Source file src/strings/strings_test.go

     1  // Copyright 2009 The Go Authors. All rights reserved.
     2  // Use of this source code is governed by a BSD-style
     3  // license that can be found in the LICENSE file.
     4  
     5  package strings_test
     6  
     7  import (
     8  	"bytes"
     9  	"fmt"
    10  	"io"
    11  	"iter"
    12  	"math"
    13  	"math/rand"
    14  	"slices"
    15  	"strconv"
    16  	. "strings"
    17  	"testing"
    18  	"unicode"
    19  	"unicode/utf8"
    20  	"unsafe"
    21  )
    22  
    23  func collect(t *testing.T, seq iter.Seq[string]) []string {
    24  	out := slices.Collect(seq)
    25  	out1 := slices.Collect(seq)
    26  	if !slices.Equal(out, out1) {
    27  		t.Fatalf("inconsistent seq:\n%s\n%s", out, out1)
    28  	}
    29  	return out
    30  }
    31  
    32  type LinesTest struct {
    33  	a string
    34  	b []string
    35  }
    36  
    37  var linesTests = []LinesTest{
    38  	{a: "abc\nabc\n", b: []string{"abc\n", "abc\n"}},
    39  	{a: "abc\r\nabc", b: []string{"abc\r\n", "abc"}},
    40  	{a: "abc\r\n", b: []string{"abc\r\n"}},
    41  	{a: "\nabc", b: []string{"\n", "abc"}},
    42  	{a: "\nabc\n\n", b: []string{"\n", "abc\n", "\n"}},
    43  }
    44  
    45  func TestLines(t *testing.T) {
    46  	for _, s := range linesTests {
    47  		result := slices.Collect(Lines(s.a))
    48  		if !slices.Equal(result, s.b) {
    49  			t.Errorf(`slices.Collect(Lines(%q)) = %q; want %q`, s.a, result, s.b)
    50  		}
    51  	}
    52  }
    53  
    54  var abcd = "abcd"
    55  var faces = "☺☻☹"
    56  var commas = "1,2,3,4"
    57  var dots = "1....2....3....4"
    58  
    59  type IndexTest struct {
    60  	s   string
    61  	sep string
    62  	out int
    63  }
    64  
    65  var indexTests = []IndexTest{
    66  	{"", "", 0},
    67  	{"", "a", -1},
    68  	{"", "foo", -1},
    69  	{"fo", "foo", -1},
    70  	{"foo", "foo", 0},
    71  	{"oofofoofooo", "f", 2},
    72  	{"oofofoofooo", "foo", 4},
    73  	{"barfoobarfoo", "foo", 3},
    74  	{"foo", "", 0},
    75  	{"foo", "o", 1},
    76  	{"abcABCabc", "A", 3},
    77  	{"jrzm6jjhorimglljrea4w3rlgosts0w2gia17hno2td4qd1jz", "jz", 47},
    78  	{"ekkuk5oft4eq0ocpacknhwouic1uua46unx12l37nioq9wbpnocqks6", "ks6", 52},
    79  	{"999f2xmimunbuyew5vrkla9cpwhmxan8o98ec", "98ec", 33},
    80  	{"9lpt9r98i04k8bz6c6dsrthb96bhi", "96bhi", 24},
    81  	{"55u558eqfaod2r2gu42xxsu631xf0zobs5840vl", "5840vl", 33},
    82  	// cases with one byte strings - test special case in Index()
    83  	{"", "a", -1},
    84  	{"x", "a", -1},
    85  	{"x", "x", 0},
    86  	{"abc", "a", 0},
    87  	{"abc", "b", 1},
    88  	{"abc", "c", 2},
    89  	{"abc", "x", -1},
    90  	// test special cases in Index() for short strings
    91  	{"", "ab", -1},
    92  	{"bc", "ab", -1},
    93  	{"ab", "ab", 0},
    94  	{"xab", "ab", 1},
    95  	{"xab"[:2], "ab", -1},
    96  	{"", "abc", -1},
    97  	{"xbc", "abc", -1},
    98  	{"abc", "abc", 0},
    99  	{"xabc", "abc", 1},
   100  	{"xabc"[:3], "abc", -1},
   101  	{"xabxc", "abc", -1},
   102  	{"", "abcd", -1},
   103  	{"xbcd", "abcd", -1},
   104  	{"abcd", "abcd", 0},
   105  	{"xabcd", "abcd", 1},
   106  	{"xyabcd"[:5], "abcd", -1},
   107  	{"xbcqq", "abcqq", -1},
   108  	{"abcqq", "abcqq", 0},
   109  	{"xabcqq", "abcqq", 1},
   110  	{"xyabcqq"[:6], "abcqq", -1},
   111  	{"xabxcqq", "abcqq", -1},
   112  	{"xabcqxq", "abcqq", -1},
   113  	{"", "01234567", -1},
   114  	{"32145678", "01234567", -1},
   115  	{"01234567", "01234567", 0},
   116  	{"x01234567", "01234567", 1},
   117  	{"x0123456x01234567", "01234567", 9},
   118  	{"xx01234567"[:9], "01234567", -1},
   119  	{"", "0123456789", -1},
   120  	{"3214567844", "0123456789", -1},
   121  	{"0123456789", "0123456789", 0},
   122  	{"x0123456789", "0123456789", 1},
   123  	{"x012345678x0123456789", "0123456789", 11},
   124  	{"xyz0123456789"[:12], "0123456789", -1},
   125  	{"x01234567x89", "0123456789", -1},
   126  	{"", "0123456789012345", -1},
   127  	{"3214567889012345", "0123456789012345", -1},
   128  	{"0123456789012345", "0123456789012345", 0},
   129  	{"x0123456789012345", "0123456789012345", 1},
   130  	{"x012345678901234x0123456789012345", "0123456789012345", 17},
   131  	{"", "01234567890123456789", -1},
   132  	{"32145678890123456789", "01234567890123456789", -1},
   133  	{"01234567890123456789", "01234567890123456789", 0},
   134  	{"x01234567890123456789", "01234567890123456789", 1},
   135  	{"x0123456789012345678x01234567890123456789", "01234567890123456789", 21},
   136  	{"xyz01234567890123456789"[:22], "01234567890123456789", -1},
   137  	{"", "0123456789012345678901234567890", -1},
   138  	{"321456788901234567890123456789012345678911", "0123456789012345678901234567890", -1},
   139  	{"0123456789012345678901234567890", "0123456789012345678901234567890", 0},
   140  	{"x0123456789012345678901234567890", "0123456789012345678901234567890", 1},
   141  	{"x012345678901234567890123456789x0123456789012345678901234567890", "0123456789012345678901234567890", 32},
   142  	{"xyz0123456789012345678901234567890"[:33], "0123456789012345678901234567890", -1},
   143  	{"", "01234567890123456789012345678901", -1},
   144  	{"32145678890123456789012345678901234567890211", "01234567890123456789012345678901", -1},
   145  	{"01234567890123456789012345678901", "01234567890123456789012345678901", 0},
   146  	{"x01234567890123456789012345678901", "01234567890123456789012345678901", 1},
   147  	{"x0123456789012345678901234567890x01234567890123456789012345678901", "01234567890123456789012345678901", 33},
   148  	{"xyz01234567890123456789012345678901"[:34], "01234567890123456789012345678901", -1},
   149  	{"xxxxxx012345678901234567890123456789012345678901234567890123456789012", "012345678901234567890123456789012345678901234567890123456789012", 6},
   150  	{"", "0123456789012345678901234567890123456789", -1},
   151  	{"xx012345678901234567890123456789012345678901234567890123456789012", "0123456789012345678901234567890123456789", 2},
   152  	{"xx012345678901234567890123456789012345678901234567890123456789012"[:41], "0123456789012345678901234567890123456789", -1},
   153  	{"xx012345678901234567890123456789012345678901234567890123456789012", "0123456789012345678901234567890123456xxx", -1},
   154  	{"xx0123456789012345678901234567890123456789012345678901234567890120123456789012345678901234567890123456xxx", "0123456789012345678901234567890123456xxx", 65},
   155  	// test fallback to Rabin-Karp.
   156  	{"oxoxoxoxoxoxoxoxoxoxoxoy", "oy", 22},
   157  	{"oxoxoxoxoxoxoxoxoxoxoxox", "oy", -1},
   158  	// test fallback to IndexRune
   159  	{"oxoxoxoxoxoxoxoxoxoxox☺", "☺", 22},
   160  	// invalid UTF-8 byte sequence (must be longer than bytealg.MaxBruteForce to
   161  	// test that we don't use IndexRune)
   162  	{"xx0123456789012345678901234567890123456789012345678901234567890120123456789012345678901234567890123456xxx\xed\x9f\xc0", "\xed\x9f\xc0", 105},
   163  }
   164  
   165  var lastIndexTests = []IndexTest{
   166  	{"", "", 0},
   167  	{"", "a", -1},
   168  	{"", "foo", -1},
   169  	{"fo", "foo", -1},
   170  	{"foo", "foo", 0},
   171  	{"foo", "f", 0},
   172  	{"oofofoofooo", "f", 7},
   173  	{"oofofoofooo", "foo", 7},
   174  	{"barfoobarfoo", "foo", 9},
   175  	{"foo", "", 3},
   176  	{"foo", "o", 2},
   177  	{"abcABCabc", "A", 3},
   178  	{"abcABCabc", "a", 6},
   179  }
   180  
   181  var indexAnyTests = []IndexTest{
   182  	{"", "", -1},
   183  	{"", "a", -1},
   184  	{"", "abc", -1},
   185  	{"a", "", -1},
   186  	{"a", "a", 0},
   187  	{"\x80", "\xffb", 0},
   188  	{"aaa", "a", 0},
   189  	{"abc", "xyz", -1},
   190  	{"abc", "xcz", 2},
   191  	{"ab☺c", "x☺yz", 2},
   192  	{"a☺b☻c☹d", "cx", len("a☺b☻")},
   193  	{"a☺b☻c☹d", "uvw☻xyz", len("a☺b")},
   194  	{"aRegExp*", ".(|)*+?^$[]", 7},
   195  	{dots + dots + dots, " ", -1},
   196  	{"012abcba210", "\xffb", 4},
   197  	{"012\x80bcb\x80210", "\xffb", 3},
   198  	{"0123456\xcf\x80abc", "\xcfb\x80", 10},
   199  }
   200  
   201  var lastIndexAnyTests = []IndexTest{
   202  	{"", "", -1},
   203  	{"", "a", -1},
   204  	{"", "abc", -1},
   205  	{"a", "", -1},
   206  	{"a", "a", 0},
   207  	{"\x80", "\xffb", 0},
   208  	{"aaa", "a", 2},
   209  	{"abc", "xyz", -1},
   210  	{"abc", "ab", 1},
   211  	{"ab☺c", "x☺yz", 2},
   212  	{"a☺b☻c☹d", "cx", len("a☺b☻")},
   213  	{"a☺b☻c☹d", "uvw☻xyz", len("a☺b")},
   214  	{"a.RegExp*", ".(|)*+?^$[]", 8},
   215  	{dots + dots + dots, " ", -1},
   216  	{"012abcba210", "\xffb", 6},
   217  	{"012\x80bcb\x80210", "\xffb", 7},
   218  	{"0123456\xcf\x80abc", "\xcfb\x80", 10},
   219  }
   220  
   221  // Execute f on each test case.  funcName should be the name of f; it's used
   222  // in failure reports.
