Newline handling

[lilyball]

Newlines right now are defined as ('\r' '\n') | '\n'. I think we should follow the Unicode spec and treat a bare CR as a line separator too. Unicode also recommends that for “a readline function” it should stop at CR, LF, CRLF, NEL (U+85), LS (U+2028), PS (U+2029), and even FF (U+0C). This is documented in §5.8 Newline Guidelines.

Treating a bare CR as a line separator will simplify parsing. For example, right now kdl-rs will incorrectly parse ”//hello\nworld\r\n as it reads until the \r\n instead of stopping at the \n. But I think it will also just make more sense to follow the Unicode standard and treat all of these characters as line terminators.

[zkat]

oh hey! This is really interesting! I think this is definitely a change I'd want to make, especially since I'm working on making the language more unicode-friendly (#11)

[lilyball]

Fixed in fa60bc6 and #32.