2002-03-27  Gareth Jones  <gareth.jones@stud.man.ac.uk>

		 * Name.pm (as_pdf): Use #XX for any characters outside !-~,
		 plus delimiters.  Nul is not a valid character in a name.
		 (name_from_string): PDF versions below 1.2 did not have # escapes.
		 Use the PDF version to determine how to interpret #.

		 * File.pm (readval): Rationalised whitespace in regexps so that
		 they will work for any valid PDF file.
		 (readval): Use name_to_string to localise hex-munging in Name.
		 (readval): Make sure that tokens end with an irregular character.
		 This is not needed for objects in the body of the PDF file but may
		 be needed in content streams.
		 (open): Set version.

		 * Dict.pm (outobjdeep): Use #XX for any characters outside !-~,
		 plus delimiters.  Nul is not a valid character in a name.

2002-03-19  Gareth Jones  <gareth.jones@stud.man.ac.uk>

		 * String.pm (convert): White space should be ignored in hex-strings.

		 * File.pm (update): Remove leading whitespace, otherwise readval
		 can loop for ever.
		 (readval): endstream does not need to be followed by a $cr.
		 (readval): \s does not include \0 which is a whitespace char in
		 PDF so we should use $ws_char rather than \s.
		 (readval): The `stream' keyword cannot be followed by a carrige
		 return alone.

		 * Null.pm: Having a specific null object avoids special cases when
		 reading files that have nulls in dictionaries because
		 $foo->{'bar'}->realise will now always work.

		 * Dict.pm (outobjdeep): This can create invalid dictionaries if
		 the keys aren't valid PDF names.

2002-03-18  Gareth Jones  <gareth.jones@stud.man.ac.uk>

		 * Pages.pm (new): Corrected documentation.

		 * File.pm: Corrected documentation.
		 (open): Test if seek fails.
		 (open): Don't start seek from negative number.
		 (readval): Ignore balanced parens when reading string.
		 (readval): Replace long regexps with sensible variables
		 (eg. $regular_char).  This means that all characters are regular
		 chars *except* those excluded.  The previous versions would have
		 rejected non-ASCII chars.
* 0.8

** Pages
    Adding a page updates the Pages tree to account for the added page

** pdfaddpg.pl
    Sort out page insertion bugs

* 0.7
Lots of changes in the area of fixing file updating (rather than creating)

* 0.5
Lots of changes and introduction of this document

** SFont
    Remove dependence upon Font::Metrics, introduce dependence on 
    Compress::Zlib, which is more likely to be available.
    Store width tables in class

** File
    Restructure saving a lot to allow incremental output with emptying