2015-07-07 03:54:22 +03:00
|
|
|
// Copyright 2014 The go-ethereum Authors
|
2015-07-22 19:48:40 +03:00
|
|
|
// This file is part of the go-ethereum library.
|
2015-07-07 03:54:22 +03:00
|
|
|
//
|
2015-07-23 19:35:11 +03:00
|
|
|
// The go-ethereum library is free software: you can redistribute it and/or modify
|
2015-07-07 03:54:22 +03:00
|
|
|
// it under the terms of the GNU Lesser General Public License as published by
|
|
|
|
// the Free Software Foundation, either version 3 of the License, or
|
|
|
|
// (at your option) any later version.
|
|
|
|
//
|
2015-07-22 19:48:40 +03:00
|
|
|
// The go-ethereum library is distributed in the hope that it will be useful,
|
2015-07-07 03:54:22 +03:00
|
|
|
// but WITHOUT ANY WARRANTY; without even the implied warranty of
|
2015-07-22 19:48:40 +03:00
|
|
|
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
2015-07-07 03:54:22 +03:00
|
|
|
// GNU Lesser General Public License for more details.
|
|
|
|
//
|
|
|
|
// You should have received a copy of the GNU Lesser General Public License
|
2015-07-22 19:48:40 +03:00
|
|
|
// along with the go-ethereum library. If not, see <http://www.gnu.org/licenses/>.
|
2015-07-07 03:54:22 +03:00
|
|
|
|
2014-11-13 21:31:48 +02:00
|
|
|
package rlp
|
|
|
|
|
|
|
|
import (
|
|
|
|
"bytes"
|
|
|
|
"encoding/hex"
|
2015-08-13 11:12:38 +03:00
|
|
|
"errors"
|
2014-11-13 21:31:48 +02:00
|
|
|
"fmt"
|
|
|
|
"io"
|
|
|
|
"math/big"
|
|
|
|
"reflect"
|
2015-07-01 03:09:02 +03:00
|
|
|
"strings"
|
2014-11-13 21:31:48 +02:00
|
|
|
"testing"
|
2021-05-22 16:10:16 +03:00
|
|
|
|
|
|
|
"github.com/ethereum/go-ethereum/common/math"
|
2023-03-17 13:51:55 +03:00
|
|
|
"github.com/holiman/uint256"
|
2014-11-13 21:31:48 +02:00
|
|
|
)
|
|
|
|
|
|
|
|
func TestStreamKind(t *testing.T) {
|
|
|
|
tests := []struct {
|
|
|
|
input string
|
|
|
|
wantKind Kind
|
|
|
|
wantLen uint64
|
|
|
|
}{
|
|
|
|
{"00", Byte, 0},
|
|
|
|
{"01", Byte, 0},
|
|
|
|
{"7F", Byte, 0},
|
|
|
|
{"80", String, 0},
|
|
|
|
{"B7", String, 55},
|
|
|
|
{"B90400", String, 1024},
|
|
|
|
{"BFFFFFFFFFFFFFFFFF", String, ^uint64(0)},
|
|
|
|
{"C0", List, 0},
|
|
|
|
{"C8", List, 8},
|
|
|
|
{"F7", List, 55},
|
|
|
|
{"F90400", List, 1024},
|
|
|
|
{"FFFFFFFFFFFFFFFFFF", List, ^uint64(0)},
|
|
|
|
}
|
|
|
|
|
|
|
|
for i, test := range tests {
|
2015-03-23 16:02:55 +02:00
|
|
|
// using plainReader to inhibit input limit errors.
|
|
|
|
s := NewStream(newPlainReader(unhex(test.input)), 0)
|
2014-11-13 21:31:48 +02:00
|
|
|
kind, len, err := s.Kind()
|
|
|
|
if err != nil {
|
2015-03-20 23:33:40 +02:00
|
|
|
t.Errorf("test %d: Kind returned error: %v", i, err)
|
2014-11-13 21:31:48 +02:00
|
|
|
continue
|
|
|
|
}
|
|
|
|
if kind != test.wantKind {
|
|
|
|
t.Errorf("test %d: kind mismatch: got %d, want %d", i, kind, test.wantKind)
|
|
|
|
}
|
|
|
|
if len != test.wantLen {
|
|
|
|
t.Errorf("test %d: len mismatch: got %d, want %d", i, len, test.wantLen)
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2014-11-24 20:02:04 +02:00
|
|
|
func TestNewListStream(t *testing.T) {
|
|
|
|
ls := NewListStream(bytes.NewReader(unhex("0101010101")), 3)
|
|
|
|
if k, size, err := ls.Kind(); k != List || size != 3 || err != nil {
|
|
|
|
t.Errorf("Kind() returned (%v, %d, %v), expected (List, 3, nil)", k, size, err)
|
|
|
|
}
|
|
|
|
if size, err := ls.List(); size != 3 || err != nil {
|
|
|
|
t.Errorf("List() returned (%d, %v), expected (3, nil)", size, err)
|
|
|
|
}
|
|
|
|
for i := 0; i < 3; i++ {
|
|
|
|
if val, err := ls.Uint(); val != 1 || err != nil {
|
|
|
|
t.Errorf("Uint() returned (%d, %v), expected (1, nil)", val, err)
|
|
|
|
}
|
|
|
|
}
|
|
|
|
if err := ls.ListEnd(); err != nil {
|
|
|
|
t.Errorf("ListEnd() returned %v, expected (3, nil)", err)
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2014-11-13 21:31:48 +02:00
|
|
|
func TestStreamErrors(t *testing.T) {
|
2015-03-23 16:02:55 +02:00
|
|
|
withoutInputLimit := func(b []byte) *Stream {
|
|
|
|
return NewStream(newPlainReader(b), 0)
|
|
|
|
}
|
|
|
|
withCustomInputLimit := func(limit uint64) func([]byte) *Stream {
|
|
|
|
return func(b []byte) *Stream {
|
|
|
|
return NewStream(bytes.NewReader(b), limit)
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2014-11-13 21:31:48 +02:00
|
|
|
type calls []string
|
|
|
|
tests := []struct {
|
|
|
|
string
|
|
|
|
calls
|
2015-03-23 16:02:55 +02:00
|
|
|
newStream func([]byte) *Stream // uses bytes.Reader if nil
|
2015-04-14 17:09:06 +03:00
|
|
|
error error
|
2014-11-13 21:31:48 +02:00
|
|
|
}{
|
2015-03-23 16:02:55 +02:00
|
|
|
{"C0", calls{"Bytes"}, nil, ErrExpectedString},
|
|
|
|
{"C0", calls{"Uint"}, nil, ErrExpectedString},
|
|
|
|
{"89000000000000000001", calls{"Uint"}, nil, errUintOverflow},
|
|
|
|
{"00", calls{"List"}, nil, ErrExpectedList},
|
|
|
|
{"80", calls{"List"}, nil, ErrExpectedList},
|
|
|
|
{"C0", calls{"List", "Uint"}, nil, EOL},
|
|
|
|
{"C8C9010101010101010101", calls{"List", "Kind"}, nil, ErrElemTooLarge},
|
|
|
|
{"C3C2010201", calls{"List", "List", "Uint", "Uint", "ListEnd", "Uint"}, nil, EOL},
|
|
|
|
{"00", calls{"ListEnd"}, nil, errNotInList},
|
|
|
|
{"C401020304", calls{"List", "Uint", "ListEnd"}, nil, errNotAtEOL},
|
|
|
|
|
2015-04-16 11:15:14 +03:00
|
|
|
// Non-canonical integers (e.g. leading zero bytes).
|
2015-04-14 17:09:06 +03:00
|
|
|
{"00", calls{"Uint"}, nil, ErrCanonInt},
|
|
|
|
{"820002", calls{"Uint"}, nil, ErrCanonInt},
|
2015-04-16 11:15:14 +03:00
|
|
|
{"8133", calls{"Uint"}, nil, ErrCanonSize},
|
2015-07-18 17:13:20 +03:00
|
|
|
{"817F", calls{"Uint"}, nil, ErrCanonSize},
|
|
|
|
{"8180", calls{"Uint"}, nil, nil},
|
2015-04-14 17:09:06 +03:00
|
|
|
|
2015-08-13 11:12:38 +03:00
|
|
|
// Non-valid boolean
|
|
|
|
{"02", calls{"Bool"}, nil, errors.New("rlp: invalid boolean value: 2")},
|
|
|
|
|
2015-04-14 17:09:06 +03:00
|
|
|
// Size tags must use the smallest possible encoding.
|
|
|
|
// Leading zero bytes in the size tag are also rejected.
|
2015-04-16 11:15:14 +03:00
|
|
|
{"8100", calls{"Uint"}, nil, ErrCanonSize},
|
|
|
|
{"8100", calls{"Bytes"}, nil, ErrCanonSize},
|
2015-07-18 17:13:20 +03:00
|
|
|
{"8101", calls{"Bytes"}, nil, ErrCanonSize},
|
|
|
|
{"817F", calls{"Bytes"}, nil, ErrCanonSize},
|
|
|
|
{"8180", calls{"Bytes"}, nil, nil},
|
2015-04-14 17:09:06 +03:00
|
|
|
{"B800", calls{"Kind"}, withoutInputLimit, ErrCanonSize},
|
|
|
|
{"B90000", calls{"Kind"}, withoutInputLimit, ErrCanonSize},
|
|
|
|
{"B90055", calls{"Kind"}, withoutInputLimit, ErrCanonSize},
|
|
|
|
{"BA0002FFFF", calls{"Bytes"}, withoutInputLimit, ErrCanonSize},
|
|
|
|
{"F800", calls{"Kind"}, withoutInputLimit, ErrCanonSize},
|
|
|
|
{"F90000", calls{"Kind"}, withoutInputLimit, ErrCanonSize},
|
|
|
|
{"F90055", calls{"Kind"}, withoutInputLimit, ErrCanonSize},
|
|
|
|
{"FA0002FFFF", calls{"List"}, withoutInputLimit, ErrCanonSize},
|
|
|
|
|
2015-03-23 16:02:55 +02:00
|
|
|
// Expected EOF
|
|
|
|
{"", calls{"Kind"}, nil, io.EOF},
|
|
|
|
{"", calls{"Uint"}, nil, io.EOF},
|
|
|
|
{"", calls{"List"}, nil, io.EOF},
|
2015-07-18 17:13:20 +03:00
|
|
|
{"8180", calls{"Uint", "Uint"}, nil, io.EOF},
|
2015-03-23 16:02:55 +02:00
|
|
|
{"C0", calls{"List", "ListEnd", "List"}, nil, io.EOF},
|
|
|
|
|
2015-04-28 11:28:15 +03:00
|
|
|
{"", calls{"List"}, withoutInputLimit, io.EOF},
|
2015-07-18 17:13:20 +03:00
|
|
|
{"8180", calls{"Uint", "Uint"}, withoutInputLimit, io.EOF},
|
2015-04-28 11:28:15 +03:00
|
|
|
{"C0", calls{"List", "ListEnd", "List"}, withoutInputLimit, io.EOF},
|
|
|
|
|
2015-03-23 16:02:55 +02:00
|
|
|
// Input limit errors.
