Greatly improve case-insensitive lookup speed

By using hash tables, the algorithmic complexity of the case-insensitive
lookup has been reduced from O(n * k) to amortized O(k), where n is the
number of files in each level and k is the number of directory levels
that need to be traversed.

This massively improves performance on Linux when loading characters,
especially when there are many missing characters, since it is no longer
necessary to scan the entire character folder repeatedly.

0 files changed, 0 insertions, 0 deletions