Googletest export

Print std::u8string, std::u16string, and std::u32string as string literals

Previously, these types were printed as "{ U+123, U+456, U+789 }". However,
printed output in that form is difficult to compare against any literals that
might be defined in code. Instead, just treat these types like std::string
and std::wstring, escaping non-ASCII characters with a hexadecimal escape
sequence.

The tests have also been updated to cover the new functionality: as a bonus,
the tests now also pass with the MSVC toolchain.

Internally, the code has been reorganized to primarily operate in terms of
char32_t, under the assumption that char32_t will always be at least as big
as wchar_t. While that assumption is currently true, perhaps it won't be in
the future...

PiperOrigin-RevId: 364033132

BUILD.bazel

@@ -43,6 +43,14 @@ config_setting(
     constraint_values = ["@platforms//os:windows"],
 )
 
+config_setting(
+    name = "msvc_compiler",
+    flag_values = {
+        "@bazel_tools//tools/cpp:compiler": "msvc-cl",
+    },
+    visibility = [":__subpackages__"],
+)
+
 config_setting(
     name = "has_absl",
     values = {"define": "absl=1"},

googletest/cmake/internal_utils.cmake

@@ -81,6 +81,8 @@ macro(config_compiler_and_linker)
     # Suppress "unreachable code" warning
     # http://stackoverflow.com/questions/3232669 explains the issue.
     set(cxx_base_flags "${cxx_base_flags} -wd4702")
+    # Ensure MSVC treats source files as UTF-8 encoded.
+    set(cxx_base_flags "${cxx_base_flags} -utf-8")
   elseif (CMAKE_CXX_COMPILER_ID STREQUAL "Clang")
     set(cxx_base_flags "-Wall -Wshadow -Werror -Wconversion")
     set(cxx_exception_flags "-fexceptions")

googletest/include/gtest/gtest-printers.h

@@ -505,24 +505,21 @@ inline void PrintTo(unsigned char* s, ::std::ostream* os) {
   PrintTo(ImplicitCast_<const void*>(s), os);
 }
 #ifdef __cpp_char8_t
-inline void PrintTo(const char8_t* s, ::std::ostream* os) {
+// Overloads for u8 strings.
-  PrintTo(ImplicitCast_<const void*>(s), os);
+void PrintTo(const char8_t* s, ::std::ostream* os);
-}
 inline void PrintTo(char8_t* s, ::std::ostream* os) {
-  PrintTo(ImplicitCast_<const void*>(s), os);
+  PrintTo(ImplicitCast_<const char8_t*>(s), os);
 }
 #endif
-inline void PrintTo(const char16_t* s, ::std::ostream* os) {
+// Overloads for u16 strings.
-  PrintTo(ImplicitCast_<const void*>(s), os);
+void PrintTo(const char16_t* s, ::std::ostream* os);
-}
 inline void PrintTo(char16_t* s, ::std::ostream* os) {
-  PrintTo(ImplicitCast_<const void*>(s), os);
+  PrintTo(ImplicitCast_<const char16_t*>(s), os);
-}
-inline void PrintTo(const char32_t* s, ::std::ostream* os) {
-  PrintTo(ImplicitCast_<const void*>(s), os);
 }
+// Overloads for u32 strings.
+void PrintTo(const char32_t* s, ::std::ostream* os);
 inline void PrintTo(char32_t* s, ::std::ostream* os) {
-  PrintTo(ImplicitCast_<const void*>(s), os);
+  PrintTo(ImplicitCast_<const char32_t*>(s), os);
 }
 
 // MSVC can be configured to define wchar_t as a typedef of unsigned
@@ -558,6 +555,26 @@ inline void PrintTo(const ::std::string& s, ::std::ostream* os) {
   PrintStringTo(s, os);
 }
 
+// Overloads for ::std::u8string
+#ifdef __cpp_char8_t
+GTEST_API_ void PrintU8StringTo(const ::std::u8string& s, ::std::ostream* os);
+inline void PrintTo(const ::std::u8string& s, ::std::ostream* os) {
+  PrintU8StringTo(s, os);
+}
+#endif
+
+// Overloads for ::std::u16string
+GTEST_API_ void PrintU16StringTo(const ::std::u16string& s, ::std::ostream* os);
+inline void PrintTo(const ::std::u16string& s, ::std::ostream* os) {
+  PrintU16StringTo(s, os);
+}
+
+// Overloads for ::std::u32string
+GTEST_API_ void PrintU32StringTo(const ::std::u32string& s, ::std::ostream* os);
+inline void PrintTo(const ::std::u32string& s, ::std::ostream* os) {
+  PrintU32StringTo(s, os);
+}
+
 // Overloads for ::std::wstring.
 #if GTEST_HAS_STD_WSTRING
 GTEST_API_ void PrintWideStringTo(const ::std::wstring&s, ::std::ostream* os);
@@ -805,6 +822,20 @@ void UniversalPrintArray(const T* begin, size_t len, ::std::ostream* os) {
 GTEST_API_ void UniversalPrintArray(
     const char* begin, size_t len, ::std::ostream* os);
 
