[libc] mbtowc implementation

[sribee8]

Implemented mbtowcs and tests for the function.

[llvmbot]

@llvm/pr-subscribers-libc

Author: None (sribee8)

Changes

Implemented mbtowcs and tests for the function.

---

Full diff: https://github.com/llvm/llvm-project/pull/145405.diff

7 Files Affected:

• (modified) libc/config/linux/x86_64/entrypoints.txt (+1)
• (modified) libc/include/wchar.yaml (+8)
• (modified) libc/src/wchar/CMakeLists.txt (+15)
• (added) libc/src/wchar/mbtowc.cpp (+34)
• (added) libc/src/wchar/mbtowc.h (+22)
• (modified) libc/test/src/wchar/CMakeLists.txt (+12)
• (added) libc/test/src/wchar/mbtowc_test.cpp (+132)

diff --git a/libc/config/linux/x86_64/entrypoints.txt b/libc/config/linux/x86_64/entrypoints.txt
index e1a2a26479de9..f0e17d6a2544f 100644
--- a/libc/config/linux/x86_64/entrypoints.txt
+++ b/libc/config/linux/x86_64/entrypoints.txt
@@ -1248,6 +1248,7 @@ if(LLVM_LIBC_FULL_BUILD)
 
     # wchar.h entrypoints
     libc.src.wchar.mbrtowc
+    libc.src.wchar.mbtowc
     libc.src.wchar.wcrtomb
   )
 endif()
diff --git a/libc/include/wchar.yaml b/libc/include/wchar.yaml
index 0a6a75ebbbf91..d5044e1728f80 100644
--- a/libc/include/wchar.yaml
+++ b/libc/include/wchar.yaml
@@ -38,6 +38,14 @@ functions:
       - type: const char *__restrict
       - type: size_t
       - type: mbstate_t *__restrict
+  - name: mbtowc
+    standards:
+      - stdc
+    return_type: int
+    arguments:
+      - type: wchar_t *__restrict
+      - type: const char *__restrict
+      - type: size_t
   - name: wmemset
     standards:
       - stdc
diff --git a/libc/src/wchar/CMakeLists.txt b/libc/src/wchar/CMakeLists.txt
index 476cf38f4662e..fe1a9e7cc6e17 100644
--- a/libc/src/wchar/CMakeLists.txt
+++ b/libc/src/wchar/CMakeLists.txt
@@ -65,6 +65,21 @@ add_entrypoint_object(
     libc.src.__support.wchar.mbstate
 )
 
+add_entrypoint_object(
+  mbtowc
+  SRCS
+    mbtowc.cpp
+  HDRS
+    mbtowc.h
+  DEPENDS
+    libc.hdr.types.size_t
+    libc.hdr.types.wchar_t
+    libc.src.__support.common
+    libc.src.__support.macros.config
+    libc.src.__support.wchar.mbrtowc
+    libc.src.__support.wchar.mbstate
+)
+
 add_entrypoint_object(
   wmemset
   SRCS
diff --git a/libc/src/wchar/mbtowc.cpp b/libc/src/wchar/mbtowc.cpp
new file mode 100644
index 0000000000000..128fe9e84a201
--- /dev/null
+++ b/libc/src/wchar/mbtowc.cpp
@@ -0,0 +1,34 @@
+//===-- Implementation of mbtowc -----------------------------------------===//
+//
+// 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
+//
+//===----------------------------------------------------------------------===//
+
+#include "src/wchar/mbtowc.h"
+
+#include "hdr/types/size_t.h"
+#include "hdr/types/wchar_t.h"
+#include "src/__support/common.h"
+#include "src/__support/macros/config.h"
+#include "src/__support/wchar/mbrtowc.h"
+#include "src/__support/wchar/mbstate.h"
+
+namespace LIBC_NAMESPACE_DECL {
+
+LLVM_LIBC_FUNCTION(int, mbtowc,
+                   (wchar_t *__restrict pwc, const char *__restrict s,
+                    size_t n)) {
+  if (s == nullptr)
+    return 0;
+  internal::mbstate internal_mbstate;
+  auto ret = internal::mbrtowc(pwc, s, n, &internal_mbstate);
+  if (!ret.has_value() || static_cast<int>(ret.value()) == -2) {
+    // Encoding failure
+    return -1;
+  }
+  return static_cast<int>(ret.value());
+}
+
+} // namespace LIBC_NAMESPACE_DECL
diff --git a/libc/src/wchar/mbtowc.h b/libc/src/wchar/mbtowc.h
new file mode 100644
index 0000000000000..f974197f81b58
--- /dev/null
+++ b/libc/src/wchar/mbtowc.h
@@ -0,0 +1,22 @@
+//===-- Implementation header for mbtowc ---------------------------------===//
+//
+// 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
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_LIBC_SRC_WCHAR_MBTOWC_H
+#define LLVM_LIBC_SRC_WCHAR_MBTOWC_H
+
+#include "hdr/types/size_t.h"
+#include "hdr/types/wchar_t.h"
+#include "src/__support/macros/config.h"
+
+namespace LIBC_NAMESPACE_DECL {
+
+int mbtowc(wchar_t *__restrict pwc, const char *__restrict s, size_t n);
+
+} // namespace LIBC_NAMESPACE_DECL
+
+#endif // LLVM_LIBC_SRC_WCHAR_MBTOWC_H
diff --git a/libc/test/src/wchar/CMakeLists.txt b/libc/test/src/wchar/CMakeLists.txt
index c932f3632c7ab..510f01d33fbeb 100644
--- a/libc/test/src/wchar/CMakeLists.txt
+++ b/libc/test/src/wchar/CMakeLists.txt
@@ -37,6 +37,18 @@ add_libc_test(
     libc.hdr.types.wchar_t
 )
 
