libgo/go/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         "math/rand"
  12         "reflect"
  13         "runtime"
  14         . "strings"
  15         "testing"
  16         "unicode"
  17         "unicode/utf8"
  18         "unsafe"
  19 )
  20
  21 func eq(a, b []string) bool {
  22         if len(a) != len(b) {
  23                 return false
  24         }
  25         for i := 0; i < len(a); i++ {
  26                 if a[i] != b[i] {
  27                         return false
  28                 }
  29         }
  30         return true
  31 }
  32
  33 var abcd = "abcd"
  34 var faces = "☺☻☹"
  35 var commas = "1,2,3,4"
  36 var dots = "1....2....3....4"
  37
  38 type IndexTest struct {
  39         s   string
  40         sep string
  41         out int
  42 }
  43
  44 var indexTests = []IndexTest{
  45         {"", "", 0},
  46         {"", "a", -1},
  47         {"", "foo", -1},
  48         {"fo", "foo", -1},
  49         {"foo", "foo", 0},
  50         {"oofofoofooo", "f", 2},
  51         {"oofofoofooo", "foo", 4},
  52         {"barfoobarfoo", "foo", 3},
  53         {"foo", "", 0},
  54         {"foo", "o", 1},
  55         {"abcABCabc", "A", 3},
  56         // cases with one byte strings - test special case in Index()
  57         {"", "a", -1},
  58         {"x", "a", -1},
  59         {"x", "x", 0},
  60         {"abc", "a", 0},
  61         {"abc", "b", 1},
  62         {"abc", "c", 2},
  63         {"abc", "x", -1},
  64         // test special cases in Index() for short strings
  65         {"", "ab", -1},
  66         {"bc", "ab", -1},
  67         {"ab", "ab", 0},
  68         {"xab", "ab", 1},
  69         {"xab"[:2], "ab", -1},
  70         {"", "abc", -1},
  71         {"xbc", "abc", -1},
  72         {"abc", "abc", 0},
  73         {"xabc", "abc", 1},
  74         {"xabc"[:3], "abc", -1},
  75         {"xabxc", "abc", -1},
  76         {"", "abcd", -1},
  77         {"xbcd", "abcd", -1},
  78         {"abcd", "abcd", 0},
  79         {"xabcd", "abcd", 1},
  80         {"xyabcd"[:5], "abcd", -1},
  81         {"xbcqq", "abcqq", -1},
  82         {"abcqq", "abcqq", 0},
  83         {"xabcqq", "abcqq", 1},
  84         {"xyabcqq"[:6], "abcqq", -1},
  85         {"xabxcqq", "abcqq", -1},
  86         {"xabcqxq", "abcqq", -1},
  87         {"", "01234567", -1},
  88         {"32145678", "01234567", -1},
  89         {"01234567", "01234567", 0},
  90         {"x01234567", "01234567", 1},
  91         {"x0123456x01234567", "01234567", 9},
  92         {"xx01234567"[:9], "01234567", -1},
  93         {"", "0123456789", -1},
  94         {"3214567844", "0123456789", -1},
  95         {"0123456789", "0123456789", 0},
  96         {"x0123456789", "0123456789", 1},
  97         {"x012345678x0123456789", "0123456789", 11},
  98         {"xyz0123456789"[:12], "0123456789", -1},
  99         {"x01234567x89", "0123456789", -1},
 100         {"", "0123456789012345", -1},
 101         {"3214567889012345", "0123456789012345", -1},
 102         {"0123456789012345", "0123456789012345", 0},
 103         {"x0123456789012345", "0123456789012345", 1},
 104         {"x012345678901234x0123456789012345", "0123456789012345", 17},
 105         {"", "01234567890123456789", -1},
 106         {"32145678890123456789", "01234567890123456789", -1},
 107         {"01234567890123456789", "01234567890123456789", 0},
 108         {"x01234567890123456789", "01234567890123456789", 1},
 109         {"x0123456789012345678x01234567890123456789", "01234567890123456789", 21},
 110         {"xyz01234567890123456789"[:22], "01234567890123456789", -1},
 111         {"", "0123456789012345678901234567890", -1},
 112         {"321456788901234567890123456789012345678911", "0123456789012345678901234567890", -1},
 113         {"0123456789012345678901234567890", "0123456789012345678901234567890", 0},
 114         {"x0123456789012345678901234567890", "0123456789012345678901234567890", 1},
 115         {"x012345678901234567890123456789x0123456789012345678901234567890", "0123456789012345678901234567890", 32},
 116         {"xyz0123456789012345678901234567890"[:33], "0123456789012345678901234567890", -1},
 117         {"", "01234567890123456789012345678901", -1},
 118         {"32145678890123456789012345678901234567890211", "01234567890123456789012345678901", -1},
 119         {"01234567890123456789012345678901", "01234567890123456789012345678901", 0},
 120         {"x01234567890123456789012345678901", "01234567890123456789012345678901", 1},
 121         {"x0123456789012345678901234567890x01234567890123456789012345678901", "01234567890123456789012345678901", 33},
 122         {"xyz01234567890123456789012345678901"[:34], "01234567890123456789012345678901", -1},
 123         {"xxxxxx012345678901234567890123456789012345678901234567890123456789012", "012345678901234567890123456789012345678901234567890123456789012", 6},
 124         {"", "0123456789012345678901234567890123456789", -1},
 125         {"xx012345678901234567890123456789012345678901234567890123456789012", "0123456789012345678901234567890123456789", 2},
 126         {"xx012345678901234567890123456789012345678901234567890123456789012"[:41], "0123456789012345678901234567890123456789", -1},
 127         {"xx012345678901234567890123456789012345678901234567890123456789012", "0123456789012345678901234567890123456xxx", -1},
 128         {"xx0123456789012345678901234567890123456789012345678901234567890120123456789012345678901234567890123456xxx", "0123456789012345678901234567890123456xxx", 65},
 129 }
 130
 131 var lastIndexTests = []IndexTest{
 132         {"", "", 0},
 133         {"", "a", -1},
 134         {"", "foo", -1},
 135         {"fo", "foo", -1},
 136         {"foo", "foo", 0},
 137         {"foo", "f", 0},
 138         {"oofofoofooo", "f", 7},
 139         {"oofofoofooo", "foo", 7},
 140         {"barfoobarfoo", "foo", 9},
 141         {"foo", "", 3},
 142         {"foo", "o", 2},
 143         {"abcABCabc", "A", 3},
 144         {"abcABCabc", "a", 6},
 145 }
 146
 147 var indexAnyTests = []IndexTest{
 148         {"", "", -1},
 149         {"", "a", -1},
 150         {"", "abc", -1},
 151         {"a", "", -1},
 152         {"a", "a", 0},
 153         {"aaa", "a", 0},
 154         {"abc", "xyz", -1},
 155         {"abc", "xcz", 2},
 156         {"ab☺c", "x☺yz", 2},
 157         {"a☺b☻c☹d", "cx", len("a☺b☻")},
 158         {"a☺b☻c☹d", "uvw☻xyz", len("a☺b")},
 159         {"aRegExp*", ".(|)*+?^$[]", 7},
 160         {dots + dots + dots, " ", -1},
 161         {"012abcba210", "\xffb", 4},
 162         {"012\x80bcb\x80210", "\xffb", 3},
 163 }
 164
 165 var lastIndexAnyTests = []IndexTest{
 166         {"", "", -1},
 167         {"", "a", -1},
 168         {"", "abc", -1},
 169         {"a", "", -1},
 170         {"a", "a", 0},
 171         {"aaa", "a", 2},
 172         {"abc", "xyz", -1},
 173         {"abc", "ab", 1},
 174         {"ab☺c", "x☺yz", 2},
 175         {"a☺b☻c☹d", "cx", len("a☺b☻")},
 176         {"a☺b☻c☹d", "uvw☻xyz", len("a☺b")},
 177         {"a.RegExp*", ".(|)*+?^$[]", 8},
 178         {dots + dots + dots, " ", -1},
 179         {"012abcba210", "\xffb", 6},
 180         {"012\x80bcb\x80210", "\xffb", 7},
 181 }
 182
 183 // Execute f on each test case.  funcName should be the name of f; it's used
 184 // in failure reports.
