Character �

["Fred L. Drake, Jr." <fdrake@acm.org>]

  Perhaps this is a poor question, but there's a bug filed against
Expat that I'd like to be able to deal with in the right way.  A user
managed to insert some U+0000 characters in their data as NUL bytes,
and has found that the MS parser interprets them as whitespace, and
Expat doesn't.  I'd like to determine what the right thing to do would
be from the perspective that strict conformance is the right thing.
  Looking at XML 1.0 (2e), section 2.2 ("Characters"), at the "Char"
production, I'm led to believe that U+0000 is not a legal XML
character.  The comment attached to that production makes things less
clear (at least to me), in that it seems to imply that all by the
surrogates, U+FFFE, and U+FFFF are legal.
  So, should U+0000 be interpreted as whitespace, non-whitespace, or
an error?  (References to something I've missed would be quite
welcome!)
  Thanks!


  -Fred

-- 
Fred L. Drake, Jr.  <fdrake at acm.org>
PythonLabs at Zope Corporation