BTLZAZIP - cr88192/bgbtech_misc GitHub Wiki

File contents consist of a series of markers. These markers express compressed file contents in-series. Markers are byte aligned within the file.

Types:

• WORD: 16 bit word, big endian.
• WORD24: 24 bit word, big endian.

Markers:

• 0xE1, len:WORD24, data:BYTE[len-4]
  • Compressed data block, contents use a Faux Zlib header.
  • This is basically Zlib format, but with 9 for Deflate64, and A for BTLZH.
  • Each E1 block is encoded independently of the preceding blocks.
• 0xE3, len:WORD24, tag:TWOCC, data:BYTE[len-6]
  • TWOCC Marker.
  • TWOCC='ST' means stored data.
• 0xE4, len:WORD24, tag:FOURCC, data:BYTE[len-8]
  • FOURCC Marker.
  • FOURCC='BTLZ' means header.
    • flags:WORD
    • blockSize:BYTE; log2 of logical block size
    • windowSize:BYTE; log2 of window size.
    • if(flags&1)
      • name:ASCIIZ4; ASCII name padded to multiple of 4 bytes.
      • This is the name of the original file which was compressed.
• 0xEC, len:WORD40, data:BYTE[len-6]
  • Compressed data block, contents use a Faux Zlib header.
  • Basically like E1 but with a larger maximum block size.
• 0xED, len:WORD40, tag:TWOCC, data:BYTE[len-8]
  • TWOCC Marker.
  • TWOCC='ST' means stored data.

Possible:

• TWOCC='BI'
  • Incremental Block
  • With this block, the end of the prior block serves as a preset dictionary.