[29] | 1 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 2 | /// \file regex_primitives.hpp |
---|
| 3 | /// Contains the syntax elements for writing static regular expressions. |
---|
| 4 | // |
---|
| 5 | // Copyright 2004 Eric Niebler. Distributed under the Boost |
---|
| 6 | // Software License, Version 1.0. (See accompanying file |
---|
| 7 | // LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt) |
---|
| 8 | |
---|
| 9 | #ifndef BOOST_XPRESSIVE_REGEX_PRIMITIVES_HPP_EAN_10_04_2005 |
---|
| 10 | #define BOOST_XPRESSIVE_REGEX_PRIMITIVES_HPP_EAN_10_04_2005 |
---|
| 11 | |
---|
| 12 | #include <climits> |
---|
| 13 | #include <boost/mpl/assert.hpp> |
---|
| 14 | #include <boost/preprocessor/cat.hpp> |
---|
| 15 | #include <boost/xpressive/proto/proto.hpp> |
---|
| 16 | #include <boost/xpressive/detail/detail_fwd.hpp> |
---|
| 17 | #include <boost/xpressive/detail/core/icase.hpp> |
---|
| 18 | #include <boost/xpressive/detail/core/action.hpp> |
---|
| 19 | #include <boost/xpressive/detail/core/matchers.hpp> |
---|
| 20 | #include <boost/xpressive/detail/static/as_xpr.hpp> |
---|
| 21 | #include <boost/xpressive/detail/static/compile.hpp> |
---|
| 22 | #include <boost/xpressive/detail/static/modifier.hpp> |
---|
| 23 | #include <boost/xpressive/detail/static/regex_operators.hpp> |
---|
| 24 | #include <boost/xpressive/detail/static/productions/productions.hpp> |
---|
| 25 | |
---|
| 26 | namespace boost { namespace xpressive { namespace detail |
---|
| 27 | { |
---|
| 28 | |
---|
| 29 | typedef assert_word_placeholder<word_boundary<true> > assert_word_boundary; |
---|
| 30 | typedef assert_word_placeholder<word_begin> assert_word_begin; |
---|
| 31 | typedef assert_word_placeholder<word_end> assert_word_end; |
---|
| 32 | |
---|
| 33 | /* |
---|
| 34 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 35 | /// INTERNAL ONLY |
---|
| 36 | // BOOST_XPRESSIVE_GLOBAL |
---|
| 37 | // for defining globals that neither violate the One Definition Rule nor |
---|
| 38 | // lead to undefined behavior due to global object initialization order. |
---|
| 39 | //#define BOOST_XPRESSIVE_GLOBAL(type, name, init) \ |
---|
| 40 | // namespace detail \ |
---|
| 41 | // { \ |
---|
| 42 | // template<int Dummy> \ |
---|
| 43 | // struct BOOST_PP_CAT(global_pod_, name) \ |
---|
| 44 | // { \ |
---|
| 45 | // static type const value; \ |
---|
| 46 | // private: \ |
---|
| 47 | // union type_must_be_pod \ |
---|
| 48 | // { \ |
---|
| 49 | // type t; \ |
---|
| 50 | // char ch; \ |
---|
| 51 | // } u; \ |
---|
| 52 | // }; \ |
---|
| 53 | // template<int Dummy> \ |
---|
| 54 | // type const BOOST_PP_CAT(global_pod_, name)<Dummy>::value = init; \ |
---|
| 55 | // } \ |
---|
| 56 | // type const &name = detail::BOOST_PP_CAT(global_pod_, name)<0>::value |
---|
| 57 | */ |
---|
| 58 | |
---|
| 59 | } // namespace detail |
---|
| 60 | |
---|
| 61 | /// INTERNAL ONLY (for backwards compatibility) |
---|
| 62 | unsigned int const repeat_max = UINT_MAX-1; |
---|
| 63 | |
---|
| 64 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 65 | /// \brief For infinite repetition of a sub-expression. |
---|
| 66 | /// |
---|
| 67 | /// Magic value used with the repeat\<\>() function template |
---|
| 68 | /// to specify an unbounded repeat. Use as: repeat<17, inf>('a'). |
---|
| 69 | /// The equivalent in perl is /a{17,}/. |
---|
| 70 | unsigned int const inf = UINT_MAX-1; |
---|
| 71 | |
---|
| 72 | /// INTERNAL ONLY (for backwards compatibility) |
---|
| 73 | proto::op_proxy< |
---|
| 74 | proto::unary_op<detail::epsilon_matcher, proto::noop_tag> |
---|
| 75 | > const epsilon = {}; |
---|
| 76 | |
---|
| 77 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 78 | /// \brief Successfully matches nothing. |
---|
| 79 | /// |
---|
| 80 | /// Successfully matches a zero-width sequence. nil always succeeds and |
---|
| 81 | /// never consumes any characters. |
---|
| 82 | proto::op_proxy< |
---|
| 83 | proto::unary_op<detail::epsilon_matcher, proto::noop_tag> |
---|
| 84 | > const nil = {}; |
---|
| 85 | |
---|
| 86 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 87 | /// \brief Matches an alpha-numeric character. |
---|
| 88 | /// |
