1 // class template regex -*- C++ -*-
3 // Copyright (C) 2013 Free Software Foundation, Inc.
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)
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.
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.
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/>.
26 * @file bits/regex.tcc
27 * This is an internal header file, included by other library headers.
28 * Do not attempt to use it directly. @headername{regex}
31 // See below __regex_algo_impl to get what this is talking about. The default
32 // value 1 indicated a conservative optimization without giving up worst case
34 #ifndef _GLIBCXX_REGEX_DFS_QUANTIFIERS_LIMIT
35 #define _GLIBCXX_REGEX_DFS_QUANTIFIERS_LIMIT 1
38 namespace std _GLIBCXX_VISIBILITY(default)
40 _GLIBCXX_BEGIN_NAMESPACE_VERSION
44 _GLIBCXX_BEGIN_NAMESPACE_VERSION
46 // Result of merging regex_match and regex_search.
48 // __policy now can be _S_auto (auto dispatch) and _S_alternate (use
49 // the other one if possible, for test purpose).
51 // That __match_mode is true means regex_match, else regex_search.
52 template<typename _BiIter, typename _Alloc,
53 typename _CharT, typename _TraitsT,
54 _RegexExecutorPolicy __policy,
57 __regex_algo_impl(_BiIter __s,
59 match_results<_BiIter, _Alloc>& __m,
60 const basic_regex<_CharT, _TraitsT>& __re,
61 regex_constants::match_flag_type __flags)
63 if (__re._M_automaton == nullptr)
66 typename match_results<_BiIter, _Alloc>::_Base_type& __res = __m;
67 __res.resize(__re._M_automaton->_M_sub_count() + 2);
68 for (auto& __it : __res)
71 // This function decide which executor to use under given circumstances.
72 // The _S_auto policy now is the following: if a NFA has no
73 // back-references and has more than _GLIBCXX_REGEX_DFS_QUANTIFIERS_LIMIT
74 // quantifiers (*, +, ?), the BFS executor will be used, other wise
75 // DFS executor. This is because DFS executor has a exponential upper
76 // bound, but better best-case performace. Meanwhile, BFS executor can
77 // effectively prevent from exponential-long time matching (which must
78 // contains many quantifiers), but it's slower in average.
80 // For simple regex, BFS executor could be 2 or more times slower than
83 // Of course, BFS executor cannot handle back-references.
85 if (!__re._M_automaton->_M_has_backref
86 && (__policy == _RegexExecutorPolicy::_S_alternate
87 || __re._M_automaton->_M_quant_count
88 > _GLIBCXX_REGEX_DFS_QUANTIFIERS_LIMIT))
90 _Executor<_BiIter, _Alloc, _TraitsT, false>
91 __executor(__s, __e, __m, __re, __flags);
93 __ret = __executor._M_match();
95 __ret = __executor._M_search();
99 _Executor<_BiIter, _Alloc, _TraitsT, true>
100 __executor(__s, __e, __m, __re, __flags);
102 __ret = __executor._M_match();
104 __ret = __executor._M_search();
108 for (auto __it : __res)
110 __it.first = __it.second = __e;
111 auto& __pre = __res[__res.size()-2];
112 auto& __suf = __res[__res.size()-1];
115 __pre.matched = false;
118 __suf.matched = false;
125 __pre.second = __res[0].first;
126 __pre.matched = (__pre.first != __pre.second);
127 __suf.first = __res[0].second;
129 __suf.matched = (__suf.first != __suf.second);
131 if (__re.flags() & regex_constants::nosubs)
137 _GLIBCXX_END_NAMESPACE_VERSION
140 template<typename _Ch_type>
141 template<typename _Fwd_iter>
