| 1 | /*============================================================================= |
|---|
| 2 | Copyright (c) 2002-2003 Hartmut Kaiser |
|---|
| 3 | http://spirit.sourceforge.net/ |
|---|
| 4 | |
|---|
| 5 | Use, modification and distribution is subject to the Boost Software |
|---|
| 6 | License, Version 1.0. (See accompanying file LICENSE_1_0.txt or copy at |
|---|
| 7 | http://www.boost.org/LICENSE_1_0.txt) |
|---|
| 8 | =============================================================================*/ |
|---|
| 9 | #ifndef BOOST_SPIRIT_CONFIX_HPP |
|---|
| 10 | #define BOOST_SPIRIT_CONFIX_HPP |
|---|
| 11 | |
|---|
| 12 | /////////////////////////////////////////////////////////////////////////////// |
|---|
| 13 | #include <boost/config.hpp> |
|---|
| 14 | #include <boost/spirit/meta/as_parser.hpp> |
|---|
| 15 | #include <boost/spirit/core/composite/operators.hpp> |
|---|
| 16 | #include <boost/spirit/utility/impl/confix.ipp> |
|---|
| 17 | |
|---|
| 18 | /////////////////////////////////////////////////////////////////////////////// |
|---|
| 19 | namespace boost { namespace spirit { |
|---|
| 20 | |
|---|
| 21 | /////////////////////////////////////////////////////////////////////////////// |
|---|
| 22 | // |
|---|
| 23 | // confix_parser class |
|---|
| 24 | // |
|---|
| 25 | // Parses a sequence of 3 sub-matches. This class may |
|---|
| 26 | // be used to parse structures, where the opening part is possibly |
|---|
| 27 | // contained in the expression part and the whole sequence is only |
|---|
| 28 | // parsed after seeing the closing part matching the first opening |
|---|
| 29 | // subsequence. Example: C-comments: |
|---|
| 30 | // |
|---|
| 31 | // /* This is a C-comment */ |
|---|
| 32 | // |
|---|
| 33 | /////////////////////////////////////////////////////////////////////////////// |
|---|
| 34 | |
|---|
| 35 | template<typename NestedT = non_nested, typename LexemeT = non_lexeme> |
|---|
| 36 | struct confix_parser_gen; |
|---|
| 37 | |
|---|
| 38 | template < |
|---|
| 39 | typename OpenT, typename ExprT, typename CloseT, typename CategoryT, |
|---|
| 40 | typename NestedT = non_nested, typename LexemeT = non_lexeme |
|---|
| 41 | > |
|---|
| 42 | struct confix_parser : |
|---|
| 43 | public parser< |
|---|
| 44 | confix_parser<OpenT, ExprT, CloseT, CategoryT, NestedT, LexemeT> |
|---|
| 45 | > |
|---|
| 46 | { |
|---|
| 47 | typedef |
|---|
| 48 | confix_parser<OpenT, ExprT, CloseT, CategoryT, NestedT, LexemeT> |
|---|
| 49 | self_t; |
|---|
| 50 | |
|---|
| 51 | confix_parser(OpenT const &open_, ExprT const &expr_, CloseT const &close_) |
|---|
| 52 | : open(open_), expr(expr_), close(close_) |
|---|
| 53 | {} |
|---|
| 54 | |
|---|
| 55 | template <typename ScannerT> |
|---|
| 56 | typename parser_result<self_t, ScannerT>::type |
|---|
| 57 | parse(ScannerT const& scan) const |
|---|
| 58 | { |
|---|
| 59 | return impl::confix_parser_type<CategoryT>:: |
|---|
| 60 | parse(NestedT(), LexemeT(), *this, scan, open, expr, close); |
|---|
| 61 | } |
|---|
| 62 | |
|---|
| 63 | private: |
|---|
| 64 | |
|---|
| 65 | typename as_parser<OpenT>::type::embed_t open; |
|---|
| 66 | typename as_parser<ExprT>::type::embed_t expr; |
|---|
| 67 | typename as_parser<CloseT>::type::embed_t close; |
|---|
| 68 | }; |
|---|
| 69 | |
|---|
| 70 | /////////////////////////////////////////////////////////////////////////////// |
|---|
| 71 | // |
|---|
| 72 | // Confix parser generator template |
|---|
| 73 | // |
|---|
| 74 | // This is a helper for generating a correct confix_parser<> from |
|---|
| 75 | // auxiliary parameters. There are the following types supported as |
