pammon created this revision.
Herald added a reviewer: EricWF.
In regex, forward assertions like '(?=stuff)' are implemented by
constructing a child regular expression 'stuff' and matching that.
If the child regular expression contains a backreference, this would
trip an assertion or reference the wrong capture group, because the
child was ignorant of the capture groups of its parent. For example,
/(x)(?=\1)/ would trip an assertion.
Address this by propagating submatches into the child, so that
backreferences reference the correct capture groups. This also allows us
to eliminate the mexp_ field, because the child expression shares the
entire submatch array with the parent.
https://reviews.llvm.org/D32635
Files:
include/regex
test/std/re/re.alg/re.alg.match/ecma.pass.cpp
Index: test/std/re/re.alg/re.alg.match/ecma.pass.cpp
===================================================================
--- test/std/re/re.alg/re.alg.match/ecma.pass.cpp
+++ test/std/re/re.alg/re.alg.match/ecma.pass.cpp
@@ -637,6 +637,22 @@
assert(m.str(0) == s);
}
{
+ std::cmatch m;
+ const char s[] = "abcabc";
+ assert(std::regex_match(s, m, std::regex("(.+)(?=\\1)(\\1)")));
+ assert(m.size() == 3);
+ assert(m.str(1) == "abc");
+ assert(m.str(2) == "abc");
+ }
+ {
+ std::cmatch m;
+ const char s[] = "aa";
+ assert(std::regex_match(s, m, std::regex("(a+)(?!\\1)(a*)")));
+ assert(m.size() == 3);
+ assert(m.str(1) == "aa");
+ assert(m.str(2) == "");
+ }
+ {
std::cmatch m;
const char s[] = "foobar";
assert(std::regex_match(s, m, std::regex("[^\\0]*")));
Index: include/regex
===================================================================
--- include/regex
+++ include/regex
@@ -2826,7 +2826,7 @@
void __push_end_marked_subexpression(unsigned);
void __push_empty();
void __push_word_boundary(bool);
- void __push_lookahead(const basic_regex&, bool, unsigned);
+ void __push_lookahead(basic_regex, bool);
template <class _Allocator>
bool
@@ -2843,6 +2843,7 @@
bool
__match_at_start_ecma(const _CharT* __first, const _CharT* __last,
match_results<const _CharT*, _Allocator>& __m,
+ const vector<sub_match<const _CharT *>> &incoming_sub_matches,
regex_constants::match_flag_type __flags, bool) const;
template <class _Allocator>
bool
@@ -2964,17 +2965,16 @@
typedef __owns_one_state<_CharT> base;
basic_regex<_CharT, _Traits> __exp_;
- unsigned __mexp_;
bool __invert_;
__lookahead(const __lookahead&);
__lookahead& operator=(const __lookahead&);
public:
typedef _VSTD::__state<_CharT> __state;
_LIBCPP_INLINE_VISIBILITY
- __lookahead(const basic_regex<_CharT, _Traits>& __exp, bool __invert, __node<_CharT>* __s, unsigned __mexp)
- : base(__s), __exp_(__exp), __mexp_(__mexp), __invert_(__invert) {}
+ __lookahead(basic_regex<_CharT, _Traits> __exp, bool __invert, __node<_CharT>* __s)
+ : base(__s), __exp_(move(__exp)), __invert_(__invert) {}
virtual void __exec(__state&) const;
};
@@ -2987,16 +2987,18 @@
__m.__init(1 + __exp_.mark_count(), __s.__current_, __s.__last_);
bool __matched = __exp_.__match_at_start_ecma(
__s.__current_, __s.__last_,
- __m,
+ __m, __s.__sub_matches_,
(__s.__flags_ | regex_constants::match_continuous) &
~regex_constants::__full_match,
__s.__at_first_ && __s.__current_ == __s.__first_);
if (__matched != __invert_)
{
