cpp_grammar.hpp

Boost provides free peer-reviewed portable C++ source libraries. We emphasize libraries that work

/*=============================================================================
    Boost.Wave: A Standard compliant C++ preprocessor library

    http://www.boost.org/

    Copyright (c) 2001-2008 Hartmut Kaiser. Distributed under the Boost
    Software License, Version 1.0. (See accompanying file
    LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
=============================================================================*/

#if !defined(CPP_GRAMMAR_HPP_FEAEBC2E_2734_428B_A7CA_85E5A415E23E_INCLUDED)
#define CPP_GRAMMAR_HPP_FEAEBC2E_2734_428B_A7CA_85E5A415E23E_INCLUDED

#include <boost/spirit/include/classic_core.hpp>
#include <boost/spirit/include/classic_parse_tree.hpp>
#include <boost/spirit/include/classic_parse_tree_utils.hpp>
#include <boost/spirit/include/classic_confix.hpp>
#include <boost/spirit/include/classic_lists.hpp>

#include <boost/wave/wave_config.hpp>
#include <boost/pool/pool_alloc.hpp>

#if BOOST_WAVE_DUMP_PARSE_TREE != 0
#include <map>
#include <boost/spirit/include/classic_tree_to_xml.hpp>
#endif

#include <boost/wave/token_ids.hpp>
#include <boost/wave/grammars/cpp_grammar_gen.hpp>
#include <boost/wave/util/pattern_parser.hpp>

#include <boost/wave/cpp_exceptions.hpp>

// this must occur after all of the includes and before any code appears
#ifdef BOOST_HAS_ABI_HEADERS
#include BOOST_ABI_PREFIX
#endif

///////////////////////////////////////////////////////////////////////////////
namespace boost {
namespace wave { 
namespace grammars {

namespace impl {

///////////////////////////////////////////////////////////////////////////////
//
//  store_found_eof
//
//      The store_found_eof functor sets a given flag if the T_EOF token was 
//      found during the parsing process
//
///////////////////////////////////////////////////////////////////////////////

    struct store_found_eof {

        store_found_eof(bool &found_eof_) : found_eof(found_eof_) {}
        
        template <typename TokenT>
        void operator()(TokenT const &/*token*/) const
        {
            found_eof = true;
        }
        
        bool &found_eof;
    };

///////////////////////////////////////////////////////////////////////////////
//
//  store_found_directive
//
//      The store_found_directive functor stores the token_id of the recognized
//      pp directive
//
///////////////////////////////////////////////////////////////////////////////

    template <typename TokenT>
    struct store_found_directive {

        store_found_directive(TokenT &found_directive_) 
        :   found_directive(found_directive_) {}
        
        void operator()(TokenT const &token) const
        {
            found_directive = token;
        }
        
        TokenT &found_directive;
    };

///////////////////////////////////////////////////////////////////////////////
//
//  store_found_eoltokens
//
//      The store_found_eoltokens functor stores the token sequence of the 
//      line ending for a particular pp directive
//
///////////////////////////////////////////////////////////////////////////////

    template <typename ContainerT>
    struct store_found_eoltokens {

        store_found_eoltokens(ContainerT &found_eoltokens_) 
        :   found_eoltokens(found_eoltokens_) {}
        
        template <typename IteratorT>
        void operator()(IteratorT const &first, IteratorT const& last) const
        {
            std::copy(first, last, 
                std::inserter(found_eoltokens, found_eoltokens.end()));
        }
        
        ContainerT &found_eoltokens;
    };

///////////////////////////////////////////////////////////////////////////////
//
//  flush_underlying_parser
//
//      The flush_underlying_parser flushes the underlying
//      multi_pass_iterator during the normal parsing process. This is
//      used at certain points during the parsing process, when it is
//      clear, that no backtracking is needed anymore and the input
//      gathered so far may be discarded.
//
///////////////////////////////////////////////////////////////////////////////
    struct flush_underlying_parser
    :   public boost::spirit::classic::parser<flush_underlying_parser>
    {
        typedef flush_underlying_parser this_t;

        template <typename ScannerT>
        typename boost::spirit::classic::parser_result<this_t, ScannerT>::type
        parse(ScannerT const& scan) const
        {
            scan.first.clear_queue();
            return scan.empty_match();  
        }
    };

    flush_underlying_parser const 
        flush_underlying_parser_p = flush_underlying_parser();

}   // anonymous namespace

///////////////////////////////////////////////////////////////////////////////
//  define, whether the rule's should generate some debug output
#define TRACE_CPP_GRAMMAR \
    bool(BOOST_SPIRIT_DEBUG_FLAGS_CPP & BOOST_SPIRIT_DEBUG_FLAGS_CPP_GRAMMAR) \
    /**/

