platform_bionic/tests/wchar_test.cpp
Elliott Hughes 1f462dec34 Add %b and %B support to the scanf/wscanf and strto*/wcsto* families.
Coming to C23 via WG14 N2630.

This one is a little interesting, because it actually changes existing
behavior. Previously "0b101" would be parsed as "0", "b", "101" by these
functions. I'm led to believe that glibc plans to actually have separate
versions of these functions for C23 and pre-C23, so callers can have the
behavior they (implicitly) specify by virtue of which -std= they compile
with. Android has never really done anything like that, and I'm pretty
sure app developers have more than enough to worry about with API levels
without having to deal with the cartesian product of API level and C
standard.

Therefore, my plan A is "if you're running on Android >= U, you get C23
behavior". My plan B in the (I think unlikely) event that that actually
causes trouble for anyone is "if you're _targeting_ Android >= U, you
get C23 behavior". I don't think we'd actually want to have two versions
of each of these functions under any circumstances --- that seems by far
the most confusing option.

Test: treehugger
Change-Id: I0bbb30315d3fabd306905ad1484361f5d8745935
2022-08-11 00:25:08 +00:00

1224 lines
35 KiB
C++

/*
* Copyright (C) 2014 The Android Open Source Project
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include <gtest/gtest.h>
#include <errno.h>
#include <inttypes.h>
#include <limits.h>
#include <locale.h>
#include <math.h>
#include <stdint.h>
#include <sys/cdefs.h>
#include <wchar.h>
#include "utils.h"
#define NUM_WCHARS(num_bytes) ((num_bytes)/sizeof(wchar_t))
TEST(wchar, sizeof_wchar_t) {
EXPECT_EQ(4U, sizeof(wchar_t));
EXPECT_EQ(4U, sizeof(wint_t));
}
TEST(wchar, mbrlen) {
char bytes[] = { 'h', 'e', 'l', 'l', 'o', '\0' };
EXPECT_EQ(0U, mbrlen(&bytes[0], 0, nullptr));
EXPECT_EQ(1U, mbrlen(&bytes[0], 1, nullptr));
EXPECT_EQ(1U, mbrlen(&bytes[4], 1, nullptr));
EXPECT_EQ(0U, mbrlen(&bytes[5], 1, nullptr));
}
TEST(wchar, wctomb_wcrtomb) {
// wctomb and wcrtomb behave differently when s == NULL.
EXPECT_EQ(0, wctomb(nullptr, L'h'));
EXPECT_EQ(0, wctomb(nullptr, L'\0'));
EXPECT_EQ(1U, wcrtomb(nullptr, L'\0', nullptr));
EXPECT_EQ(1U, wcrtomb(nullptr, L'h', nullptr));
char bytes[MB_LEN_MAX];
// wctomb and wcrtomb behave similarly for the null wide character.
EXPECT_EQ(1, wctomb(bytes, L'\0'));
EXPECT_EQ(1U, wcrtomb(bytes, L'\0', nullptr));
// ...and for regular characters.
memset(bytes, 0, sizeof(bytes));
EXPECT_EQ(1, wctomb(bytes, L'h'));
EXPECT_EQ('h', bytes[0]);
memset(bytes, 0, sizeof(bytes));
EXPECT_EQ(1U, wcrtomb(bytes, L'h', nullptr));
EXPECT_EQ('h', bytes[0]);
ASSERT_STREQ("C.UTF-8", setlocale(LC_CTYPE, "C.UTF-8"));
uselocale(LC_GLOBAL_LOCALE);
// 1-byte UTF-8.
memset(bytes, 0, sizeof(bytes));
EXPECT_EQ(1U, wcrtomb(bytes, L'h', nullptr));
EXPECT_EQ('h', bytes[0]);
// 2-byte UTF-8.
memset(bytes, 0, sizeof(bytes));
EXPECT_EQ(2U, wcrtomb(bytes, 0x00a2, nullptr));
EXPECT_EQ('\xc2', bytes[0]);
EXPECT_EQ('\xa2', bytes[1]);
// 3-byte UTF-8.
memset(bytes, 0, sizeof(bytes));
EXPECT_EQ(3U, wcrtomb(bytes, 0x20ac, nullptr));
EXPECT_EQ('\xe2', bytes[0]);
EXPECT_EQ('\x82', bytes[1]);
EXPECT_EQ('\xac', bytes[2]);
// 4-byte UTF-8.
memset(bytes, 0, sizeof(bytes));
EXPECT_EQ(4U, wcrtomb(bytes, 0x24b62, nullptr));
EXPECT_EQ('\xf0', bytes[0]);
EXPECT_EQ('\xa4', bytes[1]);
EXPECT_EQ('\xad', bytes[2]);
EXPECT_EQ('\xa2', bytes[3]);
// Invalid code point.
EXPECT_EQ(static_cast<size_t>(-1), wcrtomb(bytes, 0xffffffff, nullptr));
EXPECT_EQ(EILSEQ, errno);
}
TEST(wchar, wcrtomb_start_state) {
char out[MB_LEN_MAX];
mbstate_t ps;
// Any non-initial state is invalid when calling wcrtomb.
memset(&ps, 0, sizeof(ps));
EXPECT_EQ(static_cast<size_t>(-2), mbrtowc(nullptr, "\xc2", 1, &ps));
EXPECT_EQ(static_cast<size_t>(-1), wcrtomb(out, 0x00a2, &ps));
EXPECT_EQ(EILSEQ, errno);
// If the first argument to wcrtomb is NULL or the second is L'\0' the shift
// state should be reset.
memset(&ps, 0, sizeof(ps));
EXPECT_EQ(static_cast<size_t>(-2), mbrtowc(nullptr, "\xc2", 1, &ps));
EXPECT_EQ(1U, wcrtomb(nullptr, 0x00a2, &ps));
EXPECT_TRUE(mbsinit(&ps));
memset(&ps, 0, sizeof(ps));
EXPECT_EQ(static_cast<size_t>(-2), mbrtowc(nullptr, "\xf0\xa4", 1, &ps));
EXPECT_EQ(1U, wcrtomb(out, L'\0', &ps));
EXPECT_TRUE(mbsinit(&ps));
}
TEST(wchar, wcstombs_wcrtombs) {
const wchar_t chars[] = { L'h', L'e', L'l', L'l', L'o', 0 };
const wchar_t bad_chars[] = { L'h', L'i', static_cast<wchar_t>(0xffffffff), 0 };
const wchar_t* src;
char bytes[BUFSIZ];
// Given a NULL destination, these functions count valid characters.
