[libc] Implement towupper and towlower entrypoints (#198659)

Following up on #187670 to add public entrypoints for the wctype
conversion functions.

Assisted-by: Automated tooling, human reviewed.

---------

Co-authored-by: Muhammad Bassiouni <60100307+bassiounix@users.noreply.github.com>

libc/config/linux/aarch64/entrypoints.txt

@@ -417,6 +417,8 @@ set(TARGET_LIBC_ENTRYPOINTS
     libc.src.wctype.iswprint
     libc.src.wctype.iswctype
     libc.src.wctype.wctype
+    libc.src.wctype.towlower
+    libc.src.wctype.towupper
 
     # sys/uio.h entrypoints
     libc.src.sys.uio.writev

libc/config/linux/riscv/entrypoints.txt

@@ -421,6 +421,8 @@ set(TARGET_LIBC_ENTRYPOINTS
     libc.src.wctype.iswprint
     libc.src.wctype.iswctype
     libc.src.wctype.wctype
+    libc.src.wctype.towlower
+    libc.src.wctype.towupper
 
     # sys/uio.h entrypoints
     libc.src.sys.uio.writev

libc/config/linux/x86_64/entrypoints.txt

@@ -481,6 +481,8 @@ set(TARGET_LIBC_ENTRYPOINTS
     libc.src.wctype.iswprint
     libc.src.wctype.iswctype
     libc.src.wctype.wctype
+    libc.src.wctype.towlower
+    libc.src.wctype.towupper
 
     # sys/uio.h entrypoints
     libc.src.sys.uio.writev

libc/include/wctype.yaml

@@ -1,4 +1,6 @@
 header: wctype.h
+standards:
+  - stdc
 types:
   - type_name: wint_t
   - type_name: wctype_t
@@ -88,3 +90,15 @@ functions:
     return_type: wctype_t
     arguments:
       - type: const char*
+  - name: towlower
+    standards:
+      - stdc
+    return_type: wint_t
+    arguments:
+      - type: wint_t
+  - name: towupper
+    standards:
+      - stdc
+    return_type: wint_t
+    arguments:
+      - type: wint_t

libc/src/__support/CMakeLists.txt

@@ -165,6 +165,7 @@ list(APPEND wctype_utils_deps
 if ("${LIBC_CONF_WCTYPE_MODE}" STREQUAL "LIBC_WCTYPE_MODE_UTF8")
   list(APPEND wctype_utils_deps
       libc.src.__support.wctype.wctype_classification_utils
+      libc.src.__support.wctype.wctype_conversion_utils
   )
 endif()
 

libc/src/__support/wctype/CMakeLists.txt

@@ -49,3 +49,21 @@ add_object_library(
     libc.src.__support.CPP.limits
     libc.src.__support.libc_assert
 )
+
+add_object_library(
+  wctype_conversion_utils
+  HDRS
+    wctype_conversion_utils.h
+  SRCS
+    wctype_conversion_utils.cpp
+  COMPILE_OPTIONS
+    "-DLIBC_ENABLE_CONSTEXPR=1"
+  DEPENDS
+    libc.hdr.types.wchar_t
+    libc.hdr.types.wint_t
+    libc.src.__support.macros.attributes
+    libc.src.__support.macros.config
+    .lower_to_upper
+    .upper_to_lower
+)
+

libc/src/__support/wctype/perfect_hash_map.h

@@ -9,8 +9,6 @@
 #ifndef LLVM_LIBC_SRC___SUPPORT_WCTYPE_PERFECT_HASH_MAP_H
 #define LLVM_LIBC_SRC___SUPPORT_WCTYPE_PERFECT_HASH_MAP_H
 
-#define LIBC_ENABLE_CONSTEXPR 1
-
 #include "hdr/types/size_t.h"
 #include "hdr/types/wint_t.h"
 #include "src/__support/CPP/array.h"
@@ -29,8 +27,6 @@
 #include "src/__support/OSUtil/io.h"
 #endif
 
