src/strings/strings_test.go

// Copyright 2009 The Go Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.

package strings_test

import (
	"bytes"
	"io"
	"math/rand"
	"reflect"
	. "strings"
	"testing"
	"unicode"
	"unicode/utf8"
	"unsafe"
)

func eq(a, b []string) bool {
	if len(a) != len(b) {
		return false
	}
	for i := 0; i < len(a); i++ {
		if a[i] != b[i] {
			return false
		}
	}
	return true
}

var abcd = "abcd"
var faces = "☺☻☹"
var commas = "1,2,3,4"
var dots = "1....2....3....4"

type IndexTest struct {
	s   string
	sep string
	out int
}

var indexTests = []IndexTest{
	{"", "", 0},
	{"", "a", -1},
	{"", "foo", -1},
	{"fo", "foo", -1},
	{"foo", "foo", 0},
	{"oofofoofooo", "f", 2},
	{"oofofoofooo", "foo", 4},
	{"barfoobarfoo", "foo", 3},
	{"foo", "", 0},
	{"foo", "o", 1},
	{"abcABCabc", "A", 3},
	// cases with one byte strings - test special case in Index()
	{"", "a", -1},
	{"x", "a", -1},
	{"x", "x", 0},
	{"abc", "a", 0},
	{"abc", "b", 1},
	{"abc", "c", 2},
	{"abc", "x", -1},
	// test special cases in Index() for short strings
	{"", "ab", -1},
	{"bc", "ab", -1},
	{"ab", "ab", 0},
	{"xab", "ab", 1},
	{"xab"[:2], "ab", -1},
	{"", "abc", -1},
	{"xbc", "abc", -1},
	{"abc", "abc", 0},
	{"xabc", "abc", 1},
	{"xabc"[:3], "abc", -1},
	{"xabxc", "abc", -1},
	{"", "abcd", -1},
	{"xbcd", "abcd", -1},
	{"abcd", "abcd", 0},
	{"xabcd", "abcd", 1},
	{"xyabcd"[:5], "abcd", -1},
	{"xbcqq", "abcqq", -1},
	{"abcqq", "abcqq", 0},
	{"xabcqq", "abcqq", 1},
	{"xyabcqq"[:6], "abcqq", -1},
	{"xabxcqq", "abcqq", -1},
	{"xabcqxq", "abcqq", -1},
	{"", "01234567", -1},
	{"32145678", "01234567", -1},
	{"01234567", "01234567", 0},
	{"x01234567", "01234567", 1},
	{"xx01234567"[:9], "01234567", -1},
	{"", "0123456789", -1},
	{"3214567844", "0123456789", -1},
	{"0123456789", "0123456789", 0},
	{"x0123456789", "0123456789", 1},
	{"xyz0123456789"[:12], "0123456789", -1},
	{"x01234567x89", "0123456789", -1},
	{"", "0123456789012345", -1},
	{"3214567889012345", "0123456789012345", -1},
	{"0123456789012345", "0123456789012345", 0},
	{"x0123456789012345", "0123456789012345", 1},
	{"", "01234567890123456789", -1},
	{"32145678890123456789", "01234567890123456789", -1},
	{"01234567890123456789", "01234567890123456789", 0},
	{"x01234567890123456789", "01234567890123456789", 1},
	{"xyz01234567890123456789"[:22], "01234567890123456789", -1},
	{"", "0123456789012345678901234567890", -1},
	{"321456788901234567890123456789012345678911", "0123456789012345678901234567890", -1},
	{"0123456789012345678901234567890", "0123456789012345678901234567890", 0},
	{"x0123456789012345678901234567890", "0123456789012345678901234567890", 1},
	{"xyz0123456789012345678901234567890"[:33], "0123456789012345678901234567890", -1},
	{"", "01234567890123456789012345678901", -1},
	{"32145678890123456789012345678901234567890211", "01234567890123456789012345678901", -1},
	{"01234567890123456789012345678901", "01234567890123456789012345678901", 0},
	{"x01234567890123456789012345678901", "01234567890123456789012345678901", 1},
	{"xyz01234567890123456789012345678901"[:34], "01234567890123456789012345678901", -1},
}

var lastIndexTests = []IndexTest{
	{"", "", 0},
	{"", "a", -1},
	{"", "foo", -1},
	{"fo", "foo", -1},
	{"foo", "foo", 0},
	{"foo", "f", 0},
	{"oofofoofooo", "f", 7},
	{"oofofoofooo", "foo", 7},
	{"barfoobarfoo", "foo", 9},
	{"foo", "", 3},
	{"foo", "o", 2},
	{"abcABCabc", "A", 3},
	{"abcABCabc", "a", 6},
}

var indexAnyTests = []IndexTest{
	{"", "", -1},
	{"", "a", -1},
	{"", "abc", -1},
	{"a", "", -1},
	{"a", "a", 0},
	{"aaa", "a", 0},
	{"abc", "xyz", -1},
	{"abc", "xcz", 2},
	{"a☺b☻c☹d", "uvw☻xyz", 2 + len("☺")},
	{"aRegExp*", ".(|)*+?^$[]", 7},
	{dots + dots + dots, " ", -1},
}
var lastIndexAnyTests = []IndexTest{
	{"", "", -1},
	{"", "a", -1},
	{"", "abc", -1},
	{"a", "", -1},
	{"a", "a", 0},
	{"aaa", "a", 2},
	{"abc", "xyz", -1},
	{"abc", "ab", 1},
	{"a☺b☻c☹d", "uvw☻xyz", 2 + len("☺")},
	{"a.RegExp*", ".(|)*+?^$[]", 8},
	{dots + dots + dots, " ", -1},
}