EXPECT_EQ(5U, wcstombs(nullptr, chars, 0));
EXPECT_EQ(5U, wcstombs(nullptr, chars, 4));
EXPECT_EQ(5U, wcstombs(nullptr, chars, 256));
src = chars;
EXPECT_EQ(5U, wcsrtombs(nullptr, &src, 0, nullptr));
EXPECT_EQ(&chars[0], src);
src = chars;
EXPECT_EQ(5U, wcsrtombs(nullptr, &src, 4, nullptr));
EXPECT_EQ(&chars[0], src);
src = chars;
EXPECT_EQ(5U, wcsrtombs(nullptr, &src, 256, nullptr));
EXPECT_EQ(&chars[0], src);
// An unrepresentable char just returns an error from wcstombs...
errno = 0;
EXPECT_EQ(static_cast<size_t>(-1), wcstombs(nullptr, bad_chars, 0));
EXPECT_EQ(EILSEQ, errno);
errno = 0;
EXPECT_EQ(static_cast<size_t>(-1), wcstombs(nullptr, bad_chars, 256));
EXPECT_EQ(EILSEQ, errno);
// And wcsrtombs doesn't tell us where it got stuck because we didn't ask it
// to actually convert anything...
errno = 0;
src = bad_chars;
EXPECT_EQ(static_cast<size_t>(-1), wcsrtombs(nullptr, &src, 0, nullptr));
EXPECT_EQ(&bad_chars[0], src);
EXPECT_EQ(EILSEQ, errno);
errno = 0;
src = bad_chars;
EXPECT_EQ(static_cast<size_t>(-1), wcsrtombs(nullptr, &src, 256, nullptr));
EXPECT_EQ(&bad_chars[0], src);
EXPECT_EQ(EILSEQ, errno);
// Okay, now let's test actually converting something...
memset(bytes, 'x', sizeof(bytes));
EXPECT_EQ(0U, wcstombs(bytes, chars, 0));
memset(bytes, 'x', sizeof(bytes));
EXPECT_EQ(4U, wcstombs(bytes, chars, 4));
bytes[5] = 0;
EXPECT_STREQ("hellx", bytes);
memset(bytes, 'x', sizeof(bytes));
EXPECT_EQ(5U, wcstombs(bytes, chars, 256));
EXPECT_STREQ("hello", bytes);
memset(bytes, 'x', sizeof(bytes));
EXPECT_EQ(5U, wcstombs(bytes, chars, 6));
EXPECT_STREQ("hello", bytes);
errno = 0;
memset(bytes, 'x', sizeof(bytes));
EXPECT_EQ(static_cast<size_t>(-1), wcstombs(bytes, bad_chars, 256));
EXPECT_EQ(EILSEQ, errno);
bytes[3] = 0;
EXPECT_STREQ("hix", bytes);
// wcsrtombs is a bit more informative...
memset(bytes, 'x', sizeof(bytes));
src = chars;
EXPECT_EQ(0U, wcsrtombs(bytes, &src, 0, nullptr));
EXPECT_EQ(&chars[0], src); // No input consumed.
EXPECT_EQ(EILSEQ, errno);
memset(bytes, 'x', sizeof(bytes));
src = chars;
EXPECT_EQ(4U, wcsrtombs(bytes, &src, 4, nullptr));
EXPECT_EQ(&chars[4], src); // Some input consumed.
EXPECT_EQ(EILSEQ, errno);
bytes[5] = 0;
EXPECT_STREQ("hellx", bytes);
memset(bytes, 'x', sizeof(bytes));
src = chars;
EXPECT_EQ(5U, wcsrtombs(bytes, &src, 256, nullptr));
EXPECT_EQ(nullptr, src); // All input consumed!
EXPECT_EQ(EILSEQ, errno);
EXPECT_STREQ("hello", bytes);
memset(bytes, 'x', sizeof(bytes));
src = chars;
EXPECT_EQ(5U, wcsrtombs(bytes, &src, 6, nullptr));
EXPECT_EQ(nullptr, src); // All input consumed.
EXPECT_EQ(EILSEQ, errno);
EXPECT_STREQ("hello", bytes);
memset(bytes, 'x', sizeof(bytes));
src = bad_chars;
EXPECT_EQ(static_cast<size_t>(-1), wcsrtombs(bytes, &src, 256, nullptr));
EXPECT_EQ(&bad_chars[2], src);
EXPECT_EQ(EILSEQ, errno);
bytes[3] = 0;
EXPECT_STREQ("hix", bytes);
// Any non-initial state is invalid when calling wcsrtombs.
mbstate_t ps;
src = chars;
memset(&ps, 0, sizeof(ps));
ASSERT_EQ(static_cast<size_t>(-2), mbrtowc(nullptr, "\xc2", 1, &ps));
EXPECT_EQ(static_cast<size_t>(-1), wcsrtombs(nullptr, &src, 0, &ps));
EXPECT_EQ(EILSEQ, errno);
}
TEST(wchar, limits) {
ASSERT_LT(WCHAR_MIN, WCHAR_MAX);
}
TEST(wchar, wcsstr) {
const wchar_t* haystack = L"big daddy/giant haystacks!";
const wchar_t* empty_haystack = L"";
// The empty needle is a special case.
ASSERT_EQ(haystack, wcsstr(haystack, L""));
ASSERT_EQ(empty_haystack, wcsstr(empty_haystack, L""));
ASSERT_EQ(haystack, wcsstr(haystack, L"b"));
ASSERT_EQ(haystack, wcsstr(haystack, L"big"));
ASSERT_EQ(haystack + 9, wcsstr(haystack, L"/"));
ASSERT_EQ(haystack + 9, wcsstr(haystack, L"/giant"));
ASSERT_EQ(haystack + 25, wcsstr(haystack, L"!"));
ASSERT_EQ(haystack + 19, wcsstr(haystack, L"stacks!"));
ASSERT_EQ(nullptr, wcsstr(haystack, L"monkey"));
ASSERT_EQ(nullptr, wcsstr(empty_haystack, L"monkey"));
}
TEST(wchar, wcsstr_80199) {
// https://code.google.com/p/android/issues/detail?id=80199
ASSERT_TRUE(wcsstr(L"romrom", L"rom") != nullptr);
}
TEST(wchar, mbtowc) {
wchar_t out[8];
out[0] = 'x';
ASSERT_EQ(0, mbtowc(out, "hello", 0));
ASSERT_EQ('x', out[0]);
ASSERT_EQ(0, mbtowc(out, "hello", 0));
ASSERT_EQ(0, mbtowc(out, "", 0));
ASSERT_EQ(1, mbtowc(out, "hello", 1));
ASSERT_EQ(L'h', out[0]);
ASSERT_EQ(0, mbtowc(nullptr, "hello", 0));
ASSERT_EQ(0, mbtowc(nullptr, "", 0));
ASSERT_EQ(1, mbtowc(nullptr, "hello", 1));
ASSERT_EQ(0, mbtowc(nullptr, nullptr, 0));
}
TEST(wchar, mbrtowc) {
wchar_t out[8];
out[0] = 'x';
ASSERT_EQ(0U, mbrtowc(out, "hello", 0, nullptr));
ASSERT_EQ('x', out[0]);
ASSERT_EQ(0U, mbrtowc(out, "hello", 0, nullptr));
ASSERT_EQ(0U, mbrtowc(out, "", 0, nullptr));
ASSERT_EQ(1U, mbrtowc(out, "hello", 1, nullptr));
ASSERT_EQ(L'h', out[0]);
ASSERT_EQ(0U, mbrtowc(nullptr, "hello", 0, nullptr));
ASSERT_EQ(0U, mbrtowc(nullptr, "", 0, nullptr));
ASSERT_EQ(1U, mbrtowc(nullptr, "hello", 1, nullptr));
ASSERT_EQ(0U, mbrtowc(nullptr, nullptr, 0, nullptr));
ASSERT_STREQ("C.UTF-8", setlocale(LC_CTYPE, "C.UTF-8"));
uselocale(LC_GLOBAL_LOCALE);
// 1-byte UTF-8.
