Re: 10206 lexical ambiguity

On 19/01/17 01:57 PM, Bastiaan Veelo wrote:
>> On 19 Jan 2017, at 19:48, Paul Isaacs <[email protected]> wrote:
>>
>> Hello,
>>
>> Am I correct that, without semantic type tie-ins, that the production
>>
>> new-ordinal-type          = enumerated-type
>>                                       | subrange-type
>>                                       ;
>>
>> is ambiguous because (a) could be both an enumerated type and the expression for the lower bound of a subrange?
> Where do you see the ambiguity? enumerated-type is lexically different from subrange-type.
>
> Bastiaan.

Hello Bastiaan,

x = (a);         enumerated-type
y = (a)..5     subrange-type where (a) is a subrange-bound, the 
subrange-bound production = expression
                     and (a) is a lexically valid expression.

                     This would require a lookahead of 3 tokens to resolve.
                     I believe the case is unique because "( a," can not 
be an expression so must be an enumerated.
                    This still requires a lookahead of 2 tokens to resolve.

The 10206 spec does not contain either lookahead or lalr so maybe 
requiring more than 1 token of lookahead is allowed.

Regards,

Paul


_______________________________________________
Gpc mailing list
[email protected]
https://www.g-n-u.de/mailman/listinfo/gpc