[libc] mbrtowc implementation (#144760)

implemented the internal and public mbrtowc as well as tests for the
public function.

---------

Co-authored-by: Sriya Pratipati <sriyap@google.com>

libc/config/linux/x86_64/entrypoints.txt

@@ -1244,6 +1244,9 @@ if(LLVM_LIBC_FULL_BUILD)
     libc.src.sys.socket.recv
     libc.src.sys.socket.recvfrom
     libc.src.sys.socket.recvmsg
+
+    # wchar.h entrypoints
+    libc.src.wchar.mbrtowc
   )
 endif()
 

libc/hdr/types/CMakeLists.txt

@@ -20,6 +20,14 @@ add_proxy_header_library(
     libc.include.uchar
 )
 
+add_proxy_header_library(
+  mbstate_t
+  HDRS
+    mbstate_t.h
+  DEPENDS
+    libc.include.llvm-libc-types.mbstate_t
+)
+
 add_proxy_header_library(
   div_t
   HDRS

libc/hdr/types/mbstate_t.h

@@ -0,0 +1,22 @@
+//===-- Definition of macros from mbstate_t.h -----------------------------===//
+//
+// 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_HDR_TYPES_MBSTATE_T_H
+#define LLVM_LIBC_HDR_TYPES_MBSTATE_T_H
+
+#ifdef LIBC_FULL_BUILD
+
+#include "include/llvm-libc-types/mbstate_t.h"
+
+#else // Overlay mode
+
+#error "Cannot overlay mbstate_t
+
+#endif // LLVM_LIBC_FULL_BUILD
+
+#endif // LLVM_LIBC_HDR_TYPES_MBSTATE_T_H

libc/include/llvm-libc-types/mbstate_t.h

@@ -9,8 +9,12 @@
 #ifndef LLVM_LIBC_TYPES_MBSTATE_T_H
 #define LLVM_LIBC_TYPES_MBSTATE_T_H
 
-// TODO: Complete this once we implement functions that operate on this type.
+#include "../llvm-libc-macros/stdint-macros.h"
+
 typedef struct {
+  uint32_t __field1;
+  uint8_t __field2;
+  uint8_t __field3;
 } mbstate_t;
 
 #endif // LLVM_LIBC_TYPES_MBSTATE_T_H

libc/include/wchar.yaml

@@ -29,6 +29,15 @@ functions:
     return_type: wint_t
     arguments:
       - type: int
+  - name: mbrtowc
+    standards:
+      - stdc
+    return_type: size_t
+    arguments:
+      - type: wchar_t *__restrict
+      - type: const char *__restrict
+      - type: size_t
+      - type: mbstate_t *__restrict
   - name: wmemset
     standards:
       - stdc

libc/src/__support/wchar/CMakeLists.txt

@@ -19,3 +19,19 @@ add_object_library(
     libc.src.__support.math_extras
     .mbstate
 )
+
+add_object_library(
+  mbrtowc
+  HDRS
+    mbrtowc.h
+  SRCS
+    mbrtowc.cpp
+  DEPENDS
+  libc.hdr.types.wchar_t
+  libc.hdr.types.size_t
+  libc.src.__support.common
+  libc.src.__support.error_or
+  libc.src.__support.macros.config
+  .character_converter
+  .mbstate
+)

libc/src/__support/wchar/mbrtowc.cpp

@@ -0,0 +1,49 @@
+//===-- Implementation for mbrtowc function ---------------------*- C++ -*-===//
+//
+// 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/__support/wchar/mbrtowc.h"
+#include "hdr/types/mbstate_t.h"
+#include "hdr/types/size_t.h"
+#include "hdr/types/wchar_t.h"
+#include "src/__support/common.h"
+#include "src/__support/error_or.h"
+#include "src/__support/macros/config.h"
+#include "src/__support/wchar/character_converter.h"
+#include "src/__support/wchar/mbstate.h"
+
+namespace LIBC_NAMESPACE_DECL {
+namespace internal {
+
+ErrorOr<size_t> mbrtowc(wchar_t *__restrict pwc, const char *__restrict s,
+                        size_t n, mbstate *__restrict ps) {
+  CharacterConverter char_conv(ps);
+  if (s == nullptr)
+    return 0;
+  size_t i = 0;
+  // Reading in bytes until we have a complete wc or error
+  for (; i < n && !char_conv.isFull(); ++i) {
+    int err = char_conv.push(static_cast<char8_t>(s[i]));
+    // Encoding error
+    if (err == -1)
+      return Error(-1);
+  }
+  auto wc = char_conv.pop_utf32();
+  if (wc.has_value()) {
+    *pwc = wc.value();
+    // null terminator -> return 0
+    if (wc.value() == L'\0')
+      return 0;
+    return i;
+  }
+  // Incomplete but potentially valid
+  return -2;
+}
+
+} // namespace internal
+
+} // namespace LIBC_NAMESPACE_DECL