|---|
| 76 | // parameters yet: parsers, single characters and strings (see |
|---|
| 77 | // as_parser). |
|---|
| 78 | // |
|---|
| 79 | // If the body parser is an action_parser_category type parser (a parser |
|---|
| 80 | // with an attached semantic action) we have to do something special. This |
|---|
| 81 | // happens, if the user wrote something like: |
|---|
| 82 | // |
|---|
| 83 | // confix_p(open, body[f], close) |
|---|
| 84 | // |
|---|
| 85 | // where 'body' is the parser matching the body of the confix sequence |
|---|
| 86 | // and 'f' is a functor to be called after matching the body. If we would |
|---|
| 87 | // do nothing, the resulting code would parse the sequence as follows: |
|---|
| 88 | // |
|---|
| 89 | // start >> (body[f] - close) >> close |
|---|
| 90 | // |
|---|
| 91 | // what in most cases is not what the user expects. |
|---|
| 92 | // (If this _is_ what you've expected, then please use the confix_p |
|---|
| 93 | // generator function 'direct()', which will inhibit |
|---|
| 94 | // re-attaching the actor to the body parser). |
|---|
| 95 | // |
|---|
| 96 | // To make the confix parser behave as expected: |
|---|
| 97 | // |
|---|
| 98 | // start >> (body - close)[f] >> close |
|---|
| 99 | // |
|---|
| 100 | // the actor attached to the 'body' parser has to be re-attached to the |
|---|
| 101 | // (body - close) parser construct, which will make the resulting confix |
|---|
| 102 | // parser 'do the right thing'. This refactoring is done by the help of |
|---|
| 103 | // the refactoring parsers (see the files refactoring.[hi]pp). |
|---|
| 104 | // |
|---|
| 105 | // Additionally special care must be taken, if the body parser is a |
|---|
| 106 | // unary_parser_category type parser as |
|---|
| 107 | // |
|---|
| 108 | // confix_p(open, *anychar_p, close) |
|---|
| 109 | // |
|---|
| 110 | // which without any refactoring would result in |
|---|
| 111 | // |
|---|
| 112 | // start >> (*anychar_p - close) >> close |
|---|
| 113 | // |
|---|
| 114 | // and will not give the expected result (*anychar_p will eat up all the |
|---|
| 115 | // input up to the end of the input stream). So we have to refactor this |
|---|
| 116 | // into: |
|---|
| 117 | // |
|---|
| 118 | // start >> *(anychar_p - close) >> close |
|---|
| 119 | // |
|---|
| 120 | // what will give the correct result. |
|---|
| 121 | // |
|---|
| 122 | // The case, where the body parser is a combination of the two mentioned |
|---|
| 123 | // problems (i.e. the body parser is a unary parser with an attached |
|---|
| 124 | // action), is handled accordingly too: |
|---|
| 125 | // |
|---|
| 126 | // confix_p(start, (*anychar_p)[f], end) |
|---|
| 127 | // |
|---|
| 128 | // will be parsed as expected: |
|---|
| 129 | // |
|---|
| 130 | // start >> (*(anychar_p - end))[f] >> end. |
|---|
| 131 | // |
|---|
| 132 | /////////////////////////////////////////////////////////////////////////////// |
|---|
| 133 | |
|---|
| 134 | template<typename NestedT, typename LexemeT> |
|---|
| 135 | struct confix_parser_gen |
|---|
| 136 | { |
|---|
| 137 | // Generic generator function for creation of concrete confix parsers |
|---|
| 138 | |
|---|
| 139 | template<typename StartT, typename ExprT, typename EndT> |
|---|
| 140 | confix_parser< |
|---|
| 141 | typename as_parser<StartT>::type, |
|---|
| 142 | typename as_parser<ExprT>::type, |
|---|
| 143 | typename as_parser<EndT>::type, |
|---|
| 144 | typename as_parser<ExprT>::type::parser_category_t, |
|---|
| 145 | NestedT, |
|---|
| 146 | LexemeT |
|---|
| 147 | > |
|---|
| 148 | operator()( |
|---|
