[ceph.git] / ceph / src / boost / libs / wave / include / boost / wave / cpplexer / re2clex / cpp_re2c_lexer.hpp

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

    Re2C based C++ lexer

    http://www.boost.org/

    Copyright (c) 2001-2012 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_RE2C_LEXER_HPP_B81A2629_D5B1_4944_A97D_60254182B9A8_INCLUDED)
#define CPP_RE2C_LEXER_HPP_B81A2629_D5B1_4944_A97D_60254182B9A8_INCLUDED

#include <string>
#include <cstdio>
#include <cstdarg>
#if defined(BOOST_SPIRIT_DEBUG)
#include <iostream>
#endif // defined(BOOST_SPIRIT_DEBUG)

#include <boost/concept_check.hpp>
#include <boost/assert.hpp>
#include <boost/spirit/include/classic_core.hpp>

#include <boost/wave/wave_config.hpp>
#include <boost/wave/language_support.hpp>
#include <boost/wave/token_ids.hpp>
#include <boost/wave/util/file_position.hpp>
#include <boost/wave/cpplexer/validate_universal_char.hpp>
#include <boost/wave/cpplexer/cpplexer_exceptions.hpp>
#include <boost/wave/cpplexer/token_cache.hpp>
#include <boost/wave/cpplexer/convert_trigraphs.hpp>

#include <boost/wave/cpplexer/cpp_lex_interface.hpp>
#include <boost/wave/cpplexer/re2clex/scanner.hpp>
#include <boost/wave/cpplexer/re2clex/cpp_re.hpp>
#if BOOST_WAVE_SUPPORT_PRAGMA_ONCE != 0
#include <boost/wave/cpplexer/detect_include_guards.hpp>
#endif

#include <boost/wave/cpplexer/cpp_lex_interface_generator.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 cpplexer {
namespace re2clex {

///////////////////////////////////////////////////////////////////////////////
//
//  encapsulation of the re2c based cpp lexer
//
///////////////////////////////////////////////////////////////////////////////

template <typename IteratorT,
    typename PositionT = boost::wave::util::file_position_type,
    typename TokenT = lex_token<PositionT> >
class lexer
{
public:
    typedef TokenT token_type;
    typedef typename token_type::string_type  string_type;

    lexer(IteratorT const &first, IteratorT const &last,
        PositionT const &pos, boost::wave::language_support language_);
    ~lexer();

    token_type& get(token_type&);
    void set_position(PositionT const &pos)
    {
        // set position has to change the file name and line number only
        filename = pos.get_file();
        scanner.line = pos.get_line();
//        scanner.column = scanner.curr_column = pos.get_column();
        scanner.file_name = filename.c_str();
    }
#if BOOST_WAVE_SUPPORT_PRAGMA_ONCE != 0
    bool has_include_guards(std::string& guard_name) const
    {
        return guards.detected(guard_name);
    }
#endif

// error reporting from the re2c generated lexer
    static int report_error(Scanner const* s, int code, char const *, ...);

private:
    static char const *tok_names[];

    Scanner scanner;
    string_type filename;
    string_type value;
    bool at_eof;
    boost::wave::language_support language;
#if BOOST_WAVE_SUPPORT_PRAGMA_ONCE != 0
    include_guards<token_type> guards;
#endif

#if BOOST_WAVE_SUPPORT_THREADING == 0
    static token_cache<string_type> const cache;
#else
    token_cache<string_type> const cache;
#endif
};

///////////////////////////////////////////////////////////////////////////////
// initialize cpp lexer
template <typename IteratorT, typename PositionT, typename TokenT>
inline
lexer<IteratorT, PositionT, TokenT>::lexer(IteratorT const &first,
        IteratorT const &last, PositionT const &pos,
        boost::wave::language_support language_)
  : filename(pos.get_file()), at_eof(false), language(language_)
#if BOOST_WAVE_SUPPORT_THREADING != 0
  , cache()
#endif
{
    using namespace std;        // some systems have memset in std
    memset(&scanner, '\0', sizeof(Scanner));
    scanner.eol_offsets = aq_create();
    if (first != last) {
        scanner.first = scanner.act = (uchar *)&(*first);
        scanner.last = scanner.first + std::distance(first, last);
    }
    scanner.line = pos.get_line();
    scanner.column = scanner.curr_column = pos.get_column();
    scanner.error_proc = report_error;
    scanner.file_name = filename.c_str();

