1 // class template regex -*- C++ -*-
3 // Copyright (C) 2013-2018 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 namespace std _GLIBCXX_VISIBILITY(default)
33 _GLIBCXX_BEGIN_NAMESPACE_VERSION
37 // Result of merging regex_match and regex_search.
39 // __policy now can be _S_auto (auto dispatch) and _S_alternate (use
40 // the other one if possible, for test purpose).
42 // That __match_mode is true means regex_match, else regex_search.
43 template<typename _BiIter, typename _Alloc,
44 typename _CharT, typename _TraitsT,
45 _RegexExecutorPolicy __policy,
48 __regex_algo_impl(_BiIter __s,
50 match_results<_BiIter, _Alloc>& __m,
51 const basic_regex<_CharT, _TraitsT>& __re,
52 regex_constants::match_flag_type __flags)
54 if (__re._M_automaton == nullptr)
57 typename match_results<_BiIter, _Alloc>::_Base_type& __res = __m;
59 __m._M_resize(__re._M_automaton->_M_sub_count());
60 for (auto& __it : __res)
64 if ((__re.flags() & regex_constants::__polynomial)
65 || (__policy == _RegexExecutorPolicy::_S_alternate
66 && !__re._M_automaton->_M_has_backref))
68 _Executor<_BiIter, _Alloc, _TraitsT, false>
69 __executor(__s, __e, __m, __re, __flags);
71 __ret = __executor._M_match();
73 __ret = __executor._M_search();
77 _Executor<_BiIter, _Alloc, _TraitsT, true>
78 __executor(__s, __e, __m, __re, __flags);
80 __ret = __executor._M_match();
82 __ret = __executor._M_search();
86 for (auto& __it : __res)
88 __it.first = __it.second = __e;
89 auto& __pre = __m._M_prefix();
90 auto& __suf = __m._M_suffix();
93 __pre.matched = false;
96 __suf.matched = false;
103 __pre.second = __res[0].first;
104 __pre.matched = (__pre.first != __pre.second);
105 __suf.first = __res[0].second;
107 __suf.matched = (__suf.first != __suf.second);
113 for (auto& __it : __res)
115 __it.matched = false;
116 __it.first = __it.second = __e;
123 template<typename _Ch_type>
124 template<typename _Fwd_iter>
125 typename regex_traits<_Ch_type>::string_type
126 regex_traits<_Ch_type>::
127 lookup_collatename(_Fwd_iter __first, _Fwd_iter __last) const
129 typedef std::ctype<char_type> __ctype_type;
130 const __ctype_type& __fctyp(use_facet<__ctype_type>(_M_locale));
132 static const char* __collatenames[] =
225 "left-square-bracket",
227 "right-square-bracket",
257 "left-curly-bracket",
259 "right-curly-bracket",
265 for (; __first != __last; ++__first)
266 __s += __fctyp.narrow(*__first, 0);
268 for (const auto& __it : __collatenames)
270 return string_type(1, __fctyp.widen(
271 static_cast<char>(&__it - __collatenames)));
273 // TODO Add digraph support:
274 // http://boost.sourceforge.net/libs/regex/doc/collating_names.html
276 return string_type();
279 template<typename _Ch_type>
280 template<typename _Fwd_iter>
281 typename regex_traits<_Ch_type>::char_class_type
282 regex_traits<_Ch_type>::
283 lookup_classname(_Fwd_iter __first, _Fwd_iter __last, bool __icase) const
285 typedef std::ctype<char_type> __ctype_type;
286 const __ctype_type& __fctyp(use_facet<__ctype_type>(_M_locale));
288 // Mappings from class name to class mask.
