go/src/encoding/json/decode_test.go

2337 lines
56 KiB
Go
Raw Normal View History

// Copyright 2010 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"
"encoding"
"errors"
"fmt"
"image"
"math"
"math/big"
"net"
"reflect"
"strconv"
"strings"
"testing"
"time"
)
type T struct {
X string
Y int
Z int `json:"-"`
}
type U struct {
Alphabet string `json:"alpha"`
}
type V struct {
F1 interface{}
F2 int32
F3 Number
F4 *VOuter
}
type VOuter struct {
V V
}
type W struct {
S SS
}
type P struct {
PP PP
}
type PP struct {
encoding/json: fix performance regression in the decoder In golang.org/cl/145218, a feature was added where the JSON decoder would keep track of the entire path to a field when reporting an UnmarshalTypeError. However, we all failed to check if this affected the benchmarks - myself included, as a reviewer. Below are the numbers comparing the CL's parent with itself, once it was merged: name old time/op new time/op delta CodeDecoder-8 12.9ms ± 1% 28.2ms ± 2% +119.33% (p=0.002 n=6+6) name old speed new speed delta CodeDecoder-8 151MB/s ± 1% 69MB/s ± 3% -54.40% (p=0.002 n=6+6) name old alloc/op new alloc/op delta CodeDecoder-8 2.74MB ± 0% 109.39MB ± 0% +3891.83% (p=0.002 n=6+6) name old allocs/op new allocs/op delta CodeDecoder-8 77.5k ± 0% 168.5k ± 0% +117.30% (p=0.004 n=6+5) The reason why the decoder got twice as slow is because it now allocated ~40x as many objects, which puts a lot of pressure on the garbage collector. The reason is that the CL concatenated strings every time a nested field was decoded. In other words, practically every field generated garbage when decoded. This is hugely wasteful, especially considering that the vast majority of JSON decoding inputs won't return UnmarshalTypeError. Instead, use a stack of fields, and make sure to always use the same backing array, to ensure we only need to grow the slice to the maximum depth once. The original CL also introduced a bug. The field stack string wasn't reset to its original state when reaching "d.opcode == scanEndObject", so the last field in a decoded struct could leak. For example, an added test decodes a list of structs, and encoding/json before this CL would fail: got: cannot unmarshal string into Go struct field T.Ts.Y.Y.Y of type int want: cannot unmarshal string into Go struct field T.Ts.Y of type int To fix that, simply reset the stack after decoding every field, even if it's the last. Below is the original performance versus this CL. There's a tiny performance hit, probably due to the append for every decoded field, but at least we're back to the usual ~150MB/s. name old time/op new time/op delta CodeDecoder-8 12.9ms ± 1% 13.0ms ± 1% +1.25% (p=0.009 n=6+6) name old speed new speed delta CodeDecoder-8 151MB/s ± 1% 149MB/s ± 1% -1.24% (p=0.009 n=6+6) name old alloc/op new alloc/op delta CodeDecoder-8 2.74MB ± 0% 2.74MB ± 0% +0.00% (p=0.002 n=6+6) name old allocs/op new allocs/op delta CodeDecoder-8 77.5k ± 0% 77.5k ± 0% +0.00% (p=0.002 n=6+6) Finally, make all of these benchmarks report allocs by default. The decoder ones are pretty sensitive to generated garbage, so ReportAllocs would have made the performance regression more obvious. Change-Id: I67b50f86b2e72f55539429450c67bfb1a9464b67 Reviewed-on: https://go-review.googlesource.com/c/go/+/167978 Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org> Run-TryBot: Daniel Martí <mvdan@mvdan.cc> TryBot-Result: Gobot Gobot <gobot@golang.org>
2019-03-17 22:45:30 +00:00
T T
Ts []T
}
type SS string
func (*SS) UnmarshalJSON(data []byte) error {
return &UnmarshalTypeError{Value: "number", Type: reflect.TypeOf(SS(""))}
}
// ifaceNumAsFloat64/ifaceNumAsNumber are used to test unmarshaling with and
// without UseNumber
var ifaceNumAsFloat64 = map[string]interface{}{
"k1": float64(1),
"k2": "s",
"k3": []interface{}{float64(1), float64(2.0), float64(3e-3)},
"k4": map[string]interface{}{"kk1": "s", "kk2": float64(2)},
}
var ifaceNumAsNumber = map[string]interface{}{
"k1": Number("1"),
"k2": "s",
"k3": []interface{}{Number("1"), Number("2.0"), Number("3e-3")},
"k4": map[string]interface{}{"kk1": "s", "kk2": Number("2")},
}
type tx struct {
x int
}
type u8 uint8
// A type that can unmarshal itself.
type unmarshaler struct {
T bool
}
func (u *unmarshaler) UnmarshalJSON(b []byte) error {
*u = unmarshaler{true} // All we need to see that UnmarshalJSON is called.
return nil
}
type ustruct struct {
M unmarshaler
}
type unmarshalerText struct {
A, B string
}
// needed for re-marshaling tests
func (u unmarshalerText) MarshalText() ([]byte, error) {
return []byte(u.A + ":" + u.B), nil
}
func (u *unmarshalerText) UnmarshalText(b []byte) error {
pos := bytes.IndexByte(b, ':')
if pos == -1 {
return errors.New("missing separator")
}
u.A, u.B = string(b[:pos]), string(b[pos+1:])
return nil
}
var _ encoding.TextUnmarshaler = (*unmarshalerText)(nil)
type ustructText struct {
M unmarshalerText
}
// u8marshal is an integer type that can marshal/unmarshal itself.
type u8marshal uint8
func (u8 u8marshal) MarshalText() ([]byte, error) {
return []byte(fmt.Sprintf("u%d", u8)), nil
}
var errMissingU8Prefix = errors.New("missing 'u' prefix")
func (u8 *u8marshal) UnmarshalText(b []byte) error {
if !bytes.HasPrefix(b, []byte{'u'}) {
return errMissingU8Prefix
}
n, err := strconv.Atoi(string(b[1:]))
if err != nil {
return err
}
*u8 = u8marshal(n)
return nil
}
var _ encoding.TextUnmarshaler = (*u8marshal)(nil)
var (
um0, um1 unmarshaler // target2 of unmarshaling
ump = &um1
umtrue = unmarshaler{true}
umslice = []unmarshaler{{true}}
umslicep = new([]unmarshaler)
umstruct = ustruct{unmarshaler{true}}
um0T, um1T unmarshalerText // target2 of unmarshaling
umpType = &um1T
umtrueXY = unmarshalerText{"x", "y"}
umsliceXY = []unmarshalerText{{"x", "y"}}
umslicepType = new([]unmarshalerText)
umstructType = new(ustructText)
umstructXY = ustructText{unmarshalerText{"x", "y"}}
ummapType = map[unmarshalerText]bool{}
encoding/json: encode struct field names ahead of time Struct field names are static, so we can run HTMLEscape on them when building each struct type encoder. Then, when running the struct encoder, we can select either the original or the escaped field name to write directly. When the encoder is not escaping HTML, using the original string works because neither Go struct field names nor JSON tags allow any characters that would need to be escaped, like '"', '\\', or '\n'. When the encoder is escaping HTML, the only difference is that '<', '>', and '&' are allowed via JSON struct field tags, hence why we use HTMLEscape to properly escape them. All of the above lets us encode field names with a simple if/else and WriteString calls, which are considerably simpler and faster than encoding an arbitrary string. While at it, also include the quotes and colon in these strings, to avoid three WriteByte calls in the loop hot path. Also added a few tests, to ensure that the behavior in these edge cases is not broken. The output of the tests is the same if this optimization is reverted. name old time/op new time/op delta CodeEncoder-4 7.12ms ± 0% 6.14ms ± 0% -13.85% (p=0.004 n=6+5) name old speed new speed delta CodeEncoder-4 272MB/s ± 0% 316MB/s ± 0% +16.08% (p=0.004 n=6+5) name old alloc/op new alloc/op delta CodeEncoder-4 91.9kB ± 0% 93.2kB ± 0% +1.43% (p=0.002 n=6+6) name old allocs/op new allocs/op delta CodeEncoder-4 0.00 0.00 ~ (all equal) Updates #5683. Change-Id: I6f6a340d0de4670799ce38cf95b2092822d2e3ef Reviewed-on: https://go-review.googlesource.com/122460 Run-TryBot: Daniel Martí <mvdan@mvdan.cc> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
2018-07-07 15:59:20 +01:00
ummapXY = map[unmarshalerText]bool{{"x", "y"}: true}
)
// Test data structures for anonymous fields.
type Point struct {
Z int
}
type Top struct {
Level0 int
Embed0
*Embed0a
*Embed0b `json:"e,omitempty"` // treated as named
Embed0c `json:"-"` // ignored
Loop
Embed0p // has Point with X, Y, used
Embed0q // has Point with Z, used
embed // contains exported field
}
type Embed0 struct {
Level1a int // overridden by Embed0a's Level1a with json tag
Level1b int // used because Embed0a's Level1b is renamed
Level1c int // used because Embed0a's Level1c is ignored
Level1d int // annihilated by Embed0a's Level1d
Level1e int `json:"x"` // annihilated by Embed0a.Level1e
}
type Embed0a struct {
Level1a int `json:"Level1a,omitempty"`
Level1b int `json:"LEVEL1B,omitempty"`
Level1c int `json:"-"`
Level1d int // annihilated by Embed0's Level1d
Level1f int `json:"x"` // annihilated by Embed0's Level1e
}
type Embed0b Embed0
type Embed0c Embed0
type Embed0p struct {
image.Point
}
type Embed0q struct {
Point
}
type embed struct {
Q int
}
type Loop struct {
Loop1 int `json:",omitempty"`
Loop2 int `json:",omitempty"`
*Loop
}
// From reflect test:
// The X in S6 and S7 annihilate, but they also block the X in S8.S9.
type S5 struct {
S6
S7
S8
}
type S6 struct {
X int
}
type S7 S6
type S8 struct {
S9
}
type S9 struct {
X int
Y int
}
// From reflect test:
// The X in S11.S6 and S12.S6 annihilate, but they also block the X in S13.S8.S9.
type S10 struct {
S11
S12
S13
}
type S11 struct {
S6
}
type S12 struct {
S6
}
type S13 struct {
S8
}
type Ambig struct {
// Given "hello", the first match should win.
