1// class template regex -*- C++ -*-
2
3// Copyright (C) 2013-2025 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
31namespace std _GLIBCXX_VISIBILITY(default)
32{
33_GLIBCXX_BEGIN_NAMESPACE_VERSION
34
35#pragma GCC diagnostic push
36#pragma GCC diagnostic ignored "-Wc++17-extensions" // if constexpr
37namespace __detail
38{
39 template<typename _BiIter, typename _Alloc, typename _TraitsT,
40 bool __dfs_mode>
41 bool _Executor<_BiIter, _Alloc, _TraitsT, __dfs_mode>::
42 _M_search()
43 {
44 if (_M_search_from_first())
45 return true;
46 if (_M_flags & regex_constants::match_continuous)
47 return false;
48 _M_flags |= regex_constants::match_prev_avail;
49 while (_M_begin != _M_end)
50 {
51 ++_M_begin;
52 if (_M_search_from_first())
53 return true;
54 }
55 return false;
56 }
57
58 // The _M_main function operates in different modes, DFS mode or BFS mode,
59 // indicated by template parameter __dfs_mode, and dispatches to one of the
60 // _M_main_dispatch overloads.
61 //
62 // ------------------------------------------------------------
63 //
64 // DFS mode:
65 //
66 // It applies a Depth-First-Search (aka backtracking) on given NFA and input
67 // string.
68 // At the very beginning the executor stands in the start state, then it
69 // tries every possible state transition in current state recursively. Some
70 // state transitions consume input string, say, a single-char-matcher or a
71 // back-reference matcher; some don't, like assertion or other anchor nodes.
72 // When the input is exhausted and/or the current state is an accepting
73 // state, the whole executor returns true.
74 //
75 // TODO: This approach is exponentially slow for certain input.
76 // Try to compile the NFA to a DFA.
77 //
78 // Time complexity: \Omega(match_length), O(2^(_M_nfa.size()))
79 // Space complexity: \theta(match_results.size() + match_length)
80 //
81 template<typename _BiIter, typename _Alloc, typename _TraitsT,
82 bool __dfs_mode>
83 bool _Executor<_BiIter, _Alloc, _TraitsT, __dfs_mode>::
84 _M_main_dispatch(_Match_mode __match_mode, __dfs)
85 {
86 _M_has_sol = false;
87 *_M_states._M_get_sol_pos() = _BiIter();
88 _M_cur_results = _M_results;
89 _M_dfs(__match_mode, start: _M_states._M_start);
90 return _M_has_sol;
91 }
92
93 // ------------------------------------------------------------
94 //
95 // BFS mode:
96 //
97 // Russ Cox's article (http://swtch.com/~rsc/regexp/regexp1.html)
98 // explained this algorithm clearly.
99 //
100 // It first computes epsilon closure (states that can be achieved without
101 // consuming characters) for every state that's still matching,
102 // using the same DFS algorithm, but doesn't re-enter states (using
103 // _M_states._M_visited to check), nor follow _S_opcode_match.
104 //
105 // Then apply DFS using every _S_opcode_match (in _M_states._M_match_queue)
106 // as the start state.
107 //
108 // It significantly reduces potential duplicate states, so has a better
109 // upper bound; but it requires more overhead.
110 //
111 // Time complexity: \Omega(match_length * match_results.size())
112 // O(match_length * _M_nfa.size() * match_results.size())
113 // Space complexity: \Omega(_M_nfa.size() + match_results.size())
114 // O(_M_nfa.size() * match_results.size())
115 template<typename _BiIter, typename _Alloc, typename _TraitsT,
116 bool __dfs_mode>
117 bool _Executor<_BiIter, _Alloc, _TraitsT, __dfs_mode>::
118 _M_main_dispatch(_Match_mode __match_mode, __bfs)
119 {
120 _M_states._M_queue(_M_states._M_start, _M_results);
121 bool __ret = false;
122 while (1)
123 {
124 _M_has_sol = false;
125 if (_M_states._M_match_queue.empty())
126 break;
127 std::fill_n(_M_states._M_visited_states, _M_nfa.size(), false);
128 auto __old_queue = std::move(_M_states._M_match_queue);
129 auto __alloc = _M_cur_results.get_allocator();
130 for (auto& __task : __old_queue)
131 {
132 _M_cur_results = _ResultsVec(std::move(__task.second), __alloc);
133 _M_dfs(__match_mode, start: __task.first);
134 }
135 if (__match_mode == _Match_mode::_Prefix)
136 __ret |= _M_has_sol;
137 if (_M_current == _M_end)
138 break;
139 ++_M_current;
140 }
141 if (__match_mode == _Match_mode::_Exact)
142 __ret = _M_has_sol;
143 _M_states._M_match_queue.clear();
144 return __ret;
145 }
146
147 // Return whether now match the given sub-NFA.
