CVS commit by eros: 

PDF 1.6 Annotation Parser and DSs: annotations are loaded from PDF file.
 AFAIK the kpdf team is once again the first opensource reader to get
support for a cool feature. All annotations in pdf file are read and
converted to our internal data structures. Then they'll be rendered on
screen and changed/saved to our xml file.

 Annotations completely supported and parsed from PDF:  [markup], popup,
text, freetext, line, polygon, polyline, highlight, underline, squiggly,
strikeout, stamp, ink. Partial support for: [base], geom.

Annotations:
  Added/Changed the passive annotation data structures to contain and
  generalize the attributes of annotation classes. Missing attributes
  (won't be handled by kpdf): base{P,AP,AS,A,AA,StructPar,OC}, geom{RD}.
  Little changes to the PDF1.6 summary in header.

PDFGenerator:
  Internal parser (discussed with Albert) directly in top of xpdf. Read
  object nodes of type 'Annot' and parse the whole set of attributes as
  defined in the PDF1.6 spec (for supported annotations).


  M +152 -15   annotations.cpp   1.2.2.5
  M +250 -158  annotations.h   1.2.2.5
  M +661 -1    generator_pdf/generator_pdf.cpp   1.20.2.3 [POSSIBLY UNSAFE: scanf]
  M +2 -0      generator_pdf/generator_pdf.h   1.9.2.2