metalgearsloth

I think priorGain in the Gain setter should be interchangeable with what the occlusion setter is doing but they largely don't seem to be the same operation occurring so you...

MIDI lock contention

2

ATM if I want the audio position + occlusion + velocity updates to be as frequent as standard audio this isn't possible because the thread and the parallel job will...

Mainly a reminder for me.

You can run into event ordering issues. Ideally we either: - Get a commandbuffer for events (massive work). - CollisionWake hooks in directly after the events for awake / runs...

_locationChangeBuffer has a high chance of just bulldozing data if you register multiple updates in the same tick so whether or not it works depends entirely upon how you call...

NetMessages allocate a lot

1

https://github.com/space-wizards/RobustToolbox/pull/4582 is a bandaid but ideally we would just have stackalloced Span or something and write into NetBuffer without 1 morbillion allocations and not have to use RecycleableIOMemoryStream.

It ends up spamming logs, ideally it would get the entity treatment.

This is more of a content thing but when you stop playing an instrument it creates an instrument that renders a couple extra notes.

a

To be honest Dynamictree having approx is stupid af. My issue is that all it does is call the AABB method again, when it's mostly going to be the same...