---|
| 89 | /// The regex traits are used to determine which characters are alpha-numeric. |
---|
| 90 | /// To match any character that is not alpha-numeric, use ~alnum. |
---|
| 91 | /// |
---|
| 92 | /// \attention alnum is equivalent to /[[:alnum:]]/ in perl. ~alnum is equivalent |
---|
| 93 | /// to /[[:^alnum:]]/ in perl. |
---|
| 94 | proto::op_proxy< |
---|
| 95 | proto::unary_op<detail::posix_charset_placeholder, proto::noop_tag> |
---|
| 96 | , char const * |
---|
| 97 | > const alnum = {"alnum"}; |
---|
| 98 | |
---|
| 99 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 100 | /// \brief Matches an alphabetic character. |
---|
| 101 | /// |
---|
| 102 | /// The regex traits are used to determine which characters are alphabetic. |
---|
| 103 | /// To match any character that is not alphabetic, use ~alpha. |
---|
| 104 | /// |
---|
| 105 | /// \attention alpha is equivalent to /[[:alpha:]]/ in perl. ~alpha is equivalent |
---|
| 106 | /// to /[[:^alpha:]]/ in perl. |
---|
| 107 | proto::op_proxy< |
---|
| 108 | proto::unary_op<detail::posix_charset_placeholder, proto::noop_tag> |
---|
| 109 | , char const * |
---|
| 110 | > const alpha = {"alpha"}; |
---|
| 111 | |
---|
| 112 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 113 | /// \brief Matches a blank (horizonal white-space) character. |
---|
| 114 | /// |
---|
| 115 | /// The regex traits are used to determine which characters are blank characters. |
---|
| 116 | /// To match any character that is not blank, use ~blank. |
---|
| 117 | /// |
---|
| 118 | /// \attention blank is equivalent to /[[:blank:]]/ in perl. ~blank is equivalent |
---|
| 119 | /// to /[[:^blank:]]/ in perl. |
---|
| 120 | proto::op_proxy< |
---|
| 121 | proto::unary_op<detail::posix_charset_placeholder, proto::noop_tag> |
---|
| 122 | , char const * |
---|
| 123 | > const blank = {"blank"}; |
---|
| 124 | |
---|
| 125 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 126 | /// \brief Matches a control character. |
---|
| 127 | /// |
---|
| 128 | /// The regex traits are used to determine which characters are control characters. |
---|
| 129 | /// To match any character that is not a control character, use ~cntrl. |
---|
| 130 | /// |
---|
| 131 | /// \attention cntrl is equivalent to /[[:cntrl:]]/ in perl. ~cntrl is equivalent |
---|
| 132 | /// to /[[:^cntrl:]]/ in perl. |
---|
| 133 | proto::op_proxy< |
---|
| 134 | proto::unary_op<detail::posix_charset_placeholder, proto::noop_tag> |
---|
| 135 | , char const * |
---|
| 136 | > const cntrl = {"cntrl"}; |
---|
| 137 | |
---|
| 138 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 139 | /// \brief Matches a digit character. |
---|
| 140 | /// |
---|
| 141 | /// The regex traits are used to determine which characters are digits. |
---|
| 142 | /// To match any character that is not a digit, use ~digit. |
---|
| 143 | /// |
---|
| 144 | /// \attention digit is equivalent to /[[:digit:]]/ in perl. ~digit is equivalent |
---|
| 145 | /// to /[[:^digit:]]/ in perl. |
---|
| 146 | proto::op_proxy< |
---|
| 147 | proto::unary_op<detail::posix_charset_placeholder, proto::noop_tag> |
---|
| 148 | , char const * |
---|
| 149 | > const digit = {"digit"}; |
---|
| 150 | |
---|
| 151 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 152 | /// \brief Matches a graph character. |
---|
| 153 | /// |
---|
| 154 | /// The regex traits are used to determine which characters are graphable. |
---|
| 155 | /// To match any character that is not graphable, use ~graph. |
---|
| 156 | /// |
---|
| 157 | /// \attention graph is equivalent to /[[:graph:]]/ in perl. ~graph is equivalent |
---|
| 158 | /// to /[[:^graph:]]/ in perl. |
---|
| 159 | proto::op_proxy< |
---|
| 160 | proto::unary_op<detail::posix_charset_placeholder, proto::noop_tag> |
---|
| 161 | , char const * |
---|
| 162 | > const graph = {"graph"}; |
---|
| 163 | |
---|
| 164 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 165 | /// \brief Matches a lower-case character. |
---|
| 166 | /// |
---|
| 167 | /// The regex traits are used to determine which characters are lower-case. |
---|
| 168 | /// To match any character that is not a lower-case character, use ~lower. |
