grpc 1.34.0 → 1.42.0

data/third_party/re2/re2/re2.h

@@ -40,6 +40,9 @@
 //   R"((?i)hello)"          -- (?i) turns on case-insensitive matching
 //   R"(/\*(.*?)\*/)"        -- .*? matches . minimum no. of times possible
 //
+// When using UTF-8 encoding, case-insensitive matching will perform
+// simple case folding, not full case folding.
+//
 // -----------------------------------------------------------------------
 // MATCHING INTERFACE:
 //
@@ -205,6 +208,7 @@
 #include <map>
 #include <mutex>
 #include <string>
+#include <type_traits>
 #include <vector>
 
 #if defined(__APPLE__)
@@ -244,6 +248,7 @@ class RE2 {
     ErrorBadCharRange,       // bad character class range
     ErrorMissingBracket,     // missing closing ]
     ErrorMissingParen,       // missing closing )
+    ErrorUnexpectedParen,    // unexpected closing )
     ErrorTrailingBackslash,  // trailing \ at end of regexp
     ErrorRepeatArgument,     // repeat argument missing, e.g. "*"
     ErrorRepeatSize,         // bad repetition argument
@@ -364,12 +369,12 @@ class RE2 {
   //    (void*)NULL     (the corresponding matched sub-pattern is not copied)
   //
   // Returns true iff all of the following conditions are satisfied:
-  //   a. "text" matches "re" exactly
-  //   b. The number of matched sub-patterns is >= number of supplied pointers
+  //   a. "text" matches "re" fully - from the beginning to the end of "text".
+  //   b. The number of matched sub-patterns is >= number of supplied pointers.
   //   c. The "i"th argument has a suitable type for holding the
   //      string captured as the "i"th sub-pattern.  If you pass in
   //      NULL for the "i"th argument, or pass fewer arguments than
-  //      number of sub-patterns, "i"th captured sub-pattern is
+  //      number of sub-patterns, the "i"th captured sub-pattern is
   //      ignored.
   //
   // CAVEAT: An optional sub-pattern that does not exist in the
@@ -383,8 +388,17 @@ class RE2 {
     return Apply(FullMatchN, text, re, Arg(std::forward<A>(a))...);
   }
 
-  // Exactly like FullMatch(), except that "re" is allowed to match
-  // a substring of "text".
+  // Like FullMatch(), except that "re" is allowed to match a substring
+  // of "text".
+  //
+  // Returns true iff all of the following conditions are satisfied:
+  //   a. "text" matches "re" partially - for some substring of "text".
+  //   b. The number of matched sub-patterns is >= number of supplied pointers.
+  //   c. The "i"th argument has a suitable type for holding the
+  //      string captured as the "i"th sub-pattern.  If you pass in
+  //      NULL for the "i"th argument, or pass fewer arguments than
+  //      number of sub-patterns, the "i"th captured sub-pattern is
+  //      ignored.
   template <typename... A>
   static bool PartialMatch(const StringPiece& text, const RE2& re, A&&... a) {
     return Apply(PartialMatchN, text, re, Arg(std::forward<A>(a))...);
@@ -393,7 +407,16 @@ class RE2 {
   // Like FullMatch() and PartialMatch(), except that "re" has to match
   // a prefix of the text, and "input" is advanced past the matched
   // text.  Note: "input" is modified iff this routine returns true
-  // and "re" matched a non-empty substring of "text".
+  // and "re" matched a non-empty substring of "input".
+  //
+  // Returns true iff all of the following conditions are satisfied:
+  //   a. "input" matches "re" partially - for some prefix of "input".
+  //   b. The number of matched sub-patterns is >= number of supplied pointers.
+  //   c. The "i"th argument has a suitable type for holding the
+  //      string captured as the "i"th sub-pattern.  If you pass in
+  //      NULL for the "i"th argument, or pass fewer arguments than
+  //      number of sub-patterns, the "i"th captured sub-pattern is
+  //      ignored.
   template <typename... A>
   static bool Consume(StringPiece* input, const RE2& re, A&&... a) {
     return Apply(ConsumeN, input, re, Arg(std::forward<A>(a))...);
@@ -403,6 +426,15 @@ class RE2 {
   // the text.  That is, "re" need not start its match at the beginning
   // of "input".  For example, "FindAndConsume(s, "(\\w+)", &word)" finds
   // the next word in "s" and stores it in "word".
+  //
+  // Returns true iff all of the following conditions are satisfied:
+  //   a. "input" matches "re" partially - for some substring of "input".
+  //   b. The number of matched sub-patterns is >= number of supplied pointers.
+  //   c. The "i"th argument has a suitable type for holding the
+  //      string captured as the "i"th sub-pattern.  If you pass in
+  //      NULL for the "i"th argument, or pass fewer arguments than
+  //      number of sub-patterns, the "i"th captured sub-pattern is
+  //      ignored.
   template <typename... A>
   static bool FindAndConsume(StringPiece* input, const RE2& re, A&&... a) {
     return Apply(FindAndConsumeN, input, re, Arg(std::forward<A>(a))...);
@@ -703,32 +735,12 @@ class RE2 {
   const Options& options() const { return options_; }
 
