* [May/02/09]

        Documentation moved to docbook/xml format.
        Webpage updated.

 * [Apr/22/09] 

        Apply patches from folks at Ksplice. 
        Thank You Tim Abbott and Anders Kaseorg!

        - Add prototypes for ud_set_user_opaque_data/ud_get_user_opaque_data.
        - Remove autogenerated files from version control.
        - autogen.sh wraps autoreconf.
        - opgen.py takes path to optable.
        - Use packing in instruction tables.
        - Add support for setting the vendor to UD_VENDOR_ANY.
        - Support for build as part of linux kernel.
        - Support for 3 byte opcodes.
        - Fix/Add instructions.

 * [May/03/08] Fix: Makefile for HP-UX build.
	           Checked build on HP-UX, Solaris, FreeBSD, MacOS, Linux.
	           Potentially broken win32 build (need help.)
 * [Apr/29/08] Fix: operand size cast for rcl, ror, etc. (Thanks to Hans)
	           Fix: support for implicit operands and segment 
	                overrides (Thanks to Hans)
 * [Mar/27/08] Add: new XML based optable generator.
	           Add: new build system based on libtool.
 * [Dec/01/07] Fix: documentation buf: prefix fields
               Fix: bufoverrun.c to include the right header.
 * [Aug/05/07] Patch for DJGPP (Thanks to Robert Riebisch)
               Removed "test" Target in tot Makefile (Thanks to Robert Riebisch)
               New target:tests.
               Misc. clean up top-of-tree Makefile.
               Use config.h for package version
               Review/cleanup FP Opcode map. 
 * [Dec/18/06] Fixed: Printf warnings.
               Fixed: set_input_buffered to set_input_buffer in documentation.
               Added support for INTEL VMX instructions.
               API addition: ud_set_vendor(): UD_VENDOR_{INTEL, AMD}.
               Fixed: Handling of invalid combination of bytes. No more 
               input re-syncing.
               Added Makefiles for win32 build. (Thanks to Xi Hang).
               Added Makefile.standalone for "Standalone Udis86".
               Fixed: u->inp_end initialization bug.
               Fixed: swapgs only in 64bit mode.
 * [Oct/27/06] 0xF1 is now int1.
               Fixed: cast prefix for Group1_op80/82/C0/D0 instructions.
               Fixed: lmsw
               Added support for Udis86 Standalone.
 * [Sep/18/06] Fixed: Missing support for o32 for in/out instructions.
 * [Jul/26/06] Fixed: Missing REX.W in Group1_op83 instructions.
 * [Jul/16/06] Fixed: rex.b on [rip+disp]
 * [Jul/15/06] Fixed: Wrongly printing "loopnz" as "loopn" (Thanks to Sanjay Patel)
               Fixed: Wrongly decoding "out Ib, AL" (Thanks to Sanjay Patel)
 * [Jun/8/06]  Missing instructions added. (As per AMD64 Manual Vol3 Rev 3.11 
               Dec 2005)
               Fixed minor decoder issues. (Thanks to Sanjay Patel)
               Fixed issues with input stream buffer. (Thanks to Sanjay Patel)
 * [May/18/06] Fixed issues with handling of 64-bit values. 
               Fixed bug in syntax generation for immediate operands in branch
               instructions.
               Fixed decoding of (G, W) operands.
 * [May/12/06] Added support for data types in MS VC++.
 * [May/11/06] Fixed portability issues concerning inp_uintXX functions in 
               input.c and 64-bit casting in syn-intel/syn-att.c (Thanks to 
               Sanjay Patel)