github.com/twelsh-aw/go/src@v0.0.0-20230516233729-a56fe86a7c81/bytes/bytes_test.go (about) 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 bytes_test 6 7 import ( 8 . "bytes" 9 "fmt" 10 "internal/testenv" 11 "math" 12 "math/rand" 13 "reflect" 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 func sliceOfString(s [][]byte) []string { 34 result := make([]string, len(s)) 35 for i, v := range s { 36 result[i] = string(v) 37 } 38 return result 39 } 40 41 // For ease of reading, the test cases use strings that are converted to byte 42 // slices before invoking the functions. 43 44 var abcd = "abcd" 45 var faces = "☺☻☹" 46 var commas = "1,2,3,4" 47 var dots = "1....2....3....4" 48 49 type BinOpTest struct { 50 a string 51 b string 52 i int 53 } 54 55 func TestEqual(t *testing.T) { 56 // Run the tests and check for allocation at the same time. 57 allocs := testing.AllocsPerRun(10, func() { 58 for _, tt := range compareTests { 59 eql := Equal(tt.a, tt.b) 60 if eql != (tt.i == 0) { 61 t.Errorf(`Equal(%q, %q) = %v`, tt.a, tt.b, eql) 62 } 63 } 64 }) 65 if allocs > 0 { 66 t.Errorf("Equal allocated %v times", allocs) 67 } 68 } 69 70 func TestEqualExhaustive(t *testing.T) { 71 var size = 128 72 if testing.Short() { 73 size = 32 74 } 75 a := make([]byte, size) 76 b := make([]byte, size) 77 b_init := make([]byte, size) 78 // randomish but deterministic data 79 for i := 0; i < size; i++ { 80 a[i] = byte(17 * i) 81 b_init[i] = byte(23*i + 100) 82 } 83 84 for len := 0; len <= size; len++ { 85 for x := 0; x <= size-len; x++ { 86 for y := 0; y <= size-len; y++ { 87 copy(b, b_init) 88 copy(b[y:y+len], a[x:x+len]) 89 if !Equal(a[x:x+len], b[y:y+len]) || !Equal(b[y:y+len], a[x:x+len]) { 90 t.Errorf("Equal(%d, %d, %d) = false", len, x, y) 91 } 92 } 93 } 94 } 95 } 96 97 // make sure Equal returns false for minimally different strings. The data 98 // is all zeros except for a single one in one location. 99 func TestNotEqual(t *testing.T) { 100 var size = 128 101 if testing.Short() { 102 size = 32 103 } 104 a := make([]byte, size) 105 b := make([]byte, size) 106 107 for len := 0; len <= size; len++ { 108 for x := 0; x <= size-len; x++ { 109 for y := 0; y <= size-len; y++ { 110 for diffpos := x; diffpos < x+len; diffpos++ { 111 a[diffpos] = 1 112 if Equal(a[x:x+len], b[y:y+len]) || Equal(b[y:y+len], a[x:x+len]) { 113 t.Errorf("NotEqual(%d, %d, %d, %d) = true", len, x, y, diffpos) 114 } 115 a[diffpos] = 0 116 } 117 } 118 } 119 } 120 } 121 122 var indexTests = []BinOpTest{ 123 {"", "", 0}, 124 {"", "a", -1}, 125 {"", "foo", -1}, 126 {"fo", "foo", -1}, 127 {"foo", "baz", -1}, 128 {"foo", "foo", 0}, 129 {"oofofoofooo", "f", 2}, 130 {"oofofoofooo", "foo", 4}, 131 {"barfoobarfoo", "foo", 3}, 132 {"foo", "", 0}, 133 {"foo", "o", 1}, 134 {"abcABCabc", "A", 3}, 135 // cases with one byte strings - test IndexByte and special case in Index() 136 {"", "a", -1}, 137 {"x", "a", -1}, 138 {"x", "x", 0}, 139 {"abc", "a", 0}, 140 {"abc", "b", 1}, 141 {"abc", "c", 2}, 142 {"abc", "x", -1}, 143 {"barfoobarfooyyyzzzyyyzzzyyyzzzyyyxxxzzzyyy", "x", 33}, 144 {"fofofofooofoboo", "oo", 7}, 145 {"fofofofofofoboo", "ob", 11}, 146 {"fofofofofofoboo", "boo", 12}, 147 {"fofofofofofoboo", "oboo", 11}, 148 {"fofofofofoooboo", "fooo", 8}, 149 {"fofofofofofoboo", "foboo", 10}, 150 {"fofofofofofoboo", "fofob", 8}, 151 {"fofofofofofofoffofoobarfoo", "foffof", 12}, 152 {"fofofofofoofofoffofoobarfoo", "foffof", 13}, 153 {"fofofofofofofoffofoobarfoo", "foffofo", 12}, 154 {"fofofofofoofofoffofoobarfoo", "foffofo", 13}, 155 {"fofofofofoofofoffofoobarfoo", "foffofoo", 13}, 156 {"fofofofofofofoffofoobarfoo", "foffofoo", 12}, 157 {"fofofofofoofofoffofoobarfoo", "foffofoob", 13}, 158 {"fofofofofofofoffofoobarfoo", "foffofoob", 12}, 159 {"fofofofofoofofoffofoobarfoo", "foffofooba", 13}, 160 {"fofofofofofofoffofoobarfoo", "foffofooba", 12}, 161 {"fofofofofoofofoffofoobarfoo", "foffofoobar", 13}, 162 {"fofofofofofofoffofoobarfoo", "foffofoobar", 12}, 163 {"fofofofofoofofoffofoobarfoo", "foffofoobarf", 13}, 164 {"fofofofofofofoffofoobarfoo", "foffofoobarf", 12}, 165 {"fofofofofoofofoffofoobarfoo", "foffofoobarfo", 13}, 166 {"fofofofofofofoffofoobarfoo", "foffofoobarfo", 12}, 167 {"fofofofofoofofoffofoobarfoo", "foffofoobarfoo", 13}, 168 {"fofofofofofofoffofoobarfoo", "foffofoobarfoo", 12}, 169 {"fofofofofoofofoffofoobarfoo", "ofoffofoobarfoo", 12}, 170 {"fofofofofofofoffofoobarfoo", "ofoffofoobarfoo", 11}, 171 {"fofofofofoofofoffofoobarfoo", "fofoffofoobarfoo", 11}, 172 {"fofofofofofofoffofoobarfoo", "fofoffofoobarfoo", 10}, 173 {"fofofofofoofofoffofoobarfoo", "foobars", -1}, 174 {"foofyfoobarfoobar", "y", 4}, 175 {"oooooooooooooooooooooo", "r", -1}, 176 {"oxoxoxoxoxoxoxoxoxoxoxoy", "oy", 22}, 177 {"oxoxoxoxoxoxoxoxoxoxoxox", "oy", -1}, 178 // test fallback to Rabin-Karp. 179 {"000000000000000000000000000000000000000000000000000000000000000000000001", "0000000000000000000000000000000000000000000000000000000000000000001", 5}, 180 } 181 182 var lastIndexTests = []BinOpTest{ 183 {"", "", 0}, 184 {"", "a", -1}, 185 {"", "foo", -1}, 186 {"fo", "foo", -1}, 187 {"foo", "foo", 0}, 188 {"foo", "f", 0}, 189 {"oofofoofooo", "f", 7}, 190 {"oofofoofooo", "foo", 7}, 191 {"barfoobarfoo", "foo", 9}, 192 {"foo", "", 3}, 193 {"foo", "o", 2}, 194 {"abcABCabc", "A", 3}, 195 {"abcABCabc", "a", 6}, 196 } 197 198 var indexAnyTests = []BinOpTest{ 199 {"", "", -1}, 200 {"", "a", -1}, 201 {"", "abc", -1}, 202 {"a", "", -1}, 203 {"a", "a", 0}, 204 {"\x80", "\xffb", 0}, 205 {"aaa", "a", 0}, 206 {"abc", "xyz", -1}, 207 {"abc", "xcz", 2}, 208 {"ab☺c", "x☺yz", 2}, 209 {"a☺b☻c☹d", "cx", len("a☺b☻")}, 210 {"a☺b☻c☹d", "uvw☻xyz", len("a☺b")}, 211 {"aRegExp*", ".(|)*+?^$[]", 7}, 212 {dots + dots + dots, " ", -1}, 213 {"012abcba210", "\xffb", 4}, 214 {"012\x80bcb\x80210", "\xffb", 3}, 215 {"0123456\xcf\x80abc", "\xcfb\x80", 10}, 216 } 217 218 var lastIndexAnyTests = []BinOpTest{ 219 {"", "", -1}, 220 {"", "a", -1}, 221 {"", "abc", -1}, 222 {"a", "", -1}, 223 {"a", "a", 0}, 224 {"\x80", "\xffb", 0}, 225 {"aaa", "a", 2}, 226 {"abc", "xyz", -1}, 227 {"abc", "ab", 1}, 228 {"ab☺c", "x☺yz", 2}, 229 {"a☺b☻c☹d", "cx", len("a☺b☻")}, 230 {"a☺b☻c☹d", "uvw☻xyz", len("a☺b")}, 231 {"a.RegExp*", ".(|)*+?^$[]", 8}, 232 {dots + dots + dots, " ", -1}, 233 {"012abcba210", "\xffb", 6}, 234 {"012\x80bcb\x80210", "\xffb", 7}, 235 {"0123456\xcf\x80abc", "\xcfb\x80", 10}, 236 } 237 238 // Execute f on each test case. funcName should be the name of f; it's used 239 // in failure reports. 240 func runIndexTests(t *testing.T, f func(s, sep []byte) int, funcName string, testCases []BinOpTest) { 241 for _, test := range testCases { 242 a := []byte(test.a) 243 b := []byte(test.b) 244 actual := f(a, b) 245 if actual != test.i { 246 t.Errorf("%s(%q,%q) = %v; want %v", funcName, a, b, actual, test.i) 247 } 248 } 249 var allocTests = []struct { 250 a []byte 251 b []byte 252 i int 253 }{ 254 // case for function Index. 255 {[]byte("000000000000000000000000000000000000000000000000000000000000000000000001"), []byte("0000000000000000000000000000000000000000000000000000000000000000001"), 5}, 256 // case for function LastIndex. 