 185 func runIndexTests(t *testing.T, f func(s, sep string) int, funcName string, testCases []IndexTest) {
 186         for _, test := range testCases {
 187                 actual := f(test.s, test.sep)
 188                 if actual != test.out {
 189                         t.Errorf("%s(%q,%q) = %v; want %v", funcName, test.s, test.sep, actual, test.out)
 190                 }
 191         }
 192 }
 193
 194 func TestIndex(t *testing.T)        { runIndexTests(t, Index, "Index", indexTests) }
 195 func TestLastIndex(t *testing.T)    { runIndexTests(t, LastIndex, "LastIndex", lastIndexTests) }
 196 func TestIndexAny(t *testing.T)     { runIndexTests(t, IndexAny, "IndexAny", indexAnyTests) }
 197 func TestLastIndexAny(t *testing.T) { runIndexTests(t, LastIndexAny, "LastIndexAny", lastIndexAnyTests) }
 198
 199 func TestLastIndexByte(t *testing.T) {
 200         testCases := []IndexTest{
 201                 {"", "q", -1},
 202                 {"abcdef", "q", -1},
 203                 {"abcdefabcdef", "a", len("abcdef")},      // something in the middle
 204                 {"abcdefabcdef", "f", len("abcdefabcde")}, // last byte
 205                 {"zabcdefabcdef", "z", 0},                 // first byte
 206                 {"a☺b☻c☹d", "b", len("a☺")},               // non-ascii
 207         }
 208         for _, test := range testCases {
 209                 actual := LastIndexByte(test.s, test.sep[0])
 210                 if actual != test.out {
 211                         t.Errorf("LastIndexByte(%q,%c) = %v; want %v", test.s, test.sep[0], actual, test.out)
 212                 }
 213         }
 214 }
 215
 216 func simpleIndex(s, sep string) int {
 217         n := len(sep)
 218         for i := n; i <= len(s); i++ {
 219                 if s[i-n:i] == sep {
 220                         return i - n
 221                 }
 222         }
 223         return -1
 224 }
 225
 226 func TestIndexRandom(t *testing.T) {
 227         const chars = "abcdefghijklmnopqrstuvwxyz0123456789"
 228         for times := 0; times < 10; times++ {
 229                 for strLen := 5 + rand.Intn(5); strLen < 140; strLen += 10 { // Arbitrary
 230                         s1 := make([]byte, strLen)
 231                         for i := range s1 {
 232                                 s1[i] = chars[rand.Intn(len(chars))]
 233                         }
 234                         s := string(s1)
 235                         for i := 0; i < 50; i++ {
 236                                 begin := rand.Intn(len(s) + 1)
 237                                 end := begin + rand.Intn(len(s)+1-begin)
 238                                 sep := s[begin:end]
 239                                 if i%4 == 0 {
 240                                         pos := rand.Intn(len(sep) + 1)
 241                                         sep = sep[:pos] + "A" + sep[pos:]
 242                                 }
 243                                 want := simpleIndex(s, sep)
 244                                 res := Index(s, sep)
 245                                 if res != want {
 246                                         t.Errorf("Index(%s,%s) = %d; want %d", s, sep, res, want)
 247                                 }
 248                         }
 249                 }
 250         }
 251 }
 252
 253 func TestIndexRune(t *testing.T) {
 254         tests := []struct {
 255                 in   string
 256                 rune rune
 257                 want int
 258         }{
 259                 {"", 'a', -1},
 260                 {"", '☺', -1},
 261                 {"foo", '☹', -1},
 262                 {"foo", 'o', 1},
 263                 {"foo☺bar", '☺', 3},
 264                 {"foo☺☻☹bar", '☹', 9},
 265                 {"a A x", 'A', 2},
 266                 {"some_text=some_value", '=', 9},
 267                 {"☺a", 'a', 3},
 268                 {"a☻☺b", '☺', 4},
 269
 270                 // RuneError should match any invalid UTF-8 byte sequence.
 271                 {"�", '�', 0},
 272                 {"\xff", '�', 0},
 273                 {"☻x�", '�', len("☻x")},
 274                 {"☻x\xe2\x98", '�', len("☻x")},
 275                 {"☻x\xe2\x98�", '�', len("☻x")},
 276                 {"☻x\xe2\x98x", '�', len("☻x")},
 277
 278                 // Invalid rune values should never match.
 279                 {"a☺b☻c☹d\xe2\x98�\xff�\xed\xa0\x80", -1, -1},
 280                 {"a☺b☻c☹d\xe2\x98�\xff�\xed\xa0\x80", 0xD800, -1}, // Surrogate pair
 281                 {"a☺b☻c☹d\xe2\x98�\xff�\xed\xa0\x80", utf8.MaxRune + 1, -1},
 282         }
 283         for _, tt := range tests {
 284                 if got := IndexRune(tt.in, tt.rune); got != tt.want {
 285                         t.Errorf("IndexRune(%q, %d) = %v; want %v", tt.in, tt.rune, got, tt.want)
 286                 }
 287         }
 288
 289         haystack := "test世界"
 290         allocs := testing.AllocsPerRun(1000, func() {
 291                 if i := IndexRune(haystack, 's'); i != 2 {
 292                         t.Fatalf("'s' at %d; want 2", i)
 293                 }
 294                 if i := IndexRune(haystack, '世'); i != 4 {
 295                         t.Fatalf("'世' at %d; want 4", i)
 296                 }
 297         })
 298         if runtime.Compiler == "gccgo" {
 299                 t.Skip("skipping allocations test for gccgo until escape analysis is enabled")
 300         }
 301         if allocs != 0 && testing.CoverMode() == "" {
 302                 t.Errorf("expected no allocations, got %f", allocs)
 303         }
 304 }
 305
 306 const benchmarkString = "some_text=some☺value"
 307
 308 func BenchmarkIndexRune(b *testing.B) {
 309         if got := IndexRune(benchmarkString, '☺'); got != 14 {
 310                 b.Fatalf("wrong index: expected 14, got=%d", got)
 311         }
 312         for i := 0; i < b.N; i++ {
 313                 IndexRune(benchmarkString, '☺')
 314         }
 315 }
 316
 317 var benchmarkLongString = Repeat(" ", 100) + benchmarkString
 318
 319 func BenchmarkIndexRuneLongString(b *testing.B) {
 320         if got := IndexRune(benchmarkLongString, '☺'); got != 114 {
 321                 b.Fatalf("wrong index: expected 114, got=%d", got)
 322         }
 323         for i := 0; i < b.N; i++ {
 324                 IndexRune(benchmarkLongString, '☺')
 325         }
 326 }
 327
 328 func BenchmarkIndexRuneFastPath(b *testing.B) {
 329         if got := IndexRune(benchmarkString, 'v'); got != 17 {
 330                 b.Fatalf("wrong index: expected 17, got=%d", got)
 331         }
 332         for i := 0; i < b.N; i++ {
 333                 IndexRune(benchmarkString, 'v')
 334         }
 335 }
 336
 337 func BenchmarkIndex(b *testing.B) {
 338         if got := Index(benchmarkString, "v"); got != 17 {
 339                 b.Fatalf("wrong index: expected 17, got=%d", got)
 340         }
 341         for i := 0; i < b.N; i++ {
 342                 Index(benchmarkString, "v")
 343         }
 344 }
 345
 346 func BenchmarkLastIndex(b *testing.B) {
 347         if got := Index(benchmarkString, "v"); got != 17 {
 348                 b.Fatalf("wrong index: expected 17, got=%d", got)
 349         }
 350         for i := 0; i < b.N; i++ {
 351                 LastIndex(benchmarkString, "v")
 352         }
 353 }
 354
 355 func BenchmarkIndexByte(b *testing.B) {
 356         if got := IndexByte(benchmarkString, 'v'); got != 17 {
 357                 b.Fatalf("wrong index: expected 17, got=%d", got)
 358         }
 359         for i := 0; i < b.N; i++ {
 360                 IndexByte(benchmarkString, 'v')
 361         }
 362 }
 363
 364 type SplitTest struct {
 365         s   string
 366         sep string
 367         n   int
 368         a   []string
 369 }
 370
 371 var splittests = []SplitTest{
 372         {"", "", -1, []string{}},
 373         {abcd, "", 2, []string{"a", "bcd"}},
 374         {abcd, "", 4, []string{"a", "b", "c", "d"}},
 375         {abcd, "", -1, []string{"a", "b", "c", "d"}},
 376         {faces, "", -1, []string{"☺", "☻", "☹"}},
 377         {faces, "", 3, []string{"☺", "☻", "☹"}},
