[libc] add strncmp to strings

[michaelrj]

Summary

Add strncmp as a function to strings.h. Also adds unit tests, and adds
strncmp as an entrypoint for all current platforms.

[sivachandra]

libc/src/string/strncmp.cpp

19

Nitty comment: Why can it not be just n == 0?

22

Another nitty comment - This for statement is hard to read. Can we make it simpler like this:

for (; n > 0; --n, ++left, ++right) {
  char lc = *left;
  if (lc == '\0' || lc != *right)
    break;
}

24

Just pointing out as I am not sure what is right here: casting to unsigned char can remove the potential sign and lead to -1 > 0. Even strcmp is set up this way so may be clean up both of them in a follow up patch.

libc/test/src/string/strncmp_test.cpp

159

Fix.

[cheng.w]

libc/src/string/strncmp.cpp

19

nit: We could omit the braces since the body of if is simple.

[michaelrj]

address comments

libc/src/string/strncmp.cpp

24

given that these are supposed to be strings, I would assume that unsigned char is correct. Testing with my system libc, it appears that this is what it does as well. I ran strncmp on the following char array: [-1, 'a', '\0'] compared to "abc" with length 3 and got the value 158, which is 255 - 97 ('a' = 97).

[sivachandra]

libc/src/string/strncmp.cpp

24

Thanks for checking!

libc/config/linux/aarch64/entrypoints.txt

set(TARGET_LIBC_ENTRYPOINTS
     libc.src.string.memset
     libc.src.string.memrchr
     libc.src.string.strcat
     libc.src.string.strchr
     libc.src.string.strcpy
     libc.src.string.strcmp
     libc.src.string.strcspn
     libc.src.string.strlen
+    libc.src.string.strncmp
     libc.src.string.strncpy
     libc.src.string.strnlen
     libc.src.string.strpbrk
     libc.src.string.strrchr
     libc.src.string.strspn
     libc.src.string.strstr
     libc.src.string.strtok
     libc.src.string.strtok_r

libc/config/linux/x86_64/entrypoints.txt

set(TARGET_LIBC_ENTRYPOINTS
     libc.src.string.memrchr
     libc.src.string.memset
     libc.src.string.strcat
     libc.src.string.strchr
     libc.src.string.strcmp
     libc.src.string.strcpy
     libc.src.string.strcspn
     libc.src.string.strlen
+    libc.src.string.strncmp
     libc.src.string.strncpy
     libc.src.string.strnlen
     libc.src.string.strpbrk
     libc.src.string.strrchr
     libc.src.string.strspn
     libc.src.string.strstr
     libc.src.string.strtok
     libc.src.string.strtok_r

libc/config/windows/entrypoints.txt

set(TARGET_LIBC_ENTRYPOINTS
     libc.src.string.memset
     libc.src.string.memrchr
     libc.src.string.strcat
     libc.src.string.strchr
     libc.src.string.strcpy
     libc.src.string.strcmp
     libc.src.string.strcspn
     libc.src.string.strlen
+    libc.src.string.strncmp
     libc.src.string.strncpy
     libc.src.string.strnlen
     libc.src.string.strpbrk
     libc.src.string.strrchr
     libc.src.string.strspn
     libc.src.string.strstr
     libc.src.string.strtok
     libc.src.string.strtok_r

libc/src/string/CMakeLists.txt

add_entrypoint_object(
   strcmp
   SRCS
     strcmp.cpp
   HDRS
     strcmp.h
 )
 
 add_entrypoint_object(
+  strncmp
+  SRCS
+    strncmp.cpp
+  HDRS
+    strncmp.h
+)
+
+add_entrypoint_object(
   memchr
   SRCS
     memchr.cpp
   HDRS
     memchr.h
   DEPENDS
     .string_utils
 )