148 template<typename _BiIter, typename _Alloc, typename _TraitsT,
149 bool __dfs_mode>
150 bool _Executor<_BiIter, _Alloc, _TraitsT, __dfs_mode>::
151 _M_lookahead(_StateIdT __next)
152 {
153 // Backreferences may refer to captured content.
154 // We may want to make this faster by not copying,
155 // but let's not be clever prematurely.
156 _ResultsVec __what(_M_cur_results);
157 _Executor __sub(_M_current, _M_end, __what, _M_re, _M_flags);
158 __sub._M_states._M_start = __next;
159 if (__sub._M_search_from_first())
160 {
161 for (size_t __i = 0; __i < __what.size(); __i++)
162 if (__what[__i].matched)
163 _M_cur_results[__i] = __what[__i];
164 return true;
165 }
166 return false;
167 }
168
169 // __rep_count records how many times (__rep_count.second)
170 // this node is visited under certain input iterator
171 // (__rep_count.first). This prevent the executor from entering
172 // infinite loop by refusing to continue when it's already been
173 // visited more than twice. It's `twice` instead of `once` because
174 // we need to spare one more time for potential group capture.
175 template<typename _BiIter, typename _Alloc, typename _TraitsT,
176 bool __dfs_mode>
177 void _Executor<_BiIter, _Alloc, _TraitsT, __dfs_mode>::
178 _M_rep_once_more(_Match_mode __match_mode, _StateIdT __i)
179 {
180 const auto& __state = _M_nfa[__i];
181 auto& __rep_count = _M_rep_count[__i];
182 if (__rep_count.second == 0 || __rep_count.first != _M_current)
183 {
184 auto __back = __rep_count;
185 __rep_count.first = _M_current;
186 __rep_count.second = 1;
187 _M_dfs(__match_mode, start: __state._M_alt);
188 __rep_count = __back;
189 }
190 else
191 {
192 if (__rep_count.second < 2)
193 {
194 __rep_count.second++;
195 _M_dfs(__match_mode, start: __state._M_alt);
196 __rep_count.second--;
197 }
198 }
199 }
200
201 // _M_alt branch is "match once more", while _M_next is "get me out
202 // of this quantifier". Executing _M_next first or _M_alt first don't
203 // mean the same thing, and we need to choose the correct order under
204 // given greedy mode.
205 template<typename _BiIter, typename _Alloc, typename _TraitsT,
206 bool __dfs_mode>
207 void _Executor<_BiIter, _Alloc, _TraitsT, __dfs_mode>::
208 _M_handle_repeat(_Match_mode __match_mode, _StateIdT __i)
209 {
210 const auto& __state = _M_nfa[__i];
211
212 // Greedy.
213 if (!__state._M_neg)
214 {
215 _M_rep_once_more(__match_mode, __i);
216 // If it's DFS executor and already accepted, we're done.
217 if (!__dfs_mode || !_M_has_sol)
218 _M_dfs(__match_mode, start: __state._M_next);
219 }
220 else // Non-greedy mode
221 {
222 if constexpr (__dfs_mode)
223 {
224 // vice-versa.
225 _M_dfs(__match_mode, start: __state._M_next);
226 if (!_M_has_sol)
227 _M_rep_once_more(__match_mode, __i);
228 }
229 else
230 {
231 // DON'T attempt anything, because there's already another
232 // state with higher priority accepted. This state cannot
233 // be better by attempting its next node.
234 if (!_M_has_sol)
235 {
236 _M_dfs(__match_mode, start: __state._M_next);
237 // DON'T attempt anything if it's already accepted. An
238 // accepted state *must* be better than a solution that
239 // matches a non-greedy quantifier one more time.
