A rather massive check in will follow these comments. A new content parser has been cut into the core. New features include but not limited to.

• handling of malformed content streams
• operands are now collected as Integers and switched accordingly.
• PDF-13 has been solved and multiple threads can now access a document's pages.
• Vectors and HashTables have been replaced with Lists and Maps where appropriate.
• Name objects are now preserved between parse dictionary look ups.
• Shapes painting has now been rewritten to use a command pattern for paints operations which avoid timely "if" calls.
• Image loading has been moved off onto a worker thread that will allow the content parsing to continue.
• Optional image loading models for scaled and MipMapped images.
• configurable initial shapes size for CAD drawings.
• Removal of memory manager class, References class are now used resulting in lower memory overhead.
• creation of ImageStream object to handle image data, still some more work need here to speed up image loading, memory consumption and generic application of various masks.
• Updated the document class so that it will rebuild the cross reference table when doing a linear traversal of the document. This will insure that when page are resources are garbage collected the new crossreference entries can reget the respective objects from the file when needed.

New content parser has been checked in and is now part of the PRO release. The NContentParser and the OContentParser are created via a new ContentParserFactory which detects if the PRO jars is available. The parser logic has been abstracted with regards to postScript to Java2d but the lexers differ quite substantially.