257 {[]byte("000000000000000000000000000000000000000000000000000000000000000010000"), []byte("00000000000000000000000000000000000000000000000000000000000001"), 3}, 258 } 259 allocs := testing.AllocsPerRun(100, func() { 260 if i := Index(allocTests[1].a, allocTests[1].b); i != allocTests[1].i { 261 t.Errorf("Index([]byte(%q), []byte(%q)) = %v; want %v", allocTests[1].a, allocTests[1].b, i, allocTests[1].i) 262 } 263 if i := LastIndex(allocTests[0].a, allocTests[0].b); i != allocTests[0].i { 264 t.Errorf("LastIndex([]byte(%q), []byte(%q)) = %v; want %v", allocTests[0].a, allocTests[0].b, i, allocTests[0].i) 265 } 266 }) 267 if allocs != 0 { 268 t.Errorf("expected no allocations, got %f", allocs) 269 } 270 } 271 272 func runIndexAnyTests(t *testing.T, f func(s []byte, chars string) int, funcName string, testCases []BinOpTest) { 273 for _, test := range testCases { 274 a := []byte(test.a) 275 actual := f(a, test.b) 276 if actual != test.i { 277 t.Errorf("%s(%q,%q) = %v; want %v", funcName, a, test.b, actual, test.i) 278 } 279 } 280 } 281 282 func TestIndex(t *testing.T) { runIndexTests(t, Index, "Index", indexTests) } 283 func TestLastIndex(t *testing.T) { runIndexTests(t, LastIndex, "LastIndex", lastIndexTests) } 284 func TestIndexAny(t *testing.T) { runIndexAnyTests(t, IndexAny, "IndexAny", indexAnyTests) } 285 func TestLastIndexAny(t *testing.T) { 286 runIndexAnyTests(t, LastIndexAny, "LastIndexAny", lastIndexAnyTests) 287 } 288 289 func TestIndexByte(t *testing.T) { 290 for _, tt := range indexTests { 291 if len(tt.b) != 1 { 292 continue 293 } 294 a := []byte(tt.a) 295 b := tt.b[0] 296 pos := IndexByte(a, b) 297 if pos != tt.i { 298 t.Errorf(`IndexByte(%q, '%c') = %v`, tt.a, b, pos) 299 } 300 posp := IndexBytePortable(a, b) 301 if posp != tt.i { 302 t.Errorf(`indexBytePortable(%q, '%c') = %v`, tt.a, b, posp) 303 } 304 } 305 } 306 307 func TestLastIndexByte(t *testing.T) { 308 testCases := []BinOpTest{ 309 {"", "q", -1}, 310 {"abcdef", "q", -1}, 311 {"abcdefabcdef", "a", len("abcdef")}, // something in the middle 312 {"abcdefabcdef", "f", len("abcdefabcde")}, // last byte 313 {"zabcdefabcdef", "z", 0}, // first byte 314 {"a☺b☻c☹d", "b", len("a☺")}, // non-ascii 315 } 316 for _, test := range testCases { 317 actual := LastIndexByte([]byte(test.a), test.b[0]) 318 if actual != test.i { 319 t.Errorf("LastIndexByte(%q,%c) = %v; want %v", test.a, test.b[0], actual, test.i) 320 } 321 } 322 } 323 324 // test a larger buffer with different sizes and alignments 325 func TestIndexByteBig(t *testing.T) { 326 var n = 1024 327 if testing.Short() { 328 n = 128 329 } 330 b := make([]byte, n) 331 for i := 0; i < n; i++ { 332 // different start alignments 333 b1 := b[i:] 334 for j := 0; j < len(b1); j++ { 335 b1[j] = 'x' 336 pos := IndexByte(b1, 'x') 337 if pos != j { 338 t.Errorf("IndexByte(%q, 'x') = %v", b1, pos) 339 } 340 b1[j] = 0 341 pos = IndexByte(b1, 'x') 342 if pos != -1 { 343 t.Errorf("IndexByte(%q, 'x') = %v", b1, pos) 344 } 345 } 346 // different end alignments 347 b1 = b[:i] 348 for j := 0; j < len(b1); j++ { 349 b1[j] = 'x' 350 pos := IndexByte(b1, 'x') 351 if pos != j { 352 t.Errorf("IndexByte(%q, 'x') = %v", b1, pos) 353 } 354 b1[j] = 0 355 pos = IndexByte(b1, 'x') 356 if pos != -1 { 357 t.Errorf("IndexByte(%q, 'x') = %v", b1, pos) 358 } 359 } 360 // different start and end alignments 361 b1 = b[i/2 : n-(i+1)/2] 362 for j := 0; j < len(b1); j++ { 363 b1[j] = 'x' 364 pos := IndexByte(b1, 'x') 365 if pos != j { 366 t.Errorf("IndexByte(%q, 'x') = %v", b1, pos) 367 } 368 b1[j] = 0 369 pos = IndexByte(b1, 'x') 370 if pos != -1 { 371 t.Errorf("IndexByte(%q, 'x') = %v", b1, pos) 372 } 373 } 374 } 375 } 376 377 // test a small index across all page offsets 378 func TestIndexByteSmall(t *testing.T) { 379 b := make([]byte, 5015) // bigger than a page 380 // Make sure we find the correct byte even when straddling a page. 381 for i := 0; i <= len(b)-15; i++ { 382 for j := 0; j < 15; j++ { 383 b[i+j] = byte(100 + j) 384 } 385 for j := 0; j < 15; j++ { 386 p := IndexByte(b[i:i+15], byte(100+j)) 387 if p != j { 388 t.Errorf("IndexByte(%q, %d) = %d", b[i:i+15], 100+j, p) 389 } 390 } 391 for j := 0; j < 15; j++ { 392 b[i+j] = 0 393 } 394 } 395 // Make sure matches outside the slice never trigger. 396 for i := 0; i <= len(b)-15; i++ { 397 for j := 0; j < 15; j++ { 398 b[i+j] = 1 399 } 400 for j := 0; j < 15; j++ { 401 p := IndexByte(b[i:i+15], byte(0)) 402 if p != -1 { 403 t.Errorf("IndexByte(%q, %d) = %d", b[i:i+15], 0, p) 404 } 405 } 406 for j := 0; j < 15; j++ { 407 b[i+j] = 0 408 } 409 } 410 } 411 412 func TestIndexRune(t *testing.T) { 413 tests := []struct { 414 in string 415 rune rune 416 want int 417 }{ 418 {"", 'a', -1}, 419 {"", '☺', -1}, 420 {"foo", '☹', -1}, 421 {"foo", 'o', 1}, 422 {"foo☺bar", '☺', 3}, 423 {"foo☺☻☹bar", '☹', 9}, 424 {"a A x", 'A', 2}, 425 {"some_text=some_value", '=', 9}, 426 {"☺a", 'a', 3}, 427 {"a☻☺b", '☺', 4}, 428 429 // RuneError should match any invalid UTF-8 byte sequence. 430 {"�", '�', 0}, 431 {"\xff", '�', 0}, 432 {"☻x�", '�', len("☻x")}, 433 {"☻x\xe2\x98", '�', len("☻x")}, 434 {"☻x\xe2\x98�", '�', len("☻x")}, 435 {"☻x\xe2\x98x", '�', len("☻x")}, 436 437 // Invalid rune values should never match. 438 {"a☺b☻c☹d\xe2\x98�\xff�\xed\xa0\x80", -1, -1}, 439 {"a☺b☻c☹d\xe2\x98�\xff�\xed\xa0\x80", 0xD800, -1}, // Surrogate pair 440 {"a☺b☻c☹d\xe2\x98�\xff�\xed\xa0\x80", utf8.MaxRune + 1, -1}, 441 } 442 for _, tt := range tests { 443 if got := IndexRune([]byte(tt.in), tt.rune); got != tt.want { 444 t.Errorf("IndexRune(%q, %d) = %v; want %v", tt.in, tt.rune, got, tt.want) 445 } 446 } 447 448 haystack := []byte("test世界") 449 allocs := testing.AllocsPerRun(1000, func() { 450 if i := IndexRune(haystack, 's'); i != 2 { 451 t.Fatalf("'s' at %d; want 2", i) 452 } 453 if i := IndexRune(haystack, '世'); i != 4 { 454 t.Fatalf("'世' at %d; want 4", i) 455 } 456 }) 457 if allocs != 0 { 458 t.Errorf("expected no allocations, got %f", allocs) 459 } 460 } 461 462 // test count of a single byte across page offsets 463 func TestCountByte(t *testing.T) { 464 b := make([]byte, 5015) // bigger than a page 465 windows := []int{1, 2, 3, 4, 15, 16, 17, 31, 32, 33, 63, 64, 65, 128} 466 testCountWindow := func(i, window int) { 467 for j := 0; j < window; j++ { 468 b[i+j] = byte(100) 469 p := Count(b[i:i+window], []byte{100}) 470 if p != j+1 { 471 t.Errorf("TestCountByte.Count(%q, 100) = %d", b[i:i+window], p) 472 } 473 } 474 } 475 476 maxWnd := windows[len(windows)-1] 477 478 for i := 0; i <= 2*maxWnd; i++ { 479 for _, window := range windows { 480 if window > len(b[i:]) { 481 window = len(b[i:]) 482 } 483 testCountWindow(i, window) 484 for j := 0; j < window; j++ { 485 b[i+j] = byte(0) 486 } 487 } 488 } 489 for i := 4096 - (maxWnd + 1); i < len(b); i++ { 490 for _, window := range windows { 491 if window > len(b[i:]) { 492 window = len(b[i:]) 493 } 494 testCountWindow(i, window) 495 for j := 0; j < window; j++ { 496 b[i+j] = byte(0) 497 } 498 } 499 } 500 } 501 502 // Make sure we don't count bytes outside our window 503 func TestCountByteNoMatch(t *testing.T) { 504 b := make([]byte, 5015) 505 windows := []int{1, 2, 3, 4, 15, 16, 17, 31, 32, 33, 63, 64, 65, 128} 506 for i := 0; i <= len(b); i++ { 507 for _, window := range windows { 508 if window > len(b[i:]) { 509 window = len(b[i:]) 510 } 511 // Fill the window with non-match 512 for j := 0; j < window; j++ { 513 b[i+j] = byte(100) 514 } 515 // Try to find something that doesn't exist 516 p := Count(b[i:i+window], []byte{0}) 517 if p != 0 { 518 t.Errorf("TestCountByteNoMatch(%q, 0) = %d", b[i:i+window], p) 519 } 520 for j := 0; j < window; j++ { 521 b[i+j] = byte(0) 522 } 523 } 524 } 525 } 526 527 var bmbuf []byte 528 529 func valName(x int) string { 530 if s := x >> 20; s<<20 == x { 531 return fmt.Sprintf("%dM", s) 532 } 533 if s := x >> 10; s<<10 == x { 534 return fmt.Sprintf("%dK", s) 535 } 536 return fmt.Sprint(x) 537 } 538 539 func benchBytes(b *testing.B, sizes []int, f func(b *testing.B, n int)) { 540 for _, n := range sizes { 541 if isRaceBuilder && n > 4<<10 { 542 continue 543 } 544 b.Run(valName(n), func(b *testing.B) { 545 if len(bmbuf) < n { 546 bmbuf = make([]byte, n) 547 } 548 b.SetBytes(int64(n)) 549 f(b, n) 550 }) 551 } 552 } 553 554 var indexSizes = []int{10, 32, 4 << 10, 4 << 20, 64 << 20} 555 556 var isRaceBuilder = strings.HasSuffix(testenv.Builder(), "-race") 557 558 func BenchmarkIndexByte(b *testing.B) { 559 benchBytes(b, indexSizes, bmIndexByte(IndexByte)) 