libc/src/__support/wchar/mbrtowc.h

@@ -0,0 +1,29 @@
+//===-- Implementation header for mbrtowc function --------------*- C++ -*-===//
+//
+// 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___SUPPORT_WCHAR_MBRTOWC
+#define LLVM_LIBC_SRC___SUPPORT_WCHAR_MBRTOWC
+
+#include "hdr/types/size_t.h"
+#include "hdr/types/wchar_t.h"
+#include "src/__support/common.h"
+#include "src/__support/error_or.h"
+#include "src/__support/macros/config.h"
+#include "src/__support/wchar/mbstate.h"
+
+namespace LIBC_NAMESPACE_DECL {
+namespace internal {
+
+ErrorOr<size_t> mbrtowc(wchar_t *__restrict pwc, const char *__restrict s,
+                        size_t n, mbstate *__restrict ps);
+
+} // namespace internal
+
+} // namespace LIBC_NAMESPACE_DECL
+
+#endif // LLVM_LIBC_SRC___SUPPORT_WCHAR_MBRTOWC

libc/src/__support/wchar/mbstate.h

@@ -18,17 +18,17 @@ namespace internal {
 
 struct mbstate {
   // store a partial codepoint (in UTF-32)
-  char32_t partial;
+  char32_t partial = 0;
 
   /*
   Progress towards a conversion
     Increases with each push(...) until it reaches total_bytes
     Decreases with each pop(...) until it reaches 0
   */
-  uint8_t bytes_stored;
+  uint8_t bytes_stored = 0;
 
   // Total number of bytes that will be needed to represent this character
-  uint8_t total_bytes;
+  uint8_t total_bytes = 0;
 };
 
 } // namespace internal

libc/src/wchar/CMakeLists.txt

@@ -34,6 +34,23 @@ add_entrypoint_object(
     libc.src.__support.wctype_utils
 )
 
+add_entrypoint_object(
+  mbrtowc
+  SRCS
+    mbrtowc.cpp
+  HDRS
+    mbrtowc.h
+  DEPENDS
+    libc.hdr.types.size_t
+    libc.hdr.types.mbstate_t
+    libc.hdr.types.wchar_t
+    libc.src.__support.common
+    libc.src.__support.macros.config
+    libc.src.__support.wchar.mbrtowc
+    libc.src.__support.libc_errno
+    libc.src.__support.wchar.mbstate
+)
+
 add_entrypoint_object(
   wmemset
   SRCS

libc/src/wchar/mbrtowc.cpp

@@ -0,0 +1,38 @@
+//===-- Implementation of mbrtowc -----------------------------------------===//
+//
+// 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/mbrtowc.h"
+
+#include "hdr/types/mbstate_t.h"
+#include "hdr/types/size_t.h"
+#include "hdr/types/wchar_t.h"
+#include "src/__support/common.h"
+#include "src/__support/libc_errno.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(size_t, mbrtowc,
+                   (wchar_t *__restrict pwc, const char *__restrict s, size_t n,
+                    mbstate_t *__restrict ps)) {
+  static internal::mbstate internal_mbstate;
+  auto ret = internal::mbrtowc(pwc, s, n,
+                               ps == nullptr
+                                   ? &internal_mbstate
+                                   : reinterpret_cast<internal::mbstate *>(ps));
+  if (!ret.has_value()) {
+    // Encoding failure
+    libc_errno = EILSEQ;
+    return -1;
+  }
+  return ret.value();
+}
+
+} // namespace LIBC_NAMESPACE_DECL

libc/src/wchar/mbrtowc.h

@@ -0,0 +1,24 @@
+//===-- Implementation header for mbrtowc ---------------------------------===//
+//
+// 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_MBRTOWC_H
+#define LLVM_LIBC_SRC_WCHAR_MBRTOWC_H
+
+#include "hdr/types/mbstate_t.h"
+#include "hdr/types/size_t.h"
+#include "hdr/types/wchar_t.h"
+#include "src/__support/macros/config.h"
+
+namespace LIBC_NAMESPACE_DECL {
+
+size_t mbrtowc(wchar_t *__restrict pwc, const char *__restrict s, size_t n,
+               mbstate_t *__restrict ps);
+
+} // namespace LIBC_NAMESPACE_DECL
+
+#endif // LLVM_LIBC_SRC_WCHAR_MBRTOWC_H

libc/test/src/wchar/CMakeLists.txt

@@ -23,6 +23,20 @@ add_libc_test(
     libc.src.wchar.btowc
 )
 