240 if (!_M_has_sol)
241 _M_rep_once_more(__match_mode, __i);
242 }
243 }
244 }
245 }
246
247 template<typename _BiIter, typename _Alloc, typename _TraitsT,
248 bool __dfs_mode>
249 void _Executor<_BiIter, _Alloc, _TraitsT, __dfs_mode>::
250 _M_handle_subexpr_begin(_Match_mode __match_mode, _StateIdT __i)
251 {
252 const auto& __state = _M_nfa[__i];
253
254 auto& __res = _M_cur_results[__state._M_subexpr];
255 auto __back = __res.first;
256 __res.first = _M_current;
257 _M_dfs(__match_mode, start: __state._M_next);
258 __res.first = __back;
259 }
260
261 template<typename _BiIter, typename _Alloc, typename _TraitsT,
262 bool __dfs_mode>
263 void _Executor<_BiIter, _Alloc, _TraitsT, __dfs_mode>::
264 _M_handle_subexpr_end(_Match_mode __match_mode, _StateIdT __i)
265 {
266 const auto& __state = _M_nfa[__i];
267
268 auto& __res = _M_cur_results[__state._M_subexpr];
269 auto __back = __res;
270 __res.second = _M_current;
271 __res.matched = true;
272 _M_dfs(__match_mode, start: __state._M_next);
273 __res = __back;
274 }
275
276 template<typename _BiIter, typename _Alloc, typename _TraitsT,
277 bool __dfs_mode>
278 inline void _Executor<_BiIter, _Alloc, _TraitsT, __dfs_mode>::
279 _M_handle_line_begin_assertion(_Match_mode __match_mode, _StateIdT __i)
280 {
281 const auto& __state = _M_nfa[__i];
282 if (_M_at_begin())
283 _M_dfs(__match_mode, start: __state._M_next);
284 }
285
286 template<typename _BiIter, typename _Alloc, typename _TraitsT,
287 bool __dfs_mode>
288 inline void _Executor<_BiIter, _Alloc, _TraitsT, __dfs_mode>::
289 _M_handle_line_end_assertion(_Match_mode __match_mode, _StateIdT __i)
290 {
291 const auto& __state = _M_nfa[__i];
292 if (_M_at_end())
293 _M_dfs(__match_mode, start: __state._M_next);
294 }
295
296 template<typename _BiIter, typename _Alloc, typename _TraitsT,
297 bool __dfs_mode>
298 inline void _Executor<_BiIter, _Alloc, _TraitsT, __dfs_mode>::
299 _M_handle_word_boundary(_Match_mode __match_mode, _StateIdT __i)
300 {
301 const auto& __state = _M_nfa[__i];
302 if (_M_word_boundary() == !__state._M_neg)
303 _M_dfs(__match_mode, start: __state._M_next);
304 }
305
306 // Here __state._M_alt offers a single start node for a sub-NFA.
307 // We recursively invoke our algorithm to match the sub-NFA.
308 template<typename _BiIter, typename _Alloc, typename _TraitsT,
309 bool __dfs_mode>
310 void _Executor<_BiIter, _Alloc, _TraitsT, __dfs_mode>::
311 _M_handle_subexpr_lookahead(_Match_mode __match_mode, _StateIdT __i)
312 {
313 const auto& __state = _M_nfa[__i];
314 if (_M_lookahead(next: __state._M_alt) == !__state._M_neg)
315 _M_dfs(__match_mode, start: __state._M_next);
316 }
317
318 template<typename _BiIter, typename _Alloc, typename _TraitsT,
319 bool __dfs_mode>
320 void _Executor<_BiIter, _Alloc, _TraitsT, __dfs_mode>::
321 _M_handle_match(_Match_mode __match_mode, _StateIdT __i)
322 {
323 const auto& __state = _M_nfa[__i];
324
325 if (_M_current == _M_end)
326 return;
327 if constexpr (__dfs_mode)
328 {
329 if (__state._M_matches(*_M_current))
330 {
331 ++_M_current;
332 _M_dfs(__match_mode, start: __state._M_next);
333 --_M_current;
334 }
335 }
336 else
337 if (__state._M_matches(*_M_current))
338 _M_states._M_queue(__state._M_next, _M_cur_results);
339 }
340
341 template<typename _BiIter, typename _TraitsT>
342 struct _Backref_matcher
343 {
344 _Backref_matcher(bool /* __icase */, const _TraitsT& __traits)
345 : _M_traits(__traits) { }
346
347 bool
348 _M_apply(_BiIter __expected_begin,
349 _BiIter __expected_end, _BiIter __actual_begin,
