What's new in hachoir-parser 1.1?

Main changes: add "EFI Platform Initialization Firmware Volume" (PIFV) and "Microsoft Windows Help" (HLP) parsers. Details:

• MPEG audio:
  • add createContentSize() to support hachoir-subfile
  • support file starting with ID3v1
  • if file doesn't contain any frame, use ID3v1 or ID3v2 to create the description
• EXIF:
  • use "count" field value
  • create RationalInt32 and RationalUInt32
  • fix for empty value
  • add GPS tags
• JPEG:
  • support Ducky (APP12) chunk
  • support Comment chunk
  • improve validate(): make sure that first 3 chunk types are known
• RPM: use bzip2 or gzip handler to decompress content
• S3M: fix some parser bugs
• OLE2: reject negative block index (or special block index)
• ip2name(): catch KeybordInterrupt and don't resolve next addresses
• ELF: support big endian
• PE: createContentSize() works on PE program, improve resource section detection
• AMF: stop mixed array parser on empty key

What's new in hachoir-parser 1.0?

Changes:

• OLE2: Support file bigger than 6 MB (support many DIFAT blocks)
• OLE2: Add createContentSize() to guess content size
• LNK: Improve parser (now able to parse the whole file)
• EXE PE: Add more subsystem names
• PYC: Support Python 2.5c2
• Fix many spelling mistakes

Minor changes:

• PYC: Fix long integer parser (negative number), add (disabled) code to disassemble bytecode, use self.code_info to avoid replacing self.info
• OLE2: Add ".msi" file extension
• OLE2: Fix to support documents generated on Mac
• EXIF: set max IFD entry count to 1000 (instead of 200)
• EXIF: don't limit BYTE/UNDEFINED IFD entry count
• EXIF: add "User comment" tag
• GIF: fix image and screen description
• bzip2: catch decompressor error to be able to read trailing data
• Fix file extensions of AIFF
• Windows GUID use new TimestampUUID60 field type
• RIFF: convert class constant names to upper case
• Fix RIFF: don't replace self.info method
• ISO9660: Write parser for terminator content

What's new in hachoir-parser 0.10?

New parsers:

• Microsoft Archive parser (.mar)
• Microsoft Windows animated icon (.ani): based on RIFF file format
• Microsoft's HTML Help (.chm)
• Windows Shortcut (.lnk)
• X11 Portable Compiled Font (pcf)
• Adobe Portable Document Format (PDF)

Major changes:

• Convert many constants to Unicode
• Set charset to ISO-8859-1 for many strings with no charset. Examples: filename in gzip, strings in ID3v1
• MIME type is now in Unicode
• Timestamp are stored as datetime.datetime() object
• Add MAC48_Address and NIC24 parser
• Add IEEE 24-bit Organizationally unique identifiers list

Changes:

• Disable QueryParser fallback feature
• QueryParser accepts "class" tag
• Split Parser in HachoirParser and Parser classes
• OLE2: * Rewrite most of the code using SeekableFieldSet * Support FAT block chain * Able to parse fragmented streams * Add parser for component object and document summary
• MKV: add method to convert date value to datetime.datetime() object
• OGG: validate() checks magic string
• Write PascalStringWin32 class
• Add Win32 LANGUAGE_ID dictionary
• Rewrite GUID class using RFC 4122:
  • Supports differents GUID format versions
  • Able to read timestamp
  • Able to read network address
• iTunesDB: support sort index type and playlist
• BMP: move code to parse image data in a separated function, so code can be reused; fix magic regex (reserved may be not nul)
• EXIF/TIFF: reject IFD entry with more than 300 values
• MPEG audio:
  • Frame.isValid() also checks sync field
  • Add getNbChannel() method
  • findSyncrhonizeBits() uses stronger validation to avoid false positive
  • validate() checks first field name and not just if stream starts with bytes "ID3"
• RIFF: text: truncate to nul byte and use ISO-8859-1 charset
• JPEG: reject invalid component id or quantization index (instead of using a warning message)
• JPEG: support all sort of start of scan (especially progressive jpeg)
• JPEG: add magic string of JPEG starting with Adobe chunk
• Photoshop metadata: add parser for version information
• PNG: add method to get number of bits per pixel and use do not format timestamp value
• PNG: support transparency color
• TTF: Reject chunk with more than 300 names
• EXE: Reject PE program with more than 50 sections
• EXE resource: * PE_Resource now uses SeekableFieldSet * Parse file flags * Read file subtype (for driver or font) * Reject header with more than 300 entries * Stop parser at depth 5 * Write version information parser for NE program

Minor changes:

• GIF: replace image marker warning with a parser error
• IPTC: use charset UTF-8 and not ISO-8859-15
• CAB: validate() rejects file with more than 30 folders and fix misuse of seekBit()
• AU: fix end padding size

What's new in hachoir-parser 0.9?