#if BOOST_WAVE_SUPPORT_MS_EXTENSIONS != 0
    scanner.enable_ms_extensions = true;
#else
    scanner.enable_ms_extensions = false;
#endif

#if BOOST_WAVE_SUPPORT_VARIADICS_PLACEMARKERS != 0
    scanner.act_in_c99_mode = boost::wave::need_c99(language_);
#endif

#if BOOST_WAVE_SUPPORT_IMPORT_KEYWORD != 0
    scanner.enable_import_keyword = !boost::wave::need_c99(language_);
#else
    scanner.enable_import_keyword = false;
#endif

    scanner.detect_pp_numbers = boost::wave::need_prefer_pp_numbers(language_);
    scanner.single_line_only = boost::wave::need_single_line(language_);

#if BOOST_WAVE_SUPPORT_CPP0X != 0
    scanner.act_in_cpp0x_mode = boost::wave::need_cpp0x(language_);
#else
    scanner.act_in_cpp0x_mode = false;
#endif
}

template <typename IteratorT, typename PositionT, typename TokenT>
inline
lexer<IteratorT, PositionT, TokenT>::~lexer()
{
    using namespace std;        // some systems have free in std
    aq_terminate(scanner.eol_offsets);
    free(scanner.bot);
}

///////////////////////////////////////////////////////////////////////////////
//  get the next token from the input stream
template <typename IteratorT, typename PositionT, typename TokenT>
inline TokenT&
lexer<IteratorT, PositionT, TokenT>::get(TokenT& result)
{
    if (at_eof)
        return result = token_type();  // return T_EOI

    std::size_t actline = scanner.line;
    token_id id = token_id(scan(&scanner));

    switch (static_cast<unsigned int>(id)) {
    case T_IDENTIFIER:
    // test identifier characters for validity (throws if invalid chars found)
        value = string_type((char const *)scanner.tok,
            scanner.cur-scanner.tok);
        if (!boost::wave::need_no_character_validation(language))
            impl::validate_identifier_name(value, actline, scanner.column, filename);
        break;

    case T_STRINGLIT:
    case T_CHARLIT:
    case T_RAWSTRINGLIT:
    // test literal characters for validity (throws if invalid chars found)
        value = string_type((char const *)scanner.tok,
            scanner.cur-scanner.tok);
        if (boost::wave::need_convert_trigraphs(language))
            value = impl::convert_trigraphs(value);
        if (!boost::wave::need_no_character_validation(language))
            impl::validate_literal(value, actline, scanner.column, filename);
        break;

#if BOOST_WAVE_SUPPORT_INCLUDE_NEXT != 0
    case T_PP_HHEADER:
    case T_PP_QHEADER:
    case T_PP_INCLUDE:
    // convert to the corresponding ..._next token, if appropriate
      {
          value = string_type((char const *)scanner.tok,
              scanner.cur-scanner.tok);

      // Skip '#' and whitespace and see whether we find an 'include_next' here.
          typename string_type::size_type start = value.find("include");
          if (value.compare(start, 12, "include_next", 12) == 0)
              id = token_id(id | AltTokenType);
          break;
      }
#endif

    case T_LONGINTLIT:  // supported in C++11, C99 and long_long mode
        value = string_type((char const *)scanner.tok,
            scanner.cur-scanner.tok);
        if (!boost::wave::need_long_long(language)) {
        // syntax error: not allowed in C++ mode
            BOOST_WAVE_LEXER_THROW(lexing_exception, invalid_long_long_literal,
                value.c_str(), actline, scanner.column, filename.c_str());
        }
        break;