-#undef LIBC_ENABLE_CONSTEXPR
-
 namespace LIBC_NAMESPACE_DECL {
 namespace wctype_internal {
 

libc/src/__support/wctype/wctype_conversion_utils.cpp

@@ -0,0 +1,46 @@
+//===----------------------------------------------------------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+///
+/// \file
+/// Implementation of wctype conversion helpers.
+///
+//===----------------------------------------------------------------------===//
+
+#include "wctype_conversion_utils.h"
+#include "lower_to_upper.h"
+#include "upper_to_lower.h"
+
+namespace LIBC_NAMESPACE_DECL {
+namespace wctype_internal {
+
+wint_t tolower(wchar_t wch) {
+  if (auto mapped = UPPER_TO_LOWER_MAP.find(wch)) {
+    return mapped.value();
+  }
+#if WINT_MAX > 0xFFFF
+  if (auto mapped = UPPER_TO_LOWER_MAP_32.find(wch)) {
+    return mapped.value();
+  }
+#endif
+  return wch;
+}
+
+wint_t toupper(wchar_t wch) {
+  if (auto mapped = LOWER_TO_UPPER_MAP.find(wch)) {
+    return mapped.value();
+  }
+#if WINT_MAX > 0xFFFF
+  if (auto mapped = LOWER_TO_UPPER_MAP_32.find(wch)) {
+    return mapped.value();
+  }
+#endif
+  return wch;
+}
+
+} // namespace wctype_internal
+} // namespace LIBC_NAMESPACE_DECL

libc/src/__support/wctype/wctype_conversion_utils.h

@@ -0,0 +1,34 @@
+//===----------------------------------------------------------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+///
+/// \file
+/// Interface for wctype conversion functions.
+///
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_LIBC_SRC___SUPPORT_WCTYPE_WCTYPE_CONVERSION_UTILS_H
+#define LLVM_LIBC_SRC___SUPPORT_WCTYPE_WCTYPE_CONVERSION_UTILS_H
+
+#include "hdr/types/wchar_t.h"
+#include "hdr/types/wint_t.h"
+#include "src/__support/macros/attributes.h"
+#include "src/__support/macros/config.h"
+
+namespace LIBC_NAMESPACE_DECL {
+namespace wctype_internal {
+
+// Helper functions for non-ASCII case conversions.
+// These are implemented in wctype_conversion_utils.cpp using the generated
+// maps.
+wint_t tolower(wchar_t wch);
+wint_t toupper(wchar_t wch);
+
+} // namespace wctype_internal
+} // namespace LIBC_NAMESPACE_DECL
+
+#endif // LLVM_LIBC_SRC___SUPPORT_WCTYPE_WCTYPE_CONVERSION_UTILS_H

libc/src/__support/wctype_utils.h

@@ -22,6 +22,7 @@
 
 #if LIBC_CONF_WCTYPE_MODE == LIBC_WCTYPE_MODE_UTF8
 #include "src/__support/wctype/wctype_classification_utils.h"
+#include "src/__support/wctype/wctype_conversion_utils.h"
 #endif
 
 namespace LIBC_NAMESPACE_DECL {
@@ -585,8 +586,7 @@ LIBC_INLINE constexpr wchar_t tolower(wchar_t wch) {
   if (static_cast<uint32_t>(wch) < 128) {
     return ascii::tolower(wch);
   }
-  // TODO: Add UTF8 implementation.
-  return wch;
+  return static_cast<wchar_t>(wctype_internal::tolower(wch));
 #endif
 }
 
@@ -597,8 +597,7 @@ LIBC_INLINE constexpr wchar_t toupper(wchar_t wch) {
   if (static_cast<uint32_t>(wch) < 128) {
     return ascii::toupper(wch);
   }
-  // TODO: Add UTF8 implementation.
-  return wch;
+  return static_cast<wchar_t>(wctype_internal::toupper(wch));
 #endif
 }
 

libc/src/wctype/CMakeLists.txt

@@ -153,3 +153,27 @@ add_entrypoint_object(
     libc.src.__support.wctype_impl
     libc.hdr.types.wctype_t
 )
+
+add_entrypoint_object(
+  towlower
+  SRCS
+    towlower.cpp
+  HDRS
+    towlower.h
+  DEPENDS
+    libc.hdr.types.wint_t
+    libc.src.__support.common
+    libc.src.__support.wctype_utils
+)
+
+add_entrypoint_object(
+  towupper
+  SRCS
+    towupper.cpp
+  HDRS
+    towupper.h
+  DEPENDS
+    libc.hdr.types.wint_t
+    libc.src.__support.common
+    libc.src.__support.wctype_utils
+)