// Execute f on each test case.  funcName should be the name of f; it's used
// in failure reports.
func runIndexTests(t *testing.T, f func(s, sep string) int, funcName string, testCases []IndexTest) {
	for _, test := range testCases {
		actual := f(test.s, test.sep)
		if actual != test.out {
			t.Errorf("%s(%q,%q) = %v; want %v", funcName, test.s, test.sep, actual, test.out)
		}
	}
}

func TestIndex(t *testing.T)        { runIndexTests(t, Index, "Index", indexTests) }
func TestLastIndex(t *testing.T)    { runIndexTests(t, LastIndex, "LastIndex", lastIndexTests) }
func TestIndexAny(t *testing.T)     { runIndexTests(t, IndexAny, "IndexAny", indexAnyTests) }
func TestLastIndexAny(t *testing.T) { runIndexTests(t, LastIndexAny, "LastIndexAny", lastIndexAnyTests) }

func TestLastIndexByte(t *testing.T) {
	testCases := []IndexTest{
		{"", "q", -1},
		{"abcdef", "q", -1},
		{"abcdefabcdef", "a", len("abcdef")},      // something in the middle
		{"abcdefabcdef", "f", len("abcdefabcde")}, // last byte
		{"zabcdefabcdef", "z", 0},                 // first byte
		{"a☺b☻c☹d", "b", len("a☺")},               // non-ascii
	}
	for _, test := range testCases {
		actual := LastIndexByte(test.s, test.sep[0])
		if actual != test.out {
			t.Errorf("LastIndexByte(%q,%c) = %v; want %v", test.s, test.sep[0], actual, test.out)
		}
	}
}

var indexRuneTests = []struct {
	s    string
	rune rune
	out  int
}{
	{"a A x", 'A', 2},
	{"some_text=some_value", '=', 9},
	{"☺a", 'a', 3},
	{"a☻☺b", '☺', 4},
}

func TestIndexRune(t *testing.T) {
	for _, test := range indexRuneTests {
		if actual := IndexRune(test.s, test.rune); actual != test.out {
			t.Errorf("IndexRune(%q,%d)= %v; want %v", test.s, test.rune, actual, test.out)
		}
	}
}

const benchmarkString = "some_text=some☺value"

func BenchmarkIndexRune(b *testing.B) {
	if got := IndexRune(benchmarkString, '☺'); got != 14 {
		b.Fatalf("wrong index: expected 14, got=%d", got)
	}
	for i := 0; i < b.N; i++ {
		IndexRune(benchmarkString, '☺')
	}
}

func BenchmarkIndexRuneFastPath(b *testing.B) {
	if got := IndexRune(benchmarkString, 'v'); got != 17 {
		b.Fatalf("wrong index: expected 17, got=%d", got)
	}
	for i := 0; i < b.N; i++ {
		IndexRune(benchmarkString, 'v')
	}
}

func BenchmarkIndex(b *testing.B) {
	if got := Index(benchmarkString, "v"); got != 17 {
		b.Fatalf("wrong index: expected 17, got=%d", got)
	}
	for i := 0; i < b.N; i++ {
		Index(benchmarkString, "v")
	}
}

func BenchmarkLastIndex(b *testing.B) {
	if got := Index(benchmarkString, "v"); got != 17 {
		b.Fatalf("wrong index: expected 17, got=%d", got)
	}
	for i := 0; i < b.N; i++ {
		LastIndex(benchmarkString, "v")
	}
}

func BenchmarkIndexByte(b *testing.B) {
	if got := IndexByte(benchmarkString, 'v'); got != 17 {
		b.Fatalf("wrong index: expected 17, got=%d", got)
	}
	for i := 0; i < b.N; i++ {
		IndexByte(benchmarkString, 'v')
	}
}

var explodetests = []struct {
	s string
	n int
	a []string
}{
	{"", -1, []string{}},
	{abcd, 4, []string{"a", "b", "c", "d"}},
	{faces, 3, []string{"☺", "☻", "☹"}},
	{abcd, 2, []string{"a", "bcd"}},
}

func TestExplode(t *testing.T) {
	for _, tt := range explodetests {
		a := SplitN(tt.s, "", tt.n)
		if !eq(a, tt.a) {
			t.Errorf("explode(%q, %d) = %v; want %v", tt.s, tt.n, a, tt.a)
			continue
		}
		s := Join(a, "")
		if s != tt.s {
			t.Errorf(`Join(explode(%q, %d), "") = %q`, tt.s, tt.n, s)
		}
	}
}

type SplitTest struct {
	s   string
	sep string
	n   int
	a   []string
}

var splittests = []SplitTest{
	{abcd, "a", 0, nil},
	{abcd, "a", -1, []string{"", "bcd"}},
	{abcd, "z", -1, []string{"abcd"}},
	{abcd, "", -1, []string{"a", "b", "c", "d"}},
	{commas, ",", -1, []string{"1", "2", "3", "4"}},
	{dots, "...", -1, []string{"1", ".2", ".3", ".4"}},
	{faces, "☹", -1, []string{"☺☻", ""}},
	{faces, "~", -1, []string{faces}},
	{faces, "", -1, []string{"☺", "☻", "☹"}},
	{"1 2 3 4", " ", 3, []string{"1", "2", "3 4"}},
	{"1 2", " ", 3, []string{"1", "2"}},
	{"123", "", 2, []string{"1", "23"}},
	{"123", "", 17, []string{"1", "2", "3"}},
}