    case T_OCTALINT:
    case T_DECIMALINT:
    case T_HEXAINT:
    case T_INTLIT:
    case T_FLOATLIT:
    case T_FIXEDPOINTLIT:
    case T_CCOMMENT:
    case T_CPPCOMMENT:
    case T_SPACE:
    case T_SPACE2:
    case T_ANY:
    case T_PP_NUMBER:
        value = string_type((char const *)scanner.tok,
            scanner.cur-scanner.tok);
        break;

    case T_EOF:
    // T_EOF is returned as a valid token, the next call will return T_EOI,
    // i.e. the actual end of input
        at_eof = true;
        value.clear();
        break;

    case T_OR_TRIGRAPH:
    case T_XOR_TRIGRAPH:
    case T_LEFTBRACE_TRIGRAPH:
    case T_RIGHTBRACE_TRIGRAPH:
    case T_LEFTBRACKET_TRIGRAPH:
    case T_RIGHTBRACKET_TRIGRAPH:
    case T_COMPL_TRIGRAPH:
    case T_POUND_TRIGRAPH:
        if (boost::wave::need_convert_trigraphs(language)) {
            value = cache.get_token_value(BASEID_FROM_TOKEN(id));
        }
        else {
            value = string_type((char const *)scanner.tok,
                scanner.cur-scanner.tok);
        }
        break;

    case T_ANY_TRIGRAPH:
        if (boost::wave::need_convert_trigraphs(language)) {
            value = impl::convert_trigraph(
                string_type((char const *)scanner.tok));
        }
        else {
            value = string_type((char const *)scanner.tok,
                scanner.cur-scanner.tok);
        }
        break;

    default:
        if (CATEGORY_FROM_TOKEN(id) != EXTCATEGORY_FROM_TOKEN(id) ||
            IS_CATEGORY(id, UnknownTokenType))
        {
            value = string_type((char const *)scanner.tok,
                scanner.cur-scanner.tok);
        }
        else {
            value = cache.get_token_value(id);
        }
        break;
    }

//     std::cerr << boost::wave::get_token_name(id) << ": " << value << std::endl;

    // the re2c lexer reports the new line number for newline tokens
    result = token_type(id, value, PositionT(filename, actline, scanner.column));

#if BOOST_WAVE_SUPPORT_PRAGMA_ONCE != 0
    return guards.detect_guard(result);
#else
    return result;
#endif
}

template <typename IteratorT, typename PositionT, typename TokenT>
inline int
lexer<IteratorT, PositionT, TokenT>::report_error(Scanner const *s, int errcode,
    char const *msg, ...)
{
    BOOST_ASSERT(0 != s);
    BOOST_ASSERT(0 != msg);

    using namespace std;    // some system have vsprintf in namespace std

    char buffer[200];           // should be large enough
    va_list params;
    va_start(params, msg);
    vsprintf(buffer, msg, params);
    va_end(params);

    BOOST_WAVE_LEXER_THROW_VAR(lexing_exception, errcode, buffer, s->line,
        s->column, s->file_name);
//    BOOST_UNREACHABLE_RETURN(0);
    return 0;
}

///////////////////////////////////////////////////////////////////////////////
//
//  lex_functor
//
///////////////////////////////////////////////////////////////////////////////

template <typename IteratorT,
    typename PositionT = boost::wave::util::file_position_type,
    typename TokenT = typename lexer<IteratorT, PositionT>::token_type>
class lex_functor
:   public lex_input_interface_generator<TokenT>
{
public:
    typedef TokenT token_type;

    lex_functor(IteratorT const &first, IteratorT const &last,
            PositionT const &pos, boost::wave::language_support language)
    :   re2c_lexer(first, last, pos, language)
    {}
    virtual ~lex_functor() {}

// get the next token from the input stream
    token_type& get(token_type& result) { return re2c_lexer.get(result); }
    void set_position(PositionT const &pos) { re2c_lexer.set_position(pos); }
#if BOOST_WAVE_SUPPORT_PRAGMA_ONCE != 0
    bool has_include_guards(std::string& guard_name) const
        { return re2c_lexer.has_include_guards(guard_name); }
#endif

private:
    lexer<IteratorT, PositionT, TokenT> re2c_lexer;
};