__s.__do_ = __state::__accept_but_not_consume;
__s.__node_ = this->first();
- for (unsigned __i = 1; __i < __m.size(); ++__i) {
- __s.__sub_matches_[__mexp_ + __i - 1] = __m.__matches_[__i];
+ if (__matched) {
+ for (unsigned __i = 1; __i < __m.size(); ++__i) {
+ __s.__sub_matches_[__i - 1] = __m.__matches_[__i];
+ }
}
}
else
@@ -4168,26 +4170,16 @@
switch (*__temp)
{
case '=':
- {
- basic_regex __exp;
- __exp.__flags_ = __flags_;
- __temp = __exp.__parse(++__temp, __last);
- unsigned __mexp = __exp.__marked_count_;
- __push_lookahead(_VSTD::move(__exp), false, __marked_count_);
- __marked_count_ += __mexp;
- if (__temp == __last || *__temp != ')')
- __throw_regex_error<regex_constants::error_paren>();
- __first = ++__temp;
- }
- break;
case '!':
{
+ bool __invert = (*__temp == '!');
basic_regex __exp;
__exp.__flags_ = __flags_;
+ __exp.__marked_count_ = __marked_count_;
__temp = __exp.__parse(++__temp, __last);
unsigned __mexp = __exp.__marked_count_;
- __push_lookahead(_VSTD::move(__exp), true, __marked_count_);
- __marked_count_ += __mexp;
+ __push_lookahead(_VSTD::move(__exp), __invert);
+ __marked_count_ = __mexp;
if (__temp == __last || *__temp != ')')
__throw_regex_error<regex_constants::error_paren>();
__first = ++__temp;
@@ -4749,12 +4741,11 @@
template <class _CharT, class _Traits>
void
-basic_regex<_CharT, _Traits>::__push_lookahead(const basic_regex& __exp,
- bool __invert,
- unsigned __mexp)
+basic_regex<_CharT, _Traits>::__push_lookahead(basic_regex __exp,
+ bool __invert)
{
- __end_->first() = new __lookahead<_CharT, _Traits>(__exp, __invert,
- __end_->first(), __mexp);
+ __end_->first() = new __lookahead<_CharT, _Traits>(move(__exp), __invert,
+ __end_->first());
__end_ = static_cast<__owns_one_state<_CharT>*>(__end_->first());
}
@@ -5529,6 +5520,7 @@
basic_regex<_CharT, _Traits>::__match_at_start_ecma(
const _CharT* __first, const _CharT* __last,
match_results<const _CharT*, _Allocator>& __m,
+ const vector<sub_match<const _CharT *>> &incoming_sub_matches,
regex_constants::match_flag_type __flags, bool __at_first) const
{
vector<__state> __states;
@@ -5539,13 +5531,15 @@
__unmatched.first = __last;
__unmatched.second = __last;
__unmatched.matched = false;
+ vector<sub_match<const _CharT *>> sub_matches = incoming_sub_matches;
+ sub_matches.resize(mark_count(), __unmatched);
__states.push_back(__state());
__states.back().__do_ = 0;
__states.back().__first_ = __first;
__states.back().__current_ = __first;
__states.back().__last_ = __last;
- __states.back().__sub_matches_.resize(mark_count(), __unmatched);
+ __states.back().__sub_matches_ = move(sub_matches);
__states.back().__loop_data_.resize(__loop_count());
__states.back().__node_ = __st;
__states.back().__flags_ = __flags;
@@ -5798,7 +5792,7 @@
regex_constants::match_flag_type __flags, bool __at_first) const
{
if ((__flags_ & 0x1F0) == ECMAScript)
- return __match_at_start_ecma(__first, __last, __m, __flags, __at_first);
+ return __match_at_start_ecma(__first, __last, __m, {}, __flags, __at_first);
if (mark_count() == 0)
return __match_at_start_posix_nosubs(__first, __last, __m, __flags, __at_first);
return __match_at_start_posix_subs(__first, __last, __m, __flags, __at_first);
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
