Update FileSystem search cache to support cache misses & be faster #108
Description
On Windows saving the search cache can be quite slow, and there cache misses are also utterly expensive.
So maybe a way for a cache to skip files where it knows they don't exist would be useful
- Implement Cache miss support
A per session cache (not saved) that keeps track of when a file failed to be found.
If a file system write is done, we should check if the file is in our cache miss list, and if so, we remove it to ensure that if a file is written that was missing before, it then can be loaded, avoiding possibly blocking perfectly fine reads. - Implement cached material loading
By default a model has multiple paths for materials to search through, we could save this into the search cache so that when a model is loaded we already know which model path to use for the material saving lots of cache misses.