#if BOOST_WAVE_SUPPORT_THREADING == 0
///////////////////////////////////////////////////////////////////////////////
template <typename IteratorT, typename PositionT, typename TokenT>
token_cache<typename lexer<IteratorT, PositionT, TokenT>::string_type> const
    lexer<IteratorT, PositionT, TokenT>::cache =
        token_cache<typename lexer<IteratorT, PositionT, TokenT>::string_type>();
#endif

}   // namespace re2clex

///////////////////////////////////////////////////////////////////////////////
//
//  The new_lexer_gen<>::new_lexer function (declared in cpp_lex_interface.hpp)
//  should be defined inline, if the lex_functor shouldn't be instantiated
//  separately from the lex_iterator.
//
//  Separate (explicit) instantiation helps to reduce compilation time.
//
///////////////////////////////////////////////////////////////////////////////

#if BOOST_WAVE_SEPARATE_LEXER_INSTANTIATION != 0
#define BOOST_WAVE_RE2C_NEW_LEXER_INLINE
#else
#define BOOST_WAVE_RE2C_NEW_LEXER_INLINE inline
#endif

///////////////////////////////////////////////////////////////////////////////
//
//  The 'new_lexer' function allows the opaque generation of a new lexer object.
//  It is coupled to the iterator type to allow to decouple the lexer/iterator
//  configurations at compile time.
//
//  This function is declared inside the cpp_lex_token.hpp file, which is
//  referenced by the source file calling the lexer and the source file, which
//  instantiates the lex_functor. But it is defined here, so it will be
//  instantiated only while compiling the source file, which instantiates the
//  lex_functor. While the cpp_re2c_token.hpp file may be included everywhere,
//  this file (cpp_re2c_lexer.hpp) should be included only once. This allows
//  to decouple the lexer interface from the lexer implementation and reduces
//  compilation time.
//
///////////////////////////////////////////////////////////////////////////////

template <typename IteratorT, typename PositionT, typename TokenT>
BOOST_WAVE_RE2C_NEW_LEXER_INLINE
lex_input_interface<TokenT> *
new_lexer_gen<IteratorT, PositionT, TokenT>::new_lexer(IteratorT const &first,
    IteratorT const &last, PositionT const &pos,
    boost::wave::language_support language)
{
    using re2clex::lex_functor;
    return new lex_functor<IteratorT, PositionT, TokenT>(first, last, pos, language);
}

#undef BOOST_WAVE_RE2C_NEW_LEXER_INLINE

///////////////////////////////////////////////////////////////////////////////
}   // namespace cpplexer
}   // namespace wave
}   // namespace boost

// the suffix header occurs after all of the code
#ifdef BOOST_HAS_ABI_HEADERS
#include BOOST_ABI_SUFFIX
#endif