---|
| 169 | /// |
---|
| 170 | /// \attention lower is equivalent to /[[:lower:]]/ in perl. ~lower is equivalent |
---|
| 171 | /// to /[[:^lower:]]/ in perl. |
---|
| 172 | proto::op_proxy< |
---|
| 173 | proto::unary_op<detail::posix_charset_placeholder, proto::noop_tag> |
---|
| 174 | , char const * |
---|
| 175 | > const lower = {"lower"}; |
---|
| 176 | |
---|
| 177 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 178 | /// \brief Matches a printable character. |
---|
| 179 | /// |
---|
| 180 | /// The regex traits are used to determine which characters are printable. |
---|
| 181 | /// To match any character that is not printable, use ~print. |
---|
| 182 | /// |
---|
| 183 | /// \attention print is equivalent to /[[:print:]]/ in perl. ~print is equivalent |
---|
| 184 | /// to /[[:^print:]]/ in perl. |
---|
| 185 | proto::op_proxy< |
---|
| 186 | proto::unary_op<detail::posix_charset_placeholder, proto::noop_tag> |
---|
| 187 | , char const * |
---|
| 188 | > const print = {"print"}; |
---|
| 189 | |
---|
| 190 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 191 | /// \brief Matches a punctuation character. |
---|
| 192 | /// |
---|
| 193 | /// The regex traits are used to determine which characters are punctuation. |
---|
| 194 | /// To match any character that is not punctuation, use ~punct. |
---|
| 195 | /// |
---|
| 196 | /// \attention punct is equivalent to /[[:punct:]]/ in perl. ~punct is equivalent |
---|
| 197 | /// to /[[:^punct:]]/ in perl. |
---|
| 198 | proto::op_proxy< |
---|
| 199 | proto::unary_op<detail::posix_charset_placeholder, proto::noop_tag> |
---|
| 200 | , char const * |
---|
| 201 | > const punct = {"punct"}; |
---|
| 202 | |
---|
| 203 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 204 | /// \brief Matches a space character. |
---|
| 205 | /// |
---|
| 206 | /// The regex traits are used to determine which characters are space characters. |
---|
| 207 | /// To match any character that is not white-space, use ~space. |
---|
| 208 | /// |
---|
| 209 | /// \attention space is equivalent to /[[:space:]]/ in perl. ~space is equivalent |
---|
| 210 | /// to /[[:^space:]]/ in perl. |
---|
| 211 | proto::op_proxy< |
---|
| 212 | proto::unary_op<detail::posix_charset_placeholder, proto::noop_tag> |
---|
| 213 | , char const * |
---|
| 214 | > const space = {"space"}; |
---|
| 215 | |
---|
| 216 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 217 | /// \brief Matches an upper-case character. |
---|
| 218 | /// |
---|
| 219 | /// The regex traits are used to determine which characters are upper-case. |
---|
| 220 | /// To match any character that is not upper-case, use ~upper. |
---|
| 221 | /// |
---|
| 222 | /// \attention upper is equivalent to /[[:upper:]]/ in perl. ~upper is equivalent |
---|
| 223 | /// to /[[:^upper:]]/ in perl. |
---|
| 224 | proto::op_proxy< |
---|
| 225 | proto::unary_op<detail::posix_charset_placeholder, proto::noop_tag> |
---|
| 226 | , char const * |
---|
| 227 | > const upper = {"upper"}; |
---|
| 228 | |
---|
| 229 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 230 | /// \brief Matches a hexadecimal digit character. |
---|
| 231 | /// |
---|
| 232 | /// The regex traits are used to determine which characters are hex digits. |
---|
| 233 | /// To match any character that is not a hex digit, use ~xdigit. |
---|
| 234 | /// |
---|
| 235 | /// \attention xdigit is equivalent to /[[:xdigit:]]/ in perl. ~xdigit is equivalent |
---|
| 236 | /// to /[[:^xdigit:]]/ in perl. |
---|
| 237 | proto::op_proxy< |
---|
| 238 | proto::unary_op<detail::posix_charset_placeholder, proto::noop_tag> |
---|
| 239 | , char const * |
---|
| 240 | > const xdigit = {"xdigit"}; |
---|
| 241 | |
---|
| 242 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 243 | /// \brief Beginning of sequence assertion. |
---|
| 244 | /// |
---|
| 245 | /// For the character sequence [begin, end), 'bos' matches the |
---|
| 246 | /// zero-width sub-sequence [begin, begin). |
---|
| 247 | proto::op_proxy< |
---|
| 248 | proto::unary_op<detail::assert_bos_matcher, proto::noop_tag> |
---|
| 249 | > const bos = {}; |
---|
| 250 | |
---|