142 typename regex_traits<_Ch_type>::string_type
143 regex_traits<_Ch_type>::
144 lookup_collatename(_Fwd_iter __first, _Fwd_iter __last) const
146 typedef std::ctype<char_type> __ctype_type;
147 const __ctype_type& __fctyp(use_facet<__ctype_type>(_M_locale));
149 static const char* __collatenames[] =
242 "left-square-bracket",
244 "right-square-bracket",
274 "left-curly-bracket",
276 "right-curly-bracket",
283 //static const char* __digraphs[] =
309 std::string __s(__last - __first, '?');
310 __fctyp.narrow(__first, __last, '?', &*__s.begin());
312 for (unsigned int __i = 0; *__collatenames[__i]; __i++)
313 if (__s == __collatenames[__i])
314 return string_type(1, __fctyp.widen(static_cast<char>(__i)));
316 //for (unsigned int __i = 0; *__digraphs[__i]; __i++)
318 // const char* __now = __digraphs[__i];
321 // string_type ret(__s.size(), __fctyp.widen('?'));
322 // __fctyp.widen(__now, __now + 2/* ouch */, &*ret.begin());
326 return string_type();
329 template<typename _Ch_type>
330 template<typename _Fwd_iter>
331 typename regex_traits<_Ch_type>::char_class_type
332 regex_traits<_Ch_type>::
333 lookup_classname(_Fwd_iter __first, _Fwd_iter __last, bool __icase) const
335 typedef std::ctype<char_type> __ctype_type;
336 typedef std::ctype<char> __cctype_type;
337 typedef const pair<const char*, char_class_type> _ClassnameEntry;
338 const __ctype_type& __fctyp(use_facet<__ctype_type>(_M_locale));
339 const __cctype_type& __cctyp(use_facet<__cctype_type>(_M_locale));
341 static _ClassnameEntry __classnames[] =
343 {"d", ctype_base::digit},
344 {"w", {ctype_base::alnum, _RegexMask::_S_under}},
345 {"s", ctype_base::space},
346 {"alnum", ctype_base::alnum},
347 {"alpha", ctype_base::alpha},
348 {"blank", {0, _RegexMask::_S_blank}},
349 {"cntrl", ctype_base::cntrl},
350 {"digit", ctype_base::digit},
351 {"graph", ctype_base::graph},
352 {"lower", ctype_base::lower},
353 {"print", ctype_base::print},
354 {"punct", ctype_base::punct},
355 {"space", ctype_base::space},
356 {"upper", ctype_base::upper},
357 {"xdigit", ctype_base::xdigit},
360 std::string __s(__last - __first, '?');
361 __fctyp.narrow(__first, __last, '?', &__s[0]);
362 __cctyp.tolower(&*__s.begin(), &*__s.begin() + __s.size());
363 for (_ClassnameEntry* __it = __classnames;
364 __it < *(&__classnames + 1);
367 if (__s == __it->first)
371 & (ctype_base::lower | ctype_base::upper)) != 0))
372 return ctype_base::alpha;
379 template<typename _Ch_type>
381 regex_traits<_Ch_type>::
382 isctype(_Ch_type __c, char_class_type __f) const
384 typedef std::ctype<char_type> __ctype_type;
385 const __ctype_type& __fctyp(use_facet<__ctype_type>(_M_locale));
387 return __fctyp.is(__f._M_base, __c)
389 || ((__f._M_extended & _RegexMask::_S_under)
390 && __c == __fctyp.widen('_'))
392 || ((__f._M_extended & _RegexMask::_S_blank)
393 && (__c == __fctyp.widen(' ')
394 || __c == __fctyp.widen('\t')));
397 template<typename _Ch_type>
399 regex_traits<_Ch_type>::
400 value(_Ch_type __ch, int __radix) const
402 std::basic_istringstream<char_type> __is(string_type(1, __ch));
406 else if (__radix == 16)
409 return __is.fail() ? -1 : __v;
412 template<typename _Bi_iter, typename _Alloc>
413 template<typename _Out_iter>
414 _Out_iter match_results<_Bi_iter, _Alloc>::
415 format(_Out_iter __out,
