1 // class template regex -*- C++ -*-
3 // Copyright (C) 2013-2015 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 // A non-standard switch to let the user pick the matching algorithm.
32 // If _GLIBCXX_REGEX_USE_THOMPSON_NFA is defined, the thompson NFA
33 // algorithm will be used. This algorithm is not enabled by default,
34 // and cannot be used if the regex contains back-references, but has better
35 // (polynomial instead of exponential) worst case performance.
36 // See __regex_algo_impl below.
38 namespace std _GLIBCXX_VISIBILITY(default)
42 _GLIBCXX_BEGIN_NAMESPACE_VERSION
44 // Result of merging regex_match and regex_search.
46 // __policy now can be _S_auto (auto dispatch) and _S_alternate (use
47 // the other one if possible, for test purpose).
49 // That __match_mode is true means regex_match, else regex_search.
50 template<typename _BiIter, typename _Alloc,
51 typename _CharT, typename _TraitsT,
52 _RegexExecutorPolicy __policy,
55 __regex_algo_impl(_BiIter __s,
57 match_results<_BiIter, _Alloc>& __m,
58 const basic_regex<_CharT, _TraitsT>& __re,
59 regex_constants::match_flag_type __flags)
61 if (__re._M_automaton == nullptr)
64 typename match_results<_BiIter, _Alloc>::_Base_type& __res = __m;
66 __m._M_resize(__re._M_automaton->_M_sub_count());
67 for (auto& __it : __res)
70 // __policy is used by testsuites so that they can use Thompson NFA
71 // without defining a macro. Users should define
72 // _GLIBCXX_REGEX_USE_THOMPSON_NFA if they need to use this approach.
74 if (!__re._M_automaton->_M_has_backref
75 && !(__re._M_flags & regex_constants::ECMAScript)
76 #ifndef _GLIBCXX_REGEX_USE_THOMPSON_NFA
77 && __policy == _RegexExecutorPolicy::_S_alternate
81 _Executor<_BiIter, _Alloc, _TraitsT, false>
82 __executor(__s, __e, __m, __re, __flags);
84 __ret = __executor._M_match();
86 __ret = __executor._M_search();
90 _Executor<_BiIter, _Alloc, _TraitsT, true>
91 __executor(__s, __e, __m, __re, __flags);
93 __ret = __executor._M_match();
95 __ret = __executor._M_search();
99 for (auto& __it : __res)
101 __it.first = __it.second = __e;
102 auto& __pre = __m._M_prefix();
103 auto& __suf = __m._M_suffix();
106 __pre.matched = false;
109 __suf.matched = false;
116 __pre.second = __res[0].first;
117 __pre.matched = (__pre.first != __pre.second);
118 __suf.first = __res[0].second;
120 __suf.matched = (__suf.first != __suf.second);
126 for (auto& __it : __res)
128 __it.matched = false;
129 __it.first = __it.second = __e;
135 _GLIBCXX_END_NAMESPACE_VERSION
138 _GLIBCXX_BEGIN_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",
282 for (; __first != __last; ++__first)
283 __s += __fctyp.narrow(*__first, 0);
285 for (const auto& __it : __collatenames)
287 return string_type(1, __fctyp.widen(
288 static_cast<char>(&__it - __collatenames)));
290 // TODO Add digraph support:
291 // http://boost.sourceforge.net/libs/regex/doc/collating_names.html
293 return string_type();
296 template<typename _Ch_type>
297 template<typename _Fwd_iter>
298 typename regex_traits<_Ch_type>::char_class_type
299 regex_traits<_Ch_type>::
300 lookup_classname(_Fwd_iter __first, _Fwd_iter __last, bool __icase) const
302 typedef std::ctype<char_type> __ctype_type;
303 const __ctype_type& __fctyp(use_facet<__ctype_type>(_M_locale));
305 // Mappings from class name to class mask.