+#ifdef __cpp_char8_t
+// This overload prints a (const) char8_t array compactly.
+GTEST_API_ void UniversalPrintArray(const char8_t* begin, size_t len,
+                                    ::std::ostream* os);
+#endif
+
+// This overload prints a (const) char16_t array compactly.
+GTEST_API_ void UniversalPrintArray(const char16_t* begin, size_t len,
+                                    ::std::ostream* os);
+
+// This overload prints a (const) char32_t array compactly.
+GTEST_API_ void UniversalPrintArray(const char32_t* begin, size_t len,
+                                    ::std::ostream* os);
+
 // This overload prints a (const) wchar_t array compactly.
 GTEST_API_ void UniversalPrintArray(
     const wchar_t* begin, size_t len, ::std::ostream* os);
@@ -877,12 +908,55 @@ class UniversalTersePrinter<const char*> {
   }
 };
 template <>
-class UniversalTersePrinter<char*> {
+class UniversalTersePrinter<char*> : public UniversalTersePrinter<const char*> {
+};
+
+#ifdef __cpp_char8_t
+template <>
+class UniversalTersePrinter<const char8_t*> {
  public:
-  static void Print(char* str, ::std::ostream* os) {
+  static void Print(const char8_t* str, ::std::ostream* os) {
-    UniversalTersePrinter<const char*>::Print(str, os);
+    if (str == nullptr) {
+      *os << "NULL";
+    } else {
+      UniversalPrint(::std::u8string(str), os);
+    }
   }
 };
+template <>
+class UniversalTersePrinter<char8_t*>
+    : public UniversalTersePrinter<const char8_t*> {};
+#endif
+
+template <>
+class UniversalTersePrinter<const char16_t*> {
+ public:
+  static void Print(const char16_t* str, ::std::ostream* os) {
+    if (str == nullptr) {
+      *os << "NULL";
+    } else {
+      UniversalPrint(::std::u16string(str), os);
+    }
+  }
+};
+template <>
+class UniversalTersePrinter<char16_t*>
+    : public UniversalTersePrinter<const char16_t*> {};
+
+template <>
+class UniversalTersePrinter<const char32_t*> {
+ public:
+  static void Print(const char32_t* str, ::std::ostream* os) {
+    if (str == nullptr) {
+      *os << "NULL";
+    } else {
+      UniversalPrint(::std::u32string(str), os);
+    }
+  }
+};
+template <>
+class UniversalTersePrinter<char32_t*>
+    : public UniversalTersePrinter<const char32_t*> {};
 
 #if GTEST_HAS_STD_WSTRING
 template <>

googletest/include/gtest/internal/gtest-port.h

@@ -1936,6 +1936,19 @@ inline bool IsUpper(char ch) {
 inline bool IsXDigit(char ch) {
   return isxdigit(static_cast<unsigned char>(ch)) != 0;
 }
+#ifdef __cpp_char8_t
+inline bool IsXDigit(char8_t ch) {
+  return isxdigit(static_cast<unsigned char>(ch)) != 0;
+}
+#endif
+inline bool IsXDigit(char16_t ch) {
+  const unsigned char low_byte = static_cast<unsigned char>(ch);
+  return ch == low_byte && isxdigit(low_byte) != 0;
+}
+inline bool IsXDigit(char32_t ch) {
+  const unsigned char low_byte = static_cast<unsigned char>(ch);
+  return ch == low_byte && isxdigit(low_byte) != 0;
+}
 inline bool IsXDigit(wchar_t ch) {
   const unsigned char low_byte = static_cast<unsigned char>(ch);
   return ch == low_byte && isxdigit(low_byte) != 0;

googletest/src/gtest-printers.cc

@@ -42,12 +42,16 @@
 // defines Foo.
 
