libstdc++-v3/include/bits/regex_executor.tcc

   1 // class template regex -*- C++ -*-
   2
   3 // Copyright (C) 2013-2014 Free Software Foundation, Inc.
   4 //
   5 // This file is part of the GNU ISO C++ Library.  This library is free
   6 // software; you can redistribute it and/or modify it under the
   7 // terms of the GNU General Public License as published by the
   8 // Free Software Foundation; either version 3, or (at your option)
   9 // any later version.
  10
  11 // This library is distributed in the hope that it will be useful,
  12 // but WITHOUT ANY WARRANTY; without even the implied warranty of
  13 // MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  14 // GNU General Public License for more details.
  15
  16 // Under Section 7 of GPL version 3, you are granted additional
  17 // permissions described in the GCC Runtime Library Exception, version
  18 // 3.1, as published by the Free Software Foundation.
  19
  20 // You should have received a copy of the GNU General Public License and
  21 // a copy of the GCC Runtime Library Exception along with this program;
  22 // see the files COPYING3 and COPYING.RUNTIME respectively.  If not, see
  23 // <http://www.gnu.org/licenses/>.
  24
  25 /**
  26  *  @file bits/regex_executor.tcc
  27  *  This is an internal header file, included by other library headers.
  28  *  Do not attempt to use it directly. @headername{regex}
  29  */
  30
  31 namespace std _GLIBCXX_VISIBILITY(default)
  32 {
  33 namespace __detail
  34 {
  35 _GLIBCXX_BEGIN_NAMESPACE_VERSION
  36
  37   template<typename _BiIter, typename _Alloc, typename _TraitsT,
  38            bool __dfs_mode>
  39     bool _Executor<_BiIter, _Alloc, _TraitsT, __dfs_mode>::
  40     _M_search()
  41     {
  42       if (_M_search_from_first())
  43         return true;
  44       if (_M_flags & regex_constants::match_continuous)
  45         return false;
  46       _M_flags |= regex_constants::match_prev_avail;
  47       while (_M_begin != _M_end)
  48         {
  49           ++_M_begin;
  50           if (_M_search_from_first())
  51             return true;
  52         }
  53       return false;
  54     }
  55
  56   // The _M_main function operates in different modes, DFS mode or BFS mode,
  57   // indicated by template parameter __dfs_mode, and dispatches to one of the
  58   // _M_main_dispatch overloads.
  59   //
  60   // ------------------------------------------------------------
  61   //
  62   // DFS mode:
  63   //
  64   // It applies a Depth-First-Search (aka backtracking) on given NFA and input
  65   // string.
  66   // At the very beginning the executor stands in the start state, then it
  67   // tries every possible state transition in current state recursively. Some
  68   // state transitions consume input string, say, a single-char-matcher or a
  69   // back-reference matcher; some don't, like assertion or other anchor nodes.
  70   // When the input is exhausted and/or the current state is an accepting
  71   // state, the whole executor returns true.
  72   //
  73   // TODO: This approach is exponentially slow for certain input.
  74   //       Try to compile the NFA to a DFA.
  75   //
  76   // Time complexity: \Omega(match_length), O(2^(_M_nfa.size()))
  77   // Space complexity: \theta(match_results.size() + match_length)
  78   //
  79   template<typename _BiIter, typename _Alloc, typename _TraitsT,
  80            bool __dfs_mode>
  81     bool _Executor<_BiIter, _Alloc, _TraitsT, __dfs_mode>::
  82     _M_main_dispatch(_Match_mode __match_mode, __dfs)
  83     {
  84       _M_has_sol = false;
  85       *_M_states._M_get_sol_pos() = _BiIter();
  86       _M_cur_results = _M_results;
  87       _M_dfs(__match_mode, _M_states._M_start);
  88       return _M_has_sol;
  89     }
  90
  91   // ------------------------------------------------------------
  92   //
  93   // BFS mode:
  94   //
  95   // Russ Cox's article (http://swtch.com/~rsc/regexp/regexp1.html)
  96   // explained this algorithm clearly.
  97   //
  98   // It first computes epsilon closure (states that can be achieved without
  99   // consuming characters) for every state that's still matching,
 100   // using the same DFS algorithm, but doesn't re-enter states (using
 101   // _M_states._M_visited to check), nor follow _S_opcode_match.
 102   //
 103   // Then apply DFS using every _S_opcode_match (in _M_states._M_match_queue)
 104   // as the start state.