| 251 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 252 | /// \brief End of sequence assertion. |
---|
| 253 | /// |
---|
| 254 | /// For the character sequence [begin, end), |
---|
| 255 | /// 'eos' matches the zero-width sub-sequence [end, end). |
---|
| 256 | /// |
---|
| 257 | /// \attention Unlike the perl end of sequence assertion \$, 'eos' will |
---|
| 258 | /// not match at the position [end-1, end-1) if *(end-1) is '\\n'. To |
---|
| 259 | /// get that behavior, use (!_n >> eos). |
---|
| 260 | proto::op_proxy< |
---|
| 261 | proto::unary_op<detail::assert_eos_matcher, proto::noop_tag> |
---|
| 262 | > const eos = {}; |
---|
| 263 | |
---|
| 264 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 265 | /// \brief Beginning of line assertion. |
---|
| 266 | /// |
---|
| 267 | /// 'bol' matches the zero-width sub-sequence |
---|
| 268 | /// immediately following a logical newline sequence. The regex traits |
---|
| 269 | /// is used to determine what constitutes a logical newline sequence. |
---|
| 270 | proto::op_proxy< |
---|
| 271 | proto::unary_op<detail::assert_bol_placeholder, proto::noop_tag> |
---|
| 272 | > const bol = {}; |
---|
| 273 | |
---|
| 274 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 275 | /// \brief End of line assertion. |
---|
| 276 | /// |
---|
| 277 | /// 'eol' matches the zero-width sub-sequence |
---|
| 278 | /// immediately preceeding a logical newline sequence. The regex traits |
---|
| 279 | /// is used to determine what constitutes a logical newline sequence. |
---|
| 280 | proto::op_proxy< |
---|
| 281 | proto::unary_op<detail::assert_eol_placeholder, proto::noop_tag> |
---|
| 282 | > const eol = {}; |
---|
| 283 | |
---|
| 284 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 285 | /// \brief Beginning of word assertion. |
---|
| 286 | /// |
---|
| 287 | /// 'bow' matches the zero-width sub-sequence |
---|
| 288 | /// immediately following a non-word character and preceeding a word character. |
---|
| 289 | /// The regex traits are used to determine what constitutes a word character. |
---|
| 290 | proto::op_proxy< |
---|
| 291 | proto::unary_op<detail::assert_word_begin, proto::noop_tag> |
---|
| 292 | > const bow = {}; |
---|
| 293 | |
---|
| 294 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 295 | /// \brief End of word assertion. |
---|
| 296 | /// |
---|
| 297 | /// 'eow' matches the zero-width sub-sequence |
---|
| 298 | /// immediately following a word character and preceeding a non-word character. |
---|
| 299 | /// The regex traits are used to determine what constitutes a word character. |
---|
| 300 | proto::op_proxy< |
---|
| 301 | proto::unary_op<detail::assert_word_end, proto::noop_tag> |
---|
| 302 | > const eow = {}; |
---|
| 303 | |
---|
| 304 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 305 | /// \brief Word boundary assertion. |
---|
| 306 | /// |
---|
| 307 | /// '_b' matches the zero-width sub-sequence at the beginning or the end of a word. |
---|
| 308 | /// It is equivalent to (bow | eow). The regex traits are used to determine what |
---|
| 309 | /// constitutes a word character. To match a non-word boundary, use ~_b. |
---|
| 310 | /// |
---|
| 311 | /// \attention _b is like \\b in perl. ~_b is like \\B in perl. |
---|
| 312 | proto::op_proxy< |
---|
| 313 | proto::unary_op<detail::assert_word_boundary, proto::noop_tag> |
---|
| 314 | > const _b = {}; |
---|
| 315 | |
---|
| 316 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 317 | /// \brief Matches a word character. |
---|
| 318 | /// |
---|
| 319 | /// '_w' matches a single word character. The regex traits are used to determine which |
---|
| 320 | /// characters are word characters. Use ~_w to match a character that is not a word |
---|
| 321 | /// character. |
---|
| 322 | /// |
---|
| 323 | /// \attention _w is like \\w in perl. ~_w is like \\W in perl. |
---|
| 324 | proto::op_proxy< |
---|
| 325 | proto::unary_op<detail::posix_charset_placeholder, proto::noop_tag> |
---|
| 326 | , char const * |
---|
| 327 | > const _w = {"w"}; |
---|
| 328 | |
---|
| 329 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 330 | /// \brief Matches a digit character. |
---|