|
|
|
|
{"81", calls{"Bytes"}, nil, ErrValueTooLarge},
|
|
|
|
{"81", calls{"Uint"}, nil, ErrValueTooLarge},
|
|
|
|
{"81", calls{"Raw"}, nil, ErrValueTooLarge},
|
|
|
|
{"BFFFFFFFFFFFFFFFFFFF", calls{"Bytes"}, nil, ErrValueTooLarge},
|
|
|
|
{"C801", calls{"List"}, nil, ErrValueTooLarge},
|
|
|
|
|
2015-04-14 01:54:12 +03:00
|
|
|
// Test for list element size check overflow.
|
|
|
|
{"CD04040404FFFFFFFFFFFFFFFFFF0303", calls{"List", "Uint", "Uint", "Uint", "Uint", "List"}, nil, ErrElemTooLarge},
|
|
|
|
|
2015-03-23 16:02:55 +02:00
|
|
|
// Test for input limit overflow. Since we are counting the limit
|
|
|
|
// down toward zero in Stream.remaining, reading too far can overflow
|
|
|
|
// remaining to a large value, effectively disabling the limit.
|
|
|
|
{"C40102030401", calls{"Raw", "Uint"}, withCustomInputLimit(5), io.EOF},
|
2015-07-18 17:13:20 +03:00
|
|
|
{"C4010203048180", calls{"Raw", "Uint"}, withCustomInputLimit(6), ErrValueTooLarge},
|
2015-03-23 16:02:55 +02:00
|
|
|
|
|
|
|
// Check that the same calls are fine without a limit.
|
|
|
|
{"C40102030401", calls{"Raw", "Uint"}, withoutInputLimit, nil},
|
2015-07-18 17:13:20 +03:00
|
|
|
{"C4010203048180", calls{"Raw", "Uint"}, withoutInputLimit, nil},
|
2015-03-23 16:02:55 +02:00
|
|
|
|
|
|
|
// Unexpected EOF. This only happens when there is
|
|
|
|
// no input limit, so the reader needs to be 'dumbed down'.
|
|
|
|
{"81", calls{"Bytes"}, withoutInputLimit, io.ErrUnexpectedEOF},
|
|
|
|
{"81", calls{"Uint"}, withoutInputLimit, io.ErrUnexpectedEOF},
|
|
|
|
{"BFFFFFFFFFFFFFFF", calls{"Bytes"}, withoutInputLimit, io.ErrUnexpectedEOF},
|
|
|
|
{"C801", calls{"List", "Uint", "Uint"}, withoutInputLimit, io.ErrUnexpectedEOF},
|
2015-03-20 23:33:40 +02:00
|
|
|
|
|
|
|
// This test verifies that the input position is advanced
|
|
|
|
// correctly when calling Bytes for empty strings. Kind can be called
|
|
|
|
// any number of times in between and doesn't advance.
|
|
|
|
{"C3808080", calls{
|
|
|
|
"List", // enter the list
|
|
|
|
"Bytes", // past first element
|
|
|
|
|
|
|
|
"Kind", "Kind", "Kind", // this shouldn't advance
|
|
|
|
|
|
|
|
"Bytes", // past second element
|
|
|
|
|
|
|
|
"Kind", "Kind", // can't hurt to try
|
|
|
|
|
|
|
|
"Bytes", // past final element
|
|
|
|
"Bytes", // this one should fail
|
2015-03-23 16:02:55 +02:00
|
|
|
}, nil, EOL},
|
2014-11-13 21:31:48 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
testfor:
|
|
|
|
for i, test := range tests {
|
2015-03-23 16:02:55 +02:00
|
|
|
if test.newStream == nil {
|
|
|
|
test.newStream = func(b []byte) *Stream { return NewStream(bytes.NewReader(b), 0) }
|
|
|
|
}
|
|
|
|
s := test.newStream(unhex(test.string))
|
2014-11-13 21:31:48 +02:00
|
|
|
rs := reflect.ValueOf(s)
|
|
|
|
for j, call := range test.calls {
|
|
|
|
fval := rs.MethodByName(call)
|
|
|
|
ret := fval.Call(nil)
|
|
|
|
err := "<nil>"
|
|
|
|
if lastret := ret[len(ret)-1].Interface(); lastret != nil {
|
|
|
|
err = lastret.(error).Error()
|
|
|
|
}
|
|
|
|
if j == len(test.calls)-1 {
|
2015-03-23 16:02:55 +02:00
|
|
|
want := "<nil>"
|
|
|
|
if test.error != nil {
|
|
|
|
want = test.error.Error()
|
|
|
|
}
|
|
|
|
if err != want {
|
2015-04-14 17:09:06 +03:00
|
|
|
t.Log(test)
|
2015-03-23 16:02:55 +02:00
|
|
|
t.Errorf("test %d: last call (%s) error mismatch\ngot: %s\nwant: %s",
|
2014-11-13 21:31:48 +02:00
|
|
|
i, call, err, test.error)
|
|
|
|
}
|
|
|
|
} else if err != "<nil>" {
|
2015-04-14 17:09:06 +03:00
|
|
|
t.Log(test)
|
2014-11-13 21:31:48 +02:00
|
|
|
t.Errorf("test %d: call %d (%s) unexpected error: %q", i, j, call, err)
|
|
|
|
continue testfor
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
func TestStreamList(t *testing.T) {
|
2015-03-23 16:02:55 +02:00
|
|
|
s := NewStream(bytes.NewReader(unhex("C80102030405060708")), 0)
|
2014-11-13 21:31:48 +02:00
|
|
|
|
|
|
|
len, err := s.List()
|
|
|
|
if err != nil {
|
|
|
|
t.Fatalf("List error: %v", err)
|
|
|
|
}
|
|
|
|
if len != 8 {
|
|
|
|
t.Fatalf("List returned invalid length, got %d, want 8", len)
|
|
|
|
}
|
|
|
|
|
|
|
|
for i := uint64(1); i <= 8; i++ {
|
|
|
|
v, err := s.Uint()
|
|
|
|
if err != nil {
|
|
|
|
t.Fatalf("Uint error: %v", err)
|
|
|
|
}
|
|
|
|
if i != v {
|
|
|
|
t.Errorf("Uint returned wrong value, got %d, want %d", v, i)
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if _, err := s.Uint(); err != EOL {
|
|
|
|
t.Errorf("Uint error mismatch, got %v, want %v", err, EOL)
|
|
|
|
}
|
|
|
|
if err = s.ListEnd(); err != nil {
|
|
|
|
t.Fatalf("ListEnd error: %v", err)
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2015-03-21 01:49:31 +02:00
|
|
|
func TestStreamRaw(t *testing.T) {
|
2017-08-24 13:10:50 +03:00
|
|
|
tests := []struct {
|
|
|
|
input string
|
|
|
|
output string
|
|
|
|
}{
|
|
|
|
{
|
|
|
|
"C58401010101",
|
|
|
|
"8401010101",
|
|
|
|
},
|
|
|
|
{
|
|
|
|
"F842B84001010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101",
|
|
|
|
"B84001010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101",
|
|
|
|
},
|
2015-03-21 01:49:31 +02:00
|
|
|
}
|
2017-08-24 13:10:50 +03:00
|
|
|
for i, tt := range tests {
|
|
|
|
s := NewStream(bytes.NewReader(unhex(tt.input)), 0)
|
|
|
|
s.List()
|
|
|
|
|
|
|
|
want := unhex(tt.output)
|
|
|
|
raw, err := s.Raw()
|
|
|
|
if err != nil {
|
|
|
|
t.Fatal(err)
|
|
|
|
}
|
|
|
|
if !bytes.Equal(want, raw) {
|
|
|
|
t.Errorf("test %d: raw mismatch: got %x, want %x", i, raw, want)
|
|
|
|
}
|
2015-03-21 01:49:31 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
rlp/rlpgen: RLP encoder code generator (#24251)
This change adds a code generator tool for creating EncodeRLP method
implementations. The generated methods will behave identically to the
reflect-based encoder, but run faster because there is no reflection overhead.
Package rlp now provides the EncoderBuffer type for incremental encoding. This
is used by generated code, but the new methods can also be useful for
hand-written encoders.
There is also experimental support for generating DecodeRLP, and some new
methods have been added to the existing Stream type to support this. Creating
decoders with rlpgen is not recommended at this time because the generated
methods create very poor error reporting.
More detail about package rlp changes:
* rlp: externalize struct field processing / validation
This adds a new package, rlp/internal/rlpstruct, in preparation for the
RLP encoder generator.
I think the struct field rules are subtle enough to warrant extracting
this into their own package, even though it means that a bunch of
adapter code is needed for converting to/from rlpstruct.Type.