 105   //
 106   // It significantly reduces potential duplicate states, so has a better
 107   // upper bound; but it requires more overhead.
 108   //
 109   // Time complexity: \Omega(match_length * match_results.size())
 110   //                  O(match_length * _M_nfa.size() * match_results.size())
 111   // Space complexity: \Omega(_M_nfa.size() + match_results.size())
 112   //                   O(_M_nfa.size() * match_results.size())
 113   template<typename _BiIter, typename _Alloc, typename _TraitsT,
 114            bool __dfs_mode>
 115     bool _Executor<_BiIter, _Alloc, _TraitsT, __dfs_mode>::
 116     _M_main_dispatch(_Match_mode __match_mode, __bfs)
 117     {
 118       _M_states._M_queue(_M_states._M_start, _M_results);
 119       bool __ret = false;
 120       while (1)
 121         {
 122           _M_has_sol = false;
 123           if (_M_states._M_match_queue.empty())
 124             break;
 125           std::fill_n(_M_states._M_visited_states.get(), _M_nfa.size(), false);
 126           auto __old_queue = std::move(_M_states._M_match_queue);
 127           for (auto& __task : __old_queue)
 128             {
 129               _M_cur_results = std::move(__task.second);
 130               _M_dfs(__match_mode, __task.first);
 131             }
 132           if (__match_mode == _Match_mode::_Prefix)
 133             __ret |= _M_has_sol;
 134           if (_M_current == _M_end)
 135             break;
 136           ++_M_current;
 137         }
 138       if (__match_mode == _Match_mode::_Exact)
 139         __ret = _M_has_sol;
 140       _M_states._M_match_queue.clear();
 141       return __ret;
 142     }
 143
 144   // Return whether now match the given sub-NFA.
 145   template<typename _BiIter, typename _Alloc, typename _TraitsT,
 146            bool __dfs_mode>
 147     bool _Executor<_BiIter, _Alloc, _TraitsT, __dfs_mode>::
 148     _M_lookahead(_State<_TraitsT> __state)
 149     {
 150       _ResultsVec __what(_M_cur_results.size());
 151       _Executor __sub(_M_current, _M_end, __what, _M_re, _M_flags);
 152       __sub._M_states._M_start = __state._M_alt;
 153       if (__sub._M_search_from_first())
 154         {
 155           for (size_t __i = 0; __i < __what.size(); __i++)
 156             if (__what[__i].matched)
 157               _M_cur_results[__i] = __what[__i];
 158           return true;
 159         }
 160       return false;
 161     }
 162
 163   // __rep_count records how many times (__rep_count.second)
 164   // this node is visited under certain input iterator
 165   // (__rep_count.first). This prevent the executor from entering
 166   // infinite loop by refusing to continue when it's already been
 167   // visited more than twice. It's `twice` instead of `once` because
 168   // we need to spare one more time for potential group capture.
 169   template<typename _BiIter, typename _Alloc, typename _TraitsT,
 170     bool __dfs_mode>
 171     void _Executor<_BiIter, _Alloc, _TraitsT, __dfs_mode>::
 172     _M_rep_once_more(_Match_mode __match_mode, _StateIdT __i)
 173     {
 174       const auto& __state = _M_nfa[__i];
 175       auto& __rep_count = _M_rep_count[__i];
 176       if (__rep_count.second == 0 || __rep_count.first != _M_current)
 177         {
 178           auto __back = __rep_count;
 179           __rep_count.first = _M_current;
 180           __rep_count.second = 1;
 181           _M_dfs(__match_mode, __state._M_alt);
 182           __rep_count = __back;
 183         }
 184       else
 185         {
 186           if (__rep_count.second < 2)
 187             {
 188               __rep_count.second++;
 189               _M_dfs(__match_mode, __state._M_alt);
 190               __rep_count.second--;
 191             }
 192         }
 193     };
 194
 195   template<typename _BiIter, typename _Alloc, typename _TraitsT,
 196            bool __dfs_mode>
 197     void _Executor<_BiIter, _Alloc, _TraitsT, __dfs_mode>::
 198     _M_dfs(_Match_mode __match_mode, _StateIdT __i)
 199     {
 200       if (_M_states._M_visited(__i))
 201         return;
 202
 203       const auto& __state = _M_nfa[__i];
 204       // Every change on _M_cur_results and _M_current will be rolled back after
 205       // finishing the recursion step.