560 } 561 562 func BenchmarkIndexBytePortable(b *testing.B) { 563 benchBytes(b, indexSizes, bmIndexByte(IndexBytePortable)) 564 } 565 566 func bmIndexByte(index func([]byte, byte) int) func(b *testing.B, n int) { 567 return func(b *testing.B, n int) { 568 buf := bmbuf[0:n] 569 buf[n-1] = 'x' 570 for i := 0; i < b.N; i++ { 571 j := index(buf, 'x') 572 if j != n-1 { 573 b.Fatal("bad index", j) 574 } 575 } 576 buf[n-1] = '\x00' 577 } 578 } 579 580 func BenchmarkIndexRune(b *testing.B) { 581 benchBytes(b, indexSizes, bmIndexRune(IndexRune)) 582 } 583 584 func BenchmarkIndexRuneASCII(b *testing.B) { 585 benchBytes(b, indexSizes, bmIndexRuneASCII(IndexRune)) 586 } 587 588 func bmIndexRuneASCII(index func([]byte, rune) int) func(b *testing.B, n int) { 589 return func(b *testing.B, n int) { 590 buf := bmbuf[0:n] 591 buf[n-1] = 'x' 592 for i := 0; i < b.N; i++ { 593 j := index(buf, 'x') 594 if j != n-1 { 595 b.Fatal("bad index", j) 596 } 597 } 598 buf[n-1] = '\x00' 599 } 600 } 601 602 func bmIndexRune(index func([]byte, rune) int) func(b *testing.B, n int) { 603 return func(b *testing.B, n int) { 604 buf := bmbuf[0:n] 605 utf8.EncodeRune(buf[n-3:], '世') 606 for i := 0; i < b.N; i++ { 607 j := index(buf, '世') 608 if j != n-3 { 609 b.Fatal("bad index", j) 610 } 611 } 612 buf[n-3] = '\x00' 613 buf[n-2] = '\x00' 614 buf[n-1] = '\x00' 615 } 616 } 617 618 func BenchmarkEqual(b *testing.B) { 619 b.Run("0", func(b *testing.B) { 620 var buf [4]byte 621 buf1 := buf[0:0] 622 buf2 := buf[1:1] 623 for i := 0; i < b.N; i++ { 624 eq := Equal(buf1, buf2) 625 if !eq { 626 b.Fatal("bad equal") 627 } 628 } 629 }) 630 631 sizes := []int{1, 6, 9, 15, 16, 20, 32, 4 << 10, 4 << 20, 64 << 20} 632 benchBytes(b, sizes, bmEqual(Equal)) 633 } 634 635 func bmEqual(equal func([]byte, []byte) bool) func(b *testing.B, n int) { 636 return func(b *testing.B, n int) { 637 if len(bmbuf) < 2*n { 638 bmbuf = make([]byte, 2*n) 639 } 640 buf1 := bmbuf[0:n] 641 buf2 := bmbuf[n : 2*n] 642 buf1[n-1] = 'x' 643 buf2[n-1] = 'x' 644 for i := 0; i < b.N; i++ { 645 eq := equal(buf1, buf2) 646 if !eq { 647 b.Fatal("bad equal") 648 } 649 } 650 buf1[n-1] = '\x00' 651 buf2[n-1] = '\x00' 652 } 653 } 654 655 func BenchmarkIndex(b *testing.B) { 656 benchBytes(b, indexSizes, func(b *testing.B, n int) { 657 buf := bmbuf[0:n] 658 buf[n-1] = 'x' 659 for i := 0; i < b.N; i++ { 660 j := Index(buf, buf[n-7:]) 661 if j != n-7 { 662 b.Fatal("bad index", j) 663 } 664 } 665 buf[n-1] = '\x00' 666 }) 667 } 668 669 func BenchmarkIndexEasy(b *testing.B) { 670 benchBytes(b, indexSizes, func(b *testing.B, n int) { 671 buf := bmbuf[0:n] 672 buf[n-1] = 'x' 673 buf[n-7] = 'x' 674 for i := 0; i < b.N; i++ { 675 j := Index(buf, buf[n-7:]) 676 if j != n-7 { 677 b.Fatal("bad index", j) 678 } 679 } 680 buf[n-1] = '\x00' 681 buf[n-7] = '\x00' 682 }) 683 } 684 685 func BenchmarkCount(b *testing.B) { 686 benchBytes(b, indexSizes, func(b *testing.B, n int) { 687 buf := bmbuf[0:n] 688 buf[n-1] = 'x' 689 for i := 0; i < b.N; i++ { 690 j := Count(buf, buf[n-7:]) 691 if j != 1 { 692 b.Fatal("bad count", j) 693 } 694 } 695 buf[n-1] = '\x00' 696 }) 697 } 698 699 func BenchmarkCountEasy(b *testing.B) { 700 benchBytes(b, indexSizes, func(b *testing.B, n int) { 701 buf := bmbuf[0:n] 702 buf[n-1] = 'x' 703 buf[n-7] = 'x' 704 for i := 0; i < b.N; i++ { 705 j := Count(buf, buf[n-7:]) 706 if j != 1 { 707 b.Fatal("bad count", j) 708 } 709 } 710 buf[n-1] = '\x00' 711 buf[n-7] = '\x00' 712 }) 713 } 714 715 func BenchmarkCountSingle(b *testing.B) { 716 benchBytes(b, indexSizes, func(b *testing.B, n int) { 717 buf := bmbuf[0:n] 718 step := 8 719 for i := 0; i < len(buf); i += step { 720 buf[i] = 1 721 } 722 expect := (len(buf) + (step - 1)) / step 723 for i := 0; i < b.N; i++ { 724 j := Count(buf, []byte{1}) 725 if j != expect { 726 b.Fatal("bad count", j, expect) 727 } 728 } 729 for i := 0; i < len(buf); i++ { 730 buf[i] = 0 731 } 732 }) 733 } 734 735 type SplitTest struct { 736 s string 737 sep string 738 n int 739 a []string 740 } 741 742 var splittests = []SplitTest{ 743 {"", "", -1, []string{}}, 744 {abcd, "a", 0, nil}, 745 {abcd, "", 2, []string{"a", "bcd"}}, 746 {abcd, "a", -1, []string{"", "bcd"}}, 747 {abcd, "z", -1, []string{"abcd"}}, 748 {abcd, "", -1, []string{"a", "b", "c", "d"}}, 749 {commas, ",", -1, []string{"1", "2", "3", "4"}}, 750 {dots, "...", -1, []string{"1", ".2", ".3", ".4"}}, 751 {faces, "☹", -1, []string{"☺☻", ""}}, 752 {faces, "~", -1, []string{faces}}, 753 {faces, "", -1, []string{"☺", "☻", "☹"}}, 754 {"1 2 3 4", " ", 3, []string{"1", "2", "3 4"}}, 755 {"1 2", " ", 3, []string{"1", "2"}}, 756 {"123", "", 2, []string{"1", "23"}}, 757 {"123", "", 17, []string{"1", "2", "3"}}, 758 {"bT", "T", math.MaxInt / 4, []string{"b", ""}}, 759 {"\xff-\xff", "", -1, []string{"\xff", "-", "\xff"}}, 760 {"\xff-\xff", "-", -1, []string{"\xff", "\xff"}}, 761 } 762 763 func TestSplit(t *testing.T) { 764 for _, tt := range splittests { 765 a := SplitN([]byte(tt.s), []byte(tt.sep), tt.n) 766 767 // Appending to the results should not change future results. 768 var x []byte 769 for _, v := range a { 770 x = append(v, 'z') 771 } 772 773 result := sliceOfString(a) 774 if !eq(result, tt.a) { 775 t.Errorf(`Split(%q, %q, %d) = %v; want %v`, tt.s, tt.sep, tt.n, result, tt.a) 776 continue 777 } 778 if tt.n == 0 || len(a) == 0 { 779 continue 780 } 781 782 if want := tt.a[len(tt.a)-1] + "z"; string(x) != want { 783 t.Errorf("last appended result was %s; want %s", x, want) 784 } 785 786 s := Join(a, []byte(tt.sep)) 787 if string(s) != tt.s { 788 t.Errorf(`Join(Split(%q, %q, %d), %q) = %q`, tt.s, tt.sep, tt.n, tt.sep, s) 789 } 790 if tt.n < 0 { 791 b := Split([]byte(tt.s), []byte(tt.sep)) 792 if !reflect.DeepEqual(a, b) { 793 t.Errorf("Split disagrees withSplitN(%q, %q, %d) = %v; want %v", tt.s, tt.sep, tt.n, b, a) 794 } 795 } 796 if len(a) > 0 { 797 in, out := a[0], s 798 if cap(in) == cap(out) && &in[:1][0] == &out[:1][0] { 799 t.Errorf("Join(%#v, %q) didn't copy", a, tt.sep) 800 } 801 } 802 } 803 } 804 805 var splitaftertests = []SplitTest{ 806 {abcd, "a", -1, []string{"a", "bcd"}}, 807 {abcd, "z", -1, []string{"abcd"}}, 808 {abcd, "", -1, []string{"a", "b", "c", "d"}}, 809 {commas, ",", -1, []string{"1,", "2,", "3,", "4"}}, 810 {dots, "...", -1, []string{"1...", ".2...", ".3...", ".4"}}, 811 {faces, "☹", -1, []string{"☺☻☹", ""}}, 812 {faces, "~", -1, []string{faces}}, 813 {faces, "", -1, []string{"☺", "☻", "☹"}}, 814 {"1 2 3 4", " ", 3, []string{"1 ", "2 ", "3 4"}}, 815 {"1 2 3", " ", 3, []string{"1 ", "2 ", "3"}}, 816 {"1 2", " ", 3, []string{"1 ", "2"}}, 817 {"123", "", 2, []string{"1", "23"}}, 818 {"123", "", 17, []string{"1", "2", "3"}}, 819 } 820 821 func TestSplitAfter(t *testing.T) { 822 for _, tt := range splitaftertests { 823 a := SplitAfterN([]byte(tt.s), []byte(tt.sep), tt.n) 824 825 // Appending to the results should not change future results. 826 var x []byte 827 for _, v := range a { 828 x = append(v, 'z') 829 } 830 831 result := sliceOfString(a) 832 if !eq(result, tt.a) { 833 t.Errorf(`Split(%q, %q, %d) = %v; want %v`, tt.s, tt.sep, tt.n, result, tt.a) 834 continue 835 } 836 837 if want := tt.a[len(tt.a)-1] + "z"; string(x) != want { 838 t.Errorf("last appended result was %s; want %s", x, want) 839 } 840 841 s := Join(a, nil) 842 if string(s) != tt.s { 843 t.Errorf(`Join(Split(%q, %q, %d), %q) = %q`, tt.s, tt.sep, tt.n, tt.sep, s) 844 } 845 if tt.n < 0 { 846 b := SplitAfter([]byte(tt.s), []byte(tt.sep)) 847 if !reflect.DeepEqual(a, b) { 848 t.Errorf("SplitAfter disagrees withSplitAfterN(%q, %q, %d) = %v; want %v", tt.s, tt.sep, tt.n, b, a) 849 } 850 } 851 } 852 } 853 854 type FieldsTest struct { 855 s string 856 a []string 857 } 858 859 var fieldstests = []FieldsTest{ 860 {"", []string{}}, 861 {" ", []string{}}, 862 {" \t ", []string{}}, 863 {" abc ", []string{"abc"}}, 864 {"1 2 3 4", []string{"1", "2", "3", "4"}}, 865 {"1 2 3 4", []string{"1", "2", "3", "4"}}, 866 {"1\t\t2\t\t3\t4", []string{"1", "2", "3", "4"}}, 867 {"1\u20002\u20013\u20024", []string{"1", "2", "3", "4"}}, 868 {"\u2000\u2001\u2002", []string{}}, 869 {"\n™\t™\n", []string{"™", "™"}}, 870 {faces, []string{faces}}, 871 } 872 873 func TestFields(t *testing.T) { 874 for _, tt := range fieldstests { 875 b := []byte(tt.s) 876 a := Fields(b) 877 878 // Appending to the results should not change future results. 