Changes in 1.1.6 (Oct 30, 2011)
- Fixed bug - deadlock due to unsynchronized broadcasts (bug #876686)
- Prevent deletion of input files on error (bug #874543)
- Document how to compress/decompress from standard input (bug #820525)
- Added more detailed kernel error messages (bug #874605)
- Error-handling fixes for multi-archive scenarios (bug #883782)
Changes in 1.1.5 (Jul 16, 2011)
- Fixed excessive output permissions while compress/decompress
  is in progress (bug #807536)
Changes in 1.1.4 (Apr 22, 2011)
- Fixed hang on decompress with --ignore-trailing-garbage=1 when
  producer is interrupted on trailing garbage (bug #762464)
- Added example to help for decompression piped to tar (bug #746806)
- Fixed typo in Trailing Garbage printed message (bug #746799)
Changes in 1.1.3 (Mar 27, 2011)
- Print trailing garbage errors even when in quiet mode (bug #743635)
- Default extension on decompress of .tbz2 changed to .tar for
  bzip2 compatibility (bug #743639)
- Fixed hang on decompress with --ignore-trailing-garbage=1 and higher
  numCPU (e.g. > 2) (bug #740502)
Changes in 1.1.2 (Feb 19, 2011)
- Fix directdecompress segfault when destination file can't be
  opened (e.g. read-only) (bug #717852)
- Implemented --ignore-trailing-garbage feature (bug #594868)
- Fixed hang on decompress of some truncated archives (bug #590225)
- Pulled an error check out of normal logic block for clarity
- Debug print added after BZ2_bzDecompress to track it's return code.
- A debug print fixed in queue::remove
- Increased max memory usage limit from 1GB to 2GB
- If no -m switch given on command line, default max memory limit
  will now automatically increase from 100 MB to minimum amount
  of memory required to support the number of CPUs requested
- Improved performance when output buffer is full
- Fixed bug which caused hang while decompressing prematurely
  truncated bzip2 stream
- Consumer_decompress throttling modified to prevent potential
  deadlock/infinite loop in certain situations (Thanks to Laszlo
  Ersek for finding and helping track down the cause of this bug)
- Fixed deadlock bug and performance issue when consumer working
  with long bzip2 sequences (Thanks to Tanguy Fautre for finding)
- Fixed error message for block size range (max size was wrong)
- Moved #include <pthread.h> from pbzip2.cpp to pbzip2.h to fix
  OS/2 compiler issue

Changes in 1.1.1 (Apr 17, 2010)
- Modified decompression to use low-level libbz2 API to improve
  performance of long bzip2 streams of large single-stream bzip2
  blocks
- This release should now decompress files created with bzip2
  at least as fast as bzip2 or slightly faster
- Handle decompression of long bzip2 streams incrementally
  instead of loading whole stream in memory at once
- Fixed issue in safe_cond_timed_wait which caused segmentation
  fault when compiled in DEBUG mode
- Fixed issue with Sun Studio compiler - required explicit
  declaration of static const members in .cpp

Changes in 1.1.0 (Mar 13, 2010)
- Added support for multi-threaded decompression using STDIN/pipes
- Added code to support throttling compression to prevent memory
  exhaustion with slow output pipe
- Added -m switch to specify max memory usage amount before
  throttling starts (default 100MB)
- Fixed bug that did not allow command line parameters to be used
  when compressing data from STDIN
- Added long options to man page and -h output
- Added --loadavg, --read long options
- Added support for CPU detection on Win32
- Major improvements to protection of shared variables, error and
  signal handling, program termination
- Added -S switch for thread stack size customization
  (needs USE_STACKSIZE_CUSTOMIZATION defined when compiling)
- Fixed command line parsing bug for -b, -p, -m switches
- Fixed infinite loop when fileWriter fails to create output
  file at start
- Fixed bug that deleted input filename (with .bz2 extension for 
  compression and without .bz2 extension for decompression) when
  a user interrupts process with CTRL-C while outputting to STDOUT
- Fixed bug where 0 byte files were not processed properly when
  data input from STDIN
- Ignores fwrite return and passes chown errors in writeFileMetaData
  if effective uid root
- OutputBuffer usage redesigned as fixed-size circular buffer
- Lots of minor bugs fixed (see AUTHORS or pbzip2.cpp for full details)
- Special thanks to Yavor Nikolov for providing the majority of
  contributions to this release and a significant amount of testing

Changes in 1.0.5 (Jan 08, 2009)
- Now properly complains about trying to read or write compressed
  data to terminal, and exits
- Further fixed CPU detection crash
- Updated Makefile to force deletion when cleaning

Changes in 1.0.4 (Dec 21, 2008)
- Added support to use pbzip2 with tar (--use-compress-prog=pbzip2).
- Added support for all remaining bzip2 command line options so
  pbzip2 can be used as a drop-in replacement for bzip2.
- Fixed bug that would cause pbzip2 to crash if detection of the
  number of processors failed.
- Now prevents uclibc from being exposed to its lack of a
  getloadavg function.