 #include "gtest/gtest-printers.h"
+
 #include <stdio.h>
+
 #include <cctype>
 #include <cstdint>
 #include <cwchar>
 #include <ostream>  // NOLINT
 #include <string>
+#include <type_traits>
+
 #include "gtest/internal/gtest-port.h"
 #include "src/gtest-internal-inl.h"
 
@@ -103,6 +107,16 @@ void PrintBytesInObjectToImpl(const unsigned char* obj_bytes, size_t count,
   *os << ">";
 }
 
+// Helpers for widening a character to char32_t. Since the standard does not
+// specify if char / wchar_t is signed or unsigned, it is important to first
+// convert it to the unsigned type of the same width before widening it to
+// char32_t.
+template <typename CharType>
+char32_t ToChar32(CharType in) {
+  return static_cast<char32_t>(
+      static_cast<typename std::make_unsigned<CharType>::type>(in));
+}
+
 }  // namespace
 
 namespace internal {
@@ -131,18 +145,15 @@ enum CharFormat {
 // Returns true if c is a printable ASCII character.  We test the
 // value of c directly instead of calling isprint(), which is buggy on
 // Windows Mobile.
-inline bool IsPrintableAscii(wchar_t c) {
+inline bool IsPrintableAscii(char32_t c) { return 0x20 <= c && c <= 0x7E; }
-  return 0x20 <= c && c <= 0x7E;
-}
 
-// Prints a wide or narrow char c as a character literal without the
+// Prints c (of type char, char8_t, char16_t, char32_t, or wchar_t) as a
-// quotes, escaping it when necessary; returns how c was formatted.
+// character literal without the quotes, escaping it when necessary; returns how
-// The template argument UnsignedChar is the unsigned version of Char,
+// c was formatted.
-// which is the type of c.
+template <typename Char>
-template <typename UnsignedChar, typename Char>
 static CharFormat PrintAsCharLiteralTo(Char c, ostream* os) {
-  wchar_t w_c = static_cast<wchar_t>(c);
+  const char32_t u_c = ToChar32(c);
-  switch (w_c) {
+  switch (u_c) {
     case L'\0':
       *os << "\\0";
       break;
@@ -174,13 +185,12 @@ static CharFormat PrintAsCharLiteralTo(Char c, ostream* os) {
       *os << "\\v";
       break;
     default:
-      if (IsPrintableAscii(w_c)) {
+      if (IsPrintableAscii(u_c)) {
         *os << static_cast<char>(c);
         return kAsIs;
       } else {
         ostream::fmtflags flags = os->flags();
-        *os << "\\x" << std::hex << std::uppercase
+        *os << "\\x" << std::hex << std::uppercase << static_cast<int>(u_c);
-            << static_cast<int>(static_cast<UnsignedChar>(c));
         os->flags(flags);
         return kHexEscape;
       }
@@ -188,9 +198,9 @@ static CharFormat PrintAsCharLiteralTo(Char c, ostream* os) {
   return kSpecialEscape;
 }
 
-// Prints a wchar_t c as if it's part of a string literal, escaping it when
+// Prints a char32_t c as if it's part of a string literal, escaping it when
 // necessary; returns how c was formatted.
-static CharFormat PrintAsStringLiteralTo(wchar_t c, ostream* os) {
+static CharFormat PrintAsStringLiteralTo(char32_t c, ostream* os) {
   switch (c) {
     case L'\'':
       *os << "'";
@@ -199,26 +209,68 @@ static CharFormat PrintAsStringLiteralTo(wchar_t c, ostream* os) {
       *os << "\\\"";
       return kSpecialEscape;
     default:
-      return PrintAsCharLiteralTo<wchar_t>(c, os);
+      return PrintAsCharLiteralTo(c, os);
   }
 }
 
+static const char* GetCharWidthPrefix(char) {
+  return "";
+}
+
+static const char* GetCharWidthPrefix(signed char) {
+  return "";
+}
+
+static const char* GetCharWidthPrefix(unsigned char) {
+  return "";
+}
+
+#ifdef __cpp_char8_t
+static const char* GetCharWidthPrefix(char8_t) {
+  return "u8";
+}
+#endif
+
+static const char* GetCharWidthPrefix(char16_t) {
+  return "u";
+}
+
+static const char* GetCharWidthPrefix(char32_t) {
+  return "U";
+}
+
+static const char* GetCharWidthPrefix(wchar_t) {
+  return "L";
+}
+
 // Prints a char c as if it's part of a string literal, escaping it when
 // necessary; returns how c was formatted.
 static CharFormat PrintAsStringLiteralTo(char c, ostream* os) {
-  return PrintAsStringLiteralTo(
+  return PrintAsStringLiteralTo(ToChar32(c), os);
-      static_cast<wchar_t>(static_cast<unsigned char>(c)), os);
 }
 
