While they may be only 'visual' in terms of influence/cell grid,
they all do update CenterPosition, which is essentially the
actual world position of the actor.
'Visual' would imply that it only affects the position where the
actor is drawn, which is inaccurate.
Furthermore, using the term 'Visual' here would make
naming future methods/properties related to visual interpolation
unnecessarily complicated, because that's where we might
need a real 'Visual(Only)Position'.
While individual trait look-ups may be cheap,
if a large army that is currently standing still gets
its first move-including order, the look-ups of dozens
or even hundreds of actors may happen on the same tick.
Therefore this may help reducing that first-order lag spike,
at least a little bit.
At the end of L-turns, actors often end up with an internal facing not 100% matching the direction of the next cell on their path.
As a result, if they haven't reached their destination yet, Move queues a quick Turn as ChildActivity, which previously was not considered as IsMoving.
However, we don't want those mini-turns to interrupt move animations, so we now consider them a move as well. Additionally, to avoid any issues, we make these mini-turns non-interruptible, just like the MovePart activities already are.
While the first tick of the MoveFirstHalf child would run at the parent Move tick (see 2nd-to-last line in Move.Tick), this was not the case for Turn.
As a result, this Move tick would get wasted if a Turn was necessary, which at least contibuted to that visible jerk at the end of each L-turn (actors usually don't have the exact facing needed for the next move at the end of an L-turn).
Add GrantConditionOn*Layer traits
This allows to
- drop some booleans from Locomotor
- drop a good part of the subterranean- and jumpjet-specific code/hacks from Mobile
- grant more than 1 condition per layer type (via multiple traits)
- easily add more traits of this kind for other layers