#endif // !defined(CPP_RE2C_LEXER_HPP_B81A2629_D5B1_4944_A97D_60254182B9A8_INCLUDED)
Commit	Line	Data
7c673cae FG	1	/*=============================================================================
	2	Boost.Wave: A Standard compliant C++ preprocessor library
	3
	4	Re2C based C++ lexer
	5
	6	http://www.boost.org/
	7
	8	Copyright (c) 2001-2012 Hartmut Kaiser. Distributed under the Boost
	9	Software License, Version 1.0. (See accompanying file
	10	LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
	11	=============================================================================*/
	12
	13	#if !defined(CPP_RE2C_LEXER_HPP_B81A2629_D5B1_4944_A97D_60254182B9A8_INCLUDED)
	14	#define CPP_RE2C_LEXER_HPP_B81A2629_D5B1_4944_A97D_60254182B9A8_INCLUDED
	15
	16	#include <string>
	17	#include <cstdio>
	18	#include <cstdarg>
	19	#if defined(BOOST_SPIRIT_DEBUG)
	20	#include <iostream>
	21	#endif // defined(BOOST_SPIRIT_DEBUG)
	22
	23	#include <boost/concept_check.hpp>
	24	#include <boost/assert.hpp>
	25	#include <boost/spirit/include/classic_core.hpp>
	26
	27	#include <boost/wave/wave_config.hpp>
	28	#include <boost/wave/language_support.hpp>
	29	#include <boost/wave/token_ids.hpp>
	30	#include <boost/wave/util/file_position.hpp>
	31	#include <boost/wave/cpplexer/validate_universal_char.hpp>
	32	#include <boost/wave/cpplexer/cpplexer_exceptions.hpp>
	33	#include <boost/wave/cpplexer/token_cache.hpp>
	34	#include <boost/wave/cpplexer/convert_trigraphs.hpp>
	35
	36	#include <boost/wave/cpplexer/cpp_lex_interface.hpp>
	37	#include <boost/wave/cpplexer/re2clex/scanner.hpp>
	38	#include <boost/wave/cpplexer/re2clex/cpp_re.hpp>
	39	#if BOOST_WAVE_SUPPORT_PRAGMA_ONCE != 0
	40	#include <boost/wave/cpplexer/detect_include_guards.hpp>
	41	#endif
	42
	43	#include <boost/wave/cpplexer/cpp_lex_interface_generator.hpp>
	44
	45	// this must occur after all of the includes and before any code appears
	46	#ifdef BOOST_HAS_ABI_HEADERS
	47	#include BOOST_ABI_PREFIX
	48	#endif
	49
	50	///////////////////////////////////////////////////////////////////////////////
	51	namespace boost {
	52	namespace wave {
	53	namespace cpplexer {
	54	namespace re2clex {
	55
	56	///////////////////////////////////////////////////////////////////////////////
	57	//
	58	// encapsulation of the re2c based cpp lexer
	59	//
	60	///////////////////////////////////////////////////////////////////////////////
	61
	62	template <typename IteratorT,
	63	typename PositionT = boost::wave::util::file_position_type,
	64	typename TokenT = lex_token<PositionT> >
65	class lexer
66	{
67	public:
68	typedef TokenT token_type;
69	typedef typename token_type::string_type string_type;
70
71	lexer(IteratorT const &first, IteratorT const &last,
72	PositionT const &pos, boost::wave::language_support language_);
73	~lexer();
74
75	token_type& get(token_type&);
76	void set_position(PositionT const &pos)
77	{
78	// set position has to change the file name and line number only
79	filename = pos.get_file();
80	scanner.line = pos.get_line();
81	// scanner.column = scanner.curr_column = pos.get_column();
82	scanner.file_name = filename.c_str();
83	}
84	#if BOOST_WAVE_SUPPORT_PRAGMA_ONCE != 0
85	bool has_include_guards(std::string& guard_name) const
86	{
87	return guards.detected(guard_name);
88	}
89	#endif
90
91	// error reporting from the re2c generated lexer
92	static int report_error(Scanner const* s, int code, char const *, ...);
93
94	private:
95	static char const *tok_names[];
96
97	Scanner scanner;
98	string_type filename;
99	string_type value;
100	bool at_eof;
101	boost::wave::language_support language;
102	#if BOOST_WAVE_SUPPORT_PRAGMA_ONCE != 0
103	include_guards<token_type> guards;
104	#endif
105
106	#if BOOST_WAVE_SUPPORT_THREADING == 0
