// class template regex -*- C++ -*-
-// Copyright (C) 2013 Free Software Foundation, Inc.
+// Copyright (C) 2013-2017 Free Software Foundation, Inc.
//
// This file is part of the GNU ISO C++ Library. This library is free
// software; you can redistribute it and/or modify it under the
* Do not attempt to use it directly. @headername{regex}
*/
+// This macro defines the maximal state number a NFA can have.
+#ifndef _GLIBCXX_REGEX_STATE_LIMIT
+#define _GLIBCXX_REGEX_STATE_LIMIT 100000
+#endif
+
namespace std _GLIBCXX_VISIBILITY(default)
{
namespace __detail
*/
typedef long _StateIdT;
- typedef std::set<_StateIdT> _StateSet;
static const _StateIdT _S_invalid_state_id = -1;
template<typename _CharT>
{
_S_opcode_unknown,
_S_opcode_alternative,
+ _S_opcode_repeat,
_S_opcode_backref,
_S_opcode_line_begin_assertion,
_S_opcode_line_end_assertion,
struct _State_base
{
+ protected:
_Opcode _M_opcode; // type of outgoing transition
+
+ public:
_StateIdT _M_next; // outgoing transition
union // Since they are mutually exclusive.
{
size_t _M_backref_index; // for _S_opcode_backref
struct
{
- // for _S_opcode_alternative.
- _StateIdT _M_quant_index;
- // for _S_opcode_alternative or _S_opcode_subexpr_lookahead
+ // for _S_opcode_alternative, _S_opcode_repeat and
+ // _S_opcode_subexpr_lookahead
_StateIdT _M_alt;
// for _S_opcode_word_boundary or _S_opcode_subexpr_lookahead or
// quantifiers (ungreedy if set true)
bool _M_neg;
};
+ // For _S_opcode_match
+ __gnu_cxx::__aligned_membuf<_Matcher<char>> _M_matcher_storage;
};
+ protected:
explicit _State_base(_Opcode __opcode)
: _M_opcode(__opcode), _M_next(_S_invalid_state_id)
{ }
- protected:
- ~_State_base() = default;
-
public:
+ bool
+ _M_has_alt()
+ {
+ return _M_opcode == _S_opcode_alternative
+ || _M_opcode == _S_opcode_repeat
+ || _M_opcode == _S_opcode_subexpr_lookahead;
+ }
+
#ifdef _GLIBCXX_DEBUG
std::ostream&
_M_print(std::ostream& ostr) const;
#endif
};
- template<typename _CharT, typename _TraitsT>
+ template<typename _Char_type>
struct _State : _State_base
{
- typedef _Matcher<_CharT> _MatcherT;
+ typedef _Matcher<_Char_type> _MatcherT;
+ static_assert(sizeof(_MatcherT) == sizeof(_Matcher<char>),
+ "std::function<bool(T)> has the same size as "
+ "std::function<bool(char)>");
+ static_assert(alignof(_MatcherT) == alignof(_Matcher<char>),
+ "std::function<bool(T)> has the same alignment as "
+ "std::function<bool(char)>");
+
+ explicit
+ _State(_Opcode __opcode) : _State_base(__opcode)
+ {
+ if (_M_opcode() == _S_opcode_match)
+ new (this->_M_matcher_storage._M_addr()) _MatcherT();
+ }
- _MatcherT _M_matches; // for _S_opcode_match
+ _State(const _State& __rhs) : _State_base(__rhs)
+ {
+ if (__rhs._M_opcode() == _S_opcode_match)
+ new (this->_M_matcher_storage._M_addr())
+ _MatcherT(__rhs._M_get_matcher());
+ }
- explicit _State(_Opcode __opcode) : _State_base(__opcode) { }
+ _State(_State&& __rhs) : _State_base(__rhs)
+ {
+ if (__rhs._M_opcode() == _S_opcode_match)
+ new (this->_M_matcher_storage._M_addr())
+ _MatcherT(std::move(__rhs._M_get_matcher()));
+ }
+
+ _State&
+ operator=(const _State&) = delete;
+
+ ~_State()
+ {
+ if (_M_opcode() == _S_opcode_match)
+ _M_get_matcher().~_MatcherT();
+ }
+
+ // Since correct ctor and dtor rely on _M_opcode, it's better not to
+ // change it over time.