| 149 | StartT const &start_, ExprT const &expr_, EndT const &end_) const |
|---|
| 150 | { |
|---|
| 151 | typedef typename as_parser<StartT>::type start_t; |
|---|
| 152 | typedef typename as_parser<ExprT>::type expr_t; |
|---|
| 153 | typedef typename as_parser<EndT>::type end_t; |
|---|
| 154 | typedef |
|---|
| 155 | typename as_parser<ExprT>::type::parser_category_t |
|---|
| 156 | parser_category_t; |
|---|
| 157 | |
|---|
| 158 | typedef |
|---|
| 159 | confix_parser< |
|---|
| 160 | start_t, expr_t, end_t, parser_category_t, NestedT, LexemeT |
|---|
| 161 | > |
|---|
| 162 | return_t; |
|---|
| 163 | |
|---|
| 164 | return return_t( |
|---|
| 165 | as_parser<StartT>::convert(start_), |
|---|
| 166 | as_parser<ExprT>::convert(expr_), |
|---|
| 167 | as_parser<EndT>::convert(end_) |
|---|
| 168 | ); |
|---|
| 169 | } |
|---|
| 170 | |
|---|
| 171 | // Generic generator function for creation of concrete confix parsers |
|---|
| 172 | // which have an action directly attached to the ExprT part of the |
|---|
| 173 | // parser (see comment above, no automatic refactoring) |
|---|
| 174 | |
|---|
| 175 | template<typename StartT, typename ExprT, typename EndT> |
|---|
| 176 | confix_parser< |
|---|
| 177 | typename as_parser<StartT>::type, |
|---|
| 178 | typename as_parser<ExprT>::type, |
|---|
| 179 | typename as_parser<EndT>::type, |
|---|
| 180 | plain_parser_category, // do not re-attach action |
|---|
| 181 | NestedT, |
|---|
| 182 | LexemeT |
|---|
| 183 | > |
|---|
| 184 | direct(StartT const &start_, ExprT const &expr_, EndT const &end_) const |
|---|
| 185 | { |
|---|
| 186 | typedef typename as_parser<StartT>::type start_t; |
|---|
| 187 | typedef typename as_parser<ExprT>::type expr_t; |
|---|
| 188 | typedef typename as_parser<EndT>::type end_t; |
|---|
| 189 | typedef plain_parser_category parser_category_t; |
|---|
| 190 | |
|---|
| 191 | typedef |
|---|
| 192 | confix_parser< |
|---|
| 193 | start_t, expr_t, end_t, parser_category_t, NestedT, LexemeT |
|---|
| 194 | > |
|---|
| 195 | return_t; |
|---|
| 196 | |
|---|
| 197 | return return_t( |
|---|
| 198 | as_parser<StartT>::convert(start_), |
|---|
| 199 | as_parser<ExprT>::convert(expr_), |
|---|
| 200 | as_parser<EndT>::convert(end_) |
|---|
| 201 | ); |
|---|
| 202 | } |
|---|
| 203 | }; |
|---|
| 204 | |
|---|
| 205 | /////////////////////////////////////////////////////////////////////////////// |
|---|
| 206 | // |
|---|
| 207 | // Predefined non_nested confix parser generators |
|---|
| 208 | // |
|---|
| 209 | /////////////////////////////////////////////////////////////////////////////// |
|---|
| 210 | |
|---|
| 211 | const confix_parser_gen<non_nested, non_lexeme> confix_p = |
|---|
| 212 | confix_parser_gen<non_nested, non_lexeme>(); |
|---|
| 213 | |
|---|
| 214 | /////////////////////////////////////////////////////////////////////////////// |
|---|
| 215 | // |
|---|
| 216 | // Comments are special types of confix parsers |
|---|
| 217 | // |
|---|
| 218 | // Comment parser generator template. This is a helper for generating a |
|---|
| 219 | // correct confix_parser<> from auxiliary parameters, which is able to |
|---|
| 220 | // parse comment constructs: (StartToken >> Comment text >> EndToken). |
|---|
| 221 | // |
|---|
| 222 | // There are the following types supported as parameters yet: parsers, |
|---|
| 223 | // single characters and strings (see as_parser). |
|---|
| 224 | // |
|---|
| 225 | // There are two diffenerent predefined comment parser generators |
|---|
| 226 | // (comment_p and comment_nest_p, see below), which may be used for |
|---|
| 227 | // creating special comment parsers in two different ways. |
