Plain name IRI fragments can use unicode

[handrews]

Fixes #1314 as discussed in yesterday's community call. This just brings plain name fragment support into parity with the rest of our IRI support, and continues to rely on XML's NCName rule for the syntax. NCName always supported unicode, we just excluded all of it except for the US-ASCII characters.

I think I converted the W3C's grammar notation into ABNF properly, but if anyone is comfortable with ABNF (@awwright ?) a check would be appreciated. The W3C's grammar involves NCName, NCNameChar, and NCStartNameChar, which are just Name, NameChar, and StartNameChar minus the : character. I simplified the ABNF since we did not need the : variation at all.

[jdesrosiers]

I'd rather not include the NCName reproduction, but I don't feel strongly enough about it to ask for it to change. Consider it a suggestion and I'm ok with keeping others want to see it included.

[awwright]

@handrews Like this:

NCName          = NCNameStartChar *NCNameChar

NCNameStartChar = "_" / ALPHA
                / %xC0-D6 / %xD8-F6 / %xF8-2FF
                / %x370-37D / %x37F-1FFF
                / %x200C-200D / %x2070-218F
                / %x2C00-2FEF / %x3001-D7FF
                / %xF900-FDCF / %xFDF0-FFFD
                / %x10000-EFFFF

NCNameChar      = NCNameStartChar / "-" / "." / DIGIT
                / %xB7 / %x0300-036F / %x203F-2040

However if we're not going to use ABNF for making a normative requirement, then maybe it'd be better to just give the regular expression. Or both.

[gregsdennis]

Replaced by ☝️