From 6130bce7397918cd5563e4b404c4f86c9dfd94a9 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Julian=20M=C3=BCller?= <julian.mueller@uni-konstanz.de>
Date: Sun, 18 May 2025 13:16:18 +0200
Subject: [PATCH 01/21] rename _Parser to _Parser2

---
 stl/inc/regex | 71 ++++++++++++++++++++++++++-------------------------
 1 file changed, 36 insertions(+), 35 deletions(-)

diff --git a/stl/inc/regex b/stl/inc/regex
index 46dab2b06d0..1a319380297 100644
--- a/stl/inc/regex
+++ b/stl/inc/regex
@@ -1574,7 +1574,7 @@ enum class _Rx_char_class_kind : int { // must be aligned with corresponding _No
 };
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-class _Builder { // provides operations used by _Parser to build the nfa
+class _Builder { // provides operations used by _Parser2 to build the nfa
 public:
     _Builder(const _RxTraits& _Tr, regex_constants::syntax_option_type);
     void _Setlong();
@@ -1785,11 +1785,11 @@ enum _Prs_ret { // indicate class element type
 };
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-class _Parser { // parse a regular expression
+class _Parser2 { // parse a regular expression
 public:
     using char_class_type = typename _RxTraits::char_class_type;
 
-    _Parser(const _RxTraits& _Tr, _FwdIt _Pfirst, _FwdIt _Plast, regex_constants::syntax_option_type _Fx);
+    _Parser2(const _RxTraits& _Tr, _FwdIt _Pfirst, _FwdIt _Plast, regex_constants::syntax_option_type _Fx);
     _Root_node* _Compile();
 
     unsigned int _Mark_count() const noexcept {
@@ -2121,7 +2121,7 @@ private:
             _Visualization.assign(_First, _Last);
 #endif // _ENHANCED_REGEX_VISUALIZER
 
-            _Parser<_InIt, _Elem, _RxTraits> _Prs(_Traits, _First, _Last, _Flags);
+            _Parser2<_InIt, _Elem, _RxTraits> _Prs(_Traits, _First, _Last, _Flags);
             _Root_node* _Rx = _Prs._Compile();
             _Reset(_Rx);
         } else {
@@ -4160,19 +4160,19 @@ _BidIt _Matcher2<_BidIt, _Elem, _RxTraits, _It, _Alloc>::_Skip(_BidIt _First_arg
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-[[noreturn]] void _Parser<_FwdIt, _Elem, _RxTraits>::_Error(regex_constants::error_type _Code) { // handle error
+[[noreturn]] void _Parser2<_FwdIt, _Elem, _RxTraits>::_Error(regex_constants::error_type _Code) { // handle error
     _Xregex_error(_Code);
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-bool _Parser<_FwdIt, _Elem, _RxTraits>::_Is_esc(_FwdIt _Ch0) const { // assumes _Ch0 != _End
+bool _Parser2<_FwdIt, _Elem, _RxTraits>::_Is_esc(_FwdIt _Ch0) const { // assumes _Ch0 != _End
     return ++_Ch0 != _End
         && ((!(_L_flags & _L_nex_grp) && (*_Ch0 == _Meta_lpar || *_Ch0 == _Meta_rpar))
             || (!(_L_flags & _L_nex_rep) && (*_Ch0 == _Meta_lbr || *_Ch0 == _Meta_rbr)));
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Parser<_FwdIt, _Elem, _RxTraits>::_Trans() { // map character to meta-character
+void _Parser2<_FwdIt, _Elem, _RxTraits>::_Trans() { // map character to meta-character
     static constexpr char _Meta_map[] = {_Meta_lpar, _Meta_rpar, _Meta_dlr, _Meta_caret, _Meta_dot, _Meta_star,
         _Meta_plus, _Meta_query, _Meta_lsq, _Meta_rsq, _Meta_bar, _Meta_esc, _Meta_dash, _Meta_lbr, _Meta_rbr,
         _Meta_comma, _Meta_colon, _Meta_equal, _Meta_exc, _Meta_nl, _Meta_cr, _Meta_bsp, 0}; // array of meta chars
@@ -4278,7 +4278,7 @@ void _Parser<_FwdIt, _Elem, _RxTraits>::_Trans() { // map character to meta-char
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Parser<_FwdIt, _Elem, _RxTraits>::_Next() { // advance to next input character
+void _Parser2<_FwdIt, _Elem, _RxTraits>::_Next() { // advance to next input character
     if (_Pat != _End) { // advance
         if (*_Pat == _Meta_esc && _Is_esc(_Pat)) {
             ++_Pat;
@@ -4290,7 +4290,7 @@ void _Parser<_FwdIt, _Elem, _RxTraits>::_Next() { // advance to next input chara
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Parser<_FwdIt, _Elem, _RxTraits>::_Expect(_Meta_type _St, regex_constants::error_type _Code) {
+void _Parser2<_FwdIt, _Elem, _RxTraits>::_Expect(_Meta_type _St, regex_constants::error_type _Code) {
     // check whether current meta-character is _St
     if (_Mchar != _St) {
         _Error(_Code);
@@ -4300,7 +4300,7 @@ void _Parser<_FwdIt, _Elem, _RxTraits>::_Expect(_Meta_type _St, regex_constants:
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-int _Parser<_FwdIt, _Elem, _RxTraits>::_Do_digits(
+int _Parser2<_FwdIt, _Elem, _RxTraits>::_Do_digits(
     int _Base, int _Initial, int _Count, regex_constants::error_type _Error_type) { // translate digits to numeric value
     int _Chv;
     _Val = _Initial;
@@ -4317,25 +4317,25 @@ int _Parser<_FwdIt, _Elem, _RxTraits>::_Do_digits(
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-bool _Parser<_FwdIt, _Elem, _RxTraits>::_DecimalDigits3(
+bool _Parser2<_FwdIt, _Elem, _RxTraits>::_DecimalDigits3(
     const regex_constants::error_type _Error_type, const int _Initial /* = 0 */) { // check for decimal value
     return _Do_digits(10, _Initial, INT_MAX, _Error_type) != INT_MAX;
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Parser<_FwdIt, _Elem, _RxTraits>::_HexDigits(int _Count) { // check for _Count hex digits
+void _Parser2<_FwdIt, _Elem, _RxTraits>::_HexDigits(int _Count) { // check for _Count hex digits
     if (_Do_digits(16, 0, _Count, regex_constants::error_escape) != 0) {
         _Error(regex_constants::error_escape);
     }
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-bool _Parser<_FwdIt, _Elem, _RxTraits>::_OctalDigits() { // check for up to 3 octal digits
+bool _Parser2<_FwdIt, _Elem, _RxTraits>::_OctalDigits() { // check for up to 3 octal digits
     return _Do_digits(8, 0, 3, regex_constants::error_escape) != 3;
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-_Prs_ret _Parser<_FwdIt, _Elem, _RxTraits>::_Do_ex_class2(
+_Prs_ret _Parser2<_FwdIt, _Elem, _RxTraits>::_Do_ex_class2(
     _Meta_type _End_arg) { // handle delimited expressions within bracket expression
     const regex_constants::error_type _Errtype =
         _End_arg == _Meta_colon ? regex_constants::error_ctype : regex_constants::error_collate;
@@ -4404,7 +4404,7 @@ _Prs_ret _Parser<_FwdIt, _Elem, _RxTraits>::_Do_ex_class2(
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-bool _Parser<_FwdIt, _Elem, _RxTraits>::_CharacterClassEscape(bool _Addit) { // check for character class escape
+bool _Parser2<_FwdIt, _Elem, _RxTraits>::_CharacterClassEscape(bool _Addit) { // check for character class escape
     typename _RxTraits::char_class_type _Cls;
     _FwdIt _Ch0 = _Pat;
     if (_Ch0 == _End || (_Cls = _Traits.lookup_classname(_Pat, ++_Ch0, (_Flags & regex_constants::icase) != 0)) == 0) {
@@ -4433,7 +4433,7 @@ bool _Parser<_FwdIt, _Elem, _RxTraits>::_CharacterClassEscape(bool _Addit) { //
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-_Prs_ret _Parser<_FwdIt, _Elem, _RxTraits>::_ClassEscape3() { // check for class escape
+_Prs_ret _Parser2<_FwdIt, _Elem, _RxTraits>::_ClassEscape3() { // check for class escape
     if ((_L_flags & _L_esc_bsp) && _Char == _Esc_ctrl_b) { // handle backspace escape
         _Next();
         _Val = _Meta_bsp;
@@ -4456,7 +4456,7 @@ _Prs_ret _Parser<_FwdIt, _Elem, _RxTraits>::_ClassEscape3() { // check for class
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-_Prs_ret _Parser<_FwdIt, _Elem, _RxTraits>::_ClassAtom(const bool _Initial) { // check for class atom
+_Prs_ret _Parser2<_FwdIt, _Elem, _RxTraits>::_ClassAtom(const bool _Initial) { // check for class atom
     if (_Mchar == _Meta_esc && (_L_flags & (_L_grp_esc | _L_ident_awk))) { // check for valid escape sequence
         _Next();
         return _ClassEscape3();
@@ -4483,7 +4483,7 @@ _Prs_ret _Parser<_FwdIt, _Elem, _RxTraits>::_ClassAtom(const bool _Initial) { //
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Parser<_FwdIt, _Elem, _RxTraits>::_ClassRanges() { // check for valid class ranges
+void _Parser2<_FwdIt, _Elem, _RxTraits>::_ClassRanges() { // check for valid class ranges
     _Prs_ret _Ret;
 
     bool _Initial = true;
@@ -4532,7 +4532,7 @@ void _Parser<_FwdIt, _Elem, _RxTraits>::_ClassRanges() { // check for valid clas
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Parser<_FwdIt, _Elem, _RxTraits>::_CharacterClass() { // add bracket expression
+void _Parser2<_FwdIt, _Elem, _RxTraits>::_CharacterClass() { // add bracket expression
     _Nfa._Add_class();
     if (_Mchar == _Meta_caret) { // negate bracket expression
         _Nfa._Negate();
@@ -4543,7 +4543,7 @@ void _Parser<_FwdIt, _Elem, _RxTraits>::_CharacterClass() { // add bracket expre
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Parser<_FwdIt, _Elem, _RxTraits>::_Do_capture_group() { // add capture group
+void _Parser2<_FwdIt, _Elem, _RxTraits>::_Do_capture_group() { // add capture group
     ++_Grp_idx;
 
     if (_Grp_idx >= 1000) { // hardcoded limit
@@ -4558,21 +4558,21 @@ void _Parser<_FwdIt, _Elem, _RxTraits>::_Do_capture_group() { // add capture gro
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Parser<_FwdIt, _Elem, _RxTraits>::_Do_noncapture_group() { // add non-capture group
+void _Parser2<_FwdIt, _Elem, _RxTraits>::_Do_noncapture_group() { // add non-capture group
     _Node_base* _Pos1 = _Nfa._Begin_group();
     _Disjunction();
     _Nfa._End_group(_Pos1);
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Parser<_FwdIt, _Elem, _RxTraits>::_Do_assert_group(bool _Neg) { // add assert group
+void _Parser2<_FwdIt, _Elem, _RxTraits>::_Do_assert_group(bool _Neg) { // add assert group
     _Node_base* _Pos1 = _Nfa._Begin_assert_group(_Neg);
     _Disjunction();
     _Nfa._End_assert_group(_Pos1);
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-bool _Parser<_FwdIt, _Elem, _RxTraits>::_Wrapped_disjunction() { // add disjunction inside group
+bool _Parser2<_FwdIt, _Elem, _RxTraits>::_Wrapped_disjunction() { // add disjunction inside group
     ++_Disj_count;
     if (!(_L_flags & _L_empty_grp) && _Mchar == _Meta_rpar) {
         _Error(regex_constants::error_paren);
@@ -4604,7 +4604,7 @@ bool _Parser<_FwdIt, _Elem, _RxTraits>::_Wrapped_disjunction() { // add disjunct
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-bool _Parser<_FwdIt, _Elem, _RxTraits>::_IsIdentityEscape(bool _In_character_class) const {
+bool _Parser2<_FwdIt, _Elem, _RxTraits>::_IsIdentityEscape(bool _In_character_class) const {
     // check for valid identity escape
 