+add_libc_test(
+  mbrtowc_test
+  SUITE
+    libc_wchar_unittests
+  SRCS
+    mbrtowc_test.cpp
+  DEPENDS
+    libc.src.__support.libc_errno
+    libc.src.string.memset
+    libc.src.wchar.mbrtowc
+    libc.hdr.types.mbstate_t
+    libc.hdr.types.wchar_t
+)
+
 add_libc_test(
   wctob_test
   SUITE

libc/test/src/wchar/mbrtowc_test.cpp

@@ -0,0 +1,172 @@
+//===-- Unittests for mbrtowc ---------------------------------------------===//
+//
+// 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/mbstate_t.h"
+#include "hdr/types/wchar_t.h"
+#include "src/__support/libc_errno.h"
+#include "src/string/memset.h"
+#include "src/wchar/mbrtowc.h"
+#include "test/UnitTest/Test.h"
+
+TEST(LlvmLibcMBRToWC, OneByte) {
+  const char *ch = "A";
+  wchar_t dest[2];
+  // Testing if it works with nullptr mbstate_t
+  mbstate_t *mb = nullptr;
+  size_t n = LIBC_NAMESPACE::mbrtowc(dest, ch, 1, mb);
+  ASSERT_EQ(static_cast<char>(*dest), 'A');
+  ASSERT_EQ(static_cast<int>(n), 1);
+
+  // Should fail since we have not read enough
+  n = LIBC_NAMESPACE::mbrtowc(dest, ch, 0, mb);
+  ASSERT_EQ(static_cast<int>(n), -2);
+}
+
+TEST(LlvmLibcMBRToWC, TwoByte) {
+  const char ch[2] = {static_cast<char>(0xC2),
+                      static_cast<char>(0x8E)}; // Ž car symbol
+  wchar_t dest[2];
+  mbstate_t *mb;
+  LIBC_NAMESPACE::memset(&mb, 0, sizeof(mbstate_t));
+  size_t n = LIBC_NAMESPACE::mbrtowc(dest, ch, 2, mb);
+  ASSERT_EQ(static_cast<int>(*dest), 142);
+  ASSERT_EQ(static_cast<int>(n), 2);
+
+  // Should fail since we have not read enough
+  n = LIBC_NAMESPACE::mbrtowc(dest, ch, 1, mb);
+  ASSERT_EQ(static_cast<int>(n), -2);
+  // Should pass after reading one more byte
+  n = LIBC_NAMESPACE::mbrtowc(dest, ch + 1, 1, mb);
+  ASSERT_EQ(static_cast<int>(n), 1);
+  ASSERT_EQ(static_cast<int>(*dest), 142);
+}
+
+TEST(LlvmLibcMBRToWC, ThreeByte) {
+  const char ch[3] = {static_cast<char>(0xE2), static_cast<char>(0x88),
+                      static_cast<char>(0x91)}; // ∑ sigma symbol
+  wchar_t dest[2];
+  mbstate_t *mb;
+  LIBC_NAMESPACE::memset(&mb, 0, sizeof(mbstate_t));
+  size_t n = LIBC_NAMESPACE::mbrtowc(dest, ch, 3, mb);
+  ASSERT_EQ(static_cast<int>(*dest), 8721);
+  ASSERT_EQ(static_cast<int>(n), 3);
+
+  // Should fail since we have not read enough
+  n = LIBC_NAMESPACE::mbrtowc(dest, ch, 1, mb);
+  ASSERT_EQ(static_cast<int>(n), -2);
+  // Should pass after reading two more bytes
+  n = LIBC_NAMESPACE::mbrtowc(dest, ch + 1, 2, mb);
+  ASSERT_EQ(static_cast<int>(n), 2);
+  ASSERT_EQ(static_cast<int>(*dest), 8721);
+}
+
+TEST(LlvmLibcMBRToWC, 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];
+  mbstate_t *mb;
+  LIBC_NAMESPACE::memset(&mb, 0, sizeof(mbstate_t));
+  size_t n = LIBC_NAMESPACE::mbrtowc(dest, ch, 4, mb);
+  ASSERT_EQ(static_cast<int>(*dest), 129313);
+  ASSERT_EQ(static_cast<int>(n), 4);
+
+  // Should fail since we have not read enough
+  n = LIBC_NAMESPACE::mbrtowc(dest, ch, 2, mb);
+  ASSERT_EQ(static_cast<int>(n), -2);