350 _BiIter __actual_end)
351 {
352 return _M_traits.transform(__expected_begin, __expected_end)
353 == _M_traits.transform(__actual_begin, __actual_end);
354 }
355
356 const _TraitsT& _M_traits;
357 };
358
359 template<typename _BiIter, typename _CharT>
360 struct _Backref_matcher<_BiIter, std::regex_traits<_CharT>>
361 {
362 using _TraitsT = std::regex_traits<_CharT>;
363 _Backref_matcher(bool __icase, const _TraitsT& __traits)
364 : _M_icase(__icase), _M_traits(__traits) { }
365
366 bool
367 _M_apply(_BiIter __expected_begin,
368 _BiIter __expected_end, _BiIter __actual_begin,
369 _BiIter __actual_end)
370 {
371 if (!_M_icase)
372 return _GLIBCXX_STD_A::__equal4(__expected_begin, __expected_end,
373 __actual_begin, __actual_end);
374 typedef std::ctype<_CharT> __ctype_type;
375 const auto& __fctyp = use_facet<__ctype_type>(_M_traits.getloc());
376 return _GLIBCXX_STD_A::__equal4(__expected_begin, __expected_end,
377 __actual_begin, __actual_end,
378 [this, &__fctyp](_CharT __lhs, _CharT __rhs)
379 {
380 return __fctyp.tolower(__lhs)
381 == __fctyp.tolower(__rhs);
382 });
383 }
384
385 bool _M_icase;
386 const _TraitsT& _M_traits;
387 };
388
389 // First fetch the matched result from _M_cur_results as __submatch;
390 // then compare it with
391 // (_M_current, _M_current + (__submatch.second - __submatch.first)).
392 // If matched, keep going; else just return and try another state.
393 template<typename _BiIter, typename _Alloc, typename _TraitsT,
394 bool __dfs_mode>
395 void _Executor<_BiIter, _Alloc, _TraitsT, __dfs_mode>::
396 _M_handle_backref(_Match_mode __match_mode, _StateIdT __i)
397 {
398 static_assert(__dfs_mode, "this should never be instantiated");
399
400 const auto& __state = _M_nfa[__i];
401 auto& __submatch = _M_cur_results[__state._M_backref_index];
402 if (!__submatch.matched)
403 return;
404 auto __last = _M_current;
405 for (auto __tmp = __submatch.first;
406 __last != _M_end && __tmp != __submatch.second;
407 ++__tmp)
408 ++__last;
409 if (_Backref_matcher<_BiIter, _TraitsT>(
410 _M_re.flags() & regex_constants::icase,
411 _M_re._M_automaton->_M_traits)._M_apply(
412 __submatch.first, __submatch.second, _M_current, __last))
413 {
414 if (__last != _M_current)
415 {
416 auto __backup = _M_current;
417 _M_current = __last;
418 _M_dfs(__match_mode, start: __state._M_next);
419 _M_current = __backup;
420 }
421 else
422 _M_dfs(__match_mode, start: __state._M_next);
423 }
424 }
425
426 template<typename _BiIter, typename _Alloc, typename _TraitsT,
427 bool __dfs_mode>
428 void _Executor<_BiIter, _Alloc, _TraitsT, __dfs_mode>::
429 _M_handle_accept(_Match_mode __match_mode, _StateIdT)
430 {
431 if constexpr (__dfs_mode)
432 {
433 __glibcxx_assert(!_M_has_sol);
434 if (__match_mode == _Match_mode::_Exact)
435 _M_has_sol = _M_current == _M_end;
436 else
437 _M_has_sol = true;
438 if (_M_current == _M_begin
439 && (_M_flags & regex_constants::match_not_null))
440 _M_has_sol = false;
441 if (_M_has_sol)
442 {
443 if (_M_nfa._M_flags & regex_constants::ECMAScript)
444 _M_results = _M_cur_results;
445 else // POSIX
446 {
447 __glibcxx_assert(_M_states._M_get_sol_pos());
448 // Here's POSIX's logic: match the longest one. However
449 // we never know which one (lhs or rhs of "|") is longer
450 // unless we try both of them and compare the results.
451 // The member variable _M_sol_pos records the end
452 // position of the last successful match. It's better
453 // to be larger, because POSIX regex is always greedy.
454 // TODO: This could be slow.
