summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
Diffstat
-rw-r--r--compat/json/compat.go61
-rw-r--r--decode_scan.go9
-rw-r--r--decode_scan_test.go5
-rw-r--r--encode.go61
-rw-r--r--errors.go10
-rw-r--r--misc.go21
-rw-r--r--parse.go6
-rw-r--r--reencode.go17
-rw-r--r--test_export.go6
9 files changed, 126 insertions, 70 deletions
diff --git a/compat/json/compat.go b/compat/json/compat.go
index 37a0d5b..8dcb266 100644
--- a/compat/json/compat.go
+++ b/compat/json/compat.go
@@ -71,40 +71,46 @@ func Marshal(v any) ([]byte, error) {
}
type Encoder struct {
- inner lowmemjson.ReEncoder
+ encoder lowmemjson.Encoder
+ formatter lowmemjson.ReEncoder
}
func NewEncoder(w io.Writer) *Encoder {
- return &Encoder{
- inner: lowmemjson.ReEncoder{
- Out: w,
- Compact: true,
+ ret := &Encoder{
+ formatter: lowmemjson.ReEncoder{
+ Out: w,
+
+ AllowMultipleValues: true,
+
+ Compact: true,
+ ForceTrailingNewlines: true,
},
}
+ ret.encoder.W = &ret.formatter
+ return ret
}
func (enc *Encoder) Encode(v any) error {
- if err := convertEncodeError(lowmemjson.Encode(&enc.inner, v)); err != nil {
- return err
- }
- if err := convertEncodeError(enc.inner.Close()); err != nil {
- return err
+ err := enc.encoder.Encode(v)
+ if err == nil {
+ err = enc.formatter.Close()
}
- return nil
+ return convertEncodeError(err)
}
func (enc *Encoder) SetEscapeHTML(on bool) {
- if on {
- enc.inner.BackslashEscape = nil
- } else {
- enc.inner.BackslashEscape = lowmemjson.EscapeJSSafe
+ var escaper lowmemjson.BackslashEscaper
+ if !on {
+ escaper = lowmemjson.EscapeDefaultNonHTMLSafe
}
+ enc.encoder.BackslashEscape = escaper
+ enc.formatter.BackslashEscape = escaper
}
func (enc *Encoder) SetIndent(prefix, indent string) {
- enc.inner.Compact = prefix == "" && indent == ""
- enc.inner.Prefix = prefix
- enc.inner.Indent = indent
+ enc.formatter.Compact = prefix == "" && indent == ""
+ enc.formatter.Prefix = prefix
+ enc.formatter.Indent = indent
}
// ReEncode wrappers /////////////////////////////////////////////////
@@ -169,12 +175,19 @@ func convertDecodeError(err error) error {
if derr, ok := err.(*lowmemjson.DecodeError); ok {
switch terr := derr.Err.(type) {
case *lowmemjson.DecodeSyntaxError:
- err = &SyntaxError{
- msg: terr.Err.Error(),
- Offset: terr.Offset,
- }
- if errors.Is(terr.Err, io.ErrUnexpectedEOF) {
- err.(*SyntaxError).msg = "unexpected end of JSON input"
+ switch {
+ case errors.Is(terr.Err, io.EOF):
+ err = io.EOF
+ case errors.Is(terr.Err, io.ErrUnexpectedEOF):
+ err = &SyntaxError{
+ msg: "unexpected end of JSON input",
+ Offset: terr.Offset,
+ }
+ default:
+ err = &SyntaxError{
+ msg: terr.Err.Error(),
+ Offset: terr.Offset,
+ }
}
case *lowmemjson.DecodeTypeError:
if typeErr, ok := terr.Err.(*json.UnmarshalTypeError); ok {
diff --git a/decode_scan.go b/decode_scan.go
index 9fa6181..2830d0b 100644
--- a/decode_scan.go
+++ b/decode_scan.go
@@ -13,9 +13,10 @@ type runeTypeScanner interface {
// The returned error is a *ReadError, a *SyntaxError, or nil.