   // Argument converters; see below.
-  static inline Arg CRadix(short* x);
-  static inline Arg CRadix(unsigned short* x);
-  static inline Arg CRadix(int* x);
-  static inline Arg CRadix(unsigned int* x);
-  static inline Arg CRadix(long* x);
-  static inline Arg CRadix(unsigned long* x);
-  static inline Arg CRadix(long long* x);
-  static inline Arg CRadix(unsigned long long* x);
-
-  static inline Arg Hex(short* x);
-  static inline Arg Hex(unsigned short* x);
-  static inline Arg Hex(int* x);
-  static inline Arg Hex(unsigned int* x);
-  static inline Arg Hex(long* x);
-  static inline Arg Hex(unsigned long* x);
-  static inline Arg Hex(long long* x);
-  static inline Arg Hex(unsigned long long* x);
-
-  static inline Arg Octal(short* x);
-  static inline Arg Octal(unsigned short* x);
-  static inline Arg Octal(int* x);
-  static inline Arg Octal(unsigned int* x);
-  static inline Arg Octal(long* x);
-  static inline Arg Octal(unsigned long* x);
-  static inline Arg Octal(long long* x);
-  static inline Arg Octal(unsigned long long* x);
+  template <typename T>
+  static Arg CRadix(T* ptr);
+  template <typename T>
+  static Arg Hex(T* ptr);
+  template <typename T>
+  static Arg Octal(T* ptr);
 
  private:
   void Init(const StringPiece& pattern, const Options& options);
@@ -771,130 +783,130 @@ class RE2 {
 
 /***** Implementation details *****/
 
-// Hex/Octal/Binary?
-
-// Special class for parsing into objects that define a ParseFrom() method
-template <class T>
-class _RE2_MatchObject {
- public:
-  static inline bool Parse(const char* str, size_t n, void* dest) {
-    if (dest == NULL) return true;
-    T* object = reinterpret_cast<T*>(dest);
-    return object->ParseFrom(str, n);
-  }
-};
+namespace re2_internal {
+
+// Types for which the 3-ary Parse() function template has specializations.
+template <typename T> struct Parse3ary : public std::false_type {};
+template <> struct Parse3ary<void> : public std::true_type {};
+template <> struct Parse3ary<std::string> : public std::true_type {};
+template <> struct Parse3ary<StringPiece> : public std::true_type {};
+template <> struct Parse3ary<char> : public std::true_type {};
+template <> struct Parse3ary<signed char> : public std::true_type {};
+template <> struct Parse3ary<unsigned char> : public std::true_type {};
+template <> struct Parse3ary<float> : public std::true_type {};
+template <> struct Parse3ary<double> : public std::true_type {};
+
+template <typename T>
+bool Parse(const char* str, size_t n, T* dest);
+
+// Types for which the 4-ary Parse() function template has specializations.
+template <typename T> struct Parse4ary : public std::false_type {};
+template <> struct Parse4ary<long> : public std::true_type {};
+template <> struct Parse4ary<unsigned long> : public std::true_type {};
+template <> struct Parse4ary<short> : public std::true_type {};
+template <> struct Parse4ary<unsigned short> : public std::true_type {};
+template <> struct Parse4ary<int> : public std::true_type {};
+template <> struct Parse4ary<unsigned int> : public std::true_type {};
+template <> struct Parse4ary<long long> : public std::true_type {};
+template <> struct Parse4ary<unsigned long long> : public std::true_type {};
+
+template <typename T>
+bool Parse(const char* str, size_t n, T* dest, int radix);
+
+}  // namespace re2_internal
 