* rlp: add more decoder methods (for rlpgen)
This adds new methods on rlp.Stream:
- Uint64, Uint32, Uint16, Uint8, BigInt
- ReadBytes for decoding into []byte
- MoreDataInList - useful for optional list elements
* rlp: expose encoder buffer (for rlpgen)
This exposes the internal encoder buffer type for use in EncodeRLP
implementations.
The new EncoderBuffer type is a sort-of 'opaque handle' for a pointer to
encBuffer. It is implemented this way to ensure the global encBuffer pool
is handled correctly.
2022-02-16 20:14:12 +03:00
|
|
|
func TestStreamReadBytes(t *testing.T) {
|
|
|
|
tests := []struct {
|
|
|
|
input string
|
|
|
|
size int
|
|
|
|
err string
|
|
|
|
}{
|
|
|
|
// kind List
|
|
|
|
{input: "C0", size: 1, err: "rlp: expected String or Byte"},
|
|
|
|
// kind Byte
|
|
|
|
{input: "04", size: 0, err: "input value has wrong size 1, want 0"},
|
|
|
|
{input: "04", size: 1},
|
|
|
|
{input: "04", size: 2, err: "input value has wrong size 1, want 2"},
|
|
|
|
// kind String
|
|
|
|
{input: "820102", size: 0, err: "input value has wrong size 2, want 0"},
|
|
|
|
{input: "820102", size: 1, err: "input value has wrong size 2, want 1"},
|
|
|
|
{input: "820102", size: 2},
|
|
|
|
{input: "820102", size: 3, err: "input value has wrong size 2, want 3"},
|
|
|
|
}
|
|
|
|
|
|
|
|
for _, test := range tests {
|
|
|
|
test := test
|
|
|
|
name := fmt.Sprintf("input_%s/size_%d", test.input, test.size)
|
|
|
|
t.Run(name, func(t *testing.T) {
|
|
|
|
s := NewStream(bytes.NewReader(unhex(test.input)), 0)
|
|
|
|
b := make([]byte, test.size)
|
|
|
|
err := s.ReadBytes(b)
|
|
|
|
if test.err == "" {
|
|
|
|
if err != nil {
|
|
|
|
t.Errorf("unexpected error %q", err)
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
if err == nil {
|
|
|
|
t.Errorf("expected error, got nil")
|
|
|
|
} else if err.Error() != test.err {
|
|
|
|
t.Errorf("wrong error %q", err)
|
|
|
|
}
|
|
|
|
}
|
|
|
|
})
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2014-11-13 21:31:48 +02:00
|
|
|
func TestDecodeErrors(t *testing.T) {
|
|
|
|
r := bytes.NewReader(nil)
|
|
|
|
|
|
|
|
if err := Decode(r, nil); err != errDecodeIntoNil {
|
|
|
|
t.Errorf("Decode(r, nil) error mismatch, got %q, want %q", err, errDecodeIntoNil)
|
|
|
|
}
|
|
|
|
|
|
|
|
var nilptr *struct{}
|
|
|
|
if err := Decode(r, nilptr); err != errDecodeIntoNil {
|
|
|
|
t.Errorf("Decode(r, nilptr) error mismatch, got %q, want %q", err, errDecodeIntoNil)
|
|
|
|
}
|
|
|
|
|
|
|
|
if err := Decode(r, struct{}{}); err != errNoPointer {
|
|
|
|
t.Errorf("Decode(r, struct{}{}) error mismatch, got %q, want %q", err, errNoPointer)
|
|
|
|
}
|
|
|
|
|
|
|
|
expectErr := "rlp: type chan bool is not RLP-serializable"
|
|
|
|
if err := Decode(r, new(chan bool)); err == nil || err.Error() != expectErr {
|
|
|
|
t.Errorf("Decode(r, new(chan bool)) error mismatch, got %q, want %q", err, expectErr)
|
|
|
|
}
|
|
|
|
|
2014-12-09 11:41:47 +02:00
|
|
|
if err := Decode(r, new(uint)); err != io.EOF {
|
2014-11-13 21:31:48 +02:00
|
|
|
t.Errorf("Decode(r, new(int)) error mismatch, got %q, want %q", err, io.EOF)
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
type decodeTest struct {
|
|
|
|
input string
|
|
|
|
ptr interface{}
|
|
|
|
value interface{}
|
2014-11-25 17:00:48 +02:00
|
|
|
error string
|
2014-11-13 21:31:48 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
type simplestruct struct {
|
2014-12-09 11:41:47 +02:00
|
|
|
A uint
|
2014-11-13 21:31:48 +02:00
|
|
|
B string
|
|
|
|
}
|
|
|
|
|
|
|
|
type recstruct struct {
|
2014-12-09 11:41:47 +02:00
|
|
|
I uint
|
2015-04-17 02:16:46 +03:00
|
|
|
Child *recstruct `rlp:"nil"`
|
2014-11-13 21:31:48 +02:00
|
|
|
}
|
|
|
|
|
2021-05-25 22:56:25 +03:00
|
|
|
type bigIntStruct struct {
|
|
|
|
I *big.Int
|
|
|
|
B string
|
|
|
|
}
|
|
|
|
|
rlp: improve nil pointer handling (#20064)
* rlp: improve nil pointer handling
In both encoder and decoder, the rules for encoding nil pointers were a
bit hard to understand, and didn't leave much choice. Since RLP allows
two empty values (empty list, empty string), any protocol built on RLP
must choose either of these values to represent the null value in a
certain context.
This change adds choice in the form of two new struct tags, "nilString"
and "nilList". These can be used to specify how a nil pointer value is
encoded. The "nil" tag still exists, but its implementation is now
explicit and defines exactly how nil pointers are handled in a single
place.
Another important change in this commit is how nil pointers and the
Encoder interface interact. The EncodeRLP method was previously called
even on nil values, which was supposed to give users a choice of how
their value would be handled when nil. It turns out this is a stupid
idea. If you create a network protocol containing an object defined in
another package, it's better to be able to say that the object should be
a list or string when nil in the definition of the protocol message
rather than defining the encoding of nil on the object itself.
As of this commit, the encoding rules for pointers now take precedence
over the Encoder interface rule. I think the "nil" tag will work fine
for most cases. For special kinds of objects which are a struct in Go
but strings in RLP, code using the object can specify the desired
encoding of nil using the "nilString" and "nilList" tags.
* rlp: propagate struct field type errors
If a struct contained fields of undecodable type, the encoder and
decoder would panic instead of returning an error. Fix this by
propagating type errors in makeStruct{Writer,Decoder} and add a test.
2019-09-13 12:10:57 +03:00
|
|
|
type invalidNilTag struct {
|
|
|
|
X []byte `rlp:"nil"`
|
|
|
|
}
|
|
|
|
|
2015-12-21 22:05:20 +02:00
|
|
|
type invalidTail1 struct {
|
|
|
|
A uint `rlp:"tail"`
|
|
|
|
B string
|
|
|
|
}
|
|
|
|
|
|
|
|
type invalidTail2 struct {
|
|
|
|
A uint
|
|
|
|
B string `rlp:"tail"`
|
|
|
|
}
|
|
|
|
|
|
|
|
type tailRaw struct {
|
|
|
|
A uint
|
|
|
|
Tail []RawValue `rlp:"tail"`
|
|
|
|
}
|
|
|
|
|
|
|
|
type tailUint struct {
|
|
|
|
A uint
|
|
|
|
Tail []uint `rlp:"tail"`
|
|
|
|
}
|
|
|
|
|
2019-05-14 16:09:56 +03:00
|
|
|
type tailPrivateFields struct {
|
|
|
|
A uint
|
|
|
|
Tail []uint `rlp:"tail"`
|
2019-11-25 15:41:53 +03:00
|
|
|
x, y bool //lint:ignore U1000 unused fields required for testing purposes.
|
2019-05-14 16:09:56 +03:00
|
|
|
}
|
|
|
|
|
rlp: improve nil pointer handling (#20064)
* rlp: improve nil pointer handling
In both encoder and decoder, the rules for encoding nil pointers were a
bit hard to understand, and didn't leave much choice. Since RLP allows
two empty values (empty list, empty string), any protocol built on RLP
must choose either of these values to represent the null value in a
certain context.
This change adds choice in the form of two new struct tags, "nilString"
and "nilList". These can be used to specify how a nil pointer value is
encoded. The "nil" tag still exists, but its implementation is now
explicit and defines exactly how nil pointers are handled in a single
place.
Another important change in this commit is how nil pointers and the
Encoder interface interact. The EncodeRLP method was previously called
even on nil values, which was supposed to give users a choice of how
their value would be handled when nil. It turns out this is a stupid
idea. If you create a network protocol containing an object defined in
another package, it's better to be able to say that the object should be
a list or string when nil in the definition of the protocol message
rather than defining the encoding of nil on the object itself.
As of this commit, the encoding rules for pointers now take precedence
over the Encoder interface rule. I think the "nil" tag will work fine
for most cases. For special kinds of objects which are a struct in Go
but strings in RLP, code using the object can specify the desired
encoding of nil using the "nilString" and "nilList" tags.
* rlp: propagate struct field type errors
If a struct contained fields of undecodable type, the encoder and
decoder would panic instead of returning an error. Fix this by
propagating type errors in makeStruct{Writer,Decoder} and add a test.