   223  func runIndexTests(t *testing.T, f func(s, sep string) int, funcName string, testCases []IndexTest) {
   224  	for _, test := range testCases {
   225  		actual := f(test.s, test.sep)
   226  		if actual != test.out {
   227  			t.Errorf("%s(%q,%q) = %v; want %v", funcName, test.s, test.sep, actual, test.out)
   228  		}
   229  	}
   230  }
   231  
   232  func TestIndex(t *testing.T)     { runIndexTests(t, Index, "Index", indexTests) }
   233  func TestLastIndex(t *testing.T) { runIndexTests(t, LastIndex, "LastIndex", lastIndexTests) }
   234  func TestIndexAny(t *testing.T)  { runIndexTests(t, IndexAny, "IndexAny", indexAnyTests) }
   235  func TestLastIndexAny(t *testing.T) {
   236  	runIndexTests(t, LastIndexAny, "LastIndexAny", lastIndexAnyTests)
   237  }
   238  
   239  func TestIndexByte(t *testing.T) {
   240  	for _, tt := range indexTests {
   241  		if len(tt.sep) != 1 {
   242  			continue
   243  		}
   244  		pos := IndexByte(tt.s, tt.sep[0])
   245  		if pos != tt.out {
   246  			t.Errorf(`IndexByte(%q, %q) = %v; want %v`, tt.s, tt.sep[0], pos, tt.out)
   247  		}
   248  	}
   249  }
   250  
   251  func TestLastIndexByte(t *testing.T) {
   252  	testCases := []IndexTest{
   253  		{"", "q", -1},
   254  		{"abcdef", "q", -1},
   255  		{"abcdefabcdef", "a", len("abcdef")},      // something in the middle
   256  		{"abcdefabcdef", "f", len("abcdefabcde")}, // last byte
   257  		{"zabcdefabcdef", "z", 0},                 // first byte
   258  		{"a☺b☻c☹d", "b", len("a☺")},               // non-ascii
   259  	}
   260  	for _, test := range testCases {
   261  		actual := LastIndexByte(test.s, test.sep[0])
   262  		if actual != test.out {
   263  			t.Errorf("LastIndexByte(%q,%c) = %v; want %v", test.s, test.sep[0], actual, test.out)
   264  		}
   265  	}
   266  }
   267  
   268  func simpleIndex(s, sep string) int {
   269  	n := len(sep)
   270  	for i := n; i <= len(s); i++ {
   271  		if s[i-n:i] == sep {
   272  			return i - n
   273  		}
   274  	}
   275  	return -1
   276  }
   277  
   278  func TestIndexRandom(t *testing.T) {
   279  	const chars = "abcdefghijklmnopqrstuvwxyz0123456789"
   280  	for times := 0; times < 10; times++ {
   281  		for strLen := 5 + rand.Intn(5); strLen < 140; strLen += 10 { // Arbitrary
   282  			s1 := make([]byte, strLen)
   283  			for i := range s1 {
   284  				s1[i] = chars[rand.Intn(len(chars))]
   285  			}
   286  			s := string(s1)
   287  			for i := 0; i < 50; i++ {
   288  				begin := rand.Intn(len(s) + 1)
   289  				end := begin + rand.Intn(len(s)+1-begin)
   290  				sep := s[begin:end]
   291  				if i%4 == 0 {
   292  					pos := rand.Intn(len(sep) + 1)
   293  					sep = sep[:pos] + "A" + sep[pos:]
   294  				}
   295  				want := simpleIndex(s, sep)
   296  				res := Index(s, sep)
   297  				if res != want {
   298  					t.Errorf("Index(%s,%s) = %d; want %d", s, sep, res, want)
   299  				}
   300  			}
   301  		}
   302  	}
   303  }
   304  
   305  func TestIndexRune(t *testing.T) {
   306  	tests := []struct {
   307  		in   string
   308  		rune rune
   309  		want int
   310  	}{
   311  		{"", 'a', -1},
   312  		{"", '☺', -1},
   313  		{"foo", '☹', -1},
   314  		{"foo", 'o', 1},
   315  		{"foo☺bar", '☺', 3},
   316  		{"foo☺☻☹bar", '☹', 9},
   317  		{"a A x", 'A', 2},
   318  		{"some_text=some_value", '=', 9},
   319  		{"☺a", 'a', 3},
   320  		{"a☻☺b", '☺', 4},
   321  
   322  		// RuneError should match any invalid UTF-8 byte sequence.
   323  		{"�", '�', 0},
   324  		{"\xff", '�', 0},
   325  		{"☻x�", '�', len("☻x")},
   326  		{"☻x\xe2\x98", '�', len("☻x")},
   327  		{"☻x\xe2\x98�", '�', len("☻x")},
   328  		{"☻x\xe2\x98x", '�', len("☻x")},
   329  
   330  		// Invalid rune values should never match.
   331  		{"a☺b☻c☹d\xe2\x98�\xff�\xed\xa0\x80", -1, -1},
   332  		{"a☺b☻c☹d\xe2\x98�\xff�\xed\xa0\x80", 0xD800, -1}, // Surrogate pair
   333  		{"a☺b☻c☹d\xe2\x98�\xff�\xed\xa0\x80", utf8.MaxRune + 1, -1},
   334  
   335  		// 2 bytes
   336  		{"ӆ", 'ӆ', 0},
   337  		{"a", 'ӆ', -1},
   338  		{"  ӆ", 'ӆ', 2},
   339  		{"  a", 'ӆ', -1},
   340  		{Repeat("ц", 64) + "ӆ", 'ӆ', 128}, // test cutover
   341  		{Repeat("Ꙁ", 64) + "Ꚁ", '䚀', -1},  // 'Ꚁ' and '䚀' share the same last two bytes
   342  
   343  		// 3 bytes
   344  		{"Ꚁ", 'Ꚁ', 0},
   345  		{"a", 'Ꚁ', -1},
   346  		{"  Ꚁ", 'Ꚁ', 2},
   347  		{"  a", 'Ꚁ', -1},
   348  		{Repeat("Ꙁ", 64) + "Ꚁ", 'Ꚁ', 192}, // test cutover
   349  		{Repeat("𡋀", 64) + "𡌀", '𣌀', -1},  // '𡌀' and '𣌀' share the same last two bytes
   350  
   351  		// 4 bytes
   352  		{"𡌀", '𡌀', 0},
   353  		{"a", '𡌀', -1},
   354  		{"  𡌀", '𡌀', 2},
   355  		{"  a", '𡌀', -1},
   356  		{Repeat("𡋀", 64) + "𡌀", '𡌀', 256}, // test cutover
   357  		{Repeat("𡋀", 64), '𡌀', -1},
   358  
   359  		// Test the cutover to bytealg.IndexString when it is triggered in
   360  		// the middle of rune that contains consecutive runs of equal bytes.
   361  		{"aaaaaKKKK\U000bc104", '\U000bc104', 17}, // cutover: (n + 16) / 8
   362  		{"aaaaaKKKK鄄", '鄄', 17},
   363  		{"aaKKKKKa\U000bc104", '\U000bc104', 18}, // cutover: 4 + n>>4
   364  		{"aaKKKKKa鄄", '鄄', 18},
   365  	}
   366  	for _, tt := range tests {
   367  		if got := IndexRune(tt.in, tt.rune); got != tt.want {
   368  			t.Errorf("IndexRune(%q, %d) = %v; want %v", tt.in, tt.rune, got, tt.want)
   369  		}
   370  	}
   371  
   372  	// Make sure we trigger the cutover and string(rune) conversion.
   373  	haystack := "test" + Repeat("𡋀", 32) + "𡌀"
   374  	allocs := testing.AllocsPerRun(1000, func() {
   375  		if i := IndexRune(haystack, 's'); i != 2 {
   376  			t.Fatalf("'s' at %d; want 2", i)
   377  		}
   378  		if i := IndexRune(haystack, '𡌀'); i != 132 {
   379  			t.Fatalf("'𡌀' at %d; want 4", i)
   380  		}
   381  	})
   382  	if allocs != 0 && testing.CoverMode() == "" {
   383  		t.Errorf("expected no allocations, got %f", allocs)
   384  	}
   385  }
   386  
   387  const benchmarkString = "some_text=some☺value"
   388  
   389  func BenchmarkIndexRune(b *testing.B) {
   390  	if got := IndexRune(benchmarkString, '☺'); got != 14 {
   391  		b.Fatalf("wrong index: expected 14, got=%d", got)
   392  	}
   393  	for i := 0; i < b.N; i++ {
   394  		IndexRune(benchmarkString, '☺')
   395  	}
   396  }
   397  
   398  var benchmarkLongString = Repeat(" ", 100) + benchmarkString
   399  
   400  func BenchmarkIndexRuneLongString(b *testing.B) {
   401  	if got := IndexRune(benchmarkLongString, '☺'); got != 114 {
   402  		b.Fatalf("wrong index: expected 114, got=%d", got)
   403  	}
   404  	for i := 0; i < b.N; i++ {
   405  		IndexRune(benchmarkLongString, '☺')
   406  	}
   407  }
   408  
   409  func BenchmarkIndexRuneFastPath(b *testing.B) {
   410  	if got := IndexRune(benchmarkString, 'v'); got != 17 {
   411  		b.Fatalf("wrong index: expected 17, got=%d", got)
   412  	}
   413  	for i := 0; i < b.N; i++ {
   414  		IndexRune(benchmarkString, 'v')
   415  	}
   416  }
   417  
   418  func BenchmarkIndex(b *testing.B) {
   419  	if got := Index(benchmarkString, "v"); got != 17 {
   420  		b.Fatalf("wrong index: expected 17, got=%d", got)
   421  	}
   422  	for i := 0; i < b.N; i++ {
   423  		Index(benchmarkString, "v")
   424  	}
   425  }
   426  
   427  func BenchmarkLastIndex(b *testing.B) {
   428  	if got := Index(benchmarkString, "v"); got != 17 {
   429  		b.Fatalf("wrong index: expected 17, got=%d", got)
   430  	}
   431  	for i := 0; i < b.N; i++ {
   432  		LastIndex(benchmarkString, "v")
   433  	}
   434  }
   435  
   436  func BenchmarkIndexByte(b *testing.B) {
   437  	if got := IndexByte(benchmarkString, 'v'); got != 17 {
   438  		b.Fatalf("wrong index: expected 17, got=%d", got)
   439  	}
   440  	for i := 0; i < b.N; i++ {
   441  		IndexByte(benchmarkString, 'v')
   442  	}
   443  }
   444  
   445  type SplitTest struct {
   446  	s   string
   447  	sep string
   448  	n   int
   449  	a   []string
   450  }
   451  
   452  var splittests = []SplitTest{
   453  	{"", "", -1, []string{}},
   454  	{abcd, "", 2, []string{"a", "bcd"}},
   455  	{abcd, "", 4, []string{"a", "b", "c", "d"}},
   456  	{abcd, "", -1, []string{"a", "b", "c", "d"}},
   457  	{faces, "", -1, []string{"☺", "☻", "☹"}},
   458  	{faces, "", 3, []string{"☺", "☻", "☹"}},
   459  	{faces, "", 17, []string{"☺", "☻", "☹"}},
   460  	{"☺�☹", "", -1, []string{"☺", "�", "☹"}},
   461  	{abcd, "a", 0, nil},
   462  	{abcd, "a", -1, []string{"", "bcd"}},
   463  	{abcd, "z", -1, []string{"abcd"}},
   464  	{commas, ",", -1, []string{"1", "2", "3", "4"}},
   465  	{dots, "...", -1, []string{"1", ".2", ".3", ".4"}},
   466  	{faces, "☹", -1, []string{"☺☻", ""}},
   467  	{faces, "~", -1, []string{faces}},
   468  	{"1 2 3 4", " ", 3, []string{"1", "2", "3 4"}},
   469  	{"1 2", " ", 3, []string{"1", "2"}},
   470  	{"", "T", math.MaxInt / 4, []string{""}},
   471  	{"\xff-\xff", "", -1, []string{"\xff", "-", "\xff"}},
   472  	{"\xff-\xff", "-", -1, []string{"\xff", "\xff"}},
   473  }
   474  
   475  func TestSplit(t *testing.T) {
   476  	for _, tt := range splittests {
   477  		a := SplitN(tt.s, tt.sep, tt.n)
   478  		if !slices.Equal(a, tt.a) {
   479  			t.Errorf("Split(%q, %q, %d) = %v; want %v", tt.s, tt.sep, tt.n, a, tt.a)
   480  			continue
   481  		}
   482  		if tt.n < 0 {
   483  			a2 := slices.Collect(SplitSeq(tt.s, tt.sep))
   484  			if !slices.Equal(a2, tt.a) {