// An EOF condition is represented as one of:
//
- // end of value but not file: (_, >0, RuneTypeEOF, nil)
- // end of both value and file: (_, 0, RuneTypeEOF, nil)
- // end of file but not value: (_, 0, RuneTypeError, &DecodeSyntaxError{Offset: offset: Err: io.ErrUnexepctedEOF})
+ // end of value but not file: (_, >0, RuneTypeEOF, nil)
+ // end of both value and file: (_, 0, RuneTypeEOF, nil)
+ // end of file in middle of value: (_, 0, RuneTypeError, &DecodeSyntaxError{Offset: offset: Err: io.ErrUnexepctedEOF})
+ // end of file at start of value: (_, 0, RuneTypeError, &DecodeSyntaxError{Offset: offset: Err: io.EOF})
ReadRuneType() (rune, int, RuneType, error)
// The returned error is a *DecodeReadError, a *DecodeSyntaxError, io.EOF, or nil.
ReadRune() (rune, int, error)
@@ -127,8 +128,6 @@ func (sc *runeTypeScannerImpl) ReadRune() (rune, int, error) {
}
}
-var ErrInvalidUnreadRune = errors.New("lowmemjson: invalid use of UnreadRune")
-
// UnreadRune undoes a call to .ReadRune() or .ReadRuneType().
//
// If the last call to .ReadRune() or .ReadRuneType() has already been
diff --git a/decode_scan_test.go b/decode_scan_test.go
index 6fd9369..27b60c0 100644
--- a/decode_scan_test.go
+++ b/decode_scan_test.go
@@ -152,6 +152,11 @@ func TestRuneTypeScanner(t *testing.T) {
{0, 0, RuneTypeError, &DecodeSyntaxError{Offset: 1, Err: io.ErrUnexpectedEOF}},
{0, 0, RuneTypeError, &DecodeSyntaxError{Offset: 1, Err: io.ErrUnexpectedEOF}},
}},
+ "empty": {``, ``, []ReadRuneTypeResult{
+ {0, 0, RuneTypeError, &DecodeSyntaxError{Offset: 0, Err: io.EOF}},
+ {0, 0, RuneTypeError, &DecodeSyntaxError{Offset: 0, Err: io.EOF}},
+ {0, 0, RuneTypeError, &DecodeSyntaxError{Offset: 0, Err: io.EOF}},
+ }},
}
testRuneTypeScanner(t, testcases, func(reader io.RuneReader) runeTypeScanner {
return &runeTypeScannerImpl{
diff --git a/encode.go b/encode.go
index f6f8f0e..a77d8aa 100644
--- a/encode.go
+++ b/encode.go
@@ -39,7 +39,12 @@ func encodeWriteString(w io.Writer, str string) {
}
}
-func Encode(w io.Writer, obj any) (err error) {
+type Encoder struct {
+ W io.Writer
+ BackslashEscape BackslashEscaper
+}
+
+func (enc *Encoder) Encode(obj any) (err error) {
defer func() {
if r := recover(); r != nil {
if e, ok := r.(encodeError); ok {
@@ -49,13 +54,18 @@ func Encode(w io.Writer, obj any) (err error) {
}
}
}()
- encode(w, reflect.ValueOf(obj), false, 0, map[any]struct{}{})
- if f, ok := w.(interface{ Flush() error }); ok {
+ encode(enc.W, reflect.ValueOf(obj), enc.BackslashEscape, false, 0, map[any]struct{}{})
+ if f, ok := enc.W.(interface{ Flush() error }); ok {
return f.Flush()
}
return nil
}
+func Encode(w io.Writer, obj any) (err error) {
+ enc := &Encoder{W: w}
+ return enc.Encode(obj)
+}
+
var (
encodableType = reflect.TypeOf((*Encodable)(nil)).Elem()
jsonMarshalerType = reflect.TypeOf((*json.Marshaler)(nil)).Elem()
@@ -64,7 +74,7 @@ var (
const startDetectingCyclesAfter = 1000
-func encode(w io.Writer, val reflect.Value, quote bool, cycleDepth uint, cycleSeen map[any]struct{}) {
+func encode(w io.Writer, val reflect.Value, escaper BackslashEscaper, quote bool, cycleDepth uint, cycleSeen map[any]struct{}) {
if !val.IsValid() {
encodeWriteString(w, "null")
return