     if (_L_flags & _L_ident_ECMA) {
@@ -4654,7 +4654,7 @@ bool _Parser<_FwdIt, _Elem, _RxTraits>::_IsIdentityEscape(bool _In_character_cla
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-bool _Parser<_FwdIt, _Elem, _RxTraits>::_IdentityEscape(bool _In_character_class) {
+bool _Parser2<_FwdIt, _Elem, _RxTraits>::_IdentityEscape(bool _In_character_class) {
     // check whether an escape is valid, and process it if so
     if (_IsIdentityEscape(_In_character_class)) {
         _Val = _Char;
@@ -4666,7 +4666,7 @@ bool _Parser<_FwdIt, _Elem, _RxTraits>::_IdentityEscape(bool _In_character_class
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-bool _Parser<_FwdIt, _Elem, _RxTraits>::_Do_ffn(_Elem _Ch) { // check for limited file format escape characters
+bool _Parser2<_FwdIt, _Elem, _RxTraits>::_Do_ffn(_Elem _Ch) { // check for limited file format escape characters
     if (_Ch == _Esc_ctrl_f) {
         _Val = '\f';
     } else if (_Ch == _Esc_ctrl_n) {
@@ -4685,7 +4685,7 @@ bool _Parser<_FwdIt, _Elem, _RxTraits>::_Do_ffn(_Elem _Ch) { // check for limite
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-bool _Parser<_FwdIt, _Elem, _RxTraits>::_Do_ffnx(_Elem _Ch) { // check for the remaining file format escape characters
+bool _Parser2<_FwdIt, _Elem, _RxTraits>::_Do_ffnx(_Elem _Ch) { // check for the remaining file format escape characters
     if (_Ch == _Esc_ctrl_a) {
         _Val = '\a';
     } else if (_Ch == _Esc_ctrl_b) {
@@ -4698,7 +4698,8 @@ bool _Parser<_FwdIt, _Elem, _RxTraits>::_Do_ffnx(_Elem _Ch) { // check for the r
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-bool _Parser<_FwdIt, _Elem, _RxTraits>::_CharacterEscape(bool _In_character_class) { // check for valid character escape
+bool _Parser2<_FwdIt, _Elem, _RxTraits>::_CharacterEscape(bool _In_character_class) {
+    // check for valid character escape
     if (_Mchar == _Meta_eos) {
         _Error(regex_constants::error_escape);
     }
@@ -4736,7 +4737,7 @@ bool _Parser<_FwdIt, _Elem, _RxTraits>::_CharacterEscape(bool _In_character_clas
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Parser<_FwdIt, _Elem, _RxTraits>::_AtomEscape() { // check for valid atom escape
+void _Parser2<_FwdIt, _Elem, _RxTraits>::_AtomEscape() { // check for valid atom escape
     if ((_L_flags & (_L_bzr_chr | _L_bckr)) && (_Val = _Traits.value(_Char, 10)) != -1) { // escaped decimal sequence
         _Next();
         if ((_L_flags & _L_bzr_chr) && _Val == 0) { // handle \0
@@ -4766,7 +4767,7 @@ void _Parser<_FwdIt, _Elem, _RxTraits>::_AtomEscape() { // check for valid atom
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Parser<_FwdIt, _Elem, _RxTraits>::_Quantifier() { // check for quantifier following atom
+void _Parser2<_FwdIt, _Elem, _RxTraits>::_Quantifier() { // check for quantifier following atom
     int _Min = 0;
     int _Max = -1;
     if (_Mchar != _Meta_star) {
@@ -4812,7 +4813,7 @@ void _Parser<_FwdIt, _Elem, _RxTraits>::_Quantifier() { // check for quantifier
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-bool _Parser<_FwdIt, _Elem, _RxTraits>::_Alternative() { // check for valid alternative
+bool _Parser2<_FwdIt, _Elem, _RxTraits>::_Alternative() { // check for valid alternative
     bool _Found = false;
     for (;;) { // concatenate valid elements
         bool _Quant = true;
@@ -4879,7 +4880,7 @@ bool _Parser<_FwdIt, _Elem, _RxTraits>::_Alternative() { // check for valid alte
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Parser<_FwdIt, _Elem, _RxTraits>::_Disjunction() { // check for valid disjunction
+void _Parser2<_FwdIt, _Elem, _RxTraits>::_Disjunction() { // check for valid disjunction
     _Node_base* _Pos1 = _Nfa._Getmark();
     if (!_Alternative()) {
         if (_Mchar != _Meta_bar) {
@@ -4904,7 +4905,7 @@ void _Parser<_FwdIt, _Elem, _RxTraits>::_Disjunction() { // check for valid disj
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Parser<_FwdIt, _Elem, _RxTraits>::_Calculate_loop_simplicity(
+void _Parser2<_FwdIt, _Elem, _RxTraits>::_Calculate_loop_simplicity(
     _Node_base* _Nx, _Node_base* _Ne, _Node_rep* _Outer_rep) {
     // walks regex NFA, calculates values of _Node_rep::_Simple_loop
     for (; _Nx != _Ne && _Nx; _Nx = _Nx->_Next) {
@@ -4990,7 +4991,7 @@ void _Parser<_FwdIt, _Elem, _RxTraits>::_Calculate_loop_simplicity(
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-_Root_node* _Parser<_FwdIt, _Elem, _RxTraits>::_Compile() { // compile regular expression
+_Root_node* _Parser2<_FwdIt, _Elem, _RxTraits>::_Compile() { // compile regular expression
     _Root_node* _Res = nullptr;
     _Tidy_guard<decltype(_Nfa)> _Guard{_STD addressof(_Nfa)};
     _Node_base* _Pos1 = _Nfa._Begin_capture_group(0);
@@ -5009,7 +5010,7 @@ _Root_node* _Parser<_FwdIt, _Elem, _RxTraits>::_Compile() { // compile regular e
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-_Parser<_FwdIt, _Elem, _RxTraits>::_Parser(
+_Parser2<_FwdIt, _Elem, _RxTraits>::_Parser2(
     const _RxTraits& _Tr, _FwdIt _Pfirst, _FwdIt _Plast, regex_constants::syntax_option_type _Fx)
     : _Pat(_Pfirst), _Begin(_Pfirst), _End(_Plast), _Nfa(_Tr, _Fx), _Traits(_Tr), _Flags(_Fx) {
 

From bcdf7875ccd2c1e6f1eddb03eb2c27c45b11eea1 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Julian=20M=C3=BCller?= <julian.mueller@uni-konstanz.de>
Date: Sun, 18 May 2025 13:18:54 +0200
Subject: [PATCH 02/21] remove unnecessary version numbers from _Parser2's
 member functions

---
 stl/inc/regex | 22 +++++++++++-----------
 1 file changed, 11 insertions(+), 11 deletions(-)

diff --git a/stl/inc/regex b/stl/inc/regex
index 1a319380297..8eabdcef15e 100644
--- a/stl/inc/regex
+++ b/stl/inc/regex
@@ -1807,12 +1807,12 @@ private:
 
     // parsing
     int _Do_digits(int _Base, int _Initial, int _Count, regex_constants::error_type _Error_type);
-    bool _DecimalDigits3(regex_constants::error_type _Error_type, int _Initial = 0);
+    bool _DecimalDigits(regex_constants::error_type _Error_type, int _Initial = 0);
     void _HexDigits(int);
     bool _OctalDigits();
-    _Prs_ret _Do_ex_class2(_Meta_type);
+    _Prs_ret _Do_ex_class(_Meta_type);
     bool _CharacterClassEscape(bool);
-    _Prs_ret _ClassEscape3();
+    _Prs_ret _ClassEscape();
     _Prs_ret _ClassAtom(bool);
     void _ClassRanges();
     void _CharacterClass();
@@ -4317,7 +4317,7 @@ int _Parser2<_FwdIt, _Elem, _RxTraits>::_Do_digits(
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-bool _Parser2<_FwdIt, _Elem, _RxTraits>::_DecimalDigits3(
+bool _Parser2<_FwdIt, _Elem, _RxTraits>::_DecimalDigits(
     const regex_constants::error_type _Error_type, const int _Initial /* = 0 */) { // check for decimal value
     return _Do_digits(10, _Initial, INT_MAX, _Error_type) != INT_MAX;
 }
@@ -4335,7 +4335,7 @@ bool _Parser2<_FwdIt, _Elem, _RxTraits>::_OctalDigits() { // check for up to 3 o
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-_Prs_ret _Parser2<_FwdIt, _Elem, _RxTraits>::_Do_ex_class2(
+_Prs_ret _Parser2<_FwdIt, _Elem, _RxTraits>::_Do_ex_class(
     _Meta_type _End_arg) { // handle delimited expressions within bracket expression
     const regex_constants::error_type _Errtype =
         _End_arg == _Meta_colon ? regex_constants::error_ctype : regex_constants::error_collate;
@@ -4433,7 +4433,7 @@ bool _Parser2<_FwdIt, _Elem, _RxTraits>::_CharacterClassEscape(bool _Addit) { //
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-_Prs_ret _Parser2<_FwdIt, _Elem, _RxTraits>::_ClassEscape3() { // check for class escape
+_Prs_ret _Parser2<_FwdIt, _Elem, _RxTraits>::_ClassEscape() { // check for class escape
     if ((_L_flags & _L_esc_bsp) && _Char == _Esc_ctrl_b) { // handle backspace escape
         _Next();
         _Val = _Meta_bsp;
@@ -4459,13 +4459,13 @@ template <class _FwdIt, class _Elem, class _RxTraits>
 _Prs_ret _Parser2<_FwdIt, _Elem, _RxTraits>::_ClassAtom(const bool _Initial) { // check for class atom
     if (_Mchar == _Meta_esc && (_L_flags & (_L_grp_esc | _L_ident_awk))) { // check for valid escape sequence
         _Next();
-        return _ClassEscape3();
+        return _ClassEscape();
     } else if (_Mchar == _Meta_lsq) { // check for valid delimited expression
         _Next();
         if (_Mchar == _Meta_colon || _Mchar == _Meta_equal || _Mchar == _Meta_dot) { // handle delimited expression
             _Meta_type _St = _Mchar;
             _Next();
-            return _Do_ex_class2(_St);
+            return _Do_ex_class(_St);
         } else { // handle ordinary [
             _Val = _Meta_lsq;
             return _Prs_chr;
@@ -4747,7 +4747,7 @@ void _Parser2<_FwdIt, _Elem, _RxTraits>::_AtomEscape() { // check for valid atom
             _Nfa._Add_char2(_Elem{});
         } else if (_L_flags & _L_bckr) { // check for valid backreference
             if (!(_L_flags & _L_lim_bckr)) {
-                (void) _DecimalDigits3(regex_constants::error_backref, _Val);
+                (void) _DecimalDigits(regex_constants::error_backref, _Val);
             }
 
             if (_Val == 0) {
@@ -4777,7 +4777,7 @@ void _Parser2<_FwdIt, _Elem, _RxTraits>::_Quantifier() { // check for quantifier
             _Max = 1;
         } else if (_Mchar == _Meta_lbr) { // check for valid bracketed value
             _Next();
-            if (!_DecimalDigits3(regex_constants::error_badbrace)) {
+            if (!_DecimalDigits(regex_constants::error_badbrace)) {
                 _Error(regex_constants::error_badbrace);
             }
 
@@ -4787,7 +4787,7 @@ void _Parser2<_FwdIt, _Elem, _RxTraits>::_Quantifier() { // check for quantifier
             } else { // check for decimal constant following comma
                 _Next();
                 if (_Mchar != _Meta_rbr) {
-                    if (!_DecimalDigits3(regex_constants::error_badbrace)) {
+                    if (!_DecimalDigits(regex_constants::error_badbrace)) {
                         _Error(regex_constants::error_badbrace);
                     }
 

From d93567b5451f393fccbf331a0eda6de83e9f8b94 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Julian=20M=C3=BCller?= <julian.mueller@uni-konstanz.de>
Date: Sun, 18 May 2025 13:19:39 +0200
Subject: [PATCH 03/21] clean up parse flags, extend range of possible parse
 flags, add new parse flag for forbidden dash at range start

---
 stl/inc/regex | 94 ++++++++++++++++++++++++++-------------------------
 1 file changed, 48 insertions(+), 46 deletions(-)

diff --git a/stl/inc/regex b/stl/inc/regex
index 8eabdcef15e..d04a9275fe5 100644
--- a/stl/inc/regex
+++ b/stl/inc/regex
@@ -1843,42 +1843,43 @@ private:
     int _Val;
     _Elem _Char;
     _Meta_type _Mchar;
-    unsigned int _L_flags;
+    unsigned long long _L_flags;
 };
 