First int `json:"HELLO"`
Second int `json:"Hello"`
}
type XYZ struct {
X interface{}
Y interface{}
Z interface{}
}
type unexportedWithMethods struct{}
func (unexportedWithMethods) F() {}
func sliceAddr(x []int) *[]int { return &x }
func mapAddr(x map[string]int) *map[string]int { return &x }
type byteWithMarshalJSON byte
func (b byteWithMarshalJSON) MarshalJSON() ([]byte, error) {
return []byte(fmt.Sprintf(`"Z%.2x"`, byte(b))), nil
}
func (b *byteWithMarshalJSON) UnmarshalJSON(data []byte) error {
if len(data) != 5 || data[0] != '"' || data[1] != 'Z' || data[4] != '"' {
return fmt.Errorf("bad quoted string")
}
i, err := strconv.ParseInt(string(data[2:4]), 16, 8)
if err != nil {
return fmt.Errorf("bad hex")
}
*b = byteWithMarshalJSON(i)
return nil
}
type byteWithPtrMarshalJSON byte
func (b *byteWithPtrMarshalJSON) MarshalJSON() ([]byte, error) {
return byteWithMarshalJSON(*b).MarshalJSON()
}
func (b *byteWithPtrMarshalJSON) UnmarshalJSON(data []byte) error {
return (*byteWithMarshalJSON)(b).UnmarshalJSON(data)
}
type byteWithMarshalText byte
func (b byteWithMarshalText) MarshalText() ([]byte, error) {
return []byte(fmt.Sprintf(`Z%.2x`, byte(b))), nil
}
func (b *byteWithMarshalText) UnmarshalText(data []byte) error {
if len(data) != 3 || data[0] != 'Z' {
return fmt.Errorf("bad quoted string")
}
i, err := strconv.ParseInt(string(data[1:3]), 16, 8)
if err != nil {
return fmt.Errorf("bad hex")
}
*b = byteWithMarshalText(i)
return nil
}
type byteWithPtrMarshalText byte
func (b *byteWithPtrMarshalText) MarshalText() ([]byte, error) {
return byteWithMarshalText(*b).MarshalText()
}
func (b *byteWithPtrMarshalText) UnmarshalText(data []byte) error {
return (*byteWithMarshalText)(b).UnmarshalText(data)
}
type intWithMarshalJSON int
func (b intWithMarshalJSON) MarshalJSON() ([]byte, error) {
return []byte(fmt.Sprintf(`"Z%.2x"`, int(b))), nil
}
func (b *intWithMarshalJSON) UnmarshalJSON(data []byte) error {
if len(data) != 5 || data[0] != '"' || data[1] != 'Z' || data[4] != '"' {
return fmt.Errorf("bad quoted string")
}
i, err := strconv.ParseInt(string(data[2:4]), 16, 8)
if err != nil {
return fmt.Errorf("bad hex")
}
*b = intWithMarshalJSON(i)
return nil
}
type intWithPtrMarshalJSON int
func (b *intWithPtrMarshalJSON) MarshalJSON() ([]byte, error) {
return intWithMarshalJSON(*b).MarshalJSON()
}
func (b *intWithPtrMarshalJSON) UnmarshalJSON(data []byte) error {
return (*intWithMarshalJSON)(b).UnmarshalJSON(data)
}
type intWithMarshalText int
func (b intWithMarshalText) MarshalText() ([]byte, error) {
return []byte(fmt.Sprintf(`Z%.2x`, int(b))), nil
}
func (b *intWithMarshalText) UnmarshalText(data []byte) error {
if len(data) != 3 || data[0] != 'Z' {
return fmt.Errorf("bad quoted string")
}
i, err := strconv.ParseInt(string(data[1:3]), 16, 8)
if err != nil {
return fmt.Errorf("bad hex")
}
*b = intWithMarshalText(i)
return nil
}
type intWithPtrMarshalText int
func (b *intWithPtrMarshalText) MarshalText() ([]byte, error) {
return intWithMarshalText(*b).MarshalText()
}
func (b *intWithPtrMarshalText) UnmarshalText(data []byte) error {
return (*intWithMarshalText)(b).UnmarshalText(data)
}
type mapStringToStringData struct {
Data map[string]string `json:"data"`
}
type unmarshalTest struct {
in string
ptr interface{}
out interface{}
err error
useNumber bool
golden bool
disallowUnknownFields bool
}
type B struct {
B bool `json:",string"`
}
var unmarshalTests = []unmarshalTest{
// basic types
{in: `true`, ptr: new(bool), out: true},
{in: `1`, ptr: new(int), out: 1},
{in: `1.2`, ptr: new(float64), out: 1.2},
{in: `-5`, ptr: new(int16), out: int16(-5)},
{in: `2`, ptr: new(Number), out: Number("2"), useNumber: true},
{in: `2`, ptr: new(Number), out: Number("2")},
{in: `2`, ptr: new(interface{}), out: float64(2.0)},
{in: `2`, ptr: new(interface{}), out: Number("2"), useNumber: true},
{in: `"a\u1234"`, ptr: new(string), out: "a\u1234"},
{in: `"http:\/\/"`, ptr: new(string), out: "http://"},
{in: `"g-clef: \uD834\uDD1E"`, ptr: new(string), out: "g-clef: \U0001D11E"},
{in: `"invalid: \uD834x\uDD1E"`, ptr: new(string), out: "invalid: \uFFFDx\uFFFD"},
{in: "null", ptr: new(interface{}), out: nil},
{in: `{"X": [1,2,3], "Y": 4}`, ptr: new(T), out: T{Y: 4}, err: &UnmarshalTypeError{"array", reflect.TypeOf(""), 7, "T", "X"}},
{in: `{"X": 23}`, ptr: new(T), out: T{}, err: &UnmarshalTypeError{"number", reflect.TypeOf(""), 8, "T", "X"}}, {in: `{"x": 1}`, ptr: new(tx), out: tx{}},
{in: `{"x": 1}`, ptr: new(tx), out: tx{}},
{in: `{"x": 1}`, ptr: new(tx), err: fmt.Errorf("json: unknown field \"x\""), disallowUnknownFields: true},
{in: `{"S": 23}`, ptr: new(W), out: W{}, err: &UnmarshalTypeError{"number", reflect.TypeOf(SS("")), 0, "W", "S"}},
{in: `{"F1":1,"F2":2,"F3":3}`, ptr: new(V), out: V{F1: float64(1), F2: int32(2), F3: Number("3")}},
{in: `{"F1":1,"F2":2,"F3":3}`, ptr: new(V), out: V{F1: Number("1"), F2: int32(2), F3: Number("3")}, useNumber: true},
{in: `{"k1":1,"k2":"s","k3":[1,2.0,3e-3],"k4":{"kk1":"s","kk2":2}}`, ptr: new(interface{}), out: ifaceNumAsFloat64},
{in: `{"k1":1,"k2":"s","k3":[1,2.0,3e-3],"k4":{"kk1":"s","kk2":2}}`, ptr: new(interface{}), out: ifaceNumAsNumber, useNumber: true},
encoding/json: improve performance of Unmarshal on primitive types Skip most of the scanning and parsing logic for simple (non-object/array) JSON values. benchmark old ns/op new ns/op delta BenchmarkUnmarshalInt 948 436 -54.01% BenchmarkUnmarshalUint 930 427 -54.09% BenchmarkUnmarshalString 1407 715 -49.18% BenchmarkUnmarshalFloat 1114 536 -51.89% BenchmarkUnmarshalBool 759 266 -64.95% BenchmarkUnmarshalStruct 8165 8181 +0.20% No significant effects on the go1 benchmarks: benchmark old ns/op new ns/op delta BenchmarkBinaryTree17 9647362752 9596196417 -0.53% BenchmarkFannkuch11 5623613048 5518694872 -1.87% BenchmarkGobDecode 32944041 33165434 +0.67% BenchmarkGobEncode 21237482 21080554 -0.74% BenchmarkGzip 750955920 749861980 -0.15% BenchmarkGunzip 197369742 197886192 +0.26% BenchmarkJSONEncode 79274091 78891137 -0.48% BenchmarkJSONDecode 180257802 175280358 -2.76% BenchmarkMandelbrot200 7396666 7388266 -0.11% BenchmarkParse 11446460 11386550 -0.52% BenchmarkRevcomp 1605152523 1599512029 -0.35% BenchmarkTemplate 204538247 207765574 +1.58% benchmark old MB/s new MB/s speedup BenchmarkGobDecode 23.30 23.14 0.99x BenchmarkGobEncode 36.14 36.41 1.01x BenchmarkGzip 25.84 25.88 1.00x BenchmarkGunzip 98.32 98.06 1.00x BenchmarkJSONEncode 24.48 24.60 1.00x BenchmarkJSONDecode 10.76 11.07 1.03x BenchmarkParse 5.06 5.09 1.01x BenchmarkRevcomp 158.34 158.90 1.00x BenchmarkTemplate 9.49 9.34 0.98x Fixes #3949. R=golang-dev, dave, bradfitz, timo CC=golang-dev https://golang.org/cl/7068043
2013-01-10 17:58:45 -08:00
// raw values with whitespace
{in: "\n true ", ptr: new(bool), out: true},
{in: "\t 1 ", ptr: new(int), out: 1},
{in: "\r 1.2 ", ptr: new(float64), out: 1.2},
{in: "\t -5 \n", ptr: new(int16), out: int16(-5)},
{in: "\t \"a\\u1234\" \n", ptr: new(string), out: "a\u1234"},
// Z has a "-" tag.