306 static const pair<const char*, char_class_type> __classnames[] =
308 {"d", ctype_base::digit},
309 {"w", {ctype_base::alnum, _RegexMask::_S_under}},
310 {"s", ctype_base::space},
311 {"alnum", ctype_base::alnum},
312 {"alpha", ctype_base::alpha},
313 {"blank", ctype_base::blank},
314 {"cntrl", ctype_base::cntrl},
315 {"digit", ctype_base::digit},
316 {"graph", ctype_base::graph},
317 {"lower", ctype_base::lower},
318 {"print", ctype_base::print},
319 {"punct", ctype_base::punct},
320 {"space", ctype_base::space},
321 {"upper", ctype_base::upper},
322 {"xdigit", ctype_base::xdigit},
326 for (; __first != __last; ++__first)
327 __s += __fctyp.narrow(__fctyp.tolower(*__first), 0);
329 for (const auto& __it : __classnames)
330 if (__s == __it.first)
334 & (ctype_base::lower | ctype_base::upper)) != 0))
335 return ctype_base::alpha;
341 template<typename _Ch_type>
343 regex_traits<_Ch_type>::
344 isctype(_Ch_type __c, char_class_type __f) const
346 typedef std::ctype<char_type> __ctype_type;
347 const __ctype_type& __fctyp(use_facet<__ctype_type>(_M_locale));
349 return __fctyp.is(__f._M_base, __c)
351 || ((__f._M_extended & _RegexMask::_S_under)
352 && __c == __fctyp.widen('_'));
355 template<typename _Ch_type>
357 regex_traits<_Ch_type>::
358 value(_Ch_type __ch, int __radix) const
360 std::basic_istringstream<char_type> __is(string_type(1, __ch));
364 else if (__radix == 16)
367 return __is.fail() ? -1 : __v;
370 template<typename _Bi_iter, typename _Alloc>
371 template<typename _Out_iter>
372 _Out_iter match_results<_Bi_iter, _Alloc>::
373 format(_Out_iter __out,
374 const match_results<_Bi_iter, _Alloc>::char_type* __fmt_first,
375 const match_results<_Bi_iter, _Alloc>::char_type* __fmt_last,
376 match_flag_type __flags) const
378 _GLIBCXX_DEBUG_ASSERT( ready() );
379 regex_traits<char_type> __traits;
380 typedef std::ctype<char_type> __ctype_type;
382 __fctyp(use_facet<__ctype_type>(__traits.getloc()));
384 auto __output = [&](size_t __idx)
386 auto& __sub = (*this)[__idx];
388 __out = std::copy(__sub.first, __sub.second, __out);
391 if (__flags & regex_constants::format_sed)
393 for (; __fmt_first != __fmt_last;)
394 if (*__fmt_first == '&')
399 else if (*__fmt_first == '\\')
401 if (++__fmt_first != __fmt_last
402 && __fctyp.is(__ctype_type::digit, *__fmt_first))
403 __output(__traits.value(*__fmt_first++, 10));
408 *__out++ = *__fmt_first++;
414 auto __next = std::find(__fmt_first, __fmt_last, '$');
415 if (__next == __fmt_last)
418 __out = std::copy(__fmt_first, __next, __out);
420 auto __eat = [&](char __ch) -> bool
430 if (++__next == __fmt_last)
438 auto& __sub = _M_prefix();
440 __out = std::copy(__sub.first, __sub.second, __out);
442 else if (__eat('\''))
444 auto& __sub = _M_suffix();
446 __out = std::copy(__sub.first, __sub.second, __out);
448 else if (__fctyp.is(__ctype_type::digit, *__next))
450 long __num = __traits.value(*__next, 10);
451 if (++__next != __fmt_last
452 && __fctyp.is(__ctype_type::digit, *__next))
455 __num += __traits.value(*__next++, 10);
457 if (0 <= __num && __num < this->size())
462 __fmt_first = __next;
464 __out = std::copy(__fmt_first, __fmt_last, __out);
469 template<typename _Out_iter, typename _Bi_iter,
470 typename _Rx_traits, typename _Ch_type>
472 regex_replace(_Out_iter __out, _Bi_iter __first, _Bi_iter __last,
473 const basic_regex<_Ch_type, _Rx_traits>& __e,
474 const _Ch_type* __fmt,
475 regex_constants::match_flag_type __flags)
477 typedef regex_iterator<_Bi_iter, _Ch_type, _Rx_traits> _IterT;
478 _IterT __i(__first, __last, __e, __flags);
482 if (!(__flags & regex_constants::format_no_copy))
483 __out = std::copy(__first, __last, __out);
487 sub_match<_Bi_iter> __last;
488 auto __len = char_traits<_Ch_type>::length(__fmt);
489 for (; __i != __end; ++__i)
491 if (!(__flags & regex_constants::format_no_copy))
492 __out = std::copy(__i->prefix().first, __i->prefix().second,
494 __out = __i->format(__out, __fmt, __fmt + __len, __flags);
495 __last = __i->suffix();
496 if (__flags & regex_constants::format_first_only)
499 if (!(__flags & regex_constants::format_no_copy))
500 __out = std::copy(__last.first, __last.second, __out);
505 template<typename _Bi_iter,
509 regex_iterator<_Bi_iter, _Ch_type, _Rx_traits>::
510 operator==(const regex_iterator& __rhs) const
512 return (_M_match.empty() && __rhs._M_match.empty())
513 || (_M_begin == __rhs._M_begin
514 && _M_end == __rhs._M_end
515 && _M_pregex == __rhs._M_pregex
516 && _M_flags == __rhs._M_flags
517 && _M_match[0] == __rhs._M_match[0]);
520 template<typename _Bi_iter,
523 regex_iterator<_Bi_iter, _Ch_type, _Rx_traits>&
524 regex_iterator<_Bi_iter, _Ch_type, _Rx_traits>::
527 // In all cases in which the call to regex_search returns true,
528 // match.prefix().first shall be equal to the previous value of
529 // match[0].second, and for each index i in the half-open range
530 // [0, match.size()) for which match[i].matched is true,
531 // match[i].position() shall return distance(begin, match[i].first).
