engine-collision

OBB collision detection is implemented by attaching a Collider component to entities and running a CollisionSystem every frame to detect overlaps via SAT and populate Colliding data. This supports movement, triggers, and physical interactions.

Layer-based collision filtering uses Layer and LayerMask structures to restrict which objects interact. This allows developers to separate trigger zones from solid colliders, filtering collisions so only designated layers respond to each other.

Sliding movement and push-back responses are supported through movement integration with collision handling. When entities collide, the system constrains movement, causing objects to slide along surfaces or apply push-back behavior based on solid collider interactions.

Trigger zones are set up by configuring a Collider component in trigger mode rather than solid mode. The CollisionSystem detects overlaps and signals the Colliding component, enabling interactive environments without applying physical movement constraints.

Debug visualization is supported for collider shapes. The system provides debug rendering of OBB-based colliders, allowing developers to visually inspect collision bounds, layer assignments, and trigger zones during runtime.

SAT, or Separating Axis Theorem, is used for collision detection because it reliably determines OBB overlaps by checking for separating axes between shapes. This provides precise collision data needed for accurate movement constraints and physical interactions.