sorry about that. My testing was done with 11836/11837.
I was still able to end up inside other objects, but only when going out of my way trying to do so -- namely, repeatedly bouncing myself off of non-convex areas of the base that tossed me back into the base, or, somewhat more difficult to achieve, bouncing around in a strut-rich area of the factory model until the ship ended up impaled on one of the struts. As I said, however, this required fairly pathological persistence. I was not, in my testing, ever able to end up inside another object based on my first collision with said object. I'm sufficiently satisfied with the above statement that I wouldn't spend too much time worrying (at least at this juncture) about handling any remaining corner cases (unless there's a demolition-derby mod, the sort of behavior required to cause the undesired effects would result in player termination anyway, so I don't think it'll be a big deal).
Collision detection is noticeably improved over the old RAPID collider. I saw neither spurious collision detections when not colliding nor was I able, as with the RAPID collider, to often put 10-15% of my ship through the side of a base before it registered a collision
. As far as I can tell, the new collider is in every way superior to the old one.
I'd be very surprised if the above reports of damage are different when using the old collider, as the damage portion of collision resolution hasn't changed in ages. I seem to recall dying with some frequency when attempting to dock to the Ataraxia fighter barracks, but if someone else can regress and confirm this, that'd be appreciated
. Magnitude of damage can be affected by changes to the collision detection, but only to the degree that the normal returned from the collider is more or less aligned with the relative motion of the objects than the normal returned before (roughly, how glancing the blow is). While it is possible that the new collision algorithm is picking faces that better represent the actual angle of impact, and thus changing the resulting damage, I'd be surprised if the magnitudes were significantly different than before. But -- we don't need to suppose, because we can test
If no one else gets around to it, I can do it later this week.