func TestSplit(t *testing.T) {
	for _, tt := range splittests {
		a := SplitN(tt.s, tt.sep, tt.n)
		if !eq(a, tt.a) {
			t.Errorf("Split(%q, %q, %d) = %v; want %v", tt.s, tt.sep, tt.n, a, tt.a)
			continue
		}
		if tt.n == 0 {
			continue
		}
		s := Join(a, tt.sep)
		if s != tt.s {
			t.Errorf("Join(Split(%q, %q, %d), %q) = %q", tt.s, tt.sep, tt.n, tt.sep, s)
		}
		if tt.n < 0 {
			b := Split(tt.s, tt.sep)
			if !reflect.DeepEqual(a, b) {
				t.Errorf("Split disagrees with SplitN(%q, %q, %d) = %v; want %v", tt.s, tt.sep, tt.n, b, a)
			}
		}
	}
}

var splitaftertests = []SplitTest{
	{abcd, "a", -1, []string{"a", "bcd"}},
	{abcd, "z", -1, []string{"abcd"}},
	{abcd, "", -1, []string{"a", "b", "c", "d"}},
	{commas, ",", -1, []string{"1,", "2,", "3,", "4"}},
	{dots, "...", -1, []string{"1...", ".2...", ".3...", ".4"}},
	{faces, "☹", -1, []string{"☺☻☹", ""}},
	{faces, "~", -1, []string{faces}},
	{faces, "", -1, []string{"☺", "☻", "☹"}},
	{"1 2 3 4", " ", 3, []string{"1 ", "2 ", "3 4"}},
	{"1 2 3", " ", 3, []string{"1 ", "2 ", "3"}},
	{"1 2", " ", 3, []string{"1 ", "2"}},
	{"123", "", 2, []string{"1", "23"}},
	{"123", "", 17, []string{"1", "2", "3"}},
}

func TestSplitAfter(t *testing.T) {
	for _, tt := range splitaftertests {
		a := SplitAfterN(tt.s, tt.sep, tt.n)
		if !eq(a, tt.a) {
			t.Errorf(`Split(%q, %q, %d) = %v; want %v`, tt.s, tt.sep, tt.n, a, tt.a)
			continue
		}
		s := Join(a, "")
		if s != tt.s {
			t.Errorf(`Join(Split(%q, %q, %d), %q) = %q`, tt.s, tt.sep, tt.n, tt.sep, s)
		}
		if tt.n < 0 {
			b := SplitAfter(tt.s, tt.sep)
			if !reflect.DeepEqual(a, b) {
				t.Errorf("SplitAfter disagrees with SplitAfterN(%q, %q, %d) = %v; want %v", tt.s, tt.sep, tt.n, b, a)
			}
		}
	}
}

type FieldsTest struct {
	s string
	a []string
}

var fieldstests = []FieldsTest{
	{"", []string{}},
	{" ", []string{}},
	{" \t ", []string{}},
	{"  abc  ", []string{"abc"}},
	{"1 2 3 4", []string{"1", "2", "3", "4"}},
	{"1  2  3  4", []string{"1", "2", "3", "4"}},
	{"1\t\t2\t\t3\t4", []string{"1", "2", "3", "4"}},
	{"1\u20002\u20013\u20024", []string{"1", "2", "3", "4"}},
	{"\u2000\u2001\u2002", []string{}},
	{"\n™\t™\n", []string{"™", "™"}},
	{faces, []string{faces}},
}

func TestFields(t *testing.T) {
	for _, tt := range fieldstests {
		a := Fields(tt.s)
		if !eq(a, tt.a) {
			t.Errorf("Fields(%q) = %v; want %v", tt.s, a, tt.a)
			continue
		}
	}
}

var FieldsFuncTests = []FieldsTest{
	{"", []string{}},
	{"XX", []string{}},
	{"XXhiXXX", []string{"hi"}},
	{"aXXbXXXcX", []string{"a", "b", "c"}},
}

func TestFieldsFunc(t *testing.T) {
	for _, tt := range fieldstests {
		a := FieldsFunc(tt.s, unicode.IsSpace)
		if !eq(a, tt.a) {
			t.Errorf("FieldsFunc(%q, unicode.IsSpace) = %v; want %v", tt.s, a, tt.a)
			continue
		}
	}
	pred := func(c rune) bool { return c == 'X' }
	for _, tt := range FieldsFuncTests {
		a := FieldsFunc(tt.s, pred)
		if !eq(a, tt.a) {
			t.Errorf("FieldsFunc(%q) = %v, want %v", tt.s, a, tt.a)
		}
	}
}

// Test case for any function which accepts and returns a single string.
type StringTest struct {
	in, out string
}

// Execute f on each test case.  funcName should be the name of f; it's used
// in failure reports.
func runStringTests(t *testing.T, f func(string) string, funcName string, testCases []StringTest) {
	for _, tc := range testCases {
		actual := f(tc.in)
		if actual != tc.out {
			t.Errorf("%s(%q) = %q; want %q", funcName, tc.in, actual, tc.out)
		}
	}
}

var upperTests = []StringTest{
	{"", ""},
	{"abc", "ABC"},
	{"AbC123", "ABC123"},
	{"azAZ09_", "AZAZ09_"},
	{"\u0250\u0250\u0250\u0250\u0250", "\u2C6F\u2C6F\u2C6F\u2C6F\u2C6F"}, // grows one byte per char
}

var lowerTests = []StringTest{
	{"", ""},
	{"abc", "abc"},
	{"AbC123", "abc123"},
	{"azAZ09_", "azaz09_"},
	{"\u2C6D\u2C6D\u2C6D\u2C6D\u2C6D", "\u0251\u0251\u0251\u0251\u0251"}, // shrinks one byte per char
}

const space = "\t\v\r\f\n\u0085\u00a0\u2000\u3000"

var trimSpaceTests = []StringTest{
	{"", ""},
	{"abc", "abc"},
	{space + "abc" + space, "abc"},
	{" ", ""},
	{" \t\r\n \t\t\r\r\n\n ", ""},
	{" \t\r\n x\t\t\r\r\n\n ", "x"},
	{" \u2000\t\r\n x\t\t\r\r\ny\n \u3000", "x\t\t\r\r\ny"},
	{"1 \t\r\n2", "1 \t\r\n2"},
	{" x\x80", "x\x80"},
	{" x\xc0", "x\xc0"},
	{"x \xc0\xc0 ", "x \xc0\xc0"},
	{"x \xc0", "x \xc0"},
	{"x \xc0 ", "x \xc0"},
	{"x \xc0\xc0 ", "x \xc0\xc0"},
	{"x ☺\xc0\xc0 ", "x ☺\xc0\xc0"},
	{"x ☺ ", "x ☺"},
}

func tenRunes(ch rune) string {
	r := make([]rune, 10)
	for i := range r {
		r[i] = ch
	}
	return string(r)
}