libc/src/wctype/towlower.cpp

@@ -0,0 +1,27 @@
+//===----------------------------------------------------------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+///
+/// \file
+/// Implementation of towlower.
+///
+//===----------------------------------------------------------------------===//
+
+#include "src/wctype/towlower.h"
+#include "hdr/types/wint_t.h"
+#include "src/__support/common.h"
+#include "src/__support/wctype_utils.h"
+
+namespace LIBC_NAMESPACE_DECL {
+
+LLVM_LIBC_FUNCTION(wint_t, towlower, (wint_t c)) {
+  if (c == static_cast<wint_t>(static_cast<wchar_t>(c)))
+    return static_cast<wint_t>(internal::tolower(static_cast<wchar_t>(c)));
+  return c;
+}
+
+} // namespace LIBC_NAMESPACE_DECL

libc/src/wctype/towlower.h

@@ -0,0 +1,26 @@
+//===----------------------------------------------------------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+///
+/// \file
+/// Implementation header for towlower.
+///
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_LIBC_SRC_WCTYPE_TOWLOWER_H
+#define LLVM_LIBC_SRC_WCTYPE_TOWLOWER_H
+
+#include "hdr/types/wint_t.h"
+#include "src/__support/common.h"
+
+namespace LIBC_NAMESPACE_DECL {
+
+wint_t towlower(wint_t c);
+
+} // namespace LIBC_NAMESPACE_DECL
+
+#endif // LLVM_LIBC_SRC_WCTYPE_TOWLOWER_H

libc/src/wctype/towupper.cpp

@@ -0,0 +1,27 @@
+//===----------------------------------------------------------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+///
+/// \file
+/// Implementation of towupper.
+///
+//===----------------------------------------------------------------------===//
+
+#include "src/wctype/towupper.h"
+#include "hdr/types/wint_t.h"
+#include "src/__support/common.h"
+#include "src/__support/wctype_utils.h"
+
+namespace LIBC_NAMESPACE_DECL {
+
+LLVM_LIBC_FUNCTION(wint_t, towupper, (wint_t c)) {
+  if (c == static_cast<wint_t>(static_cast<wchar_t>(c)))
+    return static_cast<wint_t>(internal::toupper(static_cast<wchar_t>(c)));
+  return c;
+}
+
+} // namespace LIBC_NAMESPACE_DECL

libc/src/wctype/towupper.h

@@ -0,0 +1,26 @@
+//===----------------------------------------------------------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+///
+/// \file
+/// Implementation header for towupper.
+///
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_LIBC_SRC_WCTYPE_TOWUPPER_H
+#define LLVM_LIBC_SRC_WCTYPE_TOWUPPER_H
+
+#include "hdr/types/wint_t.h"
+#include "src/__support/common.h"
+
+namespace LIBC_NAMESPACE_DECL {
+
+wint_t towupper(wint_t c);
+
+} // namespace LIBC_NAMESPACE_DECL
+
+#endif // LLVM_LIBC_SRC_WCTYPE_TOWUPPER_H

libc/test/src/__support/wctype/CMakeLists.txt

@@ -23,6 +23,7 @@ add_libc_test(
   SRCS
     wctype_perfect_hash_test.cpp
   COMPILE_OPTIONS
+    -DLIBC_ENABLE_CONSTEXPR=1
     ${CONSTEXPR_FLAGS}
   DEPENDS
     libc.src.__support.wctype.lower_to_upper

libc/test/src/__support/wctype_utils_test.cpp

@@ -9,6 +9,7 @@
 #include "hdr/types/wctype_t.h"
 #include "src/__support/macros/config.h"
 #include "src/__support/wctype/wctype_classification_utils.h"
+#include "src/__support/wctype/wctype_conversion_utils.h"
 #include "test/UnitTest/Test.h"
 namespace {
 
@@ -32,6 +33,10 @@ namespace utf8_mode {
 namespace LIBC_NAMESPACE_DECL {
 using ::LIBC_NAMESPACE::lookup_properties;
 using ::LIBC_NAMESPACE::PropertyFlag;
+namespace wctype_internal {
+using ::LIBC_NAMESPACE::wctype_internal::tolower;
+using ::LIBC_NAMESPACE::wctype_internal::toupper;
+} // namespace wctype_internal
 namespace cpp = ::LIBC_NAMESPACE::cpp;
 } // namespace LIBC_NAMESPACE_DECL
 

libc/test/src/wctype/CMakeLists.txt

@@ -42,7 +42,7 @@ add_libc_test(
     libc.src.wctype.iswgraph
 )
 