   485  				t.Errorf(`collect(SplitSeq(%q, %q)) = %v; want %v`, tt.s, tt.sep, a2, tt.a)
   486  			}
   487  		}
   488  		if tt.n == 0 {
   489  			continue
   490  		}
   491  		s := Join(a, tt.sep)
   492  		if s != tt.s {
   493  			t.Errorf("Join(Split(%q, %q, %d), %q) = %q", tt.s, tt.sep, tt.n, tt.sep, s)
   494  		}
   495  		if tt.n < 0 {
   496  			b := Split(tt.s, tt.sep)
   497  			if !slices.Equal(a, b) {
   498  				t.Errorf("Split disagrees with SplitN(%q, %q, %d) = %v; want %v", tt.s, tt.sep, tt.n, b, a)
   499  			}
   500  		}
   501  	}
   502  }
   503  
   504  var splitaftertests = []SplitTest{
   505  	{abcd, "a", -1, []string{"a", "bcd"}},
   506  	{abcd, "z", -1, []string{"abcd"}},
   507  	{abcd, "", -1, []string{"a", "b", "c", "d"}},
   508  	{commas, ",", -1, []string{"1,", "2,", "3,", "4"}},
   509  	{dots, "...", -1, []string{"1...", ".2...", ".3...", ".4"}},
   510  	{faces, "☹", -1, []string{"☺☻☹", ""}},
   511  	{faces, "~", -1, []string{faces}},
   512  	{faces, "", -1, []string{"☺", "☻", "☹"}},
   513  	{"1 2 3 4", " ", 3, []string{"1 ", "2 ", "3 4"}},
   514  	{"1 2 3", " ", 3, []string{"1 ", "2 ", "3"}},
   515  	{"1 2", " ", 3, []string{"1 ", "2"}},
   516  	{"123", "", 2, []string{"1", "23"}},
   517  	{"123", "", 17, []string{"1", "2", "3"}},
   518  }
   519  
   520  func TestSplitAfter(t *testing.T) {
   521  	for _, tt := range splitaftertests {
   522  		a := SplitAfterN(tt.s, tt.sep, tt.n)
   523  		if !slices.Equal(a, tt.a) {
   524  			t.Errorf(`Split(%q, %q, %d) = %v; want %v`, tt.s, tt.sep, tt.n, a, tt.a)
   525  			continue
   526  		}
   527  		if tt.n < 0 {
   528  			a2 := slices.Collect(SplitAfterSeq(tt.s, tt.sep))
   529  			if !slices.Equal(a2, tt.a) {
   530  				t.Errorf(`collect(SplitAfterSeq(%q, %q)) = %v; want %v`, tt.s, tt.sep, a2, tt.a)
   531  			}
   532  		}
   533  		s := Join(a, "")
   534  		if s != tt.s {
   535  			t.Errorf(`Join(Split(%q, %q, %d), %q) = %q`, tt.s, tt.sep, tt.n, tt.sep, s)
   536  		}
   537  		if tt.n < 0 {
   538  			b := SplitAfter(tt.s, tt.sep)
   539  			if !slices.Equal(a, b) {
   540  				t.Errorf("SplitAfter disagrees with SplitAfterN(%q, %q, %d) = %v; want %v", tt.s, tt.sep, tt.n, b, a)
   541  			}
   542  		}
   543  	}
   544  }
   545  
   546  type FieldsTest struct {
   547  	s string
   548  	a []string
   549  }
   550  
   551  var fieldstests = []FieldsTest{
   552  	{"", []string{}},
   553  	{" ", []string{}},
   554  	{" \t ", []string{}},
   555  	{"\u2000", []string{}},
   556  	{"  abc  ", []string{"abc"}},
   557  	{"1 2 3 4", []string{"1", "2", "3", "4"}},
   558  	{"1  2  3  4", []string{"1", "2", "3", "4"}},
   559  	{"1\t\t2\t\t3\t4", []string{"1", "2", "3", "4"}},
   560  	{"1\u20002\u20013\u20024", []string{"1", "2", "3", "4"}},
   561  	{"\u2000\u2001\u2002", []string{}},
   562  	{"\n™\t™\n", []string{"™", "™"}},
   563  	{"\n\u20001™2\u2000 \u2001 ™", []string{"1™2", "™"}},
   564  	{"\n1\uFFFD \uFFFD2\u20003\uFFFD4", []string{"1\uFFFD", "\uFFFD2", "3\uFFFD4"}},
   565  	{"1\xFF\u2000\xFF2\xFF \xFF", []string{"1\xFF", "\xFF2\xFF", "\xFF"}},
   566  	{faces, []string{faces}},
   567  }
   568  
   569  func TestFields(t *testing.T) {
   570  	for _, tt := range fieldstests {
   571  		a := Fields(tt.s)
   572  		if !slices.Equal(a, tt.a) {
   573  			t.Errorf("Fields(%q) = %v; want %v", tt.s, a, tt.a)
   574  			continue
   575  		}
   576  		a2 := collect(t, FieldsSeq(tt.s))
   577  		if !slices.Equal(a2, tt.a) {
   578  			t.Errorf(`collect(FieldsSeq(%q)) = %v; want %v`, tt.s, a2, tt.a)
   579  		}
   580  	}
   581  }
   582  
   583  var FieldsFuncTests = []FieldsTest{
   584  	{"", []string{}},
   585  	{"XX", []string{}},
   586  	{"XXhiXXX", []string{"hi"}},
   587  	{"aXXbXXXcX", []string{"a", "b", "c"}},
   588  }
   589  
   590  func TestFieldsFunc(t *testing.T) {
   591  	for _, tt := range fieldstests {
   592  		a := FieldsFunc(tt.s, unicode.IsSpace)
   593  		if !slices.Equal(a, tt.a) {
   594  			t.Errorf("FieldsFunc(%q, unicode.IsSpace) = %v; want %v", tt.s, a, tt.a)
   595  			continue
   596  		}
   597  	}
   598  	pred := func(c rune) bool { return c == 'X' }
   599  	for _, tt := range FieldsFuncTests {
   600  		a := FieldsFunc(tt.s, pred)
   601  		if !slices.Equal(a, tt.a) {
   602  			t.Errorf("FieldsFunc(%q) = %v, want %v", tt.s, a, tt.a)
   603  		}
   604  		a2 := collect(t, FieldsFuncSeq(tt.s, pred))
   605  		if !slices.Equal(a2, tt.a) {
   606  			t.Errorf(`collect(FieldsFuncSeq(%q)) = %v; want %v`, tt.s, a2, tt.a)
   607  		}
   608  	}
   609  }
   610  
   611  // Test case for any function which accepts and returns a single string.
   612  type StringTest struct {
   613  	in, out string
   614  }
   615  
   616  // Execute f on each test case.  funcName should be the name of f; it's used
   617  // in failure reports.
   618  func runStringTests(t *testing.T, f func(string) string, funcName string, testCases []StringTest) {
   619  	for _, tc := range testCases {
   620  		actual := f(tc.in)
   621  		if actual != tc.out {
   622  			t.Errorf("%s(%q) = %q; want %q", funcName, tc.in, actual, tc.out)
   623  		}
   624  	}
   625  }
   626  
   627  var upperTests = []StringTest{
   628  	{"", ""},
   629  	{"ONLYUPPER", "ONLYUPPER"},
   630  	{"abc", "ABC"},
   631  	{"AbC123", "ABC123"},
   632  	{"azAZ09_", "AZAZ09_"},
   633  	{"longStrinGwitHmixofsmaLLandcAps", "LONGSTRINGWITHMIXOFSMALLANDCAPS"},
   634  	{"RENAN BASTOS 93 AOSDAJDJAIDJAIDAJIaidsjjaidijadsjiadjiOOKKO", "RENAN BASTOS 93 AOSDAJDJAIDJAIDAJIAIDSJJAIDIJADSJIADJIOOKKO"},
   635  	{"long\u0250string\u0250with\u0250nonascii\u2C6Fchars", "LONG\u2C6FSTRING\u2C6FWITH\u2C6FNONASCII\u2C6FCHARS"},
   636  	{"\u0250\u0250\u0250\u0250\u0250", "\u2C6F\u2C6F\u2C6F\u2C6F\u2C6F"}, // grows one byte per char
   637  	{"a\u0080\U0010FFFF", "A\u0080\U0010FFFF"},                           // test utf8.RuneSelf and utf8.MaxRune
   638  }
   639  
   640  var lowerTests = []StringTest{
   641  	{"", ""},
   642  	{"abc", "abc"},
   643  	{"AbC123", "abc123"},
   644  	{"azAZ09_", "azaz09_"},
   645  	{"longStrinGwitHmixofsmaLLandcAps", "longstringwithmixofsmallandcaps"},
   646  	{"renan bastos 93 AOSDAJDJAIDJAIDAJIaidsjjaidijadsjiadjiOOKKO", "renan bastos 93 aosdajdjaidjaidajiaidsjjaidijadsjiadjiookko"},
   647  	{"LONG\u2C6FSTRING\u2C6FWITH\u2C6FNONASCII\u2C6FCHARS", "long\u0250string\u0250with\u0250nonascii\u0250chars"},
   648  	{"\u2C6D\u2C6D\u2C6D\u2C6D\u2C6D", "\u0251\u0251\u0251\u0251\u0251"}, // shrinks one byte per char
   649  	{"A\u0080\U0010FFFF", "a\u0080\U0010FFFF"},                           // test utf8.RuneSelf and utf8.MaxRune
   650  }
   651  
   652  const space = "\t\v\r\f\n\u0085\u00a0\u2000\u3000"
   653  
   654  var trimSpaceTests = []StringTest{
   655  	{"", ""},
   656  	{"abc", "abc"},
   657  	{space + "abc" + space, "abc"},
   658  	{" ", ""},
   659  	{" \t\r\n \t\t\r\r\n\n ", ""},
   660  	{" \t\r\n x\t\t\r\r\n\n ", "x"},
   661  	{" \u2000\t\r\n x\t\t\r\r\ny\n \u3000", "x\t\t\r\r\ny"},
   662  	{"1 \t\r\n2", "1 \t\r\n2"},
   663  	{" x\x80", "x\x80"},
   664  	{" x\xc0", "x\xc0"},
   665  	{"x \xc0\xc0 ", "x \xc0\xc0"},
   666  	{"x \xc0", "x \xc0"},
   667  	{"x \xc0 ", "x \xc0"},
   668  	{"x \xc0\xc0 ", "x \xc0\xc0"},
   669  	{"x ☺\xc0\xc0 ", "x ☺\xc0\xc0"},
   670  	{"x ☺ ", "x ☺"},
   671  }
   672  
   673  func tenRunes(ch rune) string {
   674  	r := make([]rune, 10)
   675  	for i := range r {
   676  		r[i] = ch
   677  	}
   678  	return string(r)
   679  }
   680  
   681  // User-defined self-inverse mapping function
   682  func rot13(r rune) rune {
   683  	step := rune(13)
   684  	if r >= 'a' && r <= 'z' {
   685  		return ((r - 'a' + step) % 26) + 'a'
   686  	}
   687  	if r >= 'A' && r <= 'Z' {
   688  		return ((r - 'A' + step) % 26) + 'A'
   689  	}
   690  	return r
   691  }
   692  
   693  func TestMap(t *testing.T) {
   694  	// Run a couple of awful growth/shrinkage tests
   695  	a := tenRunes('a')
   696  	// 1.  Grow. This triggers two reallocations in Map.
   697  	maxRune := func(rune) rune { return unicode.MaxRune }
   698  	m := Map(maxRune, a)
   699  	expect := tenRunes(unicode.MaxRune)
   700  	if m != expect {
   701  		t.Errorf("growing: expected %q got %q", expect, m)
   702  	}
   703  
   704  	// 2. Shrink
   705  	minRune := func(rune) rune { return 'a' }
   706  	m = Map(minRune, tenRunes(unicode.MaxRune))
   707  	expect = a
   708  	if m != expect {
   709  		t.Errorf("shrinking: expected %q got %q", expect, m)
   710  	}
   711  
   712  	// 3. Rot13
   713  	m = Map(rot13, "a to zed")
   714  	expect = "n gb mrq"
   715  	if m != expect {
   716  		t.Errorf("rot13: expected %q got %q", expect, m)
   717  	}
   718  
   719  	// 4. Rot13^2
   720  	m = Map(rot13, Map(rot13, "a to zed"))
   721  	expect = "a to zed"
   722  	if m != expect {
   723  		t.Errorf("rot13: expected %q got %q", expect, m)
   724  	}
   725  
   726  	// 5. Drop
   727  	dropNotLatin := func(r rune) rune {
   728  		if unicode.Is(unicode.Latin, r) {
   729  			return r
   730  		}
   731  		return -1
   732  	}
   733  	m = Map(dropNotLatin, "Hello, 세계")
   734  	expect = "Hello"
   735  	if m != expect {
   736  		t.Errorf("drop: expected %q got %q", expect, m)
   737  	}
   738  
   739  	// 6. Identity
   740  	identity := func(r rune) rune {
   741  		return r
   742  	}
   743  	orig := "Input string that we expect not to be copied."
