| 1 | // Copyright (c) 2001-2011 Hartmut Kaiser |
| 2 | // |
| 3 | // Distributed under the Boost Software License, Version 1.0. (See accompanying |
| 4 | // file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt) |
| 5 | |
| 6 | #if !defined(BOOST_SPIRIT_LEXER_PARSE_NOV_17_2007_0246PM) |
| 7 | #define BOOST_SPIRIT_LEXER_PARSE_NOV_17_2007_0246PM |
| 8 | |
| 9 | #if defined(_MSC_VER) |
| 10 | #pragma once |
| 11 | #endif |
| 12 | |
| 13 | #include <boost/spirit/home/qi/skip_over.hpp> |
| 14 | #include <boost/spirit/home/qi/parse.hpp> |
| 15 | #include <boost/spirit/home/qi/nonterminal/grammar.hpp> |
| 16 | #include <boost/spirit/home/support/unused.hpp> |
| 17 | #include <boost/spirit/home/lex/lexer.hpp> |
| 18 | #include <boost/mpl/assert.hpp> |
| 19 | |
| 20 | namespace boost { namespace phoenix |
| 21 | { |
| 22 | template <typename Expr> |
| 23 | struct actor; |
| 24 | }} |
| 25 | |
| 26 | namespace boost { namespace spirit { namespace lex |
| 27 | { |
| 28 | /////////////////////////////////////////////////////////////////////////// |
| 29 | // Import skip_flag enumerator type from Qi namespace |
| 30 | using qi::skip_flag; |
| 31 | |
| 32 | /////////////////////////////////////////////////////////////////////////// |
| 33 | // |
| 34 | // The tokenize_and_parse() function is one of the main Spirit API |
| 35 | // functions. It simplifies using a lexer as the underlying token source |
| 36 | // while parsing a given input sequence. |
| 37 | // |
| 38 | // The function takes a pair of iterators spanning the underlying input |
| 39 | // stream to parse, the lexer object (built from the token definitions) |
| 40 | // and a parser object (built from the parser grammar definition). |
| 41 | // |
| 42 | // The second version of this function additionally takes an attribute to |
| 43 | // be used as the top level data structure instance the parser should use |
| 44 | // to store the recognized input to. |
| 45 | // |
| 46 | // The function returns true if the parsing succeeded (the given input |
| 47 | // sequence has been successfully matched by the given grammar). |
| 48 | // |
| 49 | // first, last: The pair of iterators spanning the underlying input |
| 50 | // sequence to parse. These iterators must at least |
| 51 | // conform to the requirements of the std::intput_iterator |
| 52 | // category. |
| 53 | // On exit the iterator 'first' will be updated to the |
| 54 | // position right after the last successfully matched |
| 55 | // token. |
| 56 | // lex: The lexer object (encoding the token definitions) to be |
| 57 | // used to convert the input sequence into a sequence of |
| 58 | // tokens. This token sequence is passed to the parsing |
| 59 | // process. The LexerExpr type must conform to the |
| 60 | // lexer interface described in the corresponding section |
| 61 | // of the documentation. |
| 62 | // xpr: The grammar object (encoding the parser grammar) to be |
| 63 | // used to match the token sequence generated by the lex |
| 64 | // object instance. The ParserExpr type must conform to |
| 65 | // the grammar interface described in the corresponding |
| 66 | // section of the documentation. |
| 67 | // attr: The top level attribute passed to the parser. It will |
| 68 | // be populated during the parsing of the input sequence. |
| 69 | // On exit it will hold the 'parser result' corresponding |
| 70 | // to the matched input sequence. |
| 71 | // |
| 72 | /////////////////////////////////////////////////////////////////////////// |
| 73 | template <typename Iterator, typename Lexer, typename ParserExpr> |
| 74 | inline bool |
| 75 | tokenize_and_parse(Iterator& first, Iterator last, Lexer const& lex, |
| 76 | ParserExpr const& xpr) |
| 77 | { |
| 78 | // Report invalid expression error as early as possible. |
| 79 | // If you got an error_invalid_expression error message here, |
