> Well, in some grammars, the base case of a left recursion may not
> consume any input. From my C grammar for Rats!
>
> variant generic DirectAbstractDeclarator =
> <FixedArray> DirectAbstractDeclarator "[":Symbol
> AssignmentExpression? void:"]":Symbol
> / <VariableArray> DirectAbstractDeclarator "[":Symbol
> VariableLength void:"]":Symbol
> / <Function> DirectAbstractDeclarator "(":Symbol
> ParameterTypeList? void:")":Symbol
> / <Parenthesized> void:"(":Symbol AttributedAbstractDeclarator
> void:")":Symbol
> / <Empty> null
> ;
>
> The last alternative does not touch the input and just returns a null
> value, which is the desired behavior for direct abstract declarators.
> So, Bryan, your condition is too restrictive. Any suggestions?
>
> Robert
>
The loop in the example code I gave is a do-while loop, so the first
iteration through doesn't invoke the length test. Thus I believe a
null match will work, it is just that any future left recursive
matches for the same rule at the same point must match with additional
length. Although looking at the code again, I notice that I forgot
the return after the loop, so perhaps that is the source of the
confusion, so with Bryan's change and the fixed return it now look
like:
// Checks for memorized version, and if not found, does the actual parse
// Returns the length of the match, or -1 if no match.
int check_parse(rule, position)
key = tuple(rule, postion);
if find(memory[key])
return memory[key];
else
cur = -1;
do
memory[key] = cur;
prev = cur;
cur = do_pase(rule, position);
while(cur != -1 && cur > prev)
return cur;
end
end
_______________________________________________
PEG mailing list
PEG@lists.csail.mit.edu
https://lists.csail.mit.edu/mailman/listinfo/peg
