rlp: add support for optional struct fields (ethereum#22832)

This adds support for a new struct tag "optional". Using this tag, structs used
for RLP encoding/decoding can be extended in a backwards-compatible way,
by adding new fields at the end.

rlp/decode.go

@@ -229,7 +229,7 @@ func decodeBigInt(s *Stream, val reflect.Value) error {
 		i = new(big.Int)
 		val.Set(reflect.ValueOf(i))
 	}
-	// Reject leading zero bytes
+	// Reject leading zero bytes.
 	if len(b) > 0 && b[0] == 0 {
 		return wrapStreamError(ErrCanonInt, val.Type())
 	}
@@ -394,9 +394,16 @@ func makeStructDecoder(typ reflect.Type) (decoder, error) {
 		if _, err := s.List(); err != nil {
 			return wrapStreamError(err, typ)
 		}
-		for _, f := range fields {
+		for i, f := range fields {
 			err := f.info.decoder(s, val.Field(f.index))
 			if err == EOL {
+				if f.optional {
+					// The field is optional, so reaching the end of the list before
+					// reaching the last field is acceptable. All remaining undecoded
+					// fields are zeroed.
+					zeroFields(val, fields[i:])
+					break
+				}
 				return &decodeError{msg: "too few elements", typ: typ}
 			} else if err != nil {
 				return addErrorContext(err, "."+typ.Field(f.index).Name)
@@ -407,6 +414,13 @@ func makeStructDecoder(typ reflect.Type) (decoder, error) {
 	return dec, nil
 }
 
+func zeroFields(structval reflect.Value, fields []field) {
+	for _, f := range fields {
+		fv := structval.Field(f.index)
+		fv.Set(reflect.Zero(fv.Type()))
+	}
+}
+
 // makePtrDecoder creates a decoder that decodes into the pointer's element type.
 func makePtrDecoder(typ reflect.Type, tag tags) (decoder, error) {
 	etype := typ.Elem()

rlp/decode_test.go

@@ -369,19 +369,46 @@ type intField struct {
 	X int
 }
 
+type optionalFields struct {
+	A uint
+	B uint `rlp:"optional"`
+	C uint `rlp:"optional"`
+}
+
+type optionalAndTailField struct {
+	A    uint
+	B    uint   `rlp:"optional"`
+	Tail []uint `rlp:"tail"`
+}
+
+type optionalBigIntField struct {
+	A uint
+	B *big.Int `rlp:"optional"`
+}
+
+type optionalPtrField struct {
+	A uint
+	B *[3]byte `rlp:"optional"`
+}
+
+type optionalPtrFieldNil struct {
+	A uint
+	B *[3]byte `rlp:"optional,nil"`
+}
+
+type ignoredField struct {
+	A uint
+	B uint `rlp:"-"`
+	C uint
+}
+
 var (
 	veryBigInt = big.NewInt(0).Add(
 		big.NewInt(0).Lsh(big.NewInt(0xFFFFFFFFFFFFFF), 16),
 		big.NewInt(0xFFFF),
 	)
 )
 
-type hasIgnoredField struct {
-	A uint
-	B uint `rlp:"-"`
-	C uint
-}
-
 var decodeTests = []decodeTest{
 	// booleans
 	{input: "01", ptr: new(bool), value: true},
@@ -551,8 +578,8 @@ var decodeTests = []decodeTest{
 	// struct tag "-"
 	{
 		input: "C20102",
-		ptr:   new(hasIgnoredField),
-		value: hasIgnoredField{A: 1, C: 2},
+		ptr:   new(ignoredField),
+		value: ignoredField{A: 1, C: 2},
 	},
 
 	// struct tag "nilList"
@@ -592,6 +619,110 @@ var decodeTests = []decodeTest{
 		value: nilStringSlice{X: &[]uint{3}},
 	},
 