-// Prints a wide or narrow character c and its code.  '\0' is printed
+#ifdef __cpp_char8_t
-// as "'\\0'", other unprintable characters are also properly escaped
+static CharFormat PrintAsStringLiteralTo(char8_t c, ostream* os) {
-// using the standard C++ escape sequence.  The template argument
+  return PrintAsStringLiteralTo(ToChar32(c), os);
-// UnsignedChar is the unsigned version of Char, which is the type of c.
+}
-template <typename UnsignedChar, typename Char>
+#endif
+
+static CharFormat PrintAsStringLiteralTo(char16_t c, ostream* os) {
+  return PrintAsStringLiteralTo(ToChar32(c), os);
+}
+
+static CharFormat PrintAsStringLiteralTo(wchar_t c, ostream* os) {
+  return PrintAsStringLiteralTo(ToChar32(c), os);
+}
+
+// Prints a character c (of type char, char8_t, char16_t, char32_t, or wchar_t)
+// and its code. '\0' is printed as "'\\0'", other unprintable characters are
+// also properly escaped using the standard C++ escape sequence.
+template <typename Char>
 void PrintCharAndCodeTo(Char c, ostream* os) {
   // First, print c as a literal in the most readable form we can find.
-  *os << ((sizeof(c) > 1) ? "L'" : "'");
+  *os << GetCharWidthPrefix(c) << "'";
-  const CharFormat format = PrintAsCharLiteralTo<UnsignedChar>(c, os);
+  const CharFormat format = PrintAsCharLiteralTo(c, os);
   *os << "'";
 
   // To aid user debugging, we also print c's code in decimal, unless
@@ -239,26 +291,21 @@ void PrintCharAndCodeTo(Char c, ostream* os) {
   *os << ")";
 }
 
-void PrintTo(unsigned char c, ::std::ostream* os) {
+void PrintTo(unsigned char c, ::std::ostream* os) { PrintCharAndCodeTo(c, os); }
-  PrintCharAndCodeTo<unsigned char>(c, os);
+void PrintTo(signed char c, ::std::ostream* os) { PrintCharAndCodeTo(c, os); }
-}
-void PrintTo(signed char c, ::std::ostream* os) {
-  PrintCharAndCodeTo<unsigned char>(c, os);
-}
 
 // Prints a wchar_t as a symbol if it is printable or as its internal
 // code otherwise and also as its code.  L'\0' is printed as "L'\\0'".
-void PrintTo(wchar_t wc, ostream* os) {
+void PrintTo(wchar_t wc, ostream* os) { PrintCharAndCodeTo(wc, os); }
-  PrintCharAndCodeTo<wchar_t>(wc, os);
-}
 
+// TODO(dcheng): Consider making this delegate to PrintCharAndCodeTo() as well.
 void PrintTo(char32_t c, ::std::ostream* os) {
   *os << std::hex << "U+" << std::uppercase << std::setfill('0') << std::setw(4)
       << static_cast<uint32_t>(c);
 }
 
 // Prints the given array of characters to the ostream.  CharType must be either
-// char or wchar_t.
+// char, char8_t, char16_t, char32_t, or wchar_t.
 // The array starts at begin, the length is len, it may include '\0' characters
 // and may not be NUL-terminated.
 template <typename CharType>
@@ -268,8 +315,8 @@ GTEST_ATTRIBUTE_NO_SANITIZE_HWADDRESS_
 GTEST_ATTRIBUTE_NO_SANITIZE_THREAD_
 static CharFormat PrintCharsAsStringTo(
     const CharType* begin, size_t len, ostream* os) {
-  const char* const kQuoteBegin = sizeof(CharType) == 1 ? "\"" : "L\"";
+  const char* const quote_prefix = GetCharWidthPrefix(*begin);
-  *os << kQuoteBegin;
+  *os << quote_prefix << "\"";
   bool is_previous_hex = false;
   CharFormat print_format = kAsIs;
   for (size_t index = 0; index < len; ++index) {
@@ -278,7 +325,7 @@ static CharFormat PrintCharsAsStringTo(
       // Previous character is of '\x..' form and this character can be
       // interpreted as another hexadecimal digit in its number. Break string to
       // disambiguate.
-      *os << "\" " << kQuoteBegin;
+      *os << "\" " << quote_prefix << "\"";
     }
     is_previous_hex = PrintAsStringLiteralTo(cur, os) == kHexEscape;
     // Remember if any characters required hex escaping.
@@ -324,22 +371,57 @@ void UniversalPrintArray(const char* begin, size_t len, ostream* os) {
   UniversalPrintCharArray(begin, len, os);
 }
 