107	static token_cache<string_type> const cache;
108	#else
109	token_cache<string_type> const cache;
110	#endif
111	};
112
113	///////////////////////////////////////////////////////////////////////////////
114	// initialize cpp lexer
115	template <typename IteratorT, typename PositionT, typename TokenT>
116	inline
117	lexer<IteratorT, PositionT, TokenT>::lexer(IteratorT const &first,
118	IteratorT const &last, PositionT const &pos,
119	boost::wave::language_support language_)
120	: filename(pos.get_file()), at_eof(false), language(language_)
121	#if BOOST_WAVE_SUPPORT_THREADING != 0
122	, cache()
123	#endif
124	{
125	using namespace std; // some systems have memset in std
126	memset(&scanner, '\0', sizeof(Scanner));
127	scanner.eol_offsets = aq_create();
128	if (first != last) {
129	scanner.first = scanner.act = (uchar )&(first);
130	scanner.last = scanner.first + std::distance(first, last);
131	}
132	scanner.line = pos.get_line();
133	scanner.column = scanner.curr_column = pos.get_column();
134	scanner.error_proc = report_error;
135	scanner.file_name = filename.c_str();
136
137	#if BOOST_WAVE_SUPPORT_MS_EXTENSIONS != 0
138	scanner.enable_ms_extensions = true;
139	#else
140	scanner.enable_ms_extensions = false;
141	#endif
142
143	#if BOOST_WAVE_SUPPORT_VARIADICS_PLACEMARKERS != 0
144	scanner.act_in_c99_mode = boost::wave::need_c99(language_);
145	#endif
146
147	#if BOOST_WAVE_SUPPORT_IMPORT_KEYWORD != 0
148	scanner.enable_import_keyword = !boost::wave::need_c99(language_);
149	#else
150	scanner.enable_import_keyword = false;
151	#endif
152
153	scanner.detect_pp_numbers = boost::wave::need_prefer_pp_numbers(language_);
154	scanner.single_line_only = boost::wave::need_single_line(language_);
155
156	#if BOOST_WAVE_SUPPORT_CPP0X != 0
157	scanner.act_in_cpp0x_mode = boost::wave::need_cpp0x(language_);
158	#else
159	scanner.act_in_cpp0x_mode = false;
160	#endif
161	}
162
163	template <typename IteratorT, typename PositionT, typename TokenT>
164	inline
165	lexer<IteratorT, PositionT, TokenT>::~lexer()
166	{
167	using namespace std; // some systems have free in std
168	aq_terminate(scanner.eol_offsets);
169	free(scanner.bot);
170	}
171
172	///////////////////////////////////////////////////////////////////////////////
173	// get the next token from the input stream
174	template <typename IteratorT, typename PositionT, typename TokenT>
175	inline TokenT&
176	lexer<IteratorT, PositionT, TokenT>::get(TokenT& result)
177	{
178	if (at_eof)
179	return result = token_type(); // return T_EOI
180
181	std::size_t actline = scanner.line;
182	token_id id = token_id(scan(&scanner));
183
184	switch (static_cast<unsigned int>(id)) {
185	case T_IDENTIFIER:
186	// test identifier characters for validity (throws if invalid chars found)
187	value = string_type((char const *)scanner.tok,
188	scanner.cur-scanner.tok);
189	if (!boost::wave::need_no_character_validation(language))
190	impl::validate_identifier_name(value, actline, scanner.column, filename);
191	break;
192
193	case T_STRINGLIT:
194	case T_CHARLIT:
195	case T_RAWSTRINGLIT:
196	// test literal characters for validity (throws if invalid chars found)
197	value = string_type((char const *)scanner.tok,
198	scanner.cur-scanner.tok);
199	if (boost::wave::need_convert_trigraphs(language))
200	value = impl::convert_trigraphs(value);
201	if (!boost::wave::need_no_character_validation(language))
202	impl::validate_literal(value, actline, scanner.column, filename);
203	break;
204
205	#if BOOST_WAVE_SUPPORT_INCLUDE_NEXT != 0
206	case T_PP_HHEADER:
207	case T_PP_QHEADER:
208	case T_PP_INCLUDE:
209	// convert to the corresponding ..._next token, if appropriate
210	{
211	value = string_type((char const *)scanner.tok,
