lzh5 implementation was based on an old implementation that implemented huffman tables lookup bit by bit, which is extremely slow. The new implementation instead does byte by byte decoding which is much faster.
We are going to remove lzh5 from official level 2 algorithm soon, but since this work was done on preview, and lzh5 is still going to be used for YM64 files, it's still worth merging.
lzh5 implementation was based on an old implementation that implemented huffman tables lookup bit by bit, which is extremely slow. The new implementation instead does byte by byte decoding which is much faster.
We are going to remove lzh5 from official level 2 algorithm soon, but since this work was done on preview, and lzh5 is still going to be used for YM64 files, it's still worth merging.