Changelist for pdf-renderer 0.9.1

CHANGES SINCE 0.9.0
--------------------------------------------------------------------------

JIRA Issues resolved

 * PDF_RENDERER-130 - Conversion of PDF into an image using pdf-renderer-0.9.0
       has text that is flipped
   PDF_RENDERER-91 -  In some PDF documents, the text is rotated
     
   Text should no longer be flipped and rotated, and the text hscale 
   operator will actually be applied properly (thanks Dave Chapman)
 
 * PDF_RENDERER-125 - PDFPage.getImage() throws IIOException on certain 
       image types
   
   All installed JPEG imageio readers are now attempted and there's also 
   now workaround in place such that YCCK CMYK images can be decoded out 
   of the box on Oracle JREs without JAI imageio (thanks Ben Day)
   
Other fixes/enhancements

 * Use of local TTF Fonts
 
   When a non-embedded TTF font is discovered the named font is searched 
   for on the local system and employed if found. Previous behaviour was to 
   immediately substitute a built-in font, which is still done if a local 
   font is not found. This behaviour can be disabled by setting the system 
   property PDFRenderer.avoidExternalTtf to true. The 
   PDFRenderer.fontSearchPath system property can be used to alter the 
   search path, though Windows and Mac OS X defaults should hopefully be 
   sensible. 
   
 * Built-in font substitution no longer subsitutes italic font when inappropriate
 
 * The system property PDFRenderer.avoidColorConvertOp can be set to true 
   to avoid use of the native-code using ColorConvertOp which has a few 
   bugs that might result in SIGSEGVs on various platforms (e.g., Linux; 
   see https://forums.oracle.com/forums/thread.jspa?threadID=1261882&tstart=225&messageID=5356357 
   where a SIGSEGV can be experienced in libpthread's pthread_join called by 
   libcmm). 
  
 * PDFDocCharsetEncoder initialisation fix
 
   Fixed bug in class initialisation as reported to dev mailing list 
   (thanks Carsten Siedentop)