
	Big fixes made to aub since the release of v2.0:


				PATCH #1	(v2.0.1)

	-- Fixed aub to run under versions of perl earlier than 4.035
	   (but not earlier than 4.0.)
	-- Lifted silly requirement that news spool directory be writable
	-- Corrected typo in error reporting code that left part of an
	   error message blank.
	-- Corrected use of wrong variable in desperatation NNTP server
	   guessing code (this code rarely used in practice.)

				PATCH #2	(v2.0.2) 

	-- Fixed aub to run under versions of perl earlier than 4.0
	-- Fixed bug in NNTP protocol to keep aub from confusing the
	   end of an NNTP transaction with the line "." in a message.
	-- Made aub understand that the 'Minimum-OS2-Version:' header
	   seen in comp.binaries.os2 does not mean that it should expect
	   immediate uudecodable stuff to follow, even though the line
	   does begin with "M".
	-- Fixed bug that required postprocessor definitions and invocations
	   to name the postprocessor exactly the same way (no mixed case)
	-- Made erorrs go to stdout instead of standard error to integrate
	   them in the debug output stream.
	-- Fixed typo in 'aub -M' output -- description of how to escape the
	   comment character was botched.

				PATCH #3	(v2.0.3)

	-- Prevented aub from generating mode 0 binaries.  RFE: 2.1 will 
	   contain the position sensitive MODE keyword.
	-- Fixed problems in the last resort search for systems programs code
	-- No longer automatically try to find hook programs specified with
	   absolute path names.
	-- Made aub marginally smarter about finding libraries (I think)
	-- Added a tad more debugging output
	
				PATCH #4	(v2.0.4)

	-- Added additional verification that suspected uuencoded data really
	   is uuencoded.  This helps aub decode binaries it previously would
	   have choken on.
	-- Now understands the '|' separator (ie, "foo.Z part 2|3")

				PATCH #5	(v2.0.5)

	-- Code introduced in 2.0.4 was too restrictive.  Some decodable
	   binaries contain more than exactly 61 characters per line. 
	   Relaxed the error checking somewhat.
	-- Fixed bug that didn't break anything until 2.0.4, when the
	   restrictive code was added.
	-- Now mapping subject lines to lower case internally to improve
	   recognition abilities.