   744  	m = Map(identity, orig)
   745  	if unsafe.StringData(orig) != unsafe.StringData(m) {
   746  		t.Error("unexpected copy during identity map")
   747  	}
   748  
   749  	// 7. Handle invalid UTF-8 sequence
   750  	replaceNotLatin := func(r rune) rune {
   751  		if unicode.Is(unicode.Latin, r) {
   752  			return r
   753  		}
   754  		return utf8.RuneError
   755  	}
   756  	m = Map(replaceNotLatin, "Hello\255World")
   757  	expect = "Hello\uFFFDWorld"
   758  	if m != expect {
   759  		t.Errorf("replace invalid sequence: expected %q got %q", expect, m)
   760  	}
   761  
   762  	// 8. Check utf8.RuneSelf and utf8.MaxRune encoding
   763  	encode := func(r rune) rune {
   764  		switch r {
   765  		case utf8.RuneSelf:
   766  			return unicode.MaxRune
   767  		case unicode.MaxRune:
   768  			return utf8.RuneSelf
   769  		}
   770  		return r
   771  	}
   772  	s := string(rune(utf8.RuneSelf)) + string(utf8.MaxRune)
   773  	r := string(utf8.MaxRune) + string(rune(utf8.RuneSelf)) // reverse of s
   774  	m = Map(encode, s)
   775  	if m != r {
   776  		t.Errorf("encoding not handled correctly: expected %q got %q", r, m)
   777  	}
   778  	m = Map(encode, r)
   779  	if m != s {
   780  		t.Errorf("encoding not handled correctly: expected %q got %q", s, m)
   781  	}
   782  
   783  	// 9. Check mapping occurs in the front, middle and back
   784  	trimSpaces := func(r rune) rune {
   785  		if unicode.IsSpace(r) {
   786  			return -1
   787  		}
   788  		return r
   789  	}
   790  	m = Map(trimSpaces, "   abc    123   ")
   791  	expect = "abc123"
   792  	if m != expect {
   793  		t.Errorf("trimSpaces: expected %q got %q", expect, m)
   794  	}
   795  }
   796  
   797  func TestToUpper(t *testing.T) { runStringTests(t, ToUpper, "ToUpper", upperTests) }
   798  
   799  func TestToLower(t *testing.T) { runStringTests(t, ToLower, "ToLower", lowerTests) }
   800  
   801  var toValidUTF8Tests = []struct {
   802  	in   string
   803  	repl string
   804  	out  string
   805  }{
   806  	{"", "\uFFFD", ""},
   807  	{"abc", "\uFFFD", "abc"},
   808  	{"\uFDDD", "\uFFFD", "\uFDDD"},
   809  	{"a\xffb", "\uFFFD", "a\uFFFDb"},
   810  	{"a\xffb\uFFFD", "X", "aXb\uFFFD"},
   811  	{"a☺\xffb☺\xC0\xAFc☺\xff", "", "a☺b☺c☺"},
   812  	{"a☺\xffb☺\xC0\xAFc☺\xff", "日本語", "a☺日本語b☺日本語c☺日本語"},
   813  	{"\xC0\xAF", "\uFFFD", "\uFFFD"},
   814  	{"\xE0\x80\xAF", "\uFFFD", "\uFFFD"},
   815  	{"\xed\xa0\x80", "abc", "abc"},
   816  	{"\xed\xbf\xbf", "\uFFFD", "\uFFFD"},
   817  	{"\xF0\x80\x80\xaf", "☺", "☺"},
   818  	{"\xF8\x80\x80\x80\xAF", "\uFFFD", "\uFFFD"},
   819  	{"\xFC\x80\x80\x80\x80\xAF", "\uFFFD", "\uFFFD"},
   820  }
   821  
   822  func TestToValidUTF8(t *testing.T) {
   823  	for _, tc := range toValidUTF8Tests {
   824  		got := ToValidUTF8(tc.in, tc.repl)
   825  		if got != tc.out {
   826  			t.Errorf("ToValidUTF8(%q, %q) = %q; want %q", tc.in, tc.repl, got, tc.out)
   827  		}
   828  	}
   829  }
   830  
   831  func BenchmarkToUpper(b *testing.B) {
   832  	for _, tc := range upperTests {
   833  		b.Run(tc.in, func(b *testing.B) {
   834  			for i := 0; i < b.N; i++ {
   835  				actual := ToUpper(tc.in)
   836  				if actual != tc.out {
   837  					b.Errorf("ToUpper(%q) = %q; want %q", tc.in, actual, tc.out)
   838  				}
   839  			}
   840  		})
   841  	}
   842  }
   843  
   844  func BenchmarkToLower(b *testing.B) {
   845  	for _, tc := range lowerTests {
   846  		b.Run(tc.in, func(b *testing.B) {
   847  			for i := 0; i < b.N; i++ {
   848  				actual := ToLower(tc.in)
   849  				if actual != tc.out {
   850  					b.Errorf("ToLower(%q) = %q; want %q", tc.in, actual, tc.out)
   851  				}
   852  			}
   853  		})
   854  	}
   855  }
   856  
   857  func BenchmarkMapNoChanges(b *testing.B) {
   858  	identity := func(r rune) rune {
   859  		return r
   860  	}
   861  	for i := 0; i < b.N; i++ {
   862  		Map(identity, "Some string that won't be modified.")
   863  	}
   864  }
   865  
   866  func TestSpecialCase(t *testing.T) {
   867  	lower := "abcçdefgğhıijklmnoöprsştuüvyz"
   868  	upper := "ABCÇDEFGĞHIİJKLMNOÖPRSŞTUÜVYZ"
   869  	u := ToUpperSpecial(unicode.TurkishCase, upper)
   870  	if u != upper {
   871  		t.Errorf("Upper(upper) is %s not %s", u, upper)
   872  	}
   873  	u = ToUpperSpecial(unicode.TurkishCase, lower)
   874  	if u != upper {
   875  		t.Errorf("Upper(lower) is %s not %s", u, upper)
   876  	}
   877  	l := ToLowerSpecial(unicode.TurkishCase, lower)
   878  	if l != lower {
   879  		t.Errorf("Lower(lower) is %s not %s", l, lower)
   880  	}
   881  	l = ToLowerSpecial(unicode.TurkishCase, upper)
   882  	if l != lower {
   883  		t.Errorf("Lower(upper) is %s not %s", l, lower)
   884  	}
   885  }
   886  
   887  func TestTrimSpace(t *testing.T) { runStringTests(t, TrimSpace, "TrimSpace", trimSpaceTests) }
   888  
   889  var trimTests = []struct {
   890  	f            string
   891  	in, arg, out string
   892  }{
   893  	{"Trim", "abba", "a", "bb"},
   894  	{"Trim", "abba", "ab", ""},
   895  	{"TrimLeft", "abba", "ab", ""},
   896  	{"TrimRight", "abba", "ab", ""},
   897  	{"TrimLeft", "abba", "a", "bba"},
   898  	{"TrimLeft", "abba", "b", "abba"},
   899  	{"TrimRight", "abba", "a", "abb"},
   900  	{"TrimRight", "abba", "b", "abba"},
   901  	{"Trim", "<tag>", "<>", "tag"},
   902  	{"Trim", "* listitem", " *", "listitem"},
   903  	{"Trim", `"quote"`, `"`, "quote"},
   904  	{"Trim", "\u2C6F\u2C6F\u0250\u0250\u2C6F\u2C6F", "\u2C6F", "\u0250\u0250"},
   905  	{"Trim", "\x80test\xff", "\xff", "test"},
   906  	{"Trim", " Ġ ", " ", "Ġ"},
   907  	{"Trim", " Ġİ0", "0 ", "Ġİ"},
   908  	//empty string tests
   909  	{"Trim", "abba", "", "abba"},
   910  	{"Trim", "", "123", ""},
   911  	{"Trim", "", "", ""},
   912  	{"TrimLeft", "abba", "", "abba"},
   913  	{"TrimLeft", "", "123", ""},
   914  	{"TrimLeft", "", "", ""},
   915  	{"TrimRight", "abba", "", "abba"},
   916  	{"TrimRight", "", "123", ""},
   917  	{"TrimRight", "", "", ""},
   918  	{"TrimRight", "☺\xc0", "☺", "☺\xc0"},
   919  	{"TrimPrefix", "aabb", "a", "abb"},
   920  	{"TrimPrefix", "aabb", "b", "aabb"},
   921  	{"TrimSuffix", "aabb", "a", "aabb"},
   922  	{"TrimSuffix", "aabb", "b", "aab"},
   923  }
   924  
   925  func TestTrim(t *testing.T) {
   926  	for _, tc := range trimTests {
   927  		name := tc.f
   928  		var f func(string, string) string
   929  		switch name {
   930  		case "Trim":
   931  			f = Trim
   932  		case "TrimLeft":
   933  			f = TrimLeft
   934  		case "TrimRight":
   935  			f = TrimRight
   936  		case "TrimPrefix":
   937  			f = TrimPrefix
   938  		case "TrimSuffix":
   939  			f = TrimSuffix
   940  		default:
   941  			t.Errorf("Undefined trim function %s", name)
   942  		}
   943  		actual := f(tc.in, tc.arg)
   944  		if actual != tc.out {
   945  			t.Errorf("%s(%q, %q) = %q; want %q", name, tc.in, tc.arg, actual, tc.out)
   946  		}
   947  	}
   948  }
   949  
   950  func BenchmarkTrim(b *testing.B) {
   951  	b.ReportAllocs()
   952  
   953  	for i := 0; i < b.N; i++ {
   954  		for _, tc := range trimTests {
   955  			name := tc.f
   956  			var f func(string, string) string
   957  			switch name {
   958  			case "Trim":
   959  				f = Trim
   960  			case "TrimLeft":
   961  				f = TrimLeft
   962  			case "TrimRight":
   963  				f = TrimRight
   964  			case "TrimPrefix":
   965  				f = TrimPrefix
   966  			case "TrimSuffix":
   967  				f = TrimSuffix
   968  			default:
   969  				b.Errorf("Undefined trim function %s", name)
   970  			}
   971  			actual := f(tc.in, tc.arg)
   972  			if actual != tc.out {
   973  				b.Errorf("%s(%q, %q) = %q; want %q", name, tc.in, tc.arg, actual, tc.out)
   974  			}
   975  		}
   976  	}
   977  }
   978  
   979  func BenchmarkToValidUTF8(b *testing.B) {
   980  	tests := []struct {
   981  		name  string
   982  		input string
   983  	}{
   984  		{"Valid", "typical"},
   985  		{"InvalidASCII", "foo\xffbar"},
   986  		{"InvalidNonASCII", "日本語\xff日本語"},
   987  	}
   988  	replacement := "\uFFFD"
   989  	b.ResetTimer()
   990  	for _, test := range tests {
   991  		b.Run(test.name, func(b *testing.B) {
   992  			for i := 0; i < b.N; i++ {
   993  				ToValidUTF8(test.input, replacement)
   994  			}
   995  		})
   996  	}
   997  }
   998  
   999  type predicate struct {
  1000  	f    func(rune) bool
  1001  	name string
  1002  }
  1003  
  1004  var isSpace = predicate{unicode.IsSpace, "IsSpace"}
  1005  var isDigit = predicate{unicode.IsDigit, "IsDigit"}
  1006  var isUpper = predicate{unicode.IsUpper, "IsUpper"}
  1007  var isValidRune = predicate{
  1008  	func(r rune) bool {
  1009  		return r != utf8.RuneError
  1010  	},
  1011  	"IsValidRune",
  1012  }
  1013  
  1014  func not(p predicate) predicate {
  1015  	return predicate{
  1016  		func(r rune) bool {
  1017  			return !p.f(r)
  1018  		},
  1019  		"not " + p.name,
  1020  	}
  1021  }
  1022  
  1023  var trimFuncTests = []struct {
  1024  	f        predicate
  1025  	in       string
  1026  	trimOut  string
  1027  	leftOut  string
  1028  	rightOut string
  1029  }{
  1030  	{isSpace, space + " hello " + space,
  1031  		"hello",
  1032  		"hello " + space,
  1033  		space + " hello"},
  1034  	{isDigit, "\u0e50\u0e5212hello34\u0e50\u0e51",
  1035  		"hello",
  1036  		"hello34\u0e50\u0e51",
  1037  		"\u0e50\u0e5212hello"},
  1038  	{isUpper, "\u2C6F\u2C6F\u2C6F\u2C6FABCDhelloEF\u2C6F\u2C6FGH\u2C6F\u2C6F",
  1039  		"hello",