-add_libc_test(  
+add_libc_test(
   iswlower_test
   SUITE
     libc_wctype_unittests
@@ -142,3 +142,27 @@ add_libc_test(
   DEPENDS
     libc.src.wctype.wctype
 )
+
+add_libc_test(
+  towlower_test
+  SUITE
+    libc_wctype_unittests
+  SRCS
+    towlower_test.cpp
+  DEPENDS
+    libc.hdr.wchar_macros
+    libc.src.wctype.towlower
+    libc.src.__support.wctype_utils
+)
+
+add_libc_test(
+  towupper_test
+  SUITE
+    libc_wctype_unittests
+  SRCS
+    towupper_test.cpp
+  DEPENDS
+    libc.hdr.wchar_macros
+    libc.src.wctype.towupper
+    libc.src.__support.wctype_utils
+)

libc/test/src/wctype/iswprint_test.cpp

@@ -6,6 +6,7 @@
 //
 //===----------------------------------------------------------------------===//
 
+#include "src/__support/wctype_utils.h"
 #include "src/wctype/iswprint.h"
 
 #include "test/UnitTest/Test.h"
@@ -15,7 +16,21 @@ TEST(LlvmLibciswprint, SimpleTest) {
     if (' ' <= ch && ch <= '~') {
       EXPECT_NE(LIBC_NAMESPACE::iswprint(ch), 0);
     } else {
+#if LIBC_CONF_WCTYPE_MODE == LIBC_WCTYPE_MODE_UTF8
+      // In UTF-8 mode, characters above 127 can be printable.
+      if (ch < 128) {
+        EXPECT_EQ(LIBC_NAMESPACE::iswprint(ch), 0);
+      }
+#else
+      // In ASCII mode, everything outside ASCII printable is not printable.
       EXPECT_EQ(LIBC_NAMESPACE::iswprint(ch), 0);
+#endif
     }
   }
+
+#if LIBC_CONF_WCTYPE_MODE == LIBC_WCTYPE_MODE_UTF8
+  EXPECT_NE(LIBC_NAMESPACE::iswprint(L'á'), 0);
+#else
+  EXPECT_EQ(LIBC_NAMESPACE::iswprint(L'á'), 0);
+#endif
 }

