Bytecode Format

Artifact contract

• Header-driven format (BC1)
• named function sections and a MAIN section
• explicit END terminators

Section decoding

• comments/blank lines are ignored
• each section has fixed semantic boundary
• malformed boundaries should fail early

Instruction normalization

• validate opcode family and arity
• validate numeric and string argument shapes
• unknown opcodes are hard errors

Strings

• quote-aware parse
• decode escapes, preserve raw representation on encode

Determinism

• encode output must be stable for same AST/bytecode
• useful for tests and caching

Practice focus

• header and section errors
• mixed opcode arg checks
• comment skipping and strict structural closing