 378         {faces, "", 17, []string{"☺", "☻", "☹"}},
 379         {"☺�☹", "", -1, []string{"☺", "�", "☹"}},
 380         {abcd, "a", 0, nil},
 381         {abcd, "a", -1, []string{"", "bcd"}},
 382         {abcd, "z", -1, []string{"abcd"}},
 383         {commas, ",", -1, []string{"1", "2", "3", "4"}},
 384         {dots, "...", -1, []string{"1", ".2", ".3", ".4"}},
 385         {faces, "☹", -1, []string{"☺☻", ""}},
 386         {faces, "~", -1, []string{faces}},
 387         {"1 2 3 4", " ", 3, []string{"1", "2", "3 4"}},
 388         {"1 2", " ", 3, []string{"1", "2"}},
 389 }
 390
 391 func TestSplit(t *testing.T) {
 392         for _, tt := range splittests {
 393                 a := SplitN(tt.s, tt.sep, tt.n)
 394                 if !eq(a, tt.a) {
 395                         t.Errorf("Split(%q, %q, %d) = %v; want %v", tt.s, tt.sep, tt.n, a, tt.a)
 396                         continue
 397                 }
 398                 if tt.n == 0 {
 399                         continue
 400                 }
 401                 s := Join(a, tt.sep)
 402                 if s != tt.s {
 403                         t.Errorf("Join(Split(%q, %q, %d), %q) = %q", tt.s, tt.sep, tt.n, tt.sep, s)
 404                 }
 405                 if tt.n < 0 {
 406                         b := Split(tt.s, tt.sep)
 407                         if !reflect.DeepEqual(a, b) {
 408                                 t.Errorf("Split disagrees with SplitN(%q, %q, %d) = %v; want %v", tt.s, tt.sep, tt.n, b, a)
 409                         }
 410                 }
 411         }
 412 }
 413
 414 var splitaftertests = []SplitTest{
 415         {abcd, "a", -1, []string{"a", "bcd"}},
 416         {abcd, "z", -1, []string{"abcd"}},
 417         {abcd, "", -1, []string{"a", "b", "c", "d"}},
 418         {commas, ",", -1, []string{"1,", "2,", "3,", "4"}},
 419         {dots, "...", -1, []string{"1...", ".2...", ".3...", ".4"}},
 420         {faces, "☹", -1, []string{"☺☻☹", ""}},
 421         {faces, "~", -1, []string{faces}},
 422         {faces, "", -1, []string{"☺", "☻", "☹"}},
 423         {"1 2 3 4", " ", 3, []string{"1 ", "2 ", "3 4"}},
 424         {"1 2 3", " ", 3, []string{"1 ", "2 ", "3"}},
 425         {"1 2", " ", 3, []string{"1 ", "2"}},
 426         {"123", "", 2, []string{"1", "23"}},
 427         {"123", "", 17, []string{"1", "2", "3"}},
 428 }
 429
 430 func TestSplitAfter(t *testing.T) {
 431         for _, tt := range splitaftertests {
 432                 a := SplitAfterN(tt.s, tt.sep, tt.n)
 433                 if !eq(a, tt.a) {
 434                         t.Errorf(`Split(%q, %q, %d) = %v; want %v`, tt.s, tt.sep, tt.n, a, tt.a)
 435                         continue
 436                 }
 437                 s := Join(a, "")
 438                 if s != tt.s {
 439                         t.Errorf(`Join(Split(%q, %q, %d), %q) = %q`, tt.s, tt.sep, tt.n, tt.sep, s)
 440                 }
 441                 if tt.n < 0 {
 442                         b := SplitAfter(tt.s, tt.sep)
 443                         if !reflect.DeepEqual(a, b) {
 444                                 t.Errorf("SplitAfter disagrees with SplitAfterN(%q, %q, %d) = %v; want %v", tt.s, tt.sep, tt.n, b, a)
 445                         }
 446                 }
 447         }
 448 }
 449
 450 type FieldsTest struct {
 451         s string
 452         a []string
 453 }
 454
 455 var fieldstests = []FieldsTest{
 456         {"", []string{}},
 457         {" ", []string{}},
 458         {" \t ", []string{}},
 459         {"\u2000", []string{}},
 460         {"  abc  ", []string{"abc"}},
 461         {"1 2 3 4", []string{"1", "2", "3", "4"}},
 462         {"1  2  3  4", []string{"1", "2", "3", "4"}},
 463         {"1\t\t2\t\t3\t4", []string{"1", "2", "3", "4"}},
 464         {"1\u20002\u20013\u20024", []string{"1", "2", "3", "4"}},
 465         {"\u2000\u2001\u2002", []string{}},
 466         {"\n™\t™\n", []string{"™", "™"}},
 467         {"\n\u20001™2\u2000 \u2001 ™", []string{"1™2", "™"}},
 468         {"\n1\uFFFD \uFFFD2\u20003\uFFFD4", []string{"1\uFFFD", "\uFFFD2", "3\uFFFD4"}},
 469         {"1\xFF\u2000\xFF2\xFF \xFF", []string{"1\xFF", "\xFF2\xFF", "\xFF"}},
 470         {faces, []string{faces}},
 471 }
 472
 473 func TestFields(t *testing.T) {
 474         for _, tt := range fieldstests {
 475                 a := Fields(tt.s)
 476                 if !eq(a, tt.a) {
 477                         t.Errorf("Fields(%q) = %v; want %v", tt.s, a, tt.a)
 478                         continue
 479                 }
 480         }
 481 }
 482
 483 var FieldsFuncTests = []FieldsTest{
 484         {"", []string{}},
 485         {"XX", []string{}},
 486         {"XXhiXXX", []string{"hi"}},
 487         {"aXXbXXXcX", []string{"a", "b", "c"}},
 488 }
 489
 490 func TestFieldsFunc(t *testing.T) {
 491         for _, tt := range fieldstests {
 492                 a := FieldsFunc(tt.s, unicode.IsSpace)
 493                 if !eq(a, tt.a) {
 494                         t.Errorf("FieldsFunc(%q, unicode.IsSpace) = %v; want %v", tt.s, a, tt.a)
 495                         continue
 496                 }
 497         }
 498         pred := func(c rune) bool { return c == 'X' }
 499         for _, tt := range FieldsFuncTests {
 500                 a := FieldsFunc(tt.s, pred)
 501                 if !eq(a, tt.a) {
 502                         t.Errorf("FieldsFunc(%q) = %v, want %v", tt.s, a, tt.a)
 503                 }
 504         }
 505 }
 506
 507 // Test case for any function which accepts and returns a single string.
 508 type StringTest struct {
 509         in, out string
 510 }
 511
 512 // Execute f on each test case.  funcName should be the name of f; it's used
 513 // in failure reports.
 514 func runStringTests(t *testing.T, f func(string) string, funcName string, testCases []StringTest) {
 515         for _, tc := range testCases {
 516                 actual := f(tc.in)
 517                 if actual != tc.out {
 518                         t.Errorf("%s(%q) = %q; want %q", funcName, tc.in, actual, tc.out)
 519                 }
 520         }
 521 }
 522
 523 var upperTests = []StringTest{
 524         {"", ""},
 525         {"abc", "ABC"},
 526         {"AbC123", "ABC123"},
 527         {"azAZ09_", "AZAZ09_"},
 528         {"\u0250\u0250\u0250\u0250\u0250", "\u2C6F\u2C6F\u2C6F\u2C6F\u2C6F"}, // grows one byte per char
 529 }
 530
 531 var lowerTests = []StringTest{
 532         {"", ""},
 533         {"abc", "abc"},
 534         {"AbC123", "abc123"},
 535         {"azAZ09_", "azaz09_"},
 536         {"\u2C6D\u2C6D\u2C6D\u2C6D\u2C6D", "\u0251\u0251\u0251\u0251\u0251"}, // shrinks one byte per char
 537 }
 538
 539 const space = "\t\v\r\f\n\u0085\u00a0\u2000\u3000"
 540
 541 var trimSpaceTests = []StringTest{
 542         {"", ""},
 543         {"abc", "abc"},
 544         {space + "abc" + space, "abc"},
 545         {" ", ""},
 546         {" \t\r\n \t\t\r\r\n\n ", ""},
 547         {" \t\r\n x\t\t\r\r\n\n ", "x"},
 548         {" \u2000\t\r\n x\t\t\r\r\ny\n \u3000", "x\t\t\r\r\ny"},
 549         {"1 \t\r\n2", "1 \t\r\n2"},
 550         {" x\x80", "x\x80"},
 551         {" x\xc0", "x\xc0"},
 552         {"x \xc0\xc0 ", "x \xc0\xc0"},
 553         {"x \xc0", "x \xc0"},
 554         {"x \xc0 ", "x \xc0"},
 555         {"x \xc0\xc0 ", "x \xc0\xc0"},
 556         {"x ☺\xc0\xc0 ", "x ☺\xc0\xc0"},
 557         {"x ☺ ", "x ☺"},
 558 }
 559
 560 func tenRunes(ch rune) string {
 561         r := make([]rune, 10)
 562         for i := range r {
 563                 r[i] = ch
 564         }
 565         return string(r)
 566 }
 567
 568 // User-defined self-inverse mapping function
 569 func rot13(r rune) rune {
 570         step := rune(13)
 571         if r >= 'a' && r <= 'z' {
 572                 return ((r - 'a' + step) % 26) + 'a'
 573         }
 574         if r >= 'A' && r <= 'Z' {
 575                 return ((r - 'A' + step) % 26) + 'A'
 576         }
 577         return r
 578 }
 579
 580 func TestMap(t *testing.T) {
 581         // Run a couple of awful growth/shrinkage tests
 582         a := tenRunes('a')
 583         // 1.  Grow. This triggers two reallocations in Map.
