Merge pull request #47 from rvasily/master

correct encode broken utf
This commit is contained in:
Victor Starodub
2016-07-14 14:56:04 +04:00
committed by GitHub
3 changed files with 78 additions and 57 deletions
+58 -37
View File
@@ -4,15 +4,15 @@ package jwriter
import (
"io"
"strconv"
"unicode/utf8"
"github.com/mailru/easyjson/buffer"
)
// Writer is a JSON writer.
type Writer struct {
EscapeLtGt bool
Error error
Buffer buffer.Buffer
Error error
Buffer buffer.Buffer
}
// Size returns the size of the data that was written out.
@@ -198,10 +198,7 @@ func (w *Writer) Bool(v bool) {
}
}
func hex(c byte) byte {
const chars = "0123456789abcdef"
return chars[c&0xf]
}
const chars = "0123456789abcdef"
func (w *Writer) String(s string) {
w.Buffer.AppendByte('"')
@@ -211,41 +208,65 @@ func (w *Writer) String(s string) {
p := 0 // last non-escape symbol
for i := 0; i < len(s); i++ {
c := s[i]
var escape byte
switch c {
case '\t':
escape = 't'
case '\r':
escape = 'r'
case '\n':
escape = 'n'
case '\\':
escape = '\\'
case '"':
escape = '"'
case '<', '>':
if !w.EscapeLtGt {
continue
for i := 0; i < len(s); {
// single-with character
if c := s[i]; c < utf8.RuneSelf {
var escape byte
switch c {
case '\t':
escape = 't'
case '\r':
escape = 'r'
case '\n':
escape = 'n'
case '\\':
escape = '\\'
case '"':
escape = '"'
case '<', '>':
// do nothing
default:
if c >= 0x20 {
// no escaping is required
i++
continue
}
}
default:
if c >= 0x20 {
// no escaping is required
continue
if escape != 0 {
w.Buffer.AppendString(s[p:i])
w.Buffer.AppendByte('\\')
w.Buffer.AppendByte(escape)
} else {
w.Buffer.AppendString(s[p:i])
w.Buffer.AppendString(`\u00`)
w.Buffer.AppendByte(chars[c>>4])
w.Buffer.AppendByte(chars[c&0xf])
}
i++
p = i
continue
}
if escape != 0 {
// broken utf
runeValue, runeWidth := utf8.DecodeRuneInString(s[i:])
if runeValue == utf8.RuneError && runeWidth == 1 {
w.Buffer.AppendString(s[p:i])
w.Buffer.AppendByte('\\')
w.Buffer.AppendByte(escape)
} else {
w.Buffer.AppendString(s[p:i])
w.Buffer.AppendString(`\u00`)
w.Buffer.AppendByte(hex(c >> 4))
w.Buffer.AppendByte(hex(c))
w.Buffer.AppendString(`\ufffd`)
i++
p = i
continue
}
p = i + 1
// jsonp stuff - tab separator and line separator
if runeValue == '\u2028' || runeValue == '\u2029' {
w.Buffer.AppendString(s[p:i])
w.Buffer.AppendString(`\u202`)
w.Buffer.AppendByte(chars[runeValue&0xf])
i += runeWidth
p = i
continue
}
i += runeWidth
}
w.Buffer.AppendString(s[p:])
w.Buffer.AppendByte('"')
+20 -16
View File
@@ -105,24 +105,28 @@ func TestParseNull(t *testing.T) {
}
}
var testCasesEncodeLtGt = []struct {
Writer *jwriter.Writer
Encoded string
var testSpecialCases = []struct {
EncodedString string
Value string
}{
{&jwriter.Writer{
EscapeLtGt: false,
}, encodeLtGtFalseWantString},
{&jwriter.Writer{
EscapeLtGt: true,
}, encodeLtGtTrueWantString},
{`"Username \u003cuser@example.com\u003e"`, `Username <user@example.com>`},
{`"Username\ufffd"`, "Username\xc5"},
{`"тестzтест"`, "тестzтест"},
{`"тест\ufffdтест"`, "тест\xc5тест"},
{`"绿茶"`, "绿茶"},
{`"绿\ufffd茶"`, "绿\xc5茶"},
{`"тест\u2028"`, "тест\xE2\x80\xA8"},
{`"\\\r\n\t\""`, "\\\r\n\t\""},
{`"ü"`, "ü"},
}
func TestEncodeLtGt(t *testing.T) {
for i, test := range testCasesEncodeLtGt {
test.Writer.String(encodeLtGtString)
got := string(test.Writer.Buffer.BuildBytes())
if got != test.Encoded {
t.Errorf("[%d] Encoded() = %+v; want %+v", i, got, test.Encoded)
func TestSpecialCases(t *testing.T) {
for i, test := range testSpecialCases {
w := jwriter.Writer{}
w.String(test.Value)
got := string(w.Buffer.BuildBytes())
if got != test.EncodedString {
t.Errorf("[%d] Encoded() = %+v; want %+v", i, got, test.EncodedString)
}
}
}
}
-4
View File
@@ -530,7 +530,3 @@ type RequiredOptionalStruct struct {
FirstName string `json:"first_name,required"`
Lastname string `json:"last_name"`
}
var encodeLtGtString = `Username <user@example.com>`
var encodeLtGtFalseWantString = `"Username <user@example.com>"`
var encodeLtGtTrueWantString = `"Username \u003cuser@example.com\u003e"`