 class RE2::Arg {
+ private:
+  template <typename T>
+  using CanParse3ary = typename std::enable_if<
+      re2_internal::Parse3ary<T>::value,
+      int>::type;
+
+  template <typename T>
+  using CanParse4ary = typename std::enable_if<
+      re2_internal::Parse4ary<T>::value,
+      int>::type;
+
+#if !defined(_MSC_VER)
+  template <typename T>
+  using CanParseFrom = typename std::enable_if<
+      std::is_member_function_pointer<
+          decltype(static_cast<bool (T::*)(const char*, size_t)>(
+              &T::ParseFrom))>::value,
+      int>::type;
+#endif
+
  public:
-  // Empty constructor so we can declare arrays of RE2::Arg
-  Arg();
+  Arg() : Arg(nullptr) {}
+  Arg(std::nullptr_t ptr) : arg_(ptr), parser_(DoNothing) {}
 
-  // Constructor specially designed for NULL arguments
-  Arg(void*);
-  Arg(std::nullptr_t);
+  template <typename T, CanParse3ary<T> = 0>
+  Arg(T* ptr) : arg_(ptr), parser_(DoParse3ary<T>) {}
+
+  template <typename T, CanParse4ary<T> = 0>
+  Arg(T* ptr) : arg_(ptr), parser_(DoParse4ary<T>) {}
+
+#if !defined(_MSC_VER)
+  template <typename T, CanParseFrom<T> = 0>
+  Arg(T* ptr) : arg_(ptr), parser_(DoParseFrom<T>) {}
+#endif
 
   typedef bool (*Parser)(const char* str, size_t n, void* dest);
 
-// Type-specific parsers
-#define MAKE_PARSER(type, name)            \
-  Arg(type* p) : arg_(p), parser_(name) {} \
-  Arg(type* p, Parser parser) : arg_(p), parser_(parser) {}
-
-  MAKE_PARSER(char,               parse_char)
-  MAKE_PARSER(signed char,        parse_schar)
-  MAKE_PARSER(unsigned char,      parse_uchar)
-  MAKE_PARSER(float,              parse_float)
-  MAKE_PARSER(double,             parse_double)
-  MAKE_PARSER(std::string,        parse_string)
-  MAKE_PARSER(StringPiece,        parse_stringpiece)
-
-  MAKE_PARSER(short,              parse_short)
-  MAKE_PARSER(unsigned short,     parse_ushort)
-  MAKE_PARSER(int,                parse_int)
-  MAKE_PARSER(unsigned int,       parse_uint)
-  MAKE_PARSER(long,               parse_long)
-  MAKE_PARSER(unsigned long,      parse_ulong)
-  MAKE_PARSER(long long,          parse_longlong)
-  MAKE_PARSER(unsigned long long, parse_ulonglong)
-
-#undef MAKE_PARSER
-
-  // Generic constructor templates
-  template <class T> Arg(T* p)
-      : arg_(p), parser_(_RE2_MatchObject<T>::Parse) { }
-  template <class T> Arg(T* p, Parser parser)
-      : arg_(p), parser_(parser) { }
-
-  // Parse the data
-  bool Parse(const char* str, size_t n) const;
+  template <typename T>
+  Arg(T* ptr, Parser parser) : arg_(ptr), parser_(parser) {}
+
+  bool Parse(const char* str, size_t n) const {
+    return (*parser_)(str, n, arg_);
+  }
 
  private:
-  void*         arg_;
-  Parser        parser_;
+  static bool DoNothing(const char* /*str*/, size_t /*n*/, void* /*dest*/) {
+    return true;
+  }
 