416 const match_results<_Bi_iter, _Alloc>::char_type* __fmt_first,
417 const match_results<_Bi_iter, _Alloc>::char_type* __fmt_last,
418 match_flag_type __flags) const
420 _GLIBCXX_DEBUG_ASSERT( ready() );
421 regex_traits<char_type> __traits;
422 typedef std::ctype<char_type> __ctype_type;
424 __fctyp(use_facet<__ctype_type>(__traits.getloc()));
426 auto __output = [&](size_t __idx)
428 auto& __sub = _Base_type::operator[](__idx);
430 std::copy(__sub.first, __sub.second, __out);
433 if (__flags & regex_constants::format_sed)
435 for (; __fmt_first != __fmt_last;)
436 if (*__fmt_first == '&')
441 else if (*__fmt_first == '\\')
443 if (++__fmt_first != __fmt_last
444 && __fctyp.is(__ctype_type::digit, *__fmt_first))
445 __output(__traits.value(*__fmt_first++, 10));
450 *__out++ = *__fmt_first++;
456 auto __next = std::find(__fmt_first, __fmt_last, '$');
457 if (__next == __fmt_last)
460 std::copy(__fmt_first, __next, __out);
462 auto __eat = [&](char __ch) -> bool
472 if (++__next == __fmt_last)
479 __output(_Base_type::size()-2);
480 else if (__eat('\''))
481 __output(_Base_type::size()-1);
482 else if (__fctyp.is(__ctype_type::digit, *__next))
484 long __num = __traits.value(*__next, 10);
485 if (++__next != __fmt_last
486 && __fctyp.is(__ctype_type::digit, *__next))
489 __num += __traits.value(*__next++, 10);
491 if (0 <= __num && __num < this->size())
496 __fmt_first = __next;
498 std::copy(__fmt_first, __fmt_last, __out);
503 template<typename _Out_iter, typename _Bi_iter,
504 typename _Rx_traits, typename _Ch_type>
506 regex_replace(_Out_iter __out, _Bi_iter __first, _Bi_iter __last,
507 const basic_regex<_Ch_type, _Rx_traits>& __e,
508 const _Ch_type* __fmt,
509 regex_constants::match_flag_type __flags)
511 typedef regex_iterator<_Bi_iter, _Ch_type, _Rx_traits> _IterT;
512 _IterT __i(__first, __last, __e, __flags);
516 if (!(__flags & regex_constants::format_no_copy))
517 std::copy(__first, __last, __out);
521 sub_match<_Bi_iter> __last;
522 auto __len = char_traits<_Ch_type>::length(__fmt);
523 for (; __i != __end; ++__i)
525 if (!(__flags & regex_constants::format_no_copy))
526 std::copy(__i->prefix().first, __i->prefix().second, __out);
527 __out = __i->format(__out, __fmt, __fmt + __len, __flags);
528 __last = __i->suffix();
529 if (__flags & regex_constants::format_first_only)
532 if (!(__flags & regex_constants::format_no_copy))
533 std::copy(__last.first, __last.second, __out);
538 template<typename _Bi_iter,
542 regex_iterator<_Bi_iter, _Ch_type, _Rx_traits>::
543 operator==(const regex_iterator& __rhs) const
545 return (_M_match.empty() && __rhs._M_match.empty())
546 || (_M_begin == __rhs._M_begin
547 && _M_end == __rhs._M_end
548 && _M_pregex == __rhs._M_pregex
549 && _M_flags == __rhs._M_flags
550 && _M_match[0] == __rhs._M_match[0]);
553 template<typename _Bi_iter,
556 regex_iterator<_Bi_iter, _Ch_type, _Rx_traits>&
557 regex_iterator<_Bi_iter, _Ch_type, _Rx_traits>::
560 // In all cases in which the call to regex_search returns true,
561 // match.prefix().first shall be equal to the previous value of
562 // match[0].second, and for each index i in the half-open range
563 // [0, match.size()) for which match[i].matched is true,
564 // match[i].position() shall return distance(begin, match[i].first).
