Scannerless Parsing

• Traditional linear-time parsers limited by fixed (e.g., one-token) lookahead

  • If we only have one lookahead token, then it's easier if tokens are big.

• Packrat parsers provide unlimited lookahead

  • No longer need to separate lexical analysis

• Why scannerless parsing?

  • Simplicity: unified grammar for entire language
  • Power: lexical elements with complex syntax