{in: `{"Y": 1, "Z": 2}`, ptr: new(T), out: T{Y: 1}},
{in: `{"Y": 1, "Z": 2}`, ptr: new(T), err: fmt.Errorf("json: unknown field \"Z\""), disallowUnknownFields: true},
{in: `{"alpha": "abc", "alphabet": "xyz"}`, ptr: new(U), out: U{Alphabet: "abc"}},
{in: `{"alpha": "abc", "alphabet": "xyz"}`, ptr: new(U), err: fmt.Errorf("json: unknown field \"alphabet\""), disallowUnknownFields: true},
{in: `{"alpha": "abc"}`, ptr: new(U), out: U{Alphabet: "abc"}},
{in: `{"alphabet": "xyz"}`, ptr: new(U), out: U{}},
{in: `{"alphabet": "xyz"}`, ptr: new(U), err: fmt.Errorf("json: unknown field \"alphabet\""), disallowUnknownFields: true},
// syntax errors
{in: `{"X": "foo", "Y"}`, err: &SyntaxError{"invalid character '}' after object key", 17}},
{in: `[1, 2, 3+]`, err: &SyntaxError{"invalid character '+' after array element", 9}},
{in: `{"X":12x}`, err: &SyntaxError{"invalid character 'x' after object key:value pair", 8}, useNumber: true},
{in: `[2, 3`, err: &SyntaxError{msg: "unexpected end of JSON input", Offset: 5}},
encoding/json: improve performance of Unmarshal on primitive types Skip most of the scanning and parsing logic for simple (non-object/array) JSON values. benchmark old ns/op new ns/op delta BenchmarkUnmarshalInt 948 436 -54.01% BenchmarkUnmarshalUint 930 427 -54.09% BenchmarkUnmarshalString 1407 715 -49.18% BenchmarkUnmarshalFloat 1114 536 -51.89% BenchmarkUnmarshalBool 759 266 -64.95% BenchmarkUnmarshalStruct 8165 8181 +0.20% No significant effects on the go1 benchmarks: benchmark old ns/op new ns/op delta BenchmarkBinaryTree17 9647362752 9596196417 -0.53% BenchmarkFannkuch11 5623613048 5518694872 -1.87% BenchmarkGobDecode 32944041 33165434 +0.67% BenchmarkGobEncode 21237482 21080554 -0.74% BenchmarkGzip 750955920 749861980 -0.15% BenchmarkGunzip 197369742 197886192 +0.26% BenchmarkJSONEncode 79274091 78891137 -0.48% BenchmarkJSONDecode 180257802 175280358 -2.76% BenchmarkMandelbrot200 7396666 7388266 -0.11% BenchmarkParse 11446460 11386550 -0.52% BenchmarkRevcomp 1605152523 1599512029 -0.35% BenchmarkTemplate 204538247 207765574 +1.58% benchmark old MB/s new MB/s speedup BenchmarkGobDecode 23.30 23.14 0.99x BenchmarkGobEncode 36.14 36.41 1.01x BenchmarkGzip 25.84 25.88 1.00x BenchmarkGunzip 98.32 98.06 1.00x BenchmarkJSONEncode 24.48 24.60 1.00x BenchmarkJSONDecode 10.76 11.07 1.03x BenchmarkParse 5.06 5.09 1.01x BenchmarkRevcomp 158.34 158.90 1.00x BenchmarkTemplate 9.49 9.34 0.98x Fixes #3949. R=golang-dev, dave, bradfitz, timo CC=golang-dev https://golang.org/cl/7068043
2013-01-10 17:58:45 -08:00
// raw value errors
{in: "\x01 42", err: &SyntaxError{"invalid character '\\x01' looking for beginning of value", 1}},
{in: " 42 \x01", err: &SyntaxError{"invalid character '\\x01' after top-level value", 5}},
{in: "\x01 true", err: &SyntaxError{"invalid character '\\x01' looking for beginning of value", 1}},
{in: " false \x01", err: &SyntaxError{"invalid character '\\x01' after top-level value", 8}},
{in: "\x01 1.2", err: &SyntaxError{"invalid character '\\x01' looking for beginning of value", 1}},
{in: " 3.4 \x01", err: &SyntaxError{"invalid character '\\x01' after top-level value", 6}},
{in: "\x01 \"string\"", err: &SyntaxError{"invalid character '\\x01' looking for beginning of value", 1}},
{in: " \"string\" \x01", err: &SyntaxError{"invalid character '\\x01' after top-level value", 11}},
// array tests
{in: `[1, 2, 3]`, ptr: new([3]int), out: [3]int{1, 2, 3}},
{in: `[1, 2, 3]`, ptr: new([1]int), out: [1]int{1}},
{in: `[1, 2, 3]`, ptr: new([5]int), out: [5]int{1, 2, 3, 0, 0}},
{in: `[1, 2, 3]`, ptr: new(MustNotUnmarshalJSON), err: errors.New("MustNotUnmarshalJSON was used")},
// empty array to interface test
{in: `[]`, ptr: new([]interface{}), out: []interface{}{}},
{in: `null`, ptr: new([]interface{}), out: []interface{}(nil)},
{in: `{"T":[]}`, ptr: new(map[string]interface{}), out: map[string]interface{}{"T": []interface{}{}}},
{in: `{"T":null}`, ptr: new(map[string]interface{}), out: map[string]interface{}{"T": interface{}(nil)}},
// composite tests
{in: allValueIndent, ptr: new(All), out: allValue},
{in: allValueCompact, ptr: new(All), out: allValue},
{in: allValueIndent, ptr: new(*All), out: &allValue},
{in: allValueCompact, ptr: new(*All), out: &allValue},
{in: pallValueIndent, ptr: new(All), out: pallValue},
{in: pallValueCompact, ptr: new(All), out: pallValue},
{in: pallValueIndent, ptr: new(*All), out: &pallValue},
{in: pallValueCompact, ptr: new(*All), out: &pallValue},
// unmarshal interface test
{in: `{"T":false}`, ptr: &um0, out: umtrue}, // use "false" so test will fail if custom unmarshaler is not called
{in: `{"T":false}`, ptr: &ump, out: &umtrue},
{in: `[{"T":false}]`, ptr: &umslice, out: umslice},
{in: `[{"T":false}]`, ptr: &umslicep, out: &umslice},
{in: `{"M":{"T":"x:y"}}`, ptr: &umstruct, out: umstruct},
// UnmarshalText interface test
{in: `"x:y"`, ptr: &um0T, out: umtrueXY},
{in: `"x:y"`, ptr: &umpType, out: &umtrueXY},
{in: `["x:y"]`, ptr: &umsliceXY, out: umsliceXY},
{in: `["x:y"]`, ptr: &umslicepType, out: &umsliceXY},
{in: `{"M":"x:y"}`, ptr: umstructType, out: umstructXY},
// integer-keyed map test
{
in: `{"-1":"a","0":"b","1":"c"}`,
ptr: new(map[int]string),
out: map[int]string{-1: "a", 0: "b", 1: "c"},
},
{
in: `{"0":"a","10":"c","9":"b"}`,
ptr: new(map[u8]string),
out: map[u8]string{0: "a", 9: "b", 10: "c"},
},
{
in: `{"-9223372036854775808":"min","9223372036854775807":"max"}`,
ptr: new(map[int64]string),
out: map[int64]string{math.MinInt64: "min", math.MaxInt64: "max"},
},
{
in: `{"18446744073709551615":"max"}`,
ptr: new(map[uint64]string),
out: map[uint64]string{math.MaxUint64: "max"},
},
{
in: `{"0":false,"10":true}`,
ptr: new(map[uintptr]bool),
out: map[uintptr]bool{0: false, 10: true},
},
// Check that MarshalText and UnmarshalText take precedence
// over default integer handling in map keys.
{
in: `{"u2":4}`,
ptr: new(map[u8marshal]int),
out: map[u8marshal]int{2: 4},
},
{
in: `{"2":4}`,
ptr: new(map[u8marshal]int),
err: errMissingU8Prefix,
},
// integer-keyed map errors
{
in: `{"abc":"abc"}`,
ptr: new(map[int]string),
err: &UnmarshalTypeError{Value: "number abc", Type: reflect.TypeOf(0), Offset: 2},
},
{
in: `{"256":"abc"}`,
ptr: new(map[uint8]string),
err: &UnmarshalTypeError{Value: "number 256", Type: reflect.TypeOf(uint8(0)), Offset: 2},
},
{
in: `{"128":"abc"}`,
ptr: new(map[int8]string),
err: &UnmarshalTypeError{Value: "number 128", Type: reflect.TypeOf(int8(0)), Offset: 2},
},
{
in: `{"-1":"abc"}`,
ptr: new(map[uint8]string),
err: &UnmarshalTypeError{Value: "number -1", Type: reflect.TypeOf(uint8(0)), Offset: 2},
},
{
in: `{"F":{"a":2,"3":4}}`,
ptr: new(map[string]map[int]int),
err: &UnmarshalTypeError{Value: "number a", Type: reflect.TypeOf(int(0)), Offset: 7},
},
{
in: `{"F":{"a":2,"3":4}}`,
ptr: new(map[string]map[uint]int),
err: &UnmarshalTypeError{Value: "number a", Type: reflect.TypeOf(uint(0)), Offset: 7},
},
// Map keys can be encoding.TextUnmarshalers.
{in: `{"x:y":true}`, ptr: &ummapType, out: ummapXY},
// If multiple values for the same key exists, only the most recent value is used.
{in: `{"x:y":false,"x:y":true}`, ptr: &ummapType, out: ummapXY},
// Overwriting of data.
// This is different from package xml, but it's what we've always done.
// Now documented and tested.
{in: `[2]`, ptr: sliceAddr([]int{1}), out: []int{2}},
{in: `{"key": 2}`, ptr: mapAddr(map[string]int{"old": 0, "key": 1}), out: map[string]int{"key": 2}},
{
in: `{
"Level0": 1,
"Level1b": 2,
"Level1c": 3,
"x": 4,
"Level1a": 5,
"LEVEL1B": 6,
"e": {
"Level1a": 8,
"Level1b": 9,
"Level1c": 10,
"Level1d": 11,
"x": 12
},
"Loop1": 13,
"Loop2": 14,
"X": 15,
"Y": 16,
"Z": 17,
"Q": 18
}`,
ptr: new(Top),
out: Top{
Level0: 1,
Embed0: Embed0{
Level1b: 2,
Level1c: 3,
},
Embed0a: &Embed0a{
Level1a: 5,
Level1b: 6,
},
Embed0b: &Embed0b{
Level1a: 8,
Level1b: 9,
Level1c: 10,
Level1d: 11,
Level1e: 12,
},
Loop: Loop{
Loop1: 13,
Loop2: 14,
},
Embed0p: Embed0p{
Point: image.Point{X: 15, Y: 16},
},
Embed0q: Embed0q{
Point: Point{Z: 17},
},
embed: embed{
Q: 18,
},
},
},
{
in: `{"hello": 1}`,
ptr: new(Ambig),
out: Ambig{First: 1},
},
{
in: `{"X": 1,"Y":2}`,
ptr: new(S5),
out: S5{S8: S8{S9: S9{Y: 2}}},
},
{
go/printer, gofmt: tuned table alignment for better results The go/printer (and thus gofmt) uses a heuristic to determine whether to break alignment between elements of an expression list which is spread across multiple lines. The heuristic only kicked in if the entry sizes (character length) was above a certain threshold (20) and the ratio between the previous and current entry size was above a certain value (4). This heuristic worked reasonably most of the time, but also led to unfortunate breaks in many cases where a single entry was suddenly much smaller (or larger) then the previous one. The behavior of gofmt was sufficiently mysterious in some of these situations that many issues were filed against it. The simplest solution to address this problem is to remove the heuristic altogether and have a programmer introduce empty lines to force different alignments if it improves readability. The problem with that approach is that the places where it really matters, very long tables with many (hundreds, or more) entries, may be machine-generated and not "post-processed" by a human (e.g., unicode/utf8/tables.go). If a single one of those entries is overlong, the result would be that the alignment would force all comments or values in key:value pairs to be adjusted to that overlong value, making the table hard to read (e.g., that entry may not even be visible on screen and all other entries seem spaced out too wide). Instead, we opted for a slightly improved heuristic that behaves much better for "normal", human-written code. 1) The threshold is increased from 20 to 40. This disables the heuristic for many common cases yet even if the alignment is not "ideal", 40 is not that many characters per line with todays screens, making it very likely that the entire line remains "visible" in an editor. 2) Changed the heuristic to not simply look at the size ratio between current and previous line, but instead considering the geometric mean of the sizes of the previous (aligned) lines. This emphasizes the "overall picture" of the previous lines, rather than a single one (which might be an outlier). 3) Changed the ratio from 4 to 2.5. Now that we ignore sizes below 40, a ratio of 4 would mean that a new entry would have to be 4 times bigger (160) or smaller (10) before alignment would be broken. A ratio of 2.5 seems more sensible. Applied updated gofmt to all of src and misc. Also tested against several former issues that complained about this and verified that the output for the given examples is satisfactory (added respective test cases). Some of the files changed because they were not gofmt-ed in the first place. For #644. For #7335. For #10392. (and probably more related issues) Fixes #22852. Change-Id: I5e48b3d3b157a5cf2d649833b7297b33f43a6f6e
2018-04-03 17:05:47 -07:00
in: `{"X": 1,"Y":2}`,
ptr: new(S5),
err: fmt.Errorf("json: unknown field \"X\""),
disallowUnknownFields: true,
},
{
in: `{"X": 1,"Y":2}`,
ptr: new(S10),
out: S10{S13: S13{S8: S8{S9: S9{Y: 2}}}},
},
{
go/printer, gofmt: tuned table alignment for better results The go/printer (and thus gofmt) uses a heuristic to determine whether to break alignment between elements of an expression list which is spread across multiple lines. The heuristic only kicked in if the entry sizes (character length) was above a certain threshold (20) and the ratio between the previous and current entry size was above a certain value (4). This heuristic worked reasonably most of the time, but also led to unfortunate breaks in many cases where a single entry was suddenly much smaller (or larger) then the previous one. The behavior of gofmt was sufficiently mysterious in some of these situations that many issues were filed against it. The simplest solution to address this problem is to remove the heuristic altogether and have a programmer introduce empty lines to force different alignments if it improves readability. The problem with that approach is that the places where it really matters, very long tables with many (hundreds, or more) entries, may be machine-generated and not "post-processed" by a human (e.g., unicode/utf8/tables.go). If a single one of those entries is overlong, the result would be that the alignment would force all comments or values in key:value pairs to be adjusted to that overlong value, making the table hard to read (e.g., that entry may not even be visible on screen and all other entries seem spaced out too wide). Instead, we opted for a slightly improved heuristic that behaves much better for "normal", human-written code. 1) The threshold is increased from 20 to 40. This disables the heuristic for many common cases yet even if the alignment is not "ideal", 40 is not that many characters per line with todays screens, making it very likely that the entire line remains "visible" in an editor. 2) Changed the heuristic to not simply look at the size ratio between current and previous line, but instead considering the geometric mean of the sizes of the previous (aligned) lines. This emphasizes the "overall picture" of the previous lines, rather than a single one (which might be an outlier). 3) Changed the ratio from 4 to 2.5. Now that we ignore sizes below 40, a ratio of 4 would mean that a new entry would have to be 4 times bigger (160) or smaller (10) before alignment would be broken. A ratio of 2.5 seems more sensible. Applied updated gofmt to all of src and misc. Also tested against several former issues that complained about this and verified that the output for the given examples is satisfactory (added respective test cases). Some of the files changed because they were not gofmt-ed in the first place. For #644. For #7335. For #10392. (and probably more related issues) Fixes #22852. Change-Id: I5e48b3d3b157a5cf2d649833b7297b33f43a6f6e
2018-04-03 17:05:47 -07:00
in: `{"X": 1,"Y":2}`,
ptr: new(S10),
err: fmt.Errorf("json: unknown field \"X\""),
disallowUnknownFields: true,
},
// invalid UTF-8 is coerced to valid UTF-8.