+#ifdef __cpp_char8_t
+// Prints a (const) char8_t array of 'len' elements, starting at address
+// 'begin'.
+void UniversalPrintArray(const char8_t* begin, size_t len, ostream* os) {
+  UniversalPrintCharArray(begin, len, os);
+}
+#endif
+
+// Prints a (const) char16_t array of 'len' elements, starting at address
+// 'begin'.
+void UniversalPrintArray(const char16_t* begin, size_t len, ostream* os) {
+  UniversalPrintCharArray(begin, len, os);
+}
+
+// Prints a (const) char32_t array of 'len' elements, starting at address
+// 'begin'.
+void UniversalPrintArray(const char32_t* begin, size_t len, ostream* os) {
+  UniversalPrintCharArray(begin, len, os);
+}
+
 // Prints a (const) wchar_t array of 'len' elements, starting at address
 // 'begin'.
 void UniversalPrintArray(const wchar_t* begin, size_t len, ostream* os) {
   UniversalPrintCharArray(begin, len, os);
 }
 
-// Prints the given C string to the ostream.
+namespace {
-void PrintTo(const char* s, ostream* os) {
+
+// Prints a null-terminated C-style string to the ostream.
+template <typename Char>
+void PrintCStringTo(const Char* s, ostream* os) {
   if (s == nullptr) {
     *os << "NULL";
   } else {
     *os << ImplicitCast_<const void*>(s) << " pointing to ";
-    PrintCharsAsStringTo(s, strlen(s), os);
+    PrintCharsAsStringTo(s, std::char_traits<Char>::length(s), os);
   }
 }
 
+}  // anonymous namespace
+
+void PrintTo(const char* s, ostream* os) { PrintCStringTo(s, os); }
+
+#ifdef __cpp_char8_t
+void PrintTo(const char8_t* s, ostream* os) { PrintCStringTo(s, os); }
+#endif
+
+void PrintTo(const char16_t* s, ostream* os) { PrintCStringTo(s, os); }
+
+void PrintTo(const char32_t* s, ostream* os) { PrintCStringTo(s, os); }
+
 // MSVC compiler can be configured to define whar_t as a typedef
 // of unsigned short. Defining an overload for const wchar_t* in that case
 // would cause pointers to unsigned shorts be printed as wide strings,
@@ -348,14 +430,7 @@ void PrintTo(const char* s, ostream* os) {
 // wchar_t is implemented as a native type.
 #if !defined(_MSC_VER) || defined(_NATIVE_WCHAR_T_DEFINED)
 // Prints the given wide C string to the ostream.
-void PrintTo(const wchar_t* s, ostream* os) {
+void PrintTo(const wchar_t* s, ostream* os) { PrintCStringTo(s, os); }
-  if (s == nullptr) {
-    *os << "NULL";
-  } else {
-    *os << ImplicitCast_<const void*>(s) << " pointing to ";
-    PrintCharsAsStringTo(s, wcslen(s), os);
-  }
-}
 #endif  // wchar_t is native
 
 namespace {
@@ -433,6 +508,20 @@ void PrintStringTo(const ::std::string& s, ostream* os) {
   }
 }
 
+#ifdef __cpp_char8_t
+void PrintU8StringTo(const ::std::u8string& s, ostream* os) {
+  PrintCharsAsStringTo(s.data(), s.size(), os);
+}
+#endif
+
+void PrintU16StringTo(const ::std::u16string& s, ostream* os) {
+  PrintCharsAsStringTo(s.data(), s.size(), os);
+}
+
+void PrintU32StringTo(const ::std::u32string& s, ostream* os) {
+  PrintCharsAsStringTo(s.data(), s.size(), os);
+}
+
 #if GTEST_HAS_STD_WSTRING
 void PrintWideStringTo(const ::std::wstring& s, ostream* os) {
   PrintCharsAsStringTo(s.data(), s.size(), os);

googletest/test/BUILD.bazel

@@ -83,6 +83,10 @@ cc_test(
     copts = select({
         "//:windows": ["-DGTEST_USE_OWN_TR1_TUPLE=0"],
         "//conditions:default": ["-DGTEST_USE_OWN_TR1_TUPLE=1"],
+    }) + select({
+        # Ensure MSVC treats source files as UTF-8 encoded.
+        "//:msvc_compiler": ["-utf-8"],
+        "//conditions:default": [],
     }),
     includes = [
         "googletest",

googletest/test/googletest-printers-test.cc

@@ -493,6 +493,92 @@ TEST(PrintCStringTest, EscapesProperly) {
             Print(p));
 }
 