-enum _Lang_flags { // describe language properties
-    _L_ext_rep    = 0x00000001, // + and ? repetitions
-    _L_alt_pipe   = 0x00000002, // uses '|' for alternation
-    _L_alt_nl     = 0x00000004, // uses '\n' for alternation (grep, egrep)
-    _L_nex_grp    = 0x00000008, // has non-escaped capture groups
-    _L_nex_rep    = 0x00000010, // has non-escaped repeats
-    _L_nc_grp     = 0x00000020, // has non-capture groups (?:xxx)
-    _L_asrt_gen   = 0x00000040, // has generalized assertions (?=xxx), (?!xxx)
-    _L_asrt_wrd   = 0x00000080, // has word boundary assertions (\b, \B)
-    _L_bckr       = 0x00000100, // has backreferences (ERE doesn't)
-    _L_lim_bckr   = 0x00000200, // has limited backreferences (BRE \1-\9)
-    _L_ngr_rep    = 0x00000400, // has non-greedy repeats
-    _L_esc_uni    = 0x00000800, // has Unicode escape sequences
-    _L_esc_hex    = 0x00001000, // has hexadecimal escape sequences
-    _L_esc_oct    = 0x00002000, // has octal escape sequences
-    _L_esc_bsp    = 0x00004000, // has backspace escape in character classes
-    _L_esc_ffnx   = 0x00008000, // has extra file escapes (\a and \b)
-    _L_esc_ffn    = 0x00010000, // has limited file escapes (\[fnrtv])
-    _L_esc_wsd    = 0x00020000, // has w, s, and d character set escapes
-    _L_esc_ctrl   = 0x00040000, // has control escape
-    _L_no_nl      = 0x00080000, // no newline in pattern or matching text
-    _L_bzr_chr    = 0x00100000, // \0 is a valid character constant
-    _L_grp_esc    = 0x00200000, // \ is special character in group
-    _L_ident_ECMA = 0x00400000, // ECMA identity escape (not identifierpart)
-    _L_ident_ERE  = 0x00800000, // ERE identity escape (.[\*^$, plus {+?}()
-    _L_ident_awk  = 0x01000000, // awk identity escape ( ERE plus "/)
-    _L_anch_rstr  = 0x02000000, // anchor restricted to beginning/end
-    _L_star_beg   = 0x04000000, // star okay at beginning of RE/expr (BRE)
-    _L_empty_grp  = 0x08000000, // empty group allowed (ERE prohibits "()")
-    _L_paren_bal  = 0x10000000, // ')'/'}' special only after '('/'{'
-    _L_brk_bal    = 0x20000000, // ']' special only after '[' (ERE, BRE); TRANSITION, ABI: same value as _L_brk_rstr
-    _L_brk_rstr   = 0x20000000, // ']' not special when first character in set
-    _L_mtch_long  = 0x40000000, // find longest match (ERE, BRE)
+enum _Lang_flags2 : unsigned long long { // describe language properties
+    _L_ext_rep    = 0x000000001ULL, // + and ? repetitions
+    _L_alt_pipe   = 0x000000002ULL, // uses '|' for alternation
+    _L_alt_nl     = 0x000000004ULL, // uses '\n' for alternation (grep, egrep)
+    _L_nex_grp    = 0x000000008ULL, // has non-escaped capture groups
+    _L_nex_rep    = 0x000000010ULL, // has non-escaped repeats
+    _L_nc_grp     = 0x000000020ULL, // has non-capture groups (?:xxx)
+    _L_asrt_gen   = 0x000000040ULL, // has generalized assertions (?=xxx), (?!xxx)
+    _L_asrt_wrd   = 0x000000080ULL, // has word boundary assertions (\b, \B)
+    _L_bckr       = 0x000000100ULL, // has backreferences (ERE doesn't)
+    _L_lim_bckr   = 0x000000200ULL, // has limited backreferences (BRE \1-\9)
+    _L_ngr_rep    = 0x000000400ULL, // has non-greedy repeats
+    _L_esc_uni    = 0x000000800ULL, // has Unicode escape sequences
+    _L_esc_hex    = 0x000001000ULL, // has hexadecimal escape sequences
+    _L_esc_oct    = 0x000002000ULL, // has octal escape sequences
+    _L_esc_bsp    = 0x000004000ULL, // has backspace escape in character classes
+    _L_esc_ffnx   = 0x000008000ULL, // has extra file escapes (\a and \b)
+    _L_esc_ffn    = 0x000010000ULL, // has limited file escapes (\[fnrtv])
+    _L_esc_wsd    = 0x000020000ULL, // has w, s, and d character set escapes
+    _L_esc_ctrl   = 0x000040000ULL, // has control escape
+    _L_no_nl      = 0x000080000ULL, // no newline in pattern or matching text
+    _L_bzr_chr    = 0x000100000ULL, // \0 is a valid character constant
+    _L_grp_esc    = 0x000200000ULL, // \ is special character in group
+    _L_ident_ECMA = 0x000400000ULL, // ECMA identity escape (not identifierpart)
+    _L_ident_ERE  = 0x000800000ULL, // ERE identity escape (.[\*^$, plus {+?}()
+    _L_ident_awk  = 0x001000000ULL, // awk identity escape ( ERE plus "/)
+    _L_anch_rstr  = 0x002000000ULL, // anchor restricted to beginning/end
+    _L_star_beg   = 0x004000000ULL, // star okay at beginning of RE/expr (BRE)
+    _L_empty_grp  = 0x008000000ULL, // empty group allowed (ERE prohibits "()")
+    _L_paren_bal  = 0x010000000ULL, // ')'/'}' special only after '('/'{'
+    _L_brk_bal    = 0x020000000ULL, // ']' special only after '[' (ERE, BRE)
+    _L_brk_rstr   = 0x040000000ULL, // ']' not special when first character in set
+    _L_dsh_rstr   = 0x080000000ULL, // '-' forbidden at range start in set except when first character (ERE, BRE)
+    _L_mtch_long  = 0x100000000ULL, // find longest match (ERE, BRE)
 };
 
 class _Regex_base : public _Container_base { // base class for basic_regex to construct and destroy proxy
@@ -4457,7 +4458,7 @@ _Prs_ret _Parser2<_FwdIt, _Elem, _RxTraits>::_ClassEscape() { // check for class
 
 template <class _FwdIt, class _Elem, class _RxTraits>
 _Prs_ret _Parser2<_FwdIt, _Elem, _RxTraits>::_ClassAtom(const bool _Initial) { // check for class atom
-    if (_Mchar == _Meta_esc && (_L_flags & (_L_grp_esc | _L_ident_awk))) { // check for valid escape sequence
+    if (_Mchar == _Meta_esc && (_L_flags & _L_grp_esc)) { // check for valid escape sequence
         _Next();
         return _ClassEscape();
     } else if (_Mchar == _Meta_lsq) { // check for valid delimited expression
@@ -5014,22 +5015,23 @@ _Parser2<_FwdIt, _Elem, _RxTraits>::_Parser2(
     const _RxTraits& _Tr, _FwdIt _Pfirst, _FwdIt _Plast, regex_constants::syntax_option_type _Fx)
     : _Pat(_Pfirst), _Begin(_Pfirst), _End(_Plast), _Nfa(_Tr, _Fx), _Traits(_Tr), _Flags(_Fx) {
 
-    constexpr unsigned int _ECMA_flags = _L_ext_rep | _L_alt_pipe | _L_nex_grp | _L_nex_rep | _L_nc_grp | _L_asrt_gen
-                                       | _L_asrt_wrd | _L_bckr | _L_ngr_rep | _L_esc_uni | _L_esc_hex | _L_esc_bsp
-                                       | _L_esc_ffn | _L_esc_wsd | _L_esc_ctrl | _L_bzr_chr | _L_grp_esc | _L_ident_ECMA
-                                       | _L_empty_grp;
+    constexpr unsigned long long _ECMA_flags = _L_ext_rep | _L_alt_pipe | _L_nex_grp | _L_nex_rep | _L_nc_grp
+                                             | _L_asrt_gen | _L_asrt_wrd | _L_bckr | _L_ngr_rep | _L_esc_uni
+                                             | _L_esc_hex | _L_esc_bsp | _L_esc_ffn | _L_esc_wsd | _L_esc_ctrl
+                                             | _L_bzr_chr | _L_grp_esc | _L_ident_ECMA | _L_empty_grp;
 
-    constexpr unsigned int _Basic_flags =
-        _L_bckr | _L_lim_bckr | _L_anch_rstr | _L_star_beg | _L_empty_grp | _L_brk_bal | _L_brk_rstr | _L_mtch_long;
+    constexpr unsigned long long _Basic_flags = _L_bckr | _L_lim_bckr | _L_anch_rstr | _L_star_beg | _L_empty_grp
+                                              | _L_brk_bal | _L_brk_rstr | _L_dsh_rstr | _L_mtch_long;
 
-    constexpr unsigned int _Grep_flags = _Basic_flags | _L_alt_nl | _L_no_nl;
+    constexpr unsigned long long _Grep_flags = _Basic_flags | _L_alt_nl | _L_no_nl;
 
-    constexpr unsigned int _Extended_flags = _L_ext_rep | _L_alt_pipe | _L_nex_grp | _L_nex_rep | _L_ident_ERE
-                                           | _L_paren_bal | _L_brk_bal | _L_brk_rstr | _L_mtch_long;
+    constexpr unsigned long long _Extended_flags = _L_ext_rep | _L_alt_pipe | _L_nex_grp | _L_nex_rep | _L_ident_ERE
+                                                 | _L_paren_bal | _L_brk_bal | _L_brk_rstr | _L_dsh_rstr | _L_mtch_long;
 
-    constexpr unsigned int _Awk_flags = _Extended_flags | _L_esc_oct | _L_esc_ffn | _L_esc_ffnx | _L_ident_awk;
+    constexpr unsigned long long _Awk_flags =
+        _Extended_flags | _L_esc_oct | _L_esc_ffn | _L_esc_ffnx | _L_grp_esc | _L_ident_awk;
 
-    constexpr unsigned int _Egrep_flags = _Extended_flags | _L_alt_nl | _L_no_nl;
+    constexpr unsigned long long _Egrep_flags = _Extended_flags | _L_alt_nl | _L_no_nl;
 
     const regex_constants::syntax_option_type _Masked = _Flags & regex_constants::_Gmask;
 

From a687fb0e8c202ae7ec94599f239789cea0a38dc3 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Julian=20M=C3=BCller?= <julian.mueller@uni-konstanz.de>
Date: Sun, 18 May 2025 13:22:17 +0200
Subject: [PATCH 04/21] remove unused member `_Begin` from `_Parser2`

---
 stl/inc/regex | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/stl/inc/regex b/stl/inc/regex
index d04a9275fe5..ad240917085 100644
--- a/stl/inc/regex
+++ b/stl/inc/regex
@@ -1832,7 +1832,6 @@ private:
     void _Calculate_loop_simplicity(_Node_base* _Nx, _Node_base* _Ne, _Node_rep* _Outer_rep);
 
     _FwdIt _Pat;
-    _FwdIt _Begin;
     _FwdIt _End;
     unsigned int _Grp_idx = 0;
     int _Disj_count       = 0;
@@ -5013,7 +5012,7 @@ _Root_node* _Parser2<_FwdIt, _Elem, _RxTraits>::_Compile() { // compile regular
 template <class _FwdIt, class _Elem, class _RxTraits>
 _Parser2<_FwdIt, _Elem, _RxTraits>::_Parser2(
     const _RxTraits& _Tr, _FwdIt _Pfirst, _FwdIt _Plast, regex_constants::syntax_option_type _Fx)
-    : _Pat(_Pfirst), _Begin(_Pfirst), _End(_Plast), _Nfa(_Tr, _Fx), _Traits(_Tr), _Flags(_Fx) {
+    : _Pat(_Pfirst), _End(_Plast), _Nfa(_Tr, _Fx), _Traits(_Tr), _Flags(_Fx) {
 
     constexpr unsigned long long _ECMA_flags = _L_ext_rep | _L_alt_pipe | _L_nex_grp | _L_nex_rep | _L_nc_grp
                                              | _L_asrt_gen | _L_asrt_wrd | _L_bckr | _L_ngr_rep | _L_esc_uni

From 8adea0f73d609095773cb01df7ba46dd1cc73307 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Julian=20M=C3=BCller?= <julian.mueller@uni-konstanz.de>
Date: Sun, 18 May 2025 13:22:48 +0200
Subject: [PATCH 05/21] reorder _Parser2 members to reduce padding

---
 stl/inc/regex | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/stl/inc/regex b/stl/inc/regex
index ad240917085..89a5e587e73 100644
--- a/stl/inc/regex
+++ b/stl/inc/regex
@@ -1838,11 +1838,11 @@ private:
     vector<bool> _Finished_grps;
     _Builder<_FwdIt, _Elem, _RxTraits> _Nfa;
     const _RxTraits& _Traits;
+    unsigned long long _L_flags;
     regex_constants::syntax_option_type _Flags;
     int _Val;
-    _Elem _Char;
     _Meta_type _Mchar;
-    unsigned long long _L_flags;
+    _Elem _Char;
 };
 
 enum _Lang_flags2 : unsigned long long { // describe language properties

From 3cefca79294d83c95ecdceb61c5b9b38646e5061 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Julian=20M=C3=BCller?= <julian.mueller@uni-konstanz.de>
Date: Sun, 18 May 2025 13:25:52 +0200
Subject: [PATCH 06/21] rename `_Builder` to `_Builder2`