{
in: "\"hello\xffworld\"",
ptr: new(string),
out: "hello\ufffdworld",
},
{
in: "\"hello\xc2\xc2world\"",
ptr: new(string),
out: "hello\ufffd\ufffdworld",
},
{
in: "\"hello\xc2\xffworld\"",
ptr: new(string),
out: "hello\ufffd\ufffdworld",
},
{
in: "\"hello\\ud800world\"",
ptr: new(string),
out: "hello\ufffdworld",
},
{
in: "\"hello\\ud800\\ud800world\"",
ptr: new(string),
out: "hello\ufffd\ufffdworld",
},
{
in: "\"hello\\ud800\\ud800world\"",
ptr: new(string),
out: "hello\ufffd\ufffdworld",
},
{
in: "\"hello\xed\xa0\x80\xed\xb0\x80world\"",
ptr: new(string),
out: "hello\ufffd\ufffd\ufffd\ufffd\ufffd\ufffdworld",
},
// Used to be issue 8305, but time.Time implements encoding.TextUnmarshaler so this works now.
{
in: `{"2009-11-10T23:00:00Z": "hello world"}`,
ptr: &map[time.Time]string{},
out: map[time.Time]string{time.Date(2009, 11, 10, 23, 0, 0, 0, time.UTC): "hello world"},
},
// issue 8305
{
in: `{"2009-11-10T23:00:00Z": "hello world"}`,
ptr: &map[Point]string{},
err: &UnmarshalTypeError{Value: "object", Type: reflect.TypeOf(map[Point]string{}), Offset: 1},
},
{
in: `{"asdf": "hello world"}`,
ptr: &map[unmarshaler]string{},
err: &UnmarshalTypeError{Value: "object", Type: reflect.TypeOf(map[unmarshaler]string{}), Offset: 1},
},
// related to issue 13783.
// Go 1.7 changed marshaling a slice of typed byte to use the methods on the byte type,
// similar to marshaling a slice of typed int.
// These tests check that, assuming the byte type also has valid decoding methods,
// either the old base64 string encoding or the new per-element encoding can be
// successfully unmarshaled. The custom unmarshalers were accessible in earlier
// versions of Go, even though the custom marshaler was not.
{
in: `"AQID"`,
ptr: new([]byteWithMarshalJSON),
out: []byteWithMarshalJSON{1, 2, 3},
},
{
in: `["Z01","Z02","Z03"]`,
ptr: new([]byteWithMarshalJSON),
out: []byteWithMarshalJSON{1, 2, 3},
golden: true,
},
{
in: `"AQID"`,
ptr: new([]byteWithMarshalText),
out: []byteWithMarshalText{1, 2, 3},
},
{
in: `["Z01","Z02","Z03"]`,
ptr: new([]byteWithMarshalText),
out: []byteWithMarshalText{1, 2, 3},
golden: true,
},
{
in: `"AQID"`,
ptr: new([]byteWithPtrMarshalJSON),
out: []byteWithPtrMarshalJSON{1, 2, 3},
},
{
in: `["Z01","Z02","Z03"]`,
ptr: new([]byteWithPtrMarshalJSON),
out: []byteWithPtrMarshalJSON{1, 2, 3},
golden: true,
},
{
in: `"AQID"`,
ptr: new([]byteWithPtrMarshalText),
out: []byteWithPtrMarshalText{1, 2, 3},
},
{
in: `["Z01","Z02","Z03"]`,
ptr: new([]byteWithPtrMarshalText),
out: []byteWithPtrMarshalText{1, 2, 3},
golden: true,
},
// ints work with the marshaler but not the base64 []byte case
{
in: `["Z01","Z02","Z03"]`,
ptr: new([]intWithMarshalJSON),
out: []intWithMarshalJSON{1, 2, 3},
golden: true,
},
{
in: `["Z01","Z02","Z03"]`,
ptr: new([]intWithMarshalText),
out: []intWithMarshalText{1, 2, 3},
golden: true,
},
{
in: `["Z01","Z02","Z03"]`,
ptr: new([]intWithPtrMarshalJSON),
out: []intWithPtrMarshalJSON{1, 2, 3},
golden: true,
},
{
in: `["Z01","Z02","Z03"]`,
ptr: new([]intWithPtrMarshalText),
out: []intWithPtrMarshalText{1, 2, 3},
golden: true,
},
{in: `0.000001`, ptr: new(float64), out: 0.000001, golden: true},
{in: `1e-7`, ptr: new(float64), out: 1e-7, golden: true},
{in: `100000000000000000000`, ptr: new(float64), out: 100000000000000000000.0, golden: true},
{in: `1e+21`, ptr: new(float64), out: 1e21, golden: true},
{in: `-0.000001`, ptr: new(float64), out: -0.000001, golden: true},
{in: `-1e-7`, ptr: new(float64), out: -1e-7, golden: true},
{in: `-100000000000000000000`, ptr: new(float64), out: -100000000000000000000.0, golden: true},
{in: `-1e+21`, ptr: new(float64), out: -1e21, golden: true},
{in: `999999999999999900000`, ptr: new(float64), out: 999999999999999900000.0, golden: true},
{in: `9007199254740992`, ptr: new(float64), out: 9007199254740992.0, golden: true},
{in: `9007199254740993`, ptr: new(float64), out: 9007199254740992.0, golden: false},
{
in: `{"V": {"F2": "hello"}}`,
ptr: new(VOuter),
err: &UnmarshalTypeError{
Value: "string",
Struct: "V",
Field: "V.F2",
Type: reflect.TypeOf(int32(0)),
Offset: 20,
},
},
{
in: `{"V": {"F4": {}, "F2": "hello"}}`,
ptr: new(VOuter),
err: &UnmarshalTypeError{
Value: "string",
Struct: "V",
Field: "V.F2",
Type: reflect.TypeOf(int32(0)),
Offset: 30,
},
},
// issue 15146.
// invalid inputs in wrongStringTests below.
{in: `{"B":"true"}`, ptr: new(B), out: B{true}, golden: true},
{in: `{"B":"false"}`, ptr: new(B), out: B{false}, golden: true},
{in: `{"B": "maybe"}`, ptr: new(B), err: errors.New(`json: invalid use of ,string struct tag, trying to unmarshal "maybe" into bool`)},
{in: `{"B": "tru"}`, ptr: new(B), err: errors.New(`json: invalid use of ,string struct tag, trying to unmarshal "tru" into bool`)},
{in: `{"B": "False"}`, ptr: new(B), err: errors.New(`json: invalid use of ,string struct tag, trying to unmarshal "False" into bool`)},
{in: `{"B": "null"}`, ptr: new(B), out: B{false}},
{in: `{"B": "nul"}`, ptr: new(B), err: errors.New(`json: invalid use of ,string struct tag, trying to unmarshal "nul" into bool`)},
{in: `{"B": [2, 3]}`, ptr: new(B), err: errors.New(`json: invalid use of ,string struct tag, trying to unmarshal unquoted value into bool`)},
// additional tests for disallowUnknownFields
{
in: `{
"Level0": 1,
"Level1b": 2,
"Level1c": 3,
"x": 4,
"Level1a": 5,
"LEVEL1B": 6,
"e": {
"Level1a": 8,
"Level1b": 9,
"Level1c": 10,
"Level1d": 11,
"x": 12
},
"Loop1": 13,
"Loop2": 14,
"X": 15,
"Y": 16,
"Z": 17,
"Q": 18,
"extra": true
}`,
go/printer, gofmt: tuned table alignment for better results The go/printer (and thus gofmt) uses a heuristic to determine whether to break alignment between elements of an expression list which is spread across multiple lines. The heuristic only kicked in if the entry sizes (character length) was above a certain threshold (20) and the ratio between the previous and current entry size was above a certain value (4). This heuristic worked reasonably most of the time, but also led to unfortunate breaks in many cases where a single entry was suddenly much smaller (or larger) then the previous one. The behavior of gofmt was sufficiently mysterious in some of these situations that many issues were filed against it. The simplest solution to address this problem is to remove the heuristic altogether and have a programmer introduce empty lines to force different alignments if it improves readability. The problem with that approach is that the places where it really matters, very long tables with many (hundreds, or more) entries, may be machine-generated and not "post-processed" by a human (e.g., unicode/utf8/tables.go). If a single one of those entries is overlong, the result would be that the alignment would force all comments or values in key:value pairs to be adjusted to that overlong value, making the table hard to read (e.g., that entry may not even be visible on screen and all other entries seem spaced out too wide). Instead, we opted for a slightly improved heuristic that behaves much better for "normal", human-written code. 1) The threshold is increased from 20 to 40. This disables the heuristic for many common cases yet even if the alignment is not "ideal", 40 is not that many characters per line with todays screens, making it very likely that the entire line remains "visible" in an editor. 2) Changed the heuristic to not simply look at the size ratio between current and previous line, but instead considering the geometric mean of the sizes of the previous (aligned) lines. This emphasizes the "overall picture" of the previous lines, rather than a single one (which might be an outlier). 3) Changed the ratio from 4 to 2.5. Now that we ignore sizes below 40, a ratio of 4 would mean that a new entry would have to be 4 times bigger (160) or smaller (10) before alignment would be broken. A ratio of 2.5 seems more sensible. Applied updated gofmt to all of src and misc. Also tested against several former issues that complained about this and verified that the output for the given examples is satisfactory (added respective test cases). Some of the files changed because they were not gofmt-ed in the first place. For #644. For #7335. For #10392. (and probably more related issues) Fixes #22852. Change-Id: I5e48b3d3b157a5cf2d649833b7297b33f43a6f6e
2018-04-03 17:05:47 -07:00
ptr: new(Top),
err: fmt.Errorf("json: unknown field \"extra\""),
disallowUnknownFields: true,
},
{
in: `{
"Level0": 1,
"Level1b": 2,
"Level1c": 3,
"x": 4,
"Level1a": 5,
"LEVEL1B": 6,
"e": {
"Level1a": 8,
"Level1b": 9,
"Level1c": 10,
"Level1d": 11,
"x": 12,
"extra": null
},
"Loop1": 13,
"Loop2": 14,
"X": 15,
"Y": 16,
"Z": 17,
"Q": 18
}`,