-  static bool parse_null          (const char* str, size_t n, void* dest);
-  static bool parse_char          (const char* str, size_t n, void* dest);
-  static bool parse_schar         (const char* str, size_t n, void* dest);
-  static bool parse_uchar         (const char* str, size_t n, void* dest);
-  static bool parse_float         (const char* str, size_t n, void* dest);
-  static bool parse_double        (const char* str, size_t n, void* dest);
-  static bool parse_string        (const char* str, size_t n, void* dest);
-  static bool parse_stringpiece   (const char* str, size_t n, void* dest);
-
-#define DECLARE_INTEGER_PARSER(name)                                       \
- private:                                                                  \
-  static bool parse_##name(const char* str, size_t n, void* dest);         \
-  static bool parse_##name##_radix(const char* str, size_t n, void* dest,  \
-                                   int radix);                             \
-                                                                           \
- public:                                                                   \
-  static bool parse_##name##_hex(const char* str, size_t n, void* dest);   \
-  static bool parse_##name##_octal(const char* str, size_t n, void* dest); \
-  static bool parse_##name##_cradix(const char* str, size_t n, void* dest);
-
-  DECLARE_INTEGER_PARSER(short)
-  DECLARE_INTEGER_PARSER(ushort)
-  DECLARE_INTEGER_PARSER(int)
-  DECLARE_INTEGER_PARSER(uint)
-  DECLARE_INTEGER_PARSER(long)
-  DECLARE_INTEGER_PARSER(ulong)
-  DECLARE_INTEGER_PARSER(longlong)
-  DECLARE_INTEGER_PARSER(ulonglong)
-
-#undef DECLARE_INTEGER_PARSER
+  template <typename T>
+  static bool DoParse3ary(const char* str, size_t n, void* dest) {
+    return re2_internal::Parse(str, n, reinterpret_cast<T*>(dest));
+  }
 
-};
+  template <typename T>
+  static bool DoParse4ary(const char* str, size_t n, void* dest) {
+    return re2_internal::Parse(str, n, reinterpret_cast<T*>(dest), 10);
+  }
 
-inline RE2::Arg::Arg() : arg_(NULL), parser_(parse_null) { }
-inline RE2::Arg::Arg(void* p) : arg_(p), parser_(parse_null) { }
-inline RE2::Arg::Arg(std::nullptr_t p) : arg_(p), parser_(parse_null) { }
+#if !defined(_MSC_VER)
+  template <typename T>
+  static bool DoParseFrom(const char* str, size_t n, void* dest) {
+    if (dest == NULL) return true;
+    return reinterpret_cast<T*>(dest)->ParseFrom(str, n);
+  }
+#endif
 
-inline bool RE2::Arg::Parse(const char* str, size_t n) const {
-  return (*parser_)(str, n, arg_);
-}
+  void*         arg_;
+  Parser        parser_;
+};
 
-// This part of the parser, appropriate only for ints, deals with bases
-#define MAKE_INTEGER_PARSER(type, name)                    \
-  inline RE2::Arg RE2::Hex(type* ptr) {                    \
-    return RE2::Arg(ptr, RE2::Arg::parse_##name##_hex);    \
-  }                                                        \
-  inline RE2::Arg RE2::Octal(type* ptr) {                  \
-    return RE2::Arg(ptr, RE2::Arg::parse_##name##_octal);  \
-  }                                                        \
-  inline RE2::Arg RE2::CRadix(type* ptr) {                 \
-    return RE2::Arg(ptr, RE2::Arg::parse_##name##_cradix); \
-  }
+template <typename T>
+inline RE2::Arg RE2::CRadix(T* ptr) {
+  return RE2::Arg(ptr, [](const char* str, size_t n, void* dest) -> bool {
+    return re2_internal::Parse(str, n, reinterpret_cast<T*>(dest), 0);
+  });
+}
 
-MAKE_INTEGER_PARSER(short,              short)
-MAKE_INTEGER_PARSER(unsigned short,     ushort)
-MAKE_INTEGER_PARSER(int,                int)
-MAKE_INTEGER_PARSER(unsigned int,       uint)
-MAKE_INTEGER_PARSER(long,               long)
-MAKE_INTEGER_PARSER(unsigned long,      ulong)
-MAKE_INTEGER_PARSER(long long,          longlong)
-MAKE_INTEGER_PARSER(unsigned long long, ulonglong)
+template <typename T>
+inline RE2::Arg RE2::Hex(T* ptr) {
+  return RE2::Arg(ptr, [](const char* str, size_t n, void* dest) -> bool {
+    return re2_internal::Parse(str, n, reinterpret_cast<T*>(dest), 16);
+  });
+}
 