2019-09-13 12:10:57 +03:00
|
|
|
type nilListUint struct {
|
|
|
|
X *uint `rlp:"nilList"`
|
|
|
|
}
|
|
|
|
|
|
|
|
type nilStringSlice struct {
|
|
|
|
X *[]uint `rlp:"nilString"`
|
|
|
|
}
|
|
|
|
|
|
|
|
type intField struct {
|
|
|
|
X int
|
|
|
|
}
|
|
|
|
|
2021-05-07 15:37:13 +03:00
|
|
|
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"`
|
|
|
|
}
|
|
|
|
|
2022-11-02 16:50:48 +03:00
|
|
|
type nonOptionalPtrField struct {
|
|
|
|
A uint
|
|
|
|
B *[3]byte
|
|
|
|
}
|
|
|
|
|
|
|
|
type multipleOptionalFields struct {
|
|
|
|
A *[3]byte `rlp:"optional"`
|
|
|
|
B *[3]byte `rlp:"optional"`
|
|
|
|
}
|
|
|
|
|
2021-05-07 15:37:13 +03:00
|
|
|
type optionalPtrFieldNil struct {
|
|
|
|
A uint
|
|
|
|
B *[3]byte `rlp:"optional,nil"`
|
|
|
|
}
|
|
|
|
|
|
|
|
type ignoredField struct {
|
|
|
|
A uint
|
|
|
|
B uint `rlp:"-"`
|
|
|
|
C uint
|
|
|
|
}
|
|
|
|
|
2014-11-13 21:31:48 +02:00
|
|
|
var (
|
2021-05-25 22:56:25 +03:00
|
|
|
veryBigInt = new(big.Int).Add(
|
2022-06-14 15:09:48 +03:00
|
|
|
new(big.Int).Lsh(big.NewInt(0xFFFFFFFFFFFFFF), 16),
|
2014-11-13 21:31:48 +02:00
|
|
|
big.NewInt(0xFFFF),
|
|
|
|
)
|
2021-05-25 22:56:25 +03:00
|
|
|
veryVeryBigInt = new(big.Int).Exp(veryBigInt, big.NewInt(8), nil)
|
2014-11-13 21:31:48 +02:00
|
|
|
)
|
|
|
|
|
2023-03-17 13:51:55 +03:00
|
|
|
var (
|
|
|
|
veryBigInt256, _ = uint256.FromBig(veryBigInt)
|
|
|
|
)
|
|
|
|
|
2014-11-13 21:31:48 +02:00
|
|
|
var decodeTests = []decodeTest{
|
2015-08-13 11:12:38 +03:00
|
|
|
// booleans
|
|
|
|
{input: "01", ptr: new(bool), value: true},
|
|
|
|
{input: "80", ptr: new(bool), value: false},
|
|
|
|
{input: "02", ptr: new(bool), error: "rlp: invalid boolean value: 2"},
|
|
|
|
|
2014-11-13 21:31:48 +02:00
|
|
|
// integers
|
|
|
|
{input: "05", ptr: new(uint32), value: uint32(5)},
|
|
|
|
{input: "80", ptr: new(uint32), value: uint32(0)},
|
|
|
|
{input: "820505", ptr: new(uint32), value: uint32(0x0505)},
|
|
|
|
{input: "83050505", ptr: new(uint32), value: uint32(0x050505)},
|
|
|
|
{input: "8405050505", ptr: new(uint32), value: uint32(0x05050505)},
|
2014-12-09 14:55:47 +02:00
|
|
|
{input: "850505050505", ptr: new(uint32), error: "rlp: input string too long for uint32"},
|
|
|
|
{input: "C0", ptr: new(uint32), error: "rlp: expected input string or byte for uint32"},
|
2015-04-14 17:09:06 +03:00
|
|
|
{input: "00", ptr: new(uint32), error: "rlp: non-canonical integer (leading zero bytes) for uint32"},
|
2015-04-16 11:15:14 +03:00
|
|
|
{input: "8105", ptr: new(uint32), error: "rlp: non-canonical size information for uint32"},
|
2015-04-14 17:09:06 +03:00
|
|
|
{input: "820004", ptr: new(uint32), error: "rlp: non-canonical integer (leading zero bytes) for uint32"},
|
|
|
|
{input: "B8020004", ptr: new(uint32), error: "rlp: non-canonical size information for uint32"},
|
2014-11-13 21:31:48 +02:00
|
|
|
|
|
|
|
// slices
|
2014-12-09 11:41:47 +02:00
|
|
|
{input: "C0", ptr: new([]uint), value: []uint{}},
|
|
|
|
{input: "C80102030405060708", ptr: new([]uint), value: []uint{1, 2, 3, 4, 5, 6, 7, 8}},
|
2015-04-14 17:09:06 +03:00
|
|
|
{input: "F8020004", ptr: new([]uint), error: "rlp: non-canonical size information for []uint"},
|
2014-11-13 21:31:48 +02:00
|
|
|
|
|
|
|
// arrays
|
2014-12-09 11:41:47 +02:00
|
|
|
{input: "C50102030405", ptr: new([5]uint), value: [5]uint{1, 2, 3, 4, 5}},
|
2015-04-17 04:11:24 +03:00
|
|
|
{input: "C0", ptr: new([5]uint), error: "rlp: input list has too few elements for [5]uint"},
|
|
|
|
{input: "C102", ptr: new([5]uint), error: "rlp: input list has too few elements for [5]uint"},
|
2014-12-09 11:41:47 +02:00
|
|
|
{input: "C6010203040506", ptr: new([5]uint), error: "rlp: input list has too many elements for [5]uint"},
|
2015-04-14 17:09:06 +03:00
|
|
|
{input: "F8020004", ptr: new([5]uint), error: "rlp: non-canonical size information for [5]uint"},
|
2014-11-13 21:31:48 +02:00
|
|
|
|
2015-04-17 04:11:24 +03:00
|
|
|
// zero sized arrays
|
|
|
|
{input: "C0", ptr: new([0]uint), value: [0]uint{}},
|
|
|
|
{input: "C101", ptr: new([0]uint), error: "rlp: input list has too many elements for [0]uint"},
|
|
|
|
|
2014-11-13 21:31:48 +02:00
|
|
|
// byte slices
|
|
|
|
{input: "01", ptr: new([]byte), value: []byte{1}},
|
|
|
|
{input: "80", ptr: new([]byte), value: []byte{}},
|
|
|
|
{input: "8D6162636465666768696A6B6C6D", ptr: new([]byte), value: []byte("abcdefghijklm")},
|
2015-04-17 03:01:38 +03:00
|
|
|
{input: "C0", ptr: new([]byte), error: "rlp: expected input string or byte for []uint8"},
|
|
|
|
{input: "8105", ptr: new([]byte), error: "rlp: non-canonical size information for []uint8"},
|
2014-11-13 21:31:48 +02:00
|
|
|
|
|
|
|
// byte arrays
|
2015-04-17 04:11:24 +03:00
|
|
|
{input: "02", ptr: new([1]byte), value: [1]byte{2}},
|
2015-07-18 17:13:20 +03:00
|
|
|
{input: "8180", ptr: new([1]byte), value: [1]byte{128}},
|
2014-11-13 21:31:48 +02:00
|
|
|
{input: "850102030405", ptr: new([5]byte), value: [5]byte{1, 2, 3, 4, 5}},
|
2015-04-17 03:01:38 +03:00
|
|
|
|
|
|
|
// byte array errors
|
2015-04-17 04:11:24 +03:00
|
|
|
{input: "02", ptr: new([5]byte), error: "rlp: input string too short for [5]uint8"},
|
|
|
|
{input: "80", ptr: new([5]byte), error: "rlp: input string too short for [5]uint8"},
|
|
|
|
{input: "820000", ptr: new([5]byte), error: "rlp: input string too short for [5]uint8"},
|
2015-04-17 03:01:38 +03:00
|
|
|
{input: "C0", ptr: new([5]byte), error: "rlp: expected input string or byte for [5]uint8"},
|
|
|
|
{input: "C3010203", ptr: new([5]byte), error: "rlp: expected input string or byte for [5]uint8"},
|
|
|
|
{input: "86010203040506", ptr: new([5]byte), error: "rlp: input string too long for [5]uint8"},
|
2015-04-17 04:11:24 +03:00
|
|
|
{input: "8105", ptr: new([1]byte), error: "rlp: non-canonical size information for [1]uint8"},
|
2015-07-18 17:13:20 +03:00
|
|
|
{input: "817F", ptr: new([1]byte), error: "rlp: non-canonical size information for [1]uint8"},
|
2014-11-13 21:31:48 +02:00
|
|
|
|
|
|
|
// zero sized byte arrays
|
|
|
|
{input: "80", ptr: new([0]byte), value: [0]byte{}},
|
2014-12-09 14:55:47 +02:00
|
|
|
{input: "01", ptr: new([0]byte), error: "rlp: input string too long for [0]uint8"},
|
|
|
|
{input: "8101", ptr: new([0]byte), error: "rlp: input string too long for [0]uint8"},
|
2014-11-13 21:31:48 +02:00
|
|
|
|
|
|
|
// strings
|
|
|
|
{input: "00", ptr: new(string), value: "\000"},
|
|
|
|
{input: "8D6162636465666768696A6B6C6D", ptr: new(string), value: "abcdefghijklm"},
|
2014-12-09 14:55:47 +02:00
|
|
|
{input: "C0", ptr: new(string), error: "rlp: expected input string or byte for string"},
|
2014-11-13 21:31:48 +02:00
|
|
|
|
|
|
|
// big ints
|
2021-05-25 22:56:25 +03:00
|
|
|
{input: "80", ptr: new(*big.Int), value: big.NewInt(0)},
|
2014-11-13 21:31:48 +02:00
|
|
|
{input: "01", ptr: new(*big.Int), value: big.NewInt(1)},
|
|
|
|
{input: "89FFFFFFFFFFFFFFFFFF", ptr: new(*big.Int), value: veryBigInt},
|
2021-05-25 22:56:25 +03:00
|
|
|
{input: "B848FFFFFFFFFFFFFFFFF800000000000000001BFFFFFFFFFFFFFFFFC8000000000000000045FFFFFFFFFFFFFFFFC800000000000000001BFFFFFFFFFFFFFFFFF8000000000000000001", ptr: new(*big.Int), value: veryVeryBigInt},
|
2014-11-13 21:31:48 +02:00
|
|
|
{input: "10", ptr: new(big.Int), value: *big.NewInt(16)}, // non-pointer also works
|
2023-03-17 13:51:55 +03:00
|
|
|
|
|
|
|
// big int errors
|
2014-12-09 14:55:47 +02:00
|
|
|
{input: "C0", ptr: new(*big.Int), error: "rlp: expected input string or byte for *big.Int"},
|
2021-05-25 22:56:25 +03:00
|
|
|
{input: "00", ptr: new(*big.Int), error: "rlp: non-canonical integer (leading zero bytes) for *big.Int"},
|
|
|
|
{input: "820001", ptr: new(*big.Int), error: "rlp: non-canonical integer (leading zero bytes) for *big.Int"},
|
|
|
|
{input: "8105", ptr: new(*big.Int), error: "rlp: non-canonical size information for *big.Int"},
|
2014-11-13 21:31:48 +02:00
|
|
|
|
2023-03-17 13:51:55 +03:00
|
|
|
// uint256
|
|
|
|
{input: "80", ptr: new(*uint256.Int), value: uint256.NewInt(0)},
|
|
|
|
{input: "01", ptr: new(*uint256.Int), value: uint256.NewInt(1)},
|
|
|
|
{input: "88FFFFFFFFFFFFFFFF", ptr: new(*uint256.Int), value: uint256.NewInt(math.MaxUint64)},
|
|
|
|
{input: "89FFFFFFFFFFFFFFFFFF", ptr: new(*uint256.Int), value: veryBigInt256},
|
|
|
|
{input: "10", ptr: new(uint256.Int), value: *uint256.NewInt(16)}, // non-pointer also works
|
|
|
|
|
|
|
|
// uint256 errors
|
|
|
|
{input: "C0", ptr: new(*uint256.Int), error: "rlp: expected input string or byte for *uint256.Int"},
|
|
|
|
{input: "00", ptr: new(*uint256.Int), error: "rlp: non-canonical integer (leading zero bytes) for *uint256.Int"},
|
|
|
|
{input: "820001", ptr: new(*uint256.Int), error: "rlp: non-canonical integer (leading zero bytes) for *uint256.Int"},
|
|
|
|
{input: "8105", ptr: new(*uint256.Int), error: "rlp: non-canonical size information for *uint256.Int"},
|
|
|
|
{input: "A1FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF00", ptr: new(*uint256.Int), error: "rlp: value too large for uint256"},
|
|
|
|
|
2014-11-13 21:31:48 +02:00
|
|
|
// structs
|
|
|
|
{
|
2015-04-17 02:16:46 +03:00
|
|
|
input: "C50583343434",
|
|
|
|
ptr: new(simplestruct),
|
|
|
|
value: simplestruct{5, "444"},
|
|
|
|
},
|
|
|
|
{
|
|
|
|
input: "C601C402C203C0",
|
2014-11-13 21:31:48 +02:00
|
|
|
ptr: new(recstruct),
|
|
|
|
value: recstruct{1, &recstruct{2, &recstruct{3, nil}}},
|
|
|
|
},
|
2021-05-25 22:56:25 +03:00
|
|
|
{
|
|
|
|
// This checks that empty big.Int works correctly in struct context. It's easy to
|
|
|
|
// miss the update of s.kind for this case, so it needs its own test.