212	scanner.cur-scanner.tok);
213
214	// Skip '#' and whitespace and see whether we find an 'include_next' here.
215	typename string_type::size_type start = value.find("include");
216	if (value.compare(start, 12, "include_next", 12) == 0)
217	id = token_id(id \| AltTokenType);
218	break;
219	}
220	#endif
221
222	case T_LONGINTLIT: // supported in C++11, C99 and long_long mode
223	value = string_type((char const *)scanner.tok,
224	scanner.cur-scanner.tok);
225	if (!boost::wave::need_long_long(language)) {
226	// syntax error: not allowed in C++ mode
227	BOOST_WAVE_LEXER_THROW(lexing_exception, invalid_long_long_literal,
228	value.c_str(), actline, scanner.column, filename.c_str());
229	}
230	break;
231
232	case T_OCTALINT:
233	case T_DECIMALINT:
234	case T_HEXAINT:
235	case T_INTLIT:
236	case T_FLOATLIT:
237	case T_FIXEDPOINTLIT:
238	case T_CCOMMENT:
239	case T_CPPCOMMENT:
240	case T_SPACE:
241	case T_SPACE2:
242	case T_ANY:
243	case T_PP_NUMBER:
244	value = string_type((char const *)scanner.tok,
245	scanner.cur-scanner.tok);
246	break;
247
248	case T_EOF:
249	// T_EOF is returned as a valid token, the next call will return T_EOI,
250	// i.e. the actual end of input
251	at_eof = true;
252	value.clear();
253	break;
254
255	case T_OR_TRIGRAPH:
256	case T_XOR_TRIGRAPH:
257	case T_LEFTBRACE_TRIGRAPH:
258	case T_RIGHTBRACE_TRIGRAPH:
259	case T_LEFTBRACKET_TRIGRAPH:
260	case T_RIGHTBRACKET_TRIGRAPH:
261	case T_COMPL_TRIGRAPH:
262	case T_POUND_TRIGRAPH:
263	if (boost::wave::need_convert_trigraphs(language)) {
264	value = cache.get_token_value(BASEID_FROM_TOKEN(id));
265	}
266	else {
267	value = string_type((char const *)scanner.tok,
268	scanner.cur-scanner.tok);
269	}
270	break;
271
272	case T_ANY_TRIGRAPH:
273	if (boost::wave::need_convert_trigraphs(language)) {
274	value = impl::convert_trigraph(
275	string_type((char const *)scanner.tok));
276	}
277	else {
278	value = string_type((char const *)scanner.tok,
279	scanner.cur-scanner.tok);
280	}
281	break;
282
283	default:
284	if (CATEGORY_FROM_TOKEN(id) != EXTCATEGORY_FROM_TOKEN(id) \|\|
285	IS_CATEGORY(id, UnknownTokenType))
286	{
287	value = string_type((char const *)scanner.tok,
288	scanner.cur-scanner.tok);
289	}
290	else {
291	value = cache.get_token_value(id);
292	}
293	break;
294	}
295
296	// std::cerr << boost::wave::get_token_name(id) << ": " << value << std::endl;
297
298	// the re2c lexer reports the new line number for newline tokens
299	result = token_type(id, value, PositionT(filename, actline, scanner.column));
300
301	#if BOOST_WAVE_SUPPORT_PRAGMA_ONCE != 0
302	return guards.detect_guard(result);
303	#else
304	return result;
305	#endif
306	}
307
308	template <typename IteratorT, typename PositionT, typename TokenT>
309	inline int
310	lexer<IteratorT, PositionT, TokenT>::report_error(Scanner const *s, int errcode,
311	char const *msg, ...)
312	{
313	BOOST_ASSERT(0 != s);
314	BOOST_ASSERT(0 != msg);
315
316	using namespace std; // some system have vsprintf in namespace std
317
318	char buffer[200]; // should be large enough
319	va_list params;
320	va_start(params, msg);
321	vsprintf(buffer, msg, params);
322	va_end(params);
323
324	BOOST_WAVE_LEXER_THROW_VAR(lexing_exception, errcode, buffer, s->line,
325	s->column, s->file_name);
326	// BOOST_UNREACHABLE_RETURN(0);
327	return 0;
328	}
329
330	///////////////////////////////////////////////////////////////////////////////
331	//
332	// lex_functor
333	//
334	///////////////////////////////////////////////////////////////////////////////
335
336	template <typename IteratorT,
337	typename PositionT = boost::wave::util::file_position_type,
338	typename TokenT = typename lexer<IteratorT, PositionT>::token_type>
339	class lex_functor