| 331 | /// |
---|
| 332 | /// '_d' matches a single digit character. The regex traits are used to determine which |
---|
| 333 | /// characters are digits. Use ~_d to match a character that is not a digit |
---|
| 334 | /// character. |
---|
| 335 | /// |
---|
| 336 | /// \attention _d is like \\d in perl. ~_d is like \\D in perl. |
---|
| 337 | proto::op_proxy< |
---|
| 338 | proto::unary_op<detail::posix_charset_placeholder, proto::noop_tag> |
---|
| 339 | , char const * |
---|
| 340 | > const _d = {"d"}; |
---|
| 341 | |
---|
| 342 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 343 | /// \brief Matches a space character. |
---|
| 344 | /// |
---|
| 345 | /// '_s' matches a single space character. The regex traits are used to determine which |
---|
| 346 | /// characters are space characters. Use ~_s to match a character that is not a space |
---|
| 347 | /// character. |
---|
| 348 | /// |
---|
| 349 | /// \attention _s is like \\s in perl. ~_s is like \\S in perl. |
---|
| 350 | proto::op_proxy< |
---|
| 351 | proto::unary_op<detail::posix_charset_placeholder, proto::noop_tag> |
---|
| 352 | , char const * |
---|
| 353 | > const _s = {"s"}; |
---|
| 354 | |
---|
| 355 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 356 | /// \brief Matches a literal newline character, '\\n'. |
---|
| 357 | /// |
---|
| 358 | /// '_n' matches a single newline character, '\\n'. Use ~_n to match a character |
---|
| 359 | /// that is not a newline. |
---|
| 360 | /// |
---|
| 361 | /// \attention ~_n is like '.' in perl without the /s modifier. |
---|
| 362 | proto::op_proxy< |
---|
| 363 | proto::unary_op<detail::literal_placeholder<char>, proto::noop_tag> |
---|
| 364 | , char |
---|
| 365 | > const _n = {'\n'}; |
---|
| 366 | |
---|
| 367 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 368 | /// \brief Matches a logical newline sequence. |
---|
| 369 | /// |
---|
| 370 | /// '_ln' matches a logical newline sequence. This can be any character in the |
---|
| 371 | /// line separator class, as determined by the regex traits, or the '\\r\\n' sequence. |
---|
| 372 | /// For the purpose of back-tracking, '\\r\\n' is treated as a unit. |
---|
| 373 | /// To match any one character that is not a logical newline, use ~_ln. |
---|
| 374 | proto::op_proxy< |
---|
| 375 | detail::logical_newline_xpression |
---|
| 376 | > const _ln = {}; |
---|
| 377 | |
---|
| 378 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 379 | /// \brief Matches any one character. |
---|
| 380 | /// |
---|
| 381 | /// Match any character, similar to '.' in perl syntax with the /s modifier. |
---|
| 382 | /// '_' matches any one character, including the newline. |
---|
| 383 | /// |
---|
| 384 | /// \attention To match any character except the newline, use ~_n |
---|
| 385 | proto::op_proxy< |
---|
| 386 | proto::unary_op<detail::any_matcher, proto::noop_tag> |
---|
| 387 | > const _ = {}; |
---|
| 388 | |
---|
| 389 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 390 | /// \brief Reference to the current regex object |
---|
| 391 | /// |
---|
| 392 | /// Useful when constructing recursive regular expression objects. The 'self' |
---|
| 393 | /// identifier is a short-hand for the current regex object. For instance, |
---|
| 394 | /// sregex rx = '(' >> (self | nil) >> ')'; will create a regex object that |
---|
| 395 | /// matches balanced parens such as "((()))". |
---|
| 396 | proto::op_proxy< |
---|
| 397 | proto::unary_op<detail::self_placeholder, proto::noop_tag> |
---|
| 398 | > const self = {}; |
---|
| 399 | |
---|
| 400 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 401 | /// \brief Used to create character sets. |
---|
| 402 | /// |
---|
| 403 | /// There are two ways to create character sets with the 'set' identifier. The |
---|
| 404 | /// easiest is to create a comma-separated list of the characters in the set, |
---|
| 405 | /// as in (set= 'a','b','c'). This set will match 'a', 'b', or 'c'. The other |
---|
| 406 | /// way is to define the set as an argument to the set subscript operator. |
---|
| 407 | /// For instance, set[ 'a' | range('b','c') | digit ] will match an 'a', 'b', |
---|
