Re: Domain name label length

On Sun, Dec 31, 2017 at 2:03 PM, Michael Powell <[email protected]> wrote:
> On Thu, Dec 28, 2017 at 12:27 AM, Joel de Guzman <[email protected]> wrote:
>> On 26/12/2017 11:38 AM, Michael Powell wrote:
>>>
>>> Hello,
>>>
>>> So, a domain name label looks something like this, in general, right?
>>>
>>> _label = _let >> *(_let | _dig | _hyp) >> (_let | _dig);
>>>
>>> However, another stipulation is that its length must be <=63 according
>>> to the RFC:
>>>
>>> https://tools.ietf.org/html/rfc1034#section-3.5
>>>
>>> So... Is there a way to constrain that using Qi?

I think I am closer in the ball park if I can manage a Phoenix
validation function, as I'd like to avoid the complexities of any form
of Bind, Phoenix or otherwise.

This at least on authority of this documentation:
https://theboostcpplibraries.com/boost.phoenix, tip recommending
avoiding Bind, and rather creating my own Phoenix functions.

However, still running into some difficulties sorting out the bits
connecting the pieces together:

namespace Validation {
    struct label_validation {
        bool operator()(std::string const& l) const {
            return l.length() <= 63;
        }
    };
}

// In my Parser struct:
boost::phoenix::function<Validation::label_validation> _validate_label;

// This would be perfectly sufficient and works with the bits that I
absolutely require:
_label = raw[_let >> -(*(_let | _dig | _hyp) >> (_let | _dig))][_pass
= _validate_label(_val)];

At the moment I am defining the following: BOOST_SPIRIT_USE_PHOENIX_V3
and BOOST_RESULT_OF_USE_DECLTPYE for both library as well as test
projects.

// ...
Severity Code Description Project File Line Suppression State
Error C2516 'boost::phoenix::evaluator::impl<const Expr
&,boost::phoenix::vector2<boost::mpl::true_,boost::phoenix::is_nullary>,boost::proto::envns_::empty_env>::result_type':
is not a legal base class Rfc3986_Uri_RegexSyntaxVerification
G:\Source\Boost.Installs\boost_1_65_1\boost\phoenix\core\is_nullary.hpp
124
Error C2039 'value': is not a member of
'boost::phoenix::result_of::is_nullary<Expr,void>'
Rfc3986_Uri_RegexSyntaxVerification
G:\Source\Boost.Installs\boost_1_65_1\boost\phoenix\core\actor.hpp 174
Error C2065 'value': undeclared identifier
Rfc3986_Uri_RegexSyntaxVerification
G:\Source\Boost.Installs\boost_1_65_1\boost\phoenix\core\actor.hpp 175
Error C2975 'C': invalid template argument for
'boost::mpl::eval_if_c', expected compile-time constant expression
Rfc3986_Uri_RegexSyntaxVerification
G:\Source\Boost.Installs\boost_1_65_1\boost\phoenix\core\actor.hpp 177

>> Yes, but if I were you, I'd probably avoid having too much smarts in
>> the parser and have those checks post-hoc.
>
> Yes, of course. The rule thus far is pretty straightforward, to do
> with length and length only. It would be inappropriate to embed
> anything that was so application specific.
>
> template<typename Iterator>
> struct MyParser : qi::grammar<Iterator, Ast::Address()> {
>
>     struct validate_label_length {
>         template<typename Context>
>         void operator()(std::string& l, Context&, bool& pass) {
>             pass = l.length() <= 63;
>         }
>     };
>
>     // ...
>
>     ... _label = raw[_let >> -(*(_let | _dig | _hyp) >> (_let |
> _dig))][validate_label_length()];
>     ... _address = _label % dot_;
>
>     ... _start = _address >> eoi;
> };
>
> But this is yielding compiler errors. I'm not sure I need to include
> something Fusion related?
>
> #include <boost/spirit/home/qi.hpp>
> #include "ast/domain_name.hpp"
>
> The structure is nested, defined in the scope of my template Parser.
>
> Severity Code Description Project File Line Suppression State
> Error C2893 Failed to specialize function template
> 'action_dispatch<Subject>::fwd_storage<unknown-type,boost::spirit::traits::action_dispatch<Subject>::fwd_none>::type
> boost::spirit::traits::action_dispatch<Subject>::do_call(F &&,...)'
> Rfc3986_Uri_RegexSyntaxVerification
> G:\Source\Boost.Installs\boost_1_65_1\boost\spirit\home\support\action_dispatch.hpp
> 142
> Error C2672 'boost::spirit::traits::action_dispatch<Subject>::do_call':
> no matching overloaded function found
> Rfc3986_Uri_RegexSyntaxVerification
> G:\Source\Boost.Installs\boost_1_65_1\boost\spirit\home\support\action_dispatch.hpp
> 142
> Error C2893 Failed to specialize function template
> 'action_dispatch<Subject>::fwd_storage<unknown-type,boost::spirit::traits::action_dispatch<Subject>::fwd_attrib_context_pass>::type
> boost::spirit::traits::action_dispatch<Subject>::do_call(F
> &&,boost::spirit::traits::action_dispatch<Subject>::fwd_tag<A>,boost::spirit::traits::action_dispatch<Subject>::fwd_tag<B>,boost::spirit::traits::action_dispatch<Subject>::fwd_tag<C>,...)'
> Rfc3986_Uri_RegexSyntaxVerification
> G:\Source\Boost.Installs\boost_1_65_1\boost\spirit\home\support\action_dispatch.hpp
> 142
> Error C2893 Failed to specialize function template
> 'action_dispatch<Subject>::fwd_storage<unknown-type,boost::spirit::traits::action_dispatch<Subject>::fwd_attrib_context>::type
> boost::spirit::traits::action_dispatch<Subject>::do_call(F
> &&,boost::spirit::traits::action_dispatch<Subject>::fwd_tag<A>,boost::spirit::traits::action_dispatch<Subject>::fwd_tag<B>,...)'
> Rfc3986_Uri_RegexSyntaxVerification
> G:\Source\Boost.Installs\boost_1_65_1\boost\spirit\home\support\action_dispatch.hpp
> 142
> Error C2893 Failed to specialize function template
> 'action_dispatch<Subject>::fwd_storage<unknown-type,boost::spirit::traits::action_dispatch<Subject>::fwd_attrib>::type
> boost::spirit::traits::action_dispatch<Subject>::do_call(F
> &&,boost::spirit::traits::action_dispatch<Subject>::fwd_tag<A>,...)'
> Rfc3986_Uri_RegexSyntaxVerification
> G:\Source\Boost.Installs\boost_1_65_1\boost\spirit\home\support\action_dispatch.hpp
> 142
>
>> Regards,
>> --
>> Joel de Guzman
>> http://www.ciere.com
>> http://boost-spirit.com
>> http://www.cycfi.com/
>>
>> ------------------------------------------------------------------------------
>> 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

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot