unleashed-firmware/documentation/file_formats/TarHeatshrinkFormat.md

20 lines
1.0 KiB
Markdown
Raw Normal View History

# Heatshrink-compressed Tarball Format {#heatshrink_file_format}
Updater: resource compression (#3716) * toolbox: compress: moved decompressor implementation to separate func * toolbox: compress: callback-based api; cli: storage unpack command * toolbox: compress: separate r/w contexts for stream api * targets: f18: sync API * compress: naming fixes & cleanup * toolbox: compress: using hs buffer size for stream buffers * toolbox: tar: heatshrink stream mode * toolbox: compress: docs & small cleanup * toolbox: tar: header support for .hs; updater: now uses .hs for resources; .hs.tar: now rewindable * toolbox: compress: fixed hs stream tail handling * updater: reworked progress for resources cleanup; rebalanced stage weights * updater: single-pass decompression; scripts: print resources compression ratio * updater: fixed warnings * toolbox: tar: doxygen * docs: update * docs: info or tarhs format; scripts: added standalone compression/decompression tool for heatshrink-formatted streams * scripts: tarhs: fixed parameter handling * cli: storage extract command; toolbox: tar: guess type based on extension * unit_tests: added test for streamed raw hs decompressor `compress_decode_streamed` * unit_tests: compress: added extraction test for .tar.hs * rpc: autodetect compressed archives * scripts: minor cleanup of common parts * scripts: update: now using in-memory intermediate tar stream * scripts: added hs.py wrapper for heatshrink-related ops (single object and directory-as-tar compression) * scripts: naming fixes * Toolbox: export compress_config_heatshrink_default as const symbol * Toolbox: fix various types naming * Toolbox: more of types naming fixes * Toolbox: use size_t in compress io callbacks and structures * UnitTests: update to match new compress API * Toolbox: proper path_extract_extension usage Co-authored-by: あく <alleteam@gmail.com>
2024-06-30 13:38:48 +03:00
Flipper supports the use of Heatshrink compression library for `.tar` archives. This allows for smaller file sizes and faster OTA updates.
Updater: resource compression (#3716) * toolbox: compress: moved decompressor implementation to separate func * toolbox: compress: callback-based api; cli: storage unpack command * toolbox: compress: separate r/w contexts for stream api * targets: f18: sync API * compress: naming fixes & cleanup * toolbox: compress: using hs buffer size for stream buffers * toolbox: tar: heatshrink stream mode * toolbox: compress: docs & small cleanup * toolbox: tar: header support for .hs; updater: now uses .hs for resources; .hs.tar: now rewindable * toolbox: compress: fixed hs stream tail handling * updater: reworked progress for resources cleanup; rebalanced stage weights * updater: single-pass decompression; scripts: print resources compression ratio * updater: fixed warnings * toolbox: tar: doxygen * docs: update * docs: info or tarhs format; scripts: added standalone compression/decompression tool for heatshrink-formatted streams * scripts: tarhs: fixed parameter handling * cli: storage extract command; toolbox: tar: guess type based on extension * unit_tests: added test for streamed raw hs decompressor `compress_decode_streamed` * unit_tests: compress: added extraction test for .tar.hs * rpc: autodetect compressed archives * scripts: minor cleanup of common parts * scripts: update: now using in-memory intermediate tar stream * scripts: added hs.py wrapper for heatshrink-related ops (single object and directory-as-tar compression) * scripts: naming fixes * Toolbox: export compress_config_heatshrink_default as const symbol * Toolbox: fix various types naming * Toolbox: more of types naming fixes * Toolbox: use size_t in compress io callbacks and structures * UnitTests: update to match new compress API * Toolbox: proper path_extract_extension usage Co-authored-by: あく <alleteam@gmail.com>
2024-06-30 13:38:48 +03:00
Heatshrink specification does not define a container format for storing compression parameters. This document describes the format used by Flipper to store Heatshrink-compressed data streams.
## Header
Header begins with a magic value, followed by a version number and compression parameters - window size and lookahead size.
Magic value consists of 4 bytes: `0x48 0x53 0x44 0x53` (ASCII "HSDS", HeatShrink DataStream).
Version number is a single byte, currently set to `0x01`.
Window size is a single byte, representing the size of the sliding window used by the compressor. It corresponds to `-w` parameter in Heatshrink CLI.
Lookahead size is a single byte, representing the size of the lookahead buffer used by the compressor. It corresponds to `-l` parameter in Heatshrink CLI.
Total header size is 7 bytes. Header is followed by compressed data.