[lex.icon] p1 does not clearly cover the (single-digit) integer literal '0'

[morinmorin]

PR #3650 slimmed down [lex.icon] p1 to

In an \grammarterm{integer-literal}, the sequence of \grammarterm{binary-digit}s, \grammarterm{octal-digit}s, \grammarterm{digit}s, or \grammarterm{hexadecimal-digit}s is interpreted as a base $N$ integer

However, the zero in the octal integer literal '0' is not a (grammar term) octal-digits. So '0' falls outside of this wording.

For reference:

\begin{bnf}
\nontermdef{octal-literal}\br
    \terminal{0}\br
    octal-literal \opt{\terminal{'}} octal-digit
\end{bnf}

\begin{bnf}
\nontermdef{octal-digit} \textnormal{one of}\br
    \terminal{0  1  2  3  4  5  6  7}
\end{bnf}

[eisenwave]

I see what you mean. The 0 is simply treated as a prefix for the whole octal-literal, but is not any of the octal-digits.

While the issue was introduced by an editorial rewrite, what you're describing is not an editorial issue, but a normative wording defect. This could be submitted under https://github.com/cplusplus/CWG/issues, or at least CWG should look at the editorial fix.

Maybe

 In an integer-literal, the sequence of
 binary-digits, octal-digits, digits,
-or
 hexadecimal-digits
+, or the octal-literal 0
 is interpreted as a base N integer.

[morinmorin]

Thank you for the guidance and the suggested resolution. This is very useful! I'll resubmit to the CWG repository (after some discussion here).

I think digits (in "the sequence of binary-digits, octal-digits, digits, or hexadecimal-digits is interpreted as a base $N$ integer") has similar problems; the leading nonzero-digit is not included. A more accurate wording would be

the sequence of
- binary-digits,
- 0 and zero or more octal-digits,
- nonzero-digit and zero or more digits, or
- hexadecimal-digits
is interpreted as a base N integer

but this may be a bit verbose.