go/printer, gofmt: tuned table alignment for better results The go/printer (and thus gofmt) uses a heuristic to determine whether to break alignment between elements of an expression list which is spread across multiple lines. The heuristic only kicked in if the entry sizes (character length) was above a certain threshold (20) and the ratio between the previous and current entry size was above a certain value (4). This heuristic worked reasonably most of the time, but also led to unfortunate breaks in many cases where a single entry was suddenly much smaller (or larger) then the previous one. The behavior of gofmt was sufficiently mysterious in some of these situations that many issues were filed against it. The simplest solution to address this problem is to remove the heuristic altogether and have a programmer introduce empty lines to force different alignments if it improves readability. The problem with that approach is that the places where it really matters, very long tables with many (hundreds, or more) entries, may be machine-generated and not "post-processed" by a human (e.g., unicode/utf8/tables.go). If a single one of those entries is overlong, the result would be that the alignment would force all comments or values in key:value pairs to be adjusted to that overlong value, making the table hard to read (e.g., that entry may not even be visible on screen and all other entries seem spaced out too wide). Instead, we opted for a slightly improved heuristic that behaves much better for "normal", human-written code. 1) The threshold is increased from 20 to 40. This disables the heuristic for many common cases yet even if the alignment is not "ideal", 40 is not that many characters per line with todays screens, making it very likely that the entire line remains "visible" in an editor. 2) Changed the heuristic to not simply look at the size ratio between current and previous line, but instead considering the geometric mean of the sizes of the previous (aligned) lines. This emphasizes the "overall picture" of the previous lines, rather than a single one (which might be an outlier). 3) Changed the ratio from 4 to 2.5. Now that we ignore sizes below 40, a ratio of 4 would mean that a new entry would have to be 4 times bigger (160) or smaller (10) before alignment would be broken. A ratio of 2.5 seems more sensible. Applied updated gofmt to all of src and misc. Also tested against several former issues that complained about this and verified that the output for the given examples is satisfactory (added respective test cases). Some of the files changed because they were not gofmt-ed in the first place. For #644. For #7335. For #10392. (and probably more related issues) Fixes #22852. Change-Id: I5e48b3d3b157a5cf2d649833b7297b33f43a6f6e
2018-04-03 17:05:47 -07:00
ptr: new(Top),
err: fmt.Errorf("json: unknown field \"extra\""),
disallowUnknownFields: true,
},
// issue 26444
// UnmarshalTypeError without field & struct values
{
in: `{"data":{"test1": "bob", "test2": 123}}`,
ptr: new(mapStringToStringData),
err: &UnmarshalTypeError{Value: "number", Type: reflect.TypeOf(""), Offset: 37, Struct: "mapStringToStringData", Field: "data"},
},
{
in: `{"data":{"test1": 123, "test2": "bob"}}`,
ptr: new(mapStringToStringData),
err: &UnmarshalTypeError{Value: "number", Type: reflect.TypeOf(""), Offset: 21, Struct: "mapStringToStringData", Field: "data"},
},
// trying to decode JSON arrays or objects via TextUnmarshaler
{
in: `[1, 2, 3]`,
ptr: new(MustNotUnmarshalText),
err: &UnmarshalTypeError{Value: "array", Type: reflect.TypeOf(&MustNotUnmarshalText{}), Offset: 1},
},
{
in: `{"foo": "bar"}`,
ptr: new(MustNotUnmarshalText),
err: &UnmarshalTypeError{Value: "object", Type: reflect.TypeOf(&MustNotUnmarshalText{}), Offset: 1},
},
// #22369
{
in: `{"PP": {"T": {"Y": "bad-type"}}}`,
ptr: new(P),
err: &UnmarshalTypeError{
Value: "string",
Struct: "T",
Field: "PP.T.Y",
Type: reflect.TypeOf(int(0)),
Offset: 29,
},
},
encoding/json: fix performance regression in the decoder In golang.org/cl/145218, a feature was added where the JSON decoder would keep track of the entire path to a field when reporting an UnmarshalTypeError. However, we all failed to check if this affected the benchmarks - myself included, as a reviewer. Below are the numbers comparing the CL's parent with itself, once it was merged: name old time/op new time/op delta CodeDecoder-8 12.9ms ± 1% 28.2ms ± 2% +119.33% (p=0.002 n=6+6) name old speed new speed delta CodeDecoder-8 151MB/s ± 1% 69MB/s ± 3% -54.40% (p=0.002 n=6+6) name old alloc/op new alloc/op delta CodeDecoder-8 2.74MB ± 0% 109.39MB ± 0% +3891.83% (p=0.002 n=6+6) name old allocs/op new allocs/op delta CodeDecoder-8 77.5k ± 0% 168.5k ± 0% +117.30% (p=0.004 n=6+5) The reason why the decoder got twice as slow is because it now allocated ~40x as many objects, which puts a lot of pressure on the garbage collector. The reason is that the CL concatenated strings every time a nested field was decoded. In other words, practically every field generated garbage when decoded. This is hugely wasteful, especially considering that the vast majority of JSON decoding inputs won't return UnmarshalTypeError. Instead, use a stack of fields, and make sure to always use the same backing array, to ensure we only need to grow the slice to the maximum depth once. The original CL also introduced a bug. The field stack string wasn't reset to its original state when reaching "d.opcode == scanEndObject", so the last field in a decoded struct could leak. For example, an added test decodes a list of structs, and encoding/json before this CL would fail: got: cannot unmarshal string into Go struct field T.Ts.Y.Y.Y of type int want: cannot unmarshal string into Go struct field T.Ts.Y of type int To fix that, simply reset the stack after decoding every field, even if it's the last. Below is the original performance versus this CL. There's a tiny performance hit, probably due to the append for every decoded field, but at least we're back to the usual ~150MB/s. name old time/op new time/op delta CodeDecoder-8 12.9ms ± 1% 13.0ms ± 1% +1.25% (p=0.009 n=6+6) name old speed new speed delta CodeDecoder-8 151MB/s ± 1% 149MB/s ± 1% -1.24% (p=0.009 n=6+6) name old alloc/op new alloc/op delta CodeDecoder-8 2.74MB ± 0% 2.74MB ± 0% +0.00% (p=0.002 n=6+6) name old allocs/op new allocs/op delta CodeDecoder-8 77.5k ± 0% 77.5k ± 0% +0.00% (p=0.002 n=6+6) Finally, make all of these benchmarks report allocs by default. The decoder ones are pretty sensitive to generated garbage, so ReportAllocs would have made the performance regression more obvious. Change-Id: I67b50f86b2e72f55539429450c67bfb1a9464b67 Reviewed-on: https://go-review.googlesource.com/c/go/+/167978 Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org> Run-TryBot: Daniel Martí <mvdan@mvdan.cc> TryBot-Result: Gobot Gobot <gobot@golang.org>
2019-03-17 22:45:30 +00:00
{
in: `{"Ts": [{"Y": 1}, {"Y": 2}, {"Y": "bad-type"}]}`,
ptr: new(PP),
err: &UnmarshalTypeError{
Value: "string",
Struct: "T",
Field: "Ts.Y",
Type: reflect.TypeOf(int(0)),
Offset: 29,
},
},
}
func TestMarshal(t *testing.T) {
b, err := Marshal(allValue)
if err != nil {
t.Fatalf("Marshal allValue: %v", err)
}
if string(b) != allValueCompact {
t.Errorf("Marshal allValueCompact")
diff(t, b, []byte(allValueCompact))
return
}
b, err = Marshal(pallValue)
if err != nil {
t.Fatalf("Marshal pallValue: %v", err)
}
if string(b) != pallValueCompact {
t.Errorf("Marshal pallValueCompact")
diff(t, b, []byte(pallValueCompact))
return
}
}
var badUTF8 = []struct {
in, out string
}{
{"hello\xffworld", `"hello\ufffdworld"`},
{"", `""`},
{"\xff", `"\ufffd"`},
{"\xff\xff", `"\ufffd\ufffd"`},
{"a\xffb", `"a\ufffdb"`},
{"\xe6\x97\xa5\xe6\x9c\xac\xff\xaa\x9e", `"日本\ufffd\ufffd\ufffd"`},
}
func TestMarshalBadUTF8(t *testing.T) {
for _, tt := range badUTF8 {
b, err := Marshal(tt.in)
if string(b) != tt.out || err != nil {
t.Errorf("Marshal(%q) = %#q, %v, want %#q, nil", tt.in, b, err, tt.out)
}
}
}
func TestMarshalNumberZeroVal(t *testing.T) {
var n Number
out, err := Marshal(n)
if err != nil {
t.Fatal(err)
}
outStr := string(out)
if outStr != "0" {
t.Fatalf("Invalid zero val for Number: %q", outStr)
}
}
func TestMarshalEmbeds(t *testing.T) {
top := &Top{
Level0: 1,
Embed0: Embed0{
Level1b: 2,
Level1c: 3,
},
Embed0a: &Embed0a{
Level1a: 5,
Level1b: 6,
},
Embed0b: &Embed0b{
Level1a: 8,
Level1b: 9,
Level1c: 10,
Level1d: 11,
Level1e: 12,
},
Loop: Loop{
Loop1: 13,
Loop2: 14,
},
Embed0p: Embed0p{
Point: image.Point{X: 15, Y: 16},
},
Embed0q: Embed0q{
Point: Point{Z: 17},
},
embed: embed{
Q: 18,
},
}
b, err := Marshal(top)
if err != nil {
t.Fatal(err)
}
want := "{\"Level0\":1,\"Level1b\":2,\"Level1c\":3,\"Level1a\":5,\"LEVEL1B\":6,\"e\":{\"Level1a\":8,\"Level1b\":9,\"Level1c\":10,\"Level1d\":11,\"x\":12},\"Loop1\":13,\"Loop2\":14,\"X\":15,\"Y\":16,\"Z\":17,\"Q\":18}"
if string(b) != want {
t.Errorf("Wrong marshal result.\n got: %q\nwant: %q", b, want)
}
}
func equalError(a, b error) bool {
if a == nil {
return b == nil
}
if b == nil {
return a == nil
}
return a.Error() == b.Error()
}
func TestUnmarshal(t *testing.T) {
for i, tt := range unmarshalTests {
var scan scanner
in := []byte(tt.in)
if err := checkValid(in, &scan); err != nil {
if !equalError(err, tt.err) {
t.Errorf("#%d: checkValid: %#v", i, err)
continue
}
}
if tt.ptr == nil {
continue
}
// v = new(right-type)
v := reflect.New(reflect.TypeOf(tt.ptr).Elem())
dec := NewDecoder(bytes.NewReader(in))
if tt.useNumber {
dec.UseNumber()
}
if tt.disallowUnknownFields {
dec.DisallowUnknownFields()
}
if err := dec.Decode(v.Interface()); !equalError(err, tt.err) {
t.Errorf("#%d: %v, want %v", i, err, tt.err)
continue
} else if err != nil {
continue
}
if !reflect.DeepEqual(v.Elem().Interface(), tt.out) {
t.Errorf("#%d: mismatch\nhave: %#+v\nwant: %#+v", i, v.Elem().Interface(), tt.out)
data, _ := Marshal(v.Elem().Interface())
println(string(data))
data, _ = Marshal(tt.out)
println(string(data))
continue
}
// Check round trip also decodes correctly.