Changes in 1.0.3 (Oct 31, 2008)
- Added support for compression using stdin and pipes!  Thanks
  to Ivan Voras for supplying the patch to enable this feature.
- Added support for decompression using stdin and pipes but
  currently limited to only a single thread
- Added support for testing bzip2 files using stdin and pipes
- Added support to directly decompress files without using
  threads when files are small or the system only has 1 CPU.
  This mode is also used if the .bz2 file contains only 1
  bzip2 stream.
- Added support to compress 0 byte files like bzip2
- Added support for pbzcat symlink to automatically specify
  decompression mode to stdout, just like bzcat
- Increased max supported path and filename from 255 to 2048
  characters
- Updated RPM spec to support SUSE Linux
- Updated help display to show how many CPUs autodetect has found
- Code cleanup

Changes in 1.0.2 (Jul 25, 2007)
- Added support to directly compress files without using threads
  when files are smaller than the specified block size or the
  system only has 1 CPU.  This will speed things up considerably
  if you are compressing many small files.  You can also force
  this behaviour by using -p1
- Added support for pbunzip2 symlink to automatically specify
  decompression mode
- Changed pbzip2 exit code behaviour to match bzip2 for all error
  states (ie: trying to compress a file that already has a .bz2
  extension)

Changes in 1.0.1 (Mar 20, 2007)
- Added #ifdef PBZIP_NO_LOADAVG to remove load average code for
  UNIX systems that do not support it such as HP-UX and OSF1

Changes in 1.0 (Mar 14, 2007)
- Official non-beta release!
- Fixed minor memory leak in queueDelete()
- Added support for UNIX systems to modify max number of CPUs used
  based on load average 

Changes in 0.9.6 (Feb 05, 2006)
- Fixed bug that caused blocks to be missed by decompression
  routine under certain conditions

Changes in 0.9.5 (Dec 31, 2005)
- Changed default output to silent like bzip2 and added -v switch
  to make verbose
- Added support to autodetect number of CPUs on OSX
- Added support to compile on Borland and other Windows compilers
  using pthreads-win32 open source library
- Added decompression throttling in case too much backlog in filewriter
- Fixed bug from patch in 0.9.4 that limited file block size to 900k
- Fixed bug that caused file output to fail with some large files
- Fixed pthreads race condition that could cause random segfaults
- Fixed pthreads resource issue that prevented pbzip2 from compressing
  a large number of files at once

Changes in 0.9.4 (Aug 30, 2005)
- Fixed major bug that prevented pbzip2 from uncompressing files
  created with a BWT block size other than default 900k (oops!)
- Fixed some bugs in the command line parser
- Code cleanup
- Updated RPM spec to match Fedora Extras policy

Changes in 0.9.3 (Aug 22, 2005)
- Added sanity check to ensure integers are the size we expect
- Ported code to compile under MinGW in Windows
- Small code cleanup
- Added ChangeLog
- Added man page
- Added AUTHORS file
- Renamed LICENCE.txt to COPYING
- Renamed README.txt to README
- Updated RPM spec file in preparation for inclusion into Fedora Extras
- Moved ChangeLog info from pbzip2.cpp into this file
- Removed extra spaces/tabs in pbzip2.cpp

Changes in 0.9.2 (Mar 12, 2005)
- Created workaround for a major pthreads problem (on Linux at least)
  that would cause segfaults and missed signals in the pthreads
- Increased size of BZIP2 header match from 56 bits to 72 bits for more
  accurate detection
- Modified some buffers to use the vector class for easier memory
  management and to eliminate one read pass of the input data for
  decompression so should be slightly faster
- Now checks if input file exists before checking if output file
  already exists (when not using -f)
- Now checks to ensure input is regular file and not a directory before
  processing
- Now deletes input file only if output file exists and is greater than
  0 bytes (-k will always preserve input file)
- Now checks input file for bzip2 header before trying to decompress
  file
- Minor code cleanup

Changes in 0.9.1 (Jan 29, 2005)
- Fixed bug: -c switch now implies -k so input file is not deleted when
  data is output to stdout

Changes in 0.9 (Jan 24, 2005)
- Added -c switch for data output to stdout
- Added -t switch to test integrity of compressed data
- Added autodetection for # of CPUs on supported platforms and made
  detected value default # CPUs to use
- Fixed bug where pbzip2 processed 0 byte files
- Tried to make program exit codes the same as bzip2

Changes in 0.8.3 (Jan 9, 2005)
- Added support to maintain file metadata
- Added signal handling to cleanup partial files
- Added support to process multiple files

Changes in 0.8.2 (Nov 30, 2004)
- Added support for large (2GB+) files
- Added version info
- Made command line input more like bzip2
- Now checks if file is already .bz2 first

Changes in 0.8.1 (Nov 27, 2004)
- Fix pthread bug, now works on OSF1, code cleanup

Changes in 0.8 (Sep 6, 2004)
- Initial public release