|
|
|
|
input: "C58083343434",
|
|
|
|
ptr: new(bigIntStruct),
|
|
|
|
value: bigIntStruct{new(big.Int), "444"},
|
|
|
|
},
|
2014-11-13 21:31:48 +02:00
|
|
|
|
2015-04-17 02:16:46 +03:00
|
|
|
// struct errors
|
|
|
|
{
|
|
|
|
input: "C0",
|
|
|
|
ptr: new(simplestruct),
|
|
|
|
error: "rlp: too few elements for rlp.simplestruct",
|
|
|
|
},
|
|
|
|
{
|
|
|
|
input: "C105",
|
|
|
|
ptr: new(simplestruct),
|
|
|
|
error: "rlp: too few elements for rlp.simplestruct",
|
|
|
|
},
|
|
|
|
{
|
|
|
|
input: "C7C50583343434C0",
|
|
|
|
ptr: new([]*simplestruct),
|
|
|
|
error: "rlp: too few elements for rlp.simplestruct, decoding into ([]*rlp.simplestruct)[1]",
|
|
|
|
},
|
2015-04-14 17:09:06 +03:00
|
|
|
{
|
|
|
|
input: "83222222",
|
|
|
|
ptr: new(simplestruct),
|
|
|
|
error: "rlp: expected input list for rlp.simplestruct",
|
|
|
|
},
|
2015-01-05 18:10:42 +02:00
|
|
|
{
|
|
|
|
input: "C3010101",
|
|
|
|
ptr: new(simplestruct),
|
|
|
|
error: "rlp: input list has too many elements for rlp.simplestruct",
|
|
|
|
},
|
|
|
|
{
|
|
|
|
input: "C501C3C00000",
|
|
|
|
ptr: new(recstruct),
|
|
|
|
error: "rlp: expected input string or byte for uint, decoding into (rlp.recstruct).Child.I",
|
|
|
|
},
|
2015-12-21 22:05:20 +02:00
|
|
|
{
|
rlp: improve nil pointer handling (#20064)
* rlp: improve nil pointer handling
In both encoder and decoder, the rules for encoding nil pointers were a
bit hard to understand, and didn't leave much choice. Since RLP allows
two empty values (empty list, empty string), any protocol built on RLP
must choose either of these values to represent the null value in a
certain context.
This change adds choice in the form of two new struct tags, "nilString"
and "nilList". These can be used to specify how a nil pointer value is
encoded. The "nil" tag still exists, but its implementation is now
explicit and defines exactly how nil pointers are handled in a single
place.
Another important change in this commit is how nil pointers and the
Encoder interface interact. The EncodeRLP method was previously called
even on nil values, which was supposed to give users a choice of how
their value would be handled when nil. It turns out this is a stupid
idea. If you create a network protocol containing an object defined in
another package, it's better to be able to say that the object should be
a list or string when nil in the definition of the protocol message
rather than defining the encoding of nil on the object itself.
As of this commit, the encoding rules for pointers now take precedence
over the Encoder interface rule. I think the "nil" tag will work fine
for most cases. For special kinds of objects which are a struct in Go
but strings in RLP, code using the object can specify the desired
encoding of nil using the "nilString" and "nilList" tags.
* rlp: propagate struct field type errors
If a struct contained fields of undecodable type, the encoder and
decoder would panic instead of returning an error. Fix this by
propagating type errors in makeStruct{Writer,Decoder} and add a test.
2019-09-13 12:10:57 +03:00
|
|
|
input: "C103",
|
|
|
|
ptr: new(intField),
|
|
|
|
error: "rlp: type int is not RLP-serializable (struct field rlp.intField.X)",
|
2015-12-21 22:05:20 +02:00
|
|
|
},
|
|
|
|
{
|
|
|
|
input: "C50102C20102",
|
|
|
|
ptr: new(tailUint),
|
|
|
|
error: "rlp: expected input string or byte for uint, decoding into (rlp.tailUint).Tail[1]",
|
|
|
|
},
|
rlp: improve nil pointer handling (#20064)
* rlp: improve nil pointer handling
In both encoder and decoder, the rules for encoding nil pointers were a
bit hard to understand, and didn't leave much choice. Since RLP allows
two empty values (empty list, empty string), any protocol built on RLP
must choose either of these values to represent the null value in a
certain context.
This change adds choice in the form of two new struct tags, "nilString"
and "nilList". These can be used to specify how a nil pointer value is
encoded. The "nil" tag still exists, but its implementation is now
explicit and defines exactly how nil pointers are handled in a single
place.
Another important change in this commit is how nil pointers and the
Encoder interface interact. The EncodeRLP method was previously called
even on nil values, which was supposed to give users a choice of how
their value would be handled when nil. It turns out this is a stupid
idea. If you create a network protocol containing an object defined in
another package, it's better to be able to say that the object should be
a list or string when nil in the definition of the protocol message
rather than defining the encoding of nil on the object itself.
As of this commit, the encoding rules for pointers now take precedence
over the Encoder interface rule. I think the "nil" tag will work fine
for most cases. For special kinds of objects which are a struct in Go
but strings in RLP, code using the object can specify the desired
encoding of nil using the "nilString" and "nilList" tags.
* rlp: propagate struct field type errors
If a struct contained fields of undecodable type, the encoder and
decoder would panic instead of returning an error. Fix this by
propagating type errors in makeStruct{Writer,Decoder} and add a test.
2019-09-13 12:10:57 +03:00
|
|
|
{
|
|
|
|
input: "C0",
|
|
|
|
ptr: new(invalidNilTag),
|
|
|
|
error: `rlp: invalid struct tag "nil" for rlp.invalidNilTag.X (field is not a pointer)`,
|
|
|
|
},
|
2015-12-21 22:05:20 +02:00
|
|
|
|
|
|
|
// struct tag "tail"
|
|
|
|
{
|
|
|
|
input: "C3010203",
|
|
|
|
ptr: new(tailRaw),
|
|
|
|
value: tailRaw{A: 1, Tail: []RawValue{unhex("02"), unhex("03")}},
|
|
|
|
},
|
|
|
|
{
|
|
|
|
input: "C20102",
|
|
|
|
ptr: new(tailRaw),
|
|
|
|
value: tailRaw{A: 1, Tail: []RawValue{unhex("02")}},
|
|
|
|
},
|
|
|
|
{
|
|
|
|
input: "C101",
|
|
|
|
ptr: new(tailRaw),
|
|
|
|
value: tailRaw{A: 1, Tail: []RawValue{}},
|
|
|
|
},
|
2019-05-14 16:09:56 +03:00
|
|
|
{
|
|
|
|
input: "C3010203",
|
|
|
|
ptr: new(tailPrivateFields),
|
|
|
|
value: tailPrivateFields{A: 1, Tail: []uint{2, 3}},
|
|
|
|
},
|
rlp: improve nil pointer handling (#20064)
* rlp: improve nil pointer handling
In both encoder and decoder, the rules for encoding nil pointers were a
bit hard to understand, and didn't leave much choice. Since RLP allows
two empty values (empty list, empty string), any protocol built on RLP
must choose either of these values to represent the null value in a
certain context.
