C++ regex parser: align with std::regex ECMAScript grammar and fix literal-location mapping

[Copilot]

This series ports the imported Ruby-oriented C++ regex parsing logic to the ECMAScript grammar used by std::regex, while preserving the required commit-by-commit migration shape. It also fixes incorrect term locations for prefixed/raw string literal forms and expands coverage for ECMAScript-specific edge cases.

• Location mapping correctness for C++ string literal spellings

  • Added regex-location coverage for plain, prefixed, raw, and custom-delimiter literals.
  • Replaced the hardcoded location offset with a content-start offset derived from StringLiteral.getValueText() (handles L/u/U/u8, raw introducer, and raw delimiters).

• Grammar narrowing: remove Ruby-only constructs

  • Removed support for Ruby-only anchors and escapes: \A \Z \z \G, \h \H.
  • Removed Ruby-only property/group/backref/comment constructs: \p{...} \P{...}, (?#...), (?'name'...), (?<name>...), \k<name>, and lookbehind (?<=...) / (?<!...).
  • Updated corpus and learned expectations to match the narrowed grammar.

• Grammar completion: add missing ECMAScript forms

  • Implemented control escapes \cX.
  • Implemented NUL escape \0 (only when not followed by a digit).
  • Implemented POSIX collating symbols [[.a.]] and equivalence classes [[=a=]] inside character classes, including delimiter accounting and tree-view exposure.

• Corpus cleanup and targeted coverage

  • Replaced non-ECMAScript \u{9879} corpus case with valid \u0061.
  • Added coverage for already-supported lazy quantifiers (*?, +?, ??, {2,3}?) and escapes \f, \v.
  • Added [\b] verification; parser treats it as an in-class escaped character (backspace semantics), not RegExpWordBoundary.

• Change note

  • Added unreleased minorAnalysis note stating the parser now targets std::regex ECMAScript grammar.

Example of the core location fix pattern:

valueText = re.getValueText() and
(
  valueText.regexpMatch("^\\s*(u8|L|u|U)?R\".*") and
  ... // prefix + delimiter aware raw literal offset
  or
  valueText.regexpMatch("^\\s*(u8|L|u|U)?\".*") and
  ... // prefixed normal literal offset
) and
startcolumn = re_start + contentStartOffset + start