289 static const pair<const char*, char_class_type> __classnames[] =
291 {"d", ctype_base::digit},
292 {"w", {ctype_base::alnum, _RegexMask::_S_under}},
293 {"s", ctype_base::space},
294 {"alnum", ctype_base::alnum},
295 {"alpha", ctype_base::alpha},
296 {"blank", ctype_base::blank},
297 {"cntrl", ctype_base::cntrl},
298 {"digit", ctype_base::digit},
299 {"graph", ctype_base::graph},
300 {"lower", ctype_base::lower},
301 {"print", ctype_base::print},
302 {"punct", ctype_base::punct},
303 {"space", ctype_base::space},
304 {"upper", ctype_base::upper},
305 {"xdigit", ctype_base::xdigit},
309 for (; __first != __last; ++__first)
310 __s += __fctyp.narrow(__fctyp.tolower(*__first), 0);
312 for (const auto& __it : __classnames)
313 if (__s == __it.first)
317 & (ctype_base::lower | ctype_base::upper)) != 0))
318 return ctype_base::alpha;
324 template<typename _Ch_type>
326 regex_traits<_Ch_type>::
327 isctype(_Ch_type __c, char_class_type __f) const
329 typedef std::ctype<char_type> __ctype_type;
330 const __ctype_type& __fctyp(use_facet<__ctype_type>(_M_locale));
332 return __fctyp.is(__f._M_base, __c)
334 || ((__f._M_extended & _RegexMask::_S_under)
335 && __c == __fctyp.widen('_'));
338 template<typename _Ch_type>
340 regex_traits<_Ch_type>::
341 value(_Ch_type __ch, int __radix) const
343 std::basic_istringstream<char_type> __is(string_type(1, __ch));
347 else if (__radix == 16)
350 return __is.fail() ? -1 : __v;
353 template<typename _Bi_iter, typename _Alloc>
354 template<typename _Out_iter>
355 _Out_iter match_results<_Bi_iter, _Alloc>::
356 format(_Out_iter __out,
357 const match_results<_Bi_iter, _Alloc>::char_type* __fmt_first,
358 const match_results<_Bi_iter, _Alloc>::char_type* __fmt_last,
359 match_flag_type __flags) const
361 __glibcxx_assert( ready() );
362 regex_traits<char_type> __traits;
363 typedef std::ctype<char_type> __ctype_type;
365 __fctyp(use_facet<__ctype_type>(__traits.getloc()));
367 auto __output = [&](size_t __idx)
369 auto& __sub = (*this)[__idx];
371 __out = std::copy(__sub.first, __sub.second, __out);
374 if (__flags & regex_constants::format_sed)
376 bool __escaping = false;
377 for (; __fmt_first != __fmt_last; __fmt_first++)
382 if (__fctyp.is(__ctype_type::digit, *__fmt_first))
383 __output(__traits.value(*__fmt_first, 10));
385 *__out++ = *__fmt_first;
388 if (*__fmt_first == '\\')
393 if (*__fmt_first == '&')
398 *__out++ = *__fmt_first;
407 auto __next = std::find(__fmt_first, __fmt_last, '$');
408 if (__next == __fmt_last)
411 __out = std::copy(__fmt_first, __next, __out);
413 auto __eat = [&](char __ch) -> bool
423 if (++__next == __fmt_last)
431 auto& __sub = _M_prefix();
433 __out = std::copy(__sub.first, __sub.second, __out);
435 else if (__eat('\''))
437 auto& __sub = _M_suffix();
439 __out = std::copy(__sub.first, __sub.second, __out);
441 else if (__fctyp.is(__ctype_type::digit, *__next))
443 long __num = __traits.value(*__next, 10);
444 if (++__next != __fmt_last
445 && __fctyp.is(__ctype_type::digit, *__next))
448 __num += __traits.value(*__next++, 10);
450 if (0 <= __num && __num < this->size())
455 __fmt_first = __next;
457 __out = std::copy(__fmt_first, __fmt_last, __out);
462 template<typename _Out_iter, typename _Bi_iter,
463 typename _Rx_traits, typename _Ch_type>
465 regex_replace(_Out_iter __out, _Bi_iter __first, _Bi_iter __last,
466 const basic_regex<_Ch_type, _Rx_traits>& __e,
467 const _Ch_type* __fmt,
468 regex_constants::match_flag_type __flags)
470 typedef regex_iterator<_Bi_iter, _Ch_type, _Rx_traits> _IterT;
471 _IterT __i(__first, __last, __e, __flags);
475 if (!(__flags & regex_constants::format_no_copy))
476 __out = std::copy(__first, __last, __out);
480 sub_match<_Bi_iter> __last;
481 auto __len = char_traits<_Ch_type>::length(__fmt);
482 for (; __i != __end; ++__i)
484 if (!(__flags & regex_constants::format_no_copy))
485 __out = std::copy(__i->prefix().first, __i->prefix().second,
487 __out = __i->format(__out, __fmt, __fmt + __len, __flags);
488 __last = __i->suffix();
489 if (__flags & regex_constants::format_first_only)
492 if (!(__flags & regex_constants::format_no_copy))
493 __out = std::copy(__last.first, __last.second, __out);
498 template<typename _Bi_iter,
502 regex_iterator<_Bi_iter, _Ch_type, _Rx_traits>::
503 operator==(const regex_iterator& __rhs) const
505 if (_M_pregex == nullptr && __rhs._M_pregex == nullptr)
507 return _M_pregex == __rhs._M_pregex
508 && _M_begin == __rhs._M_begin
509 && _M_end == __rhs._M_end
510 && _M_flags == __rhs._M_flags
511 && _M_match[0] == __rhs._M_match[0];
514 template<typename _Bi_iter,
517 regex_iterator<_Bi_iter, _Ch_type, _Rx_traits>&
518 regex_iterator<_Bi_iter, _Ch_type, _Rx_traits>::
521 // In all cases in which the call to regex_search returns true,
522 // match.prefix().first shall be equal to the previous value of
523 // match[0].second, and for each index i in the half-open range
524 // [0, match.size()) for which match[i].matched is true,
525 // match[i].position() shall return distance(begin, match[i].first).