879 var x []byte 880 for _, v := range a { 881 x = append(v, 'z') 882 } 883 884 result := sliceOfString(a) 885 if !eq(result, tt.a) { 886 t.Errorf("Fields(%q) = %v; want %v", tt.s, a, tt.a) 887 continue 888 } 889 890 if string(b) != tt.s { 891 t.Errorf("slice changed to %s; want %s", string(b), tt.s) 892 } 893 if len(tt.a) > 0 { 894 if want := tt.a[len(tt.a)-1] + "z"; string(x) != want { 895 t.Errorf("last appended result was %s; want %s", x, want) 896 } 897 } 898 } 899 } 900 901 func TestFieldsFunc(t *testing.T) { 902 for _, tt := range fieldstests { 903 a := FieldsFunc([]byte(tt.s), unicode.IsSpace) 904 result := sliceOfString(a) 905 if !eq(result, tt.a) { 906 t.Errorf("FieldsFunc(%q, unicode.IsSpace) = %v; want %v", tt.s, a, tt.a) 907 continue 908 } 909 } 910 pred := func(c rune) bool { return c == 'X' } 911 var fieldsFuncTests = []FieldsTest{ 912 {"", []string{}}, 913 {"XX", []string{}}, 914 {"XXhiXXX", []string{"hi"}}, 915 {"aXXbXXXcX", []string{"a", "b", "c"}}, 916 } 917 for _, tt := range fieldsFuncTests { 918 b := []byte(tt.s) 919 a := FieldsFunc(b, pred) 920 921 // Appending to the results should not change future results. 922 var x []byte 923 for _, v := range a { 924 x = append(v, 'z') 925 } 926 927 result := sliceOfString(a) 928 if !eq(result, tt.a) { 929 t.Errorf("FieldsFunc(%q) = %v, want %v", tt.s, a, tt.a) 930 } 931 932 if string(b) != tt.s { 933 t.Errorf("slice changed to %s; want %s", b, tt.s) 934 } 935 if len(tt.a) > 0 { 936 if want := tt.a[len(tt.a)-1] + "z"; string(x) != want { 937 t.Errorf("last appended result was %s; want %s", x, want) 938 } 939 } 940 } 941 } 942 943 // Test case for any function which accepts and returns a byte slice. 944 // For ease of creation, we write the input byte slice as a string. 945 type StringTest struct { 946 in string 947 out []byte 948 } 949 950 var upperTests = []StringTest{ 951 {"", []byte("")}, 952 {"ONLYUPPER", []byte("ONLYUPPER")}, 953 {"abc", []byte("ABC")}, 954 {"AbC123", []byte("ABC123")}, 955 {"azAZ09_", []byte("AZAZ09_")}, 956 {"longStrinGwitHmixofsmaLLandcAps", []byte("LONGSTRINGWITHMIXOFSMALLANDCAPS")}, 957 {"long\u0250string\u0250with\u0250nonascii\u2C6Fchars", []byte("LONG\u2C6FSTRING\u2C6FWITH\u2C6FNONASCII\u2C6FCHARS")}, 958 {"\u0250\u0250\u0250\u0250\u0250", []byte("\u2C6F\u2C6F\u2C6F\u2C6F\u2C6F")}, // grows one byte per char 959 {"a\u0080\U0010FFFF", []byte("A\u0080\U0010FFFF")}, // test utf8.RuneSelf and utf8.MaxRune 960 } 961 962 var lowerTests = []StringTest{ 963 {"", []byte("")}, 964 {"abc", []byte("abc")}, 965 {"AbC123", []byte("abc123")}, 966 {"azAZ09_", []byte("azaz09_")}, 967 {"longStrinGwitHmixofsmaLLandcAps", []byte("longstringwithmixofsmallandcaps")}, 968 {"LONG\u2C6FSTRING\u2C6FWITH\u2C6FNONASCII\u2C6FCHARS", []byte("long\u0250string\u0250with\u0250nonascii\u0250chars")}, 969 {"\u2C6D\u2C6D\u2C6D\u2C6D\u2C6D", []byte("\u0251\u0251\u0251\u0251\u0251")}, // shrinks one byte per char 970 {"A\u0080\U0010FFFF", []byte("a\u0080\U0010FFFF")}, // test utf8.RuneSelf and utf8.MaxRune 971 } 972 973 const space = "\t\v\r\f\n\u0085\u00a0\u2000\u3000" 974 975 var trimSpaceTests = []StringTest{ 976 {"", nil}, 977 {" a", []byte("a")}, 978 {"b ", []byte("b")}, 979 {"abc", []byte("abc")}, 980 {space + "abc" + space, []byte("abc")}, 981 {" ", nil}, 982 {"\u3000 ", nil}, 983 {" \u3000", nil}, 984 {" \t\r\n \t\t\r\r\n\n ", nil}, 985 {" \t\r\n x\t\t\r\r\n\n ", []byte("x")}, 986 {" \u2000\t\r\n x\t\t\r\r\ny\n \u3000", []byte("x\t\t\r\r\ny")}, 987 {"1 \t\r\n2", []byte("1 \t\r\n2")}, 988 {" x\x80", []byte("x\x80")}, 989 {" x\xc0", []byte("x\xc0")}, 990 {"x \xc0\xc0 ", []byte("x \xc0\xc0")}, 991 {"x \xc0", []byte("x \xc0")}, 992 {"x \xc0 ", []byte("x \xc0")}, 993 {"x \xc0\xc0 ", []byte("x \xc0\xc0")}, 994 {"x ☺\xc0\xc0 ", []byte("x ☺\xc0\xc0")}, 995 {"x ☺ ", []byte("x ☺")}, 996 } 997 998 // Execute f on each test case. funcName should be the name of f; it's used 999 // in failure reports. 1000 func runStringTests(t *testing.T, f func([]byte) []byte, funcName string, testCases []StringTest) { 1001 for _, tc := range testCases { 1002 actual := f([]byte(tc.in)) 1003 if actual == nil && tc.out != nil { 1004 t.Errorf("%s(%q) = nil; want %q", funcName, tc.in, tc.out) 1005 } 1006 if actual != nil && tc.out == nil { 1007 t.Errorf("%s(%q) = %q; want nil", funcName, tc.in, actual) 1008 } 1009 if !Equal(actual, tc.out) { 1010 t.Errorf("%s(%q) = %q; want %q", funcName, tc.in, actual, tc.out) 1011 } 1012 } 1013 } 1014 1015 func tenRunes(r rune) string { 1016 runes := make([]rune, 10) 1017 for i := range runes { 1018 runes[i] = r 1019 } 1020 return string(runes) 1021 } 1022 1023 // User-defined self-inverse mapping function 1024 func rot13(r rune) rune { 1025 const step = 13 1026 if r >= 'a' && r <= 'z' { 1027 return ((r - 'a' + step) % 26) + 'a' 1028 } 1029 if r >= 'A' && r <= 'Z' { 1030 return ((r - 'A' + step) % 26) + 'A' 1031 } 1032 return r 1033 } 1034 1035 func TestMap(t *testing.T) { 1036 // Run a couple of awful growth/shrinkage tests 1037 a := tenRunes('a') 1038 1039 // 1. Grow. This triggers two reallocations in Map. 1040 maxRune := func(r rune) rune { return unicode.MaxRune } 1041 m := Map(maxRune, []byte(a)) 1042 expect := tenRunes(unicode.MaxRune) 1043 if string(m) != expect { 1044 t.Errorf("growing: expected %q got %q", expect, m) 1045 } 1046 1047 // 2. Shrink 1048 minRune := func(r rune) rune { return 'a' } 1049 m = Map(minRune, []byte(tenRunes(unicode.MaxRune))) 1050 expect = a 1051 if string(m) != expect { 1052 t.Errorf("shrinking: expected %q got %q", expect, m) 1053 } 1054 1055 // 3. Rot13 1056 m = Map(rot13, []byte("a to zed")) 1057 expect = "n gb mrq" 1058 if string(m) != expect { 1059 t.Errorf("rot13: expected %q got %q", expect, m) 1060 } 1061 1062 // 4. Rot13^2 1063 m = Map(rot13, Map(rot13, []byte("a to zed"))) 1064 expect = "a to zed" 1065 if string(m) != expect { 1066 t.Errorf("rot13: expected %q got %q", expect, m) 1067 } 1068 1069 // 5. Drop 1070 dropNotLatin := func(r rune) rune { 1071 if unicode.Is(unicode.Latin, r) { 1072 return r 1073 } 1074 return -1 1075 } 1076 m = Map(dropNotLatin, []byte("Hello, 세계")) 1077 expect = "Hello" 1078 if string(m) != expect { 1079 t.Errorf("drop: expected %q got %q", expect, m) 1080 } 1081 1082 // 6. Invalid rune 1083 invalidRune := func(r rune) rune { 1084 return utf8.MaxRune + 1 1085 } 1086 m = Map(invalidRune, []byte("x")) 1087 expect = "\uFFFD" 1088 if string(m) != expect { 1089 t.Errorf("invalidRune: expected %q got %q", expect, m) 1090 } 1091 } 1092 1093 func TestToUpper(t *testing.T) { runStringTests(t, ToUpper, "ToUpper", upperTests) } 1094 1095 func TestToLower(t *testing.T) { runStringTests(t, ToLower, "ToLower", lowerTests) } 1096 1097 func BenchmarkToUpper(b *testing.B) { 1098 for _, tc := range upperTests { 1099 tin := []byte(tc.in) 1100 b.Run(tc.in, func(b *testing.B) { 1101 for i := 0; i < b.N; i++ { 1102 actual := ToUpper(tin) 1103 if !Equal(actual, tc.out) { 1104 b.Errorf("ToUpper(%q) = %q; want %q", tc.in, actual, tc.out) 1105 } 1106 } 1107 }) 1108 } 1109 } 1110 1111 func BenchmarkToLower(b *testing.B) { 1112 for _, tc := range lowerTests { 1113 tin := []byte(tc.in) 1114 b.Run(tc.in, func(b *testing.B) { 1115 for i := 0; i < b.N; i++ { 1116 actual := ToLower(tin) 1117 if !Equal(actual, tc.out) { 