-----------------------------------------------------------------------------
 $Id: CHANGES,v 1.5 1999/01/12 20:25:53 lindig Exp $
-----------------------------------------------------------------------------

* 0.8 -> 0.9

  The elment data type now has Econs and Enil cases to store sequences
  of elements and replaces (element list) in Eelement. The motovation
  behind this is to provide easy to use constructor functions to build
  XML documents from within programs.

  An additional Eempty case is also added. It is never used by the
  parser but intended for including break'able spaces in program
  generated XML documents. These documents tend to have too few breaks
  which makes pretty printing them difficult.


* 0.7.1 -> 0.8

  The internal representation of character data has changed (see the
  xml data type).  Data is now stored in so called chunks.  At pretty
  printing time chunks are split into words and whitespace.  In
  previous version was this done while parsing and thus the character
  data representation was much more fragmented. 

  The lexer combinators from the mylib library have changed. Saving
  results while scanning is now more flexbible because a function 
  is passed to the save fuction which can transform the scanned
  string. 
  
  The data types for prolog, xml declarations, and documents are made
  more explicit by using tagged sum types instead of tuples.

* 0.7 -> 0.7.1

  Input without an encoding declaration no longer has an encoding="-"
  in its output.

  Identifier part of declarations are no longer changed to lowercase.

  Pretty printed declarations now contain fewer spaces.