---
 stl/inc/regex | 78 +++++++++++++++++++++++++--------------------------
 1 file changed, 39 insertions(+), 39 deletions(-)

diff --git a/stl/inc/regex b/stl/inc/regex
index 89a5e587e73..24a60616f6f 100644
--- a/stl/inc/regex
+++ b/stl/inc/regex
@@ -1574,9 +1574,9 @@ enum class _Rx_char_class_kind : int { // must be aligned with corresponding _No
 };
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-class _Builder { // provides operations used by _Parser2 to build the nfa
+class _Builder2 { // provides operations used by _Parser2 to build the nfa
 public:
-    _Builder(const _RxTraits& _Tr, regex_constants::syntax_option_type);
+    _Builder2(const _RxTraits& _Tr, regex_constants::syntax_option_type);
     void _Setlong();
     // _Discard_pattern is an ABI zombie name
     void _Tidy() noexcept;
@@ -1624,7 +1624,7 @@ private:
     const int _Tmax; // TRANSITION, ABI: preserved for binary compatibility
 
 public:
-    _Builder& operator=(const _Builder&) = delete;
+    _Builder2& operator=(const _Builder2&) = delete;
 };
 
 template <class _BidIt>
@@ -1836,7 +1836,7 @@ private:
     unsigned int _Grp_idx = 0;
     int _Disj_count       = 0;
     vector<bool> _Finished_grps;
-    _Builder<_FwdIt, _Elem, _RxTraits> _Nfa;
+    _Builder2<_FwdIt, _Elem, _RxTraits> _Nfa;
     const _RxTraits& _Traits;
     unsigned long long _L_flags;
     regex_constants::syntax_option_type _Flags;
@@ -2837,28 +2837,28 @@ _EXPORT_STD using sregex_token_iterator  = regex_token_iterator<string::const_it
 _EXPORT_STD using wsregex_token_iterator = regex_token_iterator<wstring::const_iterator>;
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-_Builder<_FwdIt, _Elem, _RxTraits>::_Builder(const _RxTraits& _Tr, regex_constants::syntax_option_type _Fx)
+_Builder2<_FwdIt, _Elem, _RxTraits>::_Builder2(const _RxTraits& _Tr, regex_constants::syntax_option_type _Fx)
     : _Root(new _Root_node), _Current(_Root), _Flags(_Fx), _Traits(_Tr),
       _Bmax(static_cast<int>(_Fx & regex_constants::collate ? 0U : _Bmp_max)),
       _Tmax(static_cast<int>(_Fx & regex_constants::collate ? 0U : _ARRAY_THRESHOLD)) {}
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Builder<_FwdIt, _Elem, _RxTraits>::_Setlong() { // set flag
+void _Builder2<_FwdIt, _Elem, _RxTraits>::_Setlong() { // set flag
     _Root->_Flags |= _Fl_longest;
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Builder<_FwdIt, _Elem, _RxTraits>::_Negate() { // set flag
+void _Builder2<_FwdIt, _Elem, _RxTraits>::_Negate() { // set flag
     _Current->_Flags ^= _Fl_negate;
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-_Node_base* _Builder<_FwdIt, _Elem, _RxTraits>::_Getmark() const {
+_Node_base* _Builder2<_FwdIt, _Elem, _RxTraits>::_Getmark() const {
     return _Current;
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-_Node_base* _Builder<_FwdIt, _Elem, _RxTraits>::_Link_node(_Node_base* _Nx) { // insert _Nx at current location
+_Node_base* _Builder2<_FwdIt, _Elem, _RxTraits>::_Link_node(_Node_base* _Nx) { // insert _Nx at current location
     _Nx->_Prev = _Current;
     if (_Current->_Next) { // set back pointer
         _Nx->_Next             = _Current->_Next;
@@ -2870,7 +2870,7 @@ _Node_base* _Builder<_FwdIt, _Elem, _RxTraits>::_Link_node(_Node_base* _Nx) { //
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Builder<_FwdIt, _Elem, _RxTraits>::_Insert_node(_Node_base* _Insert_before, _Node_base* _To_insert) {
+void _Builder2<_FwdIt, _Elem, _RxTraits>::_Insert_node(_Node_base* _Insert_before, _Node_base* _To_insert) {
     // insert _To_insert into the graph before the node _Insert_before
     _Insert_before->_Prev->_Next = _To_insert;
     _To_insert->_Prev            = _Insert_before->_Prev;
@@ -2879,42 +2879,42 @@ void _Builder<_FwdIt, _Elem, _RxTraits>::_Insert_node(_Node_base* _Insert_before
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-_Node_base* _Builder<_FwdIt, _Elem, _RxTraits>::_New_node(_Node_type _Kind) { // allocate and link simple node
+_Node_base* _Builder2<_FwdIt, _Elem, _RxTraits>::_New_node(_Node_type _Kind) { // allocate and link simple node
     return _Link_node(new _Node_base(_Kind));
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Builder<_FwdIt, _Elem, _RxTraits>::_Add_nop() { // add nop node
+void _Builder2<_FwdIt, _Elem, _RxTraits>::_Add_nop() { // add nop node
     _New_node(_N_nop);
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Builder<_FwdIt, _Elem, _RxTraits>::_Add_bol() { // add bol node
+void _Builder2<_FwdIt, _Elem, _RxTraits>::_Add_bol() { // add bol node
     _New_node(_N_bol);
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Builder<_FwdIt, _Elem, _RxTraits>::_Add_eol() { // add eol node
+void _Builder2<_FwdIt, _Elem, _RxTraits>::_Add_eol() { // add eol node
     _New_node(_N_eol);
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Builder<_FwdIt, _Elem, _RxTraits>::_Add_wbound() { // add wbound node
+void _Builder2<_FwdIt, _Elem, _RxTraits>::_Add_wbound() { // add wbound node
     _New_node(_N_wbound);
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Builder<_FwdIt, _Elem, _RxTraits>::_Add_dot() { // add dot node
+void _Builder2<_FwdIt, _Elem, _RxTraits>::_Add_dot() { // add dot node
     _New_node(_N_dot);
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Builder<_FwdIt, _Elem, _RxTraits>::_Add_str_node() { // add string node
+void _Builder2<_FwdIt, _Elem, _RxTraits>::_Add_str_node() { // add string node
     _Link_node(new _Node_str<_Elem>);
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Builder<_FwdIt, _Elem, _RxTraits>::_Add_char2(_Elem _Ch) { // append character
+void _Builder2<_FwdIt, _Elem, _RxTraits>::_Add_char2(_Elem _Ch) { // append character
     if (_Current->_Kind != _N_str) {
         _Add_str_node();
     }
@@ -2930,12 +2930,12 @@ void _Builder<_FwdIt, _Elem, _RxTraits>::_Add_char2(_Elem _Ch) { // append chara
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Builder<_FwdIt, _Elem, _RxTraits>::_Add_class() { // add bracket expression node
+void _Builder2<_FwdIt, _Elem, _RxTraits>::_Add_class() { // add bracket expression node
     _Link_node(new _Node_class<_Elem, _RxTraits>);
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Builder<_FwdIt, _Elem, _RxTraits>::_Add_char_to_bitmap(_Elem _Ch) { // add character to accelerator table
+void _Builder2<_FwdIt, _Elem, _RxTraits>::_Add_char_to_bitmap(_Elem _Ch) { // add character to accelerator table
     if (_Flags & regex_constants::icase) {
         _Ch = _Traits.translate_nocase(_Ch);
     }
@@ -2950,7 +2950,7 @@ void _Builder<_FwdIt, _Elem, _RxTraits>::_Add_char_to_bitmap(_Elem _Ch) { // add
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Builder<_FwdIt, _Elem, _RxTraits>::_Add_char_to_array(_Elem _Ch) { // append character to character array
+void _Builder2<_FwdIt, _Elem, _RxTraits>::_Add_char_to_array(_Elem _Ch) { // append character to character array
     if (_Flags & regex_constants::icase) {
         _Ch = _Traits.translate_nocase(_Ch);
     }
@@ -2964,7 +2964,7 @@ void _Builder<_FwdIt, _Elem, _RxTraits>::_Add_char_to_array(_Elem _Ch) { // appe
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Builder<_FwdIt, _Elem, _RxTraits>::_Add_char_to_class(_Elem _Ch) { // add character to bracket expression
+void _Builder2<_FwdIt, _Elem, _RxTraits>::_Add_char_to_class(_Elem _Ch) { // add character to bracket expression
     if (static_cast<typename _RxTraits::_Uelem>(_Ch) < _Bmp_max) {
         _Add_char_to_bitmap(_Ch);
     } else {
@@ -2973,7 +2973,7 @@ void _Builder<_FwdIt, _Elem, _RxTraits>::_Add_char_to_class(_Elem _Ch) { // add
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Builder<_FwdIt, _Elem, _RxTraits>::_Add_range3(const _Elem _Arg0, const _Elem _Arg1) {
+void _Builder2<_FwdIt, _Elem, _RxTraits>::_Add_range3(const _Elem _Arg0, const _Elem _Arg1) {
     // add character range to set
     using string_type                    = typename _RxTraits::string_type;
     unsigned int _Ex0                    = static_cast<typename _RxTraits::_Uelem>(_Arg0);
@@ -3032,7 +3032,7 @@ void _Builder<_FwdIt, _Elem, _RxTraits>::_Add_range3(const _Elem _Arg0, const _E
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Builder<_FwdIt, _Elem, _RxTraits>::_Add_elts(
+void _Builder2<_FwdIt, _Elem, _RxTraits>::_Add_elts(
     _Node_class<_Elem, _RxTraits>* _Node, typename _RxTraits::char_class_type _Cl, bool _Negate) {
     // add characters in named class to set
     for (unsigned int _Ch = 0; _Ch < _Bmp_max; ++_Ch) { // add elements or their inverse
@@ -3048,7 +3048,7 @@ void _Builder<_FwdIt, _Elem, _RxTraits>::_Add_elts(
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Builder<_FwdIt, _Elem, _RxTraits>::_Add_named_class(
+void _Builder2<_FwdIt, _Elem, _RxTraits>::_Add_named_class(
     typename _RxTraits::char_class_type _Cl, const _Rx_char_class_kind _Kind) {
     // add contents of named class to bracket expression
     using _Char_class_type               = typename _RxTraits::char_class_type;
@@ -3073,7 +3073,7 @@ void _Builder<_FwdIt, _Elem, _RxTraits>::_Add_named_class(
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Builder<_FwdIt, _Elem, _RxTraits>::_Char_to_elts2(const _Elem* const _First, const _Elem* const _Last,
+void _Builder2<_FwdIt, _Elem, _RxTraits>::_Char_to_elts2(const _Elem* const _First, const _Elem* const _Last,
     _Sequence<_Elem>** _Cur) { // add collation element to element sequence
     auto _Diff = static_cast<unsigned int>(_Last - _First);
     while (*_Cur && _Diff < (*_Cur)->_Sz) {
@@ -3090,7 +3090,7 @@ void _Builder<_FwdIt, _Elem, _RxTraits>::_Char_to_elts2(const _Elem* const _Firs
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Builder<_FwdIt, _Elem, _RxTraits>::_Add_equiv2(const _Elem* const _First, const _Elem* const _Last) {
+void _Builder2<_FwdIt, _Elem, _RxTraits>::_Add_equiv2(const _Elem* const _First, const _Elem* const _Last) {
     // add elements of equivalence class to bracket expression
     _Node_class<_Elem, _RxTraits>* _Node = static_cast<_Node_class<_Elem, _RxTraits>*>(_Current);
     typename _RxTraits::string_type _Str = _Traits.transform_primary(_First, _Last);
@@ -3117,7 +3117,7 @@ void _Builder<_FwdIt, _Elem, _RxTraits>::_Add_equiv2(const _Elem* const _First,
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Builder<_FwdIt, _Elem, _RxTraits>::_Add_coll2(const _Elem* const _First, const _Elem* const _Last) {
+void _Builder2<_FwdIt, _Elem, _RxTraits>::_Add_coll2(const _Elem* const _First, const _Elem* const _Last) {
     // add collation element to bracket expression
     _Node_class<_Elem, _RxTraits>* _Node = static_cast<_Node_class<_Elem, _RxTraits>*>(_Current);
     _Sequence<_Elem>** _Cur              = _STD addressof(_Node->_Coll);
@@ -3125,12 +3125,12 @@ void _Builder<_FwdIt, _Elem, _RxTraits>::_Add_coll2(const _Elem* const _First, c
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-_Node_base* _Builder<_FwdIt, _Elem, _RxTraits>::_Begin_group() { // add group node
+_Node_base* _Builder2<_FwdIt, _Elem, _RxTraits>::_Begin_group() { // add group node
     return _New_node(_N_group);
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Builder<_FwdIt, _Elem, _RxTraits>::_End_group(_Node_base* _Back) { // add end of group node
+void _Builder2<_FwdIt, _Elem, _RxTraits>::_End_group(_Node_base* _Back) { // add end of group node
     _Node_type _Elt;
     if (_Back->_Kind == _N_group) {
         _Elt = _N_end_group;
@@ -3144,7 +3144,7 @@ void _Builder<_FwdIt, _Elem, _RxTraits>::_End_group(_Node_base* _Back) { // add
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-_Node_base* _Builder<_FwdIt, _Elem, _RxTraits>::_Begin_assert_group(const bool _Neg) { // add assert node
+_Node_base* _Builder2<_FwdIt, _Elem, _RxTraits>::_Begin_assert_group(const bool _Neg) { // add assert node
     auto _Node1_unique   = _STD make_unique<_Node_assert>(_Neg ? _N_neg_assert : _N_assert);
     _Node_base* _Node2   = new _Node_base(_N_nop);
     _Node_assert* _Node1 = _Node1_unique.release();
@@ -3156,23 +3156,23 @@ _Node_base* _Builder<_FwdIt, _Elem, _RxTraits>::_Begin_assert_group(const bool _
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Builder<_FwdIt, _Elem, _RxTraits>::_End_assert_group(_Node_base* _Nx) { // add end of assert node
+void _Builder2<_FwdIt, _Elem, _RxTraits>::_End_assert_group(_Node_base* _Nx) { // add end of assert node
     _End_group(_Nx);
     _Current = _Nx;
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-_Node_base* _Builder<_FwdIt, _Elem, _RxTraits>::_Begin_capture_group(unsigned int _Idx) { // add capture group node
+_Node_base* _Builder2<_FwdIt, _Elem, _RxTraits>::_Begin_capture_group(unsigned int _Idx) { // add capture group node
     return _Link_node(new _Node_capture(_Idx));
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Builder<_FwdIt, _Elem, _RxTraits>::_Add_backreference(unsigned int _Idx) { // add back reference node
+void _Builder2<_FwdIt, _Elem, _RxTraits>::_Add_backreference(unsigned int _Idx) { // add back reference node
     _Link_node(new _Node_back(_Idx));
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-_Node_base* _Builder<_FwdIt, _Elem, _RxTraits>::_Begin_if(_Node_base* _Start) { // add if node
+_Node_base* _Builder2<_FwdIt, _Elem, _RxTraits>::_Begin_if(_Node_base* _Start) { // add if node
     // append endif node
     _Node_base* _Res = new _Node_endif;
     _Link_node(_Res);
@@ -3185,7 +3185,7 @@ _Node_base* _Builder<_FwdIt, _Elem, _RxTraits>::_Begin_if(_Node_base* _Start) {
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Builder<_FwdIt, _Elem, _RxTraits>::_Else_if(_Node_base* _Start, _Node_base* _End) { // add else node
+void _Builder2<_FwdIt, _Elem, _RxTraits>::_Else_if(_Node_base* _Start, _Node_base* _End) { // add else node
     _Node_if* _Parent  = static_cast<_Node_if*>(_Start->_Next);
     _Node_base* _First = _End->_Next;
     _End->_Next        = nullptr;
@@ -3203,7 +3203,7 @@ void _Builder<_FwdIt, _Elem, _RxTraits>::_Else_if(_Node_base* _Start, _Node_base
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Builder<_FwdIt, _Elem, _RxTraits>::_Add_rep2(int _Min, int _Max, bool _Greedy) { // add repeat node
+void _Builder2<_FwdIt, _Elem, _RxTraits>::_Add_rep2(int _Min, int _Max, bool _Greedy) { // add repeat node
     if (_Current->_Kind == _N_str
         && static_cast<_Node_str<_Elem>*>(_Current)->_Data._Size() != 1) { // move final character to new string node
         _Node_str<_Elem>* _Node = static_cast<_Node_str<_Elem>*>(_Current);
@@ -3256,13 +3256,13 @@ void _Builder<_FwdIt, _Elem, _RxTraits>::_Add_rep2(int _Min, int _Max, bool _Gre
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-_Root_node* _Builder<_FwdIt, _Elem, _RxTraits>::_End_pattern() { // wrap up
+_Root_node* _Builder2<_FwdIt, _Elem, _RxTraits>::_End_pattern() { // wrap up
     _New_node(_N_end);
     return _Root;
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Builder<_FwdIt, _Elem, _RxTraits>::_Tidy() noexcept { // free memory
+void _Builder2<_FwdIt, _Elem, _RxTraits>::_Tidy() noexcept { // free memory
     _Destroy_node(_Root);
     _Root = nullptr;
 }