// User-defined self-inverse mapping function
func rot13(r rune) rune {
	step := rune(13)
	if r >= 'a' && r <= 'z' {
		return ((r - 'a' + step) % 26) + 'a'
	}
	if r >= 'A' && r <= 'Z' {
		return ((r - 'A' + step) % 26) + 'A'
	}
	return r
}

func TestMap(t *testing.T) {
	// Run a couple of awful growth/shrinkage tests
	a := tenRunes('a')
	// 1.  Grow.  This triggers two reallocations in Map.
	maxRune := func(rune) rune { return unicode.MaxRune }
	m := Map(maxRune, a)
	expect := tenRunes(unicode.MaxRune)
	if m != expect {
		t.Errorf("growing: expected %q got %q", expect, m)
	}

	// 2. Shrink
	minRune := func(rune) rune { return 'a' }
	m = Map(minRune, tenRunes(unicode.MaxRune))
	expect = a
	if m != expect {
		t.Errorf("shrinking: expected %q got %q", expect, m)
	}

	// 3. Rot13
	m = Map(rot13, "a to zed")
	expect = "n gb mrq"
	if m != expect {
		t.Errorf("rot13: expected %q got %q", expect, m)
	}

	// 4. Rot13^2
	m = Map(rot13, Map(rot13, "a to zed"))
	expect = "a to zed"
	if m != expect {
		t.Errorf("rot13: expected %q got %q", expect, m)
	}

	// 5. Drop
	dropNotLatin := func(r rune) rune {
		if unicode.Is(unicode.Latin, r) {
			return r
		}
		return -1
	}
	m = Map(dropNotLatin, "Hello, 세계")
	expect = "Hello"
	if m != expect {
		t.Errorf("drop: expected %q got %q", expect, m)
	}

	// 6. Identity
	identity := func(r rune) rune {
		return r
	}
	orig := "Input string that we expect not to be copied."
	m = Map(identity, orig)
	if (*reflect.StringHeader)(unsafe.Pointer(&orig)).Data !=
		(*reflect.StringHeader)(unsafe.Pointer(&m)).Data {
		t.Error("unexpected copy during identity map")
	}
}

func TestToUpper(t *testing.T) { runStringTests(t, ToUpper, "ToUpper", upperTests) }

func TestToLower(t *testing.T) { runStringTests(t, ToLower, "ToLower", lowerTests) }

func BenchmarkMapNoChanges(b *testing.B) {
	identity := func(r rune) rune {
		return r
	}
	for i := 0; i < b.N; i++ {
		Map(identity, "Some string that won't be modified.")
	}
}

func TestSpecialCase(t *testing.T) {
	lower := "abcçdefgğhıijklmnoöprsştuüvyz"
	upper := "ABCÇDEFGĞHIİJKLMNOÖPRSŞTUÜVYZ"
	u := ToUpperSpecial(unicode.TurkishCase, upper)
	if u != upper {
		t.Errorf("Upper(upper) is %s not %s", u, upper)
	}
	u = ToUpperSpecial(unicode.TurkishCase, lower)
	if u != upper {
		t.Errorf("Upper(lower) is %s not %s", u, upper)
	}
	l := ToLowerSpecial(unicode.TurkishCase, lower)
	if l != lower {
		t.Errorf("Lower(lower) is %s not %s", l, lower)
	}
	l = ToLowerSpecial(unicode.TurkishCase, upper)
	if l != lower {
		t.Errorf("Lower(upper) is %s not %s", l, lower)
	}
}

func TestTrimSpace(t *testing.T) { runStringTests(t, TrimSpace, "TrimSpace", trimSpaceTests) }

var trimTests = []struct {
	f            string
	in, arg, out string
}{
	{"Trim", "abba", "a", "bb"},
	{"Trim", "abba", "ab", ""},
	{"TrimLeft", "abba", "ab", ""},
	{"TrimRight", "abba", "ab", ""},
	{"TrimLeft", "abba", "a", "bba"},
	{"TrimRight", "abba", "a", "abb"},
	{"Trim", "<tag>", "<>", "tag"},
	{"Trim", "* listitem", " *", "listitem"},
	{"Trim", `"quote"`, `"`, "quote"},
	{"Trim", "\u2C6F\u2C6F\u0250\u0250\u2C6F\u2C6F", "\u2C6F", "\u0250\u0250"},
	//empty string tests
	{"Trim", "abba", "", "abba"},
	{"Trim", "", "123", ""},
	{"Trim", "", "", ""},
	{"TrimLeft", "abba", "", "abba"},
	{"TrimLeft", "", "123", ""},
	{"TrimLeft", "", "", ""},
	{"TrimRight", "abba", "", "abba"},
	{"TrimRight", "", "123", ""},
	{"TrimRight", "", "", ""},
	{"TrimRight", "☺\xc0", "☺", "☺\xc0"},
	{"TrimPrefix", "aabb", "a", "abb"},
	{"TrimPrefix", "aabb", "b", "aabb"},
	{"TrimSuffix", "aabb", "a", "aabb"},
	{"TrimSuffix", "aabb", "b", "aab"},
}