ASSERT_EQ(1U, mbrtowc(out, "abcdef", 6, nullptr));
ASSERT_EQ(L'a', out[0]);
// 2-byte UTF-8.
ASSERT_EQ(2U, mbrtowc(out, "\xc2\xa2" "cdef", 6, nullptr));
ASSERT_EQ(static_cast<wchar_t>(0x00a2), out[0]);
// 3-byte UTF-8.
ASSERT_EQ(3U, mbrtowc(out, "\xe2\x82\xac" "def", 6, nullptr));
ASSERT_EQ(static_cast<wchar_t>(0x20ac), out[0]);
// 4-byte UTF-8.
ASSERT_EQ(4U, mbrtowc(out, "\xf0\xa4\xad\xa2" "ef", 6, nullptr));
ASSERT_EQ(static_cast<wchar_t>(0x24b62), out[0]);
#if defined(__BIONIC__) // glibc allows this.
// Illegal 5-byte UTF-8.
ASSERT_EQ(static_cast<size_t>(-1), mbrtowc(out, "\xf8\xa1\xa2\xa3\xa4" "f", 6, nullptr));
ASSERT_EQ(EILSEQ, errno);
#endif
// Illegal over-long sequence.
ASSERT_EQ(static_cast<size_t>(-1), mbrtowc(out, "\xf0\x82\x82\xac" "ef", 6, nullptr));
ASSERT_EQ(EILSEQ, errno);
}
TEST(wchar, mbrtowc_valid_non_characters) {
ASSERT_STREQ("C.UTF-8", setlocale(LC_CTYPE, "C.UTF-8"));
uselocale(LC_GLOBAL_LOCALE);
wchar_t out[8] = {};
ASSERT_EQ(3U, mbrtowc(out, "\xef\xbf\xbe", 3, nullptr));
ASSERT_EQ(static_cast<wchar_t>(0xfffe), out[0]);
ASSERT_EQ(3U, mbrtowc(out, "\xef\xbf\xbf", 3, nullptr));
ASSERT_EQ(static_cast<wchar_t>(0xffff), out[0]);
}
TEST(wchar, mbrtowc_out_of_range) {
ASSERT_STREQ("C.UTF-8", setlocale(LC_CTYPE, "C.UTF-8"));
uselocale(LC_GLOBAL_LOCALE);
wchar_t out[8] = {};
errno = 0;
ASSERT_EQ(static_cast<size_t>(-1), mbrtowc(out, "\xf5\x80\x80\x80", 4, nullptr));
ASSERT_EQ(EILSEQ, errno);
}
static void test_mbrtowc_incomplete(mbstate_t* ps) {
ASSERT_STREQ("C.UTF-8", setlocale(LC_CTYPE, "C.UTF-8"));
uselocale(LC_GLOBAL_LOCALE);
wchar_t out;
// 2-byte UTF-8.
ASSERT_EQ(static_cast<size_t>(-2), mbrtowc(&out, "\xc2", 1, ps));
ASSERT_EQ(1U, mbrtowc(&out, "\xa2" "cdef", 5, ps));
ASSERT_EQ(static_cast<wchar_t>(0x00a2), out);
ASSERT_TRUE(mbsinit(ps));
// 3-byte UTF-8.
ASSERT_EQ(static_cast<size_t>(-2), mbrtowc(&out, "\xe2", 1, ps));
ASSERT_EQ(static_cast<size_t>(-2), mbrtowc(&out, "\x82", 1, ps));
ASSERT_EQ(1U, mbrtowc(&out, "\xac" "def", 4, ps));
ASSERT_EQ(static_cast<wchar_t>(0x20ac), out);
ASSERT_TRUE(mbsinit(ps));
// 4-byte UTF-8.
ASSERT_EQ(static_cast<size_t>(-2), mbrtowc(&out, "\xf0", 1, ps));
ASSERT_EQ(static_cast<size_t>(-2), mbrtowc(&out, "\xa4\xad", 2, ps));
ASSERT_EQ(1U, mbrtowc(&out, "\xa2" "ef", 3, ps));
ASSERT_EQ(static_cast<wchar_t>(0x24b62), out);
ASSERT_TRUE(mbsinit(ps));
// Invalid 2-byte
ASSERT_EQ(static_cast<size_t>(-2), mbrtowc(&out, "\xc2", 1, ps));
ASSERT_EQ(static_cast<size_t>(-1), mbrtowc(&out, "\x20" "cdef", 5, ps));
ASSERT_EQ(EILSEQ, errno);
}
TEST(wchar, mbrtowc_incomplete) {
mbstate_t ps;
memset(&ps, 0, sizeof(ps));
test_mbrtowc_incomplete(&ps);
test_mbrtowc_incomplete(nullptr);
}
static void test_mbsrtowcs(mbstate_t* ps) {
constexpr const char* VALID = "A" "\xc2\xa2" "\xe2\x82\xac" "\xf0\xa4\xad\xa2" "ef";
constexpr const char* INVALID = "A" "\xc2\x20" "ef";
constexpr const char* INCOMPLETE = "A" "\xc2";
wchar_t out[4];
const char* valid = VALID;
ASSERT_EQ(4U, mbsrtowcs(out, &valid, 4, ps));
ASSERT_EQ(L'A', out[0]);
ASSERT_EQ(static_cast<wchar_t>(0x00a2), out[1]);
ASSERT_EQ(static_cast<wchar_t>(0x20ac), out[2]);
ASSERT_EQ(static_cast<wchar_t>(0x24b62), out[3]);
// Check that valid has advanced to the next unread character.
ASSERT_EQ('e', *valid);
wmemset(out, L'x', NUM_WCHARS(sizeof(out)));
ASSERT_EQ(2U, mbsrtowcs(out, &valid, 4, ps));
ASSERT_EQ(L'e', out[0]);
ASSERT_EQ(L'f', out[1]);
ASSERT_EQ(L'\0', out[2]);
// Check that we didn't clobber the rest of out.
ASSERT_EQ(L'x', out[3]);
// Check that valid has advanced to the end of the string.
ASSERT_EQ(nullptr, valid);
const char* invalid = INVALID;
ASSERT_EQ(static_cast<size_t>(-1), mbsrtowcs(out, &invalid, 4, ps));
EXPECT_EQ(EILSEQ, errno);
ASSERT_EQ('\xc2', *invalid);
const char* incomplete = INCOMPLETE;
ASSERT_EQ(static_cast<size_t>(-1), mbsrtowcs(out, &incomplete, 2, ps));
EXPECT_EQ(EILSEQ, errno);
ASSERT_EQ('\xc2', *incomplete);
// If dst is null, *src shouldn't be updated.
