[libc] Implemented mblen functions (#150141)

Implemented mblen and mbrlen as well as tests

---------

Co-authored-by: Sriya Pratipati <[email protected]>

Author: sribee8

libc/config/linux/x86_64/entrypoints.txt

@@ -1261,6 +1261,8 @@ if(LLVM_LIBC_FULL_BUILD)
     libc.src.sys.socket.recvmsg
 
     # wchar.h entrypoints
+    libc.src.wchar.mblen
+    libc.src.wchar.mbrlen
     libc.src.wchar.mbrtowc
     libc.src.wchar.mbtowc
     libc.src.wchar.wcrtomb

libc/include/wchar.yaml

@@ -53,6 +53,21 @@ functions:
       - type: wchar_t *__restrict
       - type: const char *__restrict
       - type: size_t
+  - name: mblen
+    standards:
+      - stdc
+    return_type: int
+    arguments:
+      - type: const char *
+      - type: size_t
+  - name: mbrlen
+    standards:
+      - stdc
+    return_type: size_t
+    arguments:
+      - type: const char *__restrict
+      - type: size_t
+      - type: mbstate_t *__restrict
   - name: wmemset
     standards:
       - stdc

libc/src/wchar/CMakeLists.txt

@@ -169,6 +169,37 @@ add_entrypoint_object(
     libc.src.__support.wchar.mbstate
 )
 
+add_entrypoint_object(
+  mblen
+  SRCS
+    mblen.cpp
+  HDRS
+    mblen.h
+  DEPENDS
+    libc.hdr.types.size_t
+    libc.src.__support.common
+    libc.src.__support.macros.config
+    libc.src.__support.libc_errno
+    libc.src.__support.wchar.mbrtowc
+    libc.src.__support.wchar.mbstate
+)
+
+add_entrypoint_object(
+  mbrlen
+  SRCS
+    mbrlen.cpp
+  HDRS
+    mbrlen.h
+  DEPENDS
+    libc.hdr.types.size_t
+    libc.hdr.types.mbstate_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/mblen.cpp

@@ -0,0 +1,35 @@
+//===-- Implementation of mblen -------------------------------------------===//
+//
+// 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/mblen.h"
+
+#include "hdr/types/size_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(int, mblen, (const char *s, size_t n)) {
+  // returns 0 since UTF-8 encoding is not state-dependent
+  if (s == nullptr)
+    return 0;
+  internal::mbstate internal_mbstate;
+  auto ret = internal::mbrtowc(nullptr, s, n, &internal_mbstate);
+  if (!ret.has_value() || static_cast<int>(ret.value()) == -2) {
+    // Encoding failure
+    if (!ret.has_value())
+      libc_errno = EILSEQ;
+    return -1;
+  }
+  return static_cast<int>(ret.value());
+}
+
+} // namespace LIBC_NAMESPACE_DECL

libc/src/wchar/mblen.h

@@ -0,0 +1,21 @@
+//===-- Implementation header for mblen -----------------------------------===//
+//
+// 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_MBLEN_H
+#define LLVM_LIBC_SRC_WCHAR_MBLEN_H
+
+#include "hdr/types/size_t.h"
+#include "src/__support/macros/config.h"
+
+namespace LIBC_NAMESPACE_DECL {
+
+int mblen(const char *s, size_t n);
+
+} // namespace LIBC_NAMESPACE_DECL
+
+#endif // LLVM_LIBC_SRC_WCHAR_MBLEN_H

libc/src/wchar/mbrlen.cpp

@@ -0,0 +1,37 @@
+//===-- Implementation of mbrlen ------------------------------------------===//
+//
+// 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/mbrlen.h"
+
+#include "hdr/types/mbstate_t.h"
+#include "hdr/types/size_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, mbrlen,
+                   (const char *__restrict s, size_t n,
+                    mbstate_t *__restrict ps)) {
+  static internal::mbstate internal_mbstate;
+  auto ret = internal::mbrtowc(nullptr, s, n,
+                               ps == nullptr
+                                   ? &internal_mbstate
+                                   : reinterpret_cast<internal::mbstate *>(ps));
+  if (!ret.has_value()) {
+    // Encoding failure
+    libc_errno = ret.error();
+    return -1;
+  }
+  return ret.value();
+}
+
+} // namespace LIBC_NAMESPACE_DECL

libc/src/wchar/mbrlen.h

@@ -0,0 +1,22 @@
+//===-- Implementation header for mbrlen ----------------------------------===//
+//
+// 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_MBRLEN_H
+#define LLVM_LIBC_SRC_WCHAR_MBRLEN_H
+
+#include "hdr/types/mbstate_t.h"
+#include "hdr/types/size_t.h"
+#include "src/__support/macros/config.h"
+
+namespace LIBC_NAMESPACE_DECL {
+
+size_t mbrlen(const char *__restrict s, size_t n, mbstate_t *__restrict ps);
+
+} // namespace LIBC_NAMESPACE_DECL
+
+#endif // LLVM_LIBC_SRC_WCHAR_MBRLEN_H

libc/test/src/wchar/CMakeLists.txt

@@ -64,6 +64,33 @@ add_libc_test(
     libc.test.UnitTest.ErrnoCheckingTest
 )
 