func TestTrim(t *testing.T) {
	for _, tc := range trimTests {
		name := tc.f
		var f func(string, string) string
		switch name {
		case "Trim":
			f = Trim
		case "TrimLeft":
			f = TrimLeft
		case "TrimRight":
			f = TrimRight
		case "TrimPrefix":
			f = TrimPrefix
		case "TrimSuffix":
			f = TrimSuffix
		default:
			t.Errorf("Undefined trim function %s", name)
		}
		actual := f(tc.in, tc.arg)
		if actual != tc.out {
			t.Errorf("%s(%q, %q) = %q; want %q", name, tc.in, tc.arg, actual, tc.out)
		}
	}
}

func BenchmarkTrim(b *testing.B) {
	b.ReportAllocs()

	for i := 0; i < b.N; i++ {
		for _, tc := range trimTests {
			name := tc.f
			var f func(string, string) string
			switch name {
			case "Trim":
				f = Trim
			case "TrimLeft":
				f = TrimLeft
			case "TrimRight":
				f = TrimRight
			case "TrimPrefix":
				f = TrimPrefix
			case "TrimSuffix":
				f = TrimSuffix
			default:
				b.Errorf("Undefined trim function %s", name)
			}
			actual := f(tc.in, tc.arg)
			if actual != tc.out {
				b.Errorf("%s(%q, %q) = %q; want %q", name, tc.in, tc.arg, actual, tc.out)
			}
		}
	}
}

type predicate struct {
	f    func(rune) bool
	name string
}

var isSpace = predicate{unicode.IsSpace, "IsSpace"}
var isDigit = predicate{unicode.IsDigit, "IsDigit"}
var isUpper = predicate{unicode.IsUpper, "IsUpper"}
var isValidRune = predicate{
	func(r rune) bool {
		return r != utf8.RuneError
	},
	"IsValidRune",
}

func not(p predicate) predicate {
	return predicate{
		func(r rune) bool {
			return !p.f(r)
		},
		"not " + p.name,
	}
}

var trimFuncTests = []struct {
	f       predicate
	in, out string
}{
	{isSpace, space + " hello " + space, "hello"},
	{isDigit, "\u0e50\u0e5212hello34\u0e50\u0e51", "hello"},
	{isUpper, "\u2C6F\u2C6F\u2C6F\u2C6FABCDhelloEF\u2C6F\u2C6FGH\u2C6F\u2C6F", "hello"},
	{not(isSpace), "hello" + space + "hello", space},
	{not(isDigit), "hello\u0e50\u0e521234\u0e50\u0e51helo", "\u0e50\u0e521234\u0e50\u0e51"},
	{isValidRune, "ab\xc0a\xc0cd", "\xc0a\xc0"},
	{not(isValidRune), "\xc0a\xc0", "a"},
}

func TestTrimFunc(t *testing.T) {
	for _, tc := range trimFuncTests {
		actual := TrimFunc(tc.in, tc.f.f)
		if actual != tc.out {
			t.Errorf("TrimFunc(%q, %q) = %q; want %q", tc.in, tc.f.name, actual, tc.out)
		}
	}
}

var indexFuncTests = []struct {
	in          string
	f           predicate
	first, last int
}{
	{"", isValidRune, -1, -1},
	{"abc", isDigit, -1, -1},
	{"0123", isDigit, 0, 3},
	{"a1b", isDigit, 1, 1},
	{space, isSpace, 0, len(space) - 3}, // last rune in space is 3 bytes
	{"\u0e50\u0e5212hello34\u0e50\u0e51", isDigit, 0, 18},
	{"\u2C6F\u2C6F\u2C6F\u2C6FABCDhelloEF\u2C6F\u2C6FGH\u2C6F\u2C6F", isUpper, 0, 34},
	{"12\u0e50\u0e52hello34\u0e50\u0e51", not(isDigit), 8, 12},

	// tests of invalid UTF-8
	{"\x801", isDigit, 1, 1},
	{"\x80abc", isDigit, -1, -1},
	{"\xc0a\xc0", isValidRune, 1, 1},
	{"\xc0a\xc0", not(isValidRune), 0, 2},
	{"\xc0☺\xc0", not(isValidRune), 0, 4},
	{"\xc0☺\xc0\xc0", not(isValidRune), 0, 5},
	{"ab\xc0a\xc0cd", not(isValidRune), 2, 4},
	{"a\xe0\x80cd", not(isValidRune), 1, 2},
	{"\x80\x80\x80\x80", not(isValidRune), 0, 3},
}

func TestIndexFunc(t *testing.T) {
	for _, tc := range indexFuncTests {
		first := IndexFunc(tc.in, tc.f.f)
		if first != tc.first {
			t.Errorf("IndexFunc(%q, %s) = %d; want %d", tc.in, tc.f.name, first, tc.first)
		}
		last := LastIndexFunc(tc.in, tc.f.f)
		if last != tc.last {
			t.Errorf("LastIndexFunc(%q, %s) = %d; want %d", tc.in, tc.f.name, last, tc.last)
		}
	}
}

func equal(m string, s1, s2 string, t *testing.T) bool {
	if s1 == s2 {
		return true
	}
	e1 := Split(s1, "")
	e2 := Split(s2, "")
	for i, c1 := range e1 {
		if i >= len(e2) {
			break
		}
		r1, _ := utf8.DecodeRuneInString(c1)
		r2, _ := utf8.DecodeRuneInString(e2[i])
		if r1 != r2 {
			t.Errorf("%s diff at %d: U+%04X U+%04X", m, i, r1, r2)
		}
	}
	return false
}

func TestCaseConsistency(t *testing.T) {
	// Make a string of all the runes.
	numRunes := int(unicode.MaxRune + 1)
	if testing.Short() {
		numRunes = 1000
	}
	a := make([]rune, numRunes)
	for i := range a {
		a[i] = rune(i)
	}
	s := string(a)
	// convert the cases.
	upper := ToUpper(s)
	lower := ToLower(s)

