6. Future Extensions

There are three extensions that were considered for Yapps 1, but not implemented until Yapps 2:

1. User defined actions inside clauses would allow parsing code to be mixed in with other code, which processes parts of a clause that have been parsed so far. For example, ID {{ if ID not in self.declarations: raise InvalidIdentifier }} ":=" expr would check, at parse time, whether the identifier has been declared, and if not, raise an exception. Only if the identifier had been declared would parsing continue to the “:=” and expr.
2. Parsing forms X* and X+ would be very convenient for matching commonly occuring list patterns in languages.
3. At present, tokens must be strings. It would be convenient to also allow some standard regular expressions to be imported, most likely from the tokenize module, which defines regular expressions for Python strings, identifiers, numbers, and several other common tokens.
4. It would be nice to somehow automatically deal with end-of-string tokens. A common error I make is to forget to define END to be "\\'", and then the resulting parser is unable to parse text that seems correct.

In addition, the possibility of backtracking and LL(k) parsing were being investigated, but neither made it into Yapps 2.