+add_libc_test(
+  mbtowc_test
+  SUITE
+    libc_wchar_unittests
+  SRCS
+    mbtowc_test.cpp
+  DEPENDS
+    libc.src.__support.libc_errno
+    libc.src.wchar.mbtowc
+    libc.hdr.types.wchar_t
+)
+
 add_libc_test(
   wctob_test
   SUITE
diff --git a/libc/test/src/wchar/mbtowc_test.cpp b/libc/test/src/wchar/mbtowc_test.cpp
new file mode 100644
index 0000000000000..a3bd654d3ef04
--- /dev/null
+++ b/libc/test/src/wchar/mbtowc_test.cpp
@@ -0,0 +1,132 @@
+//===-- Unittests for mbtowc ---------------------------------------------===//
+//
+// 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
+//
+//===----------------------------------------------------------------------===//
+
+#include "hdr/types/wchar_t.h"
+#include "src/wchar/mbtowc.h"
+#include "test/UnitTest/Test.h"
+
+TEST(LlvmLibcMBToWC, OneByte) {
+  const char *ch = "A";
+  wchar_t dest[2];
+  int n = LIBC_NAMESPACE::mbtowc(dest, ch, 1);
+  ASSERT_EQ(static_cast<char>(*dest), 'A');
+  ASSERT_EQ(n, 1);
+
+  // Should fail since we have not read enough
+  n = LIBC_NAMESPACE::mbtowc(dest, ch, 0);
+  ASSERT_EQ(n, -1);
+}
+
+TEST(LlvmLibcMBToWC, TwoByte) {
+  const char ch[2] = {static_cast<char>(0xC2),
+                      static_cast<char>(0x8E)}; // � car symbol
+  wchar_t dest[2];
+  int n = LIBC_NAMESPACE::mbtowc(dest, ch, 2);
+  ASSERT_EQ(static_cast<int>(*dest), 142);
+  ASSERT_EQ(n, 2);
+
+  // Should fail since we have not read enough
+  n = LIBC_NAMESPACE::mbtowc(dest, ch, 1);
+  ASSERT_EQ(n, -1);
+  // Should fail after trying to read next byte too
+  n = LIBC_NAMESPACE::mbtowc(dest, ch + 1, 1);
+  ASSERT_EQ(n, -1);
+}
+
+TEST(LlvmLibcMBToWC, ThreeByte) {
+  const char ch[3] = {static_cast<char>(0xE2), static_cast<char>(0x88),
+                      static_cast<char>(0x91)}; // ∑ sigma symbol
+  wchar_t dest[2];
+  int n = LIBC_NAMESPACE::mbtowc(dest, ch, 3);
+  ASSERT_EQ(static_cast<int>(*dest), 8721);
+  ASSERT_EQ(n, 3);
+
+  // Should fail since we have not read enough
+  n = LIBC_NAMESPACE::mbtowc(dest, ch, 2);
+  ASSERT_EQ(n, -1);
+}
+
+TEST(LlvmLibcMBToWC, FourByte) {
+  const char ch[4] = {static_cast<char>(0xF0), static_cast<char>(0x9F),
+                      static_cast<char>(0xA4),
+                      static_cast<char>(0xA1)}; // 🤡 clown emoji
+  wchar_t dest[2];
+  int n = LIBC_NAMESPACE::mbtowc(dest, ch, 4);
+  ASSERT_EQ(static_cast<int>(*dest), 129313);
+  ASSERT_EQ(n, 4);
+
+  // Should fail since we have not read enough
+  n = LIBC_NAMESPACE::mbtowc(dest, ch, 2);
+  ASSERT_EQ(n, -1);
+}
+
+TEST(LlvmLibcMBToWC, InvalidByte) {