New parsers:

• ACE, CAB, RAR, MOD, S3M, XM, PSD, Torrent, TTF, PDF, NE, MPEG TS

Changes:

• Add unique identifier and category to each parser
• Use tags to choose the right parser
• Create ParserList and QueryParser classes
• Support magic string as regex ('magic_regex')

Improved parsers:

• 7-zip: parse a lot of headers, just not start and signature headers
• ZIP: support file without file size, support 64-bit structures
• Ogg: support "video" chunk and add function to get last page

What's new in hachoir-parser 0.8.1?

New features:

• Rewrite setup.py: uses distutils by default (instead of setuptools), doesn't depend on hachoir-core
• ICO parser: fixes to support cursors
• Parser use new HACHOIR_ERRORS constant

Bugfixes:

• gzip: fix magic string
• XCF: remove useless exceptions
• RIFF: fix fourcc handler (when fourcc is a string and not Unicode)
• FAT: catch ValueError when using string index() method
• ASF: don't create empty fields and validate() checks header minimum size
• EXE: validate() checks size_mod_512 in MSDOS header, add method to compute content size of MSDOS executable (not PE)

What's new in hachoir-parser 0.8?

New parsers:

• 7-zip archive
• Aldus Placeable Metafile (APM), variant of WMF
• Audio Interchange File Format (AIFF)
• Audio Interchange File Format Compressed (AIFC)
• Linux swap file
• LucasArts Font
• New Technology File System (NTFS)
• Microsoft Enhanced Metafile (EMF)
• Microsoft Windows Metafile (WMF)
• Musical Instrument Digital Interface (MIDI) audio file parser
• Real Audio (.ra)
• Real Media (.rm)
• Truevision Targa Graphic (TGA) picture

New features:

• Add method to compute real content size
• Add magic string to find file start
• Add method to get file extension (file name suffix)
• Add method to choose the best MIME type
• Really better file validation, sometimes use arbitrary limits to detect invalid file. Examples: 50 MB for maximum SWF file size, 6000 pixels for maximum GIF picture width, etc.

Changes:

• Lazy decompression for bzip2 and gzip parsers
• ZIP: add more MIME types and file extensions
• EXE: better PE detection
• Set constant name to upper case
• Always use a tuple for common file extensions
• Bitmap: add padding to pixels if needed, fix size of pixels field
• Tcpdump: display ARP layer info (if any) and reject file if link type is unknown

What's new in hachoir-parser 0.7?

New parsers:

• AMF metadata, used in Flash video
• Flash animation (SWF)
• Flash video (FLV)
• Java class
• Ogg/Vorbis (audio)
• Ogg/Theora (video)
• Reiser file system version 3

Important parser improvments:

• bzip2 and gzip parser are able to decompress file
• JPEG picture:
  • Parse quantization table and restart interval
  • Write stronger validate method
• GIF picture: support image comment, graphic control and netscape 2.0 extension
• ID3v1: support ID3 version 1.1 and 1.1b (track number and genre)
• MPEG audio:
  • Better file validation (less false positive), don't allow padding between frames anymore
  • Fix computation of frame size: now works with MPEG version 2 and 2.5
• RIFF: parse AVI and ODML headers
• Tcpdump: add parser for Unicast (layer 2)

Other parser improvments:

• Photoshop metadata: fix header, "reserved" is a string not four nul bytes
• Bitmap: support version 4
• PNG: add background color parser
• Sun/NeXT audio: add more codec description
• Matroska video container: add ISO 639-2 language names
• EXT2 file system: use bits for file mode (instead of 16-bit integer)

Developer changes:

• Split run_testcase.py in three: download_testcase.py, run_testcase.py for hachoir-parser and run_testcase.py for hachoir-metadata
• Update for hachoir-core 0.7:
  • Use NullBits/NullBytes for nul padding
  • Rename _createDescription() to createDescription()
  • Rename _createValue() to createValue()
• Create function parseStream() to parse a stream
• Palette is now PaletteRGB and is based on UserVector class
• New Parser class based on the simple Parser class from hachoir-core

What's new in hachoir-parser 0.6?

News of version 0.6.2:

• Fix Microsoft Office parser: misuse of new array() function
• Fix SECT.display attribute (convert integer to string)

News of version 0.6.1:

• Fix EXIF parser: SubFile import was missing

News of version 0.6:

• hachoir-parser is now a separated component so it's easier to release new versions and write small bugfix
• New parsers: * 3DO model (by Cyril Zorin) * Abstract Syntax Notation One (ASN.1) * MPEG video * Spider-Man video (by Mike Melanson) * Tcpdump: Ethernet, IPv4, ARP, ICMP, TCP, UDP * TIFF image * ZSNES save (by Jason Gorski)
• Better parsers: * MPEG audio: support padding between frames, better file validation, and guess if bit rate is constant (CBR) or variable (VBR) * Python PYC: rewritten from scratch, now support python 1.5 to 2.5 * ID3v2: support picture in v2.3.0, safer charset code
• Many small bugfixes in ID3, MPEG audio and other parsers

Since hachoir core 0.6 is able to "autofix" more bugs, hachoir-parser 0.6 is even stronger.

Parser list