looking back at this code, I think the Container mixin is pretty bad. This is obviously my fault, I basically knew nothing about PIXI nor even webGL/openGL when I wrote any of this. In any case, I am quite sure that the things the Container mixin does, are essentially done by pixi-layers also, but probably a lot cleaner. So I think that if we care to improve things in the future, we can probably exploit existing layer mechanics more, instead of monkey-patching something together like I did.
Also looking back at the code, using a "mixin" like this is quite questionable in the first place as explained before. Even if not relying more on pixi-layers, I kind of feel like it would be better to provide a helper function that performs this children pushing stuff (for flexibility) and additionally defining a Container class ourselves, which just extends the original Container class provided by pixi. Usage of pixi-shadows would then become a bit less transparent to the user though, which is a bit of a downside.
(from #10 discussion)
if staying with the current approach, consider consolidating them, with the tick logic, into one hook logic as I did here
(from #10 discussion)
if staying with the current approach, consider consolidating them, with the tick logic, into one hook logic as I did here