DEFLATE-Lite Capstone Checkpoint

1. DEFLATE combines:

LZ77 tokenization + Huffman coding BWT + arithmetic coding LZW + run-length coding Only Huffman coding

2. Canonical Huffman codes are rebuilt from:

Code lengths The original tree shape Raw probabilities The input bytes

3. An end-of-block symbol is needed to:

Tell the decoder when to stop Improve entropy Avoid prefix codes Reduce header size

4. In this pack, a distance symbol represents:

distance - 1 distance + 1 a literal byte a match length

5. Bit order must be:

Consistent between encoder and decoder MSB-first in encoder only LSB-first only Unspecified

6. The header stores:

Code lengths for literal/length and distance tables The original input size Only the Huffman tree The LZ77 tokens