|---|
| 228 | // |
|---|
| 229 | // If these are used with one parameter, a comment starting with the given |
|---|
| 230 | // first parser parameter up to the end of the line is matched. So for |
|---|
| 231 | // instance the following parser matches C++ style comments: |
|---|
| 232 | // |
|---|
| 233 | // comment_p("//"). |
|---|
| 234 | // |
|---|
| 235 | // If these are used with two parameters, a comment starting with the |
|---|
| 236 | // first parser parameter up to the second parser parameter is matched. |
|---|
| 237 | // For instance a C style comment parser should be constrcuted as: |
|---|
| 238 | // |
|---|
| 239 | // comment_p("/*", "*/"). |
|---|
| 240 | // |
|---|
| 241 | // Please note, that a comment is parsed implicitly as if the whole |
|---|
| 242 | // comment_p(...) statement were embedded into a lexeme_d[] directive. |
|---|
| 243 | // |
|---|
| 244 | /////////////////////////////////////////////////////////////////////////////// |
|---|
| 245 | |
|---|
| 246 | template<typename NestedT> |
|---|
| 247 | struct comment_parser_gen |
|---|
| 248 | { |
|---|
| 249 | // Generic generator function for creation of concrete comment parsers |
|---|
| 250 | // from an open token. The newline parser eol_p is used as the |
|---|
| 251 | // closing token. |
|---|
| 252 | |
|---|
| 253 | template<typename StartT> |
|---|
| 254 | confix_parser< |
|---|
| 255 | typename as_parser<StartT>::type, |
|---|
| 256 | kleene_star<anychar_parser>, |
|---|
| 257 | alternative<eol_parser, end_parser>, |
|---|
| 258 | unary_parser_category, // there is no action to re-attach |
|---|
| 259 | NestedT, |
|---|
| 260 | is_lexeme // insert implicit lexeme_d[] |
|---|
| 261 | > |
|---|
| 262 | operator() (StartT const &start_) const |
|---|
| 263 | { |
|---|
| 264 | typedef typename as_parser<StartT>::type start_t; |
|---|
| 265 | typedef kleene_star<anychar_parser> expr_t; |
|---|
| 266 | typedef alternative<eol_parser, end_parser> end_t; |
|---|
| 267 | typedef unary_parser_category parser_category_t; |
|---|
| 268 | |
|---|
| 269 | typedef |
|---|
| 270 | confix_parser< |
|---|
| 271 | start_t, expr_t, end_t, parser_category_t, NestedT, is_lexeme |
|---|
| 272 | > |
|---|
| 273 | return_t; |
|---|
| 274 | |
|---|
| 275 | return return_t( |
|---|
| 276 | as_parser<StartT>::convert(start_), |
|---|
| 277 | *anychar_p, |
|---|
| 278 | eol_p | end_p |
|---|
| 279 | ); |
|---|
| 280 | } |
|---|
| 281 | |
|---|
| 282 | // Generic generator function for creation of concrete comment parsers |
|---|
| 283 | // from an open and a close tokens. |
|---|
| 284 | |
|---|
| 285 | template<typename StartT, typename EndT> |
|---|
| 286 | confix_parser< |
|---|
| 287 | typename as_parser<StartT>::type, |
|---|
| 288 | kleene_star<anychar_parser>, |
|---|
| 289 | typename as_parser<EndT>::type, |
|---|
| 290 | unary_parser_category, // there is no action to re-attach |
|---|
| 291 | NestedT, |
|---|
| 292 | is_lexeme // insert implicit lexeme_d[] |
|---|
| 293 | > |
|---|
| 294 | operator() (StartT const &start_, EndT const &end_) const |
|---|
| 295 | { |
|---|
| 296 | typedef typename as_parser<StartT>::type start_t; |
|---|
| 297 | typedef kleene_star<anychar_parser> expr_t; |
|---|
| 298 | typedef typename as_parser<EndT>::type end_t; |
|---|
| 299 | typedef unary_parser_category parser_category_t; |
|---|
| 300 | |
|---|
| 301 | typedef |
|---|
| 302 | confix_parser< |
|---|
| 303 | start_t, expr_t, end_t, parser_category_t, NestedT, is_lexeme |
|---|
| 304 | > |
|---|
| 305 | return_t; |
|---|
| 306 | |
|---|
| 307 | return return_t( |
|---|
| 308 | as_parser<StartT>::convert(start_), |