+#ifdef __cpp_char8_t
+// const char8_t*.
+TEST(PrintU8StringTest, Const) {
+  const char8_t* p = u8"界";
+  EXPECT_EQ(PrintPointer(p) + " pointing to u8\"\\xE7\\x95\\x8C\"", Print(p));
+}
+
+// char8_t*.
+TEST(PrintU8StringTest, NonConst) {
+  char8_t p[] = u8"世";
+  EXPECT_EQ(PrintPointer(p) + " pointing to u8\"\\xE4\\xB8\\x96\"",
+            Print(static_cast<char8_t*>(p)));
+}
+
+// NULL u8 string.
+TEST(PrintU8StringTest, Null) {
+  const char8_t* p = nullptr;
+  EXPECT_EQ("NULL", Print(p));
+}
+
+// Tests that u8 strings are escaped properly.
+TEST(PrintU8StringTest, EscapesProperly) {
+  const char8_t* p = u8"'\"?\\\a\b\f\n\r\t\v\x7F\xFF hello 世界";
+  EXPECT_EQ(PrintPointer(p) +
+                " pointing to u8\"'\\\"?\\\\\\a\\b\\f\\n\\r\\t\\v\\x7F\\xFF "
+                "hello \\xE4\\xB8\\x96\\xE7\\x95\\x8C\"",
+            Print(p));
+}
+#endif
+
+// const char16_t*.
+TEST(PrintU16StringTest, Const) {
+  const char16_t* p = u"界";
+  EXPECT_EQ(PrintPointer(p) + " pointing to u\"\\x754C\"", Print(p));
+}
+
+// char16_t*.
+TEST(PrintU16StringTest, NonConst) {
+  char16_t p[] = u"世";
+  EXPECT_EQ(PrintPointer(p) + " pointing to u\"\\x4E16\"",
+            Print(static_cast<char16_t*>(p)));
+}
+
+// NULL u16 string.
+TEST(PrintU16StringTest, Null) {
+  const char16_t* p = nullptr;
+  EXPECT_EQ("NULL", Print(p));
+}
+
+// Tests that u16 strings are escaped properly.
+TEST(PrintU16StringTest, EscapesProperly) {
+  const char16_t* p = u"'\"?\\\a\b\f\n\r\t\v\x7F\xFF hello 世界";
+  EXPECT_EQ(PrintPointer(p) +
+                " pointing to u\"'\\\"?\\\\\\a\\b\\f\\n\\r\\t\\v\\x7F\\xFF "
+                "hello \\x4E16\\x754C\"",
+            Print(p));
+}
+
+// const char32_t*.
+TEST(PrintU32StringTest, Const) {
+  const char32_t* p = U"🗺️";
+  EXPECT_EQ(PrintPointer(p) + " pointing to U\"\\x1F5FA\\xFE0F\"", Print(p));
+}
+
+// char32_t*.
+TEST(PrintU32StringTest, NonConst) {
+  char32_t p[] = U"🌌";
+  EXPECT_EQ(PrintPointer(p) + " pointing to U\"\\x1F30C\"",
+            Print(static_cast<char32_t*>(p)));
+}
+
+// NULL u32 string.
+TEST(PrintU32StringTest, Null) {
+  const char32_t* p = nullptr;
+  EXPECT_EQ("NULL", Print(p));
+}
+
+// Tests that u32 strings are escaped properly.
+TEST(PrintU32StringTest, EscapesProperly) {
+  const char32_t* p = U"'\"?\\\a\b\f\n\r\t\v\x7F\xFF hello 🗺️";
+  EXPECT_EQ(PrintPointer(p) +
+                " pointing to U\"'\\\"?\\\\\\a\\b\\f\\n\\r\\t\\v\\x7F\\xFF "
+                "hello \\x1F5FA\\xFE0F\"",
+            Print(p));
+}
+
 // MSVC compiler can be configured to define whar_t as a typedef
 // of unsigned short. Defining an overload for const wchar_t* in that case
 // would cause pointers to unsigned shorts be printed as wide strings,
@@ -564,56 +650,6 @@ TEST(PrintCharPointerTest, ConstUnsignedChar) {
   EXPECT_EQ("NULL", Print(p));
 }
 
