On bleed, if AllowMovement is true actors with ReturnFire will actually follow the acquired target, unlike in Defend stance.
This is at least unintuitive, since ReturnFire is expected to be more passive than Defend.
- Rename Bits<T> to BitSet<T>.
- Implement set based helpers for BitSet<T>.
- When representing TargetTypes of ITargetable in various traits, use a BitSet<TargetableType> instead of HashSet<string> for better performance & reduced memory usage.
- Fix FieldLoader to trim input values when generating a BitSet<T>.
- Require T in BitSet<T> and BitSetAllocator<T> to be a class since it's just a marker value. This allows the JIT to instantiate generic code for these classes once, as they don't benefit from specialized code for T. (Typically JITs will generate shared code for all reference types, and unique code for every value type encountered).
- Cache active target query.
- Prefer .Count == 0 over Any when working with Lists.
- Use helper for GetEnabledTargetTypes.
- Don't declare target variable until actually needed.
Just a by-product from a previous commit, not meant to do anything specific other than bringing us a tiny step closer to requiring explicit implementations for these interfaces too, at some point.
Added a Damage class to pass damage value and damage(types) instead.
This removes a great amount of overhead and longterm opens possibilities to have damagetypes without warheads.
At least in theory, it's possible for an actor to have multiple AttackFollow traits, or AttackFollow is disabled initially but enabled via upgrade later.
To avoid crashes or other issues, let's play it safe and take the same approach as with AttackBase look-ups.
Most auto target scans will be conducted among groups of allied units unable to target each other because they are allied and their weapons only target enemies. Since they cannot target each other, the scan will be repeated constantly. Realising this, we can significantly reduce the performance impact of auto target scanning by bailing early in this scenario to avoid carrying out expensive targeting checks on friendly units which we know will fail anyway.
Avoid using LINQ for filtering, grouping and dictionary building. Instead, we do these ourselves to reduce overhead.
Actors are checked for AutoTargetIgnoreInfo (rather than the trait) since info checks are marginally faster. This check can now be done immediately to allow us to skip such actors right away.
The ClosestTo calculation is now only run on the group on actors for the selected armament, rather than all potential armaments.