+	// struct tag "optional"
+	{
+		input: "C101",
+		ptr:   new(optionalFields),
+		value: optionalFields{1, 0, 0},
+	},
+	{
+		input: "C20102",
+		ptr:   new(optionalFields),
+		value: optionalFields{1, 2, 0},
+	},
+	{
+		input: "C3010203",
+		ptr:   new(optionalFields),
+		value: optionalFields{1, 2, 3},
+	},
+	{
+		input: "C401020304",
+		ptr:   new(optionalFields),
+		error: "rlp: input list has too many elements for rlp.optionalFields",
+	},
+	{
+		input: "C101",
+		ptr:   new(optionalAndTailField),
+		value: optionalAndTailField{A: 1},
+	},
+	{
+		input: "C20102",
+		ptr:   new(optionalAndTailField),
+		value: optionalAndTailField{A: 1, B: 2, Tail: []uint{}},
+	},
+	{
+		input: "C401020304",
+		ptr:   new(optionalAndTailField),
+		value: optionalAndTailField{A: 1, B: 2, Tail: []uint{3, 4}},
+	},
+	{
+		input: "C101",
+		ptr:   new(optionalBigIntField),
+		value: optionalBigIntField{A: 1, B: nil},
+	},
+	{
+		input: "C20102",
+		ptr:   new(optionalBigIntField),
+		value: optionalBigIntField{A: 1, B: big.NewInt(2)},
+	},
+	{
+		input: "C101",
+		ptr:   new(optionalPtrField),
+		value: optionalPtrField{A: 1},
+	},
+	{
+		input: "C20180", // not accepted because "optional" doesn't enable "nil"
+		ptr:   new(optionalPtrField),
+		error: "rlp: input string too short for [3]uint8, decoding into (rlp.optionalPtrField).B",
+	},
+	{
+		input: "C20102",
+		ptr:   new(optionalPtrField),
+		error: "rlp: input string too short for [3]uint8, decoding into (rlp.optionalPtrField).B",
+	},
+	{
+		input: "C50183010203",
+		ptr:   new(optionalPtrField),
+		value: optionalPtrField{A: 1, B: &[3]byte{1, 2, 3}},
+	},
+	{
+		input: "C101",
+		ptr:   new(optionalPtrFieldNil),
+		value: optionalPtrFieldNil{A: 1},
+	},
+	{
+		input: "C20180", // accepted because "nil" tag allows empty input
+		ptr:   new(optionalPtrFieldNil),
+		value: optionalPtrFieldNil{A: 1},
+	},
+	{
+		input: "C20102",
+		ptr:   new(optionalPtrFieldNil),
+		error: "rlp: input string too short for [3]uint8, decoding into (rlp.optionalPtrFieldNil).B",
+	},
+
+	// struct tag "optional" field clearing
+	{
+		input: "C101",
+		ptr:   &optionalFields{A: 9, B: 8, C: 7},
+		value: optionalFields{A: 1, B: 0, C: 0},
+	},
+	{
+		input: "C20102",
+		ptr:   &optionalFields{A: 9, B: 8, C: 7},
+		value: optionalFields{A: 1, B: 2, C: 0},
+	},
+	{
+		input: "C20102",
+		ptr:   &optionalAndTailField{A: 9, B: 8, Tail: []uint{7, 6, 5}},
+		value: optionalAndTailField{A: 1, B: 2, Tail: []uint{}},
+	},
+	{
+		input: "C101",
+		ptr:   &optionalPtrField{A: 9, B: &[3]byte{8, 7, 6}},
+		value: optionalPtrField{A: 1},
+	},
+
 	// RawValue
 	{input: "01", ptr: new(RawValue), value: RawValue(unhex("01"))},
 	{input: "82FFFF", ptr: new(RawValue), value: RawValue(unhex("82FFFF"))},
@@ -822,6 +953,40 @@ func TestDecoderFunc(t *testing.T) {
 	x()
 }
 
+// This tests the validity checks for fields with struct tag "optional".
+func TestInvalidOptionalField(t *testing.T) {
+	type (
+		invalid1 struct {
+			A uint `rlp:"optional"`
+			B uint
+		}
+		invalid2 struct {
+			T []uint `rlp:"tail,optional"`
+		}
+		invalid3 struct {
+			T []uint `rlp:"optional,tail"`
+		}
+	)
+
+	tests := []struct {
+		v   interface{}
+		err string
+	}{
+		{v: new(invalid1), err: `rlp: struct field rlp.invalid1.B needs "optional" tag`},
+		{v: new(invalid2), err: `rlp: invalid struct tag "optional" for rlp.invalid2.T (also has "tail" tag)`},
+		{v: new(invalid3), err: `rlp: invalid struct tag "tail" for rlp.invalid3.T (also has "optional" tag)`},
+	}
+	for _, test := range tests {
+		err := DecodeBytes(unhex("C20102"), test.v)
+		if err == nil {
+			t.Errorf("no error for %T", test.v)
+		} else if err.Error() != test.err {
+			t.Errorf("wrong error for %T: %v", test.v, err.Error())
+		}
+	}
+
+}
+
 func ExampleDecode() {
 	input, _ := hex.DecodeString("C90A1486666F6F626172")
 

rlp/doc.go

@@ -102,29 +102,60 @@ Signed integers, floating point numbers, maps, channels and functions cannot be
 
 Struct Tags
 
-Package rlp honours certain struct tags: "-", "tail", "nil", "nilList" and "nilString".
+As with other encoding packages, the "-" tag ignores fields.
 
