This is RXP 1.3, a validating namespace-aware XML parser in C.

RXP was written by Richard Tobin at the Language Technology Group,
Human Communication Research Centre, University of Edinburgh.

RXP is distributed under the GNU Public Licence, which is in the file
COPYING.  RXP may be made available under other licensing terms;
contact M.Moens@ed.ac.uk for details.

RXP is based on the W3C XML 1.0 (second edition) recommendation of 6th
October 2000 and the Namespaces recommendation of 14th January 1999,
along with errata to those specifications.  Deviations from these
recommendations should probably be considered as bugs.

There is currently only minimal documentation.  There is a manual page
for the rxp program, and a hastily-written file called "Manual" which
describes the public functions and data structures.  The file "Threads"
contains some notes about use in a threaded environment.

The system has been developed under unix (Solaris and FreeBSD), but
has also been compiled and run under MS Windows.  It should work with
minor changes on other systems.  To compile under unix, type "make",
or "make CHAR_SIZE=8" (not recommended if you care about standards
conformance).

A simple application (called rxp) is provided that parses and writes
XML data, optionally expanding entities, defaulting attributes, and
translating to a different output encoding.

There is an RXP web page at
  http://www.cogsci.ed.ac.uk/~richard/rxp.html
Bug reports should be sent to richard@cogsci.ed.ac.uk.



Changes since 1.2:

  Several bugs and memory leaks fixed (thanks to several testers).

  Some changes corresponding to XML 1.0 and Namespaces 1.0 errata
  (allow empty xml:lang, no NOTATIONs on EMPTY elements, enforce rules
  about reserved prefixes).

  Restructuring of input functions: there are now separate functions
  for different encodings instead of a switch in the inner loop, which
  produces a 10-20% speedup on some files.  And finally the bug with
  non-UTF characters on the same line as the XML declaration is fixed.
  
