Skip to content

<regex>: Mark more loops as simple #5889

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
merged 1 commit into from
Nov 27, 2025
+12 −22
Merged

<regex>: Mark more loops as simple #5889

Show file tree
Hide file tree
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
26 changes: 11 additions & 15 deletions stl/inc/regex
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4102,16 +4102,12 @@ bool _Matcher3<_BidIt, _Elem, _RxTraits, _It, _Alloc>::_Match_pat(_Node_base* _N
}
} else if (_Sav._Loop_idx < _Nr->_Min) { // at least one more rep to reach minimum
_Next = _Nr->_Next;
// GH-5365: We have to reset the capture groups from the second iteration on.
_Tgt_state._Grp_valid = _Frames[_Sav._Loop_frame_idx]._Match_state._Grp_valid;
++_Sav._Loop_idx;
} else if (_Greedy && !_Longest && _Sav._Loop_idx != _Nr->_Max) { // one more rep to try next
// set up stack unwinding for greedy matching
_Push_frame(_Rx_unwind_ops::_Loop_simple_greedy, _Nr);

_Next = _Nr->_Next;
// GH-5365: We have to reset the capture groups from the second iteration on.
_Tgt_state._Grp_valid = _Frames[_Sav._Loop_frame_idx]._Match_state._Grp_valid;
if (_Sav._Loop_idx < INT_MAX) { // avoid overflowing _Loop_idx
++_Sav._Loop_idx;
}
Expand Down Expand Up @@ -4294,12 +4290,11 @@ bool _Matcher3<_BidIt, _Elem, _RxTraits, _It, _Alloc>::_Match_pat(_Node_base* _N
// try tail if matching one more rep failed
if (_Failed) {
auto _Node = static_cast<_Node_rep*>(_Frame._Node);
auto& _Sav = _Loop_vals[_Node->_Loop_number];

_Increase_complexity_count();
_Nx = _Node->_End_rep->_Next;
_Tgt_state._Cur = _Frame._Match_state._Cur;
_Tgt_state._Grp_valid = _Frames[_Sav._Loop_frame_idx]._Match_state._Grp_valid;
_Tgt_state._Grp_valid = _Frame._Match_state._Grp_valid;
_Failed = false;
}
break;
Expand Down Expand Up @@ -5356,14 +5351,21 @@ void _Parser2<_FwdIt, _Elem, _RxTraits>::_Calculate_loop_simplicity(
for (_Node_if* _Branch = static_cast<_Node_if*>(_Nx)->_Child; _Branch; _Branch = _Branch->_Child) {
_Calculate_loop_simplicity(_Branch->_Next, _Branch->_Endif, _Outer_rep);
}

break;

case _N_assert:
// A positive lookahead assertion inside a _Node_rep makes the rep not simple
if (_Outer_rep) {
_Outer_rep->_Simple_loop = 0;
}
_FALLTHROUGH;

case _N_neg_assert:
// visit the assertion body
// note _Outer_rep being reset: the assertion regex is completely independent
_Calculate_loop_simplicity(static_cast<_Node_assert*>(_Nx)->_Child, nullptr, nullptr);
break;

case _N_rep:
// _Node_rep inside another _Node_rep makes both not simple if _Outer_rep can be repeated more than once
// because the matcher does not reset capture group boundaries when handling simple loops.
Expand All @@ -5381,6 +5383,7 @@ void _Parser2<_FwdIt, _Elem, _RxTraits>::_Calculate_loop_simplicity(
_Outer_rep = static_cast<_Node_rep*>(_Nx);
}
break;

case _N_end_rep:
if (_Outer_rep == static_cast<_Node_end_rep*>(_Nx)->_Begin_rep) {
// if the _Node_rep is still undetermined when we reach its end, it is simple
Expand All @@ -5391,6 +5394,7 @@ void _Parser2<_FwdIt, _Elem, _RxTraits>::_Calculate_loop_simplicity(
_Outer_rep = nullptr;
}
break;

case _N_class:
if (_Outer_rep) {
// _Node_rep is not simple if a class can match character sequences of different lengths
Expand All @@ -5407,14 +5411,6 @@ void _Parser2<_FwdIt, _Elem, _RxTraits>::_Calculate_loop_simplicity(

case _N_group:
case _N_capture:
// TRANSITION, requires more research to decide on the subset of loops that we can make simple:
// - Simple mode can square the running time when matching a regex to an input string in the current matcher
// - The optimal subset of simple loops for a non-recursive rewrite of the matcher aren't clear yet
if (_Outer_rep) {
_Outer_rep->_Simple_loop = 0;
}
break;

case _N_none:
case _N_nop:
case _N_bol:
Expand Down
8 changes: 1 addition & 7 deletions tests/std/tests/VSO_0000000_regex_use/test.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -902,13 +902,7 @@ void test_gh_993() {
void test_gh_997() {
// GH-997: <regex>: Grouping within repetition causes regex stack error
// GH-1528: <regex>: regex_match gets caught in recursive loop until stack overflow occurs

try {
(void) regex_match(string(1025, 'a'), regex("(?:a)+"));
assert(false); // adjust test when matching succeeds
} catch (const regex_error& ex) {
assert(ex.code() == error_stack);
}
g_regexTester.should_match(string(1025, 'a'), "(?:a)+");

{
test_wregex rgx(&g_regexTester, LR"(^http[s]?://([^.]+\.)*example\.com/.*$)", icase);
Expand Down