-#undef MAKE_INTEGER_PARSER
+template <typename T>
+inline RE2::Arg RE2::Octal(T* ptr) {
+  return RE2::Arg(ptr, [](const char* str, size_t n, void* dest) -> bool {
+    return re2_internal::Parse(str, n, reinterpret_cast<T*>(dest), 8);
+  });
+}
 
 #ifndef SWIG
 // Silence warnings about missing initializers for members of LazyRE2.
@@ -955,8 +967,11 @@ namespace hooks {
 // thread_local, but for the sake of brevity, we lump together all versions
 // of Apple platforms that aren't macOS. If an iOS application really needs
 // the context pointee someday, we can get more specific then...
+//
+// As per https://github.com/google/re2/issues/325, thread_local support in
+// MinGW seems to be buggy. (FWIW, Abseil folks also avoid it.)
 #define RE2_HAVE_THREAD_LOCAL
-#if defined(__APPLE__) && !TARGET_OS_OSX
+#if (defined(__APPLE__) && !TARGET_OS_OSX) || defined(__MINGW32__)
 #undef RE2_HAVE_THREAD_LOCAL
 #endif
 

data/third_party/re2/re2/regexp.cc

@@ -498,6 +498,7 @@ static const char *kErrorStrings[] = {
   "invalid character class range",
   "missing ]",
   "missing )",
+  "unexpected )",
   "trailing \\",
   "no argument for repetition operator",
   "invalid repetition size",
@@ -721,8 +722,14 @@ bool Regexp::RequiredPrefixForAccel(std::string* prefix, bool* foldcase) {
   *foldcase = false;
 
   // No need for a walker: the regexp must either begin with or be
-  // a literal char or string.
+  // a literal char or string. We "see through" capturing groups,
+  // but make no effort to glue multiple prefix fragments together.
   Regexp* re = op_ == kRegexpConcat && nsub_ > 0 ? sub()[0] : this;
+  while (re->op_ == kRegexpCapture) {
+    re = re->sub()[0];
+    if (re->op_ == kRegexpConcat && re->nsub_ > 0)
+      re = re->sub()[0];
+  }
   if (re->op_ != kRegexpLiteral &&
       re->op_ != kRegexpLiteralString)
     return false;
@@ -912,7 +919,7 @@ void CharClassBuilder::Negate() {
 // The ranges are allocated in the same block as the header,
 // necessitating a special allocator and Delete method.
 
-CharClass* CharClass::New(int maxranges) {
+CharClass* CharClass::New(size_t maxranges) {
   CharClass* cc;
   uint8_t* data = new uint8_t[sizeof *cc + maxranges*sizeof cc->ranges_[0]];
   cc = reinterpret_cast<CharClass*>(data);
@@ -929,7 +936,7 @@ void CharClass::Delete() {
 }
 
 CharClass* CharClass::Negate() {
-  CharClass* cc = CharClass::New(nranges_+1);
+  CharClass* cc = CharClass::New(static_cast<size_t>(nranges_+1));
   cc->folds_ascii_ = folds_ascii_;
   cc->nrunes_ = Runemax + 1 - nrunes_;
   int n = 0;
@@ -948,7 +955,7 @@ CharClass* CharClass::Negate() {
   return cc;
 }
 
-bool CharClass::Contains(Rune r) {
+bool CharClass::Contains(Rune r) const {
   RuneRange* rr = ranges_;
   int n = nranges_;
   while (n > 0) {
@@ -966,7 +973,7 @@ bool CharClass::Contains(Rune r) {
 }
 
 CharClass* CharClassBuilder::GetCharClass() {
-  CharClass* cc = CharClass::New(static_cast<int>(ranges_.size()));
+  CharClass* cc = CharClass::New(ranges_.size());
   int n = 0;
   for (iterator it = begin(); it != end(); ++it)
     cc->ranges_[n++] = *it;

data/third_party/re2/re2/regexp.h

@@ -86,6 +86,7 @@
 // form accessible to clients, so that client code can analyze the
 // parsed regular expressions.
 