527 if (_M_match[0].matched)
529 auto __start = _M_match[0].second;
530 auto __prefix_first = _M_match[0].second;
531 if (_M_match[0].first == _M_match[0].second)
533 if (__start == _M_end)
540 if (regex_search(__start, _M_end, _M_match, *_M_pregex,
542 | regex_constants::match_not_null
543 | regex_constants::match_continuous))
545 __glibcxx_assert(_M_match[0].matched);
546 auto& __prefix = _M_match._M_prefix();
547 __prefix.first = __prefix_first;
548 __prefix.matched = __prefix.first != __prefix.second;
550 _M_match._M_begin = _M_begin;
557 _M_flags |= regex_constants::match_prev_avail;
558 if (regex_search(__start, _M_end, _M_match, *_M_pregex, _M_flags))
560 __glibcxx_assert(_M_match[0].matched);
561 auto& __prefix = _M_match._M_prefix();
562 __prefix.first = __prefix_first;
563 __prefix.matched = __prefix.first != __prefix.second;
565 _M_match._M_begin = _M_begin;
573 template<typename _Bi_iter,
576 regex_token_iterator<_Bi_iter, _Ch_type, _Rx_traits>&
577 regex_token_iterator<_Bi_iter, _Ch_type, _Rx_traits>::
578 operator=(const regex_token_iterator& __rhs)
580 _M_position = __rhs._M_position;
581 _M_subs = __rhs._M_subs;
583 _M_suffix = __rhs._M_suffix;
584 _M_has_m1 = __rhs._M_has_m1;
585 _M_normalize_result();
589 template<typename _Bi_iter,
593 regex_token_iterator<_Bi_iter, _Ch_type, _Rx_traits>::
594 operator==(const regex_token_iterator& __rhs) const
596 if (_M_end_of_seq() && __rhs._M_end_of_seq())
598 if (_M_suffix.matched && __rhs._M_suffix.matched
599 && _M_suffix == __rhs._M_suffix)
601 if (_M_end_of_seq() || _M_suffix.matched
602 || __rhs._M_end_of_seq() || __rhs._M_suffix.matched)
604 return _M_position == __rhs._M_position
605 && _M_n == __rhs._M_n
606 && _M_subs == __rhs._M_subs;
609 template<typename _Bi_iter,
612 regex_token_iterator<_Bi_iter, _Ch_type, _Rx_traits>&
613 regex_token_iterator<_Bi_iter, _Ch_type, _Rx_traits>::
616 _Position __prev = _M_position;
617 if (_M_suffix.matched)
618 *this = regex_token_iterator();
619 else if (_M_n + 1 < _M_subs.size())
622 _M_result = &_M_current_match();
628 if (_M_position != _Position())
629 _M_result = &_M_current_match();
630 else if (_M_has_m1 && __prev->suffix().length() != 0)
632 _M_suffix.matched = true;
633 _M_suffix.first = __prev->suffix().first;
634 _M_suffix.second = __prev->suffix().second;
635 _M_result = &_M_suffix;
638 *this = regex_token_iterator();
643 template<typename _Bi_iter,
647 regex_token_iterator<_Bi_iter, _Ch_type, _Rx_traits>::
648 _M_init(_Bi_iter __a, _Bi_iter __b)
651 for (auto __it : _M_subs)
657 if (_M_position != _Position())
658 _M_result = &_M_current_match();
661 _M_suffix.matched = true;
662 _M_suffix.first = __a;
663 _M_suffix.second = __b;
664 _M_result = &_M_suffix;
670 _GLIBCXX_END_NAMESPACE_VERSION