"Values" text about integer literals disagree with grammar

[lrhn]

The text states:

We only support positive, decimal integer literals; negative integers are produced by the unary negation operator.

The grammar is:

INT_LIT ::= -? DIGIT+ | -? 0x HEXDIGIT+

which includes the minus sign in the grammar and accepts hexadecimal literals.

It makes a difference whether -1.abs() is parsed as -(1.abs()) or (-1).abs(), and . binds tighter than prefix -,
so one might want the - to be included in the grammar, if there is, or can be, functions on integers.
(It has other issues to do that, though, so I understand why the text wants to say that the sign is not part
of the literal, and is just the normal unary minus operator.)

The grammar also continues:

UINT_LIT ::= INT_LIT [uU]

which reused INT_LIT, which means -42u is an unsigned integer literal. That's probably not desirable.
(There is no specification of which value a literal has, because it's "obvious", but if -42u is an unsigned literal,
it's not obvious which unsigned value that has.)

Maybe this would be more practical:

INT_NUMERAL ::= DIGIT+ | 0x HEXDIGIT+
INT_LIT ::= -? INT_NUMERAL
UINT_LIT ::= INT_NUMERAL [uU]

In any case, the text should not say that there are only decimal literals, when there are hexadecimal ones,
and it should decide whether they have a leading - as part of the grammar or not.