-The "-" tag ignores fields.
+    type StructWithIgnoredField struct{
+        Ignored uint `rlp:"-"`
+        Field   uint
+    }
+
+Go struct values encode/decode as RLP lists. There are two ways of influencing the mapping
+of fields to list elements. The "tail" tag, which may only be used on the last exported
+struct field, allows slurping up any excess list elements into a slice.
+
+    type StructWithTail struct{
+        Field   uint
+        Tail    []string `rlp:"tail"`
+    }
 
-The "tail" tag, which may only be used on the last exported struct field, allows slurping
-up any excess list elements into a slice. See examples for more details.
+The "optional" tag says that the field may be omitted if it is zero-valued. If this tag is
+used on a struct field, all subsequent public fields must also be declared optional.
 
-The "nil" tag applies to pointer-typed fields and changes the decoding rules for the field
-such that input values of size zero decode as a nil pointer. This tag can be useful when
-decoding recursive types.
+When encoding a struct with optional fields, the output RLP list contains all values up to
+the last non-zero optional field.
 
-    type StructWithOptionalFoo struct {
-        Foo *[20]byte `rlp:"nil"`
+When decoding into a struct, optional fields may be omitted from the end of the input
+list. For the example below, this means input lists of one, two, or three elements are
+accepted.
+
+   type StructWithOptionalFields struct{
+        Required  uint
+        Optional1 uint `rlp:"optional"`
+        Optional2 uint `rlp:"optional"`
+   }
+
+The "nil", "nilList" and "nilString" tags apply to pointer-typed fields only, and change
+the decoding rules for the field type. For regular pointer fields without the "nil" tag,
+input values must always match the required input length exactly and the decoder does not
+produce nil values. When the "nil" tag is set, input values of size zero decode as a nil
+pointer. This is especially useful for recursive types.
+
+    type StructWithNilField struct {
+        Field *[3]byte `rlp:"nil"`
     }
 
+In the example above, Field allows two possible input sizes. For input 0xC180 (a list
+containing an empty string) Field is set to nil after decoding. For input 0xC483000000 (a
+list containing a 3-byte string), Field is set to a non-nil array pointer.
+
 RLP supports two kinds of empty values: empty lists and empty strings. When using the
-"nil" tag, the kind of empty value allowed for a type is chosen automatically. A struct
-field whose Go type is a pointer to an unsigned integer, string, boolean or byte
-array/slice expects an empty RLP string. Any other pointer field type encodes/decodes as
-an empty RLP list.
+"nil" tag, the kind of empty value allowed for a type is chosen automatically. A field
+whose Go type is a pointer to an unsigned integer, string, boolean or byte array/slice
+expects an empty RLP string. Any other pointer field type encodes/decodes as an empty RLP
+list.
 
 The choice of null value can be made explicit with the "nilList" and "nilString" struct
-tags. Using these tags encodes/decodes a Go nil pointer value as the kind of empty
-RLP value defined by the tag.
+tags. Using these tags encodes/decodes a Go nil pointer value as the empty RLP value kind
+defined by the tag.
 */
 package rlp

rlp/encode.go

@@ -546,15 +546,40 @@ func makeStructWriter(typ reflect.Type) (writer, error) {
 			return nil, structFieldError{typ, f.index, f.info.writerErr}
 		}
 	}
-	writer := func(val reflect.Value, w *encbuf) error {
-		lh := w.list()
-		for _, f := range fields {
-			if err := f.info.writer(val.Field(f.index), w); err != nil {
-				return err
+
+	var writer writer
+	firstOptionalField := firstOptionalField(fields)
+	if firstOptionalField == len(fields) {
+		// This is the writer function for structs without any optional fields.
+		writer = func(val reflect.Value, w *encbuf) error {
+			lh := w.list()
+			for _, f := range fields {
+				if err := f.info.writer(val.Field(f.index), w); err != nil {
+					return err
+				}
 			}
+			w.listEnd(lh)
+			return nil
+		}
+	} else {
+		// If there are any "optional" fields, the writer needs to perform additional
+		// checks to determine the output list length.
+		writer = func(val reflect.Value, w *encbuf) error {
+			lastField := len(fields) - 1
+			for ; lastField >= firstOptionalField; lastField-- {
+				if !val.Field(fields[lastField].index).IsZero() {
+					break
+				}
+			}
+			lh := w.list()
+			for i := 0; i <= lastField; i++ {
+				if err := fields[i].info.writer(val.Field(fields[i].index), w); err != nil {
+					return err
+				}
+			}
+			w.listEnd(lh)
+			return nil
 		}
-		w.listEnd(lh)
-		return nil
 	}
 	return writer, nil
 }