encoding/json: speed up tokenization of literals

Decoder.Decode and Unmarshal actually scan the input bytes twice - the
first time to check for syntax errors and the length of the value, and
the second to perform the decoding.

It's in the second scan that we actually tokenize the bytes. Since
syntax errors aren't a possibility, we can take shortcuts.

In particular, literals such as quoted strings are very common in JSON,
so we can avoid a lot of work by special casing them.

name                  old time/op    new time/op    delta
CodeDecoder-8           10.3ms ± 1%     9.1ms ± 0%  -11.89%  (p=0.002 n=6+6)
UnicodeDecoder-8         342ns ± 0%     283ns ± 0%  -17.25%  (p=0.000 n=6+5)
DecoderStream-8          239ns ± 0%     230ns ± 0%   -3.90%  (p=0.000 n=6+5)
CodeUnmarshal-8         11.0ms ± 0%     9.8ms ± 0%  -11.45%  (p=0.002 n=6+6)
CodeUnmarshalReuse-8    10.3ms ± 0%     9.0ms ± 0%  -12.72%  (p=0.004 n=5+6)
UnmarshalString-8        104ns ± 0%      92ns ± 0%  -11.35%  (p=0.002 n=6+6)
UnmarshalFloat64-8      93.2ns ± 0%    87.6ns ± 0%   -6.01%  (p=0.010 n=6+4)
UnmarshalInt64-8        74.5ns ± 0%    71.5ns ± 0%   -3.91%  (p=0.000 n=5+6)

name                  old speed      new speed      delta
CodeDecoder-8          189MB/s ± 1%   214MB/s ± 0%  +13.50%  (p=0.002 n=6+6)
UnicodeDecoder-8      40.9MB/s ± 0%  49.5MB/s ± 0%  +20.96%  (p=0.002 n=6+6)
CodeUnmarshal-8        176MB/s ± 0%   199MB/s ± 0%  +12.93%  (p=0.002 n=6+6)

Updates #28923.

Change-Id: I7a5e2aef51bd4ddf2004aad24210f6f50e01eaeb
Reviewed-on: https://go-review.googlesource.com/c/go/+/151042
Run-TryBot: Daniel Martí <mvdan@mvdan.cc>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

src/encoding/json/decode.go

@@ -361,6 +361,52 @@ func (d *decodeState) scanWhile(op int) {
 	d.opcode = d.scan.eof()
 }
 
+// rescanLiteral is similar to scanWhile(scanContinue), but it specialises the
+// common case where we're decoding a literal. The decoder scans the input
+// twice, once for syntax errors and to check the length of the value, and the
+// second to perform the decoding.
+//
+// Only in the second step do we use decodeState to tokenize literals, so we
+// know there aren't any syntax errors. We can take advantage of that knowledge,
+// and scan a literal's bytes much more quickly.
+func (d *decodeState) rescanLiteral() {
+	data, i := d.data, d.off
+Switch:
+	switch data[i-1] {
+	case '"': // string
+		for ; i < len(data); i++ {
+			switch data[i] {
+			case '\\':
+				i++ // escaped char
+			case '"':
+				i++ // tokenize the closing quote too
+				break Switch
+			}
+		}
+	case '0', '1', '2', '3', '4', '5', '6', '7', '8', '9', '-': // number
+		for ; i < len(data); i++ {
+			switch data[i] {
+			case '0', '1', '2', '3', '4', '5', '6', '7', '8', '9',
+				'.', 'e', 'E', '+', '-':
+			default:
+				break Switch
+			}
+		}
+	case 't': // true
+		i += len("rue")
+	case 'f': // false
+		i += len("alse")
+	case 'n': // null
+		i += len("ull")
+	}
+	if i < len(data) {
+		d.opcode = stateEndValue(&d.scan, data[i])
+	} else {
+		d.opcode = scanEnd
+	}
+	d.off = i + 1
+}
+
 // value consumes a JSON value from d.data[d.off-1:], decoding into v, and
 // reads the following byte ahead. If v is invalid, the value is discarded.
 // The first byte of the value has been read already.
@@ -392,7 +438,7 @@ func (d *decodeState) value(v reflect.Value) error {
 	case scanBeginLiteral:
 		// All bytes inside literal return scanContinue op code.
 		start := d.readIndex()
-		d.scanWhile(scanContinue)
+		d.rescanLiteral()
 
 		if v.IsValid() {
 			if err := d.literalStore(d.data[start:d.readIndex()], v, false); err != nil {
@@ -677,7 +723,7 @@ func (d *decodeState) object(v reflect.Value) error {
 
 		// Read key.
 		start := d.readIndex()
-		d.scanWhile(scanContinue)
+		d.rescanLiteral()
 		item := d.data[start:d.readIndex()]
 		key, ok := unquoteBytes(item)
 		if !ok {
@@ -1083,7 +1129,7 @@ func (d *decodeState) objectInterface() map[string]interface{} {
 
 		// Read string key.
 		start := d.readIndex()
-		d.scanWhile(scanContinue)
+		d.rescanLiteral()
 		item := d.data[start:d.readIndex()]
 		key, ok := unquote(item)
 		if !ok {
@@ -1122,7 +1168,7 @@ func (d *decodeState) objectInterface() map[string]interface{} {
 func (d *decodeState) literalInterface() interface{} {
 	// All bytes inside literal return scanContinue op code.
 	start := d.readIndex()
-	d.scanWhile(scanContinue)
+	d.rescanLiteral()
 
 	item := d.data[start:d.readIndex()]