One Vs Many ECS Game with Source

I've put together a "one vs many" build using unity ecs.

Play it here:
https://thinkhuge.s3-us-west-2.amazonaws.com/onevsmany/index.html

Source code here:
https://github.com/areilly711/unity_ecs

Questions:
1) I'm using the UnityEngine.Bounds struct for collision in my system jobs. Whenever I update the Translation or scale component of an entity, I have to update the bounds manually as well. Is there a better way to do this?

2) I create a pool of bullet entities at start and re-use them by setting them to active or inactive. The rational behind this was because I didn't want to create a new entity every time the player fires. The problem is, I am checking my isActive flag whenever I use bullets in my jobs, slowing things down. I'm wondering what optimization experiences others are having with pooling vs creating

 

1. Are you bounds different to your mesh? If not have you considered just using the WorldRenderBounds that Unity already keeps updated. Otherwise only alternative I can think of is use local bounds and convert to world bounds when doing collision or recalculate bounds every frame (which is basically what WorldRenderBounds does.)

2. If you're going to pool, why not just used the Disabled component then it won't even appear in queries.

 

tertle:

1) Thank you, I didn't think to use that. Makes things far simpler since the worldrenderbounds will automatically be updated

2) This is a good idea. However, do disabled entities still get processed in job execute functions? It's great that the query doesn't give me the disabled ones, but I still need a way to enable disabled ones when firing

 

Disabled entities will not be returned from any query unless you add disabled to the query. So you can just query the disabled ones and remove the component when up each to spawn another.

This all said, I prefer just creating a new entity (instantiated from a prefab entity) instead of pooling entities. Don't really see what you're gaining from pooling.