Collision issues with UTProj_BioShot (and possibly static meshes in general)

[Kain-Xavier]

I derived a class from UTProj_BioShot that relies upon the data from a HitNormal to "push" away from the surface. However, I have noticed that the HitNormal becomes unreliable when the projectile is fired onto a flat portion of the ceiling or the floor. To give an example, firing straight up at a ceiling results in a HitNormal of (-1, 0, 0) which as you can imagine breaks my code. (I believe this also breaks Epic's code in regards to which hit emitter to render.)

Oddly enough, UTProj_BioGlob does not have this problem. At first I thought it had to do with the size of the collision cylinder, (UTProj_BioGlob's is smaller than UTProj_BioShot,) but that made no difference. Now I am wondering if it is because UTProj_BioGlob uses a skeletal mesh rather than a static mesh like its parent. Has anyone else run into a similar issue? Should I just go with a skeletal mesh rather than a static mesh? Or should I get rid of the collision cylinder and just enable collide actors on the static mesh?

[Kain-Xavier]

Okay first off, I was mistaken. There is not an issue with the HitNormal when it is fired at a flat surface, at least not initially.

CollisionComponent.SetTranslation( vect(16,0,0) );

The above line of code is responsible for skewing subsequent HitNormal data. Setting the translation back to vect(0,0,0) (the original value,) or even vect(-16,0,0) does not seem to alleviate the issue either. Thankfully, commenting out this line of code seems to have zero effect on collision detection and only positive effects for what I set out to do.

So my issue is solved but this begs a new question, what exactly does this function do? I'm assuming it gets called to shift the collision cylinder to the center of the goo puddle but why would that turn a HitNormal of vect(0,0,-1) into vect(-1,0,0)? And why would setting the translation back to its prior value not fix it?