  1040  		"helloEF\u2C6F\u2C6FGH\u2C6F\u2C6F",
  1041  		"\u2C6F\u2C6F\u2C6F\u2C6FABCDhello"},
  1042  	{not(isSpace), "hello" + space + "hello",
  1043  		space,
  1044  		space + "hello",
  1045  		"hello" + space},
  1046  	{not(isDigit), "hello\u0e50\u0e521234\u0e50\u0e51helo",
  1047  		"\u0e50\u0e521234\u0e50\u0e51",
  1048  		"\u0e50\u0e521234\u0e50\u0e51helo",
  1049  		"hello\u0e50\u0e521234\u0e50\u0e51"},
  1050  	{isValidRune, "ab\xc0a\xc0cd",
  1051  		"\xc0a\xc0",
  1052  		"\xc0a\xc0cd",
  1053  		"ab\xc0a\xc0"},
  1054  	{not(isValidRune), "\xc0a\xc0",
  1055  		"a",
  1056  		"a\xc0",
  1057  		"\xc0a"},
  1058  	{isSpace, "",
  1059  		"",
  1060  		"",
  1061  		""},
  1062  	{isSpace, " ",
  1063  		"",
  1064  		"",
  1065  		""},
  1066  }
  1067  
  1068  func TestTrimFunc(t *testing.T) {
  1069  	for _, tc := range trimFuncTests {
  1070  		trimmers := []struct {
  1071  			name string
  1072  			trim func(s string, f func(r rune) bool) string
  1073  			out  string
  1074  		}{
  1075  			{"TrimFunc", TrimFunc, tc.trimOut},
  1076  			{"TrimLeftFunc", TrimLeftFunc, tc.leftOut},
  1077  			{"TrimRightFunc", TrimRightFunc, tc.rightOut},
  1078  		}
  1079  		for _, trimmer := range trimmers {
  1080  			actual := trimmer.trim(tc.in, tc.f.f)
  1081  			if actual != trimmer.out {
  1082  				t.Errorf("%s(%q, %q) = %q; want %q", trimmer.name, tc.in, tc.f.name, actual, trimmer.out)
  1083  			}
  1084  		}
  1085  	}
  1086  }
  1087  
  1088  var indexFuncTests = []struct {
  1089  	in          string
  1090  	f           predicate
  1091  	first, last int
  1092  }{
  1093  	{"", isValidRune, -1, -1},
  1094  	{"abc", isDigit, -1, -1},
  1095  	{"0123", isDigit, 0, 3},
  1096  	{"a1b", isDigit, 1, 1},
  1097  	{space, isSpace, 0, len(space) - 3}, // last rune in space is 3 bytes
  1098  	{"\u0e50\u0e5212hello34\u0e50\u0e51", isDigit, 0, 18},
  1099  	{"\u2C6F\u2C6F\u2C6F\u2C6FABCDhelloEF\u2C6F\u2C6FGH\u2C6F\u2C6F", isUpper, 0, 34},
  1100  	{"12\u0e50\u0e52hello34\u0e50\u0e51", not(isDigit), 8, 12},
  1101  
  1102  	// tests of invalid UTF-8
  1103  	{"\x801", isDigit, 1, 1},
  1104  	{"\x80abc", isDigit, -1, -1},
  1105  	{"\xc0a\xc0", isValidRune, 1, 1},
  1106  	{"\xc0a\xc0", not(isValidRune), 0, 2},
  1107  	{"\xc0☺\xc0", not(isValidRune), 0, 4},
  1108  	{"\xc0☺\xc0\xc0", not(isValidRune), 0, 5},
  1109  	{"ab\xc0a\xc0cd", not(isValidRune), 2, 4},
  1110  	{"a\xe0\x80cd", not(isValidRune), 1, 2},
  1111  	{"\x80\x80\x80\x80", not(isValidRune), 0, 3},
  1112  }
  1113  
  1114  func TestIndexFunc(t *testing.T) {
  1115  	for _, tc := range indexFuncTests {
  1116  		first := IndexFunc(tc.in, tc.f.f)
  1117  		if first != tc.first {
  1118  			t.Errorf("IndexFunc(%q, %s) = %d; want %d", tc.in, tc.f.name, first, tc.first)
  1119  		}
  1120  		last := LastIndexFunc(tc.in, tc.f.f)
  1121  		if last != tc.last {
  1122  			t.Errorf("LastIndexFunc(%q, %s) = %d; want %d", tc.in, tc.f.name, last, tc.last)
  1123  		}
  1124  	}
  1125  }
  1126  
  1127  func equal(m string, s1, s2 string, t *testing.T) bool {
  1128  	if s1 == s2 {
  1129  		return true
  1130  	}
  1131  	e1 := Split(s1, "")
  1132  	e2 := Split(s2, "")
  1133  	for i, c1 := range e1 {
  1134  		if i >= len(e2) {
  1135  			break
  1136  		}
  1137  		r1, _ := utf8.DecodeRuneInString(c1)
  1138  		r2, _ := utf8.DecodeRuneInString(e2[i])
  1139  		if r1 != r2 {
  1140  			t.Errorf("%s diff at %d: U+%04X U+%04X", m, i, r1, r2)
  1141  		}
  1142  	}
  1143  	return false
  1144  }
  1145  
  1146  func TestCaseConsistency(t *testing.T) {
  1147  	// Make a string of all the runes.
  1148  	numRunes := int(unicode.MaxRune + 1)
  1149  	if testing.Short() {
  1150  		numRunes = 1000
  1151  	}
  1152  	a := make([]rune, numRunes)
  1153  	for i := range a {
  1154  		a[i] = rune(i)
  1155  	}
  1156  	s := string(a)
  1157  	// convert the cases.
  1158  	upper := ToUpper(s)
  1159  	lower := ToLower(s)
  1160  
  1161  	// Consistency checks
  1162  	if n := utf8.RuneCountInString(upper); n != numRunes {
  1163  		t.Error("rune count wrong in upper:", n)
  1164  	}
  1165  	if n := utf8.RuneCountInString(lower); n != numRunes {
  1166  		t.Error("rune count wrong in lower:", n)
  1167  	}
  1168  	if !equal("ToUpper(upper)", ToUpper(upper), upper, t) {
  1169  		t.Error("ToUpper(upper) consistency fail")
  1170  	}
  1171  	if !equal("ToLower(lower)", ToLower(lower), lower, t) {
  1172  		t.Error("ToLower(lower) consistency fail")
  1173  	}
  1174  	/*
  1175  		  These fail because of non-one-to-oneness of the data, such as multiple
  1176  		  upper case 'I' mapping to 'i'.  We comment them out but keep them for
  1177  		  interest.
  1178  		  For instance: CAPITAL LETTER I WITH DOT ABOVE:
  1179  			unicode.ToUpper(unicode.ToLower('\u0130')) != '\u0130'
  1180  
  1181  		if !equal("ToUpper(lower)", ToUpper(lower), upper, t) {
  1182  			t.Error("ToUpper(lower) consistency fail");
  1183  		}
  1184  		if !equal("ToLower(upper)", ToLower(upper), lower, t) {
  1185  			t.Error("ToLower(upper) consistency fail");
  1186  		}
  1187  	*/
  1188  }
  1189  
  1190  var longString = "a" + string(make([]byte, 1<<16)) + "z"
  1191  var longSpaces = func() string {
  1192  	b := make([]byte, 200)
  1193  	for i := range b {
  1194  		b[i] = ' '
  1195  	}
  1196  	return string(b)
  1197  }()
  1198  
  1199  var RepeatTests = []struct {
  1200  	in, out string
  1201  	count   int
  1202  }{
  1203  	{"", "", 0},
  1204  	{"", "", 1},
  1205  	{"", "", 2},
  1206  	{"-", "", 0},
  1207  	{"-", "-", 1},
  1208  	{"-", "----------", 10},
  1209  	{"abc ", "abc abc abc ", 3},
  1210  	{" ", " ", 1},
  1211  	{"--", "----", 2},
  1212  	{"===", "======", 2},
  1213  	{"000", "000000000", 3},
  1214  	{"\t\t\t\t", "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t", 4},
  1215  	{" ", longSpaces, len(longSpaces)},
  1216  	// Tests for results over the chunkLimit
  1217  	{string(rune(0)), string(make([]byte, 1<<16)), 1 << 16},
  1218  	{longString, longString + longString, 2},
  1219  }
  1220  
  1221  func TestRepeat(t *testing.T) {
  1222  	for _, tt := range RepeatTests {
  1223  		a := Repeat(tt.in, tt.count)
  1224  		if !equal("Repeat(s)", a, tt.out, t) {
  1225  			t.Errorf("Repeat(%v, %d) = %v; want %v", tt.in, tt.count, a, tt.out)
  1226  			continue
  1227  		}
  1228  	}
  1229  }
  1230  
  1231  func repeat(s string, count int) (err error) {
  1232  	defer func() {
  1233  		if r := recover(); r != nil {
  1234  			switch v := r.(type) {
  1235  			case error:
  1236  				err = v
  1237  			default:
  1238  				err = fmt.Errorf("%s", v)
  1239  			}
  1240  		}
  1241  	}()
  1242  
  1243  	Repeat(s, count)
  1244  
  1245  	return
  1246  }
  1247  
  1248  // See Issue golang.org/issue/16237
  1249  func TestRepeatCatchesOverflow(t *testing.T) {
  1250  	type testCase struct {
  1251  		s      string
  1252  		count  int
  1253  		errStr string
  1254  	}
  1255  
  1256  	runTestCases := func(prefix string, tests []testCase) {
  1257  		for i, tt := range tests {
  1258  			err := repeat(tt.s, tt.count)
  1259  			if tt.errStr == "" {
  1260  				if err != nil {
  1261  					t.Errorf("#%d panicked %v", i, err)
  1262  				}
  1263  				continue
  1264  			}
  1265  
  1266  			if err == nil || !Contains(err.Error(), tt.errStr) {
  1267  				t.Errorf("%s#%d got %q want %q", prefix, i, err, tt.errStr)
  1268  			}
  1269  		}
  1270  	}
  1271  
  1272  	const maxInt = int(^uint(0) >> 1)
  1273  
  1274  	runTestCases("", []testCase{
  1275  		0: {"--", -2147483647, "negative"},
  1276  		1: {"", maxInt, ""},
  1277  		2: {"-", 10, ""},
  1278  		3: {"gopher", 0, ""},
  1279  		4: {"-", -1, "negative"},
  1280  		5: {"--", -102, "negative"},
  1281  		6: {string(make([]byte, 255)), int((^uint(0))/255 + 1), "overflow"},
  1282  	})
  1283  
  1284  	const is64Bit = 1<<(^uintptr(0)>>63)/2 != 0
  1285  	if !is64Bit {
  1286  		return
  1287  	}
  1288  
  1289  	runTestCases("64-bit", []testCase{
  1290  		0: {"-", maxInt, "out of range"},
  1291  	})
  1292  }
  1293  
  1294  func runesEqual(a, b []rune) bool {
  1295  	if len(a) != len(b) {
  1296  		return false
  1297  	}
  1298  	for i, r := range a {
  1299  		if r != b[i] {
  1300  			return false
  1301  		}
  1302  	}
  1303  	return true
  1304  }
  1305  
  1306  var RunesTests = []struct {
  1307  	in    string
  1308  	out   []rune
  1309  	lossy bool
  1310  }{
  1311  	{"", []rune{}, false},
  1312  	{" ", []rune{32}, false},
  1313  	{"ABC", []rune{65, 66, 67}, false},
  1314  	{"abc", []rune{97, 98, 99}, false},
  1315  	{"\u65e5\u672c\u8a9e", []rune{26085, 26412, 35486}, false},
  1316  	{"ab\x80c", []rune{97, 98, 0xFFFD, 99}, true},
  1317  	{"ab\xc0c", []rune{97, 98, 0xFFFD, 99}, true},
  1318  }
  1319  
  1320  func TestRunes(t *testing.T) {
  1321  	for _, tt := range RunesTests {
  1322  		a := []rune(tt.in)
  1323  		if !runesEqual(a, tt.out) {
  1324  			t.Errorf("[]rune(%q) = %v; want %v", tt.in, a, tt.out)
  1325  			continue
  1326  		}
  1327  		if !tt.lossy {
  1328  			// can only test reassembly if we didn't lose information
  1329  			s := string(a)
  1330  			if s != tt.in {
  1331  				t.Errorf("string([]rune(%q)) = %x; want %x", tt.in, s, tt.in)
  1332  			}
  1333  		}
  1334  	}
  1335  }
  1336  
  1337  func TestReadByte(t *testing.T) {
  1338  	testStrings := []string{"", abcd, faces, commas}
  1339  	for _, s := range testStrings {
  1340  		reader := NewReader(s)