// https://code.google.com/p/android/issues/detail?id=166381
const char* mbs = VALID;
EXPECT_EQ(6U, mbsrtowcs(nullptr, &mbs, 0, ps));
EXPECT_EQ(VALID, mbs);
mbs = INVALID;
EXPECT_EQ(static_cast<size_t>(-1), mbsrtowcs(nullptr, &mbs, 0, ps));
EXPECT_EQ(INVALID, mbs);
mbs = INCOMPLETE;
EXPECT_EQ(static_cast<size_t>(-1), mbsrtowcs(nullptr, &mbs, 0, ps));
EXPECT_EQ(INCOMPLETE, mbs);
}
TEST(wchar, mbsrtowcs) {
ASSERT_STREQ("C.UTF-8", setlocale(LC_CTYPE, "C.UTF-8"));
uselocale(LC_GLOBAL_LOCALE);
mbstate_t ps;
memset(&ps, 0, sizeof(ps));
test_mbsrtowcs(&ps);
test_mbsrtowcs(nullptr);
// Invalid multi byte continuation.
const char* invalid = "\x20";
wchar_t out;
ASSERT_EQ(static_cast<size_t>(-2), mbrtowc(&out, "\xc2", 1, &ps));
ASSERT_EQ(static_cast<size_t>(-1), mbsrtowcs(&out, &invalid, 1, &ps));
EXPECT_EQ(EILSEQ, errno);
ASSERT_EQ('\x20', *invalid);
}
template <typename T>
using WcsToIntFn = T (*)(const wchar_t*, wchar_t**, int);
template <typename T>
void TestSingleWcsToInt(WcsToIntFn<T> fn, const wchar_t* str, int base,
T expected_value, ptrdiff_t expected_len) {
wchar_t* p;
ASSERT_EQ(expected_value, fn(str, &p, base));
ASSERT_EQ(expected_len, p - str) << str;
}
template <typename T>
void TestWcsToInt(WcsToIntFn<T> fn) {
TestSingleWcsToInt(fn, L"123", 10, static_cast<T>(123), 3);
TestSingleWcsToInt(fn, L"123", 0, static_cast<T>(123), 3);
TestSingleWcsToInt(fn, L"123#", 10, static_cast<T>(123), 3);
TestSingleWcsToInt(fn, L"01000", 8, static_cast<T>(512), 5);
TestSingleWcsToInt(fn, L"01000", 0, static_cast<T>(512), 5);
TestSingleWcsToInt(fn, L" 123 45", 0, static_cast<T>(123), 6);
TestSingleWcsToInt(fn, L" -123", 0, static_cast<T>(-123), 6);
TestSingleWcsToInt(fn, L"0x10000", 0, static_cast<T>(65536), 7);
TestSingleWcsToInt(fn, L"0b1011", 0, static_cast<T>(0b1011), 6);
}
template <typename T>
void TestWcsToIntLimits(WcsToIntFn<T> fn, const wchar_t* min_str,
const wchar_t* max_str) {
if (std::is_signed<T>::value) {
ASSERT_EQ(std::numeric_limits<T>::min(), fn(min_str, nullptr, 0)) << min_str;
} else {
// If the subject sequence begins with a <hyphen-minus>, the value resulting
// from the conversion shall be negated.
// http://pubs.opengroup.org/onlinepubs/9699919799/functions/strtoul.html
ASSERT_EQ(std::numeric_limits<T>::max(), fn(min_str, nullptr, 0)) << min_str;
}
ASSERT_EQ(std::numeric_limits<T>::max(), fn(max_str, nullptr, 0)) << max_str;
}
TEST(wchar, wcstol) {
TestWcsToInt(wcstol);
}
TEST(wchar, wcstol_limits) {
if (sizeof(long) == 8) {
TestWcsToIntLimits(wcstol, L"-9223372036854775809", L"9223372036854775808");
} else {
TestWcsToIntLimits(wcstol, L"-2147483649", L"2147483648");
}
}
TEST(wchar, wcstoul) {
TestWcsToInt(wcstoul);
}
TEST(wchar, wcstoul_limits) {
if (sizeof(long) == 8) {
TestWcsToIntLimits(wcstoul, L"-1", L"18446744073709551616");
} else {
TestWcsToIntLimits(wcstoul, L"-1", L"4294967296");
}
}
TEST(wchar, wcstoll) {
TestWcsToInt(wcstoll);
}
TEST(wchar, wcstoll_limits) {
TestWcsToIntLimits(wcstoll, L"-9223372036854775809", L"9223372036854775808");
}
TEST(wchar, wcstoull) {
TestWcsToInt(wcstoull);
}
TEST(wchar, wcstoull_limits) {
TestWcsToIntLimits(wcstoull, L"-1", L"18446744073709551616");
}
TEST(wchar, wcstoimax) {
TestWcsToInt(wcstoimax);
}
TEST(wchar, wcstoimax_limits) {
TestWcsToIntLimits(wcstoimax, L"-9223372036854775809",
L"9223372036854775808");
}
TEST(wchar, wcstoumax) {
TestWcsToInt(wcstoumax);
}
TEST(wchar, wcstoumax_limits) {
TestWcsToIntLimits(wcstoumax, L"-1", L"18446744073709551616");
}
TEST(wchar, mbsnrtowcs) {
wchar_t dst[128];
const char* s = "hello, world!";
const char* src;
memset(dst, 0, sizeof(dst));
src = s;
ASSERT_EQ(0U, mbsnrtowcs(dst, &src, 0, 0, nullptr));
memset(dst, 0, sizeof(dst));
src = s;
ASSERT_EQ(2U, mbsnrtowcs(dst, &src, 2, 123, nullptr)); // glibc chokes on SIZE_MAX here.
ASSERT_EQ(L'h', dst[0]);
ASSERT_EQ(L'e', dst[1]);
ASSERT_EQ(&s[2], src);
memset(dst, 0, sizeof(dst));
src = s;
ASSERT_EQ(3U, mbsnrtowcs(dst, &src, SIZE_MAX, 3, nullptr));
ASSERT_EQ(L'h', dst[0]);
ASSERT_EQ(L'e', dst[1]);
ASSERT_EQ(L'l', dst[2]);
ASSERT_EQ(&s[3], src);
memset(dst, 0, sizeof(dst));
const char* incomplete = "\xc2"; // Incomplete UTF-8 sequence.