///////////////////////////////////////////////////////////////////////////////
// Encapsulation of the C++ preprocessor grammar.
template <typename TokenT, typename ContainerT>
struct cpp_grammar : 
    public boost::spirit::classic::grammar<cpp_grammar<TokenT, ContainerT> >
{
    typedef typename TokenT::position_type  position_type;
    typedef cpp_grammar<TokenT, ContainerT> grammar_type;
    typedef impl::store_found_eof           store_found_eof_type;
    typedef impl::store_found_directive<TokenT>     store_found_directive_type;
    typedef impl::store_found_eoltokens<ContainerT> store_found_eoltokens_type;
    
    template <typename ScannerT>
    struct definition
    {
    // non-parse_tree generating rule type
        typedef typename ScannerT::iteration_policy_t iteration_policy_t;
        typedef boost::spirit::classic::match_policy match_policy_t;
        typedef typename ScannerT::action_policy_t action_policy_t;
        typedef 
            boost::spirit::classic::scanner_policies<
                iteration_policy_t, match_policy_t, action_policy_t> 
            policies_t;
        typedef 
            boost::spirit::classic::scanner<typename ScannerT::iterator_t, policies_t> 
            non_tree_scanner_t;
        typedef 
            boost::spirit::classic::rule<
                non_tree_scanner_t, boost::spirit::classic::dynamic_parser_tag> 
            no_tree_rule_type;

    // 'normal' (parse_tree generating) rule type
        typedef 
            boost::spirit::classic::rule<
                ScannerT, boost::spirit::classic::dynamic_parser_tag> 
            rule_type;

        rule_type pp_statement, macro_include_file;
//         rule_type include_file, system_include_file;
        rule_type plain_define, macro_definition, macro_parameters;
        rule_type undefine;
        rule_type ppifdef, ppifndef, ppif, ppelif;
//         rule_type ppelse, ppendif;
        rule_type ppline; 
        rule_type pperror;
        rule_type ppwarning;
        rule_type pppragma;
        rule_type illformed;
        rule_type ppqualifiedname;
        rule_type eol_tokens;
        no_tree_rule_type ppsp;
#if BOOST_WAVE_SUPPORT_MS_EXTENSIONS != 0
        rule_type ppregion;
        rule_type ppendregion;
#endif

        definition(cpp_grammar const &self) 
        {
        // import the spirit and cpplexer namespaces here
            using namespace boost::spirit::classic;
            using namespace boost::wave;
            using namespace boost::wave::util;

        // set the rule id's for later use
            pp_statement.set_id(BOOST_WAVE_PP_STATEMENT_ID);
//             include_file.set_id(BOOST_WAVE_INCLUDE_FILE_ID);
//             system_include_file.set_id(BOOST_WAVE_SYSINCLUDE_FILE_ID);
            macro_include_file.set_id(BOOST_WAVE_MACROINCLUDE_FILE_ID);
            plain_define.set_id(BOOST_WAVE_PLAIN_DEFINE_ID);
            macro_parameters.set_id(BOOST_WAVE_MACRO_PARAMETERS_ID);
            macro_definition.set_id(BOOST_WAVE_MACRO_DEFINITION_ID);
            undefine.set_id(BOOST_WAVE_UNDEFINE_ID);
            ppifdef.set_id(BOOST_WAVE_IFDEF_ID);
            ppifndef.set_id(BOOST_WAVE_IFNDEF_ID);
            ppif.set_id(BOOST_WAVE_IF_ID);
            ppelif.set_id(BOOST_WAVE_ELIF_ID);
//             ppelse.set_id(BOOST_WAVE_ELSE_ID);
//             ppendif.set_id(BOOST_WAVE_ENDIF_ID);
            ppline.set_id(BOOST_WAVE_LINE_ID);
            pperror.set_id(BOOST_WAVE_ERROR_ID);
            ppwarning.set_id(BOOST_WAVE_WARNING_ID);
            pppragma.set_id(BOOST_WAVE_PRAGMA_ID);
            illformed.set_id(BOOST_WAVE_ILLFORMED_ID);
            ppsp.set_id(BOOST_WAVE_PPSPACE_ID);
            ppqualifiedname.set_id(BOOST_WAVE_PPQUALIFIEDNAME_ID);
#if BOOST_WAVE_SUPPORT_MS_EXTENSIONS != 0
            ppregion.set_id(BOOST_WAVE_REGION_ID);
            ppendregion.set_id(BOOST_WAVE_ENDREGION_ID);
#endif

#if BOOST_WAVE_DUMP_PARSE_TREE != 0
            self.map_rule_id_to_name.init_rule_id_to_name_map(self);
#endif 

        // recognizes preprocessor directives only

        // C++ standard 16.1: A preprocessing directive consists of a sequence 
        // of preprocessing tokens. The first token in the sequence is # 
        // preprocessing token that is either the first character in the source 
        // file (optionally after white space containing no new-line 
        // characters) or that follows white space containing at least one 
        // new-line character. The last token in the sequence is the first 
        // new-line character that follows the first token in the sequence.