-#ifdef __cpp_char8_t
-// char8_t*
-TEST(PrintCharPointerTest, Char8) {
-  char8_t* p = reinterpret_cast<char8_t*>(0x1234);
-  EXPECT_EQ(PrintPointer(p), Print(p));
-  p = nullptr;
-  EXPECT_EQ("NULL", Print(p));
-}
-
-// const char8_t*
-TEST(PrintCharPointerTest, ConstChar8) {
-  const char8_t* p = reinterpret_cast<const char8_t*>(0x1234);
-  EXPECT_EQ(PrintPointer(p), Print(p));
-  p = nullptr;
-  EXPECT_EQ("NULL", Print(p));
-}
-#endif
-
-// char16_t*
-TEST(PrintCharPointerTest, Char16) {
-  char16_t* p = reinterpret_cast<char16_t*>(0x1234);
-  EXPECT_EQ(PrintPointer(p), Print(p));
-  p = nullptr;
-  EXPECT_EQ("NULL", Print(p));
-}
-
-// const char16_t*
-TEST(PrintCharPointerTest, ConstChar16) {
-  const char16_t* p = reinterpret_cast<const char16_t*>(0x1234);
-  EXPECT_EQ(PrintPointer(p), Print(p));
-  p = nullptr;
-  EXPECT_EQ("NULL", Print(p));
-}
-
-// char32_t*
-TEST(PrintCharPointerTest, Char32) {
-  char32_t* p = reinterpret_cast<char32_t*>(0x1234);
-  EXPECT_EQ(PrintPointer(p), Print(p));
-  p = nullptr;
-  EXPECT_EQ("NULL", Print(p));
-}
-
-// const char32_t*
-TEST(PrintCharPointerTest, ConstChar32) {
-  const char32_t* p = reinterpret_cast<const char32_t*>(0x1234);
-  EXPECT_EQ(PrintPointer(p), Print(p));
-  p = nullptr;
-  EXPECT_EQ("NULL", Print(p));
-}
-
 // Tests printing pointers to simple, built-in types.
 
 // bool*.
@@ -753,13 +789,58 @@ TEST(PrintArrayTest, CharArrayWithNoTerminatingNul) {
   EXPECT_EQ("\"H\\0i\" (no terminating NUL)", PrintArrayHelper(a));
 }
 
-// const char array with terminating NUL.
+// char array with terminating NUL.
-TEST(PrintArrayTest, ConstCharArrayWithTerminatingNul) {
+TEST(PrintArrayTest, CharArrayWithTerminatingNul) {
   const char a[] = "\0Hi";
   EXPECT_EQ("\"\\0Hi\"", PrintArrayHelper(a));
 }
 
-// const wchar_t array without terminating NUL.
+#ifdef __cpp_char8_t
+// char_t array without terminating NUL.
+TEST(PrintArrayTest, Char8ArrayWithNoTerminatingNul) {
+  // Array a contains '\0' in the middle and doesn't end with '\0'.
+  const char8_t a[] = {u8'H', u8'\0', u8'i'};
+  EXPECT_EQ("u8\"H\\0i\" (no terminating NUL)", PrintArrayHelper(a));
+}
+
+// char8_t array with terminating NUL.
+TEST(PrintArrayTest, Char8ArrayWithTerminatingNul) {
+  const char8_t a[] = u8"\0世界";
+  EXPECT_EQ(
+      "u8\"\\0\\xE4\\xB8\\x96\\xE7\\x95\\x8C\"",
+      PrintArrayHelper(a));
+}
+#endif
+
+// const char16_t array without terminating NUL.
+TEST(PrintArrayTest, Char16ArrayWithNoTerminatingNul) {
+  // Array a contains '\0' in the middle and doesn't end with '\0'.
+  const char16_t a[] = {u'こ', u'\0', u'ん', u'に', u'ち', u'は'};
+  EXPECT_EQ("u\"\\x3053\\0\\x3093\\x306B\\x3061\\x306F\" (no terminating NUL)",
+            PrintArrayHelper(a));
+}
+
+// char16_t array with terminating NUL.
+TEST(PrintArrayTest, Char16ArrayWithTerminatingNul) {
+  const char16_t a[] = u"\0こんにちは";
+  EXPECT_EQ("u\"\\0\\x3053\\x3093\\x306B\\x3061\\x306F\"", PrintArrayHelper(a));
+}
+
+// char32_t array without terminating NUL.
+TEST(PrintArrayTest, Char32ArrayWithNoTerminatingNul) {
+  // Array a contains '\0' in the middle and doesn't end with '\0'.
+  const char32_t a[] = {U'👋', U'\0', U'🌌'};
+  EXPECT_EQ("U\"\\x1F44B\\0\\x1F30C\" (no terminating NUL)",
+            PrintArrayHelper(a));
+}
+
+// char32_t array with terminating NUL.
+TEST(PrintArrayTest, Char32ArrayWithTerminatingNul) {
+  const char32_t a[] = U"\0👋🌌";
+  EXPECT_EQ("U\"\\0\\x1F44B\\x1F30C\"", PrintArrayHelper(a));
+}
+
+// wchar_t array without terminating NUL.
 TEST(PrintArrayTest, WCharArrayWithNoTerminatingNul) {
   // Array a contains '\0' in the middle and doesn't end with '\0'.
   const wchar_t a[] = {L'H', L'\0', L'i'};
@@ -767,50 +848,11 @@ TEST(PrintArrayTest, WCharArrayWithNoTerminatingNul) {
 }
 
