Open robinlinden opened 8 months ago
It looks like there are enough variants of LZW that writing one implementation that can be used for compress
'd data, GIFs, PDF, etc may be tricky.
Didn't notice you'd assigned yourself this. I worked a bit on this back when I was setting up the GIF stuff, but I don't recall what state it's in.
My plan back then was that archive/lzw would just be plain lzw but with support for either being big-endian or little-endian, and then just make sure that there is a method for the GIF-encoder/decoder to be able to handle the control commands and e.g. reset the dictionary or whatever else needs doing.
Oh, I didn't realize you already had some code written. I'll let you finish this one, and I'll do zstd in the meantime.
This is needed both for GIF decoding and
Content-Encoding: compress
.