MiniZinc specification of string syntax problems

[PerMildner]

String Literals and String Interpolation says: "\x[0-9a-fA-F][0-9a-fA-F]? for an arbitrary 8-bit character specified by the given hexadecimal number" (and similar for octal).

This is incorrect and contradicts the "you can use the hexadecimal or octal notation to insert the UTF-8 byte sequences" later on the page. That is, the \x.. syntax species bytes making up the UTF-8 encoding of the (Unicode) character code, the \x.. construct does not specify a "8-bit character".

Also, the BNF grammar in the same section:

<string-contents> ::= ([^"\n\] | \[0-7][0-7?][0-7]? | \x[0-9a-fA-F][0-9a-fA-F]? | \n | \t | \" | \\)*

The [0-7?] should presumably be [0-7]?.

Nit: Also, in the same grammar definition: [^"\n\] is unclear (or at least abusing notation). The first backslash is presumably part of \n whereas the second backslash means a literal backslash. The Notation says "A newline character or CRLF sequence is written \n." which is fine in EBNF but is less suitable in a regexp character set [...] notation (e.g. can a lone CR appear as itself in a string-contents?).

On the other hand, in the same grammar definition, ... | \n | ... presumably does not mean a "newline character or CRLF sequence" but rather the two character sequence \ n.

(Yes, I am of the opinion that grammars should be boringly formal. Not everyone agrees.)