340	: public lex_input_interface_generator<TokenT>
341	{
342	public:
343	typedef TokenT token_type;
344
345	lex_functor(IteratorT const &first, IteratorT const &last,
346	PositionT const &pos, boost::wave::language_support language)
347	: re2c_lexer(first, last, pos, language)
348	{}
349	virtual ~lex_functor() {}
350
351	// get the next token from the input stream
352	token_type& get(token_type& result) { return re2c_lexer.get(result); }
353	void set_position(PositionT const &pos) { re2c_lexer.set_position(pos); }
354	#if BOOST_WAVE_SUPPORT_PRAGMA_ONCE != 0
355	bool has_include_guards(std::string& guard_name) const
356	{ return re2c_lexer.has_include_guards(guard_name); }
357	#endif
358
359	private:
360	lexer<IteratorT, PositionT, TokenT> re2c_lexer;
361	};
362
363	#if BOOST_WAVE_SUPPORT_THREADING == 0
364	///////////////////////////////////////////////////////////////////////////////
365	template <typename IteratorT, typename PositionT, typename TokenT>
366	token_cache<typename lexer<IteratorT, PositionT, TokenT>::string_type> const
367	lexer<IteratorT, PositionT, TokenT>::cache =
368	token_cache<typename lexer<IteratorT, PositionT, TokenT>::string_type>();
369	#endif
370
371	} // namespace re2clex
372
373	///////////////////////////////////////////////////////////////////////////////
374	//
375	// The new_lexer_gen<>::new_lexer function (declared in cpp_lex_interface.hpp)
376	// should be defined inline, if the lex_functor shouldn't be instantiated
377	// separately from the lex_iterator.
378	//
379	// Separate (explicit) instantiation helps to reduce compilation time.
380	//
381	///////////////////////////////////////////////////////////////////////////////
382
383	#if BOOST_WAVE_SEPARATE_LEXER_INSTANTIATION != 0
384	#define BOOST_WAVE_RE2C_NEW_LEXER_INLINE
385	#else
386	#define BOOST_WAVE_RE2C_NEW_LEXER_INLINE inline
387	#endif
388
389	///////////////////////////////////////////////////////////////////////////////
390	//
391	// The 'new_lexer' function allows the opaque generation of a new lexer object.
392	// It is coupled to the iterator type to allow to decouple the lexer/iterator
393	// configurations at compile time.
394	//
395	// This function is declared inside the cpp_lex_token.hpp file, which is
396	// referenced by the source file calling the lexer and the source file, which
397	// instantiates the lex_functor. But it is defined here, so it will be
398	// instantiated only while compiling the source file, which instantiates the
399	// lex_functor. While the cpp_re2c_token.hpp file may be included everywhere,
400	// this file (cpp_re2c_lexer.hpp) should be included only once. This allows
401	// to decouple the lexer interface from the lexer implementation and reduces
402	// compilation time.
403	//
404	///////////////////////////////////////////////////////////////////////////////
405
406	template <typename IteratorT, typename PositionT, typename TokenT>
407	BOOST_WAVE_RE2C_NEW_LEXER_INLINE
408	lex_input_interface<TokenT> *
409	new_lexer_gen<IteratorT, PositionT, TokenT>::new_lexer(IteratorT const &first,
410	IteratorT const &last, PositionT const &pos,
411	boost::wave::language_support language)
412	{
413	using re2clex::lex_functor;
414	return new lex_functor<IteratorT, PositionT, TokenT>(first, last, pos, language);
415	}
416
417	#undef BOOST_WAVE_RE2C_NEW_LEXER_INLINE
418
419	///////////////////////////////////////////////////////////////////////////////
420	} // namespace cpplexer
421	} // namespace wave
422	} // namespace boost
423
424	// the suffix header occurs after all of the code
425	#ifdef BOOST_HAS_ABI_HEADERS
426	#include BOOST_ABI_SUFFIX
427	#endif
428
429	#endif // !defined(CPP_RE2C_LEXER_HPP_B81A2629_D5B1_4944_A97D_60254182B9A8_INCLUDED)