 584         maxRune := func(rune) rune { return unicode.MaxRune }
 585         m := Map(maxRune, a)
 586         expect := tenRunes(unicode.MaxRune)
 587         if m != expect {
 588                 t.Errorf("growing: expected %q got %q", expect, m)
 589         }
 590
 591         // 2. Shrink
 592         minRune := func(rune) rune { return 'a' }
 593         m = Map(minRune, tenRunes(unicode.MaxRune))
 594         expect = a
 595         if m != expect {
 596                 t.Errorf("shrinking: expected %q got %q", expect, m)
 597         }
 598
 599         // 3. Rot13
 600         m = Map(rot13, "a to zed")
 601         expect = "n gb mrq"
 602         if m != expect {
 603                 t.Errorf("rot13: expected %q got %q", expect, m)
 604         }
 605
 606         // 4. Rot13^2
 607         m = Map(rot13, Map(rot13, "a to zed"))
 608         expect = "a to zed"
 609         if m != expect {
 610                 t.Errorf("rot13: expected %q got %q", expect, m)
 611         }
 612
 613         // 5. Drop
 614         dropNotLatin := func(r rune) rune {
 615                 if unicode.Is(unicode.Latin, r) {
 616                         return r
 617                 }
 618                 return -1
 619         }
 620         m = Map(dropNotLatin, "Hello, 세계")
 621         expect = "Hello"
 622         if m != expect {
 623                 t.Errorf("drop: expected %q got %q", expect, m)
 624         }
 625
 626         // 6. Identity
 627         identity := func(r rune) rune {
 628                 return r
 629         }
 630         orig := "Input string that we expect not to be copied."
 631         m = Map(identity, orig)
 632         if (*reflect.StringHeader)(unsafe.Pointer(&orig)).Data !=
 633                 (*reflect.StringHeader)(unsafe.Pointer(&m)).Data {
 634                 t.Error("unexpected copy during identity map")
 635         }
 636
 637         // 7. Handle invalid UTF-8 sequence
 638         replaceNotLatin := func(r rune) rune {
 639                 if unicode.Is(unicode.Latin, r) {
 640                         return r
 641                 }
 642                 return '?'
 643         }
 644         m = Map(replaceNotLatin, "Hello\255World")
 645         expect = "Hello?World"
 646         if m != expect {
 647                 t.Errorf("replace invalid sequence: expected %q got %q", expect, m)
 648         }
 649 }
 650
 651 func TestToUpper(t *testing.T) { runStringTests(t, ToUpper, "ToUpper", upperTests) }
 652
 653 func TestToLower(t *testing.T) { runStringTests(t, ToLower, "ToLower", lowerTests) }
 654
 655 func BenchmarkMapNoChanges(b *testing.B) {
 656         identity := func(r rune) rune {
 657                 return r
 658         }
 659         for i := 0; i < b.N; i++ {
 660                 Map(identity, "Some string that won't be modified.")
 661         }
 662 }
 663
 664 func TestSpecialCase(t *testing.T) {
 665         lower := "abcçdefgğhıijklmnoöprsştuüvyz"
 666         upper := "ABCÇDEFGĞHIİJKLMNOÖPRSŞTUÜVYZ"
 667         u := ToUpperSpecial(unicode.TurkishCase, upper)
 668         if u != upper {
 669                 t.Errorf("Upper(upper) is %s not %s", u, upper)
 670         }
 671         u = ToUpperSpecial(unicode.TurkishCase, lower)
 672         if u != upper {
 673                 t.Errorf("Upper(lower) is %s not %s", u, upper)
 674         }
 675         l := ToLowerSpecial(unicode.TurkishCase, lower)
 676         if l != lower {
 677                 t.Errorf("Lower(lower) is %s not %s", l, lower)
 678         }
 679         l = ToLowerSpecial(unicode.TurkishCase, upper)
 680         if l != lower {
 681                 t.Errorf("Lower(upper) is %s not %s", l, lower)
 682         }
 683 }
 684
 685 func TestTrimSpace(t *testing.T) { runStringTests(t, TrimSpace, "TrimSpace", trimSpaceTests) }
 686
 687 var trimTests = []struct {
 688         f            string
 689         in, arg, out string
 690 }{
 691         {"Trim", "abba", "a", "bb"},
 692         {"Trim", "abba", "ab", ""},
 693         {"TrimLeft", "abba", "ab", ""},
 694         {"TrimRight", "abba", "ab", ""},
 695         {"TrimLeft", "abba", "a", "bba"},
 696         {"TrimRight", "abba", "a", "abb"},
 697         {"Trim", "<tag>", "<>", "tag"},
 698         {"Trim", "* listitem", " *", "listitem"},
 699         {"Trim", `"quote"`, `"`, "quote"},
 700         {"Trim", "\u2C6F\u2C6F\u0250\u0250\u2C6F\u2C6F", "\u2C6F", "\u0250\u0250"},
 701         {"Trim", "\x80test\xff", "\xff", "test"},
 702         {"Trim", " Ġ ", " ", "Ġ"},
 703         {"Trim", " Ġİ0", "0 ", "Ġİ"},
 704         //empty string tests
 705         {"Trim", "abba", "", "abba"},
 706         {"Trim", "", "123", ""},
 707         {"Trim", "", "", ""},
 708         {"TrimLeft", "abba", "", "abba"},
 709         {"TrimLeft", "", "123", ""},
 710         {"TrimLeft", "", "", ""},
 711         {"TrimRight", "abba", "", "abba"},
 712         {"TrimRight", "", "123", ""},
 713         {"TrimRight", "", "", ""},
 714         {"TrimRight", "☺\xc0", "☺", "☺\xc0"},
 715         {"TrimPrefix", "aabb", "a", "abb"},
 716         {"TrimPrefix", "aabb", "b", "aabb"},
 717         {"TrimSuffix", "aabb", "a", "aabb"},
 718         {"TrimSuffix", "aabb", "b", "aab"},
 719 }
 720
 721 func TestTrim(t *testing.T) {
 722         for _, tc := range trimTests {
 723                 name := tc.f
 724                 var f func(string, string) string
 725                 switch name {
 726                 case "Trim":
 727                         f = Trim
 728                 case "TrimLeft":
 729                         f = TrimLeft
 730                 case "TrimRight":
 731                         f = TrimRight
 732                 case "TrimPrefix":
 733                         f = TrimPrefix
 734                 case "TrimSuffix":
 735                         f = TrimSuffix
 736                 default:
 737                         t.Errorf("Undefined trim function %s", name)
 738                 }
 739                 actual := f(tc.in, tc.arg)
 740                 if actual != tc.out {
 741                         t.Errorf("%s(%q, %q) = %q; want %q", name, tc.in, tc.arg, actual, tc.out)
 742                 }
 743         }
 744 }
 745
 746 func BenchmarkTrim(b *testing.B) {
 747         b.ReportAllocs()
 748
 749         for i := 0; i < b.N; i++ {
 750                 for _, tc := range trimTests {
 751                         name := tc.f
 752                         var f func(string, string) string
 753                         switch name {
 754                         case "Trim":
 755                                 f = Trim
 756                         case "TrimLeft":
 757                                 f = TrimLeft
 758                         case "TrimRight":
 759                                 f = TrimRight
 760                         case "TrimPrefix":
 761                                 f = TrimPrefix
 762                         case "TrimSuffix":
 763                                 f = TrimSuffix
 764                         default:
 765                                 b.Errorf("Undefined trim function %s", name)
 766                         }
 767                         actual := f(tc.in, tc.arg)
 768                         if actual != tc.out {
 769                                 b.Errorf("%s(%q, %q) = %q; want %q", name, tc.in, tc.arg, actual, tc.out)
 770                         }
 771                 }
 772         }
 773 }
 774
 775 type predicate struct {
 776         f    func(rune) bool
 777         name string
 778 }
 779
 780 var isSpace = predicate{unicode.IsSpace, "IsSpace"}
 781 var isDigit = predicate{unicode.IsDigit, "IsDigit"}
 782 var isUpper = predicate{unicode.IsUpper, "IsUpper"}
 783 var isValidRune = predicate{
 784         func(r rune) bool {
 785                 return r != utf8.RuneError
 786         },
 787         "IsValidRune",
 788 }
 789
 790 func not(p predicate) predicate {
 791         return predicate{
 792                 func(r rune) bool {
 793                         return !p.f(r)