| 80 | // then the expression (expr) is not a valid spirit qi expression. |
| 81 | BOOST_SPIRIT_ASSERT_MATCH(qi::domain, ParserExpr); |
| 82 | |
| 83 | typename Lexer::iterator_type iter = lex.begin(first, last); |
| 84 | return compile<qi::domain>(xpr).parse( |
| 85 | iter, lex.end(), unused, unused, unused); |
| 86 | } |
| 87 | |
| 88 | /////////////////////////////////////////////////////////////////////////// |
| 89 | template <typename Iterator, typename Lexer, typename ParserExpr |
| 90 | , typename Attribute> |
| 91 | inline bool |
| 92 | tokenize_and_parse(Iterator& first, Iterator last, Lexer const& lex |
| 93 | , ParserExpr const& xpr, Attribute& attr) |
| 94 | { |
| 95 | // Report invalid expression error as early as possible. |
| 96 | // If you got an error_invalid_expression error message here, |
| 97 | // then the expression (expr) is not a valid spirit qi expression. |
| 98 | BOOST_SPIRIT_ASSERT_MATCH(qi::domain, ParserExpr); |
| 99 | |
| 100 | typename Lexer::iterator_type iter = lex.begin(first, last); |
| 101 | return compile<qi::domain>(xpr).parse( |
| 102 | iter, lex.end(), unused, unused, attr); |
| 103 | } |
| 104 | |
| 105 | /////////////////////////////////////////////////////////////////////////// |
| 106 | // |
| 107 | // The tokenize_and_phrase_parse() function is one of the main Spirit API |
| 108 | // functions. It simplifies using a lexer as the underlying token source |
| 109 | // while phrase parsing a given input sequence. |
| 110 | // |
| 111 | // The function takes a pair of iterators spanning the underlying input |
| 112 | // stream to parse, the lexer object (built from the token definitions) |
| 113 | // and a parser object (built from the parser grammar definition). The |
| 114 | // additional skipper parameter will be used as the skip parser during |
| 115 | // the parsing process. |
| 116 | // |
| 117 | // The second version of this function additionally takes an attribute to |
| 118 | // be used as the top level data structure instance the parser should use |
| 119 | // to store the recognized input to. |
| 120 | // |
| 121 | // The function returns true if the parsing succeeded (the given input |
| 122 | // sequence has been successfully matched by the given grammar). |
| 123 | // |
| 124 | // first, last: The pair of iterators spanning the underlying input |
| 125 | // sequence to parse. These iterators must at least |
| 126 | // conform to the requirements of the std::intput_iterator |
| 127 | // category. |
| 128 | // On exit the iterator 'first' will be updated to the |
| 129 | // position right after the last successfully matched |
| 130 | // token. |
| 131 | // lex: The lexer object (encoding the token definitions) to be |
| 132 | // used to convert the input sequence into a sequence of |
| 133 | // tokens. This token sequence is passed to the parsing |
| 134 | // process. The LexerExpr type must conform to the |
| 135 | // lexer interface described in the corresponding section |
| 136 | // of the documentation. |
| 137 | // xpr: The grammar object (encoding the parser grammar) to be |
| 138 | // used to match the token sequence generated by the lex |
| 139 | // object instance. The ParserExpr type must conform to |
| 140 | // the grammar interface described in the corresponding |
| 141 | // section of the documentation. |
| 142 | // skipper: The skip parser to be used while parsing the given |
| 143 | // input sequence. Note, the skip parser will have to |
| 144 | // act on the same token sequence as the main parser |
| 145 | // 'xpr'. |
| 146 | // post_skip: The post_skip flag controls whether the function will |
| 147 | // invoke an additional post skip after the main parser |
| 148 | // returned. |
| 149 | // attr: The top level attribute passed to the parser. It will |
| 150 | // be populated during the parsing of the input sequence. |
| 151 | // On exit it will hold the 'parser result' corresponding |
