Release 1.29.0

What's New

Bug Fixes

• Prettifier round-trips special-character identifiers — RIDDL
  identifiers may be single-quoted to carry special characters (spaces,
  /, @, etc.), e.g. user 'CI/CD Pipeline', and the parser accepts
  that form. The prettifier previously emitted such names unquoted, so
  parse → prettify → parse was not a round-trip — the regenerated RIDDL
  failed to re-parse. Identifier.format now single-quotes any name that
  is not a bare identifier, and the five prettifier sites that rendered
  names via the raw value (including user names — the reported case) now
  go through the quoting path. Bare names (including hyphenated names like
  my-entity) are emitted verbatim, so existing output is unchanged.

Features

• Whole-path quoting for path identifiers — A path reference whose
  component carries special characters now renders as a single quoted
  path — 'Domain.CI/CD Pipeline.Name' — rather than quoting each
  component. The parser gained a corresponding whole-path quoted form
  (tried after the normal dotted form, so every existing input parses
  unchanged), and the EBNF grammar and generated GBNF grammar were updated
  to match.

• New public API on Identifier — Identifier.isBareIdentifier(String)
  reports whether a value is a bare (unquoted) identifier, matching the
  parser's rule exactly ([A-Za-z][A-Za-z0-9_-]*, hyphens included), and
  Identifier.format(String) renders a value as valid RIDDL source
  (quoting when needed). Both are usable on JVM, JS, and Native.

Notes

• Error messages and diagnostics are unchanged: identify() and message
  templates continue to show names in their own single quotes without
  double-quoting (e.g. Entity 'CI/CD Pipeline'), using the raw value.
• Additive and backward compatible — no existing signatures changed.
• Consumers that previously sanitized identifiers before prettifying (to
  work around the missing quoting) can now rely on the prettifier to
  preserve original names.