ftp://ftp.jclark.com/pub/test/xmltest.zip
I changed one test case (088.xml) to reflect a change in the XML spec
and added some more tests. There are now 164 test cases which all fail
to be well formed according to the XML Proposed Tecommendation.
I've also updated my XML tokenizer/well-formedness checker at
ftp://ftp.jclark.com/pub/test/xmltok.zip
I believe this is now up to date for the XML Proposed Recommendation. I
know of one well-formedness violation it fails to detect: when the
encoding is UTF-8 it fails to detect illegal characters whose encoding
requires more than one byte (ie 0xFFFF, 0xFFFE, surrogates and
characters >= 0x10000). If you find any others, please let me know.
James