 794                 },
 795                 "not " + p.name,
 796         }
 797 }
 798
 799 var trimFuncTests = []struct {
 800         f       predicate
 801         in, out string
 802 }{
 803         {isSpace, space + " hello " + space, "hello"},
 804         {isDigit, "\u0e50\u0e5212hello34\u0e50\u0e51", "hello"},
 805         {isUpper, "\u2C6F\u2C6F\u2C6F\u2C6FABCDhelloEF\u2C6F\u2C6FGH\u2C6F\u2C6F", "hello"},
 806         {not(isSpace), "hello" + space + "hello", space},
 807         {not(isDigit), "hello\u0e50\u0e521234\u0e50\u0e51helo", "\u0e50\u0e521234\u0e50\u0e51"},
 808         {isValidRune, "ab\xc0a\xc0cd", "\xc0a\xc0"},
 809         {not(isValidRune), "\xc0a\xc0", "a"},
 810 }
 811
 812 func TestTrimFunc(t *testing.T) {
 813         for _, tc := range trimFuncTests {
 814                 actual := TrimFunc(tc.in, tc.f.f)
 815                 if actual != tc.out {
 816                         t.Errorf("TrimFunc(%q, %q) = %q; want %q", tc.in, tc.f.name, actual, tc.out)
 817                 }
 818         }
 819 }
 820
 821 var indexFuncTests = []struct {
 822         in          string
 823         f           predicate
 824         first, last int
 825 }{
 826         {"", isValidRune, -1, -1},
 827         {"abc", isDigit, -1, -1},
 828         {"0123", isDigit, 0, 3},
 829         {"a1b", isDigit, 1, 1},
 830         {space, isSpace, 0, len(space) - 3}, // last rune in space is 3 bytes
 831         {"\u0e50\u0e5212hello34\u0e50\u0e51", isDigit, 0, 18},
 832         {"\u2C6F\u2C6F\u2C6F\u2C6FABCDhelloEF\u2C6F\u2C6FGH\u2C6F\u2C6F", isUpper, 0, 34},
 833         {"12\u0e50\u0e52hello34\u0e50\u0e51", not(isDigit), 8, 12},
 834
 835         // tests of invalid UTF-8
 836         {"\x801", isDigit, 1, 1},
 837         {"\x80abc", isDigit, -1, -1},
 838         {"\xc0a\xc0", isValidRune, 1, 1},
 839         {"\xc0a\xc0", not(isValidRune), 0, 2},
 840         {"\xc0☺\xc0", not(isValidRune), 0, 4},
 841         {"\xc0☺\xc0\xc0", not(isValidRune), 0, 5},
 842         {"ab\xc0a\xc0cd", not(isValidRune), 2, 4},
 843         {"a\xe0\x80cd", not(isValidRune), 1, 2},
 844         {"\x80\x80\x80\x80", not(isValidRune), 0, 3},
 845 }
 846
 847 func TestIndexFunc(t *testing.T) {
 848         for _, tc := range indexFuncTests {
 849                 first := IndexFunc(tc.in, tc.f.f)
 850                 if first != tc.first {
 851                         t.Errorf("IndexFunc(%q, %s) = %d; want %d", tc.in, tc.f.name, first, tc.first)
 852                 }
 853                 last := LastIndexFunc(tc.in, tc.f.f)
 854                 if last != tc.last {
 855                         t.Errorf("LastIndexFunc(%q, %s) = %d; want %d", tc.in, tc.f.name, last, tc.last)
 856                 }
 857         }
 858 }
 859
 860 func equal(m string, s1, s2 string, t *testing.T) bool {
 861         if s1 == s2 {
 862                 return true
 863         }
 864         e1 := Split(s1, "")
 865         e2 := Split(s2, "")
 866         for i, c1 := range e1 {
 867                 if i >= len(e2) {
 868                         break
 869                 }
 870                 r1, _ := utf8.DecodeRuneInString(c1)
 871                 r2, _ := utf8.DecodeRuneInString(e2[i])
 872                 if r1 != r2 {
 873                         t.Errorf("%s diff at %d: U+%04X U+%04X", m, i, r1, r2)
 874                 }
 875         }
 876         return false
 877 }
 878
 879 func TestCaseConsistency(t *testing.T) {
 880         // Make a string of all the runes.
 881         numRunes := int(unicode.MaxRune + 1)
 882         if testing.Short() {
 883                 numRunes = 1000
 884         }
 885         a := make([]rune, numRunes)
 886         for i := range a {
 887                 a[i] = rune(i)
 888         }
 889         s := string(a)
 890         // convert the cases.
 891         upper := ToUpper(s)
 892         lower := ToLower(s)
 893
 894         // Consistency checks
 895         if n := utf8.RuneCountInString(upper); n != numRunes {
 896                 t.Error("rune count wrong in upper:", n)
 897         }
 898         if n := utf8.RuneCountInString(lower); n != numRunes {
 899                 t.Error("rune count wrong in lower:", n)
 900         }
 901         if !equal("ToUpper(upper)", ToUpper(upper), upper, t) {
 902                 t.Error("ToUpper(upper) consistency fail")
 903         }
 904         if !equal("ToLower(lower)", ToLower(lower), lower, t) {
 905                 t.Error("ToLower(lower) consistency fail")
 906         }
 907         /*
 908                   These fail because of non-one-to-oneness of the data, such as multiple
 909                   upper case 'I' mapping to 'i'.  We comment them out but keep them for
 910                   interest.
 911                   For instance: CAPITAL LETTER I WITH DOT ABOVE:
 912                         unicode.ToUpper(unicode.ToLower('\u0130')) != '\u0130'
 913
 914                 if !equal("ToUpper(lower)", ToUpper(lower), upper, t) {
 915                         t.Error("ToUpper(lower) consistency fail");
 916                 }
 917                 if !equal("ToLower(upper)", ToLower(upper), lower, t) {
 918                         t.Error("ToLower(upper) consistency fail");
 919                 }
 920         */
 921 }
 922
 923 var RepeatTests = []struct {
 924         in, out string
 925         count   int
 926 }{
 927         {"", "", 0},
 928         {"", "", 1},
 929         {"", "", 2},
 930         {"-", "", 0},
 931         {"-", "-", 1},
 932         {"-", "----------", 10},
 933         {"abc ", "abc abc abc ", 3},
 934 }
 935
 936 func TestRepeat(t *testing.T) {
 937         for _, tt := range RepeatTests {
 938                 a := Repeat(tt.in, tt.count)
 939                 if !equal("Repeat(s)", a, tt.out, t) {
 940                         t.Errorf("Repeat(%v, %d) = %v; want %v", tt.in, tt.count, a, tt.out)
 941                         continue
 942                 }
 943         }
 944 }
 945
 946 func repeat(s string, count int) (err error) {
 947         defer func() {
 948                 if r := recover(); r != nil {
 949                         switch v := r.(type) {
 950                         case error:
 951                                 err = v
 952                         default:
 953                                 err = fmt.Errorf("%s", v)
 954                         }
 955                 }
 956         }()
 957
 958         Repeat(s, count)
 959
 960         return
 961 }
 962
 963 // See Issue golang.org/issue/16237
 964 func TestRepeatCatchesOverflow(t *testing.T) {
 965         tests := [...]struct {
 966                 s      string
 967                 count  int
 968                 errStr string
 969         }{
 970                 0: {"--", -2147483647, "negative"},
 971                 1: {"", int(^uint(0) >> 1), ""},
 972                 2: {"-", 10, ""},
 973                 3: {"gopher", 0, ""},
 974                 4: {"-", -1, "negative"},
 975                 5: {"--", -102, "negative"},
 976                 6: {string(make([]byte, 255)), int((^uint(0))/255 + 1), "overflow"},
 977         }
 978
 979         for i, tt := range tests {
 980                 err := repeat(tt.s, tt.count)
 981                 if tt.errStr == "" {
 982                         if err != nil {
 983                                 t.Errorf("#%d panicked %v", i, err)
 984                         }
 985                         continue
 986                 }
 987
 988                 if err == nil || !Contains(err.Error(), tt.errStr) {
 989                         t.Errorf("#%d expected %q got %q", i, tt.errStr, err)
 990                 }
 991         }
 992 }
 993
 994 func runesEqual(a, b []rune) bool {
 995         if len(a) != len(b) {
 996                 return false
 997         }
 998         for i, r := range a {
 999                 if r != b[i] {
1000                         return false
1001                 }
1002         }
1003         return true
1004 }
1005
1006 var RunesTests = []struct {