+  // Should pass after reading two more bytes
+  n = LIBC_NAMESPACE::mbrtowc(dest, ch + 2, 2, mb);
+  ASSERT_EQ(static_cast<int>(n), 2);
+  ASSERT_EQ(static_cast<int>(*dest), 129313);
+}
+
+TEST(LlvmLibcMBRToWC, InvalidByte) {
+  const char ch[1] = {static_cast<char>(0x80)};
+  wchar_t dest[2];
+  mbstate_t *mb;
+  LIBC_NAMESPACE::memset(&mb, 0, sizeof(mbstate_t));
+  size_t n = LIBC_NAMESPACE::mbrtowc(dest, ch, 1, mb);
+  ASSERT_EQ(static_cast<int>(n), -1);
+  ASSERT_EQ(static_cast<int>(libc_errno), EILSEQ);
+}
+
+TEST(LlvmLibcMBRToWC, 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];
+  mbstate_t *mb;
+  LIBC_NAMESPACE::memset(&mb, 0, sizeof(mbstate_t));
+  // Trying to push all 4 should error
+  size_t n = LIBC_NAMESPACE::mbrtowc(dest, ch, 4, mb);
+  ASSERT_EQ(static_cast<int>(n), -1);
+  ASSERT_EQ(static_cast<int>(libc_errno), EILSEQ);
+  // Trying to push just the first one should error
+  n = LIBC_NAMESPACE::mbrtowc(dest, ch, 1, mb);
+  ASSERT_EQ(static_cast<int>(n), -1);
+  ASSERT_EQ(static_cast<int>(libc_errno), EILSEQ);
+  // Trying to push the second and third should correspond to null wc
+  n = LIBC_NAMESPACE::mbrtowc(dest, ch + 1, 2, mb);
+  ASSERT_EQ(static_cast<int>(n), 0);
+  ASSERT_TRUE(*dest == L'\0');
+}
+
+TEST(LlvmLibcMBRToWC, 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];
+  mbstate_t *mb;
+  LIBC_NAMESPACE::memset(&mb, 0, sizeof(mbstate_t));
+  // Trying to push all 4 should error
+  size_t n = LIBC_NAMESPACE::mbrtowc(dest, ch, 4, mb);
+  ASSERT_EQ(static_cast<int>(n), -1);
+  ASSERT_EQ(static_cast<int>(libc_errno), EILSEQ);
+}
+
+TEST(LlvmLibcMBRToWC, ValidTwoByteWithExtraRead) {
+  const char ch[3] = {static_cast<char>(0xC2), static_cast<char>(0x8E),
+                      static_cast<char>(0x80)};
+  wchar_t dest[2];
+  mbstate_t *mb;
+  LIBC_NAMESPACE::memset(&mb, 0, sizeof(mbstate_t));
+  // Trying to push all 3 should return valid 2 byte
+  size_t n = LIBC_NAMESPACE::mbrtowc(dest, ch, 3, mb);
+  ASSERT_EQ(static_cast<int>(n), 2);
+  ASSERT_EQ(static_cast<int>(*dest), 142);
+}
+
+TEST(LlvmLibcMBRToWC, 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];
+  mbstate_t *mb;
+  LIBC_NAMESPACE::memset(&mb, 0, sizeof(mbstate_t));
+  // mbstate should reset after reading first one
+  size_t n = LIBC_NAMESPACE::mbrtowc(dest, ch, 2, mb);
+  ASSERT_EQ(static_cast<int>(n), 2);
+  ASSERT_EQ(static_cast<int>(*dest), 142);
+  n = LIBC_NAMESPACE::mbrtowc(dest + 1, ch + 2, 2, mb);
+  ASSERT_EQ(static_cast<int>(n), 2);
+  ASSERT_EQ(static_cast<int>(*(dest + 1)), 460);
+}
+
+TEST(LlvmLibcMBRToWC, NullString) {
+  wchar_t dest[2] = {L'O', L'K'};
+  mbstate_t *mb;
+  LIBC_NAMESPACE::memset(&mb, 0, sizeof(mbstate_t));
+  // reading on nullptr should return 0
+  size_t n = LIBC_NAMESPACE::mbrtowc(dest, nullptr, 2, mb);
+  ASSERT_EQ(static_cast<int>(n), 0);
+  ASSERT_TRUE(dest[0] == L'O');
+  // reading a null terminator should return 0
+  const char *ch = "\0";
+  n = LIBC_NAMESPACE::mbrtowc(dest, ch, 1, mb);
+  ASSERT_EQ(static_cast<int>(n), 0);
+}