| 152 | // to the matched input sequence. |
| 153 | // |
| 154 | /////////////////////////////////////////////////////////////////////////// |
| 155 | template <typename Iterator, typename Lexer, typename ParserExpr |
| 156 | , typename Skipper> |
| 157 | inline bool |
| 158 | tokenize_and_phrase_parse(Iterator& first, Iterator last |
| 159 | , Lexer const& lex, ParserExpr const& xpr, Skipper const& skipper |
| 160 | , BOOST_SCOPED_ENUM(skip_flag) post_skip = skip_flag::postskip) |
| 161 | { |
| 162 | // Report invalid expression error as early as possible. |
| 163 | // If you got an error_invalid_expression error message here, |
| 164 | // then the expression (expr) is not a valid spirit qi expression. |
| 165 | BOOST_SPIRIT_ASSERT_MATCH(qi::domain, ParserExpr); |
| 166 | BOOST_SPIRIT_ASSERT_MATCH(qi::domain, Skipper); |
| 167 | |
| 168 | typedef |
| 169 | typename spirit::result_of::compile<qi::domain, Skipper>::type |
| 170 | skipper_type; |
| 171 | skipper_type const skipper_ = compile<qi::domain>(skipper); |
| 172 | |
| 173 | typename Lexer::iterator_type iter = lex.begin(first, last); |
| 174 | typename Lexer::iterator_type end = lex.end(); |
| 175 | if (!compile<qi::domain>(xpr).parse( |
| 176 | iter, end, unused, skipper_, unused)) |
| 177 | return false; |
| 178 | |
| 179 | // do a final post-skip |
| 180 | if (post_skip == skip_flag::postskip) |
| 181 | qi::skip_over(iter, end, skipper_); |
| 182 | return true; |
| 183 | } |
| 184 | |
| 185 | template <typename Iterator, typename Lexer, typename ParserExpr |
| 186 | , typename Skipper, typename Attribute> |
| 187 | inline bool |
| 188 | tokenize_and_phrase_parse(Iterator& first, Iterator last |
| 189 | , Lexer const& lex, ParserExpr const& xpr, Skipper const& skipper |
| 190 | , BOOST_SCOPED_ENUM(skip_flag) post_skip, Attribute& attr) |
| 191 | { |
| 192 | // Report invalid expression error as early as possible. |
| 193 | // If you got an error_invalid_expression error message here, |
| 194 | // then the expression (expr) is not a valid spirit qi expression. |
| 195 | BOOST_SPIRIT_ASSERT_MATCH(qi::domain, ParserExpr); |
| 196 | BOOST_SPIRIT_ASSERT_MATCH(qi::domain, Skipper); |
| 197 | |
| 198 | typedef |
| 199 | typename spirit::result_of::compile<qi::domain, Skipper>::type |
| 200 | skipper_type; |
| 201 | skipper_type const skipper_ = compile<qi::domain>(skipper); |
| 202 | |
| 203 | typename Lexer::iterator_type iter = lex.begin(first, last); |
| 204 | typename Lexer::iterator_type end = lex.end(); |
| 205 | if (!compile<qi::domain>(xpr).parse( |
| 206 | iter, end, unused, skipper_, attr)) |
| 207 | return false; |
| 208 | |
| 209 | // do a final post-skip |
| 210 | if (post_skip == skip_flag::postskip) |
| 211 | qi::skip_over(iter, end, skipper_); |
| 212 | return true; |
| 213 | } |
| 214 | |
| 215 | /////////////////////////////////////////////////////////////////////////// |
| 216 | template <typename Iterator, typename Lexer, typename ParserExpr |
| 217 | , typename Skipper, typename Attribute> |
| 218 | inline bool |
| 219 | tokenize_and_phrase_parse(Iterator& first, Iterator last |
| 220 | , Lexer const& lex, ParserExpr const& xpr, Skipper const& skipper |
| 221 | , Attribute& attr) |
| 222 | { |
| 223 | return tokenize_and_phrase_parse(first, last, lex, xpr, skipper |
| 224 | , skip_flag::postskip, attr); |
| 225 | } |
| 226 | |
| 227 | /////////////////////////////////////////////////////////////////////////// |
| 228 | // |
| 229 | // The tokenize() function is one of the main Spirit API functions. It |
| 230 | // simplifies using a lexer to tokenize a given input sequence. It's main |
| 231 | // purpose is to use the lexer to tokenize all the input. |
| 232 | // |
| 233 | // The second version below discards all generated tokens afterwards. |
| 234 | // This is useful whenever all the needed functionality has been |
