* = bug fix
+ = new feature or improvement
! = change that makes the new behavior incompatible with the previous behavior
--------------------------------------------------------------------------------

Release 1.1.1, 2011.04.10
================================================================================
+ Upgraded Apache Commons Compress BZip2 to version 1.1.

+ Upgraded EntityFS to version 1.2.

* Deflated Zip entries can now handle compressed entries that need an extra
  dummy byte that is not included in the entry data at the end of its data to
  inflate properly.

* ZipFileParser now ignores invalid extra field sizes for the central directory
  header if the sizes are too short to be plausible.

+ Added ZipFile.getBackingFile and TarFile.getBackingFile.

+ Added support for empty ZipFiles created from empty files.

* Some Zip files contain an empty root directory entry. That is now ignored.

! Added the postProcessDirectory method on the
  org.at4j.util.archive.EntryExtractionStrategy interface. It is used for
  setting the modification times of directories after they have been populated
  with child entities.

+ Added the preserveModificationTimes property on the
  org.at4j.util.archive.ExtractSpecification object. It can be used to preserve
  the modification times of extracted files and (for archives types supporting
  it (i.e., not Zip)) directories.

+ Added the SignedInteger class.


Release 1.1, 2009.09.22
================================================================================
+ Upgraded EntityFS to version 1.1.4.

+ Updated the documentation stylesheet.

+ Zip's BZip2CompressionMethod now uses At4J's bzip2 implementation.

+ BZip2ReadableFile and BZip2WritableFile now use BZip2InputStream and
  BZip2OutputStream by default, but can be configured to use the Apache Commons
  Compress implementations instead.

+ Added the BZip2WritableFileSettings object for configuring BZip2WritableFile
  objects.

+ Added the BZip2ReadableFileSettings object for configuring BZip2ReadableFile
  objects.

+ Minor performance enhancement in the ArchiveBuilder.

+ Added the -k argument to BZip2 and BUnzip2.

+ BZip2 and BUnzip2 now use BZip2OutputStream and BZip2InputStream by default,
  but can be configured to use the Apache Commons Compress implementations
  instead.

+ Added the BZip2OutputStream class.

+ Added the BZip2InputStream class.

* ArchiveExtractor does not try to copy more bytes than there is in a file when
  extracting it, so it does not fall prey to Java bug #4040920 anymore.

! The assumeMagicBytes property is now always true in BZip2ReadableFile.

! The writeMagicBytes property is now always true in BZip2WritableFile.

! Upgraded the Apache Commons Compress bzip2 classes to Commons Compress 1.0.
  The behavior of BZip2CompressorOutputStream has changed in this release. Now
  it assumes that a bzip2 stream starts with the magic bytes 'BZ'. 

+ Added the UnsignedInteger.fromBigEndianByteArrayToLong method.

+ Added the BitInput and BitOutput interfaces and the BitInputStream and
  BitOutputStream classes.


Release 1.0.2, 2009.07.17
================================================================================
+ The default default entry settings objects of ZipBuilderSettings are now
  read/write.

+ Added the ZipEntrySettings constructor that uses another ZipEntrySettings
  object as a template.

+ Introduced the CompressionLevel enum. Implemented support for it in
  BZip2WritableFile, LzmaOutputStreamSettings, BZip2CompressionMethod,
  DeflatedCompressionMethod and ZipEntryCompressionMethod.

+ Upgraded EntityFS to version 1.1.3.

+ Reimplemented the build script in Ruby.

! Made some filter implementations final.

+ Added the two-argument BZip2ReadableFile constructor.

+ Made it possible to run Emma coverage analysis in run_unit_tests.js.


Release 1.0.1, 2009.03.19
================================================================================
+ Upgraded EntityFS to version 1.1.1.

* The TarExtractor is now better at preserving extracted directories' last
  modification times.

* The TarExtractor now works on a locking file system.

+ Changed the signature of ArchiveExtractor.setFilter to accept a
  Filter<? super ArchiveEntry> instead of just a Filter<ArchiveEntry>.

* The ArchiveExtractor now works on locking file systems when using filters.


Release 1.0, 2009.02.07
================================================================================