@@ -84,7 +94,7 @@ func encode(w io.Writer, val reflect.Value, quote bool, cycleDepth uint, cycleSe
encodeWriteString(w, "null")
return
}
- validator := &ReEncoder{Out: w}
+ validator := &ReEncoder{Out: w, BackslashEscape: escaper}
if err := obj.EncodeJSON(validator); err != nil {
panic(encodeError{&EncodeMethodError{
Type: val.Type(),
@@ -117,7 +127,7 @@ func encode(w io.Writer, val reflect.Value, quote bool, cycleDepth uint, cycleSe
SourceFunc: "MarshalJSON",
}})
}
- validator := &ReEncoder{Out: w}
+ validator := &ReEncoder{Out: w, BackslashEscape: escaper}
if _, err := validator.Write(dat); err != nil {
panic(encodeError{err})
}
@@ -146,7 +156,7 @@ func encode(w io.Writer, val reflect.Value, quote bool, cycleDepth uint, cycleSe
SourceFunc: "MarshalText",
}})
}
- encodeStringFromBytes(w, text)
+ encodeStringFromBytes(w, escaper, text)
default:
switch val.Kind() {
@@ -202,17 +212,17 @@ func encode(w io.Writer, val reflect.Value, quote bool, cycleDepth uint, cycleSe
} else {
if quote {
var buf bytes.Buffer
- encodeStringFromString(&buf, val.String())
- encodeStringFromBytes(w, buf.Bytes())
+ encodeStringFromString(&buf, escaper, val.String())
+ encodeStringFromBytes(w, escaper, buf.Bytes())
} else {
- encodeStringFromString(w, val.String())
+ encodeStringFromString(w, escaper, val.String())
}
}
case reflect.Interface:
if val.IsNil() {
encodeWriteString(w, "null")
} else {
- encode(w, val.Elem(), quote, cycleDepth, cycleSeen)
+ encode(w, val.Elem(), escaper, quote, cycleDepth, cycleSeen)
}
case reflect.Struct:
encodeWriteByte(w, '{')
@@ -229,9 +239,9 @@ func encode(w io.Writer, val reflect.Value, quote bool, cycleDepth uint, cycleSe
encodeWriteByte(w, ',')
}
empty = false
- encodeStringFromString(w, field.Name)
+ encodeStringFromString(w, escaper, field.Name)
encodeWriteByte(w, ':')
- encode(w, fVal, field.Quote, cycleDepth, cycleSeen)
+ encode(w, fVal, escaper, field.Quote, cycleDepth, cycleSeen)
}
encodeWriteByte(w, '}')
case reflect.Map:
@@ -263,15 +273,16 @@ func encode(w io.Writer, val reflect.Value, quote bool, cycleDepth uint, cycleSe
kvs := make([]kv, val.Len())
iter := val.MapRange()
for i := 0; iter.Next(); i++ {
+ // TODO: Avoid buffering the map key
var k strings.Builder
- encode(&k, iter.Key(), false, cycleDepth, cycleSeen)
+ encode(&k, iter.Key(), escaper, false, cycleDepth, cycleSeen)
kStr := k.String()
if kStr == "null" {
kStr = `""`
}
if !strings.HasPrefix(kStr, `"`) {
k.Reset()
- encodeStringFromString(&k, kStr)
+ encodeStringFromString(&k, escaper, kStr)
kStr = k.String()
}
kvs[i].K = kStr
@@ -287,7 +298,7 @@ func encode(w io.Writer, val reflect.Value, quote bool, cycleDepth uint, cycleSe
}
encodeWriteString(w, kv.K)
encodeWriteByte(w, ':')
- encode(w, kv.V, false, cycleDepth, cycleSeen)
+ encode(w, kv.V, escaper, false, cycleDepth, cycleSeen)
}
encodeWriteByte(w, '}')
case reflect.Slice:
@@ -341,10 +352,10 @@ func encode(w io.Writer, val reflect.Value, quote bool, cycleDepth uint, cycleSe
cycleSeen[ptr] = struct{}{}
defer delete(cycleSeen, ptr)
}
- encodeArray(w, val, cycleDepth, cycleSeen)
+ encodeArray(w, val, escaper, cycleDepth, cycleSeen)
}
case reflect.Array:
- encodeArray(w, val, cycleDepth, cycleSeen)
+ encodeArray(w, val, escaper, cycleDepth, cycleSeen)
