1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249
|
// Copyright 2020 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 json
import (
"bytes"
"errors"
"io"
"math/rand"
"reflect"
"testing"
)
func FuzzCoder(f *testing.F) {
// Add a number of inputs to the corpus including valid and invalid data.
for _, td := range coderTestdata {
f.Add(int64(0), []byte(td.in))
}
for _, td := range decoderErrorTestdata {
f.Add(int64(0), []byte(td.in))
}
for _, td := range encoderErrorTestdata {
f.Add(int64(0), []byte(td.wantOut))
}
for _, td := range jsonTestdata() {
f.Add(int64(0), td.data)
}
f.Fuzz(func(t *testing.T, seed int64, b []byte) {
var tokVals []tokOrVal
rn := rand.NewSource(seed)
// Read a sequence of tokens or values. Skip the test for any errors
// since we expect this with randomly generated fuzz inputs.
src := bytes.NewReader(b)
dec := NewDecoder(src)
for {
if rn.Int63()%8 > 0 {
tok, err := dec.ReadToken()
if err != nil {
if err == io.EOF {
break
}
t.Skipf("Decoder.ReadToken error: %v", err)
}
tokVals = append(tokVals, tok.Clone())
} else {
val, err := dec.ReadValue()
if err != nil {
expectError := dec.PeekKind() == '}' || dec.PeekKind() == ']'
if expectError && errors.As(err, new(*SyntacticError)) {
continue
}
if err == io.EOF {
break
}
t.Skipf("Decoder.ReadValue error: %v", err)
}
tokVals = append(tokVals, append(zeroValue, val...))
}
}
// Write a sequence of tokens or values. Fail the test for any errors
// since the previous stage guarantees that the input is valid.
dst := new(bytes.Buffer)
enc := NewEncoder(dst)
for _, tokVal := range tokVals {
switch tokVal := tokVal.(type) {
case Token:
if err := enc.WriteToken(tokVal); err != nil {
t.Fatalf("Encoder.WriteToken error: %v", err)
}
case RawValue:
if err := enc.WriteValue(tokVal); err != nil {
t.Fatalf("Encoder.WriteValue error: %v", err)
}
}
}
// Encoded output and original input must decode to the same thing.
var got, want []Token
for dec := NewDecoder(bytes.NewReader(b)); dec.PeekKind() > 0; {
tok, err := dec.ReadToken()
if err != nil {
t.Fatalf("Decoder.ReadToken error: %v", err)
}
got = append(got, tok.Clone())
}
for dec := NewDecoder(dst); dec.PeekKind() > 0; {
tok, err := dec.ReadToken()
if err != nil {
t.Fatalf("Decoder.ReadToken error: %v", err)
}
want = append(want, tok.Clone())
}
if !equalTokens(got, want) {
t.Fatalf("mismatching output:\ngot %v\nwant %v", got, want)
}
})
}
func FuzzResumableDecoder(f *testing.F) {
for _, td := range resumableDecoderTestdata {
f.Add(int64(0), []byte(td))
}
f.Fuzz(func(t *testing.T, seed int64, b []byte) {
rn := rand.NewSource(seed)
// Regardless of how many bytes the underlying io.Reader produces,
// the provided tokens, values, and errors should always be identical.
t.Run("ReadToken", func(t *testing.T) {
decGot := NewDecoder(&FaultyBuffer{B: b, MaxBytes: 8, Rand: rn})
decWant := NewDecoder(bytes.NewReader(b))
gotTok, gotErr := decGot.ReadToken()
wantTok, wantErr := decWant.ReadToken()
if gotTok.String() != wantTok.String() || !reflect.DeepEqual(gotErr, wantErr) {
t.Errorf("Decoder.ReadToken = (%v, %v), want (%v, %v)", gotTok, gotErr, wantTok, wantErr)
}
})
t.Run("ReadValue", func(t *testing.T) {
decGot := NewDecoder(&FaultyBuffer{B: b, MaxBytes: 8, Rand: rn})
decWant := NewDecoder(bytes.NewReader(b))
gotVal, gotErr := decGot.ReadValue()
wantVal, wantErr := decWant.ReadValue()
if !reflect.DeepEqual(gotVal, wantVal) || !reflect.DeepEqual(gotErr, wantErr) {
t.Errorf("Decoder.ReadValue = (%s, %v), want (%s, %v)", gotVal, gotErr, wantVal, wantErr)
}
})
})
}
func FuzzRawValueReformat(f *testing.F) {
for _, td := range rawValueTestdata {
f.Add([]byte(td.in))
}
// isValid reports whether b is valid according to the specified options.