	// Consistency checks
	if n := utf8.RuneCountInString(upper); n != numRunes {
		t.Error("rune count wrong in upper:", n)
	}
	if n := utf8.RuneCountInString(lower); n != numRunes {
		t.Error("rune count wrong in lower:", n)
	}
	if !equal("ToUpper(upper)", ToUpper(upper), upper, t) {
		t.Error("ToUpper(upper) consistency fail")
	}
	if !equal("ToLower(lower)", ToLower(lower), lower, t) {
		t.Error("ToLower(lower) consistency fail")
	}
	/*
		  These fail because of non-one-to-oneness of the data, such as multiple
		  upper case 'I' mapping to 'i'.  We comment them out but keep them for
		  interest.
		  For instance: CAPITAL LETTER I WITH DOT ABOVE:
			unicode.ToUpper(unicode.ToLower('\u0130')) != '\u0130'

		if !equal("ToUpper(lower)", ToUpper(lower), upper, t) {
			t.Error("ToUpper(lower) consistency fail");
		}
		if !equal("ToLower(upper)", ToLower(upper), lower, t) {
			t.Error("ToLower(upper) consistency fail");
		}
	*/
}

var RepeatTests = []struct {
	in, out string
	count   int
}{
	{"", "", 0},
	{"", "", 1},
	{"", "", 2},
	{"-", "", 0},
	{"-", "-", 1},
	{"-", "----------", 10},
	{"abc ", "abc abc abc ", 3},
}

func TestRepeat(t *testing.T) {
	for _, tt := range RepeatTests {
		a := Repeat(tt.in, tt.count)
		if !equal("Repeat(s)", a, tt.out, t) {
			t.Errorf("Repeat(%v, %d) = %v; want %v", tt.in, tt.count, a, tt.out)
			continue
		}
	}
}

func runesEqual(a, b []rune) bool {
	if len(a) != len(b) {
		return false
	}
	for i, r := range a {
		if r != b[i] {
			return false
		}
	}
	return true
}

var RunesTests = []struct {
	in    string
	out   []rune
	lossy bool
}{
	{"", []rune{}, false},
	{" ", []rune{32}, false},
	{"ABC", []rune{65, 66, 67}, false},
	{"abc", []rune{97, 98, 99}, false},
	{"\u65e5\u672c\u8a9e", []rune{26085, 26412, 35486}, false},
	{"ab\x80c", []rune{97, 98, 0xFFFD, 99}, true},
	{"ab\xc0c", []rune{97, 98, 0xFFFD, 99}, true},
}

func TestRunes(t *testing.T) {
	for _, tt := range RunesTests {
		a := []rune(tt.in)
		if !runesEqual(a, tt.out) {
			t.Errorf("[]rune(%q) = %v; want %v", tt.in, a, tt.out)
			continue
		}
		if !tt.lossy {
			// can only test reassembly if we didn't lose information
			s := string(a)
			if s != tt.in {
				t.Errorf("string([]rune(%q)) = %x; want %x", tt.in, s, tt.in)
			}
		}
	}
}

func TestReadByte(t *testing.T) {
	testStrings := []string{"", abcd, faces, commas}
	for _, s := range testStrings {
		reader := NewReader(s)
		if e := reader.UnreadByte(); e == nil {
			t.Errorf("Unreading %q at beginning: expected error", s)
		}
		var res bytes.Buffer
		for {
			b, e := reader.ReadByte()
			if e == io.EOF {
				break
			}
			if e != nil {
				t.Errorf("Reading %q: %s", s, e)
				break
			}
			res.WriteByte(b)
			// unread and read again
			e = reader.UnreadByte()
			if e != nil {
				t.Errorf("Unreading %q: %s", s, e)
				break
			}
			b1, e := reader.ReadByte()
			if e != nil {
				t.Errorf("Reading %q after unreading: %s", s, e)
				break
			}
			if b1 != b {
				t.Errorf("Reading %q after unreading: want byte %q, got %q", s, b, b1)
				break
			}
		}
		if res.String() != s {
			t.Errorf("Reader(%q).ReadByte() produced %q", s, res.String())
		}
	}
}

func TestReadRune(t *testing.T) {
	testStrings := []string{"", abcd, faces, commas}
	for _, s := range testStrings {
		reader := NewReader(s)
		if e := reader.UnreadRune(); e == nil {
			t.Errorf("Unreading %q at beginning: expected error", s)
		}
		res := ""
		for {
			r, z, e := reader.ReadRune()
			if e == io.EOF {
				break
			}
			if e != nil {
				t.Errorf("Reading %q: %s", s, e)
				break
			}
			res += string(r)
			// unread and read again
			e = reader.UnreadRune()
			if e != nil {
				t.Errorf("Unreading %q: %s", s, e)
				break
			}
			r1, z1, e := reader.ReadRune()
			if e != nil {
				t.Errorf("Reading %q after unreading: %s", s, e)
				break
			}
			if r1 != r {
				t.Errorf("Reading %q after unreading: want rune %q, got %q", s, r, r1)
				break
			}
			if z1 != z {
				t.Errorf("Reading %q after unreading: want size %d, got %d", s, z, z1)
				break
			}
		}
		if res != s {
			t.Errorf("Reader(%q).ReadRune() produced %q", s, res)
		}
	}
}