src = incomplete;
errno = 0;
ASSERT_EQ(static_cast<size_t>(-1), mbsnrtowcs(dst, &src, SIZE_MAX, 3, nullptr));
ASSERT_EQ(EILSEQ, errno);
src = incomplete;
errno = 0;
ASSERT_EQ(static_cast<size_t>(-1), mbsnrtowcs(nullptr, &src, SIZE_MAX, 3, nullptr));
ASSERT_EQ(EILSEQ, errno);
}
TEST(wchar, wcsftime__wcsftime_l) {
setenv("TZ", "UTC", 1);
struct tm t;
memset(&t, 0, sizeof(tm));
t.tm_year = 200;
t.tm_mon = 2;
t.tm_mday = 10;
wchar_t buf[64];
EXPECT_EQ(24U, wcsftime(buf, sizeof(buf), L"%c", &t));
EXPECT_STREQ(L"Sun Mar 10 00:00:00 2100", buf);
EXPECT_EQ(24U, wcsftime_l(buf, sizeof(buf), L"%c", &t, LC_GLOBAL_LOCALE));
EXPECT_STREQ(L"Sun Mar 10 00:00:00 2100", buf);
}
TEST(wchar, wmemmove_smoke) {
const wchar_t const_wstr[] = L"This is a test of something or other.....";
wchar_t wstr[NUM_WCHARS(sizeof(const_wstr))];
EXPECT_EQ(wstr, wmemmove(wstr, const_wstr, NUM_WCHARS(sizeof(const_wstr))));
EXPECT_STREQ(const_wstr, wstr);
EXPECT_EQ(wstr+5, wmemmove(wstr+5, wstr, NUM_WCHARS(sizeof(const_wstr)) - 6));
EXPECT_STREQ(L"This This is a test of something or other", wstr);
}
TEST(wchar, wmemcpy_smoke) {
const wchar_t src[] = L"Source string";
wchar_t dst[NUM_WCHARS(sizeof(src))];
EXPECT_EQ(dst, wmemcpy(dst, src, NUM_WCHARS(sizeof(src))));
EXPECT_STREQ(dst, src);
}
TEST(wchar, wcpcpy_smoke) {
const wchar_t src[] = L"Source string";
wchar_t dst[NUM_WCHARS(sizeof(src))];
EXPECT_EQ(dst + NUM_WCHARS(sizeof(src)) - 1, wcpcpy(dst, src));
EXPECT_STREQ(dst, src);
}
TEST(wchar, wcpncpy_smoke) {
const wchar_t src[] = L"Source string";
wchar_t dst[NUM_WCHARS(sizeof(src)) + 5];
size_t src_len = NUM_WCHARS(sizeof(src)) - 1;
EXPECT_EQ(dst + src_len, wcpncpy(dst, src, src_len + 1));
EXPECT_STREQ(dst, src);
EXPECT_EQ(dst + 6, wcpncpy(dst, src, 6));
dst[6] = L'\0';
EXPECT_STREQ(dst, L"Source");
wmemset(dst, L'x', NUM_WCHARS(sizeof(dst)));
EXPECT_EQ(dst + src_len, wcpncpy(dst, src, src_len + 4));
EXPECT_STREQ(dst, src);
EXPECT_EQ(dst[src_len], L'\0');
EXPECT_EQ(dst[src_len+1], L'\0');
EXPECT_EQ(dst[src_len+2], L'\0');
EXPECT_EQ(dst[src_len+3], L'\0');
EXPECT_EQ(dst[src_len+4], L'x');
}
TEST(wchar, wcscpy_smoke) {
const wchar_t src[] = L"Source string";
wchar_t dst[NUM_WCHARS(sizeof(src))];
EXPECT_EQ(dst, wcscpy(dst, src));
EXPECT_STREQ(src, dst);
}
TEST(wchar, wcsncpy_smoke) {
const wchar_t src[] = L"Source string";
wchar_t dst[NUM_WCHARS(sizeof(src)) + 5];
size_t src_len = NUM_WCHARS(sizeof(src)) - 1;
EXPECT_EQ(dst, wcsncpy(dst, src, src_len + 1));
EXPECT_STREQ(dst, src);
EXPECT_EQ(dst, wcsncpy(dst, src, 6));
dst[6] = L'\0';
EXPECT_STREQ(dst, L"Source");
EXPECT_EQ(dst, wcsncpy(dst, L"clobber", 0));
EXPECT_STREQ(dst, L"Source");
wmemset(dst, L'x', NUM_WCHARS(sizeof(dst)));
EXPECT_EQ(dst, wcsncpy(dst, src, src_len + 4));
EXPECT_STREQ(dst, src);
EXPECT_EQ(dst[src_len], L'\0');
EXPECT_EQ(dst[src_len+1], L'\0');
EXPECT_EQ(dst[src_len+2], L'\0');
EXPECT_EQ(dst[src_len+3], L'\0');
EXPECT_EQ(dst[src_len+4], L'x');
}
TEST(wchar, mbrtowc_15439554) {
// http://b/15439554
ASSERT_STREQ("C.UTF-8", setlocale(LC_CTYPE, "C.UTF-8"));
uselocale(LC_GLOBAL_LOCALE);
ASSERT_GE(static_cast<size_t>(MB_LEN_MAX), MB_CUR_MAX);
ASSERT_GE(MB_CUR_MAX, 4U);
wchar_t wc;
size_t n;
// 1-byte character.
n = mbrtowc(&wc, "x", MB_CUR_MAX, nullptr);
EXPECT_EQ(1U, n);
EXPECT_EQ(L'x', wc);
// 2-byte character.
n = mbrtowc(&wc, "\xc2\xa2", MB_CUR_MAX, nullptr);
EXPECT_EQ(2U, n);
EXPECT_EQ(L'¢', wc);
// 3-byte character.
n = mbrtowc(&wc, "\xe2\x82\xac", MB_CUR_MAX, nullptr);
EXPECT_EQ(3U, n);
EXPECT_EQ(L'', wc);
// 4-byte character.
n = mbrtowc(&wc, "\xf0\xa4\xad\xa2", MB_CUR_MAX, nullptr);
EXPECT_EQ(4U, n);
EXPECT_EQ(L'𤭢', wc);
}
TEST(wchar, open_wmemstream) {
wchar_t* p = nullptr;
size_t size = 0;
FILE* fp = open_wmemstream(&p, &size);
ASSERT_NE(EOF, fputws(L"hello, world!", fp));
fclose(fp);
ASSERT_STREQ(L"hello, world!", p);
ASSERT_EQ(wcslen(L"hello, world!"), size);
free(p);
}
TEST(stdio, open_wmemstream_EINVAL) {
#if defined(__BIONIC__)
wchar_t* p;
size_t size;
// Invalid buffer.
errno = 0;
ASSERT_EQ(nullptr, open_wmemstream(nullptr, &size));
ASSERT_EQ(EINVAL, errno);
// Invalid size.
