accept unicode as bare identifier

[tbmreza]

The tests pass. What remains to be done for me is to figure out which error kinds to make.

Please have a look 🤓.

[tbmreza]

1

[Lucretiel]

Thanks you for the MR! Unfortunately I don't think it correctly implements the spec either. KDL identifiers are only concerned with code points, not grapheme clusters:

A bare Identifier is composed of any Unicode codepoint other than [decimal digits or non-identifier characters], followed by any number of Unicode codepoints other than non-identifier characters, [where a non-identifier character is]:

• Any codepoint with hexadecimal value 0x20 or below.
• Any codepoint with hexadecimal value higher than 0x10FFFF.
• Any of \/(){}<>;[]=,"

Additionally, KDL identifiers can't contain any whitespace, though that's not called out explicitly by the spec (https://github.com/kdl-org/kdl/issues/188)

[tbmreza]

whoops.

• S̶o̶ ̶I̶ ̶j̶u̶s̶t̶ ̶n̶e̶e̶d̶ ̶t̶o̶ ̶i̶t̶e̶r̶a̶t̶e̶ ̶8̶b̶i̶t̶s̶ ̶b̶y̶ ̶8̶b̶i̶t̶s̶ ̶(̶b̶e̶c̶a̶u̶s̶e̶ ̶u̶t̶f̶-̶8̶)̶.̶ ̶G̶o̶t̶ ̶i̶t̶.̶
• Yeah I mistook the unicode example in the readme. I thought if it's japanese whitespace it's part of the identifier.

I'll update my PR as soon as I can 🐱‍💻

[tbmreza]

@Lucretiel Ready for review.

[Lucretiel]

Thanks for the update, will review during the next stream (probably Monday)

[Lucretiel]

Sorry to say I totally lost track of this pull request and ended up implementing it myself. I used a similar design, but taking advantage of Chars::as_str to convert the chars iterator back into a string. Thanks you for the pull request, though!