Better formats

This commit is contained in:
Eric MacDonald
2021-09-08 21:48:48 -04:00
parent 5ccba979c6
commit fda7e0c1b5

View File

@@ -34,16 +34,16 @@ class WTVLzpf {
/** /**
* This is used to encode (one-byte) literals with no previous tracked occurence. * This is used to encode (one-byte) literals with no previous tracked occurence.
* *
* Bytes with best compression: SPACE and LF and e"/<>Tainoprst * - Bytes with best compression: SPACE and LF and e"/<>Tainoprst
* Bytes with good compression: TAB and ,-.1=ABCDEFGHILNOPRSbcdfghlmuw * - Bytes with good compression: TAB and ,-.1=ABCDEFGHILNOPRSbcdfghlmuw
* Bytes that don't change the length of the bit stream: 024:MW_kvy * - Bytes that don't change the length of the bit stream: 024:MW_kvy
* (The rest will increase the length of bit stream) * - The rest will increase the length of bit stream
* *
* I don't know what process they used to build this dictionary. * I don't know what process they used to build this dictionary. I assume they
* I assume they frequency-scanned a bunch of HTML files they had. * frequency-scanned a bunch of HTML files they had.
* *
* Using ISO-8859-1 chracter encoding. * Using ISO-8859-1 chracter encoding. Didn't seem like they used a different
* Didn't seem like they used a different dictionary for Japan builds (ISO-2022-JP). * dictionary for Japan builds (ISO-2022-JP).
**/ **/
nomatchEncode = [ nomatchEncode = [
/* [FLATTENED HUFFMAN CODE, CODE BIT LENGTH] */ /* [FLATTENED HUFFMAN CODE, CODE BIT LENGTH] */