+add_libc_test(
+  mblen_test
+  SUITE
+    libc_wchar_unittests
+  SRCS
+    mblen_test.cpp
+  DEPENDS
+    libc.src.__support.libc_errno
+    libc.src.wchar.mblen
+    libc.test.UnitTest.ErrnoCheckingTest
+)
+
+add_libc_test(
+  mbrlen_test
+  SUITE
+    libc_wchar_unittests
+  SRCS
+    mbrlen_test.cpp
+  DEPENDS
+    libc.src.__support.libc_errno
+    libc.src.__support.wchar.mbstate
+    libc.src.string.memset
+    libc.src.wchar.mbrlen
+    libc.hdr.types.mbstate_t
+    libc.test.UnitTest.ErrnoCheckingTest
+)
+
 add_libc_test(
   wctob_test
   SUITE

libc/test/src/wchar/mblen_test.cpp

@@ -0,0 +1,104 @@
+//===-- Unittests for mblen -----------------------------------------------===//
+//
+// 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/libc_errno.h"
+#include "src/wchar/mblen.h"
+#include "test/UnitTest/ErrnoCheckingTest.h"
+#include "test/UnitTest/Test.h"
+
+using LlvmLibcMBLenTest = LIBC_NAMESPACE::testing::ErrnoCheckingTest;
+
+TEST_F(LlvmLibcMBLenTest, OneByte) {
+  const char *ch = "A";
+  int n = LIBC_NAMESPACE::mblen(ch, 1);
+  ASSERT_ERRNO_SUCCESS();
+  ASSERT_EQ(n, 1);
+
+  // Should fail since we have not read enough
+  n = LIBC_NAMESPACE::mblen(ch, 0);
+  ASSERT_ERRNO_SUCCESS();
+  ASSERT_EQ(n, -1);
+}
+
+TEST_F(LlvmLibcMBLenTest, TwoByte) {
+  const char ch[2] = {static_cast<char>(0xC2),
+                      static_cast<char>(0x8E)}; // Ž car symbol
+  int n = LIBC_NAMESPACE::mblen(ch, 4);
+  ASSERT_ERRNO_SUCCESS();
+  ASSERT_EQ(n, 2);
+
+  // Should fail since we have not read enough
+  n = LIBC_NAMESPACE::mblen(ch, 1);
+  ASSERT_EQ(n, -1);
+  ASSERT_ERRNO_SUCCESS();
+  // Should fail after trying to read next byte too
+  n = LIBC_NAMESPACE::mblen(ch + 1, 1);
+  ASSERT_EQ(n, -1);
+  // This one should be an invalid starting byte so should set errno
+  ASSERT_ERRNO_EQ(EILSEQ);
+}
+
+TEST_F(LlvmLibcMBLenTest, ThreeByte) {
+  const char ch[3] = {static_cast<char>(0xE2), static_cast<char>(0x88),
+                      static_cast<char>(0x91)}; // ∑ sigma symbol
+  int n = LIBC_NAMESPACE::mblen(ch, 3);
+  ASSERT_EQ(n, 3);
+  ASSERT_ERRNO_SUCCESS();
+
+  // Should fail since we have not read enough
+  n = LIBC_NAMESPACE::mblen(ch, 2);
+  ASSERT_EQ(n, -1);
+  ASSERT_ERRNO_SUCCESS();
+}
+
+TEST_F(LlvmLibcMBLenTest, FourByte) {
+  const char ch[4] = {static_cast<char>(0xF0), static_cast<char>(0x9F),
+                      static_cast<char>(0xA4),
+                      static_cast<char>(0xA1)}; // 🤡 clown emoji
+  int n = LIBC_NAMESPACE::mblen(ch, 4);
+  ASSERT_EQ(n, 4);
+  ASSERT_ERRNO_SUCCESS();
+
+  // Should fail since we have not read enough
+  n = LIBC_NAMESPACE::mblen(ch, 2);
+  ASSERT_EQ(n, -1);
+  ASSERT_ERRNO_SUCCESS();
+}
+
+TEST_F(LlvmLibcMBLenTest, InvalidByte) {
+  const char ch[1] = {static_cast<char>(0x80)};
+  int n = LIBC_NAMESPACE::mblen(ch, 1);
+  ASSERT_EQ(n, -1);
+  ASSERT_ERRNO_EQ(EILSEQ);
+}
+
+TEST_F(LlvmLibcMBLenTest, 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
+  // Trying to push all 4 should error
+  int n = LIBC_NAMESPACE::mblen(ch, 4);
+  ASSERT_EQ(n, -1);
+  ASSERT_ERRNO_EQ(EILSEQ);
+
+  // Trying to push the second and third should correspond to null wc
+  n = LIBC_NAMESPACE::mblen(ch + 1, 2);
+  ASSERT_EQ(n, 0);
+  ASSERT_ERRNO_SUCCESS();
+}
+
+TEST_F(LlvmLibcMBLenTest, NullString) {
+  // reading on nullptr should return 0
+  int n = LIBC_NAMESPACE::mblen(nullptr, 2);
+  ASSERT_EQ(n, 0);
+  ASSERT_ERRNO_SUCCESS();
+  // reading a null terminator should return 0
+  const char *ch = "\0";
+  n = LIBC_NAMESPACE::mblen(ch, 1);
+  ASSERT_EQ(n, 0);
+}