| 1 | /* |
| 2 | * |
| 3 | * Copyright (c) 1998-2002 |
| 4 | * John Maddock |
| 5 | * |
| 6 | * Use, modification and distribution are subject to the |
| 7 | * Boost Software License, Version 1.0. (See accompanying file |
| 8 | * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt) |
| 9 | * |
| 10 | */ |
| 11 | |
| 12 | /* |
| 13 | * LOCATION: see http://www.boost.org for most recent version. |
| 14 | * FILE states.cpp |
| 15 | * VERSION see <boost/version.hpp> |
| 16 | * DESCRIPTION: Declares internal state machine structures. |
| 17 | */ |
| 18 | |
| 19 | #ifndef BOOST_REGEX_V4_STATES_HPP |
| 20 | #define BOOST_REGEX_V4_STATES_HPP |
| 21 | |
| 22 | #ifdef BOOST_MSVC |
| 23 | #pragma warning(push) |
| 24 | #pragma warning(disable: 4103) |
| 25 | #endif |
| 26 | #ifdef BOOST_HAS_ABI_HEADERS |
| 27 | # include BOOST_ABI_PREFIX |
| 28 | #endif |
| 29 | #ifdef BOOST_MSVC |
| 30 | #pragma warning(pop) |
| 31 | #endif |
| 32 | |
| 33 | namespace boost{ |
| 34 | namespace BOOST_REGEX_DETAIL_NS{ |
| 35 | |
| 36 | /*** mask_type ******************************************************* |
| 37 | Whenever we have a choice of two alternatives, we use an array of bytes |
| 38 | to indicate which of the two alternatives it is possible to take for any |
| 39 | given input character. If mask_take is set, then we can take the next |
| 40 | state, and if mask_skip is set then we can take the alternative. |
| 41 | ***********************************************************************/ |
| 42 | enum mask_type |
| 43 | { |
| 44 | mask_take = 1, |
| 45 | mask_skip = 2, |
| 46 | mask_init = 4, |
| 47 | mask_any = mask_skip | mask_take, |
| 48 | mask_all = mask_any |
| 49 | }; |
| 50 | |
| 51 | /*** helpers ********************************************************** |
| 52 | These helpers let us use function overload resolution to detect whether |
| 53 | we have narrow or wide character strings: |
| 54 | ***********************************************************************/ |
| 55 | struct _narrow_type{}; |
| 56 | struct _wide_type{}; |
| 57 | template <class charT> struct is_byte; |
| 58 | template<> struct is_byte<char> { typedef _narrow_type width_type; }; |
| 59 | template<> struct is_byte<unsigned char>{ typedef _narrow_type width_type; }; |
| 60 | template<> struct is_byte<signed char> { typedef _narrow_type width_type; }; |
| 61 | template <class charT> struct is_byte { typedef _wide_type width_type; }; |
| 62 | |
| 63 | /*** enum syntax_element_type ****************************************** |
| 64 | Every record in the state machine falls into one of the following types: |
| 65 | ***********************************************************************/ |
| 66 | enum syntax_element_type |
| 67 | { |
| 68 | // start of a marked sub-expression, or perl-style (?...) extension |
| 69 | syntax_element_startmark = 0, |
| 70 | // end of a marked sub-expression, or perl-style (?...) extension |
| 71 | syntax_element_endmark = syntax_element_startmark + 1, |
| 72 | // any sequence of literal characters |
| 73 | syntax_element_literal = syntax_element_endmark + 1, |
| 74 | // start of line assertion: ^ |
| 75 | syntax_element_start_line = syntax_element_literal + 1, |
| 76 | // end of line assertion $ |
| 77 | syntax_element_end_line = syntax_element_start_line + 1, |
| 78 | // match any character: . |
| 79 | syntax_element_wild = syntax_element_end_line + 1, |
| 80 | // end of expression: we have a match when we get here |
| 81 | syntax_element_match = syntax_element_wild + 1, |
| 82 | // perl style word boundary: \b |
| 83 | syntax_element_word_boundary = syntax_element_match + 1, |
| 84 | // perl style within word boundary: \B |
| 85 | syntax_element_within_word = syntax_element_word_boundary + 1, |
| 86 | // start of word assertion: \< |
| 87 | syntax_element_word_start = syntax_element_within_word + 1, |
| 88 | // end of word assertion: \> |
| 89 | syntax_element_word_end = syntax_element_word_start + 1, |
| 90 | // start of buffer assertion: \` |
| 91 | syntax_element_buffer_start = syntax_element_word_end + 1, |
| 92 | // end of buffer assertion: \' |
| 93 | syntax_element_buffer_end = syntax_element_buffer_start + 1, |
| 94 | // backreference to previously matched sub-expression |
| 95 | syntax_element_backref = syntax_element_buffer_end + 1, |
| 96 | // either a wide character set [..] or one with multicharacter collating elements: |
| 97 | syntax_element_long_set = syntax_element_backref + 1, |
| 98 | // narrow character set: [...] |
| 99 | syntax_element_set = syntax_element_long_set + 1, |
| 100 | // jump to a new state in the machine: |
| 101 | syntax_element_jump = syntax_element_set + 1, |
| 102 | // choose between two production states: |
| 103 | syntax_element_alt = syntax_element_jump + 1, |