  1341  		if e := reader.UnreadByte(); e == nil {
  1342  			t.Errorf("Unreading %q at beginning: expected error", s)
  1343  		}
  1344  		var res bytes.Buffer
  1345  		for {
  1346  			b, e := reader.ReadByte()
  1347  			if e == io.EOF {
  1348  				break
  1349  			}
  1350  			if e != nil {
  1351  				t.Errorf("Reading %q: %s", s, e)
  1352  				break
  1353  			}
  1354  			res.WriteByte(b)
  1355  			// unread and read again
  1356  			e = reader.UnreadByte()
  1357  			if e != nil {
  1358  				t.Errorf("Unreading %q: %s", s, e)
  1359  				break
  1360  			}
  1361  			b1, e := reader.ReadByte()
  1362  			if e != nil {
  1363  				t.Errorf("Reading %q after unreading: %s", s, e)
  1364  				break
  1365  			}
  1366  			if b1 != b {
  1367  				t.Errorf("Reading %q after unreading: want byte %q, got %q", s, b, b1)
  1368  				break
  1369  			}
  1370  		}
  1371  		if res.String() != s {
  1372  			t.Errorf("Reader(%q).ReadByte() produced %q", s, res.String())
  1373  		}
  1374  	}
  1375  }
  1376  
  1377  func TestReadRune(t *testing.T) {
  1378  	testStrings := []string{"", abcd, faces, commas}
  1379  	for _, s := range testStrings {
  1380  		reader := NewReader(s)
  1381  		if e := reader.UnreadRune(); e == nil {
  1382  			t.Errorf("Unreading %q at beginning: expected error", s)
  1383  		}
  1384  		res := ""
  1385  		for {
  1386  			r, z, e := reader.ReadRune()
  1387  			if e == io.EOF {
  1388  				break
  1389  			}
  1390  			if e != nil {
  1391  				t.Errorf("Reading %q: %s", s, e)
  1392  				break
  1393  			}
  1394  			res += string(r)
  1395  			// unread and read again
  1396  			e = reader.UnreadRune()
  1397  			if e != nil {
  1398  				t.Errorf("Unreading %q: %s", s, e)
  1399  				break
  1400  			}
  1401  			r1, z1, e := reader.ReadRune()
  1402  			if e != nil {
  1403  				t.Errorf("Reading %q after unreading: %s", s, e)
  1404  				break
  1405  			}
  1406  			if r1 != r {
  1407  				t.Errorf("Reading %q after unreading: want rune %q, got %q", s, r, r1)
  1408  				break
  1409  			}
  1410  			if z1 != z {
  1411  				t.Errorf("Reading %q after unreading: want size %d, got %d", s, z, z1)
  1412  				break
  1413  			}
  1414  		}
  1415  		if res != s {
  1416  			t.Errorf("Reader(%q).ReadRune() produced %q", s, res)
  1417  		}
  1418  	}
  1419  }
  1420  
  1421  var UnreadRuneErrorTests = []struct {
  1422  	name string
  1423  	f    func(*Reader)
  1424  }{
  1425  	{"Read", func(r *Reader) { r.Read([]byte{0}) }},
  1426  	{"ReadByte", func(r *Reader) { r.ReadByte() }},
  1427  	{"UnreadRune", func(r *Reader) { r.UnreadRune() }},
  1428  	{"Seek", func(r *Reader) { r.Seek(0, io.SeekCurrent) }},
  1429  	{"WriteTo", func(r *Reader) { r.WriteTo(&bytes.Buffer{}) }},
  1430  }
  1431  
  1432  func TestUnreadRuneError(t *testing.T) {
  1433  	for _, tt := range UnreadRuneErrorTests {
  1434  		reader := NewReader("0123456789")
  1435  		if _, _, err := reader.ReadRune(); err != nil {
  1436  			// should not happen
  1437  			t.Fatal(err)
  1438  		}
  1439  		tt.f(reader)
  1440  		err := reader.UnreadRune()
  1441  		if err == nil {
  1442  			t.Errorf("Unreading after %s: expected error", tt.name)
  1443  		}
  1444  	}
  1445  }
  1446  
  1447  var ReplaceTests = []struct {
  1448  	in       string
  1449  	old, new string
  1450  	n        int
  1451  	out      string
  1452  }{
  1453  	{"hello", "l", "L", 0, "hello"},
  1454  	{"hello", "l", "L", -1, "heLLo"},
  1455  	{"hello", "x", "X", -1, "hello"},
  1456  	{"", "x", "X", -1, ""},
  1457  	{"radar", "r", "<r>", -1, "<r>ada<r>"},
  1458  	{"", "", "<>", -1, "<>"},
  1459  	{"banana", "a", "<>", -1, "b<>n<>n<>"},
  1460  	{"banana", "a", "<>", 1, "b<>nana"},
  1461  	{"banana", "a", "<>", 1000, "b<>n<>n<>"},
  1462  	{"banana", "an", "<>", -1, "b<><>a"},
  1463  	{"banana", "ana", "<>", -1, "b<>na"},
  1464  	{"banana", "", "<>", -1, "<>b<>a<>n<>a<>n<>a<>"},
  1465  	{"banana", "", "<>", 10, "<>b<>a<>n<>a<>n<>a<>"},
  1466  	{"banana", "", "<>", 6, "<>b<>a<>n<>a<>n<>a"},
  1467  	{"banana", "", "<>", 5, "<>b<>a<>n<>a<>na"},
  1468  	{"banana", "", "<>", 1, "<>banana"},
  1469  	{"banana", "a", "a", -1, "banana"},
  1470  	{"banana", "a", "a", 1, "banana"},
  1471  	{"☺☻☹", "", "<>", -1, "<>☺<>☻<>☹<>"},
  1472  }
  1473  
  1474  func TestReplace(t *testing.T) {
  1475  	for _, tt := range ReplaceTests {
  1476  		if s := Replace(tt.in, tt.old, tt.new, tt.n); s != tt.out {
  1477  			t.Errorf("Replace(%q, %q, %q, %d) = %q, want %q", tt.in, tt.old, tt.new, tt.n, s, tt.out)
  1478  		}
  1479  		if tt.n == -1 {
  1480  			s := ReplaceAll(tt.in, tt.old, tt.new)
  1481  			if s != tt.out {
  1482  				t.Errorf("ReplaceAll(%q, %q, %q) = %q, want %q", tt.in, tt.old, tt.new, s, tt.out)
  1483  			}
  1484  		}
  1485  	}
  1486  }
  1487  
  1488  var TitleTests = []struct {
  1489  	in, out string
  1490  }{
  1491  	{"", ""},
  1492  	{"a", "A"},
  1493  	{" aaa aaa aaa ", " Aaa Aaa Aaa "},
  1494  	{" Aaa Aaa Aaa ", " Aaa Aaa Aaa "},
  1495  	{"123a456", "123a456"},
  1496  	{"double-blind", "Double-Blind"},
  1497  	{"ÿøû", "Ÿøû"},
  1498  	{"with_underscore", "With_underscore"},
  1499  	{"unicode \xe2\x80\xa8 line separator", "Unicode \xe2\x80\xa8 Line Separator"},
  1500  }
  1501  
  1502  func TestTitle(t *testing.T) {
  1503  	for _, tt := range TitleTests {
  1504  		if s := Title(tt.in); s != tt.out {
  1505  			t.Errorf("Title(%q) = %q, want %q", tt.in, s, tt.out)
  1506  		}
  1507  	}
  1508  }
  1509  
  1510  var ContainsTests = []struct {
  1511  	str, substr string
  1512  	expected    bool
  1513  }{
  1514  	{"abc", "bc", true},
  1515  	{"abc", "bcd", false},
  1516  	{"abc", "", true},
  1517  	{"", "a", false},
  1518  
  1519  	// cases to cover code in runtime/asm_amd64.s:indexShortStr
  1520  	// 2-byte needle
  1521  	{"xxxxxx", "01", false},
  1522  	{"01xxxx", "01", true},
  1523  	{"xx01xx", "01", true},
  1524  	{"xxxx01", "01", true},
  1525  	{"01xxxxx"[1:], "01", false},
  1526  	{"xxxxx01"[:6], "01", false},
  1527  	// 3-byte needle
  1528  	{"xxxxxxx", "012", false},
  1529  	{"012xxxx", "012", true},
  1530  	{"xx012xx", "012", true},
  1531  	{"xxxx012", "012", true},
  1532  	{"012xxxxx"[1:], "012", false},
  1533  	{"xxxxx012"[:7], "012", false},
  1534  	// 4-byte needle
  1535  	{"xxxxxxxx", "0123", false},
  1536  	{"0123xxxx", "0123", true},
  1537  	{"xx0123xx", "0123", true},
  1538  	{"xxxx0123", "0123", true},
  1539  	{"0123xxxxx"[1:], "0123", false},
  1540  	{"xxxxx0123"[:8], "0123", false},
  1541  	// 5-7-byte needle
  1542  	{"xxxxxxxxx", "01234", false},
  1543  	{"01234xxxx", "01234", true},
  1544  	{"xx01234xx", "01234", true},
  1545  	{"xxxx01234", "01234", true},
  1546  	{"01234xxxxx"[1:], "01234", false},
  1547  	{"xxxxx01234"[:9], "01234", false},
  1548  	// 8-byte needle
  1549  	{"xxxxxxxxxxxx", "01234567", false},
  1550  	{"01234567xxxx", "01234567", true},
  1551  	{"xx01234567xx", "01234567", true},
  1552  	{"xxxx01234567", "01234567", true},
  1553  	{"01234567xxxxx"[1:], "01234567", false},
  1554  	{"xxxxx01234567"[:12], "01234567", false},
  1555  	// 9-15-byte needle
  1556  	{"xxxxxxxxxxxxx", "012345678", false},
  1557  	{"012345678xxxx", "012345678", true},
  1558  	{"xx012345678xx", "012345678", true},
  1559  	{"xxxx012345678", "012345678", true},
  1560  	{"012345678xxxxx"[1:], "012345678", false},
  1561  	{"xxxxx012345678"[:13], "012345678", false},
  1562  	// 16-byte needle
  1563  	{"xxxxxxxxxxxxxxxxxxxx", "0123456789ABCDEF", false},
  1564  	{"0123456789ABCDEFxxxx", "0123456789ABCDEF", true},
  1565  	{"xx0123456789ABCDEFxx", "0123456789ABCDEF", true},
  1566  	{"xxxx0123456789ABCDEF", "0123456789ABCDEF", true},
  1567  	{"0123456789ABCDEFxxxxx"[1:], "0123456789ABCDEF", false},
  1568  	{"xxxxx0123456789ABCDEF"[:20], "0123456789ABCDEF", false},
  1569  	// 17-31-byte needle
  1570  	{"xxxxxxxxxxxxxxxxxxxxx", "0123456789ABCDEFG", false},
  1571  	{"0123456789ABCDEFGxxxx", "0123456789ABCDEFG", true},
  1572  	{"xx0123456789ABCDEFGxx", "0123456789ABCDEFG", true},
  1573  	{"xxxx0123456789ABCDEFG", "0123456789ABCDEFG", true},
  1574  	{"0123456789ABCDEFGxxxxx"[1:], "0123456789ABCDEFG", false},
  1575  	{"xxxxx0123456789ABCDEFG"[:21], "0123456789ABCDEFG", false},
  1576  
  1577  	// partial match cases
  1578  	{"xx01x", "012", false},                             // 3
  1579  	{"xx0123x", "01234", false},                         // 5-7
  1580  	{"xx01234567x", "012345678", false},                 // 9-15
  1581  	{"xx0123456789ABCDEFx", "0123456789ABCDEFG", false}, // 17-31, issue 15679
  1582  }
  1583  
  1584  func TestContains(t *testing.T) {
  1585  	for _, ct := range ContainsTests {
  1586  		if Contains(ct.str, ct.substr) != ct.expected {
  1587  			t.Errorf("Contains(%s, %s) = %v, want %v",
  1588  				ct.str, ct.substr, !ct.expected, ct.expected)
  1589  		}
  1590  	}
  1591  }
  1592  
  1593  var ContainsAnyTests = []struct {
  1594  	str, substr string
  1595  	expected    bool
  1596  }{
  1597  	{"", "", false},
  1598  	{"", "a", false},
  1599  	{"", "abc", false},
  1600  	{"a", "", false},
  1601  	{"a", "a", true},
  1602  	{"aaa", "a", true},
  1603  	{"abc", "xyz", false},
  1604  	{"abc", "xcz", true},
  1605  	{"a☺b☻c☹d", "uvw☻xyz", true},
  1606  	{"aRegExp*", ".(|)*+?^$[]", true},
  1607  	{dots + dots + dots, " ", false},
  1608  }
  1609  
  1610  func TestContainsAny(t *testing.T) {
  1611  	for _, ct := range ContainsAnyTests {
  1612  		if ContainsAny(ct.str, ct.substr) != ct.expected {