1007         in    string
1008         out   []rune
1009         lossy bool
1010 }{
1011         {"", []rune{}, false},
1012         {" ", []rune{32}, false},
1013         {"ABC", []rune{65, 66, 67}, false},
1014         {"abc", []rune{97, 98, 99}, false},
1015         {"\u65e5\u672c\u8a9e", []rune{26085, 26412, 35486}, false},
1016         {"ab\x80c", []rune{97, 98, 0xFFFD, 99}, true},
1017         {"ab\xc0c", []rune{97, 98, 0xFFFD, 99}, true},
1018 }
1019
1020 func TestRunes(t *testing.T) {
1021         for _, tt := range RunesTests {
1022                 a := []rune(tt.in)
1023                 if !runesEqual(a, tt.out) {
1024                         t.Errorf("[]rune(%q) = %v; want %v", tt.in, a, tt.out)
1025                         continue
1026                 }
1027                 if !tt.lossy {
1028                         // can only test reassembly if we didn't lose information
1029                         s := string(a)
1030                         if s != tt.in {
1031                                 t.Errorf("string([]rune(%q)) = %x; want %x", tt.in, s, tt.in)
1032                         }
1033                 }
1034         }
1035 }
1036
1037 func TestReadByte(t *testing.T) {
1038         testStrings := []string{"", abcd, faces, commas}
1039         for _, s := range testStrings {
1040                 reader := NewReader(s)
1041                 if e := reader.UnreadByte(); e == nil {
1042                         t.Errorf("Unreading %q at beginning: expected error", s)
1043                 }
1044                 var res bytes.Buffer
1045                 for {
1046                         b, e := reader.ReadByte()
1047                         if e == io.EOF {
1048                                 break
1049                         }
1050                         if e != nil {
1051                                 t.Errorf("Reading %q: %s", s, e)
1052                                 break
1053                         }
1054                         res.WriteByte(b)
1055                         // unread and read again
1056                         e = reader.UnreadByte()
1057                         if e != nil {
1058                                 t.Errorf("Unreading %q: %s", s, e)
1059                                 break
1060                         }
1061                         b1, e := reader.ReadByte()
1062                         if e != nil {
1063                                 t.Errorf("Reading %q after unreading: %s", s, e)
1064                                 break
1065                         }
1066                         if b1 != b {
1067                                 t.Errorf("Reading %q after unreading: want byte %q, got %q", s, b, b1)
1068                                 break
1069                         }
1070                 }
1071                 if res.String() != s {
1072                         t.Errorf("Reader(%q).ReadByte() produced %q", s, res.String())
1073                 }
1074         }
1075 }
1076
1077 func TestReadRune(t *testing.T) {
1078         testStrings := []string{"", abcd, faces, commas}
1079         for _, s := range testStrings {
1080                 reader := NewReader(s)
1081                 if e := reader.UnreadRune(); e == nil {
1082                         t.Errorf("Unreading %q at beginning: expected error", s)
1083                 }
1084                 res := ""
1085                 for {
1086                         r, z, e := reader.ReadRune()
1087                         if e == io.EOF {
1088                                 break
1089                         }
1090                         if e != nil {
1091                                 t.Errorf("Reading %q: %s", s, e)
1092                                 break
1093                         }
1094                         res += string(r)
1095                         // unread and read again
1096                         e = reader.UnreadRune()
1097                         if e != nil {
1098                                 t.Errorf("Unreading %q: %s", s, e)
1099                                 break
1100                         }
1101                         r1, z1, e := reader.ReadRune()
1102                         if e != nil {
1103                                 t.Errorf("Reading %q after unreading: %s", s, e)
1104                                 break
1105                         }
1106                         if r1 != r {
1107                                 t.Errorf("Reading %q after unreading: want rune %q, got %q", s, r, r1)
1108                                 break
1109                         }
1110                         if z1 != z {
1111                                 t.Errorf("Reading %q after unreading: want size %d, got %d", s, z, z1)
1112                                 break
1113                         }
1114                 }
1115                 if res != s {
1116                         t.Errorf("Reader(%q).ReadRune() produced %q", s, res)
1117                 }
1118         }
1119 }
1120
1121 var UnreadRuneErrorTests = []struct {
1122         name string
1123         f    func(*Reader)
1124 }{
1125         {"Read", func(r *Reader) { r.Read([]byte{0}) }},
1126         {"ReadByte", func(r *Reader) { r.ReadByte() }},
1127         {"UnreadRune", func(r *Reader) { r.UnreadRune() }},
1128         {"Seek", func(r *Reader) { r.Seek(0, io.SeekCurrent) }},
1129         {"WriteTo", func(r *Reader) { r.WriteTo(&bytes.Buffer{}) }},
1130 }
1131
1132 func TestUnreadRuneError(t *testing.T) {
1133         for _, tt := range UnreadRuneErrorTests {
1134                 reader := NewReader("0123456789")
1135                 if _, _, err := reader.ReadRune(); err != nil {
1136                         // should not happen
1137                         t.Fatal(err)
1138                 }
1139                 tt.f(reader)
1140                 err := reader.UnreadRune()
1141                 if err == nil {
1142                         t.Errorf("Unreading after %s: expected error", tt.name)
1143                 }
1144         }
1145 }
1146
1147 var ReplaceTests = []struct {
1148         in       string
1149         old, new string
1150         n        int
1151         out      string
1152 }{
1153         {"hello", "l", "L", 0, "hello"},
1154         {"hello", "l", "L", -1, "heLLo"},
1155         {"hello", "x", "X", -1, "hello"},
1156         {"", "x", "X", -1, ""},
1157         {"radar", "r", "<r>", -1, "<r>ada<r>"},
1158         {"", "", "<>", -1, "<>"},
1159         {"banana", "a", "<>", -1, "b<>n<>n<>"},
1160         {"banana", "a", "<>", 1, "b<>nana"},
1161         {"banana", "a", "<>", 1000, "b<>n<>n<>"},
1162         {"banana", "an", "<>", -1, "b<><>a"},
1163         {"banana", "ana", "<>", -1, "b<>na"},
1164         {"banana", "", "<>", -1, "<>b<>a<>n<>a<>n<>a<>"},
1165         {"banana", "", "<>", 10, "<>b<>a<>n<>a<>n<>a<>"},
1166         {"banana", "", "<>", 6, "<>b<>a<>n<>a<>n<>a"},
1167         {"banana", "", "<>", 5, "<>b<>a<>n<>a<>na"},
1168         {"banana", "", "<>", 1, "<>banana"},
1169         {"banana", "a", "a", -1, "banana"},
1170         {"banana", "a", "a", 1, "banana"},
1171         {"☺☻☹", "", "<>", -1, "<>☺<>☻<>☹<>"},
1172 }
1173
1174 func TestReplace(t *testing.T) {
1175         for _, tt := range ReplaceTests {
1176                 if s := Replace(tt.in, tt.old, tt.new, tt.n); s != tt.out {
1177                         t.Errorf("Replace(%q, %q, %q, %d) = %q, want %q", tt.in, tt.old, tt.new, tt.n, s, tt.out)
1178                 }
1179         }
1180 }
1181
1182 var TitleTests = []struct {
1183         in, out string
1184 }{
1185         {"", ""},
1186         {"a", "A"},
1187         {" aaa aaa aaa ", " Aaa Aaa Aaa "},
1188         {" Aaa Aaa Aaa ", " Aaa Aaa Aaa "},
1189         {"123a456", "123a456"},
1190         {"double-blind", "Double-Blind"},
1191         {"ÿøû", "Ÿøû"},
1192         {"with_underscore", "With_underscore"},
1193         {"unicode \xe2\x80\xa8 line separator", "Unicode \xe2\x80\xa8 Line Separator"},
1194 }
1195
1196 func TestTitle(t *testing.T) {
