Avoid iterating through all the entities about specs HOT 3 CLOSED

For note, my 'Aspects' only defined where they looked in the Aspect cache. If someone in my system creates an Aspect.match<HealthComponent, DamageComponent>, Aspect.optional>() it just returns a single pointer into the aspect cache, the cache itself was a recursive map of maps with keys of component atoms/ids and values of the matched ones with sub-matches with other components requested. You 'should' create any Aspects as early as possible as the first time a certain matchspec is defined it will iterate over all entities to see which go in, but from then on that cache is kept up to date among all aspects that match on the same matchspec.

The old system I based the style off of (Artemis, don't bother looking at it, this was the only good part about its system) used each Aspect as its own cache and they just registered event handlers to hear when components were added/removed, which caused more indirect function calls compared to mine.

Plus duplicated data

from specs.

I just looked in the Tables library a bit more (I still have so much more to learn about rust syntax, I have the ideas, but just the syntax... >.>), looks like it would work very well as a backing store while also using Aspects, they are not competing concepts but rather seem to attack two different issues in their own efficient ways. Perhaps they should both be integrated.

from specs.

Fixed by #29

from specs.