1118 b.Errorf("ToLower(%q) = %q; want %q", tc.in, actual, tc.out) 1119 } 1120 } 1121 }) 1122 } 1123 } 1124 1125 var toValidUTF8Tests = []struct { 1126 in string 1127 repl string 1128 out string 1129 }{ 1130 {"", "\uFFFD", ""}, 1131 {"abc", "\uFFFD", "abc"}, 1132 {"\uFDDD", "\uFFFD", "\uFDDD"}, 1133 {"a\xffb", "\uFFFD", "a\uFFFDb"}, 1134 {"a\xffb\uFFFD", "X", "aXb\uFFFD"}, 1135 {"a☺\xffb☺\xC0\xAFc☺\xff", "", "a☺b☺c☺"}, 1136 {"a☺\xffb☺\xC0\xAFc☺\xff", "日本語", "a☺日本語b☺日本語c☺日本語"}, 1137 {"\xC0\xAF", "\uFFFD", "\uFFFD"}, 1138 {"\xE0\x80\xAF", "\uFFFD", "\uFFFD"}, 1139 {"\xed\xa0\x80", "abc", "abc"}, 1140 {"\xed\xbf\xbf", "\uFFFD", "\uFFFD"}, 1141 {"\xF0\x80\x80\xaf", "☺", "☺"}, 1142 {"\xF8\x80\x80\x80\xAF", "\uFFFD", "\uFFFD"}, 1143 {"\xFC\x80\x80\x80\x80\xAF", "\uFFFD", "\uFFFD"}, 1144 } 1145 1146 func TestToValidUTF8(t *testing.T) { 1147 for _, tc := range toValidUTF8Tests { 1148 got := ToValidUTF8([]byte(tc.in), []byte(tc.repl)) 1149 if !Equal(got, []byte(tc.out)) { 1150 t.Errorf("ToValidUTF8(%q, %q) = %q; want %q", tc.in, tc.repl, got, tc.out) 1151 } 1152 } 1153 } 1154 1155 func TestTrimSpace(t *testing.T) { runStringTests(t, TrimSpace, "TrimSpace", trimSpaceTests) } 1156 1157 type RepeatTest struct { 1158 in, out string 1159 count int 1160 } 1161 1162 var longString = "a" + string(make([]byte, 1<<16)) + "z" 1163 1164 var RepeatTests = []RepeatTest{ 1165 {"", "", 0}, 1166 {"", "", 1}, 1167 {"", "", 2}, 1168 {"-", "", 0}, 1169 {"-", "-", 1}, 1170 {"-", "----------", 10}, 1171 {"abc ", "abc abc abc ", 3}, 1172 // Tests for results over the chunkLimit 1173 {string(rune(0)), string(make([]byte, 1<<16)), 1 << 16}, 1174 {longString, longString + longString, 2}, 1175 } 1176 1177 func TestRepeat(t *testing.T) { 1178 for _, tt := range RepeatTests { 1179 tin := []byte(tt.in) 1180 tout := []byte(tt.out) 1181 a := Repeat(tin, tt.count) 1182 if !Equal(a, tout) { 1183 t.Errorf("Repeat(%q, %d) = %q; want %q", tin, tt.count, a, tout) 1184 continue 1185 } 1186 } 1187 } 1188 1189 func repeat(b []byte, count int) (err error) { 1190 defer func() { 1191 if r := recover(); r != nil { 1192 switch v := r.(type) { 1193 case error: 1194 err = v 1195 default: 1196 err = fmt.Errorf("%s", v) 1197 } 1198 } 1199 }() 1200 1201 Repeat(b, count) 1202 1203 return 1204 } 1205 1206 // See Issue golang.org/issue/16237 1207 func TestRepeatCatchesOverflow(t *testing.T) { 1208 tests := [...]struct { 1209 s string 1210 count int 1211 errStr string 1212 }{ 1213 0: {"--", -2147483647, "negative"}, 1214 1: {"", int(^uint(0) >> 1), ""}, 1215 2: {"-", 10, ""}, 1216 3: {"gopher", 0, ""}, 1217 4: {"-", -1, "negative"}, 1218 5: {"--", -102, "negative"}, 1219 6: {string(make([]byte, 255)), int((^uint(0))/255 + 1), "overflow"}, 1220 } 1221 1222 for i, tt := range tests { 1223 err := repeat([]byte(tt.s), tt.count) 1224 if tt.errStr == "" { 1225 if err != nil { 1226 t.Errorf("#%d panicked %v", i, err) 1227 } 1228 continue 1229 } 1230 1231 if err == nil || !strings.Contains(err.Error(), tt.errStr) { 1232 t.Errorf("#%d expected %q got %q", i, tt.errStr, err) 1233 } 1234 } 1235 } 1236 1237 func runesEqual(a, b []rune) bool { 1238 if len(a) != len(b) { 1239 return false 1240 } 1241 for i, r := range a { 1242 if r != b[i] { 1243 return false 1244 } 1245 } 1246 return true 1247 } 1248 1249 type RunesTest struct { 1250 in string 1251 out []rune 1252 lossy bool 1253 } 1254 1255 var RunesTests = []RunesTest{ 1256 {"", []rune{}, false}, 1257 {" ", []rune{32}, false}, 1258 {"ABC", []rune{65, 66, 67}, false}, 1259 {"abc", []rune{97, 98, 99}, false}, 1260 {"\u65e5\u672c\u8a9e", []rune{26085, 26412, 35486}, false}, 1261 {"ab\x80c", []rune{97, 98, 0xFFFD, 99}, true}, 1262 {"ab\xc0c", []rune{97, 98, 0xFFFD, 99}, true}, 1263 } 1264 1265 func TestRunes(t *testing.T) { 1266 for _, tt := range RunesTests { 1267 tin := []byte(tt.in) 1268 a := Runes(tin) 1269 if !runesEqual(a, tt.out) { 1270 t.Errorf("Runes(%q) = %v; want %v", tin, a, tt.out) 1271 continue 1272 } 1273 if !tt.lossy { 1274 // can only test reassembly if we didn't lose information 1275 s := string(a) 1276 if s != tt.in { 1277 t.Errorf("string(Runes(%q)) = %x; want %x", tin, s, tin) 1278 } 1279 } 1280 } 1281 } 1282 1283 type TrimTest struct { 1284 f string 1285 in, arg, out string 1286 } 1287 1288 var trimTests = []TrimTest{ 1289 {"Trim", "abba", "a", "bb"}, 1290 {"Trim", "abba", "ab", ""}, 1291 {"TrimLeft", "abba", "ab", ""}, 1292 {"TrimRight", "abba", "ab", ""}, 1293 {"TrimLeft", "abba", "a", "bba"}, 1294 {"TrimLeft", "abba", "b", "abba"}, 1295 {"TrimRight", "abba", "a", "abb"}, 1296 {"TrimRight", "abba", "b", "abba"}, 1297 {"Trim", "<tag>", "<>", "tag"}, 1298 {"Trim", "* listitem", " *", "listitem"}, 1299 {"Trim", `"quote"`, `"`, "quote"}, 1300 {"Trim", "\u2C6F\u2C6F\u0250\u0250\u2C6F\u2C6F", "\u2C6F", "\u0250\u0250"}, 1301 {"Trim", "\x80test\xff", "\xff", "test"}, 1302 {"Trim", " Ġ ", " ", "Ġ"}, 1303 {"Trim", " Ġİ0", "0 ", "Ġİ"}, 1304 //empty string tests 1305 {"Trim", "abba", "", "abba"}, 1306 {"Trim", "", "123", ""}, 1307 {"Trim", "", "", ""}, 1308 {"TrimLeft", "abba", "", "abba"}, 1309 {"TrimLeft", "", "123", ""}, 1310 {"TrimLeft", "", "", ""}, 1311 {"TrimRight", "abba", "", "abba"}, 1312 {"TrimRight", "", "123", ""}, 1313 {"TrimRight", "", "", ""}, 1314 {"TrimRight", "☺\xc0", "☺", "☺\xc0"}, 1315 {"TrimPrefix", "aabb", "a", "abb"}, 1316 {"TrimPrefix", "aabb", "b", "aabb"}, 1317 {"TrimSuffix", "aabb", "a", "aabb"}, 1318 {"TrimSuffix", "aabb", "b", "aab"}, 1319 } 1320 1321 type TrimNilTest struct { 1322 f string 1323 in []byte 1324 arg string 1325 out []byte 1326 } 1327 1328 var trimNilTests = []TrimNilTest{ 1329 {"Trim", nil, "", nil}, 1330 {"Trim", []byte{}, "", nil}, 1331 {"Trim", []byte{'a'}, "a", nil}, 1332 {"Trim", []byte{'a', 'a'}, "a", nil}, 1333 {"Trim", []byte{'a'}, "ab", nil}, 1334 {"Trim", []byte{'a', 'b'}, "ab", nil}, 1335 {"Trim", []byte("☺"), "☺", nil}, 1336 {"TrimLeft", nil, "", nil}, 1337 {"TrimLeft", []byte{}, "", nil}, 1338 {"TrimLeft", []byte{'a'}, "a", nil}, 1339 {"TrimLeft", []byte{'a', 'a'}, "a", nil}, 1340 {"TrimLeft", []byte{'a'}, "ab", nil}, 1341 {"TrimLeft", []byte{'a', 'b'}, "ab", nil}, 1342 {"TrimLeft", []byte("☺"), "☺", nil}, 1343 {"TrimRight", nil, "", nil}, 1344 {"TrimRight", []byte{}, "", []byte{}}, 1345 {"TrimRight", []byte{'a'}, "a", []byte{}}, 1346 {"TrimRight", []byte{'a', 'a'}, "a", []byte{}}, 1347 {"TrimRight", []byte{'a'}, "ab", []byte{}}, 1348 {"TrimRight", []byte{'a', 'b'}, "ab", []byte{}}, 1349 {"TrimRight", []byte("☺"), "☺", []byte{}}, 1350 {"TrimPrefix", nil, "", nil}, 1351 {"TrimPrefix", []byte{}, "", []byte{}}, 1352 {"TrimPrefix", []byte{'a'}, "a", []byte{}}, 1353 {"TrimPrefix", []byte("☺"), "☺", []byte{}}, 1354 {"TrimSuffix", nil, "", nil}, 1355 {"TrimSuffix", []byte{}, "", []byte{}}, 1356 {"TrimSuffix", []byte{'a'}, "a", []byte{}}, 1357 {"TrimSuffix", []byte("☺"), "☺", []byte{}}, 1358 } 1359 1360 func TestTrim(t *testing.T) { 1361 toFn := func(name string) (func([]byte, string) []byte, func([]byte, []byte) []byte) { 1362 switch name { 1363 case "Trim": 1364 return Trim, nil 1365 case "TrimLeft": 1366 return TrimLeft, nil 1367 case "TrimRight": 1368 return TrimRight, nil 1369 case "TrimPrefix": 1370 return nil, TrimPrefix 1371 case "TrimSuffix": 1372 return nil, TrimSuffix 1373 default: 1374 t.Errorf("Undefined trim function %s", name) 1375 return nil, nil 1376 } 1377 } 1378 1379 for _, tc := range trimTests { 1380 name := tc.f 1381 f, fb := toFn(name) 1382 if f == nil && fb == nil { 1383 continue 1384 } 1385 var actual string 1386 if f != nil { 1387 actual = string(f([]byte(tc.in), tc.arg)) 1388 } else { 1389 actual = string(fb([]byte(tc.in), []byte(tc.arg))) 1390 } 1391 if actual != tc.out { 1392 t.Errorf("%s(%q, %q) = %q; want %q", name, tc.in, tc.arg, actual, tc.out) 1393 } 1394 } 1395 1396 for _, tc := range trimNilTests { 1397 name := tc.f 1398 f, fb := toFn(name) 1399 if f == nil && fb == nil { 1400 continue 1401 } 1402 var actual []byte 1403 if f != nil { 1404 actual = f(tc.in, tc.arg) 1405 } else { 1406 actual = fb(tc.in, []byte(tc.arg)) 1407 } 1408 report := func(s []byte) string { 1409 if s == nil { 1410 return "nil" 1411 } else { 1412 return fmt.Sprintf("%q", s) 1413 } 1414 } 1415 if len(actual) != 0 { 1416 t.Errorf("%s(%s, %q) returned non-empty value", name, report(tc.in), tc.arg) 1417 } else { 1418 actualNil := actual == nil 1419 outNil := tc.out == nil 1420 if actualNil != outNil { 1421 t.Errorf("%s(%s, %q) got nil %t; want nil %t", name, report(tc.in), tc.arg, actualNil, outNil) 1422 } 1423 } 1424 } 1425 } 1426 1427 type predicate struct { 1428 f func(r rune) bool 1429 name string 1430 } 1431 1432 var isSpace = predicate{unicode.IsSpace, "IsSpace"} 1433 var isDigit = predicate{unicode.IsDigit, "IsDigit"} 1434 var isUpper = predicate{unicode.IsUpper, "IsUpper"} 1435 var isValidRune = predicate{ 1436 func(r rune) bool { 1437 return r != utf8.RuneError 1438 }, 1439 "IsValidRune", 1440 } 1441 1442 type TrimFuncTest struct { 1443 f predicate 1444 in string 1445 trimOut []byte 1446 leftOut []byte 1447 rightOut []byte 1448 } 1449 1450 func not(p predicate) predicate { 1451 return predicate{ 1452 func(r rune) bool { 1453 return !p.f(r) 1454 }, 1455 "not " + p.name, 1456 } 1457 } 1458 1459 var trimFuncTests = []TrimFuncTest{ 1460 {isSpace, space + " hello " + space, 1461 []byte("hello"), 1462 []byte("hello " + space), 1463 []byte(space + " hello")}, 1464 {isDigit, "\u0e50\u0e5212hello34\u0e50\u0e51", 1465 []byte("hello"), 1466 []byte("hello34\u0e50\u0e51"), 1467 []byte("\u0e50\u0e5212hello")}, 1468 {isUpper, "\u2C6F\u2C6F\u2C6F\u2C6FABCDhelloEF\u2C6F\u2C6FGH\u2C6F\u2C6F", 1469 []byte("hello"), 1470 []byte("helloEF\u2C6F\u2C6FGH\u2C6F\u2C6F"), 1471 []byte("\u2C6F\u2C6F\u2C6F\u2C6FABCDhello")}, 1472 {not(isSpace), "hello" + space + "hello", 1473 []byte(space), 1474 []byte(space + "hello"), 1475 []byte("hello" + space)}, 1476 {not(isDigit), "hello\u0e50\u0e521234\u0e50\u0e51helo", 1477 []byte("\u0e50\u0e521234\u0e50\u0e51"), 1478 []byte("\u0e50\u0e521234\u0e50\u0e51helo"), 1479 []byte("hello\u0e50\u0e521234\u0e50\u0e51")}, 1480 {isValidRune, "ab\xc0a\xc0cd", 1481 []byte("\xc0a\xc0"), 1482 []byte("\xc0a\xc0cd"), 1483 []byte("ab\xc0a\xc0")}, 1484 {not(isValidRune), "\xc0a\xc0", 1485 []byte("a"), 1486 []byte("a\xc0"), 1487 []byte("\xc0a")}, 1488 // The nils returned by TrimLeftFunc are odd behavior, but we need 1489 // to preserve backwards compatibility. 1490 {isSpace, "", 1491 nil, 1492 nil, 1493 []byte("")}, 1494 {isSpace, " ", 1495 nil, 1496 nil, 1497 []byte("")}, 1498 } 1499 1500 func TestTrimFunc(t *testing.T) { 1501 for _, tc := range trimFuncTests { 1502 trimmers := []struct { 1503 name string 1504 trim func(s []byte, f func(r rune) bool) []byte 1505 out []byte 1506 }{ 1507 {"TrimFunc", TrimFunc, tc.trimOut}, 1508 {"TrimLeftFunc", TrimLeftFunc, tc.leftOut}, 1509 {"TrimRightFunc", TrimRightFunc, tc.rightOut}, 1510 } 1511 for _, trimmer := range trimmers { 1512 actual := trimmer.trim([]byte(tc.in), tc.f.f) 1513 if actual == nil && trimmer.out != nil { 1514 t.Errorf("%s(%q, %q) = nil; want %q", trimmer.name, tc.in, tc.f.name, trimmer.out) 1515 } 1516 if actual != nil && trimmer.out == nil { 1517 t.Errorf("%s(%q, %q) = %q; want nil", trimmer.name, tc.in, tc.f.name, actual) 1518 } 1519 if !Equal(actual, trimmer.out) { 1520 t.Errorf("%s(%q, %q) = %q; want %q", trimmer.name, tc.in, tc.f.name, actual, trimmer.out) 1521 } 1522 } 1523 } 1524 } 1525 1526 type IndexFuncTest struct { 1527 in string 1528 f predicate 1529 first, last int 1530 } 1531 1532 var indexFuncTests = []IndexFuncTest{ 1533 {"", isValidRune, -1, -1}, 1534 {"abc", isDigit, -1, -1}, 1535 {"0123", isDigit, 0, 3}, 1536 {"a1b", isDigit, 1, 1}, 1537 {space, isSpace, 0, len(space) - 3}, // last rune in space is 3 bytes 1538 {"\u0e50\u0e5212hello34\u0e50\u0e51", isDigit, 0, 18}, 1539 {"\u2C6F\u2C6F\u2C6F\u2C6FABCDhelloEF\u2C6F\u2C6FGH\u2C6F\u2C6F", isUpper, 0, 34}, 1540 {"12\u0e50\u0e52hello34\u0e50\u0e51", not(isDigit), 8, 12}, 1541 1542 // tests of invalid UTF-8 1543 {"\x801", isDigit, 1, 1}, 1544 {"\x80abc", isDigit, -1, -1}, 1545 {"\xc0a\xc0", isValidRune, 1, 1}, 1546 {"\xc0a\xc0", not(isValidRune), 0, 2}, 1547 {"\xc0☺\xc0", not(isValidRune), 0, 4}, 1548 {"\xc0☺\xc0\xc0", not(isValidRune), 0, 5}, 1549 {"ab\xc0a\xc0cd", not(isValidRune), 2, 4}, 1550 {"a\xe0\x80cd", not(isValidRune), 1, 2}, 1551 } 1552 1553 func TestIndexFunc(t *testing.T) { 1554 for _, tc := range indexFuncTests { 1555 first := IndexFunc([]byte(tc.in), tc.f.f) 1556 if first != tc.first { 1557 t.Errorf("IndexFunc(%q, %s) = %d; want %d", tc.in, tc.f.name, first, tc.first) 1558 } 1559 last := LastIndexFunc([]byte(tc.in), tc.f.f) 1560 if last != tc.last { 1561 t.Errorf("LastIndexFunc(%q, %s) = %d; want %d", tc.in, tc.f.name, last, tc.last) 1562 } 1563 } 1564 } 1565 1566 type ReplaceTest struct { 1567 in string 1568 old, new string 1569 n int 1570 out string 1571 } 1572 1573 var ReplaceTests = []ReplaceTest{ 1574 {"hello", "l", "L", 0, "hello"}, 1575 {"hello", "l", "L", -1, "heLLo"}, 1576 {"hello", "x", "X", -1, "hello"}, 1577 {"", "x", "X", -1, ""}, 1578 {"radar", "r", "<r>", -1, "<r>ada<r>"}, 1579 {"", "", "<>", -1, "<>"}, 1580 {"banana", "a", "<>", -1, "b<>n<>n<>"}, 1581 {"banana", "a", "<>", 1, "b<>nana"}, 1582 {"banana", "a", "<>", 1000, "b<>n<>n<>"}, 1583 {"banana", "an", "<>", -1, "b<><>a"}, 1584 {"banana", "ana", "<>", -1, "b<>na"}, 1585 {"banana", "", "<>", -1, "<>b<>a<>n<>a<>n<>a<>"}, 1586 {"banana", "", "<>", 10, "<>b<>a<>n<>a<>n<>a<>"}, 1587 {"banana", "", "<>", 6, "<>b<>a<>n<>a<>n<>a"}, 1588 {"banana", "", "<>", 5, "<>b<>a<>n<>a<>na"}, 1589 {"banana", "", "<>", 1, "<>banana"}, 1590 {"banana", "a", "a", -1, "banana"}, 1591 {"banana", "a", "a", 1, "banana"}, 1592 {"☺☻☹", "", "<>", -1, "<>☺<>☻<>☹<>"}, 1593 } 1594 1595 func TestReplace(t *testing.T) { 1596 for _, tt := range ReplaceTests { 1597 in := append([]byte(tt.in), "<spare>"...) 1598 in = in[:len(tt.in)] 1599 out := Replace(in, []byte(tt.old), []byte(tt.new), tt.n) 1600 if s := string(out); s != tt.out { 1601 t.Errorf("Replace(%q, %q, %q, %d) = %q, want %q", tt.in, tt.old, tt.new, tt.n, s, tt.out) 1602 } 1603 if cap(in) == cap(out) && &in[:1][0] == &out[:1][0] { 1604 t.Errorf("Replace(%q, %q, %q, %d) didn't copy", tt.in, tt.old, tt.new, tt.n) 1605 } 1606 if tt.n == -1 { 1607 out := ReplaceAll(in, []byte(tt.old), []byte(tt.new)) 1608 if s := string(out); s != tt.out { 1609 t.Errorf("ReplaceAll(%q, %q, %q) = %q, want %q", tt.in, tt.old, tt.new, s, tt.out) 1610 } 1611 } 1612 } 1613 } 1614 1615 type TitleTest struct { 1616 in, out string 1617 } 1618 1619 var TitleTests = []TitleTest{ 1620 {"", ""}, 1621 {"a", "A"}, 1622 {" aaa aaa aaa ", " Aaa Aaa Aaa "}, 1623 {" Aaa Aaa Aaa ", " Aaa Aaa Aaa "}, 1624 {"123a456", "123a456"}, 1625 {"double-blind", "Double-Blind"}, 1626 {"ÿøû", "Ÿøû"}, 1627 {"with_underscore", "With_underscore"}, 1628 {"unicode \xe2\x80\xa8 line separator", "Unicode \xe2\x80\xa8 Line Separator"}, 1629 } 1630 1631 func TestTitle(t *testing.T) { 1632 for _, tt := range TitleTests { 1633 if s := string(Title([]byte(tt.in))); s != tt.out { 1634 t.Errorf("Title(%q) = %q, want %q", tt.in, s, tt.out) 1635 } 1636 } 1637 } 1638 1639 var ToTitleTests = []TitleTest{ 1640 {"", ""}, 1641 {"a", "A"}, 1642 {" aaa aaa aaa ", " AAA AAA AAA "}, 1643 {" Aaa Aaa Aaa ", " AAA AAA AAA "}, 1644 {"123a456", "123A456"}, 1645 {"double-blind", "DOUBLE-BLIND"}, 1646 {"ÿøû", "ŸØÛ"}, 1647 } 1648 1649 func TestToTitle(t *testing.T) { 1650 for _, tt := range ToTitleTests { 1651 if s := string(ToTitle([]byte(tt.in))); s != tt.out { 1652 t.Errorf("ToTitle(%q) = %q, want %q", tt.in, s, tt.out) 