This change adds choice in the form of two new struct tags, "nilString"
and "nilList". These can be used to specify how a nil pointer value is
encoded. The "nil" tag still exists, but its implementation is now
explicit and defines exactly how nil pointers are handled in a single
place.
Another important change in this commit is how nil pointers and the
Encoder interface interact. The EncodeRLP method was previously called
even on nil values, which was supposed to give users a choice of how
their value would be handled when nil. It turns out this is a stupid
idea. If you create a network protocol containing an object defined in
another package, it's better to be able to say that the object should be
a list or string when nil in the definition of the protocol message
rather than defining the encoding of nil on the object itself.
As of this commit, the encoding rules for pointers now take precedence
over the Encoder interface rule. I think the "nil" tag will work fine
for most cases. For special kinds of objects which are a struct in Go
but strings in RLP, code using the object can specify the desired
encoding of nil using the "nilString" and "nilList" tags.
* rlp: propagate struct field type errors
If a struct contained fields of undecodable type, the encoder and
decoder would panic instead of returning an error. Fix this by
propagating type errors in makeStruct{Writer,Decoder} and add a test.
2019-09-13 12:10:57 +03:00
|
|
|
{
|
|
|
|
input: "C0",
|
|
|
|
ptr: new(invalidTail1),
|
|
|
|
error: `rlp: invalid struct tag "tail" for rlp.invalidTail1.A (must be on last field)`,
|
|
|
|
},
|
|
|
|
{
|
|
|
|
input: "C0",
|
|
|
|
ptr: new(invalidTail2),
|
|
|
|
error: `rlp: invalid struct tag "tail" for rlp.invalidTail2.B (field type is not slice)`,
|
|
|
|
},
|
2015-01-05 18:10:42 +02:00
|
|
|
|
2017-03-07 14:37:53 +03:00
|
|
|
// struct tag "-"
|
|
|
|
{
|
|
|
|
input: "C20102",
|
2021-05-07 15:37:13 +03:00
|
|
|
ptr: new(ignoredField),
|
|
|
|
value: ignoredField{A: 1, C: 2},
|
2017-03-07 14:37:53 +03:00
|
|
|
},
|
|
|
|
|
rlp: improve nil pointer handling (#20064)
* rlp: improve nil pointer handling
In both encoder and decoder, the rules for encoding nil pointers were a
bit hard to understand, and didn't leave much choice. Since RLP allows
two empty values (empty list, empty string), any protocol built on RLP
must choose either of these values to represent the null value in a
certain context.
This change adds choice in the form of two new struct tags, "nilString"
and "nilList". These can be used to specify how a nil pointer value is
encoded. The "nil" tag still exists, but its implementation is now
explicit and defines exactly how nil pointers are handled in a single
place.
Another important change in this commit is how nil pointers and the
Encoder interface interact. The EncodeRLP method was previously called
even on nil values, which was supposed to give users a choice of how
their value would be handled when nil. It turns out this is a stupid
idea. If you create a network protocol containing an object defined in
another package, it's better to be able to say that the object should be
a list or string when nil in the definition of the protocol message
rather than defining the encoding of nil on the object itself.
As of this commit, the encoding rules for pointers now take precedence
over the Encoder interface rule. I think the "nil" tag will work fine
for most cases. For special kinds of objects which are a struct in Go
but strings in RLP, code using the object can specify the desired
encoding of nil using the "nilString" and "nilList" tags.
* rlp: propagate struct field type errors
If a struct contained fields of undecodable type, the encoder and
decoder would panic instead of returning an error. Fix this by
propagating type errors in makeStruct{Writer,Decoder} and add a test.
2019-09-13 12:10:57 +03:00
|
|
|
// struct tag "nilList"
|
|
|
|
{
|
|
|
|
input: "C180",
|
|
|
|
ptr: new(nilListUint),
|
|
|
|
error: "rlp: wrong kind of empty value (got String, want List) for *uint, decoding into (rlp.nilListUint).X",
|
|
|
|
},
|
|
|
|
{
|
|
|
|
input: "C1C0",
|
|
|
|
ptr: new(nilListUint),
|
|
|
|
value: nilListUint{},
|
|
|
|
},
|
|
|
|
{
|
|
|
|
input: "C103",
|
|
|
|
ptr: new(nilListUint),
|
|
|
|
value: func() interface{} {
|
|
|
|
v := uint(3)
|
|
|
|
return nilListUint{X: &v}
|
|
|
|
}(),
|
|
|
|
},
|
|
|
|
|
|
|
|
// struct tag "nilString"
|
|
|
|
{
|
|
|
|
input: "C1C0",
|
|
|
|
ptr: new(nilStringSlice),
|
|
|
|
error: "rlp: wrong kind of empty value (got List, want String) for *[]uint, decoding into (rlp.nilStringSlice).X",
|
|
|
|
},
|
|
|
|
{
|
|
|
|
input: "C180",
|
|
|
|
ptr: new(nilStringSlice),
|
|
|
|
value: nilStringSlice{},
|
|
|
|
},
|
|
|
|
{
|
|
|
|
input: "C2C103",
|
|
|
|
ptr: new(nilStringSlice),
|
|
|
|
value: nilStringSlice{X: &[]uint{3}},
|
|
|
|
},
|
|
|
|
|
2021-05-07 15:37:13 +03:00
|
|
|
// 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}},
|
|
|
|
},
|
2022-11-02 16:50:48 +03:00
|
|
|
{
|
|
|
|
// all optional fields nil
|
|
|
|
input: "C0",
|
|
|
|
ptr: new(multipleOptionalFields),
|
|
|
|
value: multipleOptionalFields{A: nil, B: nil},
|
|
|
|
},
|
|
|
|
{
|
|
|
|
// all optional fields set
|
|
|
|
input: "C88301020383010203",
|
|
|
|
ptr: new(multipleOptionalFields),
|
|
|
|
value: multipleOptionalFields{A: &[3]byte{1, 2, 3}, B: &[3]byte{1, 2, 3}},
|
|
|
|
},
|
|
|
|
{
|
|
|
|
// nil optional field appears before a non-nil one
|
|
|
|
input: "C58083010203",
|
|
|
|
ptr: new(multipleOptionalFields),
|
|
|
|
error: "rlp: input string too short for [3]uint8, decoding into (rlp.multipleOptionalFields).A",
|
|
|
|
},
|
|
|
|
{
|
|
|
|
// decode a nil ptr into a ptr that is not nil or not optional
|
|
|
|
input: "C20180",
|
|
|
|
ptr: new(nonOptionalPtrField),
|
|
|
|
error: "rlp: input string too short for [3]uint8, decoding into (rlp.nonOptionalPtrField).B",
|
|
|
|
},
|
2021-05-07 15:37:13 +03:00
|
|
|
{
|
|
|
|
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},
|
|
|
|
},
|
|
|
|
|
2015-09-09 04:34:15 +03:00
|
|
|
// RawValue
|
|
|
|
{input: "01", ptr: new(RawValue), value: RawValue(unhex("01"))},
|
|
|
|
{input: "82FFFF", ptr: new(RawValue), value: RawValue(unhex("82FFFF"))},
|
|
|
|
{input: "C20102", ptr: new([]RawValue), value: []RawValue{unhex("01"), unhex("02")}},
|
|
|
|
|
2014-11-13 21:31:48 +02:00
|
|
|
// pointers
|
2015-04-14 17:09:06 +03:00
|
|
|
{input: "00", ptr: new(*[]byte), value: &[]byte{0}},
|
2015-04-17 02:16:46 +03:00
|
|
|
{input: "80", ptr: new(*uint), value: uintp(0)},
|
|
|
|
{input: "C0", ptr: new(*uint), error: "rlp: expected input string or byte for uint"},
|
2014-12-09 11:41:47 +02:00
|
|
|
{input: "07", ptr: new(*uint), value: uintp(7)},
|
2015-07-18 17:13:20 +03:00
|
|
|
{input: "817F", ptr: new(*uint), error: "rlp: non-canonical size information for uint"},
|
|
|
|
{input: "8180", ptr: new(*uint), value: uintp(0x80)},
|
2014-12-09 11:41:47 +02:00
|
|
|
{input: "C109", ptr: new(*[]uint), value: &[]uint{9}},
|
2014-11-13 21:31:48 +02:00
|
|
|
{input: "C58403030303", ptr: new(*[][]byte), value: &[][]byte{{3, 3, 3, 3}}},
|
|
|
|
|
2015-03-21 01:49:31 +02:00
|
|
|
// check that input position is advanced also for empty values.
