center tile vs. one given tile) would be...

Hi @Cleptomania, the current API has no dedicated method for moving, that is correct. The idea behind this design decision is that to actually move a light, you need to...

Hmmm. A brief look into the RNG implementation does not show anything suspicious: the `clone()` call correctly duplicates the state (represented by only 4 internal numbers). It would be nice...

> A simple code repro also does not seem to cause it: To clarify: this small code example *shows* the issue, or works correctly? I am browser-only at the moment,...

Note (not sure if relevant): cloning a RNG does not seem to persist the seed. In other words: ``` RNG.setSeed(123) RNG.clone().getSeed() != 123 ``` But you are not *reading* the...

The complete RNG state can be seen by calling `rng.getState()`. So you can debug by comparing states of your multiple (same-seeded) RNGs during map creation...

I would suggest storing the `weightingCallback` as a private property, not passing it to `_add` and calling `this._weigtingCallback(x, y)` instead.

Hi @uzudil, I do not exactly see the point of having a DOM-based display backend. What is its purpose? `` shall be faster and there are - AFAIK - no...