566 if (_M_match[0].matched)
568 auto __start = _M_match[0].second;
569 auto __prefix_first = _M_match[0].second;
570 if (_M_match[0].first == _M_match[0].second)
572 if (__start == _M_end)
574 _M_match = value_type();
579 if (regex_search(__start, _M_end, _M_match, *_M_pregex,
581 | regex_constants::match_not_null
582 | regex_constants::match_continuous))
584 _GLIBCXX_DEBUG_ASSERT(_M_match[0].matched);
585 _M_match.at(_M_match.size()).first = __prefix_first;
586 _M_match._M_in_iterator = true;
587 _M_match._M_begin = _M_begin;
594 _M_flags |= regex_constants::match_prev_avail;
595 if (regex_search(__start, _M_end, _M_match, *_M_pregex, _M_flags))
597 _GLIBCXX_DEBUG_ASSERT(_M_match[0].matched);
598 _M_match.at(_M_match.size()).first = __prefix_first;
599 _M_match._M_in_iterator = true;
600 _M_match._M_begin = _M_begin;
603 _M_match = value_type();
608 template<typename _Bi_iter,
611 regex_token_iterator<_Bi_iter, _Ch_type, _Rx_traits>&
612 regex_token_iterator<_Bi_iter, _Ch_type, _Rx_traits>::
613 operator=(const regex_token_iterator& __rhs)
615 _M_position = __rhs._M_position;
616 _M_subs = __rhs._M_subs;
618 _M_result = __rhs._M_result;
619 _M_suffix = __rhs._M_suffix;
620 _M_has_m1 = __rhs._M_has_m1;
621 if (__rhs._M_result == &__rhs._M_suffix)
622 _M_result = &_M_suffix;
626 template<typename _Bi_iter,
630 regex_token_iterator<_Bi_iter, _Ch_type, _Rx_traits>::
631 operator==(const regex_token_iterator& __rhs) const
633 if (_M_end_of_seq() && __rhs._M_end_of_seq())
635 if (_M_suffix.matched && __rhs._M_suffix.matched
636 && _M_suffix == __rhs._M_suffix)
638 if (_M_end_of_seq() || _M_suffix.matched
639 || __rhs._M_end_of_seq() || __rhs._M_suffix.matched)
641 return _M_position == __rhs._M_position
642 && _M_n == __rhs._M_n
643 && _M_subs == __rhs._M_subs;
646 template<typename _Bi_iter,
649 regex_token_iterator<_Bi_iter, _Ch_type, _Rx_traits>&
650 regex_token_iterator<_Bi_iter, _Ch_type, _Rx_traits>::
653 _Position __prev = _M_position;
654 if (_M_suffix.matched)
655 *this = regex_token_iterator();
656 else if (_M_n + 1 < _M_subs.size())
659 _M_result = &_M_current_match();
665 if (_M_position != _Position())
666 _M_result = &_M_current_match();
667 else if (_M_has_m1 && __prev->suffix().length() != 0)
669 _M_suffix.matched = true;
670 _M_suffix.first = __prev->suffix().first;
671 _M_suffix.second = __prev->suffix().second;
672 _M_result = &_M_suffix;
675 *this = regex_token_iterator();
680 template<typename _Bi_iter,
684 regex_token_iterator<_Bi_iter, _Ch_type, _Rx_traits>::
685 _M_init(_Bi_iter __a, _Bi_iter __b)
688 for (auto __it : _M_subs)
694 if (_M_position != _Position())
695 _M_result = &_M_current_match();
698 _M_suffix.matched = true;
699 _M_suffix.first = __a;
700 _M_suffix.second = __b;
701 _M_result = &_M_suffix;
707 _GLIBCXX_END_NAMESPACE_VERSION