Re: Attribute ignored in first rule of alternatives

I got multiple answers directly, instead of going through this list. The 
problem was indeed the semantic action inhibiting attribute propagation. 
By setting the third template argument to true, attribute propagation is 
forced:

const auto bound = x3::rule<struct _, Bound, true>("bound")
     = (x3::lit("[") >> x3::lit("<=") > uint_ > x3::lit("]"))
     | (x3::lit("[") >> x3::lit("<") > uint_[bound_dec1] > x3::lit("]"))
     ;

- Søren

On 2017-12-08 03:43, Joel de Guzman wrote:
> On 06/12/2017 11:41 PM, Søren Enevoldsen wrote:
>> Dear Spirit Community,
>>
>> I have the following rule:
>>
>> auto const bound = rule<Bound>("bound")
>>      = (x3::lit("[") >> x3::lit("<=") > uint_ > x3::lit("]"))
>>      | (x3::lit("[") >> x3::lit("<") > uint_[bound_dec1] > x3::lit("]"))
>>      ;
>>
>> Bound is a struct with a uint32_t member. I have a BOOST_ADAPT_STRUCT 
>> to ensure Spirit X3 can write to it.
>>
>> When I parse "[<5]" I get a Bound with the value 4 (bound_dec1] 
>> decrements the parsed number by one and stores that. But if I input 
>> "[<=5]", I would expect a bound with 5 after all the magic. But I 
>> always get a 0. If I add a "no-op" function to the rule, such that 
>> the number parsing becomes uint_[bound_dec0] , it parses correctly. 
>> But I don't really think that should be necessary. How do I Spirit to 
>> not zero out my attribute?
>
> Do you have a minimal cpp file we can try? It seems you are inhibiting 
> attribute
> propagation by having a semantic action, but I'm not sure.
>
> Regards,



------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Spirit-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/spirit-general