libc/src/string/strncmp.h

+//===-- Implementation header for strncmp -----------------------*- 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_STRING_STRNCMP_H
+#define LLVM_LIBC_SRC_STRING_STRNCMP_H
+
+#include <stddef.h>
+
+namespace __llvm_libc {
+
+int strncmp(const char *left, const char *right, size_t n);
+
+} // namespace __llvm_libc
+
+#endif // LLVM_LIBC_SRC_STRING_STRNCMP_H

libc/src/string/strncmp.cpp

+//===-- Implementation of strncmp -----------------------------------------===//
+//
+// 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/string/strncmp.h"
+
+#include "src/__support/common.h"
+#include <stddef.h>
+
+namespace __llvm_libc {
+
+// TODO: Look at benefits for comparing words at a time.
+LLVM_LIBC_FUNCTION(int, strncmp,
+                   (const char *left, const char *right, size_t n)) {
+  if (n < 1) {
+    return 0;
+  }
+  for (size_t i = 1; *left && *left == *right && i < n; ++left, ++right, ++i)
+    ;
+  return *reinterpret_cast<const unsigned char *>(left) -
+         *reinterpret_cast<const unsigned char *>(right);
+}
+
+} // namespace __llvm_libc

libc/test/src/string/CMakeLists.txt

SUITE
     libc_string_unittests
   SRCS
     strcmp_test.cpp
   DEPENDS
     libc.src.string.strcmp
 )
 
 add_libc_unittest(
+  strncmp_test
+  SUITE
+    libc_string_unittests
+  SRCS
+    strncmp_test.cpp
+  DEPENDS
+    libc.src.string.strncmp
+)
+
+add_libc_unittest(
   memchr_test
   SUITE
     libc_string_unittests
   SRCS
     memchr_test.cpp
   DEPENDS
     libc.src.string.memchr
 )