|
2015-04-17 02:16:46 +03:00
|
|
|
{input: "C3808005", ptr: new([]*uint), value: []*uint{uintp(0), uintp(0), uintp(5)}},
|
2014-11-17 13:02:08 +02:00
|
|
|
|
2014-11-13 21:31:48 +02:00
|
|
|
// interface{}
|
|
|
|
{input: "00", ptr: new(interface{}), value: []byte{0}},
|
|
|
|
{input: "01", ptr: new(interface{}), value: []byte{1}},
|
|
|
|
{input: "80", ptr: new(interface{}), value: []byte{}},
|
|
|
|
{input: "850505050505", ptr: new(interface{}), value: []byte{5, 5, 5, 5, 5}},
|
|
|
|
{input: "C0", ptr: new(interface{}), value: []interface{}{}},
|
|
|
|
{input: "C50183040404", ptr: new(interface{}), value: []interface{}{[]byte{1}, []byte{4, 4, 4}}},
|
2015-01-16 00:21:41 +02:00
|
|
|
{
|
|
|
|
input: "C3010203",
|
|
|
|
ptr: new([]io.Reader),
|
|
|
|
error: "rlp: type io.Reader is not RLP-serializable",
|
|
|
|
},
|
2015-04-28 11:28:15 +03:00
|
|
|
|
|
|
|
// fuzzer crashes
|
|
|
|
{
|
|
|
|
input: "c330f9c030f93030ce3030303030303030bd303030303030",
|
|
|
|
ptr: new(interface{}),
|
|
|
|
error: "rlp: element is larger than containing list",
|
|
|
|
},
|
2014-11-13 21:31:48 +02:00
|
|
|
}
|
|
|
|
|
2014-12-09 11:41:47 +02:00
|
|
|
func uintp(i uint) *uint { return &i }
|
2014-11-13 21:31:48 +02:00
|
|
|
|
2014-11-24 20:01:25 +02:00
|
|
|
func runTests(t *testing.T, decode func([]byte, interface{}) error) {
|
2014-11-13 21:31:48 +02:00
|
|
|
for i, test := range decodeTests {
|
|
|
|
input, err := hex.DecodeString(test.input)
|
|
|
|
if err != nil {
|
|
|
|
t.Errorf("test %d: invalid hex input %q", i, test.input)
|
|
|
|
continue
|
|
|
|
}
|
2014-11-24 20:01:25 +02:00
|
|
|
err = decode(input, test.ptr)
|
2014-11-25 17:00:48 +02:00
|
|
|
if err != nil && test.error == "" {
|
2014-11-13 21:31:48 +02:00
|
|
|
t.Errorf("test %d: unexpected Decode error: %v\ndecoding into %T\ninput %q",
|
|
|
|
i, err, test.ptr, test.input)
|
|
|
|
continue
|
|
|
|
}
|
2014-11-25 17:00:48 +02:00
|
|
|
if test.error != "" && fmt.Sprint(err) != test.error {
|
2014-11-13 21:31:48 +02:00
|
|
|
t.Errorf("test %d: Decode error mismatch\ngot %v\nwant %v\ndecoding into %T\ninput %q",
|
|
|
|
i, err, test.error, test.ptr, test.input)
|
|
|
|
continue
|
|
|
|
}
|
|
|
|
deref := reflect.ValueOf(test.ptr).Elem().Interface()
|
|
|
|
if err == nil && !reflect.DeepEqual(deref, test.value) {
|
|
|
|
t.Errorf("test %d: value mismatch\ngot %#v\nwant %#v\ndecoding into %T\ninput %q",
|
|
|
|
i, deref, test.value, test.ptr, test.input)
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2014-11-24 20:01:25 +02:00
|
|
|
func TestDecodeWithByteReader(t *testing.T) {
|
|
|
|
runTests(t, func(input []byte, into interface{}) error {
|
|
|
|
return Decode(bytes.NewReader(input), into)
|
|
|
|
})
|
|
|
|
}
|
|
|
|
|
2021-02-18 12:19:49 +03:00
|
|
|
func testDecodeWithEncReader(t *testing.T, n int) {
|
|
|
|
s := strings.Repeat("0", n)
|
|
|
|
_, r, _ := EncodeToReader(s)
|
|
|
|
var decoded string
|
|
|
|
err := Decode(r, &decoded)
|
|
|
|
if err != nil {
|
|
|
|
t.Errorf("Unexpected decode error with n=%v: %v", n, err)
|
|
|
|
}
|
|
|
|
if decoded != s {
|
|
|
|
t.Errorf("Decode mismatch with n=%v", n)
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
// This is a regression test checking that decoding from encReader
|
|
|
|
// works for RLP values of size 8192 bytes or more.
|
|
|
|
func TestDecodeWithEncReader(t *testing.T) {
|
|
|
|
testDecodeWithEncReader(t, 8188) // length with header is 8191
|
|
|
|
testDecodeWithEncReader(t, 8189) // length with header is 8192
|
|
|
|
}
|
|
|
|
|
2015-03-23 16:02:55 +02:00
|
|
|
// plainReader reads from a byte slice but does not
|
|
|
|
// implement ReadByte. It is also not recognized by the
|
|
|
|
// size validation. This is useful to test how the decoder
|
|
|
|
// behaves on a non-buffered input stream.
|
|
|
|
type plainReader []byte
|
|
|
|
|
|
|
|
func newPlainReader(b []byte) io.Reader {
|
|
|
|
return (*plainReader)(&b)
|
|
|
|
}
|
2014-11-24 20:01:25 +02:00
|
|
|
|
2015-03-23 16:02:55 +02:00
|
|
|
func (r *plainReader) Read(buf []byte) (n int, err error) {
|
2014-11-24 20:01:25 +02:00
|
|
|
if len(*r) == 0 {
|
|
|
|
return 0, io.EOF
|
|
|
|
}
|
|
|
|
n = copy(buf, *r)
|
|
|
|
*r = (*r)[n:]
|
|
|
|
return n, nil
|
|
|
|
}
|
|
|
|
|
|
|
|
func TestDecodeWithNonByteReader(t *testing.T) {
|
|
|
|
runTests(t, func(input []byte, into interface{}) error {
|
2015-03-23 16:02:55 +02:00
|
|
|
return Decode(newPlainReader(input), into)
|
2014-11-24 20:01:25 +02:00
|
|
|
})
|
|
|
|
}
|
|
|
|
|
|
|
|
func TestDecodeStreamReset(t *testing.T) {
|
2015-03-23 16:02:55 +02:00
|
|
|
s := NewStream(nil, 0)
|
2014-11-24 20:01:25 +02:00
|
|
|
runTests(t, func(input []byte, into interface{}) error {
|
2015-03-23 16:02:55 +02:00
|
|
|
s.Reset(bytes.NewReader(input), 0)
|
2014-11-24 20:01:25 +02:00
|
|
|
return s.Decode(into)
|
|
|
|
})
|
|
|
|
}
|
|
|
|
|
2014-11-13 21:31:48 +02:00
|
|
|
type testDecoder struct{ called bool }
|
|
|
|
|
|
|
|
func (t *testDecoder) DecodeRLP(s *Stream) error {
|
|
|
|
if _, err := s.Uint(); err != nil {
|
|
|
|
return err
|
|
|
|
}
|
|
|
|
t.called = true
|
|
|
|
return nil
|
|
|
|
}
|
|
|
|
|
|
|
|
func TestDecodeDecoder(t *testing.T) {
|
|
|
|
var s struct {
|
|
|
|
T1 testDecoder
|
|
|
|
T2 *testDecoder
|
|
|
|
T3 **testDecoder
|
|
|
|
}
|
|
|
|
if err := Decode(bytes.NewReader(unhex("C3010203")), &s); err != nil {
|
|
|
|
t.Fatalf("Decode error: %v", err)
|
|
|
|
}
|
|
|
|
|
|
|
|
if !s.T1.called {
|
|
|
|
t.Errorf("DecodeRLP was not called for (non-pointer) testDecoder")
|
|
|
|
}
|
|
|
|
|
|
|
|
if s.T2 == nil {
|
|
|
|
t.Errorf("*testDecoder has not been allocated")
|
|
|
|
} else if !s.T2.called {
|
|
|
|
t.Errorf("DecodeRLP was not called for *testDecoder")
|
|
|
|
}
|
|
|
|
|
|
|
|
if s.T3 == nil || *s.T3 == nil {
|
|
|
|
t.Errorf("**testDecoder has not been allocated")
|
|
|
|
} else if !(*s.T3).called {
|
|
|
|
t.Errorf("DecodeRLP was not called for **testDecoder")
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
rlp: improve nil pointer handling (#20064)
* rlp: improve nil pointer handling
In both encoder and decoder, the rules for encoding nil pointers were a
bit hard to understand, and didn't leave much choice. Since RLP allows
two empty values (empty list, empty string), any protocol built on RLP
must choose either of these values to represent the null value in a
certain context.
This change adds choice in the form of two new struct tags, "nilString"
and "nilList". These can be used to specify how a nil pointer value is
encoded. The "nil" tag still exists, but its implementation is now
explicit and defines exactly how nil pointers are handled in a single
place.
Another important change in this commit is how nil pointers and the
Encoder interface interact. The EncodeRLP method was previously called
even on nil values, which was supposed to give users a choice of how
their value would be handled when nil. It turns out this is a stupid
idea. If you create a network protocol containing an object defined in
another package, it's better to be able to say that the object should be
a list or string when nil in the definition of the protocol message
rather than defining the encoding of nil on the object itself.
As of this commit, the encoding rules for pointers now take precedence
over the Encoder interface rule. I think the "nil" tag will work fine
for most cases. For special kinds of objects which are a struct in Go
but strings in RLP, code using the object can specify the desired
encoding of nil using the "nilString" and "nilList" tags.
* rlp: propagate struct field type errors
If a struct contained fields of undecodable type, the encoder and
decoder would panic instead of returning an error. Fix this by
propagating type errors in makeStruct{Writer,Decoder} and add a test.
