[Clarification]: 7.2 numeric-like matchs

[aleguy02]

Question

The "numeric-like" regex pattern in section 7.2 seems too permissive or perhaps incorrect. For instance, the pattern is

/^-?\d+(?:.\d+)?(?:e[+-]?\d+)?$/i

However, this would mean the first non-capturing group (?:.\d+) uses a dot which captures all, and would capture something like a123. It seems like the intended behavior is to capture decimal portions, thus the non-capturing group should be (?:.\d+)
(note the escaped dot character).

Is this intended and I am missing something or misreading it?

Specification Section

Section 7.2 Quoting Rules for String Values

Source of Confusion

The "numeric-like" regex pattern in section 7.2 seems too permissive or perhaps incorrect. For instance, the pattern is  

/^-?\d+(?:.\d+)?(?:e[+-]?\d+)?$/i

However, this would mean the first non-capturing group (?:.\d+) uses a dot which captures all, and would capture something like a123. It seems like the intended behavior is to capture decimal portions, thus the non-capturing group *should* be (?:\.\d+)
(note the escaped dot character).

Example Case

The pattern in the spec would match 12a34, which is not a number.

Use Case / Context

No response

Reference Implementation Behavior

No response

Checklist

• I have read the relevant specification sections
• I have searched for similar questions in existing issues
• I have checked the examples in the specification

[johannschopplich]

This was generated by AI during triage.

You read the rendered page correctly – the source has the escaped dot ((?:\.\d+)), but the regex isn't wrapped in a code span, so GitHub's markdown rendering swallows the backslash and displays (?:.\d+). Good catch on the rendering. We'll wrap the inline regexes in code spans in the next editorial pass (and audit §7.3's siblings for the same problem). No grammar change needed, so closing.