| 104 | // a repeat |
| 105 | syntax_element_rep = syntax_element_alt + 1, |
| 106 | // match a combining character sequence |
| 107 | syntax_element_combining = syntax_element_rep + 1, |
| 108 | // perl style soft buffer end: \z |
| 109 | syntax_element_soft_buffer_end = syntax_element_combining + 1, |
| 110 | // perl style continuation: \G |
| 111 | syntax_element_restart_continue = syntax_element_soft_buffer_end + 1, |
| 112 | // single character repeats: |
| 113 | syntax_element_dot_rep = syntax_element_restart_continue + 1, |
| 114 | syntax_element_char_rep = syntax_element_dot_rep + 1, |
| 115 | syntax_element_short_set_rep = syntax_element_char_rep + 1, |
| 116 | syntax_element_long_set_rep = syntax_element_short_set_rep + 1, |
| 117 | // a backstep for lookbehind repeats: |
| 118 | syntax_element_backstep = syntax_element_long_set_rep + 1, |
| 119 | // an assertion that a mark was matched: |
| 120 | syntax_element_assert_backref = syntax_element_backstep + 1, |
| 121 | syntax_element_toggle_case = syntax_element_assert_backref + 1, |
| 122 | // a recursive expression: |
| 123 | syntax_element_recurse = syntax_element_toggle_case + 1, |
| 124 | // Verbs: |
| 125 | syntax_element_fail = syntax_element_recurse + 1, |
| 126 | syntax_element_accept = syntax_element_fail + 1, |
| 127 | syntax_element_commit = syntax_element_accept + 1, |
| 128 | syntax_element_then = syntax_element_commit + 1 |
| 129 | }; |
| 130 | |
| 131 | #ifdef BOOST_REGEX_DEBUG |
| 132 | // dwa 09/26/00 - This is needed to suppress warnings about an ambiguous conversion |
| 133 | std::ostream& operator<<(std::ostream&, syntax_element_type); |
| 134 | #endif |
| 135 | |
| 136 | struct re_syntax_base; |
| 137 | |
| 138 | /*** union offset_type ************************************************ |
| 139 | Points to another state in the machine. During machine construction |
| 140 | we use integral offsets, but these are converted to pointers before |
| 141 | execution of the machine. |
| 142 | ***********************************************************************/ |
| 143 | union offset_type |
| 144 | { |
| 145 | re_syntax_base* p; |
| 146 | std::ptrdiff_t i; |
| 147 | }; |
| 148 | |
| 149 | /*** struct re_syntax_base ******************************************** |
| 150 | Base class for all states in the machine. |
| 151 | ***********************************************************************/ |
| 152 | struct re_syntax_base |
| 153 | { |
| 154 | syntax_element_type type; // what kind of state this is |
| 155 | offset_type next; // next state in the machine |
| 156 | }; |
| 157 | |
| 158 | /*** struct re_brace ************************************************** |
| 159 | A marked parenthesis. |
| 160 | ***********************************************************************/ |
| 161 | struct re_brace : public re_syntax_base |
| 162 | { |
| 163 | // The index to match, can be zero (don't mark the sub-expression) |
| 164 | // or negative (for perl style (?...) extentions): |
| 165 | int index; |
| 166 | bool icase; |
| 167 | }; |
| 168 | |
| 169 | /*** struct re_dot ************************************************** |
| 170 | Match anything. |
| 171 | ***********************************************************************/ |
| 172 | enum |
| 173 | { |
| 174 | dont_care = 1, |
| 175 | force_not_newline = 0, |
| 176 | force_newline = 2, |
| 177 | |
| 178 | test_not_newline = 2, |
| 179 | test_newline = 3 |
| 180 | }; |
| 181 | struct re_dot : public re_syntax_base |
| 182 | { |
| 183 | unsigned char mask; |
| 184 | }; |
| 185 | |
| 186 | /*** struct re_literal ************************************************ |
| 187 | A string of literals, following this structure will be an |
| 188 | array of characters: charT[length] |
| 189 | ***********************************************************************/ |
| 190 | struct re_literal : public re_syntax_base |
| 191 | { |
| 192 | unsigned int length; |
| 193 | }; |
| 194 | |
| 195 | /*** struct re_case ************************************************ |
| 196 | Indicates whether we are moving to a case insensive block or not |
| 197 | ***********************************************************************/ |
| 198 | struct re_case : public re_syntax_base |
| 199 | { |
| 200 | bool icase; |
| 201 | }; |
| 202 | |
| 203 | /*** struct re_set_long *********************************************** |
| 204 | A wide character set of characters, following this structure will be |
| 205 | an array of type charT: |
| 206 | First csingles null-terminated strings |
| 207 | Then 2 * cranges NULL terminated strings |
| 208 | Then cequivalents NULL terminated strings |
| 209 | ***********************************************************************/ |
| 210 | template <class mask_type> |