1653 } 1654 } 1655 } 1656 1657 var EqualFoldTests = []struct { 1658 s, t string 1659 out bool 1660 }{ 1661 {"abc", "abc", true}, 1662 {"ABcd", "ABcd", true}, 1663 {"123abc", "123ABC", true}, 1664 {"αβδ", "ΑΒΔ", true}, 1665 {"abc", "xyz", false}, 1666 {"abc", "XYZ", false}, 1667 {"abcdefghijk", "abcdefghijX", false}, 1668 {"abcdefghijk", "abcdefghij\u212A", true}, 1669 {"abcdefghijK", "abcdefghij\u212A", true}, 1670 {"abcdefghijkz", "abcdefghij\u212Ay", false}, 1671 {"abcdefghijKz", "abcdefghij\u212Ay", false}, 1672 } 1673 1674 func TestEqualFold(t *testing.T) { 1675 for _, tt := range EqualFoldTests { 1676 if out := EqualFold([]byte(tt.s), []byte(tt.t)); out != tt.out { 1677 t.Errorf("EqualFold(%#q, %#q) = %v, want %v", tt.s, tt.t, out, tt.out) 1678 } 1679 if out := EqualFold([]byte(tt.t), []byte(tt.s)); out != tt.out { 1680 t.Errorf("EqualFold(%#q, %#q) = %v, want %v", tt.t, tt.s, out, tt.out) 1681 } 1682 } 1683 } 1684 1685 var cutTests = []struct { 1686 s, sep string 1687 before, after string 1688 found bool 1689 }{ 1690 {"abc", "b", "a", "c", true}, 1691 {"abc", "a", "", "bc", true}, 1692 {"abc", "c", "ab", "", true}, 1693 {"abc", "abc", "", "", true}, 1694 {"abc", "", "", "abc", true}, 1695 {"abc", "d", "abc", "", false}, 1696 {"", "d", "", "", false}, 1697 {"", "", "", "", true}, 1698 } 1699 1700 func TestCut(t *testing.T) { 1701 for _, tt := range cutTests { 1702 if before, after, found := Cut([]byte(tt.s), []byte(tt.sep)); string(before) != tt.before || string(after) != tt.after || found != tt.found { 1703 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) 1704 } 1705 } 1706 } 1707 1708 var cutPrefixTests = []struct { 1709 s, sep string 1710 after string 1711 found bool 1712 }{ 1713 {"abc", "a", "bc", true}, 1714 {"abc", "abc", "", true}, 1715 {"abc", "", "abc", true}, 1716 {"abc", "d", "abc", false}, 1717 {"", "d", "", false}, 1718 {"", "", "", true}, 1719 } 1720 1721 func TestCutPrefix(t *testing.T) { 1722 for _, tt := range cutPrefixTests { 1723 if after, found := CutPrefix([]byte(tt.s), []byte(tt.sep)); string(after) != tt.after || found != tt.found { 1724 t.Errorf("CutPrefix(%q, %q) = %q, %v, want %q, %v", tt.s, tt.sep, after, found, tt.after, tt.found) 1725 } 1726 } 1727 } 1728 1729 var cutSuffixTests = []struct { 1730 s, sep string 1731 before string 1732 found bool 1733 }{ 1734 {"abc", "bc", "a", true}, 1735 {"abc", "abc", "", true}, 1736 {"abc", "", "abc", true}, 1737 {"abc", "d", "abc", false}, 1738 {"", "d", "", false}, 1739 {"", "", "", true}, 1740 } 1741 1742 func TestCutSuffix(t *testing.T) { 1743 for _, tt := range cutSuffixTests { 1744 if before, found := CutSuffix([]byte(tt.s), []byte(tt.sep)); string(before) != tt.before || found != tt.found { 1745 t.Errorf("CutSuffix(%q, %q) = %q, %v, want %q, %v", tt.s, tt.sep, before, found, tt.before, tt.found) 1746 } 1747 } 1748 } 1749 1750 func TestBufferGrowNegative(t *testing.T) { 1751 defer func() { 1752 if err := recover(); err == nil { 1753 t.Fatal("Grow(-1) should have panicked") 1754 } 1755 }() 1756 var b Buffer 1757 b.Grow(-1) 1758 } 1759 1760 func TestBufferTruncateNegative(t *testing.T) { 1761 defer func() { 1762 if err := recover(); err == nil { 1763 t.Fatal("Truncate(-1) should have panicked") 1764 } 1765 }() 1766 var b Buffer 1767 b.Truncate(-1) 1768 } 1769 1770 func TestBufferTruncateOutOfRange(t *testing.T) { 1771 defer func() { 1772 if err := recover(); err == nil { 1773 t.Fatal("Truncate(20) should have panicked") 1774 } 1775 }() 1776 var b Buffer 1777 b.Write(make([]byte, 10)) 1778 b.Truncate(20) 1779 } 1780 1781 var containsTests = []struct { 1782 b, subslice []byte 1783 want bool 1784 }{ 1785 {[]byte("hello"), []byte("hel"), true}, 1786 {[]byte("日本語"), []byte("日本"), true}, 1787 {[]byte("hello"), []byte("Hello, world"), false}, 1788 {[]byte("東京"), []byte("京東"), false}, 1789 } 1790 1791 func TestContains(t *testing.T) { 1792 for _, tt := range containsTests { 1793 if got := Contains(tt.b, tt.subslice); got != tt.want { 1794 t.Errorf("Contains(%q, %q) = %v, want %v", tt.b, tt.subslice, got, tt.want) 1795 } 1796 } 1797 } 1798 1799 var ContainsAnyTests = []struct { 1800 b []byte 1801 substr string 1802 expected bool 1803 }{ 1804 {[]byte(""), "", false}, 1805 {[]byte(""), "a", false}, 1806 {[]byte(""), "abc", false}, 1807 {[]byte("a"), "", false}, 1808 {[]byte("a"), "a", true}, 1809 {[]byte("aaa"), "a", true}, 1810 {[]byte("abc"), "xyz", false}, 1811 {[]byte("abc"), "xcz", true}, 1812 {[]byte("a☺b☻c☹d"), "uvw☻xyz", true}, 1813 {[]byte("aRegExp*"), ".(|)*+?^$[]", true}, 1814 {[]byte(dots + dots + dots), " ", false}, 1815 } 1816 1817 func TestContainsAny(t *testing.T) { 1818 for _, ct := range ContainsAnyTests { 1819 if ContainsAny(ct.b, ct.substr) != ct.expected { 1820 t.Errorf("ContainsAny(%s, %s) = %v, want %v", 1821 ct.b, ct.substr, !ct.expected, ct.expected) 1822 } 1823 } 1824 } 1825 1826 var ContainsRuneTests = []struct { 1827 b []byte 1828 r rune 1829 expected bool 1830 }{ 1831 {[]byte(""), 'a', false}, 1832 {[]byte("a"), 'a', true}, 1833 {[]byte("aaa"), 'a', true}, 1834 {[]byte("abc"), 'y', false}, 1835 {[]byte("abc"), 'c', true}, 1836 {[]byte("a☺b☻c☹d"), 'x', false}, 1837 {[]byte("a☺b☻c☹d"), '☻', true}, 1838 {[]byte("aRegExp*"), '*', true}, 1839 } 1840 1841 func TestContainsRune(t *testing.T) { 1842 for _, ct := range ContainsRuneTests { 1843 if ContainsRune(ct.b, ct.r) != ct.expected { 1844 t.Errorf("ContainsRune(%q, %q) = %v, want %v", 1845 ct.b, ct.r, !ct.expected, ct.expected) 1846 } 1847 } 1848 } 1849 1850 func TestContainsFunc(t *testing.T) { 1851 for _, ct := range ContainsRuneTests { 1852 if ContainsFunc(ct.b, func(r rune) bool { 1853 return ct.r == r 1854 }) != ct.expected { 1855 t.Errorf("ContainsFunc(%q, func(%q)) = %v, want %v", 1856 ct.b, ct.r, !ct.expected, ct.expected) 1857 } 1858 } 1859 } 1860 1861 var makeFieldsInput = func() []byte { 1862 x := make([]byte, 1<<20) 1863 // Input is ~10% space, ~10% 2-byte UTF-8, rest ASCII non-space. 1864 for i := range x { 1865 switch rand.Intn(10) { 1866 case 0: 1867 x[i] = ' ' 1868 case 1: 1869 if i > 0 && x[i-1] == 'x' { 1870 copy(x[i-1:], "χ") 1871 break 1872 } 1873 fallthrough 1874 default: 1875 x[i] = 'x' 1876 } 1877 } 1878 return x 1879 } 1880 1881 var makeFieldsInputASCII = func() []byte { 1882 x := make([]byte, 1<<20) 1883 // Input is ~10% space, rest ASCII non-space. 1884 for i := range x { 1885 if rand.Intn(10) == 0 { 1886 x[i] = ' ' 1887 } else { 1888 x[i] = 'x' 1889 } 1890 } 1891 return x 1892 } 1893 1894 var bytesdata = []struct { 1895 name string 1896 data []byte 1897 }{ 1898 {"ASCII", makeFieldsInputASCII()}, 1899 {"Mixed", makeFieldsInput()}, 1900 } 1901 1902 func BenchmarkFields(b *testing.B) { 1903 for _, sd := range bytesdata { 1904 b.Run(sd.name, func(b *testing.B) { 1905 for j := 1 << 4; j <= 1<<20; j <<= 4 { 1906 b.Run(fmt.Sprintf("%d", j), func(b *testing.B) { 1907 b.ReportAllocs() 1908 b.SetBytes(int64(j)) 1909 data := sd.data[:j] 1910 for i := 0; i < b.N; i++ { 1911 Fields(data) 1912 } 1913 }) 1914 } 1915 }) 1916 } 1917 } 1918 1919 func BenchmarkFieldsFunc(b *testing.B) { 1920 for _, sd := range bytesdata { 1921 b.Run(sd.name, func(b *testing.B) { 1922 for j := 1 << 4; j <= 1<<20; j <<= 4 { 1923 b.Run(fmt.Sprintf("%d", j), func(b *testing.B) { 1924 b.ReportAllocs() 1925 b.SetBytes(int64(j)) 1926 data := sd.data[:j] 1927 for i := 0; i < b.N; i++ { 1928 FieldsFunc(data, unicode.IsSpace) 1929 } 1930 }) 1931 } 1932 }) 1933 } 1934 } 1935 1936 func BenchmarkTrimSpace(b *testing.B) { 1937 tests := []struct { 1938 name string 1939 input []byte 1940 }{ 1941 {"NoTrim", []byte("typical")}, 1942 {"ASCII", []byte(" foo bar ")}, 1943 {"SomeNonASCII", []byte(" \u2000\t\r\n x\t\t\r\r\ny\n \u3000 ")}, 1944 {"JustNonASCII", []byte("\u2000\u2000\u2000☺☺☺☺\u3000\u3000\u3000")}, 1945 } 1946 for _, test := range tests { 1947 b.Run(test.name, func(b *testing.B) { 1948 for i := 0; i < b.N; i++ { 1949 TrimSpace(test.input) 1950 } 1951 }) 1952 } 1953 } 1954 1955 func BenchmarkToValidUTF8(b *testing.B) { 1956 tests := []struct { 1957 name string 1958 input []byte 1959 }{ 1960 {"Valid", []byte("typical")}, 1961 {"InvalidASCII", []byte("foo\xffbar")}, 1962 {"InvalidNonASCII", []byte("日本語\xff日本語")}, 1963 } 1964 replacement := []byte("\uFFFD") 1965 b.ResetTimer() 1966 for _, test := range tests { 1967 b.Run(test.name, func(b *testing.B) { 1968 for i := 0; i < b.N; i++ { 1969 ToValidUTF8(test.input, replacement) 1970 } 1971 }) 1972 } 1973 } 1974 1975 func makeBenchInputHard() []byte { 1976 tokens := [...]string{ 1977 "<a>", "<p>", "<b>", "<strong>", 1978 "</a>", "</p>", "</b>", "</strong>", 1979 "hello", "world", 1980 } 1981 x := make([]byte, 0, 1<<20) 1982 for { 1983 i := rand.Intn(len(tokens)) 1984 if len(x)+len(tokens[i]) >= 1<<20 { 1985 break 1986 } 1987 x = append(x, tokens[i]...) 