mlange-42/arche: Arche v0.6.0

:rocket: Arche v0.6.0 features fast batch entity creation and deletion, cached filters, and many internal optimizations. :rocket:

• Batch creation and deletion of entities, with up to 4x and 10x speedup, respectively. Even more when combined with World.Reset().
• Cached filters for handling many archetypes and complex queries without slowdown.
• A lot of internal performance optimizations.

• Generic mappers do no longer return all components when creating entities or components (#145)
• Resources API moved out of the world, to a helper to get by World.Resources() (#150)
• World.Reset() does no longer remove the component change listener (#157)
• Removes methods filter.ALL.Not() and filter.ANY.Not(), use NoneOf() and AnyNot() instead (#160)
• World listener function takes a pointer to the EntityEvent instead of a copy as argument (#162)

• Adds method World.Reset(), to allow for more efficient systematic simulations (#138)
• Adds World.Batch() helper for performing optimized batch-creation and batch-removal of entities (#149)
• Adds method Mask.Exclusive() to create a filter matching an exact component composition (#149, #188)
• Generic mappers (Map1, ...) have methods NewEntities, NewEntitiesWith and RemoveEntities for batch operations (#151)
• Batch-creation methods (ID-based and generic) have variants like NewEntitiesQuery that return a query over the created entities (#152)
• Notification during batch-creation is delayed until the resp. query is closed (#157)
• Batch-remove methods (RemoveEntities()) return the number of removed entities (#173)
• Filters can be cached and tracked by the World to speed up queries when there are many archetypes (#178)
• Function AddResource[T](*World) returns the ID of the resource (#183)

• Speedup of archetype mask checks by 10% by checking mask before empty archetype (#139)
• Speedup of generic queries and mappers to come closer to ID-based access (#144)
• Speedup of archetype mask checks by casting filter interface to concrete type when possible (#148)
• Optimized batch creation of entities (#159)
• More efficiently clear the memory of removed components, with 2-3x speedup (#165)
• Do not clear memory when adding entities to archetypes, not required anymore as of #147 (#165)
• Speed up copying entity to archetype by getting entity pointer without reflection (#166)
• Avoid slice allocations in generic mapper methods (#170)
• Avoid type checks in query when iterating archetypes (#179)
• Speed up counting entities in queries with a cached filter (#182)
• Implements a fast and memory-efficient lookup data structure for components ID keys, to reduce the memory footprint of archetypes and the archetype graph (#192)
• Speedup of archetype creation by 40% by using a const for archetype storage page sizes (#197)

• Archetype storage buffers are "zeroed" when removing entities, to allow GC on pointers and slices in components (#147)
• Use slices instead of arrays inside paged archetype list to ensure pointer persistence (#184)

• Adds an example for batch-creation and batch-removal of entities (#173)
• Adds code examples to most public types, methods and functions (#183, #189)

• Restructure and extend benchmarks (#146, #153, #155, #156)
• Add an ECS competition benchmark for adding and removing components (#170)
• Add benchmarks for different ways to implement parent-child relations between entities (#194, #195)