isValid := func(opts DecodeOptions, b []byte) bool {
d := opts.NewDecoder(bytes.NewReader(b))
_, errVal := d.ReadValue()
_, errEOF := d.ReadToken()
return errVal == nil && errEOF == io.EOF
}
// stripWhitespace removes all JSON whitespace characters from the input.
stripWhitespace := func(in []byte) (out []byte) {
out = make([]byte, 0, len(in))
for _, c := range in {
switch c {
case ' ', '\n', '\r', '\t':
default:
out = append(out, c)
}
}
return out
}
// unmarshal unmarshals the input into an any.
unmarshal := func(in []byte) (out any) {
if err := Unmarshal(in, &out); err != nil {
return nil // ignore invalid input
}
return out
}
f.Fuzz(func(t *testing.T, b []byte) {
validRFC7159 := isValid(DecodeOptions{AllowInvalidUTF8: true, AllowDuplicateNames: true}, b)
validRFC8259 := isValid(DecodeOptions{AllowInvalidUTF8: false, AllowDuplicateNames: true}, b)
validRFC7493 := isValid(DecodeOptions{AllowInvalidUTF8: false, AllowDuplicateNames: false}, b)
switch {
case !validRFC7159 && validRFC8259:
t.Errorf("invalid input per RFC 7159 implies invalid per RFC 8259")
case !validRFC8259 && validRFC7493:
t.Errorf("invalid input per RFC 8259 implies invalid per RFC 7493")
}
gotValid := RawValue(b).IsValid()
wantValid := validRFC7493
if gotValid != wantValid {
t.Errorf("RawValue.IsValid = %v, want %v", gotValid, wantValid)
}
gotCompacted := RawValue(string(b))
gotCompactOk := gotCompacted.Compact() == nil
wantCompactOk := validRFC7159
if !bytes.Equal(stripWhitespace(gotCompacted), stripWhitespace(b)) {
t.Errorf("stripWhitespace(RawValue.Compact) = %s, want %s", stripWhitespace(gotCompacted), stripWhitespace(b))
}
if !reflect.DeepEqual(unmarshal(gotCompacted), unmarshal(b)) {
t.Errorf("unmarshal(RawValue.Compact) = %s, want %s", unmarshal(gotCompacted), unmarshal(b))
}
if gotCompactOk != wantCompactOk {
t.Errorf("RawValue.Compact success mismatch: got %v, want %v", gotCompactOk, wantCompactOk)
}
gotIndented := RawValue(string(b))
gotIndentOk := gotIndented.Indent("", " ") == nil
wantIndentOk := validRFC7159
if !bytes.Equal(stripWhitespace(gotIndented), stripWhitespace(b)) {
t.Errorf("stripWhitespace(RawValue.Indent) = %s, want %s", stripWhitespace(gotIndented), stripWhitespace(b))
}
if !reflect.DeepEqual(unmarshal(gotIndented), unmarshal(b)) {
t.Errorf("unmarshal(RawValue.Indent) = %s, want %s", unmarshal(gotIndented), unmarshal(b))
}
if gotIndentOk != wantIndentOk {
t.Errorf("RawValue.Indent success mismatch: got %v, want %v", gotIndentOk, wantIndentOk)
}
gotCanonicalized := RawValue(string(b))
gotCanonicalizeOk := gotCanonicalized.Canonicalize() == nil
wantCanonicalizeOk := validRFC7493
if !reflect.DeepEqual(unmarshal(gotCanonicalized), unmarshal(b)) {
t.Errorf("unmarshal(RawValue.Canonicalize) = %s, want %s", unmarshal(gotCanonicalized), unmarshal(b))
}
if gotCanonicalizeOk != wantCanonicalizeOk {
t.Errorf("RawValue.Canonicalize success mismatch: got %v, want %v", gotCanonicalizeOk, wantCanonicalizeOk)
}
})
}
func FuzzEqualFold(f *testing.F) {
for _, tt := range equalFoldTestdata {
f.Add([]byte(tt.in1), []byte(tt.in2))
}
equalFoldSimple := func(x, y []byte) bool {
strip := func(b []byte) []byte {
return bytes.Map(func(r rune) rune {
if r == '_' || r == '-' {
return -1 // ignore underscores and dashes
}
return r
}, b)
}
return bytes.EqualFold(strip(x), strip(y))
}
f.Fuzz(func(t *testing.T, s1, s2 []byte) {
// Compare the optimized and simplified implementations.
got := equalFold(s1, s2)
want := equalFoldSimple(s1, s2)
if got != want {
t.Errorf("equalFold(%q, %q) = %v, want %v", s1, s2, got, want)
}
})
}
|