+#include <stddef.h>
 #include <stdint.h>
 #include <map>
 #include <set>
@@ -177,6 +178,7 @@ enum RegexpStatusCode {
   kRegexpBadCharRange,       // bad character class range
   kRegexpMissingBracket,     // missing closing ]
   kRegexpMissingParen,       // missing closing )
+  kRegexpUnexpectedParen,    // unexpected closing )
   kRegexpTrailingBackslash,  // at end of regexp
   kRegexpRepeatArgument,     // repeat argument missing, e.g. "*"
   kRegexpRepeatSize,         // bad repetition argument
@@ -252,13 +254,13 @@ class CharClass {
   bool full() { return nrunes_ == Runemax+1; }
   bool FoldsASCII() { return folds_ascii_; }
 
-  bool Contains(Rune r);
+  bool Contains(Rune r) const;
   CharClass* Negate();
 
  private:
   CharClass();  // not implemented
   ~CharClass();  // not implemented
-  static CharClass* New(int maxranges);
+  static CharClass* New(size_t maxranges);
 
   friend class CharClassBuilder;
 
@@ -447,6 +449,10 @@ class Regexp {
   // regardless of the return value.
   bool RequiredPrefixForAccel(std::string* prefix, bool* foldcase);
 
+  // Controls the maximum repeat count permitted by the parser.
+  // FOR FUZZING ONLY.
+  static void FUZZING_ONLY_set_maximum_repeat_count(int i);
+
  private:
   // Constructor allocates vectors as appropriate for operator.
   explicit Regexp(RegexpOp op, ParseFlags parse_flags);

data/third_party/re2/re2/set.cc

@@ -7,6 +7,7 @@
 #include <stddef.h>
 #include <algorithm>
 #include <memory>
+#include <utility>
 
 #include "util/util.h"
 #include "util/logging.h"
@@ -18,19 +19,37 @@
 
 namespace re2 {
 
-RE2::Set::Set(const RE2::Options& options, RE2::Anchor anchor) {
-  options_.Copy(options);
+RE2::Set::Set(const RE2::Options& options, RE2::Anchor anchor)
+    : options_(options),
+      anchor_(anchor),
+      compiled_(false),
+      size_(0) {
   options_.set_never_capture(true);  // might unblock some optimisations
-  anchor_ = anchor;
-  prog_ = NULL;
-  compiled_ = false;
-  size_ = 0;
 }
 
 RE2::Set::~Set() {
   for (size_t i = 0; i < elem_.size(); i++)
     elem_[i].second->Decref();
-  delete prog_;
+}
+
+RE2::Set::Set(Set&& other)
+    : options_(other.options_),
+      anchor_(other.anchor_),
+      elem_(std::move(other.elem_)),
+      compiled_(other.compiled_),
+      size_(other.size_),
+      prog_(std::move(other.prog_)) {
+  other.elem_.clear();
+  other.elem_.shrink_to_fit();
+  other.compiled_ = false;
+  other.size_ = 0;
+  other.prog_.reset();
+}
+
+RE2::Set& RE2::Set::operator=(Set&& other) {
+  this->~Set();
+  (void) new (this) Set(std::move(other));
+  return *this;
 }
 
 int RE2::Set::Add(const StringPiece& pattern, std::string* error) {
@@ -97,9 +116,9 @@ bool RE2::Set::Compile() {
     options_.ParseFlags());
   re2::Regexp* re = re2::Regexp::Alternate(sub.data(), size_, pf);
 
-  prog_ = Prog::CompileSet(re, anchor_, options_.max_mem());
+  prog_.reset(Prog::CompileSet(re, anchor_, options_.max_mem()));
   re->Decref();
-  return prog_ != NULL;
+  return prog_ != nullptr;
 }
 
 bool RE2::Set::Match(const StringPiece& text, std::vector<int>* v) const {
@@ -114,6 +133,9 @@ bool RE2::Set::Match(const StringPiece& text, std::vector<int>* v,
       error_info->kind = kNotCompiled;
     return false;
   }
+#ifdef RE2_HAVE_THREAD_LOCAL
+  hooks::context = NULL;
+#endif
   bool dfa_failed = false;
   std::unique_ptr<SparseSet> matches;
   if (v != NULL) {

data/third_party/re2/re2/set.h

@@ -5,6 +5,7 @@
 #ifndef RE2_SET_H_
 #define RE2_SET_H_
 
+#include <memory>
 #include <string>
 #include <utility>
 #include <vector>
@@ -36,6 +37,13 @@ class RE2::Set {
   Set(const RE2::Options& options, RE2::Anchor anchor);
   ~Set();
 
