Treatment of ^ in regexes

Appendix F says: "All XML characters are valid character ranges, except as
follows:...
The ^ character is only valid at the beginning of a ·positive character
group· if it is part of a ·negative character group·; ...".  However, the
EBNF doesn't seem consistent with this. Consider

  [^X]

This is ambiguous wrt the EBNF, since "^" is an XmlCharIncDash and thus a
charRange: according to the EBNF it could be a powCharGroup containing "^"
and "X" or a negCharGroup containing "X".  Consider also

  [^]

According to the EBNF, this is unambiguously a posCharGroup containing "^",
but this is inconsistent with the prose.

There are similar problems with "-", but I understand the WG is already
aware of these.  I would suggest you need to resolve the issues with "-" and
"^" consistently.

James