 206       switch (__state._M_opcode)
 207         {
 208         // _M_alt branch is "match once more", while _M_next is "get me out
 209         // of this quantifier". Executing _M_next first or _M_alt first don't
 210         // mean the same thing, and we need to choose the correct order under
 211         // given greedy mode.
 212         case _S_opcode_repeat:
 213           {
 214             // Greedy.
 215             if (!__state._M_neg)
 216               {
 217                 _M_rep_once_more(__match_mode, __i);
 218                 // If it's DFS executor and already accepted, we're done.
 219                 if (!__dfs_mode || !_M_has_sol)
 220                   _M_dfs(__match_mode, __state._M_next);
 221               }
 222             else // Non-greedy mode
 223               {
 224                 if (__dfs_mode)
 225                   {
 226                     // vice-versa.
 227                     _M_dfs(__match_mode, __state._M_next);
 228                     if (!_M_has_sol)
 229                       _M_rep_once_more(__match_mode, __i);
 230                   }
 231                 else
 232                   {
 233                     // DON'T attempt anything, because there's already another
 234                     // state with higher priority accepted. This state cannot
 235                     // be better by attempting its next node.
 236                     if (!_M_has_sol)
 237                       {
 238                         _M_dfs(__match_mode, __state._M_next);
 239                         // DON'T attempt anything if it's already accepted. An
 240                         // accepted state *must* be better than a solution that
 241                         // matches a non-greedy quantifier one more time.
 242                         if (!_M_has_sol)
 243                           _M_rep_once_more(__match_mode, __i);
 244                       }
 245                   }
 246               }
 247             }
 248           break;
 249         case _S_opcode_subexpr_begin:
 250           {
 251             auto& __res = _M_cur_results[__state._M_subexpr];
 252             auto __back = __res.first;
 253             __res.first = _M_current;
 254             _M_dfs(__match_mode, __state._M_next);
 255             __res.first = __back;
 256           }
 257           break;
 258         case _S_opcode_subexpr_end:
 259           {
 260             auto& __res = _M_cur_results[__state._M_subexpr];
 261             auto __back = __res;
 262             __res.second = _M_current;
 263             __res.matched = true;
 264             _M_dfs(__match_mode, __state._M_next);
 265             __res = __back;
 266           }
 267           break;
 268         case _S_opcode_line_begin_assertion:
 269           if (_M_at_begin())
 270             _M_dfs(__match_mode, __state._M_next);
 271           break;
 272         case _S_opcode_line_end_assertion:
 273           if (_M_at_end())
 274             _M_dfs(__match_mode, __state._M_next);
 275           break;
 276         case _S_opcode_word_boundary:
 277           if (_M_word_boundary() == !__state._M_neg)
 278             _M_dfs(__match_mode, __state._M_next);
 279           break;
 280         // Here __state._M_alt offers a single start node for a sub-NFA.
 281         // We recursively invoke our algorithm to match the sub-NFA.
 282         case _S_opcode_subexpr_lookahead:
 283           if (_M_lookahead(__state) == !__state._M_neg)
 284             _M_dfs(__match_mode, __state._M_next);
 285           break;
 286         case _S_opcode_match:
 287           if (_M_current == _M_end)
 288             break;
 289           if (__dfs_mode)
 290             {
 291               if (__state._M_matches(*_M_current))
 292                 {
 293                   ++_M_current;
 294                   _M_dfs(__match_mode, __state._M_next);
 295                   --_M_current;
 296                 }
 297             }
 298           else
 299             if (__state._M_matches(*_M_current))
 300               _M_states._M_queue(__state._M_next, _M_cur_results);
 301           break;
 302         // First fetch the matched result from _M_cur_results as __submatch;
 303         // then compare it with
 304         // (_M_current, _M_current + (__submatch.second - __submatch.first)).
 305         // If matched, keep going; else just return and try another state.
