Skip to content

<regex>: Skip initial NFA nodes that do nothing during matching #6026

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
Stephan T. Lavavej (StephanTLavavej) merged 1 commit into from
Jan 20, 2026
Merged

<regex>: Skip initial NFA nodes that do nothing during matching #6026

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
29 changes: 17 additions & 12 deletions stl/inc/regex
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1700,7 +1700,7 @@ class _Matcher3 { // provides ways to match a regular expression to a text seque
public:
_Matcher3(_It _Pfirst, _It _Plast, const _RxTraits& _Tr, _Root_node* _Re, unsigned int _Nx,
regex_constants::syntax_option_type _Sf, regex_constants::match_flag_type _Mf)
: _Begin(_Pfirst), _End(_Plast), _Rep(_Re), _Sflags(_Sf), _Mflags(_Mf), _Ncap(_Nx),
: _Begin(_Pfirst), _End(_Plast), _Sflags(_Sf), _Mflags(_Mf), _Ncap(_Nx),
_Longest((_Re->_Flags & _Fl_longest) && !(_Mf & regex_constants::match_any)), _Traits(_Tr) {
_Loop_vals.resize(_Re->_Loops);
_Adl_verify_range(_Pfirst, _Plast);
Expand All @@ -1720,6 +1720,12 @@ public:
_Frames_limit = _Calculate_frames_limit(_Input_length);
_Complexity_limit = _Calculate_complexity_limit(_Input_length);

// TRANSITION, ABI, GH-6025:
// The first two nodes are of types _N_begin and _N_capture with capturing group 0.
// These nodes do not affect the state of the matcher and thus can be skipped immediately
// before engaging the expensive NFA interpreter loop.
_Start = _Re->_Next->_Next;

// sanitize multiline mode setting
#if _REGEX_LEGACY_MULTILINE_MODE
_Sflags |= regex_constants::multiline; // old matcher applied multiline mode for all grammars
Expand Down Expand Up @@ -1763,7 +1769,7 @@ public:

_Matched = false;

bool _Succeeded = _Match_pat(_Rep) || _Matched;
bool _Succeeded = _Match_pat(_Start) || _Matched;

if (!_Succeeded) {
return false;
Expand Down Expand Up @@ -1832,7 +1838,7 @@ private:

_It _Begin;
_It _End;
_Node_base* _Rep;
_Node_base* _Start;
regex_constants::syntax_option_type _Sflags;
regex_constants::match_flag_type _Mflags;
bool _Matched = false;
Expand Down Expand Up @@ -3997,14 +4003,13 @@ bool _Matcher3<_BidIt, _Elem, _RxTraits, _It, _Alloc>::_Match_pat(_Node_base* _N
{ // record current position
auto _Node = static_cast<_Node_capture*>(_Nx);
auto _Idx = _Node->_Idx;
if (_Idx != 0U) {
auto& _Group = _Tgt_state._Grps[_Idx];
auto _Frame_idx = _Push_frame(_Rx_unwind_ops::_Capture_restore_begin, _Node);
auto& _Frame = _Frames[_Frame_idx];
_Frame._Pos = _Group._Begin;
_Frame._Capture_idx = _Idx;
_Group._Begin = _Tgt_state._Cur;
}
_STL_INTERNAL_CHECK(_Idx != 0U);
auto& _Group = _Tgt_state._Grps[_Idx];
auto _Frame_idx = _Push_frame(_Rx_unwind_ops::_Capture_restore_begin, _Node);
auto& _Frame = _Frames[_Frame_idx];
_Frame._Pos = _Group._Begin;
_Frame._Capture_idx = _Idx;
_Group._Begin = _Tgt_state._Cur;
break;
}

Expand Down Expand Up @@ -4469,7 +4474,7 @@ _BidIt _Matcher3<_BidIt, _Elem, _RxTraits, _It, _Alloc>::_Skip(
static constexpr wchar_t _Line_terminators_wchar_t[] = {static_cast<wchar_t>(_Meta_cr),
static_cast<wchar_t>(_Meta_nl), static_cast<wchar_t>(_Meta_ls), static_cast<wchar_t>(_Meta_ps)};
constexpr unsigned int _Max_recursion_depth = 50U;
_Node_base* _Nx = _Node_arg ? _Node_arg : _Rep;
_Node_base* _Nx = _Node_arg ? _Node_arg : _Start;

while (_First_arg != _Last && _Nx) { // check current node
switch (_Nx->_Kind) { // handle current node's type
Expand Down