errno = 0;
ASSERT_EQ(nullptr, open_wmemstream(&p, nullptr));
ASSERT_EQ(EINVAL, errno);
#else
GTEST_SKIP() << "This test is bionic-specific";
#endif
}
TEST(wchar, wcstol_EINVAL) {
errno = 0;
wcstol(L"123", nullptr, -1);
ASSERT_EQ(EINVAL, errno);
errno = 0;
wcstol(L"123", nullptr, 1);
ASSERT_EQ(EINVAL, errno);
errno = 0;
wcstol(L"123", nullptr, 37);
ASSERT_EQ(EINVAL, errno);
}
TEST(wchar, wcstoll_EINVAL) {
errno = 0;
wcstoll(L"123", nullptr, -1);
ASSERT_EQ(EINVAL, errno);
errno = 0;
wcstoll(L"123", nullptr, 1);
ASSERT_EQ(EINVAL, errno);
errno = 0;
wcstoll(L"123", nullptr, 37);
ASSERT_EQ(EINVAL, errno);
}
TEST(wchar, wcstoul_EINVAL) {
errno = 0;
wcstoul(L"123", nullptr, -1);
ASSERT_EQ(EINVAL, errno);
errno = 0;
wcstoul(L"123", nullptr, 1);
ASSERT_EQ(EINVAL, errno);
errno = 0;
wcstoul(L"123", nullptr, 37);
ASSERT_EQ(EINVAL, errno);
}
TEST(wchar, wcstoull_EINVAL) {
errno = 0;
wcstoull(L"123", nullptr, -1);
ASSERT_EQ(EINVAL, errno);
errno = 0;
wcstoull(L"123", nullptr, 1);
ASSERT_EQ(EINVAL, errno);
errno = 0;
wcstoull(L"123", nullptr, 37);
ASSERT_EQ(EINVAL, errno);
}
TEST(wchar, wcstoll_l_EINVAL) {
errno = 0;
wcstoll_l(L"123", nullptr, -1, LC_GLOBAL_LOCALE);
ASSERT_EQ(EINVAL, errno);
errno = 0;
wcstoll_l(L"123", nullptr, 1, LC_GLOBAL_LOCALE);
ASSERT_EQ(EINVAL, errno);
errno = 0;
wcstoll_l(L"123", nullptr, 37, LC_GLOBAL_LOCALE);
ASSERT_EQ(EINVAL, errno);
}
TEST(wchar, wcstoull_l_EINVAL) {
errno = 0;
wcstoull_l(L"123", nullptr, -1, LC_GLOBAL_LOCALE);
ASSERT_EQ(EINVAL, errno);
errno = 0;
wcstoull_l(L"123", nullptr, 1, LC_GLOBAL_LOCALE);
ASSERT_EQ(EINVAL, errno);
errno = 0;
wcstoull_l(L"123", nullptr, 37, LC_GLOBAL_LOCALE);
ASSERT_EQ(EINVAL, errno);
}
TEST(wchar, wmempcpy) {
#if !defined(ANDROID_HOST_MUSL)
wchar_t dst[6];
ASSERT_EQ(&dst[4], wmempcpy(dst, L"hello", 4));
#else
GTEST_SKIP() << "musl doesn't have wmempcpy";
#endif
}
template <typename T>
using WcsToFloatFn = T (*)(const wchar_t*, wchar_t**);
template <typename T>
void TestSingleWcsToFloat(WcsToFloatFn<T> fn, const wchar_t* str,
T expected_value, ptrdiff_t expected_len) {
wchar_t* p;
ASSERT_EQ(expected_value, fn(str, &p));
ASSERT_EQ(expected_len, p - str);
}
template <typename T>
void TestWcsToFloat(WcsToFloatFn<T> fn) {
TestSingleWcsToFloat(fn, L"123", static_cast<T>(123.0L), 3);
TestSingleWcsToFloat(fn, L"123#", static_cast<T>(123.0L), 3);
TestSingleWcsToFloat(fn, L" 123 45", static_cast<T>(123.0L), 6);
TestSingleWcsToFloat(fn, L"9.0", static_cast<T>(9.0L), 3);
TestSingleWcsToFloat(fn, L"-9.0", static_cast<T>(-9.0L), 4);
TestSingleWcsToFloat(fn, L" \t\v\f\r\n9.0", static_cast<T>(9.0L), 9);
}
template <typename T>
void TestWcsToFloatHexFloats(WcsToFloatFn<T> fn) {
TestSingleWcsToFloat(fn, L"0.9e1", static_cast<T>(9.0L), 5);
TestSingleWcsToFloat(fn, L"0x1.2p3", static_cast<T>(9.0L), 7);
TestSingleWcsToFloat(fn, L"+1e+100", static_cast<T>(1e100L), 7);
TestSingleWcsToFloat(fn, L"0x10000.80", static_cast<T>(65536.50L), 10);
}
template <typename T>
void TestWcsToFloatInfNan(WcsToFloatFn<T> fn) {
ASSERT_TRUE(isnan(fn(L"+nan", nullptr)));
ASSERT_TRUE(isnan(fn(L"nan", nullptr)));
ASSERT_TRUE(isnan(fn(L"-nan", nullptr)));
ASSERT_TRUE(isnan(fn(L"+nan(0xff)", nullptr)));
ASSERT_TRUE(isnan(fn(L"nan(0xff)", nullptr)));
ASSERT_TRUE(isnan(fn(L"-nan(0xff)", nullptr)));
wchar_t* p;
ASSERT_TRUE(isnan(fn(L"+nanny", &p)));
ASSERT_STREQ(L"ny", p);
ASSERT_TRUE(isnan(fn(L"nanny", &p)));
ASSERT_STREQ(L"ny", p);
ASSERT_TRUE(isnan(fn(L"-nanny", &p)));
ASSERT_STREQ(L"ny", p);
ASSERT_EQ(0, fn(L"muppet", &p));
ASSERT_STREQ(L"muppet", p);
ASSERT_EQ(0, fn(L" muppet", &p));
ASSERT_STREQ(L" muppet", p);
ASSERT_EQ(std::numeric_limits<T>::infinity(), fn(L"+inf", nullptr));
ASSERT_EQ(std::numeric_limits<T>::infinity(), fn(L"inf", nullptr));
ASSERT_EQ(-std::numeric_limits<T>::infinity(), fn(L"-inf", nullptr));
ASSERT_EQ(std::numeric_limits<T>::infinity(), fn(L"+infinity", nullptr));
ASSERT_EQ(std::numeric_limits<T>::infinity(), fn(L"infinity", nullptr));
ASSERT_EQ(-std::numeric_limits<T>::infinity(), fn(L"-infinity", nullptr));
ASSERT_EQ(std::numeric_limits<T>::infinity(), fn(L"+infinitude", &p));
ASSERT_STREQ(L"initude", p);
ASSERT_EQ(std::numeric_limits<T>::infinity(), fn(L"infinitude", &p));
ASSERT_STREQ(L"initude", p);
ASSERT_EQ(-std::numeric_limits<T>::infinity(), fn(L"-infinitude", &p));
ASSERT_STREQ(L"initude", p);
// Check case-insensitivity.