  1613  			t.Errorf("ContainsAny(%s, %s) = %v, want %v",
  1614  				ct.str, ct.substr, !ct.expected, ct.expected)
  1615  		}
  1616  	}
  1617  }
  1618  
  1619  var ContainsRuneTests = []struct {
  1620  	str      string
  1621  	r        rune
  1622  	expected bool
  1623  }{
  1624  	{"", 'a', false},
  1625  	{"a", 'a', true},
  1626  	{"aaa", 'a', true},
  1627  	{"abc", 'y', false},
  1628  	{"abc", 'c', true},
  1629  	{"a☺b☻c☹d", 'x', false},
  1630  	{"a☺b☻c☹d", '☻', true},
  1631  	{"aRegExp*", '*', true},
  1632  }
  1633  
  1634  func TestContainsRune(t *testing.T) {
  1635  	for _, ct := range ContainsRuneTests {
  1636  		if ContainsRune(ct.str, ct.r) != ct.expected {
  1637  			t.Errorf("ContainsRune(%q, %q) = %v, want %v",
  1638  				ct.str, ct.r, !ct.expected, ct.expected)
  1639  		}
  1640  	}
  1641  }
  1642  
  1643  func TestContainsFunc(t *testing.T) {
  1644  	for _, ct := range ContainsRuneTests {
  1645  		if ContainsFunc(ct.str, func(r rune) bool {
  1646  			return ct.r == r
  1647  		}) != ct.expected {
  1648  			t.Errorf("ContainsFunc(%q, func(%q)) = %v, want %v",
  1649  				ct.str, ct.r, !ct.expected, ct.expected)
  1650  		}
  1651  	}
  1652  }
  1653  
  1654  var EqualFoldTests = []struct {
  1655  	s, t string
  1656  	out  bool
  1657  }{
  1658  	{"abc", "abc", true},
  1659  	{"ABcd", "ABcd", true},
  1660  	{"123abc", "123ABC", true},
  1661  	{"αβδ", "ΑΒΔ", true},
  1662  	{"abc", "xyz", false},
  1663  	{"abc", "XYZ", false},
  1664  	{"abcdefghijk", "abcdefghijX", false},
  1665  	{"abcdefghijk", "abcdefghij\u212A", true},
  1666  	{"abcdefghijK", "abcdefghij\u212A", true},
  1667  	{"abcdefghijkz", "abcdefghij\u212Ay", false},
  1668  	{"abcdefghijKz", "abcdefghij\u212Ay", false},
  1669  	{"1", "2", false},
  1670  	{"utf-8", "US-ASCII", false},
  1671  }
  1672  
  1673  func TestEqualFold(t *testing.T) {
  1674  	for _, tt := range EqualFoldTests {
  1675  		if out := EqualFold(tt.s, tt.t); out != tt.out {
  1676  			t.Errorf("EqualFold(%#q, %#q) = %v, want %v", tt.s, tt.t, out, tt.out)
  1677  		}
  1678  		if out := EqualFold(tt.t, tt.s); out != tt.out {
  1679  			t.Errorf("EqualFold(%#q, %#q) = %v, want %v", tt.t, tt.s, out, tt.out)
  1680  		}
  1681  	}
  1682  }
  1683  
  1684  func BenchmarkEqualFold(b *testing.B) {
  1685  	b.Run("Tests", func(b *testing.B) {
  1686  		for i := 0; i < b.N; i++ {
  1687  			for _, tt := range EqualFoldTests {
  1688  				if out := EqualFold(tt.s, tt.t); out != tt.out {
  1689  					b.Fatal("wrong result")
  1690  				}
  1691  			}
  1692  		}
  1693  	})
  1694  
  1695  	const s1 = "abcdefghijKz"
  1696  	const s2 = "abcDefGhijKz"
  1697  
  1698  	b.Run("ASCII", func(b *testing.B) {
  1699  		for i := 0; i < b.N; i++ {
  1700  			EqualFold(s1, s2)
  1701  		}
  1702  	})
  1703  
  1704  	b.Run("UnicodePrefix", func(b *testing.B) {
  1705  		for i := 0; i < b.N; i++ {
  1706  			EqualFold("αβδ"+s1, "ΑΒΔ"+s2)
  1707  		}
  1708  	})
  1709  
  1710  	b.Run("UnicodeSuffix", func(b *testing.B) {
  1711  		for i := 0; i < b.N; i++ {
  1712  			EqualFold(s1+"αβδ", s2+"ΑΒΔ")
  1713  		}
  1714  	})
  1715  }
  1716  
  1717  var CountTests = []struct {
  1718  	s, sep string
  1719  	num    int
  1720  }{
  1721  	{"", "", 1},
  1722  	{"", "notempty", 0},
  1723  	{"notempty", "", 9},
  1724  	{"smaller", "not smaller", 0},
  1725  	{"12345678987654321", "6", 2},
  1726  	{"611161116", "6", 3},
  1727  	{"notequal", "NotEqual", 0},
  1728  	{"equal", "equal", 1},
  1729  	{"abc1231231123q", "123", 3},
  1730  	{"11111", "11", 2},
  1731  }
  1732  
  1733  func TestCount(t *testing.T) {
  1734  	for _, tt := range CountTests {
  1735  		if num := Count(tt.s, tt.sep); num != tt.num {
  1736  			t.Errorf("Count(%q, %q) = %d, want %d", tt.s, tt.sep, num, tt.num)
  1737  		}
  1738  	}
  1739  }
  1740  
  1741  var cutTests = []struct {
  1742  	s, sep        string
  1743  	before, after string
  1744  	found         bool
  1745  }{
  1746  	{"abc", "b", "a", "c", true},
  1747  	{"abc", "a", "", "bc", true},
  1748  	{"abc", "c", "ab", "", true},
  1749  	{"abc", "abc", "", "", true},
  1750  	{"abc", "", "", "abc", true},
  1751  	{"abc", "d", "abc", "", false},
  1752  	{"", "d", "", "", false},
  1753  	{"", "", "", "", true},
  1754  }
  1755  
  1756  func TestCut(t *testing.T) {
  1757  	for _, tt := range cutTests {
  1758  		if before, after, found := Cut(tt.s, tt.sep); before != tt.before || after != tt.after || found != tt.found {
  1759  			t.Errorf("Cut(%q, %q) = %q, %q, %v, want %q, %q, %v", tt.s, tt.sep, before, after, found, tt.before, tt.after, tt.found)
  1760  		}
  1761  	}
  1762  }
  1763  
  1764  var cutPrefixTests = []struct {
  1765  	s, sep string
  1766  	after  string
  1767  	found  bool
  1768  }{
  1769  	{"abc", "a", "bc", true},
  1770  	{"abc", "abc", "", true},
  1771  	{"abc", "", "abc", true},
  1772  	{"abc", "d", "abc", false},
  1773  	{"", "d", "", false},
  1774  	{"", "", "", true},
  1775  }
  1776  
  1777  func TestCutPrefix(t *testing.T) {
  1778  	for _, tt := range cutPrefixTests {
  1779  		if after, found := CutPrefix(tt.s, tt.sep); after != tt.after || found != tt.found {
  1780  			t.Errorf("CutPrefix(%q, %q) = %q, %v, want %q, %v", tt.s, tt.sep, after, found, tt.after, tt.found)
  1781  		}
  1782  	}
  1783  }
  1784  
  1785  var cutSuffixTests = []struct {
  1786  	s, sep string
  1787  	before string
  1788  	found  bool
  1789  }{
  1790  	{"abc", "bc", "a", true},
  1791  	{"abc", "abc", "", true},
  1792  	{"abc", "", "abc", true},
  1793  	{"abc", "d", "abc", false},
  1794  	{"", "d", "", false},
  1795  	{"", "", "", true},
  1796  }
  1797  
  1798  func TestCutSuffix(t *testing.T) {
  1799  	for _, tt := range cutSuffixTests {
  1800  		if before, found := CutSuffix(tt.s, tt.sep); before != tt.before || found != tt.found {
  1801  			t.Errorf("CutSuffix(%q, %q) = %q, %v, want %q, %v", tt.s, tt.sep, before, found, tt.before, tt.found)
  1802  		}
  1803  	}
  1804  }
  1805  
  1806  func makeBenchInputHard() string {
  1807  	tokens := [...]string{
  1808  		"<a>", "<p>", "<b>", "<strong>",
  1809  		"</a>", "</p>", "</b>", "</strong>",
  1810  		"hello", "world",
  1811  	}
  1812  	x := make([]byte, 0, 1<<20)
  1813  	for {
  1814  		i := rand.Intn(len(tokens))
  1815  		if len(x)+len(tokens[i]) >= 1<<20 {
  1816  			break
  1817  		}
  1818  		x = append(x, tokens[i]...)
  1819  	}
  1820  	return string(x)
  1821  }
  1822  
  1823  var benchInputHard = makeBenchInputHard()
  1824  
  1825  func benchmarkIndexHard(b *testing.B, sep string) {
  1826  	for i := 0; i < b.N; i++ {
  1827  		Index(benchInputHard, sep)
  1828  	}
  1829  }
  1830  
  1831  func benchmarkLastIndexHard(b *testing.B, sep string) {
  1832  	for i := 0; i < b.N; i++ {
  1833  		LastIndex(benchInputHard, sep)
  1834  	}
  1835  }
  1836  
  1837  func benchmarkCountHard(b *testing.B, sep string) {
  1838  	for i := 0; i < b.N; i++ {
  1839  		Count(benchInputHard, sep)
  1840  	}
  1841  }
  1842  
  1843  func BenchmarkIndexHard1(b *testing.B) { benchmarkIndexHard(b, "<>") }
  1844  func BenchmarkIndexHard2(b *testing.B) { benchmarkIndexHard(b, "</pre>") }
  1845  func BenchmarkIndexHard3(b *testing.B) { benchmarkIndexHard(b, "<b>hello world</b>") }
  1846  func BenchmarkIndexHard4(b *testing.B) {
  1847  	benchmarkIndexHard(b, "<pre><b>hello</b><strong>world</strong></pre>")
  1848  }
  1849  
  1850  func BenchmarkLastIndexHard1(b *testing.B) { benchmarkLastIndexHard(b, "<>") }
  1851  func BenchmarkLastIndexHard2(b *testing.B) { benchmarkLastIndexHard(b, "</pre>") }
  1852  func BenchmarkLastIndexHard3(b *testing.B) { benchmarkLastIndexHard(b, "<b>hello world</b>") }
  1853  
  1854  func BenchmarkCountHard1(b *testing.B) { benchmarkCountHard(b, "<>") }
  1855  func BenchmarkCountHard2(b *testing.B) { benchmarkCountHard(b, "</pre>") }
  1856  func BenchmarkCountHard3(b *testing.B) { benchmarkCountHard(b, "<b>hello world</b>") }
  1857  
  1858  var benchInputTorture = Repeat("ABC", 1<<10) + "123" + Repeat("ABC", 1<<10)
  1859  var benchNeedleTorture = Repeat("ABC", 1<<10+1)
  1860  
  1861  func BenchmarkIndexTorture(b *testing.B) {
  1862  	for i := 0; i < b.N; i++ {
  1863  		Index(benchInputTorture, benchNeedleTorture)
  1864  	}
  1865  }
  1866  
  1867  func BenchmarkCountTorture(b *testing.B) {
  1868  	for i := 0; i < b.N; i++ {
  1869  		Count(benchInputTorture, benchNeedleTorture)
  1870  	}
  1871  }
  1872  
  1873  func BenchmarkCountTortureOverlapping(b *testing.B) {
  1874  	A := Repeat("ABC", 1<<20)
  1875  	B := Repeat("ABC", 1<<10)
  1876  	for i := 0; i < b.N; i++ {
  1877  		Count(A, B)
  1878  	}
  1879  }
  1880  
  1881  func BenchmarkCountByte(b *testing.B) {
  1882  	indexSizes := []int{10, 32, 4 << 10, 4 << 20, 64 << 20}
  1883  	benchStr := Repeat(benchmarkString,
  1884  		(indexSizes[len(indexSizes)-1]+len(benchmarkString)-1)/len(benchmarkString))
  1885  	benchFunc := func(b *testing.B, benchStr string) {
  1886  		b.SetBytes(int64(len(benchStr)))
  1887  		for i := 0; i < b.N; i++ {
  1888  			Count(benchStr, "=")
  1889  		}
  1890  	}
  1891  	for _, size := range indexSizes {
  1892  		b.Run(fmt.Sprintf("%d", size), func(b *testing.B) {
  1893  			benchFunc(b, benchStr[:size])
  1894  		})
  1895  	}
  1896  
  1897  }
  1898  
  1899  var makeFieldsInput = func() string {
  1900  	x := make([]byte, 1<<20)
  1901  	// Input is ~10% space, ~10% 2-byte UTF-8, rest ASCII non-space.
