package osbx

1.2.4 (development version)

• Adapt to Angstrom 0.7.0 API changes

  • This is completely internal

• Added --only-pick-uid option to rescue mode

  • This allows only picking sbx blocks of specific uid when rescuing
  • This is useful when you only only want to rescue one specific file
• Renamed --only-pick option to --only-pick-block in rescue mode

1.2.3 (newest version on OPAM)

• Message text polishing

• Fixed division by zero error/exception in progress reporting when number of total units is 0

  • The error/exception affects all four modes when dealing with files of size 0 bytes

• Added only-pick option to rescue mode

  • This allows asking osbx to only pick some type(s) of blocks during rescuing

• Added seek position checking code for rescue log file

  • This prevents showing "Unknown error" in rescue mode when logged bytes processed field exceeds the file size
• File UID hex is printed in upper case

• Added from, to options to rescue mode

  • This allows specifying range of file to scan for data rescuing

• Added from, to options to show mode(previously only skip-to option exists for show)

  • This allows specifying range of file to scan for metadata

• Added force-misalign option to rescue and show mode

  • This allows disabling automatic rounding down of from byte parameter

1.2.2 (published on OPAM)

• Optimized decoding code to speed up reference block scanning

  • Avoids decoding the sbx block in full when already has a block of that type

• Added no-meta option to decode mode

  • If enabled, saves time by not looking for metadata block actively

• Added code to handle negative parameters

  • decode

    • show-fail-max

  • show

    • find-max
    • skip-to

1.2.1 (published on OPAM)

• Optmized CRC-CCITT

  • Changed fake_uint16 to use int instead of int64
  • This makes encoding and decoding significantly faster
  • Other block related operations are also faster in general

• Added silent option to all modes

  • Only affects progress text printing

1.2.0 (superseded by 1.2.1, not published on OPAM)

• Fixed slightly off progress counting in Decode
• Optimized decoding code to avoid searching for a reference block twice

• Added command line option for decode and show

  • Added --show-fail-max for decode
  • Added --find-max for show
• Updated show command to also show sbx container version
• Added position tracking to show mode
• Added --skip-to for show for skipping to some byte

• Improved progress reporting

  • Replaced fraction with progress bar
  • Made the rate more readable
  • General overhaul to improve quality
• Optimized performance of progress printing code and log writing code

1.1.1 (published on OPAM)

• Removed dependency on Core_kernel

  • 1.1.1 uses a custom replacement of Core_kernel.protect and OCaml Pervasives channels

1.1.0 (superseded by 1.1.1, not published on OPAM)

• Added "show" command

• Switched from using Core to Core_kernel

  • Now osbx can be compiled on Windows

• Switched to a pure OCaml implementation of CRC-CCITT

  • Implementation is translated from libcrc implementation

• Added support for multiple hash functions

  • added support for sha1 sha256 sha512 and blake2b-512
  • in prior versions, only sha256 was supported
• Encoding, decoding now try to be "smart" (see their man pages)
• Fixed Multihash encoding, decoding of BLAKE2 hashes

1.0.1 (published on OPAM)

• Changed rescue mode to use the original bytes of the file, rather than regenerating the block bytes itself
• Fixed jbuild file to work on macOS when building

• Added graceful handling of Ctrl-C breaks in rescue mode

  • log file stays valid after being interrupted
  • in 1.0.0, log file may be empty/broken when osbx is interrupted during rescuing
• Better progress reporting for encoding, decoding and rescuing

• Fixed progress reporting issues with rescuing

  • in 1.0.0, rescue mode only does a progress report when it outputs a block, rather than after scanning the block bytes
  • this makes osbx outputs no log when the scanned file section is large and contains no valid blocks

• Fixed log writing with rescuing

  • in 1.0.0, rescue mode only writes to log file when a valid block is found
  • now it is changed to update log on bytes read, not on blocks written

• Updated file size retrieval function to act correctly when dealing with block devices

  • Changed File_utils.getsize to use Core.In_channel.length instead of Unix.LargeFile.stat
• Added progress reporting for hash in decode mode

1.0.0 (published on OPAM)

• Base version