case reflect.Pointer:
if val.IsNil() {
encodeWriteString(w, "null")
@@ -360,7 +371,7 @@ func encode(w io.Writer, val reflect.Value, quote bool, cycleDepth uint, cycleSe
cycleSeen[ptr] = struct{}{}
defer delete(cycleSeen, ptr)
}
- encode(w, val.Elem(), quote, cycleDepth, cycleSeen)
+ encode(w, val.Elem(), escaper, quote, cycleDepth, cycleSeen)
}
default:
panic(encodeError{&EncodeTypeError{
@@ -370,21 +381,21 @@ func encode(w io.Writer, val reflect.Value, quote bool, cycleDepth uint, cycleSe
}
}
-func encodeStringFromString(w io.Writer, str string) {
+func encodeStringFromString(w io.Writer, escaper BackslashEscaper, str string) {
encodeWriteByte(w, '"')
for _, c := range str {
- if _, err := writeStringChar(w, c, BackslashEscapeNone, nil); err != nil {
+ if _, err := writeStringChar(w, c, BackslashEscapeNone, escaper); err != nil {
panic(encodeError{err})
}
}
encodeWriteByte(w, '"')
}
-func encodeStringFromBytes(w io.Writer, str []byte) {
+func encodeStringFromBytes(w io.Writer, escaper BackslashEscaper, str []byte) {
encodeWriteByte(w, '"')
for i := 0; i < len(str); {
c, size := utf8.DecodeRune(str[i:])
- if _, err := writeStringChar(w, c, BackslashEscapeNone, nil); err != nil {
+ if _, err := writeStringChar(w, c, BackslashEscapeNone, escaper); err != nil {
panic(encodeError{err})
}
i += size
@@ -392,14 +403,14 @@ func encodeStringFromBytes(w io.Writer, str []byte) {
encodeWriteByte(w, '"')
}
-func encodeArray(w io.Writer, val reflect.Value, cycleDepth uint, cycleSeen map[any]struct{}) {
+func encodeArray(w io.Writer, val reflect.Value, escaper BackslashEscaper, cycleDepth uint, cycleSeen map[any]struct{}) {
encodeWriteByte(w, '[')
n := val.Len()
for i := 0; i < n; i++ {
if i > 0 {
encodeWriteByte(w, ',')
}
- encode(w, val.Index(i), false, cycleDepth, cycleSeen)
+ encode(w, val.Index(i), escaper, false, cycleDepth, cycleSeen)
}
encodeWriteByte(w, ']')
}
diff --git a/errors.go b/errors.go
index 3978d62..a2d88bf 100644
--- a/errors.go
+++ b/errors.go
@@ -12,12 +12,11 @@ import (
"strings"
)
-// parser errors ///////////////////////////////////////////////////////////////////////////////////
+var (
+ ErrInvalidUnreadRune = errors.New("lowmemjson: invalid use of UnreadRune")
+)
-type ParseError struct {
- Err error
- Offset int64
-}
+// parser errors ///////////////////////////////////////////////////////////////////////////////////
var (
ErrParserExceededMaxDepth = errors.New("exceeded max depth")
@@ -155,3 +154,4 @@ type ReEncodeSyntaxError struct {
func (e *ReEncodeSyntaxError) Error() string {
return fmt.Sprintf("json: syntax error at input byte %v: %v", e.Offset, e.Err)
}
+func (e *ReEncodeSyntaxError) Unwrap() error { return e.Err }
diff --git a/misc.go b/misc.go
index dce9dac..84d2ae1 100644
--- a/misc.go
+++ b/misc.go
@@ -67,7 +67,13 @@ const (
BackslashEscapeUnicode
)
-func EscapeJSSafe(c rune, _ BackslashEscapeMode) BackslashEscapeMode {
+type BackslashEscaper = func(rune, BackslashEscapeMode) BackslashEscapeMode
+
+func EscapePreserve(_ rune, wasEscaped BackslashEscapeMode) BackslashEscapeMode {
+ return wasEscaped
+}
+
+func EscapeJSSafe(c rune, wasEscaped BackslashEscapeMode) BackslashEscapeMode {
// JSON is notionally a JS subset, but that's not actually
// true.