ASSERT_EQ(std::numeric_limits<T>::infinity(), fn(L"InFiNiTy", nullptr));
ASSERT_TRUE(isnan(fn(L"NaN", nullptr)));
}
TEST(wchar, wcstof) {
TestWcsToFloat(wcstof);
}
TEST(wchar, wcstof_hex_floats) {
TestWcsToFloatHexFloats(wcstof);
}
TEST(wchar, wcstof_hex_inf_nan) {
TestWcsToFloatInfNan(wcstof);
}
TEST(wchar, wcstod) {
TestWcsToFloat(wcstod);
}
TEST(wchar, wcstod_hex_floats) {
TestWcsToFloatHexFloats(wcstod);
}
TEST(wchar, wcstod_hex_inf_nan) {
TestWcsToFloatInfNan(wcstod);
}
TEST(wchar, wcstold) {
TestWcsToFloat(wcstold);
}
TEST(wchar, wcstold_hex_floats) {
TestWcsToFloatHexFloats(wcstold);
}
TEST(wchar, wcstold_hex_inf_nan) {
TestWcsToFloatInfNan(wcstold);
}
TEST(wchar, wcstod_l) {
#if !defined(ANDROID_HOST_MUSL)
EXPECT_EQ(1.23, wcstod_l(L"1.23", nullptr, LC_GLOBAL_LOCALE));
#else
GTEST_SKIP() << "musl doesn't have wcstod_l";
#endif
}
TEST(wchar, wcstof_l) {
#if !defined(ANDROID_HOST_MUSL)
EXPECT_EQ(1.23f, wcstof_l(L"1.23", nullptr, LC_GLOBAL_LOCALE));
#else
GTEST_SKIP() << "musl doesn't have wcstof_l";
#endif
}
TEST(wchar, wcstol_l) {
#if !defined(ANDROID_HOST_MUSL)
EXPECT_EQ(123L, wcstol_l(L"123", nullptr, 10, LC_GLOBAL_LOCALE));
#else
GTEST_SKIP() << "musl doesn't have wcstol_l";
#endif
}
TEST(wchar, wcstold_l) {
EXPECT_EQ(1.23L, wcstold_l(L"1.23", nullptr, LC_GLOBAL_LOCALE));
}
TEST(wchar, wcstoll_l) {
EXPECT_EQ(123LL, wcstoll_l(L"123", nullptr, 10, LC_GLOBAL_LOCALE));
}
TEST(wchar, wcstoul_l) {
#if !defined(ANDROID_HOST_MUSL)
EXPECT_EQ(123UL, wcstoul_l(L"123", nullptr, 10, LC_GLOBAL_LOCALE));
#else
GTEST_SKIP() << "musl doesn't have wcstoul_l";
#endif
}
TEST(wchar, wcstoull_l) {
EXPECT_EQ(123ULL, wcstoull_l(L"123", nullptr, 10, LC_GLOBAL_LOCALE));
}
static void AssertWcwidthRange(wchar_t begin, wchar_t end, int expected) {
for (wchar_t i = begin; i < end; ++i) {
EXPECT_EQ(expected, wcwidth(i)) << static_cast<int>(i);
}
}
TEST(wchar, wcwidth_NUL) {
// NUL is defined to return 0 rather than -1, despite being a C0 control.
EXPECT_EQ(0, wcwidth(0));
}
TEST(wchar, wcwidth_ascii) {
AssertWcwidthRange(0x20, 0x7f, 1); // Non-C0 non-DEL ASCII.
}
TEST(wchar, wcwidth_controls) {
AssertWcwidthRange(0x01, 0x20, -1); // C0 controls.
EXPECT_EQ(-1, wcwidth(0x7f)); // DEL.
AssertWcwidthRange(0x80, 0xa0, -1); // C1 controls.
}
TEST(wchar, wcwidth_non_spacing_and_enclosing_marks_and_format) {
if (!have_dl()) return;
EXPECT_EQ(0, wcwidth(0x0300)); // Combining grave.
EXPECT_EQ(0, wcwidth(0x20dd)); // Combining enclosing circle.
EXPECT_EQ(0, wcwidth(0x00ad)); // Soft hyphen (SHY).
EXPECT_EQ(0, wcwidth(0x200b)); // Zero width space.
}
TEST(wchar, wcwidth_cjk) {
if (!have_dl()) return;
EXPECT_EQ(2, wcwidth(0x4e00)); // Start of CJK unified block.
EXPECT_EQ(2, wcwidth(0x9fff)); // End of CJK unified block.
EXPECT_EQ(2, wcwidth(0x3400)); // Start of CJK extension A block.
EXPECT_EQ(2, wcwidth(0x4dbf)); // End of CJK extension A block.
EXPECT_EQ(2, wcwidth(0x20000)); // Start of CJK extension B block.
EXPECT_EQ(2, wcwidth(0x2a6df)); // End of CJK extension B block.
}
TEST(wchar, wcwidth_korean_combining_jamo) {
if (!have_dl()) return;
AssertWcwidthRange(0x1160, 0x1200, 0); // Original range.
EXPECT_EQ(0, wcwidth(0xd7b0)); // Newer.
EXPECT_EQ(0, wcwidth(0xd7cb));
}
TEST(wchar, wcwidth_korean_jeongeul_syllables) {
if (!have_dl()) return;
EXPECT_EQ(2, wcwidth(0xac00)); // Start of block.
EXPECT_EQ(2, wcwidth(0xd7a3)); // End of defined code points in Unicode 7.
// Undefined characters at the end of the block have width 1.
}
TEST(wchar, wcwidth_kana) {
if (!have_dl()) return;
// Hiragana (most, not undefined).
AssertWcwidthRange(0x3041, 0x3097, 2);
// Katakana.
AssertWcwidthRange(0x30a0, 0x3100, 2);
}
TEST(wchar, wcwidth_circled_two_digit_cjk) {
if (!have_dl()) return;
// Circled two-digit CJK "speed sign" numbers are wide,
// though EastAsianWidth is ambiguous.
AssertWcwidthRange(0x3248, 0x3250, 2);
}
TEST(wchar, wcwidth_hexagrams) {
if (!have_dl()) return;
// Hexagrams are wide, though EastAsianWidth is neutral.
AssertWcwidthRange(0x4dc0, 0x4e00, 2);
}
TEST(wchar, wcwidth_default_ignorables) {
if (!have_dl()) return;
AssertWcwidthRange(0xfff0, 0xfff8, 0); // Unassigned by default ignorable.
EXPECT_EQ(0, wcwidth(0xe0000)); // ...through 0xe0fff.
}
TEST(wchar, wcwidth_korean_common_non_syllables) {
if (!have_dl()) return;
EXPECT_EQ(2, wcwidth(L'')); // Korean "crying" emoticon.
EXPECT_EQ(2, wcwidth(L'')); // Korean "laughing" emoticon.