+ _Opcode
+ _M_opcode() const
+ { return _State_base::_M_opcode; }
+
+ bool
+ _M_matches(_Char_type __char) const
+ { return _M_get_matcher()(__char); }
+
+ _MatcherT&
+ _M_get_matcher()
+ { return *static_cast<_MatcherT*>(this->_M_matcher_storage._M_addr()); }
+
+ const _MatcherT&
+ _M_get_matcher() const
+ {
+ return *static_cast<const _MatcherT*>(
+ this->_M_matcher_storage._M_addr());
+ }
};
struct _NFA_base
explicit
_NFA_base(_FlagT __f)
: _M_flags(__f), _M_start_state(0), _M_subexpr_count(0),
- _M_quant_count(0), _M_has_backref(false)
+ _M_has_backref(false)
{ }
_NFA_base(_NFA_base&&) = default;
_M_start() const
{ return _M_start_state; }
- const _StateSet&
- _M_final_states() const
- { return _M_accepting_states; }
-
_SizeT
_M_sub_count() const
{ return _M_subexpr_count; }
std::vector<size_t> _M_paren_stack;
- _StateSet _M_accepting_states;
_FlagT _M_flags;
_StateIdT _M_start_state;
_SizeT _M_subexpr_count;
- _SizeT _M_quant_count;
bool _M_has_backref;
};
- template<typename _CharT, typename _TraitsT>
+ template<typename _TraitsT>
struct _NFA
- : _NFA_base, std::vector<_State<_CharT, _TraitsT>>
+ : _NFA_base, std::vector<_State<typename _TraitsT::char_type>>
{
- typedef _State<_CharT, _TraitsT> _StateT;
- typedef _Matcher<_CharT> _MatcherT;
+ typedef typename _TraitsT::char_type _Char_type;
+ typedef _State<_Char_type> _StateT;
+ typedef _Matcher<_Char_type> _MatcherT;
- using _NFA_base::_NFA_base;
+ _NFA(const typename _TraitsT::locale_type& __loc, _FlagT __flags)
+ : _NFA_base(__flags)
+ { _M_traits.imbue(__loc); }
// for performance reasons _NFA objects should only be moved not copied
_NFA(const _NFA&) = delete;
_M_insert_accept()
{
auto __ret = _M_insert_state(_StateT(_S_opcode_accept));
- this->_M_accepting_states.insert(__ret);
return __ret;
}
_StateIdT
- _M_insert_alt(_StateIdT __next, _StateIdT __alt, bool __neg)
+ _M_insert_alt(_StateIdT __next, _StateIdT __alt,
+ bool __neg __attribute__((__unused__)))
{
_StateT __tmp(_S_opcode_alternative);
// It labels every quantifier to make greedy comparison easier in BFS
// approach.
- __tmp._M_quant_index = this->_M_quant_count++;
+ __tmp._M_next = __next;
+ __tmp._M_alt = __alt;
+ return _M_insert_state(std::move(__tmp));
+ }
+
+ _StateIdT
+ _M_insert_repeat(_StateIdT __next, _StateIdT __alt, bool __neg)
+ {
+ _StateT __tmp(_S_opcode_repeat);
+ // It labels every quantifier to make greedy comparison easier in BFS
+ // approach.
__tmp._M_next = __next;
__tmp._M_alt = __alt;
__tmp._M_neg = __neg;
_M_insert_matcher(_MatcherT __m)
{
_StateT __tmp(_S_opcode_match);
- __tmp._M_matches = std::move(__m);
+ __tmp._M_get_matcher() = std::move(__m);
return _M_insert_state(std::move(__tmp));
}
_M_insert_state(_StateT __s)
{
this->push_back(std::move(__s));
+ if (this->size() > _GLIBCXX_REGEX_STATE_LIMIT)
+ __throw_regex_error(
+ regex_constants::error_space,
+ "Number of NFA states exceeds limit. Please use shorter regex "
+ "string, or use smaller brace expression, or make "
+ "_GLIBCXX_REGEX_STATE_LIMIT larger.");
return this->size()-1;
}
std::ostream&
_M_dot(std::ostream& __ostr) const;
#endif
+ public:
+ _TraitsT _M_traits;
};
/// Describes a sequence of one or more %_State, its current start
/// and end(s). This structure contains fragments of an NFA during
/// construction.
- template<typename _CharT, typename _TraitsT>
+ template<typename _TraitsT>
class _StateSeq
{
public:
- typedef _NFA<_CharT, _TraitsT> _RegexT;
+ typedef _NFA<_TraitsT> _RegexT;
public:
_StateSeq(_RegexT& __nfa, _StateIdT __s)