Tried to sketch something using pooling but still seeing resource leaks. Would need to profile Stride again but at this point I wonder if it would be better to checkout Stride and start investigating the issue at the core.
How many instance getting created/destroyed per second/minute are we talking about here?
yes I think in my example the manager is already doing too much.
For the animation you’re making, I don’t see the point of using the pattern we were discussing. Since you could approach this with a spread of LFO or Stopwatch that would just be trigger to run when needed. No need to create/delete anything at runtime for an animation like this, to me it’s overkill.
@lecloneur totally, and actually this method is what I’m using in the original project, but then the RefCnt issue became more prominent and I decided to explore other ways, thinking that it would solve it.