From 1c0c01df65bfd16a6bac4ad302932e4d8d7d1b97 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Julian=20M=C3=BCller?= <julian.mueller@uni-konstanz.de>
Date: Sun, 18 May 2025 13:26:05 +0200
Subject: [PATCH 07/21] remove version numbers from member functions of
 `_Builder2`

---
 stl/inc/regex | 32 ++++++++++++++++----------------
 1 file changed, 16 insertions(+), 16 deletions(-)

diff --git a/stl/inc/regex b/stl/inc/regex
index 24a60616f6f..7ce568a59a6 100644
--- a/stl/inc/regex
+++ b/stl/inc/regex
@@ -1590,10 +1590,10 @@ public:
     void _Add_char2(_Elem _Ch);
     void _Add_class();
     void _Add_char_to_class(_Elem _Ch);
-    void _Add_range3(_Elem, _Elem);
+    void _Add_range(_Elem, _Elem);
     void _Add_named_class(typename _RxTraits::char_class_type, _Rx_char_class_kind);
-    void _Add_equiv2(const _Elem*, const _Elem*);
-    void _Add_coll2(const _Elem*, const _Elem*);
+    void _Add_equiv(const _Elem*, const _Elem*);
+    void _Add_coll(const _Elem*, const _Elem*);
     _Node_base* _Begin_group();
     void _End_group(_Node_base* _Back);
     _Node_base* _Begin_assert_group(bool);
@@ -1602,7 +1602,7 @@ public:
     void _Add_backreference(unsigned int _Idx);
     _Node_base* _Begin_if(_Node_base* _Start);
     void _Else_if(_Node_base*, _Node_base*);
-    void _Add_rep2(int _Min, int _Max, bool _Greedy);
+    void _Add_rep(int _Min, int _Max, bool _Greedy);
     void _Negate();
     _Root_node* _End_pattern();
 
@@ -1614,7 +1614,7 @@ private:
     void _Add_char_to_bitmap(_Elem _Ch);
     void _Add_char_to_array(_Elem _Ch);
     void _Add_elts(_Node_class<_Elem, _RxTraits>*, typename _RxTraits::char_class_type, bool);
-    void _Char_to_elts2(const _Elem*, const _Elem*, _Sequence<_Elem>**);
+    void _Char_to_elts(const _Elem*, const _Elem*, _Sequence<_Elem>**);
 
     _Root_node* _Root;
     _Node_base* _Current;
@@ -2973,7 +2973,7 @@ void _Builder2<_FwdIt, _Elem, _RxTraits>::_Add_char_to_class(_Elem _Ch) { // add
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Builder2<_FwdIt, _Elem, _RxTraits>::_Add_range3(const _Elem _Arg0, const _Elem _Arg1) {
+void _Builder2<_FwdIt, _Elem, _RxTraits>::_Add_range(const _Elem _Arg0, const _Elem _Arg1) {
     // add character range to set
     using string_type                    = typename _RxTraits::string_type;
     unsigned int _Ex0                    = static_cast<typename _RxTraits::_Uelem>(_Arg0);
@@ -3073,7 +3073,7 @@ void _Builder2<_FwdIt, _Elem, _RxTraits>::_Add_named_class(
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Builder2<_FwdIt, _Elem, _RxTraits>::_Char_to_elts2(const _Elem* const _First, const _Elem* const _Last,
+void _Builder2<_FwdIt, _Elem, _RxTraits>::_Char_to_elts(const _Elem* const _First, const _Elem* const _Last,
     _Sequence<_Elem>** _Cur) { // add collation element to element sequence
     auto _Diff = static_cast<unsigned int>(_Last - _First);
     while (*_Cur && _Diff < (*_Cur)->_Sz) {
@@ -3090,7 +3090,7 @@ void _Builder2<_FwdIt, _Elem, _RxTraits>::_Char_to_elts2(const _Elem* const _Fir
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Builder2<_FwdIt, _Elem, _RxTraits>::_Add_equiv2(const _Elem* const _First, const _Elem* const _Last) {
+void _Builder2<_FwdIt, _Elem, _RxTraits>::_Add_equiv(const _Elem* const _First, const _Elem* const _Last) {
     // add elements of equivalence class to bracket expression
     _Node_class<_Elem, _RxTraits>* _Node = static_cast<_Node_class<_Elem, _RxTraits>*>(_Current);
     typename _RxTraits::string_type _Str = _Traits.transform_primary(_First, _Last);
@@ -3112,16 +3112,16 @@ void _Builder2<_FwdIt, _Elem, _RxTraits>::_Add_equiv2(const _Elem* const _First,
     }
     if (_Bmp_max < static_cast<unsigned int>(_STD _Max_limit<_Elem>())) { // map range
         _Sequence<_Elem>** _Cur = _STD addressof(_Node->_Equiv);
-        _Char_to_elts2(_First, _Last, _Cur);
+        _Char_to_elts(_First, _Last, _Cur);
     }
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Builder2<_FwdIt, _Elem, _RxTraits>::_Add_coll2(const _Elem* const _First, const _Elem* const _Last) {
+void _Builder2<_FwdIt, _Elem, _RxTraits>::_Add_coll(const _Elem* const _First, const _Elem* const _Last) {
     // add collation element to bracket expression
     _Node_class<_Elem, _RxTraits>* _Node = static_cast<_Node_class<_Elem, _RxTraits>*>(_Current);
     _Sequence<_Elem>** _Cur              = _STD addressof(_Node->_Coll);
-    _Char_to_elts2(_First, _Last, _Cur);
+    _Char_to_elts(_First, _Last, _Cur);
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
@@ -3203,7 +3203,7 @@ void _Builder2<_FwdIt, _Elem, _RxTraits>::_Else_if(_Node_base* _Start, _Node_bas
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Builder2<_FwdIt, _Elem, _RxTraits>::_Add_rep2(int _Min, int _Max, bool _Greedy) { // add repeat node
+void _Builder2<_FwdIt, _Elem, _RxTraits>::_Add_rep(int _Min, int _Max, bool _Greedy) { // add repeat node
     if (_Current->_Kind == _N_str
         && static_cast<_Node_str<_Elem>*>(_Current)->_Data._Size() != 1) { // move final character to new string node
         _Node_str<_Elem>* _Node = static_cast<_Node_str<_Elem>*>(_Current);
@@ -4391,13 +4391,13 @@ _Prs_ret _Parser2<_FwdIt, _Elem, _RxTraits>::_Do_ex_class(
         }
 
         if (_End_arg == _Meta_equal) { // process equivalence
-            _Nfa._Add_equiv2(_Coll_elem_first, _Coll_elem_last);
+            _Nfa._Add_equiv(_Coll_elem_first, _Coll_elem_last);
             return _Prs_set;
         } else { // process collating element
 
             // Character ranges with multi-character bounds cannot be represented in NFA nodes yet (see GH-5391).
             // Provisionally treat multi-character collating elements as character sets.
-            _Nfa._Add_coll2(_Coll_elem_first, _Coll_elem_last);
+            _Nfa._Add_coll(_Coll_elem_first, _Coll_elem_last);
             return _Prs_set;
         }
     }
@@ -4524,7 +4524,7 @@ void _Parser2<_FwdIt, _Elem, _RxTraits>::_ClassRanges() { // check for valid cla
                 _Chr2 = _Traits.translate(_Chr2);
             }
 
-            _Nfa._Add_range3(_Chr1, _Chr2);
+            _Nfa._Add_range(_Chr1, _Chr2);
         } else if (_Ret == _Prs_chr) {
             _Nfa._Add_char_to_class(static_cast<_Elem>(_Val));
         }
@@ -4809,7 +4809,7 @@ void _Parser2<_FwdIt, _Elem, _RxTraits>::_Quantifier() { // check for quantifier
         _Next();
     }
 
-    _Nfa._Add_rep2(_Min, _Max, _Greedy);
+    _Nfa._Add_rep(_Min, _Max, _Greedy);
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>

From 7d9748a1b17ac2752f0fb6f9b655cc9cd9f6433d Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Julian=20M=C3=BCller?= <julian.mueller@uni-konstanz.de>
Date: Sun, 18 May 2025 13:31:23 +0200
Subject: [PATCH 08/21] remove obsolete member variables from `_Builder2`

---
 stl/inc/regex | 6 +-----
 1 file changed, 1 insertion(+), 5 deletions(-)

diff --git a/stl/inc/regex b/stl/inc/regex
index 7ce568a59a6..5e5fa084c2a 100644
--- a/stl/inc/regex
+++ b/stl/inc/regex
@@ -1620,8 +1620,6 @@ private:
     _Node_base* _Current;
     regex_constants::syntax_option_type _Flags;
     const _RxTraits& _Traits;
-    const int _Bmax; // TRANSITION, ABI: preserved for binary compatibility
-    const int _Tmax; // TRANSITION, ABI: preserved for binary compatibility
 
 public:
     _Builder2& operator=(const _Builder2&) = delete;
@@ -2838,9 +2836,7 @@ _EXPORT_STD using wsregex_token_iterator = regex_token_iterator<wstring::const_i
 
 template <class _FwdIt, class _Elem, class _RxTraits>
 _Builder2<_FwdIt, _Elem, _RxTraits>::_Builder2(const _RxTraits& _Tr, regex_constants::syntax_option_type _Fx)
-    : _Root(new _Root_node), _Current(_Root), _Flags(_Fx), _Traits(_Tr),
-      _Bmax(static_cast<int>(_Fx & regex_constants::collate ? 0U : _Bmp_max)),
-      _Tmax(static_cast<int>(_Fx & regex_constants::collate ? 0U : _ARRAY_THRESHOLD)) {}
+    : _Root(new _Root_node), _Current(_Root), _Flags(_Fx), _Traits(_Tr) {}
 
 template <class _FwdIt, class _Elem, class _RxTraits>
 void _Builder2<_FwdIt, _Elem, _RxTraits>::_Setlong() { // set flag

