ChangeLog

0.29: (not yet released)
  * Rename pdf.py to jbig2topdf.py

0.28:
  * Update to the latest Leptonica (1.68)
  * autotools support, VC++ 2008 solution
  * fix binary file open mode on Windows
  * version info (-V --version)
  * pdf.py now correctly retains DPI from input images (thanks to Steven Lee
    http://blog.rubypdf.com/2011/09/09/jbig2-pdf-py-patch-the-right-way-to-get-dpi/)
  * R. Hatlapatka: option to use autoThresholding. Improved version from
    bachelor thesis JBIG2 compression http://is.muni.cz/th/208155/fi_b/.

0.27:
  * Update to the latest Leptonica (1.58)
  * Fix comments in pdf.py (thanks Dan)

0.26:
  * Update to the latest Leptonica (1.53)

0.25:
   * Should build on Windows if you get a leptonica which does so (thanks to
     Dwight Kelly and a man known only as Dennis)

0.24:
   * Update to leptonica 1.38
   * Add ability to gray symbols but it's commented out because it does little
     good
   * Refinement support broke because of memory savings in leptonica

0.23:
   * Quick release to support leptonica 1.37 (now bundled)

0.22:
   * Added segmentation support to extract text from mixed images
   * Flyspeck removal disabled for now
   * A few minor fixes

0.21:
   * Patch from Alberto Accomazzi to fix the pdf.py script and to output more
     sensible filenames in pdf mode
   * Symbols which are only used once are now included in a per-page dictionary
     (great for scans of old books)
   * Fixed several scale bugs (one at 65K symbols and one at 100K symbols)
   * Page numbers > 256 went wrong