libc/test/src/wctype/towlower_test.cpp

@@ -0,0 +1,85 @@
+//===----------------------------------------------------------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+///
+/// \file
+/// Unit tests for towlower.
+///
+//===----------------------------------------------------------------------===//
+
+#include "hdr/wchar_macros.h" // for WEOF
+#include "src/__support/wctype_utils.h"
+#include "src/wctype/towlower.h"
+#include "test/UnitTest/Test.h"
+
+TEST(LlvmLibcTowLower, SimpleTest) {
+  // ASCII Conversions
+  EXPECT_EQ(LIBC_NAMESPACE::towlower(L'A'), static_cast<wint_t>(L'a'));
+  EXPECT_EQ(LIBC_NAMESPACE::towlower(L'Z'), static_cast<wint_t>(L'z'));
+
+  // ASCII Unchanged
+  EXPECT_EQ(LIBC_NAMESPACE::towlower(L'a'), static_cast<wint_t>(L'a'));
+  EXPECT_EQ(LIBC_NAMESPACE::towlower(L'z'), static_cast<wint_t>(L'z'));
+  EXPECT_EQ(LIBC_NAMESPACE::towlower(L'1'), static_cast<wint_t>(L'1'));
+  EXPECT_EQ(LIBC_NAMESPACE::towlower(L'\0'), static_cast<wint_t>(L'\0'));
+
+  // WEOF Test
+  EXPECT_EQ(LIBC_NAMESPACE::towlower(WEOF), static_cast<wint_t>(WEOF));
+
+  // Boundary / Out-of-domain Tests (should return unchanged)
+  EXPECT_EQ(LIBC_NAMESPACE::towlower(0xFFFF), static_cast<wint_t>(0xFFFF));
+#if WCHAR_MAX > 0xFFFF
+  EXPECT_EQ(LIBC_NAMESPACE::towlower(0x110000), static_cast<wint_t>(0x110000));
+#endif
+
+  // Non-ASCII Lowercase (Already lowercase: should remain unchanged
+  // unconditionally in BOTH modes)
+  EXPECT_EQ(LIBC_NAMESPACE::towlower(L'α'), static_cast<wint_t>(L'α'));
+  EXPECT_EQ(LIBC_NAMESPACE::towlower(L'а'), static_cast<wint_t>(L'а'));
+  EXPECT_EQ(LIBC_NAMESPACE::towlower(L'é'), static_cast<wint_t>(L'é'));
+}
+
+#if LIBC_CONF_WCTYPE_MODE == LIBC_WCTYPE_MODE_UTF8
+TEST(LlvmLibcTowLower, NonAsciiTestUtf8) {
+  // Greek conversions
+  EXPECT_EQ(LIBC_NAMESPACE::towlower(L'Α'), static_cast<wint_t>(L'α')); // alpha
+  EXPECT_EQ(LIBC_NAMESPACE::towlower(L'Ω'), static_cast<wint_t>(L'ω')); // omega
+
+  // Cyrillic conversions
+  EXPECT_EQ(LIBC_NAMESPACE::towlower(L'А'), static_cast<wint_t>(L'а')); // A
+  EXPECT_EQ(LIBC_NAMESPACE::towlower(L'Я'), static_cast<wint_t>(L'я')); // Ya
+
+  // Accented Latin
+  EXPECT_EQ(LIBC_NAMESPACE::towlower(L'É'), static_cast<wint_t>(L'é'));
+  EXPECT_EQ(LIBC_NAMESPACE::towlower(L'Ü'), static_cast<wint_t>(L'ü'));
+
+#if WCHAR_MAX > 0xFFFF
+  // Deseret (Unicode Plane 1) conversions
+  EXPECT_EQ(LIBC_NAMESPACE::towlower(L'\U00010400'),
+            static_cast<wint_t>(L'\U00010428'));
+#endif
+}
+#endif
+
+#if LIBC_CONF_WCTYPE_MODE == LIBC_WCTYPE_MODE_ASCII
+TEST(LlvmLibcTowLower, NonAsciiTestAscii) {
+  // Non-ASCII uppercase characters must return unchanged under ASCII-only mode
+  EXPECT_EQ(LIBC_NAMESPACE::towlower(L'Α'), static_cast<wint_t>(L'Α'));
+  EXPECT_EQ(LIBC_NAMESPACE::towlower(L'Ω'), static_cast<wint_t>(L'Ω'));
+
+  EXPECT_EQ(LIBC_NAMESPACE::towlower(L'А'), static_cast<wint_t>(L'А'));
+  EXPECT_EQ(LIBC_NAMESPACE::towlower(L'Я'), static_cast<wint_t>(L'Я'));
+
+  EXPECT_EQ(LIBC_NAMESPACE::towlower(L'É'), static_cast<wint_t>(L'É'));
+  EXPECT_EQ(LIBC_NAMESPACE::towlower(L'Ü'), static_cast<wint_t>(L'Ü'));
+
+#if WCHAR_MAX > 0xFFFF
+  EXPECT_EQ(LIBC_NAMESPACE::towlower(L'\U00010400'),
+            static_cast<wint_t>(L'\U00010400'));
+#endif
+}
+#endif

