This allows the LINQ spelling to be used, but benefits from the performance improvement of the specific methods for these classes that provide the same result.
Previously the StartGameNotification and MusicPlaylist traits used the IWorldLoaded interface to play an audio notification and begin music when the game started. However this interface is used by many traits to perform initial loading whilst the load screen was visible, and this loading can take time. Since the traits could run in any order, then audio notification might fire before another trait with a long loading time. This is not ideal as we want the time between the audio notification occurring and the player being able to interact to be as short and reliable as possible.
Now, we introduce a new IPostWorldLoaded which runs after all other loading activity, and we switch StartGameNotification and MusicPlaylist to use it. This allows timing sensitive traits that want to run right at the end of loading to fire reliably and with minimal delay. The player perception of hearing the notification and being able to interact is now much snappier.
The ignoreSelf flag is intended to allow the current actor to be ignored when checking for blocking actors. This check worked correctly for cells occupied by a single actor. When a cell was occupied by multiple actors, the check was only working if the current actor happened to be the first actor. This is incorrect, if the current actor is anywhere in the cell then this flag should apply.
This flag failing to be as effective as intended meant that checks in methods such as PathFinder.FindPathToTargetCells would consider the source cell inaccessible, when it should have considered the cell accessible. This is a disaster for performance as an inaccessible cell requires a slow fallback path that performs a local path search. This means pathfinding was unexpectedly slow when this occurred. One scenario is force attacking with a group of infantry sharing the same cell. They should benefit from this check to do a fast path search, but failed to benefit from this check and the search would be slow instead.
Applying the flag correctly resolves the performance impact.
When handling the Nodes collection in MiniYaml, individual nodes are located via one of two methods:
// Lookup a single key with linear search.
var node = yaml.Nodes.FirstOrDefault(n => n.Key == "SomeKey");
// Convert to dictionary, expecting many key lookups.
var dict = nodes.ToDictionary();
// Lookup a single key in the dictionary.
var node = dict["SomeKey"];
To simplify lookup of individual keys via linear search, provide helper methods NodeWithKeyOrDefault and NodeWithKey. These helpers do the equivalent of Single{OrDefault} searches. Whilst this requires checking the whole list, it provides a useful correctness check. Two duplicated keys in TS yaml are fixed as a result. We can also optimize the helpers to not use LINQ, avoiding allocation of the delegate to search for a key.
Adjust existing code to use either lnear searches or dictionary lookups based on whether it will be resolving many keys. Resolving few keys can be done with linear searches to avoid building a dictionary. Resolving many keys should be done with a dictionary to avoid quaradtic runtime from repeated linear searches.
This changeset is motivated by a simple concept - get rid of the MiniYaml.Clone and MiniYamlNode.Clone methods to avoid deep copying yaml trees during merging. MiniYaml becoming immutable allows the merge function to reuse existing yaml trees rather than cloning them, saving on memory and improving merge performance. On initial loading the YAML for all maps is processed, so this provides a small reduction in initial loading time.
The rest of the changeset is dealing with the change in the exposed API surface. Some With* helper methods are introduced to allow creating new YAML from existing YAML. Areas of code that generated small amounts of YAML are able to transition directly to the immutable model without too much ceremony. Some use cases are far less ergonomic even with these helper methods and so a MiniYamlBuilder is introduced to retain mutable creation functionality. This allows those areas to continue to use the old mutable structures. The main users are the update rules and linting capabilities.
- In FieldLoader, cache boxed bools and some boxed ints.
- In FieldLoader, presize collections when parsing a List, HashSet or Dictionary.
- In FieldLoader, don't allocate a list of missing items until required.
- In FieldLoader, when a string value is passed, avoid wrapping this in a MiniYaml object by allowing both strings and yaml to be passed in the GetValue overload that does the real work.
- In Animation, avoid allocating no-op actions.
- In VxlReader, use EnsureCapcity to better size the Dictionary.
- In VxlReader change VxlElement to a struct.
- In Locomotor, presize TerrainSpeeds dictionary.
The Harvester trait and MoveAdjacentTo activity called the pathfinder but had a single source and multiple targets. The pathfinder interface only allows for the opposite: multiple sources and a single target. To work around this they would swap the inputs. This works in most cases but not all cases. One aspect of asymmetry is that an actor may move out of an inaccessible source cell, but not onto an inaccessible target cell.
Searches that involved an inaccessible source cell and that applied this swapping method would therefore fail to return a path, when a valid path was possible. Although a rare case, once good way to reproduce is to use a production building that spawns actors on inaccessible cells around it, such as the RA naval yard. A move order uses the pathfinder correctly and the unit will move out. Using a force attack causes the unit to use the broken "swapped" mechanism in MoveAdjacentTo and it will be stuck.
This asymmetry has been longstanding but the pathfinding infrastructure only sporadically accounted for it. It is now documented and applied consistently. Create a new overload on the pathfinder trait that allows a single source and multiple targets, so callers have an overload that does what they need and won't be tempted to swap the positions and run into this issue.
Internally, this requires us to teach Locomotor to ignore the self actor when performing movement cost checks for these "in reverse" searches so the unit doesn't consider the cell blocked by itself.
