CFDL

This grammar is intentionally pragmatic for v0.1: it captures the surface syntax needed by the Core spec. It is suitable as the basis for a hand-written parser or parser-generator input after minor adaptation (token rules, whitespace/comment handling).

1) Lexical tokens (informative)

Implementations MUST support:

  • Whitespace: spaces, tabs, newlines (ignored except as separators)
  • Line comments: // ... to end of line
  • Block comments: /* ... */

Recommended tokens:

  • IDENT → [A-Za-z_][A-Za-z0-9_]*
  • QNAME → IDENT ('.' IDENT)*
  • STRING → "..." with escapes
  • INT → digits with optional _ separators
  • DECIMAL → digits . digits with optional _
  • DATE → YYYY-MM or YYYY-MM-DD

2) Grammar

The formal EBNF grammar is maintained as a standalone file for use with grammar tooling (railroad diagram generators, parser generators, etc.):

📄 Download the grammar (EBNF)

3) Notes & implementation guidance (informative)

  1. Line breaks are not significant; blocks and keywords provide structure.
  2. money_lit is syntactic sugar; implementations should normalize to Money with currency.
  3. DATE accepts YYYY-MM and YYYY-MM-DD; normalize YYYY-MM to YYYY-MM-01 during parsing.
  4. schedule_opts are order-independent; parsers should accept any order.
  5. activate contract / deactivate contract are included in grammar, but may be a no-op in early engines.