var UnreadRuneErrorTests = []struct {
	name string
	f    func(*Reader)
}{
	{"Read", func(r *Reader) { r.Read([]byte{0}) }},
	{"ReadByte", func(r *Reader) { r.ReadByte() }},
	{"UnreadRune", func(r *Reader) { r.UnreadRune() }},
	{"Seek", func(r *Reader) { r.Seek(0, 1) }},
	{"WriteTo", func(r *Reader) { r.WriteTo(&bytes.Buffer{}) }},
}

func TestUnreadRuneError(t *testing.T) {
	for _, tt := range UnreadRuneErrorTests {
		reader := NewReader("0123456789")
		if _, _, err := reader.ReadRune(); err != nil {
			// should not happen
			t.Fatal(err)
		}
		tt.f(reader)
		err := reader.UnreadRune()
		if err == nil {
			t.Errorf("Unreading after %s: expected error", tt.name)
		}
	}
}

var ReplaceTests = []struct {
	in       string
	old, new string
	n        int
	out      string
}{
	{"hello", "l", "L", 0, "hello"},
	{"hello", "l", "L", -1, "heLLo"},
	{"hello", "x", "X", -1, "hello"},
	{"", "x", "X", -1, ""},
	{"radar", "r", "<r>", -1, "<r>ada<r>"},
	{"", "", "<>", -1, "<>"},
	{"banana", "a", "<>", -1, "b<>n<>n<>"},
	{"banana", "a", "<>", 1, "b<>nana"},
	{"banana", "a", "<>", 1000, "b<>n<>n<>"},
	{"banana", "an", "<>", -1, "b<><>a"},
	{"banana", "ana", "<>", -1, "b<>na"},
	{"banana", "", "<>", -1, "<>b<>a<>n<>a<>n<>a<>"},
	{"banana", "", "<>", 10, "<>b<>a<>n<>a<>n<>a<>"},
	{"banana", "", "<>", 6, "<>b<>a<>n<>a<>n<>a"},
	{"banana", "", "<>", 5, "<>b<>a<>n<>a<>na"},
	{"banana", "", "<>", 1, "<>banana"},
	{"banana", "a", "a", -1, "banana"},
	{"banana", "a", "a", 1, "banana"},
	{"☺☻☹", "", "<>", -1, "<>☺<>☻<>☹<>"},
}

func TestReplace(t *testing.T) {
	for _, tt := range ReplaceTests {
		if s := Replace(tt.in, tt.old, tt.new, tt.n); s != tt.out {
			t.Errorf("Replace(%q, %q, %q, %d) = %q, want %q", tt.in, tt.old, tt.new, tt.n, s, tt.out)
		}
	}
}

var TitleTests = []struct {
	in, out string
}{
	{"", ""},
	{"a", "A"},
	{" aaa aaa aaa ", " Aaa Aaa Aaa "},
	{" Aaa Aaa Aaa ", " Aaa Aaa Aaa "},
	{"123a456", "123a456"},
	{"double-blind", "Double-Blind"},
	{"ÿøû", "Ÿøû"},
	{"with_underscore", "With_underscore"},
	{"unicode \xe2\x80\xa8 line separator", "Unicode \xe2\x80\xa8 Line Separator"},
}

func TestTitle(t *testing.T) {
	for _, tt := range TitleTests {
		if s := Title(tt.in); s != tt.out {
			t.Errorf("Title(%q) = %q, want %q", tt.in, s, tt.out)
		}
	}
}

var ContainsTests = []struct {
	str, substr string
	expected    bool
}{
	{"abc", "bc", true},
	{"abc", "bcd", false},
	{"abc", "", true},
	{"", "a", false},
}

func TestContains(t *testing.T) {
	for _, ct := range ContainsTests {
		if Contains(ct.str, ct.substr) != ct.expected {
			t.Errorf("Contains(%s, %s) = %v, want %v",
				ct.str, ct.substr, !ct.expected, ct.expected)
		}
	}
}

var ContainsAnyTests = []struct {
	str, substr string
	expected    bool
}{
	{"", "", false},
	{"", "a", false},
	{"", "abc", false},
	{"a", "", false},
	{"a", "a", true},
	{"aaa", "a", true},
	{"abc", "xyz", false},
	{"abc", "xcz", true},
	{"a☺b☻c☹d", "uvw☻xyz", true},
	{"aRegExp*", ".(|)*+?^$[]", true},
	{dots + dots + dots, " ", false},
}

func TestContainsAny(t *testing.T) {
	for _, ct := range ContainsAnyTests {
		if ContainsAny(ct.str, ct.substr) != ct.expected {
			t.Errorf("ContainsAny(%s, %s) = %v, want %v",
				ct.str, ct.substr, !ct.expected, ct.expected)
		}
	}
}

var ContainsRuneTests = []struct {
	str      string
	r        rune
	expected bool
}{
	{"", 'a', false},
	{"a", 'a', true},
	{"aaa", 'a', true},
	{"abc", 'y', false},
	{"abc", 'c', true},
	{"a☺b☻c☹d", 'x', false},
	{"a☺b☻c☹d", '☻', true},
	{"aRegExp*", '*', true},
}

func TestContainsRune(t *testing.T) {
	for _, ct := range ContainsRuneTests {
		if ContainsRune(ct.str, ct.r) != ct.expected {
			t.Errorf("ContainsRune(%q, %q) = %v, want %v",
				ct.str, ct.r, !ct.expected, ct.expected)
		}
	}
}

var EqualFoldTests = []struct {
	s, t string
	out  bool
}{
	{"abc", "abc", true},
	{"ABcd", "ABcd", true},
	{"123abc", "123ABC", true},
	{"αβδ", "ΑΒΔ", true},
	{"abc", "xyz", false},
	{"abc", "XYZ", false},
	{"abcdefghijk", "abcdefghijX", false},
	{"abcdefghijk", "abcdefghij\u212A", true},
	{"abcdefghijK", "abcdefghij\u212A", true},
	{"abcdefghijkz", "abcdefghij\u212Ay", false},
	{"abcdefghijKz", "abcdefghij\u212Ay", false},
}