//
@@ -76,7 +82,7 @@ func EscapeJSSafe(c rune, _ BackslashEscapeMode) BackslashEscapeMode {
case '\u2028', '\u2029':
return BackslashEscapeUnicode
default:
- return BackslashEscapeNone
+ return wasEscaped
}
}
@@ -98,8 +104,13 @@ func EscapeDefault(c rune, wasEscaped BackslashEscapeMode) BackslashEscapeMode {
}
}
-func EscapePreserve(_ rune, wasEscaped BackslashEscapeMode) BackslashEscapeMode {
- return wasEscaped
+func EscapeDefaultNonHTMLSafe(c rune, wasEscaped BackslashEscapeMode) BackslashEscapeMode {
+ switch c {
+ case '\b', '\f', utf8.RuneError:
+ return BackslashEscapeUnicode
+ default:
+ return EscapeJSSafe(c, wasEscaped)
+ }
}
func writeStringUnicodeEscape(w io.Writer, c rune) (int, error) {
@@ -134,7 +145,7 @@ func writeStringShortEscape(w io.Writer, c rune) (int, error) {
buf := [2]byte{'\\', b}
return w.Write(buf[:])
}
-func writeStringChar(w io.Writer, c rune, wasEscaped BackslashEscapeMode, escaper func(rune, BackslashEscapeMode) BackslashEscapeMode) (int, error) {
+func writeStringChar(w io.Writer, c rune, wasEscaped BackslashEscapeMode, escaper BackslashEscaper) (int, error) {
if escaper == nil {
escaper = EscapeDefault
}
diff --git a/parse.go b/parse.go
index 954d3dc..3fa6978 100644
--- a/parse.go
+++ b/parse.go
@@ -327,10 +327,14 @@ func (par *Parser) HandleEOF() (RuneType, error) {
case 0:
return RuneTypeEOF, nil
case 1:
- if par.stack[0].IsNumber() {
+ switch {
+ case par.stack[0].IsNumber():
if _, err := par.HandleRune('\n'); err == nil {
return RuneTypeEOF, nil
}
+ case par.stack[0] == RuneTypeError:
+ par.err = io.EOF
+ return RuneTypeError, par.err
}
fallthrough
default:
diff --git a/reencode.go b/reencode.go
index 7c5ce52..bcb3932 100644
--- a/reencode.go
+++ b/reencode.go
@@ -23,13 +23,16 @@ type ReEncoder struct {
Indent string
// String to put before indents.
Prefix string
+ // Whether to emit a newline after each top-level value, even
+ // if it could unambiguously be omitted.
+ ForceTrailingNewlines bool
// Returns whether a given character in a string should be
// backslash-escaped. The bool argument is whether it was
// \u-escaped in the input. This does not affect characters
// that must or must-not be escaped to be valid JSON.
//
// If not set, then EscapeDefault is used.
- BackslashEscape func(rune, BackslashEscapeMode) BackslashEscapeMode
+ BackslashEscape BackslashEscaper
// state: .Write's utf8-decoding buffer
buf [utf8.UTFMax]byte
@@ -267,8 +270,16 @@ func (enc *ReEncoder) handleRune(c rune, t RuneType) error {
rune(enc.uhex[3])<<0
return enc.emit(writeStringChar(enc.Out, c, BackslashEscapeUnicode, enc.BackslashEscape))
- case RuneTypeError: // EOF
- return nil
+ case RuneTypeEOF: // start of next top-level value
+ if !enc.ForceTrailingNewlines && (enc.Compact || enc.Indent == "") && !enc.lastNonSpace.IsNumber() {
+ return nil
+ }
+ return enc.emitByte('\n')
+ case RuneTypeError: // .Close()
+ if !enc.ForceTrailingNewlines {
+ return nil
+ }
+ return enc.emitByte('\n')
default:
return enc.emitByte(byte(c))
}
diff --git a/test_export.go b/test_export.go
index ccac2b2..76d29d2 100644
--- a/test_export.go
+++ b/test_export.go
@@ -5,12 +5,14 @@
package lowmemjson
import (
+ "io"
+
"git.lukeshu.com/go/lowmemjson/internal"
)
func init() {
- internal.EncodeStringFromString = encodeStringFromString
- internal.EncodeStringFromBytes = encodeStringFromBytes
+ internal.EncodeStringFromString = func(w io.Writer, s string) { encodeStringFromString(w, nil, s) }
+ internal.EncodeStringFromBytes = func(w io.Writer, s []byte) { encodeStringFromBytes(w, nil, s) }
}
var parseTag = internal.ParseTag
generated by cgit v1.1-4-g5e80 at 2024-06-26 18:02:23 +0000