 306         case _S_opcode_backref:
 307           {
 308             _GLIBCXX_DEBUG_ASSERT(__dfs_mode);
 309             auto& __submatch = _M_cur_results[__state._M_backref_index];
 310             if (!__submatch.matched)
 311               break;
 312             auto __last = _M_current;
 313             for (auto __tmp = __submatch.first;
 314                  __last != _M_end && __tmp != __submatch.second;
 315                  ++__tmp)
 316               ++__last;
 317             if (_M_re._M_automaton->_M_traits.transform(__submatch.first,
 318                                                         __submatch.second)
 319                 == _M_re._M_automaton->_M_traits.transform(_M_current, __last))
 320               {
 321                 if (__last != _M_current)
 322                   {
 323                     auto __backup = _M_current;
 324                     _M_current = __last;
 325                     _M_dfs(__match_mode, __state._M_next);
 326                     _M_current = __backup;
 327                   }
 328                 else
 329                   _M_dfs(__match_mode, __state._M_next);
 330               }
 331           }
 332           break;
 333         case _S_opcode_accept:
 334           if (__dfs_mode)
 335             {
 336               _GLIBCXX_DEBUG_ASSERT(!_M_has_sol);
 337               if (__match_mode == _Match_mode::_Exact)
 338                 _M_has_sol = _M_current == _M_end;
 339               else
 340                 _M_has_sol = true;
 341               if (_M_current == _M_begin
 342                   && (_M_flags & regex_constants::match_not_null))
 343                 _M_has_sol = false;
 344               if (_M_has_sol)
 345                 {
 346                   if (_M_nfa._M_flags & regex_constants::ECMAScript)
 347                     _M_results = _M_cur_results;
 348                   else // POSIX
 349                     {
 350                       _GLIBCXX_DEBUG_ASSERT(_M_states._M_get_sol_pos());
 351                       // Here's POSIX's logic: match the longest one. However
 352                       // we never know which one (lhs or rhs of "|") is longer
 353                       // unless we try both of them and compare the results.
 354                       // The member variable _M_sol_pos records the end
 355                       // position of the last successful match. It's better
 356                       // to be larger, because POSIX regex is always greedy.
 357                       // TODO: This could be slow.
 358                       if (*_M_states._M_get_sol_pos() == _BiIter()
 359                           || std::distance(_M_begin,
 360                                            *_M_states._M_get_sol_pos())
 361                              < std::distance(_M_begin, _M_current))
 362                         {
 363                           *_M_states._M_get_sol_pos() = _M_current;
 364                           _M_results = _M_cur_results;
 365                         }
 366                     }
 367                 }
 368             }
 369           else
 370             {
 371               if (_M_current == _M_begin
 372                   && (_M_flags & regex_constants::match_not_null))
 373                 break;
 374               if (__match_mode == _Match_mode::_Prefix || _M_current == _M_end)
 375                 if (!_M_has_sol)
 376                   {
 377                     _M_has_sol = true;
 378                     _M_results = _M_cur_results;
 379                   }
 380             }
 381           break;
 382         case _S_opcode_alternative:
 383           if (_M_nfa._M_flags & regex_constants::ECMAScript)
 384             {
 385               // TODO: Let BFS support ECMAScript's alternative operation.
 386               _GLIBCXX_DEBUG_ASSERT(__dfs_mode);
 387               _M_dfs(__match_mode, __state._M_alt);
 388               // Pick lhs if it matches. Only try rhs if it doesn't.
 389               if (!_M_has_sol)
 390                 _M_dfs(__match_mode, __state._M_next);
 391             }
 392           else
 393             {
 394               // Try both and compare the result.
 395               // See "case _S_opcode_accept:" handling above.
 396               _M_dfs(__match_mode, __state._M_alt);
 397               auto __has_sol = _M_has_sol;
 398               _M_has_sol = false;
 399               _M_dfs(__match_mode, __state._M_next);
 400               _M_has_sol |= __has_sol;
 401             }
 402           break;
 403         default:
 404           _GLIBCXX_DEBUG_ASSERT(false);
 405         }
 406     }
 407
 408   // Return whether now is at some word boundary.
 409   template<typename _BiIter, typename _Alloc, typename _TraitsT,
 410            bool __dfs_mode>
 411     bool _Executor<_BiIter, _Alloc, _TraitsT, __dfs_mode>::
 412     _M_word_boundary() const
 413     {
 414       bool __left_is_word = false;
 415       if (_M_current != _M_begin
 416           || (_M_flags & regex_constants::match_prev_avail))
 417         {
 418           auto __prev = _M_current;
 419           if (_M_is_word(*std::prev(__prev)))
 420             __left_is_word = true;
 421         }
 422       bool __right_is_word =
 423         _M_current != _M_end && _M_is_word(*_M_current);
 424
 425       if (__left_is_word == __right_is_word)
 426         return false;
 427       if (__left_is_word && !(_M_flags & regex_constants::match_not_eow))
 428         return true;
 429       if (__right_is_word && !(_M_flags & regex_constants::match_not_bow))
 430         return true;
 431       return false;
 432     }
 433
 434 _GLIBCXX_END_NAMESPACE_VERSION
 435 } // namespace __detail
 436 } // namespace