- Don't call string.Format in Actor.ToString since it gets called often, instead prefer a simple concat.
- In HiddenUnderFog.IsVisible, avoid a needless level of lambda indirection.
- Made Array.IndexOf available via extension method.
- Made ToHashSet extension method.
- Change collections queried often via Contains into sets.
- Avoid Count() extension if Count or Length property exist.
- Made Count() > 0 checks and variations calls to Any() instead.
- Don't call ToList/ToArray if there is no benefit to materializing the sequence.
- If the sequence does benefit from materialization, follow this general pattern:
- Collection queried often via Contains use ToHashSet to speed up lookups.
- Short lived variables use ToList. This is because ToArray requires an extra copy to output the final size.
- Collections persisted into fields or for a long time use ToArray to minimize memory overhead.
A globally shared sheet builder leaks memory and resources between mod switches. Instead, we create and inject the sheet builder during mod startup to ensure we still share the builder across all fonts, but can reclaim it when the mod is unloaded.
The HardwarePalette will now grow its palette buffer and texture in power-of-2 increments. This avoids it having to allocate memory for a full 256x256 texture up front. In practice the default mods use 22 or 23 palettes so a 32x256 texture is used. This means both the buffer and texture save neatly on memory. Additionally, HardwarePalette.ApplyModifiers sees a nice speedup as it has to transfer a much smaller amount of memory from the buffer to the texture.
To facilitate this change, the MaxPalettes constant is no more. Instead the PaletteReference deals with the calculation of the index and this is passed into the appropriate methods.
The OrderBy overload that takes an int generating key selector is faster than the one that requires a custom comparer. We extract a function from the ScreenZPosition function that determines the Z position of a WPos with an offset, but does not account for the tileset height. For the ordering function this is fine as only the relative magnitude of the comparison keys matter, so we don't need to spend time adjusting for the tileset height, as that won't affect the sort.
To resolve the ambiguity introduced when the introduction of isometric maps meant that cell and map coordinates were no longer equivalent, a new type has been introduced so they can each be represented separately.
Automatically formatted all files via VS. This generally corrects indentation, removes trailing whitespace and corrects misplaced tabs or spaces. Manually tweaked a few files where required.
The TraitsImplementing<T> performs a dictionary lookup to match up its generic type parameter with the right trait collection. Since actors are rendered so much, it is useful to cache this result to avoid looking it up repeatedly.
Avoid allocating memory and resources for graphical elements that will never be drawn when starting a dedicated server. This reduces the server memory footprint by approx 17 MiB.
The bin partitioning in ActorMap worked by dividing the map up into a few chunks of cells, each of which would contain some actors. Unfortunately, the bins were accessed directly in world coordinates which are on a scale 1024x greater than cell coordinates. This lead to all actors being placed into the bottom right bin. When checking for actors in a box, only this bottom right bin would be iterated for actors. Thanks to the fact this bin indeed contained all the actors, some clamps on the input ranges and filtering required per bin anyway, this actually returned correct results. Effectively, it was as if there was no spatial partitioning at all.
Not surprisingly however, this is fairly inefficient. By correcting the spatial partitioning to actually partition we see a 7x speedup in ActorsInBox on the RA shellmap.
