changes in 0.1.3
- added support for MS Write
- added very simple support for KOffice formats (no metadata yet)
- tarfile_nextfile(): stop on EOF
- made compile work with external objdir
- made compile work with glib versions 2.0, 2.2, 2.4 and 2.6
- added --enable-pdf option to configure script (pdf support is optional now)
- added very rudimentary AmiPro text extraction (.sam files)
- added very simple text extraction support for OOXML
- added dBase text extraction (including memo fields)
- 

changes in 0.1.2 (2008-01-23)
- fix fontparser bug for type1c fonts (BUG was introduced in 0.1.1)
- use pdf object generation numbers (fixes extraction of pdf files
  with objects which have a non-zero generation number)
- reduced use of MSG_FATAL in libtexterize
- better quality conversion of msword files
- removed most CFLAGS from configure.ac
- fix crash in html.c when reallocating the meta data array
- fix crash when opening of file failed
- added fuzz testing stream (only for testing/debugging use)
- fixed some crashes found using fuzz testing
- temporary file is not created in current working directory anymore
  (using system-wide temporary file directory)
- pdf: only use the latest xref if multiple occurences of the same
  xref are found. This fixes a bug with "updated" pdf's.
- some last minute fixes which bring linux/macosx at the same level

changes in 0.1.1 (2007-11-14)
- proper extraction of ole metadata in charsets other than CP1252
- fixed up tar file extraction (supports most tar files now)
- convert dates in pdf meta data to iso8601 formatted dates
- fixed compile in case libmagic is a static library
- fixed text extraction of msword documents in case of unicode text
- skip "fields" in msword document text extraction
- added some new flags to magic_open() to make it do less internal checks
- fixed handling of comments in pdf files
- handle Tf operator occuring before BT (Begin Text) operator in pdf files
- fixed small bug in handling multibyte strings in metadata
- fixed bugs in Microsoft Excel SST table parsing
- handle more opcodes in Microsoft Excel (MULRK, TXO, CHSTRING)
- added a basic rtf parser (does text and metadata)
- handle http-equiv meta header which specifies a character set
- fixed a small bug in pdf text extraction which made it fail to decrypt
  a stream without filters

changes in 0.1.0 (2007-10-28)
- fixed some bad bugs in the new ole file extractor
- do pdf decryption in-place to save a malloc (and fix a memory leak)

changes in 0.0.9 (2007-10-24)
- added wordperfect text and metadata extraction
- added id3 (mp3 tags) metadata extraction
- added msexcel/msword meta data extraction
- added pdf metadata extraction
- added ogg vorbis meta data extraction (tags)
- added support for pdf object streams (PDF-1.5/PDF-1.6)
- added powerpoint text and metadata extraction
- pdf text extraction performance improved (using an object cache)
- made texterize library with very simple API
- removed texterize dependency on libole2 see arlib/ole2file.c (for LGPL3)
- relicensed most parts under the LGPL3 (not txtools/ and test/ directories)

changes in 0.0.8 (2006-12-30)
- removed all cvs logging
- basic pdf support (not all pdf types are supported yet)
- initial jpeg support (very unfinished)
- added valgrind.sh and massif.sh script
- streamGetLine() added to arfile.c
- made sure all the test binaries are not installed
- made sure all the non-texterize binaries are not installed
- fixed xmlEscape() to not choke on a NULL pointer
- added memsearch functions to baselib
- added getMemLine function (in new file io.c)
- added get_* functions to binary.c where endianness is a variable
- fixed bug in hexdump() function in message.c
- added verbose/quiet options to texterize
- added format-documentation
- added msexcel text extraction