| 408 | /// 'c' or a digit character. |
---|
| 409 | /// |
---|
| 410 | /// To complement a set, apply the '~' operator. For instance, ~(set= 'a','b','c') |
---|
| 411 | /// will match any character that is not an 'a', 'b', or 'c'. |
---|
| 412 | /// |
---|
| 413 | /// Sets can be composed of other, possibly complemented, sets. For instance, |
---|
| 414 | /// set[ ~digit | ~(set= 'a','b','c') ]. |
---|
| 415 | proto::op_proxy< |
---|
| 416 | detail::set_initializer_type |
---|
| 417 | > const set = {}; |
---|
| 418 | |
---|
| 419 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 420 | /// \brief Sub-match placeholder, like $& in Perl |
---|
| 421 | proto::op_proxy<detail::mark_tag, int> const s0 = {0}; |
---|
| 422 | |
---|
| 423 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 424 | /// \brief Sub-match placeholder, like $1 in perl. |
---|
| 425 | /// |
---|
| 426 | /// To create a sub-match, assign a sub-expression to the sub-match placeholder. |
---|
| 427 | /// For instance, (s1= _) will match any one character and remember which |
---|
| 428 | /// character was matched in the 1st sub-match. Later in the pattern, you can |
---|
| 429 | /// refer back to the sub-match. For instance, (s1= _) >> s1 will match any |
---|
| 430 | /// character, and then match the same character again. |
---|
| 431 | /// |
---|
| 432 | /// After a successful regex_match() or regex_search(), the sub-match placeholders |
---|
| 433 | /// can be used to index into the match_results\<\> object to retrieve the Nth |
---|
| 434 | /// sub-match. |
---|
| 435 | proto::op_proxy<detail::mark_tag, int> const s1 = {1}; |
---|
| 436 | proto::op_proxy<detail::mark_tag, int> const s2 = {2}; |
---|
| 437 | proto::op_proxy<detail::mark_tag, int> const s3 = {3}; |
---|
| 438 | proto::op_proxy<detail::mark_tag, int> const s4 = {4}; |
---|
| 439 | proto::op_proxy<detail::mark_tag, int> const s5 = {5}; |
---|
| 440 | proto::op_proxy<detail::mark_tag, int> const s6 = {6}; |
---|
| 441 | proto::op_proxy<detail::mark_tag, int> const s7 = {7}; |
---|
| 442 | proto::op_proxy<detail::mark_tag, int> const s8 = {8}; |
---|
| 443 | proto::op_proxy<detail::mark_tag, int> const s9 = {9}; |
---|
| 444 | |
---|
| 445 | // NOTE: For the purpose of xpressive's documentation, make icase() look like an |
---|
| 446 | // ordinary function. In reality, it is a function object defined in detail/icase.hpp |
---|
| 447 | // so that it can serve double-duty as regex_constants::icase, the syntax_option_type. |
---|
| 448 | // Do the same for as_xpr(), which is actually defined in detail/static/as_xpr.hpp |
---|
| 449 | #ifdef BOOST_XPRESSIVE_DOXYGEN_INVOKED |
---|
| 450 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 451 | /// \brief Makes a literal into a regular expression. |
---|
| 452 | /// |
---|
| 453 | /// Use as_xpr() to turn a literal into a regular expression. For instance, |
---|
| 454 | /// "foo" >> "bar" will not compile because both operands to the right-shift |
---|
| 455 | /// operator are const char*, and no such operator exists. Use as_xpr("foo") >> "bar" |
---|
| 456 | /// instead. |
---|
| 457 | /// |
---|
| 458 | /// You can use as_xpr() with character literals in addition to string literals. |
---|
| 459 | /// For instance, as_xpr('a') will match an 'a'. You can also complement a |
---|
| 460 | /// character literal, as with ~as_xpr('a'). This will match any one character |
---|
| 461 | /// that is not an 'a'. |
---|
| 462 | template<typename Literal> |
---|
| 463 | inline typename detail::as_xpr_type<Literal>::const_reference |
---|
| 464 | as_xpr(Literal const &literal) |
---|
| 465 | { |
---|
| 466 | return detail::as_xpr_type<Literal>::call(xpr); |
---|
| 467 | } |
---|
| 468 | |
---|
| 469 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 470 | /// \brief Makes a sub-expression case-insensitive. |
---|
| 471 | /// |
---|
| 472 | /// Use icase() to make a sub-expression case-insensitive. For instance, |
---|
| 473 | /// "foo" >> icase(set['b'] >> "ar") will match "foo" exactly followed by |
---|
| 474 | /// "bar" irrespective of case. |
---|
| 475 | template<typename Xpr> |
---|
| 476 | inline proto::binary_op<detail::icase_modifier, typename detail::as_xpr_type<Xpr>::type, modifier_tag> const |