From 5fbd5b1f59ba4372126438e2c9b3f62692d66329 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Julian=20M=C3=BCller?= <julian.mueller@uni-konstanz.de>
Date: Sun, 18 May 2025 13:32:57 +0200
Subject: [PATCH 09/21] do not escape parentheses and braces in bracket
 expressions in basic or grep mode

---
 stl/inc/regex                                 |  9 +++--
 .../GH_005244_regex_escape_sequences/test.cpp | 35 +++++++++++++------
 2 files changed, 31 insertions(+), 13 deletions(-)

diff --git a/stl/inc/regex b/stl/inc/regex
index 5e5fa084c2a..d2e842cd309 100644
--- a/stl/inc/regex
+++ b/stl/inc/regex
@@ -1782,6 +1782,8 @@ enum _Prs_ret { // indicate class element type
     _Prs_set
 };
 
+enum class _Lex_mode : unsigned char { _Default, _Character_class };
+
 template <class _FwdIt, class _Elem, class _RxTraits>
 class _Parser2 { // parse a regular expression
 public:
@@ -1840,6 +1842,7 @@ private:
     regex_constants::syntax_option_type _Flags;
     int _Val;
     _Meta_type _Mchar;
+    _Lex_mode _Mode;
     _Elem _Char;
 };
 
@@ -4162,7 +4165,7 @@ template <class _FwdIt, class _Elem, class _RxTraits>
 
 template <class _FwdIt, class _Elem, class _RxTraits>
 bool _Parser2<_FwdIt, _Elem, _RxTraits>::_Is_esc(_FwdIt _Ch0) const { // assumes _Ch0 != _End
-    return ++_Ch0 != _End
+    return _Mode == _Lex_mode::_Default && ++_Ch0 != _End
         && ((!(_L_flags & _L_nex_grp) && (*_Ch0 == _Meta_lpar || *_Ch0 == _Meta_rpar))
             || (!(_L_flags & _L_nex_rep) && (*_Ch0 == _Meta_lbr || *_Ch0 == _Meta_rbr)));
 }
@@ -4835,8 +4838,10 @@ bool _Parser2<_FwdIt, _Elem, _RxTraits>::_Alternative() { // check for valid alt
                 _AtomEscape();
             }
         } else if (_Mchar == _Meta_lsq) { // add bracket expression
+            _Mode = _Lex_mode::_Character_class;
             _Next();
             _CharacterClass();
+            _Mode = _Lex_mode::_Default;
             _Expect(_Meta_rsq, regex_constants::error_brack);
         } else if (_Mchar == _Meta_lpar) { // check for valid group
             _Next();
@@ -5008,7 +5013,7 @@ _Root_node* _Parser2<_FwdIt, _Elem, _RxTraits>::_Compile() { // compile regular
 template <class _FwdIt, class _Elem, class _RxTraits>
 _Parser2<_FwdIt, _Elem, _RxTraits>::_Parser2(
     const _RxTraits& _Tr, _FwdIt _Pfirst, _FwdIt _Plast, regex_constants::syntax_option_type _Fx)
-    : _Pat(_Pfirst), _End(_Plast), _Nfa(_Tr, _Fx), _Traits(_Tr), _Flags(_Fx) {
+    : _Pat(_Pfirst), _End(_Plast), _Nfa(_Tr, _Fx), _Traits(_Tr), _Flags(_Fx), _Mode(_Lex_mode::_Default) {
 
     constexpr unsigned long long _ECMA_flags = _L_ext_rep | _L_alt_pipe | _L_nex_grp | _L_nex_rep | _L_nc_grp
                                              | _L_asrt_gen | _L_asrt_wrd | _L_bckr | _L_ngr_rep | _L_esc_uni
diff --git a/tests/std/tests/GH_005244_regex_escape_sequences/test.cpp b/tests/std/tests/GH_005244_regex_escape_sequences/test.cpp
index 7747befd700..cb5752a3073 100644
--- a/tests/std/tests/GH_005244_regex_escape_sequences/test.cpp
+++ b/tests/std/tests/GH_005244_regex_escape_sequences/test.cpp
@@ -632,22 +632,15 @@ void test_gh_5244_classescape_posix_not_awk(syntax_option_type option) {
     check_classescape_noescape("?", option);
     check_classescape_noescape("|", option);
 
-    // TRANSITION, GH-5379
-    if (option & (extended | egrep)) {
-        check_classescape_noescape("(", option);
-        check_classescape_noescape(")", option);
-        check_classescape_noescape("{", option);
-    }
+    check_classescape_noescape("(", option);
+    check_classescape_noescape(")", option);
+    check_classescape_noescape("{", option);
 
     // closing characters that are not considered special
     g_regexTester.should_match("\\]", R"([\]])", option);
     g_regexTester.should_not_match("]", R"([\]])", option);
     g_regexTester.should_not_match("\\", R"([\]])", option);
-
-    // TRANSITION, GH-5379
-    if (option & (extended | egrep)) {
-        check_classescape_noescape("}", option);
-    }
+    check_classescape_noescape("}", option);
 
     // awk escape sequences
     check_classescape_noescape("a", option);
@@ -882,8 +875,28 @@ void test_gh_5244() {
     test_gh_5244_classescape_awk();
 }
 
+void test_gh_5379() {
+    // GH-5379: Backslashes in character classes are sometimes not matched in basic regular expressions
+
+    // Correct handling of these backslashes at the beginning of a character class is already covered by GH-5244 tests.
+    // The following tests check that backslashes are handled correctly immediately after a bracketed character class.
+    for (syntax_option_type syntax : {basic, grep}) {
+        g_regexTester.should_throw(R"([a]\b)", error_escape, syntax);
+        g_regexTester.should_match("a[b]", R"([a]\[b])", syntax);
+        g_regexTester.should_match("a", R"(\([a]\))", syntax);
+        g_regexTester.should_match("ab", R"([a]\(b\))", syntax);
+        g_regexTester.should_match("a", R"([a]\{1\})", syntax);
+        g_regexTester.should_throw(R"([a]\})", error_brace, syntax);
+
+        // also check handling of identity escape "\]",
+        // which is supported as an extension following more recent POSIX standards
+        g_regexTester.should_match("a]", R"([a]\])", syntax);
+    }
+}
+
 int main() {
     test_gh_5244();
+    test_gh_5379();
 
     return g_regexTester.result();
 }

From ccc071b80d1ee29eba255151971cc018b00527e0 Mon Sep 17 00:00:00 2001
From: "Stephan T. Lavavej" <stl@microsoft.com>
Date: Thu, 22 May 2025 02:33:39 -0700
Subject: [PATCH 10/21] Rename `_Builder2::_Add_char2` back to `_Add_char`.

---
 stl/inc/regex | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)

diff --git a/stl/inc/regex b/stl/inc/regex
index d2e842cd309..dd4d6795819 100644
--- a/stl/inc/regex
+++ b/stl/inc/regex
@@ -1587,7 +1587,7 @@ public:
     void _Add_eol();
     void _Add_wbound();
     void _Add_dot();
-    void _Add_char2(_Elem _Ch);
+    void _Add_char(_Elem _Ch);
     void _Add_class();
     void _Add_char_to_class(_Elem _Ch);
     void _Add_range(_Elem, _Elem);
@@ -2913,7 +2913,7 @@ void _Builder2<_FwdIt, _Elem, _RxTraits>::_Add_str_node() { // add string node
 }
 
 template <class _FwdIt, class _Elem, class _RxTraits>
-void _Builder2<_FwdIt, _Elem, _RxTraits>::_Add_char2(_Elem _Ch) { // append character
+void _Builder2<_FwdIt, _Elem, _RxTraits>::_Add_char(_Elem _Ch) { // append character
     if (_Current->_Kind != _N_str) {
         _Add_str_node();
     }
@@ -3207,7 +3207,7 @@ void _Builder2<_FwdIt, _Elem, _RxTraits>::_Add_rep(int _Min, int _Max, bool _Gre
         && static_cast<_Node_str<_Elem>*>(_Current)->_Data._Size() != 1) { // move final character to new string node
         _Node_str<_Elem>* _Node = static_cast<_Node_str<_Elem>*>(_Current);
         _Add_str_node();
-        _Add_char2(_Node->_Data._Del());
+        _Add_char(_Node->_Data._Del());
     }
 
     _Node_base* _Pos = _Current;
@@ -4743,7 +4743,7 @@ void _Parser2<_FwdIt, _Elem, _RxTraits>::_AtomEscape() { // check for valid atom
             if (_Traits.value(_Char, 10) != -1) {
                 _Error(regex_constants::error_escape);
             }
-            _Nfa._Add_char2(_Elem{});
+            _Nfa._Add_char(_Elem{});
         } else if (_L_flags & _L_bckr) { // check for valid backreference
             if (!(_L_flags & _L_lim_bckr)) {
                 (void) _DecimalDigits(regex_constants::error_backref, _Val);
@@ -4759,7 +4759,7 @@ void _Parser2<_FwdIt, _Elem, _RxTraits>::_AtomEscape() { // check for valid atom
             }
         }
     } else if (_CharacterEscape(false)) {
-        _Nfa._Add_char2(static_cast<_Elem>(_Val));
+        _Nfa._Add_char(static_cast<_Elem>(_Val));
     } else if (!(_L_flags & _L_esc_wsd) || !_CharacterClassEscape(true)) {
         _Error(regex_constants::error_escape);
     }
@@ -4851,7 +4851,7 @@ bool _Parser2<_FwdIt, _Elem, _RxTraits>::_Alternative() { // check for valid alt
             _Nfa._Add_bol();
             _Next();
             if ((_L_flags & _L_star_beg) && _Mchar == _Meta_star && !_Found) {
-                _Nfa._Add_char2(_Char);
+                _Nfa._Add_char(_Char);
                 _Next();
             } else {
                 _Quant = false;
@@ -4868,7 +4868,7 @@ bool _Parser2<_FwdIt, _Elem, _RxTraits>::_Alternative() { // check for valid alt
         } else if (_Mchar == _Meta_rsq && !(_L_flags & _L_brk_bal)) {
             _Error(regex_constants::error_brack);
         } else { // add character
-            _Nfa._Add_char2(_Char);
+            _Nfa._Add_char(_Char);
             _Next();
         }
 

From 8a92b64f8cced769f3d7626b7935f1467a795b18 Mon Sep 17 00:00:00 2001
From: "Stephan T. Lavavej" <stl@microsoft.com>
Date: Thu, 22 May 2025 02:35:16 -0700
Subject: [PATCH 11/21] Drop "ABI zombie name" comment.

---
 stl/inc/regex | 1 -
 1 file changed, 1 deletion(-)

diff --git a/stl/inc/regex b/stl/inc/regex
index dd4d6795819..9d464df5f92 100644
--- a/stl/inc/regex
+++ b/stl/inc/regex
@@ -1578,7 +1578,6 @@ class _Builder2 { // provides operations used by _Parser2 to build the nfa
 public:
     _Builder2(const _RxTraits& _Tr, regex_constants::syntax_option_type);
     void _Setlong();
-    // _Discard_pattern is an ABI zombie name
     void _Tidy() noexcept;
     _Node_base* _Getmark() const;
 

From b6651d8c2dae153ea74b6046a8e1334d925ab6b0 Mon Sep 17 00:00:00 2001
From: "Stephan T. Lavavej" <stl@microsoft.com>
Date: Thu, 22 May 2025 02:42:12 -0700
Subject: [PATCH 12/21] Delete copy ctors for `_Builder2` and `_Matcher2`.

---
 stl/inc/regex | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/stl/inc/regex b/stl/inc/regex
index 9d464df5f92..1fd2309c5f4 100644
--- a/stl/inc/regex
+++ b/stl/inc/regex
@@ -1621,6 +1621,7 @@ private:
     const _RxTraits& _Traits;
 
 public:
+    _Builder2(const _Builder2&)            = delete;
     _Builder2& operator=(const _Builder2&) = delete;
 };
 
@@ -1772,6 +1773,7 @@ private:
     typename _RxTraits::char_class_type _Char_class_d{};
 
 public:
+    _Matcher2(const _Matcher2&)            = delete;
     _Matcher2& operator=(const _Matcher2&) = delete;
 };
 

From e32b28b91f87fa11d7dd36944632420db60e3101 Mon Sep 17 00:00:00 2001
From: "Stephan T. Lavavej" <stl@microsoft.com>
Date: Thu, 22 May 2025 03:06:40 -0700
Subject: [PATCH 13/21] `_Builder2::_Add_nop` was unused.