| 235 | // implemented directly inside the lexer semantic actions, which are being |
| 236 | // executed while the tokens are matched. |
| 237 | // |
| 238 | // The function takes a pair of iterators spanning the underlying input |
| 239 | // stream to scan, the lexer object (built from the token definitions), |
| 240 | // and a (optional) functor being called for each of the generated tokens. |
| 241 | // |
| 242 | // The function returns true if the scanning of the input succeeded (the |
| 243 | // given input sequence has been successfully matched by the given token |
| 244 | // definitions). |
| 245 | // |
| 246 | // first, last: The pair of iterators spanning the underlying input |
| 247 | // sequence to parse. These iterators must at least |
| 248 | // conform to the requirements of the std::intput_iterator |
| 249 | // category. |
| 250 | // On exit the iterator 'first' will be updated to the |
| 251 | // position right after the last successfully matched |
| 252 | // token. |
| 253 | // lex: The lexer object (encoding the token definitions) to be |
| 254 | // used to convert the input sequence into a sequence of |
| 255 | // tokens. The LexerExpr type must conform to the |
| 256 | // lexer interface described in the corresponding section |
| 257 | // of the documentation. |
| 258 | // f: A functor (callable object) taking a single argument of |
| 259 | // the token type and returning a bool, indicating whether |
| 260 | // the tokenization should be canceled. |
| 261 | // initial_state: The name of the state the lexer should start matching. |
| 262 | // The default value is zero, causing the lexer to start |
| 263 | // in its 'INITIAL' state. |
| 264 | // |
| 265 | /////////////////////////////////////////////////////////////////////////// |
| 266 | namespace detail |
| 267 | { |
| 268 | template <typename Token, typename F> |
| 269 | bool tokenize_callback(Token const& t, F f) |
| 270 | { |
| 271 | return f(t); |
| 272 | } |
| 273 | |
| 274 | template <typename Token, typename Eval> |
| 275 | bool tokenize_callback(Token const& t, phoenix::actor<Eval> const& f) |
| 276 | { |
| 277 | f(t); |
| 278 | return true; |
| 279 | } |
| 280 | |
| 281 | template <typename Token> |
| 282 | bool tokenize_callback(Token const& t, void (*f)(Token const&)) |
| 283 | { |
| 284 | f(t); |
| 285 | return true; |
| 286 | } |
| 287 | |
| 288 | template <typename Token> |
| 289 | bool tokenize_callback(Token const& t, bool (*f)(Token const&)) |
| 290 | { |
| 291 | return f(t); |
| 292 | } |
| 293 | } |
| 294 | |
| 295 | template <typename Iterator, typename Lexer, typename F> |
| 296 | inline bool |
| 297 | tokenize(Iterator& first, Iterator last, Lexer const& lex, F f |
| 298 | , typename Lexer::char_type const* initial_state = 0) |
| 299 | { |
| 300 | typedef typename Lexer::iterator_type iterator_type; |
| 301 | |
| 302 | iterator_type iter = lex.begin(first, last, initial_state); |
| 303 | iterator_type end = lex.end(); |
| 304 | for (/**/; iter != end && token_is_valid(*iter); ++iter) |
| 305 | { |
| 306 | if (!detail::tokenize_callback(*iter, f)) |
| 307 | return false; |
| 308 | } |
| 309 | return (iter == end) ? true : false; |
| 310 | } |
| 311 | |
| 312 | /////////////////////////////////////////////////////////////////////////// |
| 313 | template <typename Iterator, typename Lexer> |
| 314 | inline bool |
| 315 | tokenize(Iterator& first, Iterator last, Lexer const& lex |
| 316 | , typename Lexer::char_type const* initial_state = 0) |
| 317 | { |
| 318 | typedef typename Lexer::iterator_type iterator_type; |
| 319 | |
| 320 | iterator_type iter = lex.begin(first, last, initial_state); |
| 321 | iterator_type end = lex.end(); |
| 322 | |
| 323 | while (iter != end && token_is_valid(*iter)) |
| 324 | ++iter; |
| 325 | |
| 326 | return (iter == end) ? true : false; |
| 327 | } |
| 328 | |
| 329 | }}} |
| 330 | |
| 331 | #endif |
| 332 | |