---|
| 477 | icase(Xpr const &xpr) |
---|
| 478 | { |
---|
| 479 | detail::icase_modifier mod; |
---|
| 480 | return proto::make_op<modifier_tag>(mod, as_xpr(xpr)); |
---|
| 481 | } |
---|
| 482 | #endif |
---|
| 483 | |
---|
| 484 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 485 | /// \brief Embed a regex object by reference. |
---|
| 486 | /// |
---|
| 487 | /// \param rex The basic_regex object to embed by reference. |
---|
| 488 | template<typename BidiIter> |
---|
| 489 | inline proto::unary_op<detail::regex_placeholder<BidiIter, true>, proto::noop_tag> const |
---|
| 490 | by_ref(basic_regex<BidiIter> const &rex) |
---|
| 491 | { |
---|
| 492 | typedef detail::core_access<BidiIter> access; |
---|
| 493 | shared_ptr<detail::regex_impl<BidiIter> > impl = access::get_regex_impl(rex); |
---|
| 494 | return proto::noop(detail::regex_placeholder<BidiIter, true>(impl)); |
---|
| 495 | } |
---|
| 496 | |
---|
| 497 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 498 | /// \brief Match a range of characters. |
---|
| 499 | /// |
---|
| 500 | /// Match any character in the range [ch_min, ch_max]. |
---|
| 501 | /// |
---|
| 502 | /// \param ch_min The lower end of the range to match. |
---|
| 503 | /// \param ch_max The upper end of the range to match. |
---|
| 504 | template<typename Char> |
---|
| 505 | inline proto::unary_op<detail::range_placeholder<Char>, proto::noop_tag> const |
---|
| 506 | range(Char ch_min, Char ch_max) |
---|
| 507 | { |
---|
| 508 | return proto::noop(detail::range_placeholder<Char>(ch_min, ch_max)); |
---|
| 509 | } |
---|
| 510 | |
---|
| 511 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 512 | /// \brief Make a sub-expression optional. Equivalent to !as_xpr(xpr). |
---|
| 513 | /// |
---|
| 514 | /// \param xpr The sub-expression to make optional. |
---|
| 515 | template<typename Xpr> |
---|
| 516 | inline proto::unary_op |
---|
| 517 | < |
---|
| 518 | typename detail::as_xpr_type<Xpr>::type |
---|
| 519 | , proto::logical_not_tag |
---|
| 520 | > const |
---|
| 521 | optional(Xpr const &xpr) |
---|
| 522 | { |
---|
| 523 | return !as_xpr(xpr); |
---|
| 524 | } |
---|
| 525 | |
---|
| 526 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 527 | /// \brief Repeat a sub-expression multiple times. |
---|
| 528 | /// |
---|
| 529 | /// There are two forms of the repeat\<\>() function template. To match a |
---|
| 530 | /// sub-expression N times, use repeat\<N\>(xpr). To match a sub-expression |
---|
| 531 | /// from M to N times, use repeat\<M,N\>(xpr). |
---|
| 532 | /// |
---|
| 533 | /// The repeat\<\>() function creates a greedy quantifier. To make the quantifier |
---|
| 534 | /// non-greedy, apply the unary minus operator, as in -repeat\<M,N\>(xpr). |
---|
| 535 | /// |
---|
| 536 | /// \param xpr The sub-expression to repeat. |
---|
| 537 | template<unsigned int Min, unsigned int Max, typename Xpr> |
---|
| 538 | inline proto::unary_op |
---|
| 539 | < |
---|
| 540 | typename detail::as_xpr_type<Xpr>::type |
---|
| 541 | , detail::generic_quant_tag<Min, Max> |
---|
| 542 | > const |
---|
| 543 | repeat(Xpr const &xpr) |
---|
| 544 | { |
---|
| 545 | return proto::make_op<detail::generic_quant_tag<Min, Max> >(as_xpr(xpr)); |
---|
| 546 | } |
---|
| 547 | |
---|
| 548 | /// \overload |
---|
| 549 | template<unsigned int Count, typename Xpr2> |
---|
| 550 | inline proto::unary_op |
---|
| 551 | < |
---|
| 552 | typename detail::as_xpr_type<Xpr2>::type |
---|
| 553 | , detail::generic_quant_tag<Count, Count> |
---|
| 554 | > const |
---|
| 555 | repeat(Xpr2 const &xpr) |
---|
| 556 | { |
---|
| 557 | return proto::make_op<detail::generic_quant_tag<Count, Count> >(as_xpr(xpr)); |
---|
| 558 | } |
---|
| 559 | |
---|
| 560 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 561 | /// \brief Create an independent sub-expression. |
---|
| 562 | /// |
---|
| 563 | /// Turn off back-tracking for a sub-expression. Any branches or repeats within |
---|
| 564 | /// the sub-expression will match only one way, and no other alternatives are |
---|
| 565 | /// tried. |
---|
| 566 | /// |
---|
