[antlr-interest] Strange ambiguous decision warning
mail at martin-probst.com
Wed Apr 2 11:44:03 PDT 2008
> If you use ANTLRWorks, it will help you to see the following:
Well, ANTLRWorks comes with ANTLR 3.0.1, and that doesn't work at
> Alt1 starts with EMPTY_SEQUENCE
> Alt2 goes to itemType
> At3 goes to itemType
> itemType goes to atomicType
> atomicType goes to qNameOrIndent
> qNameOrIndent contains EMTPY_SEQUENCE
> hence all three alts can start with EMPTY_SEQUENCE. as
> occurenceIndicator is optional, then it is ambiguous.
Why? The parentheses following the literal EMPTY_SEQUENCE are not
optional, and the occurrence indicator after the second rule is not
optional either, so all three rules have a different second token -
k=2 should be sufficient to disambiguate.
> sequenceType : (EMPTY_SEQUENCE '(')=>(EMPTY_SEQUENCE '(' ')')
> | itemType ((occurrenceIndicator)=>occurrenceIndicator)?
Thanks, this worked. Though I'm still curious why this was ambiguous...
> Though it is somewhat irrelevant to the issue, you will find your
> grammar more maintainable if you format it a little ;-)
You don't really want to argue about indentation style, do you? :-)
More information about the antlr-interest