hachoir 3.2.0 (2022-11-27)

  • Fix hachoir-grep command line parsing.
  • PYC parser supports Python 3.12.

hachoir 3.1.3 (2022-04-04)

  • The development branch master was renamed to main. See for the rationale.
  • Replace Travis CI with GitHub Actions.
  • ttf: Support OpenType magic number (OTTO).
  • hachoir-wx: Load darkdetect and test once, fallback if not found.
  • Add hachoir-wx docs.
  • jpeg: Set the size of a JpegImageData with no terminator to the remaining length in the stream to avoid parsing subfields of the JpegImageData if possible.
  • fit: Add parser of Garmin fit files.
  • lzx: Fix LZX decompression.

hachoir 3.1.2 (2020-02-15)

  • Fix a SyntaxWarning in the PDF parser.

hachoir 3.1.1 (2020-01-06)

  • Fix hachoir-wx

hachoir 3.1.0 (2019-10-28)

  • Add __main__ submodule to urwid, metadata, strip, wx. It’s now possible to run python3 -m hachoir.metadata filename for example.
  • Remove hachoir.version submodule; add hachoir.VERSION constant.

hachoir 3.0a5 (2019-04-12)

  • metadata: Fix for calling Dict.values property from MultipleMetadata.iterGroups method.

hachoir 3.0a4 (2019-03-21)

  • Issue #36: Hachoir now requires Python 3.6 or newer.
  • Add “pip install hachoir[urwid]” extra dependency for hachoir-urwi
  • wx: Implement tree view lazy-loading.
  • enhance the documentation
  • urwid: ignore set_focus() error
  • Generate Field docs from soure code.
  • ntfs: parse update sequence array.
  • zip: use magic to identify files.
  • elf: reduce duplication of MAGIC
  • fat: handle the common case where root_start = 2 more gracefully
  • tga: Support 32-bit images. Fixes #39.
  • java_serialized: Support serialized proxy classes.
  • java_serialized: fix support for Externalizable classes

hachoir 3.0a3 (2018-06-04)

Project renamed from Hachoir3 to Hachoir:


  • Add Canon Raw v2 support
  • Add support for creation date in tiff files
  • tga: improvements to adhere more closely to the spec.
  • Add MacRoman win32 codepage
  • tiff/exif: support SubIFDs and tiled images
  • hachoir.metadata: add exportDictionary() method
  • java: parse OpcodeCPIndexShort’s CPIndex argument properly
  • java: have CPIndex opcodes show argument in display (but not value)
  • java: distinguish CPIndex/CPInfo display from value.
  • java: Create value and display for LineNumberTableEntry
  • Enhance the AR archive parser


  • Fix ELF parser (on Python 3)
  • Fix LZX substream in CAB parser
  • pdf: Python 3 compatibility, slightly improved parsing
  • rar: fix TypeError on unknown block types
  • mpeg_video: don’t attempt to parse Stream past length
  • mpeg_video: parse ESCR correctly, add SCR value
  • ext2: fix directory digging

hachoir 3.0a2 (2017-02-24)


  • Add initial parser for Mapsforge .map files (only version 3 supported)

  • Add parser for PAK files from “Project: Starfighter” game

  • Add OS X .bom parser

  • mov parser: add traf entry

  • Update iTunesDB Parser

  • 7zip: Improve and expand 7zip parser

  • tar: Support ustar prefix field in tar archives.

  • Enhance the Java class parser.

  • Added more field to exif extractor

  • Add WIP Mach-O parser

  • ntfs improvements: parse non-resident runlists

  • ext2: support ext4, massive parser improvements

    Huge changeset that may break backwards compatibility, for better consistency and deeper parsing. Basic rundown:

    • Support many more (new) flags added in ext4 and beyond
    • Create nice option displays for flags
    • Improve handling of groups using SeekableFieldSet
    • Parse (demarcate) inode data blocks
    • Consistently use lower case for flag names
  • Enhance mpeg_ts parser to support MTS/M2TS; add MIME type

New features:

  • Python parser supports Python 3.3-3.7 .pyc files.
  • metadata: get comment from ZIP
  • Support
  • Add a close() method and support for the context manager protocol (with obj: ...) to parsers, input and output streams.
  • Add more file extensions for PE program parser.
  • ZIP: add MIME type for Android APK, .apk file.
  • Add editable field for TimestampMac32 type


  • Issue #2: Fix saving a filed into a file in urwid
    • FileFromInputStream: fix comparison between None and an int
    • InputIOStream: open the file in binary mode
  • Fix OutputStream.writeBits() (was broken since the migration to Python 3)
  • Fix ResourceWarning warnings on files: use the new close() methods and context managers.
  • Fix a few pending warnings on StopIteration.
  • Fixup and relocate hachoir-wx, which now works mostly properly.
  • Fix hachoir-parser matroska SimpleBlock
  • Fix Mac timestamp name

Remove the unmaintained experimental HTTP interface.

hachoir 3.0a1 (2017-01-09)


  • metadata: support TIFF picture

Big refactoring:

  • First release of the Python 3 port
  • The 7 Hachoir subprojects (core, editor, metadata, parser, regex, subfile, urwid) which lived in different directories are merged again into one big unique Python 3 module: hachoir. For example, “hachoir_parser” becomes “hachoir.parser”.
  • The project moved from Bitbucket (Mercurial repository) to GitHub (Git repository). The Mercurial history since 2007 was kept.
  • Reorganize tests into a new tests/ subdirectory. Copy test files directly into the Git repository, instead of relying on an old FTP server which is not convenient. For example, it’s now possible to add the required test file in a Git commit. So it’s more convenient for pull requests as well.
  • Port code to Python 3: “for field in parser: yield field” becomes “yield from parser”.
  • Fix PEP 8 issues: most of the code does now respect the latest PEP 8 coding style.
  • Enable Travis CI on the project: tests are on Python 3.5, check also pep8 and documentation.
  • Copy old wiki pages and documentation splitted into many subdirectories into a single consistent Sphinx documentation in the doc/ subdirectory. Publish the documentation online at