| 567 | /// \attention keep(xpr) is equivalent to the perl (?>...) extension. |
---|
| 568 | /// |
---|
| 569 | /// \param xpr The sub-expression to modify. |
---|
| 570 | template<typename Xpr> |
---|
| 571 | inline proto::unary_op |
---|
| 572 | < |
---|
| 573 | typename detail::as_xpr_type<Xpr>::type |
---|
| 574 | , detail::keeper_tag |
---|
| 575 | > const |
---|
| 576 | keep(Xpr const &xpr) |
---|
| 577 | { |
---|
| 578 | return proto::make_op<detail::keeper_tag>(as_xpr(xpr)); |
---|
| 579 | } |
---|
| 580 | |
---|
| 581 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 582 | /// \brief Look-ahead assertion. |
---|
| 583 | /// |
---|
| 584 | /// before(xpr) succeeds if the xpr sub-expression would match at the current |
---|
| 585 | /// position in the sequence, but xpr is not included in the match. For instance, |
---|
| 586 | /// before("foo") succeeds if we are before a "foo". Look-ahead assertions can be |
---|
| 587 | /// negated with the bit-compliment operator. |
---|
| 588 | /// |
---|
| 589 | /// \attention before(xpr) is equivalent to the perl (?=...) extension. |
---|
| 590 | /// ~before(xpr) is a negative look-ahead assertion, equivalent to the |
---|
| 591 | /// perl (?!...) extension. |
---|
| 592 | /// |
---|
| 593 | /// \param xpr The sub-expression to put in the look-ahead assertion. |
---|
| 594 | template<typename Xpr> |
---|
| 595 | inline proto::unary_op |
---|
| 596 | < |
---|
| 597 | typename detail::as_xpr_type<Xpr>::type |
---|
| 598 | , detail::lookahead_tag<true> |
---|
| 599 | > const |
---|
| 600 | before(Xpr const &xpr) |
---|
| 601 | { |
---|
| 602 | return proto::make_op<detail::lookahead_tag<true> >(as_xpr(xpr)); |
---|
| 603 | } |
---|
| 604 | |
---|
| 605 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 606 | /// \brief Look-behind assertion. |
---|
| 607 | /// |
---|
| 608 | /// after(xpr) succeeds if the xpr sub-expression would match at the current |
---|
| 609 | /// position minus N in the sequence, where N is the width of xpr. xpr is not included in |
---|
| 610 | /// the match. For instance, after("foo") succeeds if we are after a "foo". Look-behind |
---|
| 611 | /// assertions can be negated with the bit-complement operator. |
---|
| 612 | /// |
---|
| 613 | /// \attention after(xpr) is equivalent to the perl (?<=...) extension. |
---|
| 614 | /// ~after(xpr) is a negative look-behind assertion, equivalent to the |
---|
| 615 | /// perl (?<!...) extension. |
---|
| 616 | /// |
---|
| 617 | /// \param xpr The sub-expression to put in the look-ahead assertion. |
---|
| 618 | /// |
---|
| 619 | /// \pre xpr cannot match a variable number of characters. |
---|
| 620 | template<typename Xpr> |
---|
| 621 | inline proto::unary_op |
---|
| 622 | < |
---|
| 623 | typename detail::as_xpr_type<Xpr>::type |
---|
| 624 | , detail::lookbehind_tag<true> |
---|
| 625 | > const |
---|
| 626 | after(Xpr const &xpr) |
---|
| 627 | { |
---|
| 628 | return proto::make_op<detail::lookbehind_tag<true> >(as_xpr(xpr)); |
---|
| 629 | } |
---|
| 630 | |
---|
| 631 | /////////////////////////////////////////////////////////////////////////////// |
---|
| 632 | /// \brief Specify a regex traits or a std::locale. |
---|
| 633 | /// |
---|
| 634 | /// imbue() instructs the regex engine to use the specified traits or locale |
---|
| 635 | /// when matching the regex. The entire expression must use the same traits/locale. |
---|
| 636 | /// For instance, the following specifies a locale for use with a regex: |
---|
| 637 | /// std::locale loc; |
---|
| 638 | /// sregex rx = imbue(loc)(+digit); |
---|
| 639 | /// |
---|
| 640 | /// \param loc The std::locale or regex traits object. |
---|
| 641 | template<typename Locale> |
---|
| 642 | inline detail::modifier_op<detail::locale_modifier<Locale> > const |
---|
| 643 | imbue(Locale const &loc) |
---|
| 644 | { |
---|
| 645 | detail::modifier_op<detail::locale_modifier<Locale> > mod = |
---|
| 646 | { |
---|
| 647 | detail::locale_modifier<Locale>(loc) |
---|
| 648 | , regex_constants::ECMAScript |
---|
| 649 | }; |
---|
| 650 | return mod; |
---|
| 651 | } |
---|
| 652 | |
---|
| 653 | }} // namespace boost::xpressive |
---|
| 654 | |
---|
| 655 | #endif |
---|