]> wimlib.net Git - wimlib/commit
lzx_compress: optimize storing information in lzx_sequence
authorEric Biggers <ebiggers3@gmail.com>
Sun, 15 Jan 2017 21:34:36 +0000 (13:34 -0800)
committerEric Biggers <ebiggers3@gmail.com>
Sun, 15 Jan 2017 23:49:39 +0000 (15:49 -0800)
commitdf84fc50295e41ea1616b68685c6d8dea8ffd84e
tree054fd2d1a7b33f4701f604bf29ecdfa1a801753d
parent831eeaad4513475ce947efaef16fc79ab3997444
lzx_compress: optimize storing information in lzx_sequence

Pack the literal run length and match length ourselves instead of using
bitfields, and store the actual match length instead of the adjusted
match length.  Also make matchlen=0 represent end-of-block, and store
the full main symbol, not just the match header.
include/wimlib/lzx_constants.h
src/lzx_compress.c