| 211 | struct re_set_long : public re_syntax_base |
| 212 | { |
| 213 | unsigned int csingles, cranges, cequivalents; |
| 214 | mask_type cclasses; |
| 215 | mask_type cnclasses; |
| 216 | bool isnot; |
| 217 | bool singleton; |
| 218 | }; |
| 219 | |
| 220 | /*** struct re_set **************************************************** |
| 221 | A set of narrow-characters, matches any of _map which is none-zero |
| 222 | ***********************************************************************/ |
| 223 | struct re_set : public re_syntax_base |
| 224 | { |
| 225 | unsigned char _map[1 << CHAR_BIT]; |
| 226 | }; |
| 227 | |
| 228 | /*** struct re_jump *************************************************** |
| 229 | Jump to a new location in the machine (not next). |
| 230 | ***********************************************************************/ |
| 231 | struct re_jump : public re_syntax_base |
| 232 | { |
| 233 | offset_type alt; // location to jump to |
| 234 | }; |
| 235 | |
| 236 | /*** struct re_alt *************************************************** |
| 237 | Jump to a new location in the machine (possibly next). |
| 238 | ***********************************************************************/ |
| 239 | struct re_alt : public re_jump |
| 240 | { |
| 241 | unsigned char _map[1 << CHAR_BIT]; // which characters can take the jump |
| 242 | unsigned int can_be_null; // true if we match a NULL string |
| 243 | }; |
| 244 | |
| 245 | /*** struct re_repeat ************************************************* |
| 246 | Repeat a section of the machine |
| 247 | ***********************************************************************/ |
| 248 | struct re_repeat : public re_alt |
| 249 | { |
| 250 | std::size_t min, max; // min and max allowable repeats |
| 251 | int state_id; // Unique identifier for this repeat |
| 252 | bool leading; // True if this repeat is at the start of the machine (lets us optimize some searches) |
| 253 | bool greedy; // True if this is a greedy repeat |
| 254 | }; |
| 255 | |
| 256 | /*** struct re_recurse ************************************************ |
| 257 | Recurse to a particular subexpression. |
| 258 | **********************************************************************/ |
| 259 | struct re_recurse : public re_jump |
| 260 | { |
| 261 | int state_id; // identifier of first nested repeat within the recursion. |
| 262 | }; |
| 263 | |
| 264 | /*** struct re_commit ************************************************* |
| 265 | Used for the PRUNE, SKIP and COMMIT verbs which basically differ only in what happens |
| 266 | if no match is found and we start searching forward. |
| 267 | **********************************************************************/ |
| 268 | enum commit_type |
| 269 | { |
| 270 | commit_prune, |
| 271 | commit_skip, |
| 272 | commit_commit |
| 273 | }; |
| 274 | struct re_commit : public re_syntax_base |
| 275 | { |
| 276 | commit_type action; |
| 277 | }; |
| 278 | |
| 279 | /*** enum re_jump_size_type ******************************************* |
| 280 | Provides compiled size of re_jump structure (allowing for trailing alignment). |
| 281 | We provide this so we know how manybytes to insert when constructing the machine |
| 282 | (The value of padding_mask is defined in regex_raw_buffer.hpp). |
| 283 | ***********************************************************************/ |
| 284 | enum re_jump_size_type |
| 285 | { |
| 286 | re_jump_size = (sizeof(re_jump) + padding_mask) & ~(padding_mask), |
| 287 | re_repeater_size = (sizeof(re_repeat) + padding_mask) & ~(padding_mask), |
| 288 | re_alt_size = (sizeof(re_alt) + padding_mask) & ~(padding_mask) |
| 289 | }; |
| 290 | |
| 291 | /*** proc re_is_set_member ********************************************* |
| 292 | Forward declaration: we'll need this one later... |
| 293 | ***********************************************************************/ |
| 294 | |
| 295 | template<class charT, class traits> |
| 296 | struct regex_data; |
| 297 | |
| 298 | template <class iterator, class charT, class traits_type, class char_classT> |
| 299 | iterator BOOST_REGEX_CALL re_is_set_member(iterator next, |
| 300 | iterator last, |
| 301 | const re_set_long<char_classT>* set_, |
| 302 | const regex_data<charT, traits_type>& e, bool icase); |
| 303 | |
| 304 | } // namespace BOOST_REGEX_DETAIL_NS |
| 305 | |
| 306 | } // namespace boost |
| 307 | |
| 308 | #ifdef BOOST_MSVC |
| 309 | #pragma warning(push) |
| 310 | #pragma warning(disable: 4103) |
| 311 | #endif |
| 312 | #ifdef BOOST_HAS_ABI_HEADERS |
| 313 | # include BOOST_ABI_SUFFIX |
| 314 | #endif |
| 315 | #ifdef BOOST_MSVC |
| 316 | #pragma warning(pop) |
| 317 | #endif |
| 318 | |
| 319 | #endif |
| 320 | |
| 321 | |
| 322 | |