---
 stl/inc/regex | 6 ------
 1 file changed, 6 deletions(-)

diff --git a/stl/inc/regex b/stl/inc/regex
index 1fd2309c5f4..e8299da406b 100644
--- a/stl/inc/regex
+++ b/stl/inc/regex
@@ -1581,7 +1581,6 @@ public:
     void _Tidy() noexcept;
     _Node_base* _Getmark() const;
 
-    void _Add_nop();
     void _Add_bol();
     void _Add_eol();
     void _Add_wbound();
@@ -2883,11 +2882,6 @@ _Node_base* _Builder2<_FwdIt, _Elem, _RxTraits>::_New_node(_Node_type _Kind) { /
     return _Link_node(new _Node_base(_Kind));
 }
 
-template <class _FwdIt, class _Elem, class _RxTraits>
-void _Builder2<_FwdIt, _Elem, _RxTraits>::_Add_nop() { // add nop node
-    _New_node(_N_nop);
-}
-
 template <class _FwdIt, class _Elem, class _RxTraits>
 void _Builder2<_FwdIt, _Elem, _RxTraits>::_Add_bol() { // add bol node
     _New_node(_N_bol);

From 260cbce022e2696fbe9ff2dc6aa919465e1fe919 Mon Sep 17 00:00:00 2001
From: "Stephan T. Lavavej" <stl@microsoft.com>
Date: Thu, 22 May 2025 03:07:03 -0700
Subject: [PATCH 14/21] Avoid shadowing in `_Builder2`: `_Negate` =>
 `_Negative`

---
 stl/inc/regex | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/stl/inc/regex b/stl/inc/regex
index e8299da406b..b3e32a62934 100644
--- a/stl/inc/regex
+++ b/stl/inc/regex
@@ -3027,11 +3027,11 @@ void _Builder2<_FwdIt, _Elem, _RxTraits>::_Add_range(const _Elem _Arg0, const _E
 
 template <class _FwdIt, class _Elem, class _RxTraits>
 void _Builder2<_FwdIt, _Elem, _RxTraits>::_Add_elts(
-    _Node_class<_Elem, _RxTraits>* _Node, typename _RxTraits::char_class_type _Cl, bool _Negate) {
+    _Node_class<_Elem, _RxTraits>* _Node, typename _RxTraits::char_class_type _Cl, bool _Negative) {
     // add characters in named class to set
     for (unsigned int _Ch = 0; _Ch < _Bmp_max; ++_Ch) { // add elements or their inverse
         bool _Matches = _Traits.isctype(static_cast<_Elem>(_Ch), _Cl);
-        if (_Matches != _Negate) { // add contents of named class to accelerator table
+        if (_Matches != _Negative) { // add contents of named class to accelerator table
             if (!_Node->_Small) {
                 _Node->_Small = new _Bitmap;
             }

From 43c3235becc7c965f9a509b7d61b767a2545bc86 Mon Sep 17 00:00:00 2001
From: "Stephan T. Lavavej" <stl@microsoft.com>
Date: Thu, 22 May 2025 03:13:36 -0700
Subject: [PATCH 15/21] `_Parser2::char_class_type` was unused.

---
 stl/inc/regex | 2 --
 1 file changed, 2 deletions(-)

diff --git a/stl/inc/regex b/stl/inc/regex
index b3e32a62934..c572fe31b4f 100644
--- a/stl/inc/regex
+++ b/stl/inc/regex
@@ -1787,8 +1787,6 @@ enum class _Lex_mode : unsigned char { _Default, _Character_class };
 template <class _FwdIt, class _Elem, class _RxTraits>
 class _Parser2 { // parse a regular expression
 public:
-    using char_class_type = typename _RxTraits::char_class_type;
-
     _Parser2(const _RxTraits& _Tr, _FwdIt _Pfirst, _FwdIt _Plast, regex_constants::syntax_option_type _Fx);
     _Root_node* _Compile();
 

From 50204cb1071d2b336ed34f2239bf6040dac92fef Mon Sep 17 00:00:00 2001
From: "Stephan T. Lavavej" <stl@microsoft.com>
Date: Thu, 22 May 2025 03:17:02 -0700
Subject: [PATCH 16/21] `_Parser2::_Mark_count` can be private.

---
 stl/inc/regex | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/stl/inc/regex b/stl/inc/regex
index c572fe31b4f..fc11ef61b50 100644
--- a/stl/inc/regex
+++ b/stl/inc/regex
@@ -1790,11 +1790,11 @@ public:
     _Parser2(const _RxTraits& _Tr, _FwdIt _Pfirst, _FwdIt _Plast, regex_constants::syntax_option_type _Fx);
     _Root_node* _Compile();
 
+private:
     unsigned int _Mark_count() const noexcept {
         return _Grp_idx + 1;
     }
 
-private:
     // lexing
     [[noreturn]] void _Error(regex_constants::error_type);
 

From 4fba27c14a8a06ca6e817e0bed4dc779311ebaa6 Mon Sep 17 00:00:00 2001
From: "Stephan T. Lavavej" <stl@microsoft.com>
Date: Thu, 22 May 2025 05:06:46 -0700
Subject: [PATCH 17/21] Use a data member initializer for `_Lex_mode _Mode`.

---
 stl/inc/regex | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/stl/inc/regex b/stl/inc/regex
index fc11ef61b50..e73454755a5 100644
--- a/stl/inc/regex
+++ b/stl/inc/regex
@@ -1840,7 +1840,7 @@ private:
     regex_constants::syntax_option_type _Flags;
     int _Val;
     _Meta_type _Mchar;
-    _Lex_mode _Mode;
+    _Lex_mode _Mode = _Lex_mode::_Default;
     _Elem _Char;
 };
 
@@ -5006,7 +5006,7 @@ _Root_node* _Parser2<_FwdIt, _Elem, _RxTraits>::_Compile() { // compile regular
 template <class _FwdIt, class _Elem, class _RxTraits>
 _Parser2<_FwdIt, _Elem, _RxTraits>::_Parser2(
     const _RxTraits& _Tr, _FwdIt _Pfirst, _FwdIt _Plast, regex_constants::syntax_option_type _Fx)
-    : _Pat(_Pfirst), _End(_Plast), _Nfa(_Tr, _Fx), _Traits(_Tr), _Flags(_Fx), _Mode(_Lex_mode::_Default) {
+    : _Pat(_Pfirst), _End(_Plast), _Nfa(_Tr, _Fx), _Traits(_Tr), _Flags(_Fx) {
 
     constexpr unsigned long long _ECMA_flags = _L_ext_rep | _L_alt_pipe | _L_nex_grp | _L_nex_rep | _L_nc_grp
                                              | _L_asrt_gen | _L_asrt_wrd | _L_bckr | _L_ngr_rep | _L_esc_uni

From 5e0151dfbe85ed97187bb245ee2cc31928df9e6a Mon Sep 17 00:00:00 2001
From: "Stephan T. Lavavej" <stl@microsoft.com>
Date: Thu, 22 May 2025 05:32:57 -0700
Subject: [PATCH 18/21] Meld `_L_nc_grp` and `_L_asrt_gen` into `_L_nc_asrt`,
 improve comments.

---
 stl/inc/regex | 19 +++++++++----------
 1 file changed, 9 insertions(+), 10 deletions(-)

diff --git a/stl/inc/regex b/stl/inc/regex
index e73454755a5..e83ee402c4e 100644
--- a/stl/inc/regex
+++ b/stl/inc/regex
@@ -1850,8 +1850,7 @@ enum _Lang_flags2 : unsigned long long { // describe language properties
     _L_alt_nl     = 0x000000004ULL, // uses '\n' for alternation (grep, egrep)
     _L_nex_grp    = 0x000000008ULL, // has non-escaped capture groups
     _L_nex_rep    = 0x000000010ULL, // has non-escaped repeats
-    _L_nc_grp     = 0x000000020ULL, // has non-capture groups (?:xxx)
-    _L_asrt_gen   = 0x000000040ULL, // has generalized assertions (?=xxx), (?!xxx)
+    _L_nc_asrt    = 0x000000020ULL, // has non-capture groups (?:xxx) and generalized assertions (?=xxx), (?!xxx)
     _L_asrt_wrd   = 0x000000080ULL, // has word boundary assertions (\b, \B)
     _L_bckr       = 0x000000100ULL, // has backreferences (ERE doesn't)
     _L_lim_bckr   = 0x000000200ULL, // has limited backreferences (BRE \1-\9)
@@ -4568,17 +4567,17 @@ bool _Parser2<_FwdIt, _Elem, _RxTraits>::_Wrapped_disjunction() { // add disjunc
     ++_Disj_count;
     if (!(_L_flags & _L_empty_grp) && _Mchar == _Meta_rpar) {
         _Error(regex_constants::error_paren);
-    } else if ((_L_flags & _L_nc_grp) && _Mchar == _Meta_query) { // check for valid ECMAScript (?x ... ) group
+    } else if ((_L_flags & _L_nc_asrt) && _Mchar == _Meta_query) { // check for valid ECMAScript (?x ... ) group
         _Next();
         _Meta_type _Ch = _Mchar;
         _Next();
-        if (_Ch == _Meta_colon) {
+        if (_Ch == _Meta_colon) { // process non-capture group (?:meow)
             _Do_noncapture_group();
-        } else if (_Ch == _Meta_exc) { // process assert group, negating
+        } else if (_Ch == _Meta_exc) { // process negative lookahead assertion (?!meow)
             _Do_assert_group(true);
             --_Disj_count;
             return false;
-        } else if (_Ch == _Meta_equal) { // process assert group
+        } else if (_Ch == _Meta_equal) { // process positive lookahead assertion (?=meow)
             _Do_assert_group(false);
             --_Disj_count;
             return false;
@@ -5008,10 +5007,10 @@ _Parser2<_FwdIt, _Elem, _RxTraits>::_Parser2(
     const _RxTraits& _Tr, _FwdIt _Pfirst, _FwdIt _Plast, regex_constants::syntax_option_type _Fx)
     : _Pat(_Pfirst), _End(_Plast), _Nfa(_Tr, _Fx), _Traits(_Tr), _Flags(_Fx) {
 
-    constexpr unsigned long long _ECMA_flags = _L_ext_rep | _L_alt_pipe | _L_nex_grp | _L_nex_rep | _L_nc_grp
-                                             | _L_asrt_gen | _L_asrt_wrd | _L_bckr | _L_ngr_rep | _L_esc_uni
-                                             | _L_esc_hex | _L_esc_bsp | _L_esc_ffn | _L_esc_wsd | _L_esc_ctrl
-                                             | _L_bzr_chr | _L_grp_esc | _L_ident_ECMA | _L_empty_grp;
+    constexpr unsigned long long _ECMA_flags = _L_ext_rep | _L_alt_pipe | _L_nex_grp | _L_nex_rep | _L_nc_asrt
+                                             | _L_asrt_wrd | _L_bckr | _L_ngr_rep | _L_esc_uni | _L_esc_hex | _L_esc_bsp
+                                             | _L_esc_ffn | _L_esc_wsd | _L_esc_ctrl | _L_bzr_chr | _L_grp_esc
+                                             | _L_ident_ECMA | _L_empty_grp;
 
     constexpr unsigned long long _Basic_flags = _L_bckr | _L_lim_bckr | _L_anch_rstr | _L_star_beg | _L_empty_grp
                                               | _L_brk_bal | _L_brk_rstr | _L_dsh_rstr | _L_mtch_long;

From cfaf2df16f37540c9b5b316bfb9edc0bdeb287b7 Mon Sep 17 00:00:00 2001
From: "Stephan T. Lavavej" <stl@microsoft.com>
Date: Thu, 22 May 2025 05:36:58 -0700
Subject: [PATCH 19/21] Renumber `_Lang_flags2`.

