Question Custom Spawn Over Distance

I tried to create custom spawn over distance based on this thread.
The problem is I fail to understand the trick with target position. To be precise I understand the concept and the problem with initial position, but I can't understand some random behaviours.

There are projectiles spawned + pooled - all of them leave trail behind. The default spawn over distance block produces something like this:

Start position of trails is almoast always the same and segments correctly spawn during movement.

I tried advanced example from the link (the last one) and it has no problems with initial particles, but start position is often randomly offseted:


Meanwhile I created my own simple version, but without targetPosition trick. For some reason it produces lines from world origin randomly (trail going up):

Spoiler: Graph

I don't understand why this happens only to ~20% of projectiles. On top of this I tried to experiment and find the cause, by coloring particles to red when newLoop is true, but there was not even one red particle.

I tried to remove initial event or created custom script that sends play event and sets position via event attributes, and this "fixed" the issue I think, but when I tried this (it runs just after projectile is spawned):

Code (CSharp):

1. eventAttribute.SetVector3("position", transform.position + Vector3.forward);

Only some of them received offset, while I expected all of them to have it:


What am I missing?

 

Trail graph

 

Thank you very much for insight,

Yeah, actually I suspected I skipped the first frame, when I forced spawn count to 1 on new loop there are red particles, however I think not all trails have it.

I expected only the first segment (or part of trail) to be red and my logic was that when I spawn/activate object, it reinits vfx and everything is set to zero. The object is spawned in the target/final position (I don't spawn in zero and move), so when initial event runs I expect newLoop to be true, so old position should be the same as current one and even if there are no particles to spawn the value should be correct in next update.

If I remove delta it condition I think it works, but in this case the whole thing becomes even more confusing - the original spawner has those checks (for treshold, but still there) and it works correctly, meanwhile recreating the same logic inside graph produces completely different results?

I will try to create some simple example as the whole thing is integrated with projectile pooling system with simulation running on jobs. Maybe I will find out it is fault of my code, no clue.

 

Ok, I created demo. There are 3 spawn over distance and here are my observations:
1. Original spawn over distance does not start exactly in the spawn place, but this is kind of expected and I think there are no huge differences. It works as expected and looks quite smooth, no errors.
2. My "Custom" spawner looks similar to original one, but has the problem with the initial particle randomly (version with deltaTime).
3. The advanced spawner (copied from your post) "Spawn Dist" has no origin problems, but it seems to "lag" in some way - the beggining and the end of trail has some kind of flickering behaviour, like it had lower framerate or spawned segments less often, also the start position is more random (offset) like I described in the first post.

My current unity 2022.2.16

 

Just in case attaching gif with flickering:

If you decrease size of projectile sphere it will be easier to notice the difference.

 

Oh my, yes this is the actual cause. The funny part is I already tried the same thing before, but I did put that block as first to execute, like this:

and all trails disappeared, and I tried it just now and the same thing happens
I fail to see how this difference affects the graph logic - is it just me or it's a bug, perhaps even the same one?

 

Phew, so it looks like I am still sane, thank you very much for help