}
TEST(wchar, wcswidth) {
EXPECT_EQ(2, wcswidth(L"abc", 2));
EXPECT_EQ(2, wcswidth(L"ab\t", 2));
EXPECT_EQ(-1, wcswidth(L"a\tb", 2));
}
TEST(wchar, wcslcpy) {
#if defined(__BIONIC__)
wchar_t dst[32];
ASSERT_EQ(11U, wcslcpy(dst, L"hello world", 3));
ASSERT_STREQ(L"he", dst);
ASSERT_EQ(11U, wcslcpy(dst, L"hello world", 32));
ASSERT_STREQ(L"hello world", dst);
#else
GTEST_SKIP() << "no wcslcpy in glibc";
#endif
}
TEST(wchar, wcscat) {
wchar_t dst[32];
ASSERT_EQ(dst, wcscat(dst, L"hello"));
ASSERT_STREQ(dst, L"hello");
ASSERT_EQ(dst, wcscat(dst, L" world"));
ASSERT_STREQ(dst, L"hello world");
}
TEST(wchar, wcscpy) {
wchar_t dst[32];
ASSERT_EQ(dst, wcscpy(dst, L"hello"));
ASSERT_STREQ(dst, L"hello");
ASSERT_EQ(dst, wcscpy(dst, L"world"));
ASSERT_STREQ(dst, L"world");
}
TEST(wchar, wcscasecmp) {
ASSERT_EQ(0, wcscasecmp(L"hello", L"HELLO"));
ASSERT_TRUE(wcscasecmp(L"hello1", L"HELLO2") < 0);
ASSERT_TRUE(wcscasecmp(L"hello2", L"HELLO1") > 0);
ASSERT_TRUE(wcscasecmp(L"hello", L"HELL") > 0);
ASSERT_TRUE(wcscasecmp(L"hell", L"HELLO") < 0);
}
TEST(wchar, wcscspn) {
ASSERT_EQ(0U, wcscspn(L"hello world", L"abcdefghijklmnopqrstuvwxyz"));
ASSERT_EQ(5U, wcscspn(L"hello world", L" "));
ASSERT_EQ(11U, wcscspn(L"hello world", L"!"));
}
TEST(wchar, wcsspn) {
ASSERT_EQ(0U, wcsspn(L"hello world", L"!"));
ASSERT_EQ(5U, wcsspn(L"hello world", L"abcdefghijklmnopqrstuvwxyz"));
ASSERT_EQ(11U, wcsspn(L"hello world", L"abcdefghijklmnopqrstuvwxyz "));
}
TEST(wchar, wcsdup) {
wchar_t* s = wcsdup(L"hello");
ASSERT_STREQ(s, L"hello");
free(s);
}
TEST(wchar, wcslcat) {
#if defined(__BIONIC__)
wchar_t dst[4] = {};
ASSERT_EQ(1U, wcslcat(dst, L"a", 4));
ASSERT_EQ(7U, wcslcat(dst, L"bcdefg", 4));
ASSERT_STREQ(dst, L"abc");
#else
GTEST_SKIP() << "no wcslcpy in glibc";
#endif
}
TEST(wchar, wcsncasecmp) {
ASSERT_EQ(0, wcsncasecmp(L"foo", L"bar", 0));
ASSERT_EQ(0, wcsncasecmp(L"hello1", L"HELLO2", 5));
ASSERT_TRUE(wcsncasecmp(L"hello1", L"HELLO2", 6) < 0);
ASSERT_TRUE(wcsncasecmp(L"hello2", L"HELLO1", 6) > 0);
ASSERT_TRUE(wcsncasecmp(L"hello", L"HELL", 5) > 0);
ASSERT_TRUE(wcsncasecmp(L"hell", L"HELLO", 5) < 0);
}
TEST(wchar, wcsncat) {
wchar_t dst[32];
ASSERT_EQ(dst, wcsncat(dst, L"hello, world!", 5));
ASSERT_STREQ(dst, L"hello");
ASSERT_EQ(dst, wcsncat(dst, L"hello, world!", 0));
ASSERT_STREQ(dst, L"hello");
ASSERT_EQ(dst, wcsncat(dst, L", world!", 8));
ASSERT_STREQ(dst, L"hello, world!");
}
TEST(wchar, wcsncmp) {
ASSERT_EQ(0, wcsncmp(L"foo", L"bar", 0));
ASSERT_EQ(0, wcsncmp(L"aaaa", L"aaab", 3));
ASSERT_TRUE(wcsncmp(L"aaaa", L"aaab", 4) < 0);
ASSERT_TRUE(wcsncmp(L"aaab", L"aaaa", 4) > 0);
}
TEST(wchar, wcsnlen) {
ASSERT_EQ(2U, wcsnlen(L"hello", 2));
ASSERT_EQ(5U, wcsnlen(L"hello", 5));
ASSERT_EQ(5U, wcsnlen(L"hello", 666));
}
TEST(wchar, wcspbrk) {
const wchar_t* s = L"hello, world!";
ASSERT_EQ(nullptr, wcspbrk(s, L"-"));
ASSERT_EQ(s, wcspbrk(s, L"abch"));
ASSERT_EQ(s + 2, wcspbrk(s, L"l"));
ASSERT_EQ(s + 5, wcspbrk(s, L",. !"));
}
TEST(wchar, wcstok) {
wchar_t s[] = L"this is\ta\nstring";
wchar_t* p;
ASSERT_EQ(s, wcstok(s, L"\t\n ", &p));
ASSERT_STREQ(s, L"this");
ASSERT_STREQ(p, L"is\ta\nstring");
ASSERT_EQ(s + 5, wcstok(nullptr, L"\t\n ", &p));
ASSERT_STREQ(s + 5, L"is");
ASSERT_STREQ(p, L"a\nstring");
ASSERT_EQ(s + 8, wcstok(nullptr, L"\t\n ", &p));
ASSERT_STREQ(s + 8, L"a");
ASSERT_STREQ(p, L"string");
ASSERT_EQ(s + 10, wcstok(nullptr, L"\t\n ", &p));
ASSERT_STREQ(s + 10, L"string");
ASSERT_EQ(nullptr, p);
}
TEST(wchar, wmemchr) {
const wchar_t* s = L"hello, world!";
ASSERT_EQ(s, wmemchr(s, L'h', 13));
ASSERT_EQ(s + 5, wmemchr(s, L',', 13));
ASSERT_EQ(s + 12, wmemchr(s, L'!', 13));
ASSERT_EQ(nullptr, wmemchr(s, L'a', 13));
}
TEST(wchar, wmemcmp) {
ASSERT_EQ(0, wmemcmp(L"aaaa", L"aaab", 3));
ASSERT_TRUE(wmemcmp(L"aaaa", L"aaab", 4) < 0);
ASSERT_TRUE(wmemcmp(L"aaab", L"aaaa", 4) > 0);
}
TEST(wchar, wmemcpy) {
wchar_t dst[32] = {};
ASSERT_EQ(dst, wmemcpy(dst, L"hello", 5));
ASSERT_STREQ(dst, L"hello");
}
TEST(wchar, wmemmove) {
wchar_t dst[32] = {};
ASSERT_EQ(dst, wmemmove(dst, L"hello", 5));
ASSERT_STREQ(dst, L"hello");
}
TEST(wchar, wmemset) {
wchar_t dst[4] = {};
ASSERT_EQ(dst, wmemset(dst, 0x12345678, 3));
ASSERT_EQ(dst[0], wchar_t(0x12345678));
ASSERT_EQ(dst[1], wchar_t(0x12345678));
ASSERT_EQ(dst[2], wchar_t(0x12345678));
ASSERT_EQ(dst[3], wchar_t(0));
ASSERT_EQ(dst, wmemset(dst, L'y', 0));
ASSERT_EQ(dst[0], wchar_t(0x12345678));
}