func TestEqualFold(t *testing.T) {
	for _, tt := range EqualFoldTests {
		if out := EqualFold(tt.s, tt.t); out != tt.out {
			t.Errorf("EqualFold(%#q, %#q) = %v, want %v", tt.s, tt.t, out, tt.out)
		}
		if out := EqualFold(tt.t, tt.s); out != tt.out {
			t.Errorf("EqualFold(%#q, %#q) = %v, want %v", tt.t, tt.s, out, tt.out)
		}
	}
}

var CountTests = []struct {
	s, sep string
	num    int
}{
	{"", "", 1},
	{"", "notempty", 0},
	{"notempty", "", 9},
	{"smaller", "not smaller", 0},
	{"12345678987654321", "6", 2},
	{"611161116", "6", 3},
	{"notequal", "NotEqual", 0},
	{"equal", "equal", 1},
	{"abc1231231123q", "123", 3},
	{"11111", "11", 2},
}

func TestCount(t *testing.T) {
	for _, tt := range CountTests {
		if num := Count(tt.s, tt.sep); num != tt.num {
			t.Errorf("Count(\"%s\", \"%s\") = %d, want %d", tt.s, tt.sep, num, tt.num)
		}
	}
}

func makeBenchInputHard() string {
	tokens := [...]string{
		"<a>", "<p>", "<b>", "<strong>",
		"</a>", "</p>", "</b>", "</strong>",
		"hello", "world",
	}
	x := make([]byte, 0, 1<<20)
	for {
		i := rand.Intn(len(tokens))
		if len(x)+len(tokens[i]) >= 1<<20 {
			break
		}
		x = append(x, tokens[i]...)
	}
	return string(x)
}

var benchInputHard = makeBenchInputHard()

func benchmarkIndexHard(b *testing.B, sep string) {
	for i := 0; i < b.N; i++ {
		Index(benchInputHard, sep)
	}
}

func benchmarkLastIndexHard(b *testing.B, sep string) {
	for i := 0; i < b.N; i++ {
		LastIndex(benchInputHard, sep)
	}
}

func benchmarkCountHard(b *testing.B, sep string) {
	for i := 0; i < b.N; i++ {
		Count(benchInputHard, sep)
	}
}

func BenchmarkIndexHard1(b *testing.B) { benchmarkIndexHard(b, "<>") }
func BenchmarkIndexHard2(b *testing.B) { benchmarkIndexHard(b, "</pre>") }
func BenchmarkIndexHard3(b *testing.B) { benchmarkIndexHard(b, "<b>hello world</b>") }

func BenchmarkLastIndexHard1(b *testing.B) { benchmarkLastIndexHard(b, "<>") }
func BenchmarkLastIndexHard2(b *testing.B) { benchmarkLastIndexHard(b, "</pre>") }
func BenchmarkLastIndexHard3(b *testing.B) { benchmarkLastIndexHard(b, "<b>hello world</b>") }

func BenchmarkCountHard1(b *testing.B) { benchmarkCountHard(b, "<>") }
func BenchmarkCountHard2(b *testing.B) { benchmarkCountHard(b, "</pre>") }
func BenchmarkCountHard3(b *testing.B) { benchmarkCountHard(b, "<b>hello world</b>") }

var benchInputTorture = Repeat("ABC", 1<<10) + "123" + Repeat("ABC", 1<<10)
var benchNeedleTorture = Repeat("ABC", 1<<10+1)

func BenchmarkIndexTorture(b *testing.B) {
	for i := 0; i < b.N; i++ {
		Index(benchInputTorture, benchNeedleTorture)
	}
}

func BenchmarkCountTorture(b *testing.B) {
	for i := 0; i < b.N; i++ {
		Count(benchInputTorture, benchNeedleTorture)
	}
}

func BenchmarkCountTortureOverlapping(b *testing.B) {
	A := Repeat("ABC", 1<<20)
	B := Repeat("ABC", 1<<10)
	for i := 0; i < b.N; i++ {
		Count(A, B)
	}
}

var makeFieldsInput = func() string {
	x := make([]byte, 1<<20)
	// Input is ~10% space, ~10% 2-byte UTF-8, rest ASCII non-space.
	for i := range x {
		switch rand.Intn(10) {
		case 0:
			x[i] = ' '
		case 1:
			if i > 0 && x[i-1] == 'x' {
				copy(x[i-1:], "χ")
				break
			}
			fallthrough
		default:
			x[i] = 'x'
		}
	}
	return string(x)
}

var fieldsInput = makeFieldsInput()

func BenchmarkFields(b *testing.B) {
	b.SetBytes(int64(len(fieldsInput)))
	for i := 0; i < b.N; i++ {
		Fields(fieldsInput)
	}
}

func BenchmarkFieldsFunc(b *testing.B) {
	b.SetBytes(int64(len(fieldsInput)))
	for i := 0; i < b.N; i++ {
		FieldsFunc(fieldsInput, unicode.IsSpace)
	}
}

func BenchmarkSplit1(b *testing.B) {
	for i := 0; i < b.N; i++ {
		Split(benchInputHard, "")
	}
}

func BenchmarkSplit2(b *testing.B) {
	for i := 0; i < b.N; i++ {
		Split(benchInputHard, "/")
	}
}

func BenchmarkSplit3(b *testing.B) {
	for i := 0; i < b.N; i++ {
		Split(benchInputHard, "hello")
	}
}

func BenchmarkRepeat(b *testing.B) {
	for i := 0; i < b.N; i++ {
		Repeat("-", 80)
	}
}