[libc-commits] [libc] [libc] mbrtowc implementation (PR #144760)

[via libc-commits]

llvmbot wrote:


<!--LLVM PR SUMMARY COMMENT-->

@llvm/pr-subscribers-libc

Author: None (sribee8)

<details>
<summary>Changes</summary>

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

---
Full diff: https://github.com/llvm/llvm-project/pull/144760.diff


13 Files Affected:

- (modified) libc/config/linux/x86_64/entrypoints.txt (+1) 
- (modified) libc/hdr/types/CMakeLists.txt (+8) 
- (added) libc/hdr/types/mbstate_t.h (+22) 
- (modified) libc/include/llvm-libc-types/mbstate_t.h (+5-1) 
- (modified) libc/include/wchar.yaml (+9) 
- (modified) libc/src/__support/wchar/CMakeLists.txt (+16) 
- (added) libc/src/__support/wchar/mbrtowc.cpp (+50) 
- (added) libc/src/__support/wchar/mbrtowc.h (+29) 
- (modified) libc/src/wchar/CMakeLists.txt (+17) 
- (added) libc/src/wchar/mbrtowc.cpp (+40) 
- (added) libc/src/wchar/mbrtowc.h (+24) 
- (modified) libc/test/src/wchar/CMakeLists.txt (+14) 
- (added) libc/test/src/wchar/mbrtowc_test.cpp (+170) 


``````````diff
diff --git a/libc/config/linux/x86_64/entrypoints.txt b/libc/config/linux/x86_64/entrypoints.txt
index aa2079faed409..10509a0c25835 100644
--- a/libc/config/linux/x86_64/entrypoints.txt
+++ b/libc/config/linux/x86_64/entrypoints.txt
@@ -365,6 +365,7 @@ set(TARGET_LIBC_ENTRYPOINTS
 
     # wchar.h entrypoints
     libc.src.wchar.btowc
+    libc.src.wchar.mbrtowc
     libc.src.wchar.wcslen
     libc.src.wchar.wctob
     libc.src.wchar.wmemmove
diff --git a/libc/hdr/types/CMakeLists.txt b/libc/hdr/types/CMakeLists.txt
index c88c357009072..e4b3cb0faa820 100644
--- a/libc/hdr/types/CMakeLists.txt
+++ b/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
diff --git a/libc/hdr/types/mbstate_t.h b/libc/hdr/types/mbstate_t.h
new file mode 100644
index 0000000000000..15b2614341d7d
--- /dev/null
+++ b/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
+
+#include "hdr/wchar_overlay.h"
+
+#endif // LLVM_LIBC_FULL_BUILD
+
+#endif // LLVM_LIBC_HDR_TYPES_MBSTATE_T_H
diff --git a/libc/include/llvm-libc-types/mbstate_t.h b/libc/include/llvm-libc-types/mbstate_t.h
index 540d50975a264..009fe57da50e2 100644
--- a/libc/include/llvm-libc-types/mbstate_t.h
+++ b/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
diff --git a/libc/include/wchar.yaml b/libc/include/wchar.yaml
index 84db73d8f01ea..06c621f59b462 100644
--- a/libc/include/wchar.yaml
+++ b/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
diff --git a/libc/src/__support/wchar/CMakeLists.txt b/libc/src/__support/wchar/CMakeLists.txt
index 6715e354e23e5..479c1dff2c6e0 100644
--- a/libc/src/__support/wchar/CMakeLists.txt
+++ b/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
+)
diff --git a/libc/src/__support/wchar/mbrtowc.cpp b/libc/src/__support/wchar/mbrtowc.cpp
new file mode 100644
index 0000000000000..969448ee60e81
--- /dev/null
+++ b/libc/src/__support/wchar/mbrtowc.cpp
@@ -0,0 +1,50 @@
+//===-- 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;
+  auto wc = char_conv.pop_utf32();
+  // Reading in bytes until we have a complete wc or error
+  for (; i < n && !wc.has_value(); ++i) {
+    int err = char_conv.push(static_cast<char8_t>(s[i]));
+    // Encoding error
+    if (err == -1)
+      return Error(-1);
+    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 Error(-2);
+}
+
+} // namespace internal
+
+} // namespace LIBC_NAMESPACE_DECL
diff --git a/libc/src/__support/wchar/mbrtowc.h b/libc/src/__support/wchar/mbrtowc.h
new file mode 100644
index 0000000000000..37329ee61beac
--- /dev/null
+++ b/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
diff --git a/libc/src/wchar/CMakeLists.txt b/libc/src/wchar/CMakeLists.txt
index 491dd5b34340a..163c29847e6a2 100644
--- a/libc/src/wchar/CMakeLists.txt
+++ b/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
diff --git a/libc/src/wchar/mbrtowc.cpp b/libc/src/wchar/mbrtowc.cpp
new file mode 100644
index 0000000000000..c29c5ee161e32
--- /dev/null
+++ b/libc/src/wchar/mbrtowc.cpp
@@ -0,0 +1,40 @@
+//===-- 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 mbstate_t internal_mbstate{0, 0, 0};
+  auto ret = internal::mbrtowc(
+      pwc, s, n, (internal::mbstate *)(ps == nullptr ? &internal_mbstate : ps));
+  if (!ret.has_value()) {
+    // Encoding failure
+    if (ret.error() == -1) {
+      libc_errno = EILSEQ;
+      return -1;
+    }
+    // Could potentially read a valid wide character.
+    return -2;
+  }
+  return ret.value();
+}
+
+} // namespace LIBC_NAMESPACE_DECL
diff --git a/libc/src/wchar/mbrtowc.h b/libc/src/wchar/mbrtowc.h
new file mode 100644
index 0000000000000..e2e3d3ebd2853
--- /dev/null
+++ b/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
diff --git a/libc/test/src/wchar/CMakeLists.txt b/libc/test/src/wchar/CMakeLists.txt
index 4990b6953348b..d4cae1f6228bd 100644
--- a/libc/test/src/wchar/CMakeLists.txt
+++ b/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
diff --git a/libc/test/src/wchar/mbrtowc_test.cpp b/libc/test/src/wchar/mbrtowc_test.cpp
new file mode 100644
index 0000000000000..6e96e7ac31f49
--- /dev/null
+++ b/libc/test/src/wchar/mbrtowc_test.cpp
@@ -0,0 +1,170 @@
+//===-- 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);
+}
+
+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];
+  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);
+  // 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);
+}

``````````

</details>


https://github.com/llvm/llvm-project/pull/144760