1197         for _, tt := range TitleTests {
1198                 if s := Title(tt.in); s != tt.out {
1199                         t.Errorf("Title(%q) = %q, want %q", tt.in, s, tt.out)
1200                 }
1201         }
1202 }
1203
1204 var ContainsTests = []struct {
1205         str, substr string
1206         expected    bool
1207 }{
1208         {"abc", "bc", true},
1209         {"abc", "bcd", false},
1210         {"abc", "", true},
1211         {"", "a", false},
1212
1213         // cases to cover code in runtime/asm_amd64.s:indexShortStr
1214         // 2-byte needle
1215         {"xxxxxx", "01", false},
1216         {"01xxxx", "01", true},
1217         {"xx01xx", "01", true},
1218         {"xxxx01", "01", true},
1219         {"01xxxxx"[1:], "01", false},
1220         {"xxxxx01"[:6], "01", false},
1221         // 3-byte needle
1222         {"xxxxxxx", "012", false},
1223         {"012xxxx", "012", true},
1224         {"xx012xx", "012", true},
1225         {"xxxx012", "012", true},
1226         {"012xxxxx"[1:], "012", false},
1227         {"xxxxx012"[:7], "012", false},
1228         // 4-byte needle
1229         {"xxxxxxxx", "0123", false},
1230         {"0123xxxx", "0123", true},
1231         {"xx0123xx", "0123", true},
1232         {"xxxx0123", "0123", true},
1233         {"0123xxxxx"[1:], "0123", false},
1234         {"xxxxx0123"[:8], "0123", false},
1235         // 5-7-byte needle
1236         {"xxxxxxxxx", "01234", false},
1237         {"01234xxxx", "01234", true},
1238         {"xx01234xx", "01234", true},
1239         {"xxxx01234", "01234", true},
1240         {"01234xxxxx"[1:], "01234", false},
1241         {"xxxxx01234"[:9], "01234", false},
1242         // 8-byte needle
1243         {"xxxxxxxxxxxx", "01234567", false},
1244         {"01234567xxxx", "01234567", true},
1245         {"xx01234567xx", "01234567", true},
1246         {"xxxx01234567", "01234567", true},
1247         {"01234567xxxxx"[1:], "01234567", false},
1248         {"xxxxx01234567"[:12], "01234567", false},
1249         // 9-15-byte needle
1250         {"xxxxxxxxxxxxx", "012345678", false},
1251         {"012345678xxxx", "012345678", true},
1252         {"xx012345678xx", "012345678", true},
1253         {"xxxx012345678", "012345678", true},
1254         {"012345678xxxxx"[1:], "012345678", false},
1255         {"xxxxx012345678"[:13], "012345678", false},
1256         // 16-byte needle
1257         {"xxxxxxxxxxxxxxxxxxxx", "0123456789ABCDEF", false},
1258         {"0123456789ABCDEFxxxx", "0123456789ABCDEF", true},
1259         {"xx0123456789ABCDEFxx", "0123456789ABCDEF", true},
1260         {"xxxx0123456789ABCDEF", "0123456789ABCDEF", true},
1261         {"0123456789ABCDEFxxxxx"[1:], "0123456789ABCDEF", false},
1262         {"xxxxx0123456789ABCDEF"[:20], "0123456789ABCDEF", false},
1263         // 17-31-byte needle
1264         {"xxxxxxxxxxxxxxxxxxxxx", "0123456789ABCDEFG", false},
1265         {"0123456789ABCDEFGxxxx", "0123456789ABCDEFG", true},
1266         {"xx0123456789ABCDEFGxx", "0123456789ABCDEFG", true},
1267         {"xxxx0123456789ABCDEFG", "0123456789ABCDEFG", true},
1268         {"0123456789ABCDEFGxxxxx"[1:], "0123456789ABCDEFG", false},
1269         {"xxxxx0123456789ABCDEFG"[:21], "0123456789ABCDEFG", false},
1270
1271         // partial match cases
1272         {"xx01x", "012", false},                             // 3
1273         {"xx0123x", "01234", false},                         // 5-7
1274         {"xx01234567x", "012345678", false},                 // 9-15
1275         {"xx0123456789ABCDEFx", "0123456789ABCDEFG", false}, // 17-31, issue 15679
1276 }
1277
1278 func TestContains(t *testing.T) {
1279         for _, ct := range ContainsTests {
1280                 if Contains(ct.str, ct.substr) != ct.expected {
1281                         t.Errorf("Contains(%s, %s) = %v, want %v",
1282                                 ct.str, ct.substr, !ct.expected, ct.expected)
1283                 }
1284         }
1285 }
1286
1287 var ContainsAnyTests = []struct {
1288         str, substr string
1289         expected    bool
1290 }{
1291         {"", "", false},
1292         {"", "a", false},
1293         {"", "abc", false},
1294         {"a", "", false},
1295         {"a", "a", true},
1296         {"aaa", "a", true},
1297         {"abc", "xyz", false},
1298         {"abc", "xcz", true},
1299         {"a☺b☻c☹d", "uvw☻xyz", true},
1300         {"aRegExp*", ".(|)*+?^$[]", true},
1301         {dots + dots + dots, " ", false},
1302 }
1303
1304 func TestContainsAny(t *testing.T) {
1305         for _, ct := range ContainsAnyTests {
1306                 if ContainsAny(ct.str, ct.substr) != ct.expected {
1307                         t.Errorf("ContainsAny(%s, %s) = %v, want %v",
1308                                 ct.str, ct.substr, !ct.expected, ct.expected)
1309                 }
1310         }
1311 }
1312
1313 var ContainsRuneTests = []struct {
1314         str      string
1315         r        rune
1316         expected bool
1317 }{
1318         {"", 'a', false},
1319         {"a", 'a', true},
1320         {"aaa", 'a', true},
1321         {"abc", 'y', false},
1322         {"abc", 'c', true},
1323         {"a☺b☻c☹d", 'x', false},
1324         {"a☺b☻c☹d", '☻', true},
1325         {"aRegExp*", '*', true},
1326 }
1327
1328 func TestContainsRune(t *testing.T) {
1329         for _, ct := range ContainsRuneTests {
1330                 if ContainsRune(ct.str, ct.r) != ct.expected {
1331                         t.Errorf("ContainsRune(%q, %q) = %v, want %v",
1332                                 ct.str, ct.r, !ct.expected, ct.expected)
1333                 }
1334         }
1335 }
1336
1337 var EqualFoldTests = []struct {
1338         s, t string
1339         out  bool
1340 }{
1341         {"abc", "abc", true},
1342         {"ABcd", "ABcd", true},
1343         {"123abc", "123ABC", true},
1344         {"αβδ", "ΑΒΔ", true},
1345         {"abc", "xyz", false},
1346         {"abc", "XYZ", false},
1347         {"abcdefghijk", "abcdefghijX", false},
1348         {"abcdefghijk", "abcdefghij\u212A", true},
1349         {"abcdefghijK", "abcdefghij\u212A", true},
1350         {"abcdefghijkz", "abcdefghij\u212Ay", false},
1351         {"abcdefghijKz", "abcdefghij\u212Ay", false},
1352 }
1353
1354 func TestEqualFold(t *testing.T) {
1355         for _, tt := range EqualFoldTests {
1356                 if out := EqualFold(tt.s, tt.t); out != tt.out {
1357                         t.Errorf("EqualFold(%#q, %#q) = %v, want %v", tt.s, tt.t, out, tt.out)
1358                 }
1359                 if out := EqualFold(tt.t, tt.s); out != tt.out {
1360                         t.Errorf("EqualFold(%#q, %#q) = %v, want %v", tt.t, tt.s, out, tt.out)
1361                 }
1362         }
1363 }
1364
1365 var CountTests = []struct {
1366         s, sep string
1367         num    int
1368 }{
1369         {"", "", 1},
1370         {"", "notempty", 0},
1371         {"notempty", "", 9},
1372         {"smaller", "not smaller", 0},
1373         {"12345678987654321", "6", 2},
1374         {"611161116", "6", 3},
1375         {"notequal", "NotEqual", 0},
1376         {"equal", "equal", 1},
1377         {"abc1231231123q", "123", 3},
1378         {"11111", "11", 2},
1379 }
1380
1381 func TestCount(t *testing.T) {
1382         for _, tt := range CountTests {
1383                 if num := Count(tt.s, tt.sep); num != tt.num {
1384                         t.Errorf("Count(\"%s\", \"%s\") = %d, want %d", tt.s, tt.sep, num, tt.num)
1385                 }
1386         }
1387 }
1388
1389 func makeBenchInputHard() string {
1390         tokens := [...]string{
1391                 "<a>", "<p>", "<b>", "<strong>",
1392                 "</a>", "</p>", "</b>", "</strong>",
1393                 "hello", "world",
1394         }
1395         x := make([]byte, 0, 1<<20)
1396         for {
1397                 i := rand.Intn(len(tokens))
1398                 if len(x)+len(tokens[i]) >= 1<<20 {
1399                         break
1400                 }
1401                 x = append(x, tokens[i]...)