533 if (_M_match[0].matched)
535 auto __start = _M_match[0].second;
536 auto __prefix_first = _M_match[0].second;
537 if (_M_match[0].first == _M_match[0].second)
539 if (__start == _M_end)
541 _M_match = value_type();
546 if (regex_search(__start, _M_end, _M_match, *_M_pregex,
548 | regex_constants::match_not_null
549 | regex_constants::match_continuous))
551 _GLIBCXX_DEBUG_ASSERT(_M_match[0].matched);
552 auto& __prefix = _M_match._M_prefix();
553 __prefix.first = __prefix_first;
554 __prefix.matched = __prefix.first != __prefix.second;
556 _M_match._M_begin = _M_begin;
563 _M_flags |= regex_constants::match_prev_avail;
564 if (regex_search(__start, _M_end, _M_match, *_M_pregex, _M_flags))
566 _GLIBCXX_DEBUG_ASSERT(_M_match[0].matched);
567 auto& __prefix = _M_match._M_prefix();
568 __prefix.first = __prefix_first;
569 __prefix.matched = __prefix.first != __prefix.second;
571 _M_match._M_begin = _M_begin;
574 _M_match = value_type();
579 template<typename _Bi_iter,
582 regex_token_iterator<_Bi_iter, _Ch_type, _Rx_traits>&
583 regex_token_iterator<_Bi_iter, _Ch_type, _Rx_traits>::
584 operator=(const regex_token_iterator& __rhs)
586 _M_position = __rhs._M_position;
587 _M_subs = __rhs._M_subs;
589 _M_suffix = __rhs._M_suffix;
590 _M_has_m1 = __rhs._M_has_m1;
591 _M_normalize_result();
595 template<typename _Bi_iter,
599 regex_token_iterator<_Bi_iter, _Ch_type, _Rx_traits>::
600 operator==(const regex_token_iterator& __rhs) const
602 if (_M_end_of_seq() && __rhs._M_end_of_seq())
604 if (_M_suffix.matched && __rhs._M_suffix.matched
605 && _M_suffix == __rhs._M_suffix)
607 if (_M_end_of_seq() || _M_suffix.matched
608 || __rhs._M_end_of_seq() || __rhs._M_suffix.matched)
610 return _M_position == __rhs._M_position
611 && _M_n == __rhs._M_n
612 && _M_subs == __rhs._M_subs;
615 template<typename _Bi_iter,
618 regex_token_iterator<_Bi_iter, _Ch_type, _Rx_traits>&
619 regex_token_iterator<_Bi_iter, _Ch_type, _Rx_traits>::
622 _Position __prev = _M_position;
623 if (_M_suffix.matched)
624 *this = regex_token_iterator();
625 else if (_M_n + 1 < _M_subs.size())
628 _M_result = &_M_current_match();
634 if (_M_position != _Position())
635 _M_result = &_M_current_match();
636 else if (_M_has_m1 && __prev->suffix().length() != 0)
638 _M_suffix.matched = true;
639 _M_suffix.first = __prev->suffix().first;
640 _M_suffix.second = __prev->suffix().second;
641 _M_result = &_M_suffix;
644 *this = regex_token_iterator();
649 template<typename _Bi_iter,
653 regex_token_iterator<_Bi_iter, _Ch_type, _Rx_traits>::
654 _M_init(_Bi_iter __a, _Bi_iter __b)
657 for (auto __it : _M_subs)
663 if (_M_position != _Position())
664 _M_result = &_M_current_match();
667 _M_suffix.matched = true;
668 _M_suffix.first = __a;
669 _M_suffix.second = __b;
670 _M_result = &_M_suffix;
676 _GLIBCXX_END_NAMESPACE_VERSION