libc/test/src/string/strncmp_test.cpp

+//===-- Unittests for strncmp ---------------------------------------------===//
+//
+// 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/string/strncmp.h"
+#include "utils/UnitTest/Test.h"
+
+// This group is just copies of the strcmp tests, since all the same cases still
+// need to be tested.
+
+TEST(LlvmLibcStrNCmpTest, EmptyStringsShouldReturnZeroWithSufficientLength) {
+  const char *s1 = "";
+  const char *s2 = "";
+  int result = __llvm_libc::strncmp(s1, s2, 1);
+  ASSERT_EQ(result, 0);
+
+  // Verify operands reversed.
+  result = __llvm_libc::strncmp(s2, s1, 1);
+  ASSERT_EQ(result, 0);
+}
+
+TEST(LlvmLibcStrNCmpTest,
+     EmptyStringShouldNotEqualNonEmptyStringWithSufficientLength) {
+  const char *empty = "";
+  const char *s2 = "abc";
+  int result = __llvm_libc::strncmp(empty, s2, 3);
+  // This should be '\0' - 'a' = -97
+  ASSERT_EQ(result, -97);
+
+  // Similar case if empty string is second argument.
+  const char *s3 = "123";
+  result = __llvm_libc::strncmp(s3, empty, 3);
+  // This should be '1' - '\0' = 49
+  ASSERT_EQ(result, 49);
+}
+
+TEST(LlvmLibcStrNCmpTest, EqualStringsShouldReturnZeroWithSufficientLength) {
+  const char *s1 = "abc";
+  const char *s2 = "abc";
+  int result = __llvm_libc::strncmp(s1, s2, 3);
+  ASSERT_EQ(result, 0);
+
+  // Verify operands reversed.
+  result = __llvm_libc::strncmp(s2, s1, 3);
+  ASSERT_EQ(result, 0);
+}
+
+TEST(LlvmLibcStrNCmpTest,
+     ShouldReturnResultOfFirstDifferenceWithSufficientLength) {
+  const char *s1 = "___B42__";
+  const char *s2 = "___C55__";
+  int result = __llvm_libc::strncmp(s1, s2, 8);
+  // This should return 'B' - 'C' = -1.
+  ASSERT_EQ(result, -1);
+
+  // Verify operands reversed.
+  result = __llvm_libc::strncmp(s2, s1, 8);
+  // This should return 'C' - 'B' = 1.
+  ASSERT_EQ(result, 1);
+}
+
+TEST(LlvmLibcStrNCmpTest,
+     CapitalizedLetterShouldNotBeEqualWithSufficientLength) {
+  const char *s1 = "abcd";
+  const char *s2 = "abCd";
+  int result = __llvm_libc::strncmp(s1, s2, 4);
+  // 'c' - 'C' = 32.
+  ASSERT_EQ(result, 32);
+
+  // Verify operands reversed.
+  result = __llvm_libc::strncmp(s2, s1, 4);
+  // 'C' - 'c' = -32.
+  ASSERT_EQ(result, -32);
+}
+
+TEST(LlvmLibcStrNCmpTest,
+     UnequalLengthStringsShouldNotReturnZeroWithSufficientLength) {
+  const char *s1 = "abc";
+  const char *s2 = "abcd";
+  int result = __llvm_libc::strncmp(s1, s2, 4);
+  // '\0' - 'd' = -100.
+  ASSERT_EQ(result, -100);
+
+  // Verify operands reversed.
+  result = __llvm_libc::strncmp(s2, s1, 4);
+  // 'd' - '\0' = 100.
+  ASSERT_EQ(result, 100);
+}
+
+TEST(LlvmLibcStrNCmpTest, StringArgumentSwapChangesSignWithSufficientLength) {
+  const char *a = "a";
+  const char *b = "b";
+  int result = __llvm_libc::strncmp(b, a, 1);
+  // 'b' - 'a' = 1.
+  ASSERT_EQ(result, 1);
+
+  result = __llvm_libc::strncmp(a, b, 1);
+  // 'a' - 'b' = -1.
+  ASSERT_EQ(result, -1);
+}
+
+// This group is actually testing strncmp functionality
+
+TEST(LlvmLibcStrNCmpTest, NonEqualStringsEqualWithLengthZero) {
+  const char *s1 = "abc";
+  const char *s2 = "def";
+  int result = __llvm_libc::strncmp(s1, s2, 0);
+  ASSERT_EQ(result, 0);
+
+  // Verify operands reversed.
+  result = __llvm_libc::strncmp(s2, s1, 0);
+  ASSERT_EQ(result, 0);
+}
+
+TEST(LlvmLibcStrNCmpTest, NonEqualStringsNotEqualWithLengthOne) {
+  const char *s1 = "abc";
+  const char *s2 = "def";
+  int result = __llvm_libc::strncmp(s1, s2, 1);
+  ASSERT_EQ(result, -3);
+
+  // Verify operands reversed.
+  result = __llvm_libc::strncmp(s2, s1, 1);
+  ASSERT_EQ(result, 3);
+}
+
+TEST(LlvmLibcStrNCmpTest, NonEqualStringsEqualWithShorterLength) {
+  const char *s1 = "___B42__";
+  const char *s2 = "___C55__";
+  int result = __llvm_libc::strncmp(s1, s2, 3);
+  ASSERT_EQ(result, 0);
+
+  // This should return 'B' - 'C' = -1.
+  result = __llvm_libc::strncmp(s1, s2, 4);
+  ASSERT_EQ(result, -1);
+
+  // Verify operands reversed.
+  result = __llvm_libc::strncmp(s2, s1, 3);
+  ASSERT_EQ(result, 0);
+
+  // This should return 'C' - 'B' = 1.
+  result = __llvm_libc::strncmp(s2, s1, 4);
+  ASSERT_EQ(result, 1);
+}
+
+TEST(LlvmLibcStrNCmpTest, StringComparisonEndsOnNullByteEvenWithLongerLength) {
+  const char *s1 = "abc\0def";
+  const char *s2 = "abc\0abc";
+  int result = __llvm_libc::strncmp(s1, s2, 7);
+  ASSERT_EQ(result, 0);
+
+  // Verify operands reversed.
+  result = __llvm_libc::strncmp(s2, s1, 7);
+  ASSERT_EQ(result, 0);
+}
 No newline at end of file