Re: [Boost-users] [Spirit] parse tree node does not get produced with Single character input.

Subject: Re: [Boost-users] [Spirit] parse tree node does not get produced with Single character input.
From: OvermindDL1 (overminddl1_at_[hidden])
Date: 2010-06-29 03:22:20

I know this is old, but it has had no response, I apologize...

On Tue, Jun 8, 2010 at 9:03 AM, Philippe forest
<philippe.forest_at_[hidden]> wrote:
> Hi,

Greetings,

You should update your Boost, the later versions are vastly better,
allow me to demonstrate:

On Tue, Jun 8, 2010 at 9:03 AM, Philippe forest
<philippe.forest_at_[hidden]> wrote:
> First of all I would like to state that
> spirit =
> Â str_p("very") >> +ch_p(' ') >>( str_p("nice") | str_p("cool") |
> str_p("greate") ) >> +ch_p(' ') >> str_p("parser")
> ;

In modern Spirit:
  spirit
    =  skip(' ')[lit("very")
    >> (lit("nice")
       |    "cool"
       |    "great")
    )>>"parser"]
    ;

> So Thanks for that.
>
> But I do face a problem. When assigning ID to nodes of the parse tree.
> ex: rule<ScannerT, parser_context<>, parser_tag<stringID> > Â String;
>
> I've attached a sample to demonstrate the problem.
>
> In short:
>
> The grammar below does not produce a parse tree node with a 'stringID'
> when the input is a single character 'a'. As soon as I have 2 or more
> characters 'aa' then I get a node stringID.
>
> Â  Â  Â  Â  definition(
> Â  Â  Â  Â  Â ExpressionNoStringIDGrammar const& /*self*/)
> Â  Â  Â  Â {
> Â  Â  Â  Â  Â Char =
> Â  Â  Â  Â  Â  Â ch_p('a')
> Â  Â  Â  Â  Â  Â ;
>
> Â  Â  Â  Â  Â String =
> Â  Â  Â  Â  Â  Â ( +Char )
> Â  Â  Â  Â  Â  Â ;
>
> Â  Â  Â  Â  Â language =
> Â  Â  Â  Â  Â  Â String
> Â  Â  Â  Â  Â  Â ;
> Â  Â  Â  Â }
>
> rule<ScannerT, parser_context<>, parser_tag<charID> > Â  Â  Â  Â  Char;
> rule<ScannerT, parser_context<>, parser_tag<stringID> > Â  Â  Â  String;
> rule<ScannerT> language;

In modern Spirit, this should work fine:
template<typename Iterator>
class definition : grammar<Iterator, std::string()>
{
    definition() : definition::base_type(language)
    {
        Char =
            char_('a')
            ;

          String =
            +Char
            ;

          language =
            String
            ;
        }

 rule<Iterator, char()>         Char;
 rule<Iterator, std::string()>  String;
 rule<Iterator, std::string()>  language;
}

That will return the parsed string of a's as an std::string.  Look at
the new Spirit, insanely more powerful, and so much easier to use as
well.

http://www.boost.org/doc/libs/1_43_0/libs/spirit/doc/html/index.html

On Tue, Jun 8, 2010 at 9:03 AM, Philippe forest
<philippe.forest_at_[hidden]> wrote:
> BUT:
> If I remove the 'Char' parser and substitute it with ch_p('a') then it
> works fine , then I always get a node stringID even with a single input
> char.
>
> Â  Â  Â  Â  definition(
> Â  Â  Â  Â  Â ExpressionNoStringIDGrammar const& /*self*/)
> Â  Â  Â  Â {
> Â  Â  Â  Â  Â String =
> Â  Â  Â  Â  Â  Â ( +ch_p('a') )
> Â  Â  Â  Â  Â  Â ;
>
> Â  Â  Â  Â  Â language =
> Â  Â  Â  Â  Â  Â String
> Â  Â  Â  Â  Â  Â ;
> Â  Â  Â  Â }
>
> rule<ScannerT, parser_context<>, parser_tag<charID> > Â  Â  Â  Â  Char;
> rule<ScannerT, parser_context<>, parser_tag<stringID> > Â  Â  Â  String;
> rule<ScannerT> language;
>
> Could anyone either explain me what is going on ?
> Â - I'm I doing something wrong or is this a bug or works as designed ?