---
 stl/inc/regex | 52 +++++++++++++++++++++++++--------------------------
 1 file changed, 26 insertions(+), 26 deletions(-)

diff --git a/stl/inc/regex b/stl/inc/regex
index e83ee402c4e..5ea107724dc 100644
--- a/stl/inc/regex
+++ b/stl/inc/regex
@@ -1851,32 +1851,32 @@ enum _Lang_flags2 : unsigned long long { // describe language properties
     _L_nex_grp    = 0x000000008ULL, // has non-escaped capture groups
     _L_nex_rep    = 0x000000010ULL, // has non-escaped repeats
     _L_nc_asrt    = 0x000000020ULL, // has non-capture groups (?:xxx) and generalized assertions (?=xxx), (?!xxx)
-    _L_asrt_wrd   = 0x000000080ULL, // has word boundary assertions (\b, \B)
-    _L_bckr       = 0x000000100ULL, // has backreferences (ERE doesn't)
-    _L_lim_bckr   = 0x000000200ULL, // has limited backreferences (BRE \1-\9)
-    _L_ngr_rep    = 0x000000400ULL, // has non-greedy repeats
-    _L_esc_uni    = 0x000000800ULL, // has Unicode escape sequences
-    _L_esc_hex    = 0x000001000ULL, // has hexadecimal escape sequences
-    _L_esc_oct    = 0x000002000ULL, // has octal escape sequences
-    _L_esc_bsp    = 0x000004000ULL, // has backspace escape in character classes
-    _L_esc_ffnx   = 0x000008000ULL, // has extra file escapes (\a and \b)
-    _L_esc_ffn    = 0x000010000ULL, // has limited file escapes (\[fnrtv])
-    _L_esc_wsd    = 0x000020000ULL, // has w, s, and d character set escapes
-    _L_esc_ctrl   = 0x000040000ULL, // has control escape
-    _L_no_nl      = 0x000080000ULL, // no newline in pattern or matching text
-    _L_bzr_chr    = 0x000100000ULL, // \0 is a valid character constant
-    _L_grp_esc    = 0x000200000ULL, // \ is special character in group
-    _L_ident_ECMA = 0x000400000ULL, // ECMA identity escape (not identifierpart)
-    _L_ident_ERE  = 0x000800000ULL, // ERE identity escape (.[\*^$, plus {+?}()
-    _L_ident_awk  = 0x001000000ULL, // awk identity escape ( ERE plus "/)
-    _L_anch_rstr  = 0x002000000ULL, // anchor restricted to beginning/end
-    _L_star_beg   = 0x004000000ULL, // star okay at beginning of RE/expr (BRE)
-    _L_empty_grp  = 0x008000000ULL, // empty group allowed (ERE prohibits "()")
-    _L_paren_bal  = 0x010000000ULL, // ')'/'}' special only after '('/'{'
-    _L_brk_bal    = 0x020000000ULL, // ']' special only after '[' (ERE, BRE)
-    _L_brk_rstr   = 0x040000000ULL, // ']' not special when first character in set
-    _L_dsh_rstr   = 0x080000000ULL, // '-' forbidden at range start in set except when first character (ERE, BRE)
-    _L_mtch_long  = 0x100000000ULL, // find longest match (ERE, BRE)
+    _L_asrt_wrd   = 0x000000040ULL, // has word boundary assertions (\b, \B)
+    _L_bckr       = 0x000000080ULL, // has backreferences (ERE doesn't)
+    _L_lim_bckr   = 0x000000100ULL, // has limited backreferences (BRE \1-\9)
+    _L_ngr_rep    = 0x000000200ULL, // has non-greedy repeats
+    _L_esc_uni    = 0x000000400ULL, // has Unicode escape sequences
+    _L_esc_hex    = 0x000000800ULL, // has hexadecimal escape sequences
+    _L_esc_oct    = 0x000001000ULL, // has octal escape sequences
+    _L_esc_bsp    = 0x000002000ULL, // has backspace escape in character classes
+    _L_esc_ffnx   = 0x000004000ULL, // has extra file escapes (\a and \b)
+    _L_esc_ffn    = 0x000008000ULL, // has limited file escapes (\[fnrtv])
+    _L_esc_wsd    = 0x000010000ULL, // has w, s, and d character set escapes
+    _L_esc_ctrl   = 0x000020000ULL, // has control escape
+    _L_no_nl      = 0x000040000ULL, // no newline in pattern or matching text
+    _L_bzr_chr    = 0x000080000ULL, // \0 is a valid character constant
+    _L_grp_esc    = 0x000100000ULL, // \ is special character in group
+    _L_ident_ECMA = 0x000200000ULL, // ECMA identity escape (not identifierpart)
+    _L_ident_ERE  = 0x000400000ULL, // ERE identity escape (.[\*^$, plus {+?}()
+    _L_ident_awk  = 0x000800000ULL, // awk identity escape ( ERE plus "/)
+    _L_anch_rstr  = 0x001000000ULL, // anchor restricted to beginning/end
+    _L_star_beg   = 0x002000000ULL, // star okay at beginning of RE/expr (BRE)
+    _L_empty_grp  = 0x004000000ULL, // empty group allowed (ERE prohibits "()")
+    _L_paren_bal  = 0x008000000ULL, // ')'/'}' special only after '('/'{'
+    _L_brk_bal    = 0x010000000ULL, // ']' special only after '[' (ERE, BRE)
+    _L_brk_rstr   = 0x020000000ULL, // ']' not special when first character in set
+    _L_dsh_rstr   = 0x040000000ULL, // '-' forbidden at range start in set except when first character (ERE, BRE)
+    _L_mtch_long  = 0x080000000ULL, // find longest match (ERE, BRE)
 };
 
 class _Regex_base : public _Container_base { // base class for basic_regex to construct and destroy proxy

From 4f84d6db86340e05ef0d227d7de4028e262c240c Mon Sep 17 00:00:00 2001
From: "Stephan T. Lavavej" <stl@microsoft.com>
Date: Thu, 22 May 2025 06:12:48 -0700
Subject: [PATCH 20/21] Improve `_Lang_flags2` comments.

Use muellerj2's superior descriptions of `_L_alt_nl` and `_L_no_nl`.

Note that `_L_no_nl` is (grep, egrep).

Note that `_L_esc_oct` and `_L_esc_ffnx` are (awk).

`_L_esc_ffn` confusingly said "(\[fnrtv])" when other comments like `_L_ident_ERE` mean square brackets literally. Spell out "(\f \n \r \t \v)" for clarity and improved searchability.

Rephrase `_L_ident_awk`'s comment for clarity.

Note that `_L_anch_rstr` is (BRE) only, `_L_paren_bal` is (ERE) only, and `_L_brk_rstr` is (ERE, BRE).
---
 stl/inc/regex | 18 +++++++++---------
 1 file changed, 9 insertions(+), 9 deletions(-)

diff --git a/stl/inc/regex b/stl/inc/regex
index 5ea107724dc..045ca9b605e 100644
--- a/stl/inc/regex
+++ b/stl/inc/regex
@@ -1847,7 +1847,7 @@ private:
 enum _Lang_flags2 : unsigned long long { // describe language properties
     _L_ext_rep    = 0x000000001ULL, // + and ? repetitions
     _L_alt_pipe   = 0x000000002ULL, // uses '|' for alternation
-    _L_alt_nl     = 0x000000004ULL, // uses '\n' for alternation (grep, egrep)
+    _L_alt_nl     = 0x000000004ULL, // newlines outside parentheses/brackets represent alternations (grep, egrep)
     _L_nex_grp    = 0x000000008ULL, // has non-escaped capture groups
     _L_nex_rep    = 0x000000010ULL, // has non-escaped repeats
     _L_nc_asrt    = 0x000000020ULL, // has non-capture groups (?:xxx) and generalized assertions (?=xxx), (?!xxx)
@@ -1857,24 +1857,24 @@ enum _Lang_flags2 : unsigned long long { // describe language properties
     _L_ngr_rep    = 0x000000200ULL, // has non-greedy repeats
     _L_esc_uni    = 0x000000400ULL, // has Unicode escape sequences
     _L_esc_hex    = 0x000000800ULL, // has hexadecimal escape sequences
-    _L_esc_oct    = 0x000001000ULL, // has octal escape sequences
+    _L_esc_oct    = 0x000001000ULL, // has octal escape sequences (awk)
     _L_esc_bsp    = 0x000002000ULL, // has backspace escape in character classes
-    _L_esc_ffnx   = 0x000004000ULL, // has extra file escapes (\a and \b)
-    _L_esc_ffn    = 0x000008000ULL, // has limited file escapes (\[fnrtv])
+    _L_esc_ffnx   = 0x000004000ULL, // has extra file escapes (\a and \b) (awk)
+    _L_esc_ffn    = 0x000008000ULL, // has limited file escapes (\f \n \r \t \v)
     _L_esc_wsd    = 0x000010000ULL, // has w, s, and d character set escapes
     _L_esc_ctrl   = 0x000020000ULL, // has control escape
-    _L_no_nl      = 0x000040000ULL, // no newline in pattern or matching text
+    _L_no_nl      = 0x000040000ULL, // newlines within parentheses/brackets are syntax errors (grep, egrep)
     _L_bzr_chr    = 0x000080000ULL, // \0 is a valid character constant
     _L_grp_esc    = 0x000100000ULL, // \ is special character in group
     _L_ident_ECMA = 0x000200000ULL, // ECMA identity escape (not identifierpart)
     _L_ident_ERE  = 0x000400000ULL, // ERE identity escape (.[\*^$, plus {+?}()
-    _L_ident_awk  = 0x000800000ULL, // awk identity escape ( ERE plus "/)
-    _L_anch_rstr  = 0x001000000ULL, // anchor restricted to beginning/end
+    _L_ident_awk  = 0x000800000ULL, // awk identity escape (awk additionally supports " and / characters)
+    _L_anch_rstr  = 0x001000000ULL, // anchor restricted to beginning/end (BRE)
     _L_star_beg   = 0x002000000ULL, // star okay at beginning of RE/expr (BRE)
     _L_empty_grp  = 0x004000000ULL, // empty group allowed (ERE prohibits "()")
-    _L_paren_bal  = 0x008000000ULL, // ')'/'}' special only after '('/'{'
+    _L_paren_bal  = 0x008000000ULL, // ')'/'}' special only after '('/'{' (ERE)
     _L_brk_bal    = 0x010000000ULL, // ']' special only after '[' (ERE, BRE)
-    _L_brk_rstr   = 0x020000000ULL, // ']' not special when first character in set
+    _L_brk_rstr   = 0x020000000ULL, // ']' not special when first character in set (ERE, BRE)
     _L_dsh_rstr   = 0x040000000ULL, // '-' forbidden at range start in set except when first character (ERE, BRE)
     _L_mtch_long  = 0x080000000ULL, // find longest match (ERE, BRE)
 };

From e9eb1c67c1d1a73ad5a35a98f9ded0986f83ec43 Mon Sep 17 00:00:00 2001
From: "Stephan T. Lavavej" <stl@microsoft.com>
Date: Thu, 22 May 2025 06:21:45 -0700
Subject: [PATCH 21/21] Rename to `_L_non_greedy`.

---
 stl/inc/regex | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/stl/inc/regex b/stl/inc/regex
index 045ca9b605e..3e7e51b2f27 100644
--- a/stl/inc/regex
+++ b/stl/inc/regex
@@ -1854,7 +1854,7 @@ enum _Lang_flags2 : unsigned long long { // describe language properties
     _L_asrt_wrd   = 0x000000040ULL, // has word boundary assertions (\b, \B)
     _L_bckr       = 0x000000080ULL, // has backreferences (ERE doesn't)
     _L_lim_bckr   = 0x000000100ULL, // has limited backreferences (BRE \1-\9)
-    _L_ngr_rep    = 0x000000200ULL, // has non-greedy repeats
+    _L_non_greedy = 0x000000200ULL, // has non-greedy repeats
     _L_esc_uni    = 0x000000400ULL, // has Unicode escape sequences
     _L_esc_hex    = 0x000000800ULL, // has hexadecimal escape sequences
     _L_esc_oct    = 0x000001000ULL, // has octal escape sequences (awk)
@@ -4795,7 +4795,7 @@ void _Parser2<_FwdIt, _Elem, _RxTraits>::_Quantifier() { // check for quantifier
     }
 
     _Next();
-    const bool _Greedy = !(_L_flags & _L_ngr_rep) || _Mchar != _Meta_query;
+    const bool _Greedy = !(_L_flags & _L_non_greedy) || _Mchar != _Meta_query;
     if (!_Greedy) { // add non-greedy repeat node
         _Next();
     }
@@ -5008,9 +5008,9 @@ _Parser2<_FwdIt, _Elem, _RxTraits>::_Parser2(
     : _Pat(_Pfirst), _End(_Plast), _Nfa(_Tr, _Fx), _Traits(_Tr), _Flags(_Fx) {
 
     constexpr unsigned long long _ECMA_flags = _L_ext_rep | _L_alt_pipe | _L_nex_grp | _L_nex_rep | _L_nc_asrt
-                                             | _L_asrt_wrd | _L_bckr | _L_ngr_rep | _L_esc_uni | _L_esc_hex | _L_esc_bsp
-                                             | _L_esc_ffn | _L_esc_wsd | _L_esc_ctrl | _L_bzr_chr | _L_grp_esc
-                                             | _L_ident_ECMA | _L_empty_grp;
+                                             | _L_asrt_wrd | _L_bckr | _L_non_greedy | _L_esc_uni | _L_esc_hex
+                                             | _L_esc_bsp | _L_esc_ffn | _L_esc_wsd | _L_esc_ctrl | _L_bzr_chr
+                                             | _L_grp_esc | _L_ident_ECMA | _L_empty_grp;
 
     constexpr unsigned long long _Basic_flags = _L_bckr | _L_lim_bckr | _L_anch_rstr | _L_star_beg | _L_empty_grp
                                               | _L_brk_bal | _L_brk_rstr | _L_dsh_rstr | _L_mtch_long;