|---|
| 309 | *anychar_p, |
|---|
| 310 | as_parser<EndT>::convert(end_) |
|---|
| 311 | ); |
|---|
| 312 | } |
|---|
| 313 | }; |
|---|
| 314 | |
|---|
| 315 | /////////////////////////////////////////////////////////////////////////////// |
|---|
| 316 | // |
|---|
| 317 | // Predefined non_nested comment parser generator |
|---|
| 318 | // |
|---|
| 319 | /////////////////////////////////////////////////////////////////////////////// |
|---|
| 320 | |
|---|
| 321 | const comment_parser_gen<non_nested> comment_p = |
|---|
| 322 | comment_parser_gen<non_nested>(); |
|---|
| 323 | |
|---|
| 324 | /////////////////////////////////////////////////////////////////////////////// |
|---|
| 325 | // |
|---|
| 326 | // comment_nest_parser class |
|---|
| 327 | // |
|---|
| 328 | // Parses a nested comments. |
|---|
| 329 | // Example: nested PASCAL-comments: |
|---|
| 330 | // |
|---|
| 331 | // { This is a { nested } PASCAL-comment } |
|---|
| 332 | // |
|---|
| 333 | /////////////////////////////////////////////////////////////////////////////// |
|---|
| 334 | |
|---|
| 335 | template<typename OpenT, typename CloseT> |
|---|
| 336 | struct comment_nest_parser: |
|---|
| 337 | public parser<comment_nest_parser<OpenT, CloseT> > |
|---|
| 338 | { |
|---|
| 339 | typedef comment_nest_parser<OpenT, CloseT> self_t; |
|---|
| 340 | |
|---|
| 341 | comment_nest_parser(OpenT const &open_, CloseT const &close_): |
|---|
| 342 | open(open_), close(close_) |
|---|
| 343 | {} |
|---|
| 344 | |
|---|
| 345 | template<typename ScannerT> |
|---|
| 346 | typename parser_result<self_t, ScannerT>::type |
|---|
| 347 | parse(ScannerT const &scan) const |
|---|
| 348 | { |
|---|
| 349 | return do_parse( |
|---|
| 350 | open >> *(*this | (anychar_p - close)) >> close, |
|---|
| 351 | scan); |
|---|
| 352 | } |
|---|
| 353 | |
|---|
| 354 | private: |
|---|
| 355 | template<typename ParserT, typename ScannerT> |
|---|
| 356 | typename parser_result<self_t, ScannerT>::type |
|---|
| 357 | do_parse(ParserT const &p, ScannerT const &scan) const |
|---|
| 358 | { |
|---|
| 359 | return |
|---|
| 360 | impl::contiguous_parser_parse< |
|---|
| 361 | BOOST_DEDUCED_TYPENAME parser_result<ParserT, ScannerT>::type |
|---|
| 362 | >(p, scan, scan); |
|---|
| 363 | } |
|---|
| 364 | |
|---|
| 365 | typename as_parser<OpenT>::type::embed_t open; |
|---|
| 366 | typename as_parser<CloseT>::type::embed_t close; |
|---|
| 367 | }; |
|---|
| 368 | |
|---|
| 369 | /////////////////////////////////////////////////////////////////////////////// |
|---|
| 370 | // |
|---|
| 371 | // Predefined nested comment parser generator |
|---|
| 372 | // |
|---|
| 373 | /////////////////////////////////////////////////////////////////////////////// |
|---|
| 374 | |
|---|
| 375 | template<typename OpenT, typename CloseT> |
|---|
| 376 | inline |
|---|
| 377 | comment_nest_parser< |
|---|
| 378 | typename as_parser<OpenT>::type, |
|---|
| 379 | typename as_parser<CloseT>::type |
|---|
| 380 | > |
|---|
| 381 | comment_nest_p(OpenT const &open, CloseT const &close) |
|---|
| 382 | { |
|---|
| 383 | return |
|---|
| 384 | comment_nest_parser< |
|---|
| 385 | BOOST_DEDUCED_TYPENAME as_parser<OpenT>::type, |
|---|
| 386 | BOOST_DEDUCED_TYPENAME as_parser<CloseT>::type |
|---|
| 387 | >( |
|---|
| 388 | as_parser<OpenT>::convert(open), |
|---|
| 389 | as_parser<CloseT>::convert(close) |
|---|
| 390 | ); |
|---|
| 391 | } |
|---|
| 392 | |
|---|
| 393 | /////////////////////////////////////////////////////////////////////////////// |
|---|
| 394 | }} // namespace boost::spirit |
|---|
| 395 | |
|---|
| 396 | #endif |
|---|