1402         }
1403         return string(x)
1404 }
1405
1406 var benchInputHard = makeBenchInputHard()
1407
1408 func benchmarkIndexHard(b *testing.B, sep string) {
1409         for i := 0; i < b.N; i++ {
1410                 Index(benchInputHard, sep)
1411         }
1412 }
1413
1414 func benchmarkLastIndexHard(b *testing.B, sep string) {
1415         for i := 0; i < b.N; i++ {
1416                 LastIndex(benchInputHard, sep)
1417         }
1418 }
1419
1420 func benchmarkCountHard(b *testing.B, sep string) {
1421         for i := 0; i < b.N; i++ {
1422                 Count(benchInputHard, sep)
1423         }
1424 }
1425
1426 func BenchmarkIndexHard1(b *testing.B) { benchmarkIndexHard(b, "<>") }
1427 func BenchmarkIndexHard2(b *testing.B) { benchmarkIndexHard(b, "</pre>") }
1428 func BenchmarkIndexHard3(b *testing.B) { benchmarkIndexHard(b, "<b>hello world</b>") }
1429 func BenchmarkIndexHard4(b *testing.B) {
1430         benchmarkIndexHard(b, "<pre><b>hello</b><strong>world</strong></pre>")
1431 }
1432
1433 func BenchmarkLastIndexHard1(b *testing.B) { benchmarkLastIndexHard(b, "<>") }
1434 func BenchmarkLastIndexHard2(b *testing.B) { benchmarkLastIndexHard(b, "</pre>") }
1435 func BenchmarkLastIndexHard3(b *testing.B) { benchmarkLastIndexHard(b, "<b>hello world</b>") }
1436
1437 func BenchmarkCountHard1(b *testing.B) { benchmarkCountHard(b, "<>") }
1438 func BenchmarkCountHard2(b *testing.B) { benchmarkCountHard(b, "</pre>") }
1439 func BenchmarkCountHard3(b *testing.B) { benchmarkCountHard(b, "<b>hello world</b>") }
1440
1441 var benchInputTorture = Repeat("ABC", 1<<10) + "123" + Repeat("ABC", 1<<10)
1442 var benchNeedleTorture = Repeat("ABC", 1<<10+1)
1443
1444 func BenchmarkIndexTorture(b *testing.B) {
1445         for i := 0; i < b.N; i++ {
1446                 Index(benchInputTorture, benchNeedleTorture)
1447         }
1448 }
1449
1450 func BenchmarkCountTorture(b *testing.B) {
1451         for i := 0; i < b.N; i++ {
1452                 Count(benchInputTorture, benchNeedleTorture)
1453         }
1454 }
1455
1456 func BenchmarkCountTortureOverlapping(b *testing.B) {
1457         A := Repeat("ABC", 1<<20)
1458         B := Repeat("ABC", 1<<10)
1459         for i := 0; i < b.N; i++ {
1460                 Count(A, B)
1461         }
1462 }
1463
1464 func BenchmarkCountByte(b *testing.B) {
1465         indexSizes := []int{10, 32, 4 << 10, 4 << 20, 64 << 20}
1466         benchStr := Repeat(benchmarkString,
1467                 (indexSizes[len(indexSizes)-1]+len(benchmarkString)-1)/len(benchmarkString))
1468         benchFunc := func(b *testing.B, benchStr string) {
1469                 b.SetBytes(int64(len(benchStr)))
1470                 for i := 0; i < b.N; i++ {
1471                         Count(benchStr, "=")
1472                 }
1473         }
1474         for _, size := range indexSizes {
1475                 b.Run(fmt.Sprintf("%d", size), func(b *testing.B) {
1476                         benchFunc(b, benchStr[:size])
1477                 })
1478         }
1479
1480 }
1481
1482 var makeFieldsInput = func() string {
1483         x := make([]byte, 1<<20)
1484         // Input is ~10% space, ~10% 2-byte UTF-8, rest ASCII non-space.
1485         for i := range x {
1486                 switch rand.Intn(10) {
1487                 case 0:
1488                         x[i] = ' '
1489                 case 1:
1490                         if i > 0 && x[i-1] == 'x' {
1491                                 copy(x[i-1:], "χ")
1492                                 break
1493                         }
1494                         fallthrough
1495                 default:
1496                         x[i] = 'x'
1497                 }
1498         }
1499         return string(x)
1500 }
1501
1502 var makeFieldsInputASCII = func() string {
1503         x := make([]byte, 1<<20)
1504         // Input is ~10% space, rest ASCII non-space.
1505         for i := range x {
1506                 if rand.Intn(10) == 0 {
1507                         x[i] = ' '
1508                 } else {
1509                         x[i] = 'x'
1510                 }
1511         }
1512         return string(x)
1513 }
1514
1515 var stringdata = []struct{ name, data string }{
1516         {"ASCII", makeFieldsInputASCII()},
1517         {"Mixed", makeFieldsInput()},
1518 }
1519
1520 func BenchmarkFields(b *testing.B) {
1521         for _, sd := range stringdata {
1522                 b.Run(sd.name, func(b *testing.B) {
1523                         for j := 1 << 4; j <= 1<<20; j <<= 4 {
1524                                 b.Run(fmt.Sprintf("%d", j), func(b *testing.B) {
1525                                         b.ReportAllocs()
1526                                         b.SetBytes(int64(j))
1527                                         data := sd.data[:j]
1528                                         for i := 0; i < b.N; i++ {
1529                                                 Fields(data)
1530                                         }
1531                                 })
1532                         }
1533                 })
1534         }
1535 }
1536
1537 func BenchmarkFieldsFunc(b *testing.B) {
1538         for _, sd := range stringdata {
1539                 b.Run(sd.name, func(b *testing.B) {
1540                         for j := 1 << 4; j <= 1<<20; j <<= 4 {
1541                                 b.Run(fmt.Sprintf("%d", j), func(b *testing.B) {
1542                                         b.ReportAllocs()
1543                                         b.SetBytes(int64(j))
1544                                         data := sd.data[:j]
1545                                         for i := 0; i < b.N; i++ {
1546                                                 FieldsFunc(data, unicode.IsSpace)
1547                                         }
1548                                 })
1549                         }
1550                 })
1551         }
1552 }
1553
1554 func BenchmarkSplitEmptySeparator(b *testing.B) {
1555         for i := 0; i < b.N; i++ {
1556                 Split(benchInputHard, "")
1557         }
1558 }
1559
1560 func BenchmarkSplitSingleByteSeparator(b *testing.B) {
1561         for i := 0; i < b.N; i++ {
1562                 Split(benchInputHard, "/")
1563         }
1564 }
1565
1566 func BenchmarkSplitMultiByteSeparator(b *testing.B) {
1567         for i := 0; i < b.N; i++ {
1568                 Split(benchInputHard, "hello")
1569         }
1570 }
1571
1572 func BenchmarkSplitNSingleByteSeparator(b *testing.B) {
1573         for i := 0; i < b.N; i++ {
1574                 SplitN(benchInputHard, "/", 10)
1575         }
1576 }
1577
1578 func BenchmarkSplitNMultiByteSeparator(b *testing.B) {
1579         for i := 0; i < b.N; i++ {
1580                 SplitN(benchInputHard, "hello", 10)
1581         }
1582 }
1583
1584 func BenchmarkRepeat(b *testing.B) {
1585         for i := 0; i < b.N; i++ {
1586                 Repeat("-", 80)
1587         }
1588 }
1589
1590 func BenchmarkIndexAnyASCII(b *testing.B) {
1591         x := Repeat("#", 4096) // Never matches set
1592         cs := "0123456789abcdef"
1593         for k := 1; k <= 4096; k <<= 4 {
1594                 for j := 1; j <= 16; j <<= 1 {
1595                         b.Run(fmt.Sprintf("%d:%d", k, j), func(b *testing.B) {
1596                                 for i := 0; i < b.N; i++ {
1597                                         IndexAny(x[:k], cs[:j])
1598                                 }
1599                         })
1600                 }
1601         }
1602 }
1603
1604 func BenchmarkTrimASCII(b *testing.B) {
1605         cs := "0123456789abcdef"
1606         for k := 1; k <= 4096; k <<= 4 {
1607                 for j := 1; j <= 16; j <<= 1 {
1608                         b.Run(fmt.Sprintf("%d:%d", k, j), func(b *testing.B) {
1609                                 x := Repeat(cs[:j], k) // Always matches set
1610                                 for i := 0; i < b.N; i++ {
1611                                         Trim(x[:k], cs[:j])
1612                                 }
1613                         })
1614                 }
1615         }
1616 }