[antlr-interest] Strange ambiguous decision warning

[Jim Idle]

> -----Original Message-----
> From: antlr-interest-bounces at antlr.org [mailto:antlr-interest-
> bounces at antlr.org] On Behalf Of Martin Probst
> Sent: Wednesday, April 02, 2008 1:19 AM
> To: antlr-interest at antlr.org
> Subject: [antlr-interest] Strange ambiguous decision warning
> 
> Hi all,
> 
> I'm running into a seemingly strange ambiguity warning with ANTLR 3
> (latest beta).
> 
> I have a rule like this:
> sequenceType	:	(EMPTY_SEQUENCE '(' ')')
> 			| (itemType occurrenceIndicator) => (itemType
> occurrenceIndicator)
> 			| itemType ;
> occurrenceIndicator
> 		:	'?' | '*' | '+';
> itemType	:	kindTest | (ITEM '(' ')') | atomicType ;
> kindTest	:	...;
> atomicType	:	qNameOrIdent ;
> qNameOrIdent	:	QNAME | EMPTY_SEQUENCE | ...;
> 
> ANTLR complains about sequenceType:
> ANTLR Parser Generator  Version 3.1b1 (??)  1989-2007
> warning(200): /Users/martin/workspace/XQPretty/src/antlr/XQuery.g:
> 235:14: Decision can match input such as "EMPTY_SEQUENCE" using
> multiple alternatives: 1, 2, 3 As a result, alternative(s) 2,3 were
> disabled for that input


If you use ANTLRWorks, it will help you to see the following:

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.


Change your rule to:

sequenceType	: (EMPTY_SEQUENCE '(')=>(EMPTY_SEQUENCE '(' ')')  
			| itemType occurrenceIndicator?
			;

Then you will get ambiguities on '+', which means that you actually need:

sequenceType	: (EMPTY_SEQUENCE '(')=>(EMPTY_SEQUENCE '(' ')')  
			| itemType ((occurrenceIndicator)=>occurrenceIndicator)?
			;

If you want to eliminate all the warnings.

Doing this removes the huge predicate that you have right, replace it with a two token sequence, which will be much faster :-)

Though it is somewhat irrelevant to the issue, you will find your grammar more maintainable if you format it a little ;-)

Jim