 // wchar_t array with terminating NUL.
-TEST(PrintArrayTest, WConstCharArrayWithTerminatingNul) {
+TEST(PrintArrayTest, WCharArrayWithTerminatingNul) {
   const wchar_t a[] = L"\0Hi";
   EXPECT_EQ("L\"\\0Hi\"", PrintArrayHelper(a));
 }
 
-#ifdef __cpp_char8_t
-// char8_t array.
-TEST(PrintArrayTest, Char8Array) {
-  const char8_t a[] = u8"Hello, world!";
-  EXPECT_EQ(
-      "{ U+0048, U+0065, U+006C, U+006C, U+006F, U+002C, U+0020, U+0077, "
-      "U+006F, U+0072, U+006C, U+0064, U+0021, U+0000 }",
-      PrintArrayHelper(a));
-}
-#endif
-
-// char16_t array.
-#ifdef _MSC_VER
-// TODO(b/173029407): Figure out why this doesn't work under MSVC.
-TEST(PrintArrayTest, DISABLED_Char16Array) {
-#else
-TEST(PrintArrayTest, Char16Array) {
-#endif
-  const char16_t a[] = u"Hello, 世界";
-  EXPECT_EQ(
-      "{ U+0048, U+0065, U+006C, U+006C, U+006F, U+002C, U+0020, U+4E16, "
-      "U+754C, U+0000 }",
-      PrintArrayHelper(a));
-}
-
-// char32_t array.
-#ifdef _MSC_VER
-// TODO(b/173029407): Figure out why this doesn't work under MSVC.
-TEST(PrintArrayTest, DISABLED_Char32Array) {
-#else
-TEST(PrintArrayTest, Char32Array) {
-#endif
-  const char32_t a[] = U"Hello, 世界";
-  EXPECT_EQ(
-      "{ U+0048, U+0065, U+006C, U+006C, U+006F, U+002C, U+0020, U+4E16, "
-      "U+754C, U+0000 }",
-      PrintArrayHelper(a));
-}
-
 // Array of objects.
 TEST(PrintArrayTest, ObjectArray) {
   std::string a[3] = {"Hi", "Hello", "Ni hao"};
@@ -872,41 +914,22 @@ TEST(PrintWideStringTest, StringAmbiguousHex) {
 
 #ifdef __cpp_char8_t
 TEST(PrintStringTest, U8String) {
-  std::u8string str = u8"Hello, world!";
+  std::u8string str = u8"Hello, 世界";
   EXPECT_EQ(str, str);  // Verify EXPECT_EQ compiles with this type.
-  EXPECT_EQ(
+  EXPECT_EQ("u8\"Hello, \\xE4\\xB8\\x96\\xE7\\x95\\x8C\"", Print(str));
-      "{ U+0048, U+0065, U+006C, U+006C, U+006F, U+002C, U+0020, U+0077, "
-      "U+006F, U+0072, U+006C, U+0064, U+0021 }",
-      Print(str));
 }
 #endif
 
-#ifdef _MSC_VER
-// TODO(b/173029407): Figure out why this doesn't work under MSVC.
-TEST(PrintStringTest, DISABLED_U16String) {
-#else
 TEST(PrintStringTest, U16String) {
-#endif
   std::u16string str = u"Hello, 世界";
   EXPECT_EQ(str, str);  // Verify EXPECT_EQ compiles with this type.
-  EXPECT_EQ(
+  EXPECT_EQ("u\"Hello, \\x4E16\\x754C\"", Print(str));
-      "{ U+0048, U+0065, U+006C, U+006C, U+006F, U+002C, U+0020, U+4E16, "
-      "U+754C }",
-      Print(str));
 }
 
-#ifdef _MSC_VER
-// TODO(b/173029407): Figure out why this doesn't work under MSVC.
-TEST(PrintStringTest, DISABLED_U32String) {
-#else
 TEST(PrintStringTest, U32String) {
-#endif
+  std::u32string str = U"Hello, 🗺️";
-  std::u32string str = U"Hello, 世界";
+  EXPECT_EQ(str, str);  // Verify EXPECT_EQ compiles with this type
-  EXPECT_EQ(str, str);  // Verify EXPECT_EQ compiles with this type.
+  EXPECT_EQ("U\"Hello, \\x1F5FA\\xFE0F\"", Print(str));
-  EXPECT_EQ(
-      "{ U+0048, U+0065, U+006C, U+006C, U+006F, U+002C, U+0020, U+4E16, "
-      "U+754C }",
-      Print(str));
 }
 
 // Tests printing types that support generic streaming (i.e. streaming