if tt.err == nil {
enc, err := Marshal(v.Interface())
if err != nil {
t.Errorf("#%d: error re-marshaling: %v", i, err)
continue
}
if tt.golden && !bytes.Equal(enc, in) {
t.Errorf("#%d: remarshal mismatch:\nhave: %s\nwant: %s", i, enc, in)
}
vv := reflect.New(reflect.TypeOf(tt.ptr).Elem())
dec = NewDecoder(bytes.NewReader(enc))
if tt.useNumber {
dec.UseNumber()
}
if err := dec.Decode(vv.Interface()); err != nil {
t.Errorf("#%d: error re-unmarshaling %#q: %v", i, enc, err)
continue
}
if !reflect.DeepEqual(v.Elem().Interface(), vv.Elem().Interface()) {
t.Errorf("#%d: mismatch\nhave: %#+v\nwant: %#+v", i, v.Elem().Interface(), vv.Elem().Interface())
t.Errorf(" In: %q", strings.Map(noSpace, string(in)))
t.Errorf("Marshal: %q", strings.Map(noSpace, string(enc)))
continue
}
}
}
}
func TestUnmarshalMarshal(t *testing.T) {
initBig()
var v interface{}
if err := Unmarshal(jsonBig, &v); err != nil {
t.Fatalf("Unmarshal: %v", err)
}
b, err := Marshal(v)
if err != nil {
t.Fatalf("Marshal: %v", err)
}
if !bytes.Equal(jsonBig, b) {
t.Errorf("Marshal jsonBig")
diff(t, b, jsonBig)
return
}
}
var numberTests = []struct {
in string
i int64
intErr string
f float64
floatErr string
}{
{in: "-1.23e1", intErr: "strconv.ParseInt: parsing \"-1.23e1\": invalid syntax", f: -1.23e1},
{in: "-12", i: -12, f: -12.0},
{in: "1e1000", intErr: "strconv.ParseInt: parsing \"1e1000\": invalid syntax", floatErr: "strconv.ParseFloat: parsing \"1e1000\": value out of range"},
}
// Independent of Decode, basic coverage of the accessors in Number
func TestNumberAccessors(t *testing.T) {
for _, tt := range numberTests {
n := Number(tt.in)
if s := n.String(); s != tt.in {
t.Errorf("Number(%q).String() is %q", tt.in, s)
}
if i, err := n.Int64(); err == nil && tt.intErr == "" && i != tt.i {
t.Errorf("Number(%q).Int64() is %d", tt.in, i)
} else if (err == nil && tt.intErr != "") || (err != nil && err.Error() != tt.intErr) {
t.Errorf("Number(%q).Int64() wanted error %q but got: %v", tt.in, tt.intErr, err)
}
if f, err := n.Float64(); err == nil && tt.floatErr == "" && f != tt.f {
t.Errorf("Number(%q).Float64() is %g", tt.in, f)
} else if (err == nil && tt.floatErr != "") || (err != nil && err.Error() != tt.floatErr) {
t.Errorf("Number(%q).Float64() wanted error %q but got: %v", tt.in, tt.floatErr, err)
}
}
}
func TestLargeByteSlice(t *testing.T) {
s0 := make([]byte, 2000)
for i := range s0 {
s0[i] = byte(i)
}
b, err := Marshal(s0)
if err != nil {
t.Fatalf("Marshal: %v", err)
}
var s1 []byte
if err := Unmarshal(b, &s1); err != nil {
t.Fatalf("Unmarshal: %v", err)
}
if !bytes.Equal(s0, s1) {
t.Errorf("Marshal large byte slice")
diff(t, s0, s1)
}
}
type Xint struct {
X int
}
func TestUnmarshalInterface(t *testing.T) {
var xint Xint
var i interface{} = &xint
if err := Unmarshal([]byte(`{"X":1}`), &i); err != nil {
t.Fatalf("Unmarshal: %v", err)
}
if xint.X != 1 {
t.Fatalf("Did not write to xint")
}
}
func TestUnmarshalPtrPtr(t *testing.T) {
var xint Xint
pxint := &xint
if err := Unmarshal([]byte(`{"X":1}`), &pxint); err != nil {
t.Fatalf("Unmarshal: %v", err)
}
if xint.X != 1 {
t.Fatalf("Did not write to xint")
}
}
func TestEscape(t *testing.T) {
const input = `"foobar"<html>` + " [\u2028 \u2029]"
const expected = `"\"foobar\"\u003chtml\u003e [\u2028 \u2029]"`
b, err := Marshal(input)
if err != nil {
t.Fatalf("Marshal error: %v", err)
}
if s := string(b); s != expected {
t.Errorf("Encoding of [%s]:\n got [%s]\nwant [%s]", input, s, expected)
}
}
// WrongString is a struct that's misusing the ,string modifier.
type WrongString struct {
Message string `json:"result,string"`
}
type wrongStringTest struct {
in, err string
}
var wrongStringTests = []wrongStringTest{
{`{"result":"x"}`, `json: invalid use of ,string struct tag, trying to unmarshal "x" into string`},
{`{"result":"foo"}`, `json: invalid use of ,string struct tag, trying to unmarshal "foo" into string`},
{`{"result":"123"}`, `json: invalid use of ,string struct tag, trying to unmarshal "123" into string`},
{`{"result":123}`, `json: invalid use of ,string struct tag, trying to unmarshal unquoted value into string`},
}
// If people misuse the ,string modifier, the error message should be
// helpful, telling the user that they're doing it wrong.
func TestErrorMessageFromMisusedString(t *testing.T) {
for n, tt := range wrongStringTests {
r := strings.NewReader(tt.in)
var s WrongString
err := NewDecoder(r).Decode(&s)
got := fmt.Sprintf("%v", err)
if got != tt.err {
t.Errorf("%d. got err = %q, want %q", n, got, tt.err)
}
}
}
func noSpace(c rune) rune {
if isSpace(byte(c)) { //only used for ascii
return -1
}
return c
}
type All struct {
Bool bool
Int int
Int8 int8
Int16 int16
Int32 int32
Int64 int64
Uint uint
Uint8 uint8
Uint16 uint16
Uint32 uint32
Uint64 uint64
Uintptr uintptr
Float32 float32
Float64 float64
Foo string `json:"bar"`
Foo2 string `json:"bar2,dummyopt"`
IntStr int64 `json:",string"`
UintptrStr uintptr `json:",string"`
PBool *bool
PInt *int
PInt8 *int8
PInt16 *int16
PInt32 *int32
PInt64 *int64
PUint *uint
PUint8 *uint8
PUint16 *uint16
PUint32 *uint32
PUint64 *uint64
PUintptr *uintptr
PFloat32 *float32
PFloat64 *float64
String string
PString *string
Map map[string]Small
MapP map[string]*Small
PMap *map[string]Small
PMapP *map[string]*Small
EmptyMap map[string]Small
NilMap map[string]Small
Slice []Small
SliceP []*Small
PSlice *[]Small
PSliceP *[]*Small
EmptySlice []Small
NilSlice []Small
StringSlice []string
ByteSlice []byte
Small Small
PSmall *Small
PPSmall **Small
Interface interface{}
PInterface *interface{}
unexported int
}
type Small struct {
Tag string
}
var allValue = All{
Bool: true,
Int: 2,
Int8: 3,
Int16: 4,
Int32: 5,
Int64: 6,
Uint: 7,
Uint8: 8,
Uint16: 9,
Uint32: 10,
Uint64: 11,
Uintptr: 12,
Float32: 14.1,
Float64: 15.1,
Foo: "foo",
Foo2: "foo2",
IntStr: 42,
UintptrStr: 44,
String: "16",
Map: map[string]Small{
"17": {Tag: "tag17"},
"18": {Tag: "tag18"},
},
MapP: map[string]*Small{
"19": {Tag: "tag19"},
"20": nil,
},
EmptyMap: map[string]Small{},
Slice: []Small{{Tag: "tag20"}, {Tag: "tag21"}},
SliceP: []*Small{{Tag: "tag22"}, nil, {Tag: "tag23"}},
EmptySlice: []Small{},
StringSlice: []string{"str24", "str25", "str26"},
ByteSlice: []byte{27, 28, 29},
Small: Small{Tag: "tag30"},
PSmall: &Small{Tag: "tag31"},
Interface: 5.2,
}
var pallValue = All{
PBool: &allValue.Bool,
PInt: &allValue.Int,
PInt8: &allValue.Int8,
PInt16: &allValue.Int16,
PInt32: &allValue.Int32,
PInt64: &allValue.Int64,
PUint: &allValue.Uint,
PUint8: &allValue.Uint8,
PUint16: &allValue.Uint16,
PUint32: &allValue.Uint32,
PUint64: &allValue.Uint64,
PUintptr: &allValue.Uintptr,
PFloat32: &allValue.Float32,
PFloat64: &allValue.Float64,
PString: &allValue.String,
PMap: &allValue.Map,
PMapP: &allValue.MapP,
PSlice: &allValue.Slice,
PSliceP: &allValue.SliceP,
PPSmall: &allValue.PSmall,
PInterface: &allValue.Interface,
}
var allValueIndent = `{
"Bool": true,
"Int": 2,
"Int8": 3,
"Int16": 4,
"Int32": 5,
"Int64": 6,
"Uint": 7,
"Uint8": 8,
"Uint16": 9,
"Uint32": 10,
"Uint64": 11,
"Uintptr": 12,
"Float32": 14.1,
"Float64": 15.1,
"bar": "foo",
"bar2": "foo2",
"IntStr": "42",
"UintptrStr": "44",
"PBool": null,
"PInt": null,
"PInt8": null,
"PInt16": null,
"PInt32": null,
"PInt64": null,
"PUint": null,
"PUint8": null,
"PUint16": null,
"PUint32": null,
"PUint64": null,
"PUintptr": null,
"PFloat32": null,
"PFloat64": null,
"String": "16",
"PString": null,
"Map": {
"17": {
"Tag": "tag17"
},
"18": {
"Tag": "tag18"
}
},
"MapP": {
"19": {
"Tag": "tag19"
},
"20": null
},
"PMap": null,
"PMapP": null,
"EmptyMap": {},
"NilMap": null,
"Slice": [
{
"Tag": "tag20"
},
{
"Tag": "tag21"
}
],
"SliceP": [
{
"Tag": "tag22"
},
null,
{
"Tag": "tag23"
}
],
"PSlice": null,
"PSliceP": null,
"EmptySlice": [],
"NilSlice": null,
"StringSlice": [
"str24",
"str25",
"str26"
],
"ByteSlice": "Gxwd",
"Small": {
"Tag": "tag30"
},
"PSmall": {
"Tag": "tag31"
},
"PPSmall": null,
"Interface": 5.2,
"PInterface": null
}`
var allValueCompact = strings.Map(noSpace, allValueIndent)
var pallValueIndent = `{
"Bool": false,
"Int": 0,
"Int8": 0,
"Int16": 0,
"Int32": 0,
"Int64": 0,
"Uint": 0,
"Uint8": 0,
"Uint16": 0,
"Uint32": 0,
"Uint64": 0,
"Uintptr": 0,
"Float32": 0,
"Float64": 0,
"bar": "",
"bar2": "",
"IntStr": "0",
"UintptrStr": "0",
"PBool": true,
"PInt": 2,
"PInt8": 3,
"PInt16": 4,
"PInt32": 5,
"PInt64": 6,
"PUint": 7,
"PUint8": 8,
"PUint16": 9,
"PUint32": 10,
"PUint64": 11,
"PUintptr": 12,
"PFloat32": 14.1,
"PFloat64": 15.1,
"String": "",
"PString": "16",
"Map": null,
"MapP": null,
"PMap": {
"17": {
"Tag": "tag17"
},
"18": {
"Tag": "tag18"
}
},
"PMapP": {
"19": {
"Tag": "tag19"
},
"20": null
},
"EmptyMap": null,
"NilMap": null,
"Slice": null,
"SliceP": null,
"PSlice": [
{
"Tag": "tag20"
},
{
"Tag": "tag21"
}
],
"PSliceP": [
{
"Tag": "tag22"
},
null,
{
"Tag": "tag23"
}
],
"EmptySlice": null,
"NilSlice": null,
"StringSlice": null,
"ByteSlice": null,
"Small": {
"Tag": ""
},
"PSmall": null,
"PPSmall": {
"Tag": "tag31"
},
"Interface": null,
"PInterface": 5.2
}`
var pallValueCompact = strings.Map(noSpace, pallValueIndent)
func TestRefUnmarshal(t *testing.T) {
type S struct {
// Ref is defined in encode_test.go.