+  const char ch[1] = {static_cast<char>(0x80)};
+  wchar_t dest[2];
+  int n = LIBC_NAMESPACE::mbtowc(dest, ch, 1);
+  ASSERT_EQ(n, -1);
+}
+
+TEST(LlvmLibcMBToWC, InvalidMultiByte) {
+  const char ch[4] = {static_cast<char>(0x80), static_cast<char>(0x00),
+                      static_cast<char>(0x80),
+                      static_cast<char>(0x00)}; // invalid sequence of bytes
+  wchar_t dest[2];
+  // Trying to push all 4 should error
+  int n = LIBC_NAMESPACE::mbtowc(dest, ch, 4);
+  ASSERT_EQ(n, -1);
+  // Trying to push the second and third should correspond to null wc
+  n = LIBC_NAMESPACE::mbtowc(dest, ch + 1, 2);
+  ASSERT_EQ(n, 0);
+  ASSERT_TRUE(*dest == L'\0');
+}
+
+TEST(LlvmLibcMBToWC, InvalidLastByte) {
+  // Last byte is invalid since it does not have correct starting sequence.
+  // 0xC0 --> 11000000 starting sequence should be 10xxxxxx
+  const char ch[4] = {static_cast<char>(0xF1), static_cast<char>(0x80),
+                      static_cast<char>(0x80), static_cast<char>(0xC0)};
+  wchar_t dest[2];
+  // Trying to push all 4 should error
+  int n = LIBC_NAMESPACE::mbtowc(dest, ch, 4);
+  ASSERT_EQ(n, -1);
+}
+
+TEST(LlvmLibcMBToWC, ValidTwoByteWithExtraRead) {
+  const char ch[3] = {static_cast<char>(0xC2), static_cast<char>(0x8E),
+                      static_cast<char>(0x80)};
+  wchar_t dest[2];
+  // Trying to push all 3 should return valid 2 byte
+  int n = LIBC_NAMESPACE::mbtowc(dest, ch, 3);
+  ASSERT_EQ(n, 2);
+  ASSERT_EQ(static_cast<int>(*dest), 142);
+}
+
+TEST(LlvmLibcMBToWC, TwoValidTwoBytes) {
+  const char ch[4] = {static_cast<char>(0xC2), static_cast<char>(0x8E),
+                      static_cast<char>(0xC7), static_cast<char>(0x8C)};
+  wchar_t dest[2];
+  int n = LIBC_NAMESPACE::mbtowc(dest, ch, 2);
+  ASSERT_EQ(n, 2);
+  ASSERT_EQ(static_cast<int>(*dest), 142);
+  n = LIBC_NAMESPACE::mbtowc(dest + 1, ch + 2, 2);
+  ASSERT_EQ(n, 2);
+  ASSERT_EQ(static_cast<int>(*(dest + 1)), 460);
+}
+
+TEST(LlvmLibcMBToWC, NullString) {
+  wchar_t dest[2] = {L'O', L'K'};
+  // reading on nullptr should return 0
+  int n = LIBC_NAMESPACE::mbtowc(dest, nullptr, 2);
+  ASSERT_EQ(n, 0);
+  ASSERT_TRUE(dest[0] == L'O');
+  // reading a null terminator should return 0
+  const char *ch = "\0";
+  n = LIBC_NAMESPACE::mbtowc(dest, ch, 1);
+  ASSERT_EQ(n, 0);
+}

[michaelrj-google]

https://pubs.opengroup.org/onlinepubs/9799919799/functions/mbtowc.html

Posix clarifies some of the behavior since the C standard is very vague.

[michaelrj-google]

LGTM, tests pass on my machine