The Locomotor IsMoving check was allowing us to consider another actor that moving as not a blocker. However for some reason it also considered the actor trying to path being mobile as sufficient for this check to pass which did not make sense. We remove that extra check and inline the method.
This was a regression from 4a609bbee8 which changed the method from IsMovingInMyDirection (which required the lookup of the mobile trait) to just IsMoving. It should have removed the lookup as not required.
This fixes a crash in HPF which was considered the location as blocked when Locomotor considered it unblocked because the logic was not aligned. Removing this check aligns the logic and resolves the crash.
Using the local pathfinder, you could not find a path to an unreachable destination cell, but it was possible to find a path from an unreachable source cell if there was a reachable cells adjacent to it.
The hierarchical pathfinder did not have this behaviour and considering an unreachable source cell to block attempts to find a path.
Now, we unify the pathfinders to use a consistent behaviour, allowing paths from unreachable source cells to be found.
Occupied cells was defined by height yet we didn't update actor map on changing height. This in some scenarios could have caused the aircraft to forget to remove its influence from actor map
- Use LongBitSetAllocator and not BitSetAllocator. Using the wrong allocator means all string based checks and displays would provide incorrect results.
- Remove LongBitSetAllocator.Mask which wasn't being calculated or Reset correctly. We can use world.AllPlayersMask to provide the same effect at use sites.
When crushables and crates change their Location/TopLeft, their crushability is cached, but when their CenterPosition is changed, their cached crushability is not refreshed. Since their CrushableBy functions depends on IsAtGroundLevel, which depends on the CenterPosition, this means that when the crushability is cached it will depend on the current height of the object. If the height of the object changes, the cache is not refreshed and now contains out of date information.
The Locomotor cache and the HPF both cache this same information, but at different times. HPF caches immediately, but Locomotor caches on demand which means there can be a delay. This means they can have inconsistent, differing views of the crushability information. This eventually surfaces in a "The abstract path should never be searched for an unreachable point." error from HPF when it detects the inconsistency.
The bug is that Locomotor was caching information without refreshing it when required. Fixing this to refresh the cache when the CenterPosition changes is likely to have negative performance impacts. As would removing crushability from the cache. These would both be fixes that address the underlying bug.
The high impacts of a proper fix lead us to a workaround instead. If we set the CenterPosition before setting the Location, then when the Location is set and the caches are refreshed, the new CenterPosition is available when caching the crushability information. This means logic depending on IsAtGroundLevel will get the new information and cache a more up-to-date view of things. This means when changing both the CenterPosition and Location together we now cache correct information. However calls that set only the CenterPosition and not the Location can still result in a bad cache state. Although this is imperfect it is an improvement over current affairs, and has less impact.
When the Land activity is run, the aircraft adds influence to the cell so it cannot be used by other actors. When the TakeOff activity runs, it removes the influence so the cell can be used by other actors.
However, when a Carryall picks up a unit, it is told to Land with a vertical offset - it never reaches ground level. When the TakeOff activity runs, it saw the aircraft was above ground level and bailed out. The means the influence is never removed. The cell is now unusable despite the fact the Carryall has left.
To fix this, TakeOff now checks if influence was applied instead of checking if the aircraft is above ground level. If so, we know the Land activity had decided that influence was required, even if the aircraft has not made it to ground level. When TakeOff runs, it will treat it as a proper take off event even though the aircraft is already above ground level. This means influence will be removed and the cell will become accessible as intended.
In ActorMap, we also fix a design flaw where disposed actors where excluded from queries. This caused cache inconsistencies with clients using ActorMap.CellUpdated event to rely on updates. This event will not get called when the actor was disposed, so the downsteam client may have cached the actors at that location, only for them to "change" when the actor is later disposed. This could cause the Locomotor and HierarchicalPathFInder to have inconsistent views of the actors on the map, causing crashes if the inconsistent state broken some internal invariants. The only reason to exclude disposed actors would be to cover up for the actors not being removed properly from the map, which is fixed now aircraft are handled correctly. If ever an actor isn't removed from the actor map, then the caller needs fixing rather than having the actor map exclude it.