R0 Ref
R1 *Ref
R2 RefText
R3 *RefText
}
want := S{
R0: 12,
R1: new(Ref),
R2: 13,
R3: new(RefText),
}
*want.R1 = 12
*want.R3 = 13
var got S
if err := Unmarshal([]byte(`{"R0":"ref","R1":"ref","R2":"ref","R3":"ref"}`), &got); err != nil {
t.Fatalf("Unmarshal: %v", err)
}
if !reflect.DeepEqual(got, want) {
t.Errorf("got %+v, want %+v", got, want)
}
}
// Test that the empty string doesn't panic decoding when ,string is specified
// Issue 3450
func TestEmptyString(t *testing.T) {
type T2 struct {
Number1 int `json:",string"`
Number2 int `json:",string"`
}
data := `{"Number1":"1", "Number2":""}`
dec := NewDecoder(strings.NewReader(data))
var t2 T2
err := dec.Decode(&t2)
if err == nil {
t.Fatal("Decode: did not return error")
}
if t2.Number1 != 1 {
t.Fatal("Decode: did not set Number1")
}
}
// Test that a null for ,string is not replaced with the previous quoted string (issue 7046).
// It should also not be an error (issue 2540, issue 8587).
func TestNullString(t *testing.T) {
type T struct {
A int `json:",string"`
B int `json:",string"`
C *int `json:",string"`
}
data := []byte(`{"A": "1", "B": null, "C": null}`)
var s T
s.B = 1
s.C = new(int)
*s.C = 2
err := Unmarshal(data, &s)
if err != nil {
t.Fatalf("Unmarshal: %v", err)
}
if s.B != 1 || s.C != nil {
t.Fatalf("after Unmarshal, s.B=%d, s.C=%p, want 1, nil", s.B, s.C)
}
}
func intp(x int) *int {
p := new(int)
*p = x
return p
}
func intpp(x *int) **int {
pp := new(*int)
*pp = x
return pp
}
var interfaceSetTests = []struct {
pre interface{}
json string
post interface{}
}{
{"foo", `"bar"`, "bar"},
{"foo", `2`, 2.0},
{"foo", `true`, true},
{"foo", `null`, nil},
{nil, `null`, nil},
{new(int), `null`, nil},
{(*int)(nil), `null`, nil},
{new(*int), `null`, new(*int)},
{(**int)(nil), `null`, nil},
{intp(1), `null`, nil},
{intpp(nil), `null`, intpp(nil)},
{intpp(intp(1)), `null`, intpp(nil)},
}
func TestInterfaceSet(t *testing.T) {
for _, tt := range interfaceSetTests {
b := struct{ X interface{} }{tt.pre}
blob := `{"X":` + tt.json + `}`
if err := Unmarshal([]byte(blob), &b); err != nil {
t.Errorf("Unmarshal %#q: %v", blob, err)
continue
}
if !reflect.DeepEqual(b.X, tt.post) {
t.Errorf("Unmarshal %#q into %#v: X=%#v, want %#v", blob, tt.pre, b.X, tt.post)
}
}
}
type NullTest struct {
Bool bool
Int int
Int8 int8
Int16 int16
Int32 int32
Int64 int64
Uint uint
Uint8 uint8
Uint16 uint16
Uint32 uint32
Uint64 uint64
Float32 float32
Float64 float64
String string
PBool *bool
Map map[string]string
Slice []string
Interface interface{}
PRaw *RawMessage
PTime *time.Time
PBigInt *big.Int
PText *MustNotUnmarshalText
PBuffer *bytes.Buffer // has methods, just not relevant ones
PStruct *struct{}
Raw RawMessage
Time time.Time
BigInt big.Int
Text MustNotUnmarshalText
Buffer bytes.Buffer
Struct struct{}
}
type NullTestStrings struct {
Bool bool `json:",string"`
Int int `json:",string"`
Int8 int8 `json:",string"`
Int16 int16 `json:",string"`
Int32 int32 `json:",string"`
Int64 int64 `json:",string"`
Uint uint `json:",string"`
Uint8 uint8 `json:",string"`
Uint16 uint16 `json:",string"`
Uint32 uint32 `json:",string"`
Uint64 uint64 `json:",string"`
Float32 float32 `json:",string"`
Float64 float64 `json:",string"`
String string `json:",string"`
PBool *bool `json:",string"`
Map map[string]string `json:",string"`
Slice []string `json:",string"`
Interface interface{} `json:",string"`
PRaw *RawMessage `json:",string"`
PTime *time.Time `json:",string"`
PBigInt *big.Int `json:",string"`
PText *MustNotUnmarshalText `json:",string"`
PBuffer *bytes.Buffer `json:",string"`
PStruct *struct{} `json:",string"`
Raw RawMessage `json:",string"`
Time time.Time `json:",string"`
BigInt big.Int `json:",string"`
Text MustNotUnmarshalText `json:",string"`
Buffer bytes.Buffer `json:",string"`
Struct struct{} `json:",string"`
}
// JSON null values should be ignored for primitives and string values instead of resulting in an error.
// Issue 2540
func TestUnmarshalNulls(t *testing.T) {
// Unmarshal docs:
// The JSON null value unmarshals into an interface, map, pointer, or slice
// by setting that Go value to nil. Because null is often used in JSON to mean
// ``not present,'' unmarshaling a JSON null into any other Go type has no effect
// on the value and produces no error.
jsonData := []byte(`{
"Bool" : null,
"Int" : null,
"Int8" : null,
"Int16" : null,
"Int32" : null,
"Int64" : null,
"Uint" : null,
"Uint8" : null,
"Uint16" : null,
"Uint32" : null,
"Uint64" : null,
"Float32" : null,
"Float64" : null,
"String" : null,
"PBool": null,
"Map": null,
"Slice": null,
"Interface": null,
"PRaw": null,
"PTime": null,
"PBigInt": null,
"PText": null,
"PBuffer": null,
"PStruct": null,
"Raw": null,
"Time": null,
"BigInt": null,
"Text": null,
"Buffer": null,
"Struct": null
}`)
nulls := NullTest{
Bool: true,
Int: 2,
Int8: 3,
Int16: 4,
Int32: 5,
Int64: 6,
Uint: 7,
Uint8: 8,
Uint16: 9,
Uint32: 10,
Uint64: 11,
Float32: 12.1,
Float64: 13.1,
String: "14",
PBool: new(bool),
Map: map[string]string{},
Slice: []string{},
Interface: new(MustNotUnmarshalJSON),
PRaw: new(RawMessage),
PTime: new(time.Time),
PBigInt: new(big.Int),
PText: new(MustNotUnmarshalText),
PStruct: new(struct{}),
PBuffer: new(bytes.Buffer),
Raw: RawMessage("123"),
Time: time.Unix(123456789, 0),
BigInt: *big.NewInt(123),
}
before := nulls.Time.String()
err := Unmarshal(jsonData, &nulls)
if err != nil {
t.Errorf("Unmarshal of null values failed: %v", err)
}
if !nulls.Bool || nulls.Int != 2 || nulls.Int8 != 3 || nulls.Int16 != 4 || nulls.Int32 != 5 || nulls.Int64 != 6 ||
nulls.Uint != 7 || nulls.Uint8 != 8 || nulls.Uint16 != 9 || nulls.Uint32 != 10 || nulls.Uint64 != 11 ||
nulls.Float32 != 12.1 || nulls.Float64 != 13.1 || nulls.String != "14" {
t.Errorf("Unmarshal of null values affected primitives")
}
if nulls.PBool != nil {
t.Errorf("Unmarshal of null did not clear nulls.PBool")
}
if nulls.Map != nil {
t.Errorf("Unmarshal of null did not clear nulls.Map")
}
if nulls.Slice != nil {
t.Errorf("Unmarshal of null did not clear nulls.Slice")
}
if nulls.Interface != nil {
t.Errorf("Unmarshal of null did not clear nulls.Interface")
}
if nulls.PRaw != nil {
t.Errorf("Unmarshal of null did not clear nulls.PRaw")
}
if nulls.PTime != nil {
t.Errorf("Unmarshal of null did not clear nulls.PTime")
}
if nulls.PBigInt != nil {
t.Errorf("Unmarshal of null did not clear nulls.PBigInt")
}
if nulls.PText != nil {
t.Errorf("Unmarshal of null did not clear nulls.PText")
}
if nulls.PBuffer != nil {
t.Errorf("Unmarshal of null did not clear nulls.PBuffer")
}
if nulls.PStruct != nil {
t.Errorf("Unmarshal of null did not clear nulls.PStruct")
}
if string(nulls.Raw) != "null" {
t.Errorf("Unmarshal of RawMessage null did not record null: %v", string(nulls.Raw))
}
if nulls.Time.String() != before {
t.Errorf("Unmarshal of time.Time null set time to %v", nulls.Time.String())
}
if nulls.BigInt.String() != "123" {
t.Errorf("Unmarshal of big.Int null set int to %v", nulls.BigInt.String())
}
}
type MustNotUnmarshalJSON struct{}
func (x MustNotUnmarshalJSON) UnmarshalJSON(data []byte) error {
return errors.New("MustNotUnmarshalJSON was used")
}
type MustNotUnmarshalText struct{}
func (x MustNotUnmarshalText) UnmarshalText(text []byte) error {
return errors.New("MustNotUnmarshalText was used")
}
func TestStringKind(t *testing.T) {
type stringKind string
var m1, m2 map[stringKind]int
m1 = map[stringKind]int{
"foo": 42,
}
data, err := Marshal(m1)
if err != nil {
t.Errorf("Unexpected error marshaling: %v", err)
}
err = Unmarshal(data, &m2)
if err != nil {
t.Errorf("Unexpected error unmarshaling: %v", err)
}
if !reflect.DeepEqual(m1, m2) {
t.Error("Items should be equal after encoding and then decoding")
}
}
// Custom types with []byte as underlying type could not be marshaled
// and then unmarshaled.
// Issue 8962.
func TestByteKind(t *testing.T) {
type byteKind []byte
a := byteKind("hello")
data, err := Marshal(a)
if err != nil {
t.Error(err)
}
var b byteKind
err = Unmarshal(data, &b)
if err != nil {
t.Fatal(err)
}
if !reflect.DeepEqual(a, b) {
t.Errorf("expected %v == %v", a, b)
}
}
// The fix for issue 8962 introduced a regression.