455 if (*_M_states._M_get_sol_pos() == _BiIter()
456 || std::distance(_M_begin,
457 *_M_states._M_get_sol_pos())
458 < std::distance(_M_begin, _M_current))
459 {
460 *_M_states._M_get_sol_pos() = _M_current;
461 _M_results = _M_cur_results;
462 }
463 }
464 }
465 }
466 else
467 {
468 if (_M_current == _M_begin
469 && (_M_flags & regex_constants::match_not_null))
470 return;
471 if (__match_mode == _Match_mode::_Prefix || _M_current == _M_end)
472 if (!_M_has_sol)
473 {
474 _M_has_sol = true;
475 _M_results = _M_cur_results;
476 }
477 }
478 }
479
480 template<typename _BiIter, typename _Alloc, typename _TraitsT,
481 bool __dfs_mode>
482 void _Executor<_BiIter, _Alloc, _TraitsT, __dfs_mode>::
483 _M_handle_alternative(_Match_mode __match_mode, _StateIdT __i)
484 {
485 const auto& __state = _M_nfa[__i];
486
487 if (_M_nfa._M_flags & regex_constants::ECMAScript)
488 {
489 // TODO: Fix BFS support. It is wrong.
490 _M_dfs(__match_mode, start: __state._M_alt);
491 // Pick lhs if it matches. Only try rhs if it doesn't.
492 if (!_M_has_sol)
493 _M_dfs(__match_mode, start: __state._M_next);
494 }
495 else
496 {
497 // Try both and compare the result.
498 // See "case _S_opcode_accept:" handling above.
499 _M_dfs(__match_mode, start: __state._M_alt);
500 auto __has_sol = _M_has_sol;
501 _M_has_sol = false;
502 _M_dfs(__match_mode, start: __state._M_next);
503 _M_has_sol |= __has_sol;
504 }
505 }
506
507 template<typename _BiIter, typename _Alloc, typename _TraitsT,
508 bool __dfs_mode>
509 void _Executor<_BiIter, _Alloc, _TraitsT, __dfs_mode>::
510 _M_dfs(_Match_mode __match_mode, _StateIdT __i)
511 {
512 if (_M_states._M_visited(__i))
513 return;
514
515 switch (_M_nfa[__i]._M_opcode())
516 {
517 case _S_opcode_repeat:
518 _M_handle_repeat(__match_mode, __i); break;
519 case _S_opcode_subexpr_begin:
520 _M_handle_subexpr_begin(__match_mode, __i); break;
521 case _S_opcode_subexpr_end:
522 _M_handle_subexpr_end(__match_mode, __i); break;
523 case _S_opcode_line_begin_assertion:
524 _M_handle_line_begin_assertion(__match_mode, __i); break;
525 case _S_opcode_line_end_assertion:
526 _M_handle_line_end_assertion(__match_mode, __i); break;
527 case _S_opcode_word_boundary:
528 _M_handle_word_boundary(__match_mode, __i); break;
529 case _S_opcode_subexpr_lookahead:
530 _M_handle_subexpr_lookahead(__match_mode, __i); break;
531 case _S_opcode_match:
532 _M_handle_match(__match_mode, __i); break;
533 case _S_opcode_backref:
534 if constexpr (__dfs_mode)
535 _M_handle_backref(__match_mode, __i);
536 else
537 __builtin_unreachable();
538 break;
539 case _S_opcode_accept:
540 _M_handle_accept(__match_mode, __i); break;
541 case _S_opcode_alternative:
542 _M_handle_alternative(__match_mode, __i); break;
543 default:
544 __glibcxx_assert(false);
545 }
546 }
547
548 // Return whether now is at some word boundary.
549 template<typename _BiIter, typename _Alloc, typename _TraitsT,
550 bool __dfs_mode>
551 bool _Executor<_BiIter, _Alloc, _TraitsT, __dfs_mode>::
552 _M_word_boundary() const
553 {
554 if (_M_current == _M_begin && (_M_flags & regex_constants::match_not_bow))
555 return false;
556 if (_M_current == _M_end && (_M_flags & regex_constants::match_not_eow))
557 return false;
558
559 bool __left_is_word = false;
560 if (_M_current != _M_begin
561 || (_M_flags & regex_constants::match_prev_avail))
562 {
563 auto __prev = _M_current;
564 if (_M_is_word(ch: *std::prev(__prev)))
565 __left_is_word = true;
566 }
567 bool __right_is_word =
568 _M_current != _M_end && _M_is_word(ch: *_M_current);
569
570 return __left_is_word != __right_is_word;
571 }
572} // namespace __detail
573#pragma GCC diagnostic pop
574
575_GLIBCXX_END_NAMESPACE_VERSION
576} // namespace
577