By tracking updates on the ActorMap the HierarchicalPathFinder can be aware of actors moving around the map. We track a subset of immovable actors that always block. These actors can be treated as impassable obstacles just like terrain. When a path needs to be found the abstract path will guide the search around this subset of immovable actors just like it can guide the search around impassable terrain. For path searches that were previously imperformant because some immovable actors created a bottleneck that needed to be routed around, these will now be performant instead. Path searches with bottlenecks created by items such as trees, walls and buildings should see a performance improvement. Bottlenecks created by other units will not benefit.
We now maintain two sets of HPFs. One is aware of immovable actors and will be used for path searches that request BlockedByActor.Immovable, BlockedByActor.Stationary and BlockedByActor.All to guide that around the immovable obstacles. The other is aware of terrain only and will be used for searches that request BlockedByActor.None, or if an ignoreActor is provided. A new UI dropdown when using the `/hpf` command will allow switching between the visuals of the two sets.
When the UpdateCellBlocking encountered a transit-only cell (the bibs around a building) it would bail from the loop. This would leave the cellCrushablePlayers set to all players. It would update the cell cache and mark that cell as a crushable location.
When CanMoveFreelyInto would later evaluate a cell, it would consider it passable because the crushable check would pass (cellCache.Crushable.Overlaps(actor.Owner.PlayerMask)) rather than because the transit check (otherActor.OccupiesSpace is Building building && building.TransitOnlyCells().Contains(cell)) would pass.
Although this meant the cell was treated as passable in either scenario, it means the cache contained incorrect data. The cell does not contain any crushable actors but the cache would indicate it did. Correcting this means we can rely on the crushability information stored in the cache to be accurate.
Since bbf5970bc1 we update frozen actors only when required.
In 8339c6843e a regression was fixed where actors created in line of sight would be invisible.
Here, we fix a related regression where cloaked units that are revealed, and then frozen when you move out of line of sight would lack tooltips.
The fix centers around the setting of the Hidden flag. In the old code this used CanBeViewedByPlayer which checks for visibility modifiers and then uses the default visibility. The bug with this code is that when a visibility modifier was not hiding the actor, then we would report the default visibility state instead. However the frozen visibility state applies here which means if the frozen actor is visible, then we consider the actor to be hidden and therefore tooltips will not appear. In the fixed version we only consider the modifiers. This means a visibility modifier such as Cloak can hide the frozen actor tooltips. But otherwise we do not consider the frozen actor to be hidden. This prevents a frozen actor from hiding its own tooltips in some unintended circular logic. Hidden now becomes just a flag to indicate if the visibility modifiers are overriding things, as intended.
Since bbf5970bc1 we only update frozen actor state on demand rather than every tick. However when the actor was initially created we were failing to set the initial visibility state if the frozen actor was invisible.
With this fix, we now set the visibility states on creation correctly. This fixes an issue where enemy actors created within line of sight would not appear.
During the refactoring to introduce HierarchicalPathFinder, custom costs were no longer applied to source locations. The logic being that if we are already in the source location, then there should be no cost added to it to get there - we are already there!
Path searches support the ability to go from multiple sources to a single target, but the reverse is not supported. Some callers that require a search from a single source to one of multiple targets perform their search in reverse, swapping the source and targets so they can run the search, then reversing the path they are given so things are the correct way around again. For callers that also use a custom cost like the harvester code that do this in order to find free refineries, they might want the custom cost to be applied to the source location. The harvester code uses this to filter out overloaded refineries. It wants to search from a harvester to multiple refineries, and thus reverses this in order to perform the path search. Without the custom cost being applied to the "source" locations, this filtering logic never gets applied.
To fix this, we now apply the custom cost to source locations. If the custom cost provides an invalid path, then the source location is excluded entirely. Although this seems unintuitive on its own, this allows searches done "in reverse" to work again.
Previously, actors that were visible would refresh their frozen actor state every tick in preparation for the actor becoming hidden, and the frozen actor appearing as a placeholder instead.
By using ICreatesFrozenActors.OnVisibilityChanged when can avoid refreshing the state constantly, and instead just refresh it the moment the frozen actor needs to appear. This provides a nice performance improvement on the cost on managing frozen actors.