+  // Not copyable.
+  Set(const Set&) = delete;
+  Set& operator=(const Set&) = delete;
+  // Movable.
+  Set(Set&& other);
+  Set& operator=(Set&& other);
+
   // Adds pattern to the set using the options passed to the constructor.
   // Returns the index that will identify the regexp in the output of Match(),
   // or -1 if the regexp cannot be parsed.
@@ -67,12 +75,9 @@ class RE2::Set {
   RE2::Options options_;
   RE2::Anchor anchor_;
   std::vector<Elem> elem_;
-  re2::Prog* prog_;
   bool compiled_;
   int size_;
-
-  Set(const Set&) = delete;
-  Set& operator=(const Set&) = delete;
+  std::unique_ptr<re2::Prog> prog_;
 };
 
 }  // namespace re2

data/third_party/re2/re2/simplify.cc

@@ -28,8 +28,6 @@ bool Regexp::SimplifyRegexp(const StringPiece& src, ParseFlags flags,
   Regexp* sre = re->Simplify();
   re->Decref();
   if (sre == NULL) {
-    // Should not happen, since Simplify never fails.
-    LOG(ERROR) << "Simplify failed on " << src;
     if (status) {
       status->set_code(kRegexpInternalError);
       status->set_error_arg(src);
@@ -180,10 +178,20 @@ Regexp* Regexp::Simplify() {
   CoalesceWalker cw;
   Regexp* cre = cw.Walk(this, NULL);
   if (cre == NULL)
-    return cre;
+    return NULL;
+  if (cw.stopped_early()) {
+    cre->Decref();
+    return NULL;
+  }
   SimplifyWalker sw;
   Regexp* sre = sw.Walk(cre, NULL);
   cre->Decref();
+  if (sre == NULL)
+    return NULL;
+  if (sw.stopped_early()) {
+    sre->Decref();
+    return NULL;
+  }
   return sre;
 }
 

data/third_party/re2/re2/tostring.cc

@@ -291,7 +291,7 @@ int ToStringWalker::PostVisit(Regexp* re, int parent_arg, int pre_arg,
       // There's no syntax accepted by the parser to generate
       // this node (it is generated by RE2::Set) so make something
       // up that is readable but won't compile.
-      t_->append("(?HaveMatch:%d)", re->match_id());
+      t_->append(StringPrintf("(?HaveMatch:%d)", re->match_id()));
       break;
   }
 

data/third_party/re2/re2/walker-inl.h

@@ -119,7 +119,7 @@ template<typename T> T Regexp::Walker<T>::Copy(T arg) {
 
 // State about a single level in the traversal.
 template<typename T> struct WalkState {
-  WalkState<T>(Regexp* re, T parent)
+  WalkState(Regexp* re, T parent)
     : re(re),
       n(-1),
       parent_arg(parent),

data/third_party/re2/util/mutex.h

@@ -12,6 +12,7 @@
 
 #ifdef _WIN32
 // Requires Windows Vista or Windows Server 2008 at minimum.
+#include <windows.h>
 #if defined(WINVER) && WINVER >= 0x0600
 #define MUTEX_IS_WIN32_SRWLOCK
 #endif
@@ -26,7 +27,6 @@
 #endif
 
 #if defined(MUTEX_IS_WIN32_SRWLOCK)
-#include <windows.h>
 typedef SRWLOCK MutexType;
 #elif defined(MUTEX_IS_PTHREAD_RWLOCK)
 #include <pthread.h>
@@ -66,7 +66,7 @@ class Mutex {
 
 #if defined(MUTEX_IS_WIN32_SRWLOCK)
 
-Mutex::Mutex()             { InitializeSRWLock(&mutex_); }
+Mutex::Mutex()             : mutex_(SRWLOCK_INIT) { }
 Mutex::~Mutex()            { }
 void Mutex::Lock()         { AcquireSRWLockExclusive(&mutex_); }
 void Mutex::Unlock()       { ReleaseSRWLockExclusive(&mutex_); }