// Issue 12921.
func TestSliceOfCustomByte(t *testing.T) {
type Uint8 uint8
a := []Uint8("hello")
data, err := Marshal(a)
if err != nil {
t.Fatal(err)
}
var b []Uint8
err = Unmarshal(data, &b)
if err != nil {
t.Fatal(err)
}
if !reflect.DeepEqual(a, b) {
t.Fatalf("expected %v == %v", a, b)
}
}
var decodeTypeErrorTests = []struct {
dest interface{}
src string
}{
{new(string), `{"user": "name"}`}, // issue 4628.
{new(error), `{}`}, // issue 4222
{new(error), `[]`},
{new(error), `""`},
{new(error), `123`},
{new(error), `true`},
}
func TestUnmarshalTypeError(t *testing.T) {
for _, item := range decodeTypeErrorTests {
err := Unmarshal([]byte(item.src), item.dest)
if _, ok := err.(*UnmarshalTypeError); !ok {
t.Errorf("expected type error for Unmarshal(%q, type %T): got %T",
item.src, item.dest, err)
}
}
}
var unmarshalSyntaxTests = []string{
"tru",
"fals",
"nul",
"123e",
`"hello`,
`[1,2,3`,
`{"key":1`,
`{"key":1,`,
}
func TestUnmarshalSyntax(t *testing.T) {
var x interface{}
for _, src := range unmarshalSyntaxTests {
err := Unmarshal([]byte(src), &x)
if _, ok := err.(*SyntaxError); !ok {
t.Errorf("expected syntax error for Unmarshal(%q): got %T", src, err)
}
}
}
// Test handling of unexported fields that should be ignored.
// Issue 4660
type unexportedFields struct {
Name string
m map[string]interface{} `json:"-"`
m2 map[string]interface{} `json:"abcd"`
s []int `json:"-"`
}
func TestUnmarshalUnexported(t *testing.T) {
input := `{"Name": "Bob", "m": {"x": 123}, "m2": {"y": 456}, "abcd": {"z": 789}, "s": [2, 3]}`
want := &unexportedFields{Name: "Bob"}
out := &unexportedFields{}
err := Unmarshal([]byte(input), out)
if err != nil {
t.Errorf("got error %v, expected nil", err)
}
if !reflect.DeepEqual(out, want) {
t.Errorf("got %q, want %q", out, want)
}
}
// Time3339 is a time.Time which encodes to and from JSON
// as an RFC 3339 time in UTC.
type Time3339 time.Time
func (t *Time3339) UnmarshalJSON(b []byte) error {
if len(b) < 2 || b[0] != '"' || b[len(b)-1] != '"' {
return fmt.Errorf("types: failed to unmarshal non-string value %q as an RFC 3339 time", b)
}
tm, err := time.Parse(time.RFC3339, string(b[1:len(b)-1]))
if err != nil {
return err
}
*t = Time3339(tm)
return nil
}
func TestUnmarshalJSONLiteralError(t *testing.T) {
var t3 Time3339
err := Unmarshal([]byte(`"0000-00-00T00:00:00Z"`), &t3)
if err == nil {
t.Fatalf("expected error; got time %v", time.Time(t3))
}
if !strings.Contains(err.Error(), "range") {
t.Errorf("got err = %v; want out of range error", err)
}
}
// Test that extra object elements in an array do not result in a
// "data changing underfoot" error.
// Issue 3717
func TestSkipArrayObjects(t *testing.T) {
json := `[{}]`
var dest [0]interface{}
err := Unmarshal([]byte(json), &dest)
if err != nil {
t.Errorf("got error %q, want nil", err)
}
}
// Test semantics of pre-filled struct fields and pre-filled map fields.
// Issue 4900.
func TestPrefilled(t *testing.T) {
ptrToMap := func(m map[string]interface{}) *map[string]interface{} { return &m }
// Values here change, cannot reuse table across runs.
var prefillTests = []struct {
in string
ptr interface{}
out interface{}
}{
{
in: `{"X": 1, "Y": 2}`,
ptr: &XYZ{X: float32(3), Y: int16(4), Z: 1.5},
out: &XYZ{X: float64(1), Y: float64(2), Z: 1.5},
},
{
in: `{"X": 1, "Y": 2}`,
ptr: ptrToMap(map[string]interface{}{"X": float32(3), "Y": int16(4), "Z": 1.5}),
out: ptrToMap(map[string]interface{}{"X": float64(1), "Y": float64(2), "Z": 1.5}),
},
}
for _, tt := range prefillTests {
ptrstr := fmt.Sprintf("%v", tt.ptr)
err := Unmarshal([]byte(tt.in), tt.ptr) // tt.ptr edited here
if err != nil {
t.Errorf("Unmarshal: %v", err)
}
if !reflect.DeepEqual(tt.ptr, tt.out) {
t.Errorf("Unmarshal(%#q, %s): have %v, want %v", tt.in, ptrstr, tt.ptr, tt.out)
}
}
}
var invalidUnmarshalTests = []struct {
v interface{}
want string
}{
{nil, "json: Unmarshal(nil)"},
{struct{}{}, "json: Unmarshal(non-pointer struct {})"},
{(*int)(nil), "json: Unmarshal(nil *int)"},
}
func TestInvalidUnmarshal(t *testing.T) {
buf := []byte(`{"a":"1"}`)
for _, tt := range invalidUnmarshalTests {
err := Unmarshal(buf, tt.v)
if err == nil {
t.Errorf("Unmarshal expecting error, got nil")
continue
}
if got := err.Error(); got != tt.want {
t.Errorf("Unmarshal = %q; want %q", got, tt.want)
}
}
}
var invalidUnmarshalTextTests = []struct {
v interface{}
want string
}{
{nil, "json: Unmarshal(nil)"},
{struct{}{}, "json: Unmarshal(non-pointer struct {})"},
{(*int)(nil), "json: Unmarshal(nil *int)"},
{new(net.IP), "json: cannot unmarshal number into Go value of type *net.IP"},
}
func TestInvalidUnmarshalText(t *testing.T) {
buf := []byte(`123`)
for _, tt := range invalidUnmarshalTextTests {
err := Unmarshal(buf, tt.v)
if err == nil {
t.Errorf("Unmarshal expecting error, got nil")
continue
}
if got := err.Error(); got != tt.want {
t.Errorf("Unmarshal = %q; want %q", got, tt.want)
}
}
}
// Test that string option is ignored for invalid types.
// Issue 9812.
func TestInvalidStringOption(t *testing.T) {
num := 0
item := struct {
T time.Time `json:",string"`
M map[string]string `json:",string"`
S []string `json:",string"`
A [1]string `json:",string"`
I interface{} `json:",string"`
P *int `json:",string"`
}{M: make(map[string]string), S: make([]string, 0), I: num, P: &num}
data, err := Marshal(item)
if err != nil {
t.Fatalf("Marshal: %v", err)
}
err = Unmarshal(data, &item)
if err != nil {
t.Fatalf("Unmarshal: %v", err)
}
}
// Test unmarshal behavior with regards to embedded unexported structs.
//
// (Issue 21357) If the embedded struct is a pointer and is unallocated,
// this returns an error because unmarshal cannot set the field.
//
// (Issue 24152) If the embedded struct is given an explicit name,
// ensure that the normal unmarshal logic does not panic in reflect.
//
// (Issue 28145) If the embedded struct is given an explicit name and has
// exported methods, don't cause a panic trying to get its value.
func TestUnmarshalEmbeddedUnexported(t *testing.T) {
type (
embed1 struct{ Q int }
embed2 struct{ Q int }
embed3 struct {
Q int64 `json:",string"`
}
S1 struct {
*embed1
R int
}
S2 struct {
*embed1
Q int
}
S3 struct {
embed1
R int
}
S4 struct {
*embed1
embed2
}
S5 struct {
*embed3
R int
}
S6 struct {
embed1 `json:"embed1"`
}
S7 struct {
embed1 `json:"embed1"`
embed2
}
S8 struct {
embed1 `json:"embed1"`
embed2 `json:"embed2"`
Q int
}
S9 struct {
unexportedWithMethods `json:"embed"`
}
)
tests := []struct {
in string
ptr interface{}
out interface{}
err error
}{{
// Error since we cannot set S1.embed1, but still able to set S1.R.
in: `{"R":2,"Q":1}`,
ptr: new(S1),
out: &S1{R: 2},
err: fmt.Errorf("json: cannot set embedded pointer to unexported struct: json.embed1"),
}, {
// The top level Q field takes precedence.
in: `{"Q":1}`,
ptr: new(S2),
out: &S2{Q: 1},
}, {
// No issue with non-pointer variant.
in: `{"R":2,"Q":1}`,
ptr: new(S3),
out: &S3{embed1: embed1{Q: 1}, R: 2},
}, {
// No error since both embedded structs have field R, which annihilate each other.
// Thus, no attempt is made at setting S4.embed1.
in: `{"R":2}`,
ptr: new(S4),
out: new(S4),
}, {
// Error since we cannot set S5.embed1, but still able to set S5.R.
in: `{"R":2,"Q":1}`,
ptr: new(S5),
out: &S5{R: 2},
err: fmt.Errorf("json: cannot set embedded pointer to unexported struct: json.embed3"),
}, {
// Issue 24152, ensure decodeState.indirect does not panic.
in: `{"embed1": {"Q": 1}}`,
ptr: new(S6),
out: &S6{embed1{1}},
}, {
// Issue 24153, check that we can still set forwarded fields even in
// the presence of a name conflict.
//
// This relies on obscure behavior of reflect where it is possible
// to set a forwarded exported field on an unexported embedded struct
// even though there is a name conflict, even when it would have been
// impossible to do so according to Go visibility rules.
// Go forbids this because it is ambiguous whether S7.Q refers to
// S7.embed1.Q or S7.embed2.Q. Since embed1 and embed2 are unexported,
// it should be impossible for an external package to set either Q.
//
// It is probably okay for a future reflect change to break this.
in: `{"embed1": {"Q": 1}, "Q": 2}`,
ptr: new(S7),
out: &S7{embed1{1}, embed2{2}},
}, {
// Issue 24153, similar to the S7 case.
in: `{"embed1": {"Q": 1}, "embed2": {"Q": 2}, "Q": 3}`,
ptr: new(S8),
out: &S8{embed1{1}, embed2{2}, 3},
}, {
// Issue 228145, similar to the cases above.
in: `{"embed": {}}`,
ptr: new(S9),
out: &S9{},
}}
for i, tt := range tests {
err := Unmarshal([]byte(tt.in), tt.ptr)
if !equalError(err, tt.err) {
t.Errorf("#%d: %v, want %v", i, err, tt.err)
}
if !reflect.DeepEqual(tt.ptr, tt.out) {
t.Errorf("#%d: mismatch\ngot: %#+v\nwant: %#+v", i, tt.ptr, tt.out)
}
}
}
type unmarshalPanic struct{}
func (unmarshalPanic) UnmarshalJSON([]byte) error { panic(0xdead) }
func TestUnmarshalPanic(t *testing.T) {
defer func() {
if got := recover(); !reflect.DeepEqual(got, 0xdead) {
t.Errorf("panic() = (%T)(%v), want 0xdead", got, got)
}
}()
Unmarshal([]byte("{}"), &unmarshalPanic{})
t.Fatalf("Unmarshal should have panicked")
}