1988 } 1989 return x 1990 } 1991 1992 var benchInputHard = makeBenchInputHard() 1993 1994 func benchmarkIndexHard(b *testing.B, sep []byte) { 1995 for i := 0; i < b.N; i++ { 1996 Index(benchInputHard, sep) 1997 } 1998 } 1999 2000 func benchmarkLastIndexHard(b *testing.B, sep []byte) { 2001 for i := 0; i < b.N; i++ { 2002 LastIndex(benchInputHard, sep) 2003 } 2004 } 2005 2006 func benchmarkCountHard(b *testing.B, sep []byte) { 2007 for i := 0; i < b.N; i++ { 2008 Count(benchInputHard, sep) 2009 } 2010 } 2011 2012 func BenchmarkIndexHard1(b *testing.B) { benchmarkIndexHard(b, []byte("<>")) } 2013 func BenchmarkIndexHard2(b *testing.B) { benchmarkIndexHard(b, []byte("</pre>")) } 2014 func BenchmarkIndexHard3(b *testing.B) { benchmarkIndexHard(b, []byte("<b>hello world</b>")) } 2015 func BenchmarkIndexHard4(b *testing.B) { 2016 benchmarkIndexHard(b, []byte("<pre><b>hello</b><strong>world</strong></pre>")) 2017 } 2018 2019 func BenchmarkLastIndexHard1(b *testing.B) { benchmarkLastIndexHard(b, []byte("<>")) } 2020 func BenchmarkLastIndexHard2(b *testing.B) { benchmarkLastIndexHard(b, []byte("</pre>")) } 2021 func BenchmarkLastIndexHard3(b *testing.B) { benchmarkLastIndexHard(b, []byte("<b>hello world</b>")) } 2022 2023 func BenchmarkCountHard1(b *testing.B) { benchmarkCountHard(b, []byte("<>")) } 2024 func BenchmarkCountHard2(b *testing.B) { benchmarkCountHard(b, []byte("</pre>")) } 2025 func BenchmarkCountHard3(b *testing.B) { benchmarkCountHard(b, []byte("<b>hello world</b>")) } 2026 2027 func BenchmarkSplitEmptySeparator(b *testing.B) { 2028 for i := 0; i < b.N; i++ { 2029 Split(benchInputHard, nil) 2030 } 2031 } 2032 2033 func BenchmarkSplitSingleByteSeparator(b *testing.B) { 2034 sep := []byte("/") 2035 for i := 0; i < b.N; i++ { 2036 Split(benchInputHard, sep) 2037 } 2038 } 2039 2040 func BenchmarkSplitMultiByteSeparator(b *testing.B) { 2041 sep := []byte("hello") 2042 for i := 0; i < b.N; i++ { 2043 Split(benchInputHard, sep) 2044 } 2045 } 2046 2047 func BenchmarkSplitNSingleByteSeparator(b *testing.B) { 2048 sep := []byte("/") 2049 for i := 0; i < b.N; i++ { 2050 SplitN(benchInputHard, sep, 10) 2051 } 2052 } 2053 2054 func BenchmarkSplitNMultiByteSeparator(b *testing.B) { 2055 sep := []byte("hello") 2056 for i := 0; i < b.N; i++ { 2057 SplitN(benchInputHard, sep, 10) 2058 } 2059 } 2060 2061 func BenchmarkRepeat(b *testing.B) { 2062 for i := 0; i < b.N; i++ { 2063 Repeat([]byte("-"), 80) 2064 } 2065 } 2066 2067 func BenchmarkRepeatLarge(b *testing.B) { 2068 s := Repeat([]byte("@"), 8*1024) 2069 for j := 8; j <= 30; j++ { 2070 for _, k := range []int{1, 16, 4097} { 2071 s := s[:k] 2072 n := (1 << j) / k 2073 if n == 0 { 2074 continue 2075 } 2076 b.Run(fmt.Sprintf("%d/%d", 1<<j, k), func(b *testing.B) { 2077 for i := 0; i < b.N; i++ { 2078 Repeat(s, n) 2079 } 2080 b.SetBytes(int64(n * len(s))) 2081 }) 2082 } 2083 } 2084 } 2085 2086 func BenchmarkBytesCompare(b *testing.B) { 2087 for n := 1; n <= 2048; n <<= 1 { 2088 b.Run(fmt.Sprint(n), func(b *testing.B) { 2089 var x = make([]byte, n) 2090 var y = make([]byte, n) 2091 2092 for i := 0; i < n; i++ { 2093 x[i] = 'a' 2094 } 2095 2096 for i := 0; i < n; i++ { 2097 y[i] = 'a' 2098 } 2099 2100 b.ResetTimer() 2101 for i := 0; i < b.N; i++ { 2102 Compare(x, y) 2103 } 2104 }) 2105 } 2106 } 2107 2108 func BenchmarkIndexAnyASCII(b *testing.B) { 2109 x := Repeat([]byte{'#'}, 2048) // Never matches set 2110 cs := "0123456789abcdefghijklmnopqrstuvwxyz0123456789abcdefghijklmnopqrstuvwxyz" 2111 for k := 1; k <= 2048; k <<= 4 { 2112 for j := 1; j <= 64; j <<= 1 { 2113 b.Run(fmt.Sprintf("%d:%d", k, j), func(b *testing.B) { 2114 for i := 0; i < b.N; i++ { 2115 IndexAny(x[:k], cs[:j]) 2116 } 2117 }) 2118 } 2119 } 2120 } 2121 2122 func BenchmarkIndexAnyUTF8(b *testing.B) { 2123 x := Repeat([]byte{'#'}, 2048) // Never matches set 2124 cs := "你好世界, hello world. 你好世界, hello world. 你好世界, hello world." 2125 for k := 1; k <= 2048; k <<= 4 { 2126 for j := 1; j <= 64; j <<= 1 { 2127 b.Run(fmt.Sprintf("%d:%d", k, j), func(b *testing.B) { 2128 for i := 0; i < b.N; i++ { 2129 IndexAny(x[:k], cs[:j]) 2130 } 2131 }) 2132 } 2133 } 2134 } 2135 2136 func BenchmarkLastIndexAnyASCII(b *testing.B) { 2137 x := Repeat([]byte{'#'}, 2048) // Never matches set 2138 cs := "0123456789abcdefghijklmnopqrstuvwxyz0123456789abcdefghijklmnopqrstuvwxyz" 2139 for k := 1; k <= 2048; k <<= 4 { 2140 for j := 1; j <= 64; j <<= 1 { 2141 b.Run(fmt.Sprintf("%d:%d", k, j), func(b *testing.B) { 2142 for i := 0; i < b.N; i++ { 2143 LastIndexAny(x[:k], cs[:j]) 2144 } 2145 }) 2146 } 2147 } 2148 } 2149 2150 func BenchmarkLastIndexAnyUTF8(b *testing.B) { 2151 x := Repeat([]byte{'#'}, 2048) // Never matches set 2152 cs := "你好世界, hello world. 你好世界, hello world. 你好世界, hello world." 2153 for k := 1; k <= 2048; k <<= 4 { 2154 for j := 1; j <= 64; j <<= 1 { 2155 b.Run(fmt.Sprintf("%d:%d", k, j), func(b *testing.B) { 2156 for i := 0; i < b.N; i++ { 2157 LastIndexAny(x[:k], cs[:j]) 2158 } 2159 }) 2160 } 2161 } 2162 } 2163 2164 func BenchmarkTrimASCII(b *testing.B) { 2165 cs := "0123456789abcdef" 2166 for k := 1; k <= 4096; k <<= 4 { 2167 for j := 1; j <= 16; j <<= 1 { 2168 b.Run(fmt.Sprintf("%d:%d", k, j), func(b *testing.B) { 2169 x := Repeat([]byte(cs[:j]), k) // Always matches set 2170 for i := 0; i < b.N; i++ { 2171 Trim(x[:k], cs[:j]) 2172 } 2173 }) 2174 } 2175 } 2176 } 2177 2178 func BenchmarkTrimByte(b *testing.B) { 2179 x := []byte(" the quick brown fox ") 2180 for i := 0; i < b.N; i++ { 2181 Trim(x, " ") 2182 } 2183 } 2184 2185 func BenchmarkIndexPeriodic(b *testing.B) { 2186 key := []byte{1, 1} 2187 for _, skip := range [...]int{2, 4, 8, 16, 32, 64} { 2188 b.Run(fmt.Sprintf("IndexPeriodic%d", skip), func(b *testing.B) { 2189 buf := make([]byte, 1<<16) 2190 for i := 0; i < len(buf); i += skip { 2191 buf[i] = 1 2192 } 2193 for i := 0; i < b.N; i++ { 2194 Index(buf, key) 2195 } 2196 }) 2197 } 2198 } 2199 2200 func TestClone(t *testing.T) { 2201 var cloneTests = [][]byte{ 2202 []byte(nil), 2203 []byte{}, 2204 Clone([]byte{}), 2205 []byte(strings.Repeat("a", 42))[:0], 2206 []byte(strings.Repeat("a", 42))[:0:0], 2207 []byte("short"), 2208 []byte(strings.Repeat("a", 42)), 2209 } 2210 for _, input := range cloneTests { 2211 clone := Clone(input) 2212 if !Equal(clone, input) { 2213 t.Errorf("Clone(%q) = %q; want %q", input, clone, input) 2214 } 2215 2216 if input == nil && clone != nil { 2217 t.Errorf("Clone(%#v) return value should be equal to nil slice.", input) 2218 } 2219 2220 if input != nil && clone == nil { 2221 t.Errorf("Clone(%#v) return value should not be equal to nil slice.", input) 2222 } 2223 2224 if cap(input) != 0 && unsafe.SliceData(input) == unsafe.SliceData(clone) { 2225 t.Errorf("Clone(%q) return value should not reference inputs backing memory.", input) 2226 } 2227 } 2228 }