            pp_statement
                =   (   plain_define
//                     |   include_file
//                     |   system_include_file
                    |   ppif
                    |   ppelif
                    |   ppifndef
                    |   ppifdef
                    |   undefine
//                     |   ppelse
                    |   macro_include_file
                    |   ppline
                    |   pppragma
                    |   pperror
                    |   ppwarning
//                     |   ppendif
#if BOOST_WAVE_SUPPORT_MS_EXTENSIONS != 0
                    |   ppregion
                    |   ppendregion
#endif
                    |   illformed
                    )
                    >> eol_tokens
                       [ store_found_eoltokens_type(self.found_eoltokens) ]
//  In parser debug mode it is useful not to flush the underlying stream
//  to allow its investigation in the debugger and to see the correct
//  output in the printed debug log..
//  Note: this may break the parser, though.
#if !(defined(BOOST_SPIRIT_DEBUG) && \
      (BOOST_SPIRIT_DEBUG_FLAGS_CPP & BOOST_SPIRIT_DEBUG_FLAGS_CPP_GRAMMAR) \
     )
                   >>  impl::flush_underlying_parser_p
#endif // !(defined(BOOST_SPIRIT_DEBUG) &&
                ;

//         // #include ...
//             include_file            // include "..."
//                 =   ch_p(T_PP_QHEADER) 
//                     [ store_found_directive_type(self.found_directive) ]
// #if BOOST_WAVE_SUPPORT_INCLUDE_NEXT != 0
//                 |   ch_p(T_PP_QHEADER_NEXT)
//                     [ store_found_directive_type(self.found_directive) ]
// #endif 
//                 ;

//             system_include_file     // include <...>
//                 =   ch_p(T_PP_HHEADER) 
//                     [ store_found_directive_type(self.found_directive) ]
// #if BOOST_WAVE_SUPPORT_INCLUDE_NEXT != 0
//                 |   ch_p(T_PP_HHEADER_NEXT)
//                     [ store_found_directive_type(self.found_directive) ]
// #endif 
//                 ;

            macro_include_file      // include ...anything else...
                =   no_node_d
                    [
                        ch_p(T_PP_INCLUDE)
                        [ store_found_directive_type(self.found_directive) ]
#if BOOST_WAVE_SUPPORT_INCLUDE_NEXT != 0
                    |   ch_p(T_PP_INCLUDE_NEXT)
                        [ store_found_directive_type(self.found_directive) ]
#endif
                    ]
                    >> *(   anychar_p -
                            (ch_p(T_NEWLINE) | ch_p(T_CPPCOMMENT) | ch_p(T_EOF)) 
                        )
                ;

        // #define FOO foo (with optional parameters)
            plain_define
                =   no_node_d
                    [
                        ch_p(T_PP_DEFINE) 
                        [ store_found_directive_type(self.found_directive) ]
                        >> +ppsp
                    ]
                    >>  (   ch_p(T_IDENTIFIER) 
                        |   pattern_p(KeywordTokenType, TokenTypeMask)
                        |   pattern_p(OperatorTokenType|AltExtTokenType, 
                                ExtTokenTypeMask)   // and, bit_and etc.
                        |   pattern_p(BoolLiteralTokenType, TokenTypeMask)  // true/false
                        )
                    >>  (   (   no_node_d[eps_p(ch_p(T_LEFTPAREN))]
                                >>  macro_parameters
                                >> !macro_definition
                            )
                        |  !(   no_node_d[+ppsp] 
                                >>  macro_definition
                            )
                        )
                ;

        // parameter list
        // normal C++ mode
            macro_parameters
                =   confix_p(
                        no_node_d[ch_p(T_LEFTPAREN) >> *ppsp],
                       !list_p(
                            (   ch_p(T_IDENTIFIER) 
                            |   pattern_p(KeywordTokenType, TokenTypeMask)
                            |   pattern_p(OperatorTokenType|AltExtTokenType, 
                                    ExtTokenTypeMask)   // and, bit_and etc.
                            |   pattern_p(BoolLiteralTokenType, TokenTypeMask)  // true/false
#if BOOST_WAVE_SUPPORT_VARIADICS_PLACEMARKERS != 0
                            |   ch_p(T_ELLIPSIS)
#endif
                            ), 
                            no_node_d[*ppsp >> ch_p(T_COMMA) >> *ppsp]
                        ),
                        no_node_d[*ppsp >> ch_p(T_RIGHTPAREN)]
                    )
                ;
            
        // macro body (anything left until eol)
            macro_definition
                =   no_node_d[*ppsp]
                    >> *(   anychar_p -
                            (ch_p(T_NEWLINE) | ch_p(T_CPPCOMMENT) | ch_p(T_EOF)) 
                        )
                ;

        // #undef FOO 
            undefine