  1902  	for i := range x {
  1903  		switch rand.Intn(10) {
  1904  		case 0:
  1905  			x[i] = ' '
  1906  		case 1:
  1907  			if i > 0 && x[i-1] == 'x' {
  1908  				copy(x[i-1:], "χ")
  1909  				break
  1910  			}
  1911  			fallthrough
  1912  		default:
  1913  			x[i] = 'x'
  1914  		}
  1915  	}
  1916  	return string(x)
  1917  }
  1918  
  1919  var makeFieldsInputASCII = func() string {
  1920  	x := make([]byte, 1<<20)
  1921  	// Input is ~10% space, rest ASCII non-space.
  1922  	for i := range x {
  1923  		if rand.Intn(10) == 0 {
  1924  			x[i] = ' '
  1925  		} else {
  1926  			x[i] = 'x'
  1927  		}
  1928  	}
  1929  	return string(x)
  1930  }
  1931  
  1932  var stringdata = []struct{ name, data string }{
  1933  	{"ASCII", makeFieldsInputASCII()},
  1934  	{"Mixed", makeFieldsInput()},
  1935  }
  1936  
  1937  func BenchmarkFields(b *testing.B) {
  1938  	for _, sd := range stringdata {
  1939  		b.Run(sd.name, func(b *testing.B) {
  1940  			for j := 1 << 4; j <= 1<<20; j <<= 4 {
  1941  				b.Run(fmt.Sprintf("%d", j), func(b *testing.B) {
  1942  					b.ReportAllocs()
  1943  					b.SetBytes(int64(j))
  1944  					data := sd.data[:j]
  1945  					for i := 0; i < b.N; i++ {
  1946  						Fields(data)
  1947  					}
  1948  				})
  1949  			}
  1950  		})
  1951  	}
  1952  }
  1953  
  1954  func BenchmarkFieldsFunc(b *testing.B) {
  1955  	for _, sd := range stringdata {
  1956  		b.Run(sd.name, func(b *testing.B) {
  1957  			for j := 1 << 4; j <= 1<<20; j <<= 4 {
  1958  				b.Run(fmt.Sprintf("%d", j), func(b *testing.B) {
  1959  					b.ReportAllocs()
  1960  					b.SetBytes(int64(j))
  1961  					data := sd.data[:j]
  1962  					for i := 0; i < b.N; i++ {
  1963  						FieldsFunc(data, unicode.IsSpace)
  1964  					}
  1965  				})
  1966  			}
  1967  		})
  1968  	}
  1969  }
  1970  
  1971  func BenchmarkSplitEmptySeparator(b *testing.B) {
  1972  	for i := 0; i < b.N; i++ {
  1973  		Split(benchInputHard, "")
  1974  	}
  1975  }
  1976  
  1977  func BenchmarkSplitSingleByteSeparator(b *testing.B) {
  1978  	for i := 0; i < b.N; i++ {
  1979  		Split(benchInputHard, "/")
  1980  	}
  1981  }
  1982  
  1983  func BenchmarkSplitMultiByteSeparator(b *testing.B) {
  1984  	for i := 0; i < b.N; i++ {
  1985  		Split(benchInputHard, "hello")
  1986  	}
  1987  }
  1988  
  1989  func BenchmarkSplitNSingleByteSeparator(b *testing.B) {
  1990  	for i := 0; i < b.N; i++ {
  1991  		SplitN(benchInputHard, "/", 10)
  1992  	}
  1993  }
  1994  
  1995  func BenchmarkSplitNMultiByteSeparator(b *testing.B) {
  1996  	for i := 0; i < b.N; i++ {
  1997  		SplitN(benchInputHard, "hello", 10)
  1998  	}
  1999  }
  2000  
  2001  func BenchmarkRepeat(b *testing.B) {
  2002  	s := "0123456789"
  2003  	for _, n := range []int{5, 10} {
  2004  		for _, c := range []int{0, 1, 2, 6} {
  2005  			b.Run(fmt.Sprintf("%dx%d", n, c), func(b *testing.B) {
  2006  				for i := 0; i < b.N; i++ {
  2007  					Repeat(s[:n], c)
  2008  				}
  2009  			})
  2010  		}
  2011  	}
  2012  }
  2013  
  2014  func BenchmarkRepeatLarge(b *testing.B) {
  2015  	s := Repeat("@", 8*1024)
  2016  	for j := 8; j <= 30; j++ {
  2017  		for _, k := range []int{1, 16, 4097} {
  2018  			s := s[:k]
  2019  			n := (1 << j) / k
  2020  			if n == 0 {
  2021  				continue
  2022  			}
  2023  			b.Run(fmt.Sprintf("%d/%d", 1<<j, k), func(b *testing.B) {
  2024  				for i := 0; i < b.N; i++ {
  2025  					Repeat(s, n)
  2026  				}
  2027  				b.SetBytes(int64(n * len(s)))
  2028  			})
  2029  		}
  2030  	}
  2031  }
  2032  
  2033  func BenchmarkRepeatSpaces(b *testing.B) {
  2034  	b.ReportAllocs()
  2035  	for i := 0; i < b.N; i++ {
  2036  		Repeat(" ", 2)
  2037  	}
  2038  }
  2039  
  2040  func BenchmarkIndexAnyASCII(b *testing.B) {
  2041  	x := Repeat("#", 2048) // Never matches set
  2042  	cs := "0123456789abcdefghijklmnopqrstuvwxyz0123456789abcdefghijklmnopqrstuvwxyz"
  2043  	for k := 1; k <= 2048; k <<= 4 {
  2044  		for j := 1; j <= 64; j <<= 1 {
  2045  			b.Run(fmt.Sprintf("%d:%d", k, j), func(b *testing.B) {
  2046  				for i := 0; i < b.N; i++ {
  2047  					IndexAny(x[:k], cs[:j])
  2048  				}
  2049  			})
  2050  		}
  2051  	}
  2052  }
  2053  
  2054  func BenchmarkIndexAnyUTF8(b *testing.B) {
  2055  	x := Repeat("#", 2048) // Never matches set
  2056  	cs := "你好世界, hello world. 你好世界, hello world. 你好世界, hello world."
  2057  	for k := 1; k <= 2048; k <<= 4 {
  2058  		for j := 1; j <= 64; j <<= 1 {
  2059  			b.Run(fmt.Sprintf("%d:%d", k, j), func(b *testing.B) {
  2060  				for i := 0; i < b.N; i++ {
  2061  					IndexAny(x[:k], cs[:j])
  2062  				}
  2063  			})
  2064  		}
  2065  	}
  2066  }
  2067  
  2068  func BenchmarkLastIndexAnyASCII(b *testing.B) {
  2069  	x := Repeat("#", 2048) // Never matches set
  2070  	cs := "0123456789abcdefghijklmnopqrstuvwxyz0123456789abcdefghijklmnopqrstuvwxyz"
  2071  	for k := 1; k <= 2048; k <<= 4 {
  2072  		for j := 1; j <= 64; j <<= 1 {
  2073  			b.Run(fmt.Sprintf("%d:%d", k, j), func(b *testing.B) {
  2074  				for i := 0; i < b.N; i++ {
  2075  					LastIndexAny(x[:k], cs[:j])
  2076  				}
  2077  			})
  2078  		}
  2079  	}
  2080  }
  2081  
  2082  func BenchmarkLastIndexAnyUTF8(b *testing.B) {
  2083  	x := Repeat("#", 2048) // Never matches set
  2084  	cs := "你好世界, hello world. 你好世界, hello world. 你好世界, hello world."
  2085  	for k := 1; k <= 2048; k <<= 4 {
  2086  		for j := 1; j <= 64; j <<= 1 {
  2087  			b.Run(fmt.Sprintf("%d:%d", k, j), func(b *testing.B) {
  2088  				for i := 0; i < b.N; i++ {
  2089  					LastIndexAny(x[:k], cs[:j])
  2090  				}
  2091  			})
  2092  		}
  2093  	}
  2094  }
  2095  
  2096  func BenchmarkTrimASCII(b *testing.B) {
  2097  	cs := "0123456789abcdef"
  2098  	for k := 1; k <= 4096; k <<= 4 {
  2099  		for j := 1; j <= 16; j <<= 1 {
  2100  			b.Run(fmt.Sprintf("%d:%d", k, j), func(b *testing.B) {
  2101  				x := Repeat(cs[:j], k) // Always matches set
  2102  				for i := 0; i < b.N; i++ {
  2103  					Trim(x[:k], cs[:j])
  2104  				}
  2105  			})
  2106  		}
  2107  	}
  2108  }
  2109  
  2110  func BenchmarkTrimByte(b *testing.B) {
  2111  	x := "  the quick brown fox   "
  2112  	for i := 0; i < b.N; i++ {
  2113  		Trim(x, " ")
  2114  	}
  2115  }
  2116  
  2117  func BenchmarkIndexPeriodic(b *testing.B) {
  2118  	key := "aa"
  2119  	for _, skip := range [...]int{2, 4, 8, 16, 32, 64} {
  2120  		b.Run(fmt.Sprintf("IndexPeriodic%d", skip), func(b *testing.B) {
  2121  			s := Repeat("a"+Repeat(" ", skip-1), 1<<16/skip)
  2122  			for i := 0; i < b.N; i++ {
  2123  				Index(s, key)
  2124  			}
  2125  		})
  2126  	}
  2127  }
  2128  
  2129  func BenchmarkJoin(b *testing.B) {
  2130  	vals := []string{"red", "yellow", "pink", "green", "purple", "orange", "blue"}
  2131  	for l := 0; l <= len(vals); l++ {
  2132  		b.Run(strconv.Itoa(l), func(b *testing.B) {
  2133  			b.ReportAllocs()
  2134  			vals := vals[:l]
  2135  			for i := 0; i < b.N; i++ {
  2136  				Join(vals, " and ")
  2137  			}
  2138  		})
  2139  	}
  2140  }
  2141  
  2142  func BenchmarkTrimSpace(b *testing.B) {
  2143  	tests := []struct{ name, input string }{
  2144  		{"NoTrim", "typical"},
  2145  		{"ASCII", "  foo bar  "},
  2146  		{"SomeNonASCII", "    \u2000\t\r\n x\t\t\r\r\ny\n \u3000    "},
  2147  		{"JustNonASCII", "\u2000\u2000\u2000☺☺☺☺\u3000\u3000\u3000"},
  2148  	}
  2149  	for _, test := range tests {
  2150  		b.Run(test.name, func(b *testing.B) {
  2151  			for i := 0; i < b.N; i++ {
  2152  				TrimSpace(test.input)
  2153  			}
  2154  		})
  2155  	}
  2156  }
  2157  
  2158  var stringSink string
  2159  
  2160  func BenchmarkReplaceAll(b *testing.B) {
  2161  	b.ReportAllocs()
  2162  	for i := 0; i < b.N; i++ {
  2163  		stringSink = ReplaceAll("banana", "a", "<>")
  2164  	}
  2165  }
  2166  

View as plain text