New in 1.3:
* Bugs in the MS-ZIP and Quantum decompressors have been fixed.
* MS-ZIP recovery has been improved.
* cabextract should now compile with MinGW

New in 1.2:
* "-t" option: Allows you to test that every file can be extracted without
  having to extract it. This also prints the MD5 checksum of the file.
* UTF-8 filenames are now decoded correctly.
* Large files (>2GB) are now searched correctly for cabinets.
* Problems with the GNU library code have been fixed.
* Several bugs in the LZX and Quantum decompressors have been fixed.

New in 1.1:
* A security vulnerability has been fixed. If the files within a cabinet
  file include the directory ".." in their filename, this will be changed to
  "xx", to prevent directory traversal by a malicious cabinet (e.g. a cabinet
  with "..\etc\passwd", "..\..\etc\passwd" and so on)
* The GNU library code used has been upgraded to the latest from libiberty,
  so it should now work better with Cygwin.
* cabextract should now compile cleanly on AIX.

New in 1.0:
* The libmspack library is used for all the hard work. You should find
  cabextract more robust, and fractionally faster. See the libmspack website
  at http://www.cabextract.org.uk/libmspack/ for more on this library.
* Fixed a bug (now moved to libmspack) where CAB files that had two kinds of
  compression would crash.
* CAB files should no longer get "/" and "\" mixed up, even with non
  MS-DOS cabinet files created by tools like Powerarchiver and Cablinux.
* Now ignores CAB files listed on the command line that have already been
  used as part of a multipart set. You can now type "cabextract *.cab" on
  your Win95 cabinet set, and cabextract will not extract all the files
  sixty times over.
* Now supports reading and extracting large (more than 2Gb) files.
* "-L" option: viewing the archive pays attention to the "-L" flag, and
  shows the filenames in lowercase.
* "-L" option: cabextract now correctly lowercases cabinet files with
  Unicode filenames.
* "-l" option: viewing the archive should always show forward slashes as
  directory seperators.
* "-F" option: Now you can choose which files are extracted from the cab.
  For example, "cabextract -F '*.wav' music.cab" extracts only WAV files.
* "-s" option: Now you can limit which multipart cabinets are used. For
  example, "cabextract -s WIN98_5*.CAB" will only extract files from
  WIN98_50.CAB through WIN98_59.CAB.
* "-p" option: Now you can extract files to stdout rather than disk.
* Now comes with an experimental "wince_info" and "wince_rename" perl
  scripts in the src/ directory, for examining and processing Windows CE
  installation cabinet header files. There is also the "cabinfo" command
  for examining the structure of a cab file. As this information is not
  very useful to most people, none of these three tools are installed.
* Both the English and Japanese manual page have been updated.
* cabextract now alerts you if you try to unpack InstallShield cabinets.
* The definition of Microsoft cabinet files for UNIX's "magic" file, in
  doc/magic, has been fixed, and a definition of Windows CE install cabinet
  header files has also been added.

New in 0.6:
* Quantum method is now supported! Many thanks to Matthew Russotto for this.
* Fixed more build problems on non-glibc systems. I have now followed the
  'FSF approved' method for including getopt_long, so hopefully Solaris
  and all the other OSes won't have a problem building cabextract any
  more.
* Japanese manual page included -- you will have to install the file
  ja/cabextract.1 by yourself, it's not part of the install process.
* Now searches for spanning cabinet files in a case-insensitive way. It
  also searches in both directions for spanned cabinets (each spanned
  cabinet can include a 'previous' and 'next' cabinet filename), so you
  can also specify a cabinet in the middle of a set and get the full set,
  rather than have to work out what the first cabinet is.
* An off-by-one month error in file dates was fixed.
* Cabinets with files beginning with "/" (or rather, "\") no longer
  extract to your root directory.
* A better embeded cabinet search algorithm was introduced for finding
  those elusive cabinets in files and executables :)
* Now you can find more than one cabinet in a file (if there is
  more than one to find), and it will be extracted or listed as normal.
* Files and directories created now honour your umask settings.
* Errors and warnings are now printed to stderr.

New in 0.5:
* Fixed build problems on non-glibc systems.

New in 0.4:
* Fixed a new bug I added by myself to 0.3... :) It corrupts all MS-ZIP
  compressed cabinets after the first 32k, and it should be very obvious
  that corruption has taken place.
* cabextract should be happier building on Cygwin and other such
  architectures -- you should be able to "./configure; make; make install"
  again. This was broken in 0.3 due to the getopt.c / getopt1.c weirdness
  I copied from GNU hello.
* Now prints 'Finished processing cabinet.' after completing the extraction
  a cabinet.

New in 0.3:
* Fixed very rare, but invisible decrunching bug... if you have any
  important things you extracted with cabextract, extract them again
  to be sure they're not corrupt.
* cabextract now tells you if a file isn't a cabinet file.
* cabextract now goes on to the next file, if extracting one fails.
* cabextract now goes on to the next cabinet, if extracting one fails.
* cabextract lets you try to 'fix' some cabinets by skipping over bad
  blocks rather than failing on them. If you have a corrupt cabinet, try
  the '-f' option, and see how it goes.
* Use the new '-d dir' option to extract cabinets to a given directory.
* Use the new '-L' option to turn the extracted filenames to lowercase.
* Use the new '-q' option to be quiet while extracting archives.
* cabextract now prints more information while it's probing multi-part
  cabinets

New in 0.2:
* new even-easier installation: ./configure; make; make install
* Extracted files now have their timestamps set
* Fixed bug which occured when extracting tiny files 
* Fixed completely invisible intel E8 decoding bug... if you have any
  important things you extracted with cabextract, extract them again
  to be sure they're not corrupt.
* cabinet search extended to look through *any* file for cabinets, not
  just MS-DOS/Windows executables.
* Now looks for multipart cabinets in the same directory as the base cabinet.
  This means you can do stuff like 'cd /tmp && cabextract /cdrom/part01.cab'

New in 0.1:
* supports MSZIP and LZX compression methods
* supports split cabs and cabs embedded in executables
* initial release