213 lines · cpp
1//===-- Unittests for mbsetowcs -------------------------------------------===//2//3// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.4// See https://llvm.org/LICENSE.txt for license information.5// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception6//7//===----------------------------------------------------------------------===//8 9#include "hdr/errno_macros.h"10#include "hdr/types/mbstate_t.h"11#include "hdr/types/wchar_t.h"12#include "src/__support/macros/null_check.h"13#include "src/__support/wchar/mbstate.h"14#include "src/string/memset.h"15#include "src/wchar/mbsnrtowcs.h"16#include "test/UnitTest/ErrnoCheckingTest.h"17#include "test/UnitTest/Test.h"18 19using LlvmLibcMBSNRToWCSTest = LIBC_NAMESPACE::testing::ErrnoCheckingTest;20 21TEST_F(LlvmLibcMBSNRToWCSTest, OneByteOneChar) {22 const char *ch = "A";23 const char *original = ch;24 wchar_t dest[2];25 mbstate_t mb;26 LIBC_NAMESPACE::memset(&mb, 0, sizeof(mbstate_t));27 size_t n = LIBC_NAMESPACE::mbsnrtowcs(dest, &ch, 1, 1, &mb);28 ASSERT_EQ(static_cast<char>(*dest), 'A');29 ASSERT_EQ(static_cast<int>(n), 1);30 // Should point to null terminator now31 ASSERT_EQ(ch, original + 1);32 ASSERT_ERRNO_SUCCESS();33 34 n = LIBC_NAMESPACE::mbsnrtowcs(dest + 1, &ch, 1, 1, &mb);35 ASSERT_EQ(static_cast<char>(dest[1]), '\0');36 // Should not include null terminator37 ASSERT_EQ(static_cast<int>(n), 0);38 // Should now be a nullptr39 ASSERT_EQ(ch, nullptr);40 ASSERT_ERRNO_SUCCESS();41}42 43TEST_F(LlvmLibcMBSNRToWCSTest, FourByteOneChar) {44 const char *src = "\xf0\x9f\x98\xb9"; // laughing cat emoji 😹45 const char *original = src;46 wchar_t dest[2];47 mbstate_t mb;48 LIBC_NAMESPACE::memset(&mb, 0, sizeof(mbstate_t));49 // Not enough bytes for the full character50 size_t n = LIBC_NAMESPACE::mbsnrtowcs(dest, &src, 3, 2, &mb);51 ASSERT_ERRNO_SUCCESS();52 ASSERT_EQ(static_cast<int>(n), 0);53 ASSERT_EQ(src, original + 3);54 // Needs 2 more bytes (last byte of cat + null terminator)55 n = LIBC_NAMESPACE::mbsnrtowcs(dest, &src, 2, 2, &mb);56 ASSERT_ERRNO_SUCCESS();57 // Does not include null terminator58 ASSERT_EQ(static_cast<int>(n), 1);59 ASSERT_EQ(src, nullptr);60 ASSERT_EQ(static_cast<int>(dest[0]), 128569);61 ASSERT_TRUE(dest[1] == L'\0');62}63 64TEST_F(LlvmLibcMBSNRToWCSTest, MixedNumberOfBytes) {65 // 'A', sigma symbol 'Σ', recycling symbol '♻', laughing cat emoji '😹'66 const char *src = "A\xce\xa3\xe2\x99\xbb\xf0\x9f\x98\xb9";67 const char *original = src;68 wchar_t dest[5];69 mbstate_t mb;70 LIBC_NAMESPACE::memset(&mb, 0, sizeof(mbstate_t));71 72 // Read 'A'73 size_t n = LIBC_NAMESPACE::mbsnrtowcs(dest, &src, 1, 1, &mb);74 ASSERT_ERRNO_SUCCESS();75 ASSERT_EQ(static_cast<char>(dest[0]), 'A');76 ASSERT_EQ(static_cast<int>(n), 1);77 ASSERT_EQ(src, original + 1);78 79 // Read sigma 'Σ'80 n = LIBC_NAMESPACE::mbsnrtowcs(dest + 1, &src, 2, 1, &mb);81 ASSERT_ERRNO_SUCCESS();82 ASSERT_EQ(static_cast<int>(dest[1]), 931);83 ASSERT_EQ(static_cast<int>(n), 1);84 ASSERT_EQ(src, original + 3);85 86 // Read recycling '♻'87 n = LIBC_NAMESPACE::mbsnrtowcs(dest + 2, &src, 2, 5, &mb);88 ASSERT_ERRNO_SUCCESS();89 ASSERT_EQ(static_cast<int>(n), 0);90 ASSERT_EQ(src, original + 5);91 n = LIBC_NAMESPACE::mbsnrtowcs(dest + 2, &src, 1, 1, &mb);92 ASSERT_ERRNO_SUCCESS();93 ASSERT_EQ(static_cast<int>(n), 1);94 ASSERT_EQ(src, original + 6);95 ASSERT_EQ(static_cast<int>(dest[2]), 9851);96 97 // Read laughing cat emoji '😹'98 n = LIBC_NAMESPACE::mbsnrtowcs(dest + 3, &src, 4, 5, &mb);99 ASSERT_ERRNO_SUCCESS();100 ASSERT_EQ(static_cast<int>(n), 1);101 ASSERT_EQ(src, original + 10);102 ASSERT_EQ(static_cast<int>(dest[3]), 128569);103 104 n = LIBC_NAMESPACE::mbsnrtowcs(dest + 4, &src, 4, 4, nullptr);105 ASSERT_TRUE(dest[4] == L'\0');106 ASSERT_ERRNO_SUCCESS();107 // Should not count null terminator in number108 ASSERT_EQ(static_cast<int>(n), 0);109 // Should now be a nullptr110 ASSERT_EQ(src, nullptr);111}112 113TEST_F(LlvmLibcMBSNRToWCSTest, ReadLessThanStringLength) {114 // Four laughing cat emojis "😹😹😹😹"115 const char *src =116 "\xf0\x9f\x98\xb9\xf0\x9f\x98\xb9\xf0\x9f\x98\xb9\xf0\x9f\x98\xb9";117 const char *original = src;118 wchar_t dest[5] = {L'a', L'b', L'c', L'd', L'e'};119 size_t n = LIBC_NAMESPACE::mbsnrtowcs(dest, &src, 100, 3, nullptr);120 ASSERT_ERRNO_SUCCESS();121 // Should have read 3 emojis122 ASSERT_EQ(static_cast<int>(n), 3);123 ASSERT_EQ(static_cast<int>(dest[0]), 128569);124 ASSERT_EQ(static_cast<int>(dest[1]), 128569);125 ASSERT_EQ(static_cast<int>(dest[2]), 128569);126 ASSERT_TRUE(dest[3] == L'd');127 ASSERT_TRUE(dest[4] == L'e');128 // Read three laughing cat emojis, 12 bytes129 ASSERT_EQ(src, original + 12);130}131 132TEST_F(LlvmLibcMBSNRToWCSTest, InvalidFirstByte) {133 // 0x80 is invalid first byte of mb character134 const char *src =135 "\x80\x9f\x98\xb9\xf0\x9f\x98\xb9\xf0\x9f\x98\xb9\xf0\x9f\x98\xb9";136 const char *original = src;137 wchar_t dest[3];138 size_t n = LIBC_NAMESPACE::mbsnrtowcs(dest, &src, 88, 88, nullptr);139 // Should return error and set errno140 ASSERT_EQ(static_cast<int>(n), -1);141 ASSERT_ERRNO_EQ(EILSEQ);142 // Should not update pointer143 ASSERT_EQ(src, original);144}145 146TEST_F(LlvmLibcMBSNRToWCSTest, InvalidMiddleByte) {147 // The 7th byte is invalid for a 4 byte character148 const char *src =149 "\xf0\x9f\x98\xb9\xf0\x9f\xf0\xb9\xf0\x9f\x98\xb9\xf0\x9f\x98\xb9";150 const char *original = src;151 wchar_t dest[3];152 mbstate_t mb;153 LIBC_NAMESPACE::memset(&mb, 0, sizeof(mbstate_t));154 // Successfully read one character and first byte of the second character155 size_t n = LIBC_NAMESPACE::mbsnrtowcs(dest, &src, 5, 88, &mb);156 ASSERT_EQ(static_cast<int>(n), 1);157 ASSERT_ERRNO_SUCCESS();158 ASSERT_EQ(src, original + 5);159 ASSERT_EQ(static_cast<int>(dest[0]), 128569);160 161 n = LIBC_NAMESPACE::mbsnrtowcs(dest + 1, &src, 5, 88, &mb);162 // Should return error, set errno, and not update the pointer163 ASSERT_EQ(static_cast<int>(n), -1);164 ASSERT_ERRNO_EQ(EILSEQ);165 ASSERT_EQ(src, original + 5);166}167 168TEST_F(LlvmLibcMBSNRToWCSTest, NullDestination) {169 // Four laughing cat emojis "😹😹😹😹"170 const char *src =171 "\xf0\x9f\x98\xb9\xf0\x9f\x98\xb9\xf0\x9f\x98\xb9\xf0\x9f\x98\xb9";172 const char *original = src;173 size_t n = LIBC_NAMESPACE::mbsnrtowcs(nullptr, &src, 88, 88, nullptr);174 ASSERT_ERRNO_SUCCESS();175 // Null destination should ignore len and read till end of string176 ASSERT_EQ(static_cast<int>(n), 4);177 // It should also not change the src pointer178 ASSERT_EQ(src, original);179}180 181TEST_F(LlvmLibcMBSNRToWCSTest, ErrnoChecks) {182 // Two laughing cat emojis and invalid 3rd mb char (3rd byte of it)183 const char *src =184 "\xf0\x9f\x98\xb9\xf0\x9f\x98\xb9\xf0\x9f\xf0\xb9\xf0\x9f\x98\xb9";185 const char *original = src;186 wchar_t dest[5];187 mbstate_t mb;188 LIBC_NAMESPACE::memset(&mb, 0, sizeof(mbstate_t));189 // First two bytes are valid --> should not set errno190 size_t n = LIBC_NAMESPACE::mbsnrtowcs(dest, &src, 80, 2, &mb);191 ASSERT_ERRNO_SUCCESS();192 ASSERT_EQ(static_cast<int>(n), 2);193 ASSERT_EQ(static_cast<int>(dest[0]), 128569);194 ASSERT_EQ(static_cast<int>(dest[1]), 128569);195 ASSERT_EQ(src, original + 8);196 197 // Trying to read the 3rd byte should set errno198 n = LIBC_NAMESPACE::mbsnrtowcs(dest + 2, &src, 4, 2, &mb);199 ASSERT_ERRNO_EQ(EILSEQ);200 ASSERT_EQ(static_cast<int>(n), -1);201 // Should not move the pointer202 ASSERT_EQ(src, original + 8);203}204 205#if defined(LIBC_ADD_NULL_CHECKS)206TEST(LlvmLibcMBSNRToWCSTest, NullptrCrash) {207 // Passing in a nullptr should crash the program.208 EXPECT_DEATH(209 [] { LIBC_NAMESPACE::mbsnrtowcs(nullptr, nullptr, 1, 1, nullptr); },210 WITH_SIGNAL(-1));211}212#endif // LIBC_ADD_NULL_CHECKS213