libc/test/src/wctype/towupper_test.cpp

@@ -0,0 +1,85 @@
+//===----------------------------------------------------------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+///
+/// \file
+/// Unit tests for towupper.
+///
+//===----------------------------------------------------------------------===//
+
+#include "hdr/wchar_macros.h" // for WEOF
+#include "src/__support/wctype_utils.h"
+#include "src/wctype/towupper.h"
+#include "test/UnitTest/Test.h"
+
+TEST(LlvmLibcTowUpper, SimpleTest) {
+  // ASCII Conversions
+  EXPECT_EQ(LIBC_NAMESPACE::towupper(L'a'), static_cast<wint_t>(L'A'));
+  EXPECT_EQ(LIBC_NAMESPACE::towupper(L'z'), static_cast<wint_t>(L'Z'));
+
+  // ASCII Unchanged
+  EXPECT_EQ(LIBC_NAMESPACE::towupper(L'A'), static_cast<wint_t>(L'A'));
+  EXPECT_EQ(LIBC_NAMESPACE::towupper(L'Z'), static_cast<wint_t>(L'Z'));
+  EXPECT_EQ(LIBC_NAMESPACE::towupper(L'1'), static_cast<wint_t>(L'1'));
+  EXPECT_EQ(LIBC_NAMESPACE::towupper(L'\0'), static_cast<wint_t>(L'\0'));
+
+  // WEOF Test
+  EXPECT_EQ(LIBC_NAMESPACE::towupper(WEOF), static_cast<wint_t>(WEOF));
+
+  // Boundary / Out-of-domain Tests (should return unchanged)
+  EXPECT_EQ(LIBC_NAMESPACE::towupper(0xFFFF), static_cast<wint_t>(0xFFFF));
+#if WCHAR_MAX > 0xFFFF
+  EXPECT_EQ(LIBC_NAMESPACE::towupper(0x110000), static_cast<wint_t>(0x110000));
+#endif
+
+  // Non-ASCII Uppercase (Already uppercase: should remain unchanged
+  // unconditionally in BOTH modes)
+  EXPECT_EQ(LIBC_NAMESPACE::towupper(L'Α'), static_cast<wint_t>(L'Α'));
+  EXPECT_EQ(LIBC_NAMESPACE::towupper(L'А'), static_cast<wint_t>(L'А'));
+  EXPECT_EQ(LIBC_NAMESPACE::towupper(L'É'), static_cast<wint_t>(L'É'));
+}
+
+#if LIBC_CONF_WCTYPE_MODE == LIBC_WCTYPE_MODE_UTF8
+TEST(LlvmLibcTowUpper, NonAsciiTestUtf8) {
+  // Greek conversions
+  EXPECT_EQ(LIBC_NAMESPACE::towupper(L'α'), static_cast<wint_t>(L'Α')); // alpha
+  EXPECT_EQ(LIBC_NAMESPACE::towupper(L'ω'), static_cast<wint_t>(L'Ω')); // omega
+
+  // Cyrillic conversions
+  EXPECT_EQ(LIBC_NAMESPACE::towupper(L'а'), static_cast<wint_t>(L'А')); // A
+  EXPECT_EQ(LIBC_NAMESPACE::towupper(L'я'), static_cast<wint_t>(L'Я')); // Ya
+
+  // Accented Latin
+  EXPECT_EQ(LIBC_NAMESPACE::towupper(L'é'), static_cast<wint_t>(L'É'));
+  EXPECT_EQ(LIBC_NAMESPACE::towupper(L'ü'), static_cast<wint_t>(L'Ü'));
+
+#if WCHAR_MAX > 0xFFFF
+  // Deseret (Unicode Plane 1) conversions
+  EXPECT_EQ(LIBC_NAMESPACE::towupper(L'\U00010428'),
+            static_cast<wint_t>(L'\U00010400'));
+#endif
+}
+#endif
+
+#if LIBC_CONF_WCTYPE_MODE == LIBC_WCTYPE_MODE_ASCII
+TEST(LlvmLibcTowUpper, NonAsciiTestAscii) {
+  // Non-ASCII lowercase characters must return unchanged under ASCII-only mode
+  EXPECT_EQ(LIBC_NAMESPACE::towupper(L'α'), static_cast<wint_t>(L'α'));
+  EXPECT_EQ(LIBC_NAMESPACE::towupper(L'ω'), static_cast<wint_t>(L'ω'));
+
+  EXPECT_EQ(LIBC_NAMESPACE::towupper(L'а'), static_cast<wint_t>(L'а'));
+  EXPECT_EQ(LIBC_NAMESPACE::towupper(L'я'), static_cast<wint_t>(L'я'));
+
+  EXPECT_EQ(LIBC_NAMESPACE::towupper(L'é'), static_cast<wint_t>(L'é'));
+  EXPECT_EQ(LIBC_NAMESPACE::towupper(L'ü'), static_cast<wint_t>(L'ü'));
+
+#if WCHAR_MAX > 0xFFFF
+  EXPECT_EQ(LIBC_NAMESPACE::towupper(L'\U00010428'),
+            static_cast<wint_t>(L'\U00010428'));
+#endif
+}
+#endif