2019-09-13 12:10:57 +03:00
|
|
|
func TestDecodeDecoderNilPointer(t *testing.T) {
|
|
|
|
var s struct {
|
|
|
|
T1 *testDecoder `rlp:"nil"`
|
|
|
|
T2 *testDecoder
|
|
|
|
}
|
|
|
|
if err := Decode(bytes.NewReader(unhex("C2C002")), &s); err != nil {
|
|
|
|
t.Fatalf("Decode error: %v", err)
|
|
|
|
}
|
|
|
|
if s.T1 != nil {
|
|
|
|
t.Errorf("decoder T1 allocated for empty input (called: %v)", s.T1.called)
|
|
|
|
}
|
|
|
|
if s.T2 == nil || !s.T2.called {
|
|
|
|
t.Errorf("decoder T2 not allocated/called")
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2014-11-13 21:31:48 +02:00
|
|
|
type byteDecoder byte
|
|
|
|
|
|
|
|
func (bd *byteDecoder) DecodeRLP(s *Stream) error {
|
|
|
|
_, err := s.Uint()
|
|
|
|
*bd = 255
|
|
|
|
return err
|
|
|
|
}
|
|
|
|
|
|
|
|
func (bd byteDecoder) called() bool {
|
|
|
|
return bd == 255
|
|
|
|
}
|
|
|
|
|
|
|
|
// This test verifies that the byte slice/byte array logic
|
|
|
|
// does not kick in for element types implementing Decoder.
|
|
|
|
func TestDecoderInByteSlice(t *testing.T) {
|
|
|
|
var slice []byteDecoder
|
|
|
|
if err := Decode(bytes.NewReader(unhex("C101")), &slice); err != nil {
|
|
|
|
t.Errorf("unexpected Decode error %v", err)
|
|
|
|
} else if !slice[0].called() {
|
|
|
|
t.Errorf("DecodeRLP not called for slice element")
|
|
|
|
}
|
|
|
|
|
|
|
|
var array [1]byteDecoder
|
|
|
|
if err := Decode(bytes.NewReader(unhex("C101")), &array); err != nil {
|
|
|
|
t.Errorf("unexpected Decode error %v", err)
|
|
|
|
} else if !array[0].called() {
|
|
|
|
t.Errorf("DecodeRLP not called for array element")
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2019-05-14 16:09:56 +03:00
|
|
|
type unencodableDecoder func()
|
|
|
|
|
|
|
|
func (f *unencodableDecoder) DecodeRLP(s *Stream) error {
|
|
|
|
if _, err := s.List(); err != nil {
|
|
|
|
return err
|
|
|
|
}
|
|
|
|
if err := s.ListEnd(); err != nil {
|
|
|
|
return err
|
|
|
|
}
|
|
|
|
*f = func() {}
|
|
|
|
return nil
|
|
|
|
}
|
|
|
|
|
|
|
|
func TestDecoderFunc(t *testing.T) {
|
|
|
|
var x func()
|
|
|
|
if err := DecodeBytes([]byte{0xC0}, (*unencodableDecoder)(&x)); err != nil {
|
|
|
|
t.Fatal(err)
|
|
|
|
}
|
|
|
|
x()
|
|
|
|
}
|
|
|
|
|
2021-05-07 15:37:13 +03:00
|
|
|
// 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
|
|
|
|
}{
|
rlp/rlpgen: RLP encoder code generator (#24251)
This change adds a code generator tool for creating EncodeRLP method
implementations. The generated methods will behave identically to the
reflect-based encoder, but run faster because there is no reflection overhead.
Package rlp now provides the EncoderBuffer type for incremental encoding. This
is used by generated code, but the new methods can also be useful for
hand-written encoders.
There is also experimental support for generating DecodeRLP, and some new
methods have been added to the existing Stream type to support this. Creating
decoders with rlpgen is not recommended at this time because the generated
methods create very poor error reporting.
More detail about package rlp changes:
* rlp: externalize struct field processing / validation
This adds a new package, rlp/internal/rlpstruct, in preparation for the
RLP encoder generator.
I think the struct field rules are subtle enough to warrant extracting
this into their own package, even though it means that a bunch of
adapter code is needed for converting to/from rlpstruct.Type.
* rlp: add more decoder methods (for rlpgen)
This adds new methods on rlp.Stream:
- Uint64, Uint32, Uint16, Uint8, BigInt
- ReadBytes for decoding into []byte
- MoreDataInList - useful for optional list elements
* rlp: expose encoder buffer (for rlpgen)
This exposes the internal encoder buffer type for use in EncodeRLP
implementations.
The new EncoderBuffer type is a sort-of 'opaque handle' for a pointer to
encBuffer. It is implemented this way to ensure the global encBuffer pool
is handled correctly.
2022-02-16 20:14:12 +03:00
|
|
|
{v: new(invalid1), err: `rlp: invalid struct tag "" for rlp.invalid1.B (must be optional because preceding field "A" is optional)`},
|
2021-05-07 15:37:13 +03:00
|
|
|
{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())
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2014-11-13 21:31:48 +02:00
|
|
|
func ExampleDecode() {
|
|
|
|
input, _ := hex.DecodeString("C90A1486666F6F626172")
|
|
|
|
|
|
|
|
type example struct {
|
2019-11-25 15:41:53 +03:00
|
|
|
A, B uint
|
|
|
|
String string
|
2014-11-13 21:31:48 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
var s example
|
|
|
|
err := Decode(bytes.NewReader(input), &s)
|
|
|
|
if err != nil {
|
|
|
|
fmt.Printf("Error: %v\n", err)
|
|
|
|
} else {
|
|
|
|
fmt.Printf("Decoded value: %#v\n", s)
|
|
|
|
}
|
|
|
|
// Output:
|
2019-11-25 15:41:53 +03:00
|
|
|
// Decoded value: rlp.example{A:0xa, B:0x14, String:"foobar"}
|
2014-11-13 21:31:48 +02:00
|
|
|
}
|
|
|
|
|
2015-04-17 02:16:46 +03:00
|
|
|
func ExampleDecode_structTagNil() {
|
|
|
|
// In this example, we'll use the "nil" struct tag to change
|
|
|
|
// how a pointer-typed field is decoded. The input contains an RLP
|
|
|
|
// list of one element, an empty string.
|
|
|
|
input := []byte{0xC1, 0x80}
|
|
|
|
|
|
|
|
// This type uses the normal rules.
|
|
|
|
// The empty input string is decoded as a pointer to an empty Go string.
|
|
|
|
var normalRules struct {
|
|
|
|
String *string
|
|
|
|
}
|
|
|
|
Decode(bytes.NewReader(input), &normalRules)
|
|
|
|
fmt.Printf("normal: String = %q\n", *normalRules.String)
|
|
|
|
|
|
|
|
// This type uses the struct tag.
|
|
|
|
// The empty input string is decoded as a nil pointer.
|
|
|
|
var withEmptyOK struct {
|
|
|
|
String *string `rlp:"nil"`
|
|
|
|
}
|
|
|
|
Decode(bytes.NewReader(input), &withEmptyOK)
|
|
|
|
fmt.Printf("with nil tag: String = %v\n", withEmptyOK.String)
|
|
|
|
|
|
|
|
// Output:
|
|
|
|
// normal: String = ""
|
|
|
|
// with nil tag: String = <nil>
|
|
|
|
}
|
|
|
|
|
2014-11-13 21:31:48 +02:00
|
|
|
func ExampleStream() {
|
|
|
|
input, _ := hex.DecodeString("C90A1486666F6F626172")
|
2015-03-23 16:02:55 +02:00
|
|
|
s := NewStream(bytes.NewReader(input), 0)
|
2014-11-13 21:31:48 +02:00
|
|
|
|
|
|
|
// Check what kind of value lies ahead
|
|
|
|
kind, size, _ := s.Kind()
|
|
|
|
fmt.Printf("Kind: %v size:%d\n", kind, size)
|
|
|
|
|
|
|
|
// Enter the list
|
|
|
|
if _, err := s.List(); err != nil {
|
|
|
|
fmt.Printf("List error: %v\n", err)
|
|
|
|
return
|
|
|
|
}
|
|
|
|
|
|
|
|
// Decode elements
|
|
|
|
fmt.Println(s.Uint())
|
|
|
|
fmt.Println(s.Uint())
|
|
|
|
fmt.Println(s.Bytes())
|
|
|
|
|
|
|
|
// Acknowledge end of list
|
|
|
|
if err := s.ListEnd(); err != nil {
|
|
|
|
fmt.Printf("ListEnd error: %v\n", err)
|
|
|
|
}
|
|
|
|
// Output:
|
|
|
|
// Kind: List size:9
|
|
|
|
// 10 <nil>
|
|
|
|
// 20 <nil>
|
|
|
|
// [102 111 111 98 97 114] <nil>
|
|
|
|
}
|
|
|
|
|
2021-05-22 16:10:16 +03:00
|
|
|
func BenchmarkDecodeUints(b *testing.B) {
|
2014-12-30 12:40:32 +02:00
|
|
|
enc := encodeTestSlice(90000)
|
2014-11-13 21:31:48 +02:00
|
|
|
b.SetBytes(int64(len(enc)))
|
|
|
|
b.ReportAllocs()
|
|
|
|
b.ResetTimer()
|
|
|
|
|
|
|
|
for i := 0; i < b.N; i++ {
|
2015-01-15 12:12:39 +02:00
|
|
|
var s []uint
|
2014-11-13 21:31:48 +02:00
|
|
|
r := bytes.NewReader(enc)
|
|
|
|
if err := Decode(r, &s); err != nil {
|
|
|
|
b.Fatalf("Decode error: %v", err)
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2021-05-22 16:10:16 +03:00
|
|
|
func BenchmarkDecodeUintsReused(b *testing.B) {
|
2014-12-30 12:40:32 +02:00
|
|
|
enc := encodeTestSlice(100000)
|
2014-11-13 21:31:48 +02:00
|
|
|
b.SetBytes(int64(len(enc)))
|
|
|
|
b.ReportAllocs()
|
|
|
|
b.ResetTimer()
|
|
|
|
|
2015-01-15 12:12:39 +02:00
|
|
|
var s []uint
|
2014-11-13 21:31:48 +02:00
|
|
|
for i := 0; i < b.N; i++ {
|
|
|
|
r := bytes.NewReader(enc)
|
|
|
|
if err := Decode(r, &s); err != nil {
|
|
|
|
b.Fatalf("Decode error: %v", err)
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2021-05-22 16:10:16 +03:00
|
|
|
func BenchmarkDecodeByteArrayStruct(b *testing.B) {
|
|
|
|
enc, err := EncodeToBytes(&byteArrayStruct{})
|
|
|
|
if err != nil {
|
|
|
|
b.Fatal(err)
|
|
|
|
}
|
|
|
|
b.SetBytes(int64(len(enc)))
|
|
|
|
b.ReportAllocs()
|
|
|
|
b.ResetTimer()
|
|
|
|
|
|
|
|
var out byteArrayStruct
|
|
|
|
for i := 0; i < b.N; i++ {
|
|
|
|
if err := DecodeBytes(enc, &out); err != nil {
|
|
|
|
b.Fatal(err)
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
func BenchmarkDecodeBigInts(b *testing.B) {
|
|
|
|
ints := make([]*big.Int, 200)
|
|
|
|
for i := range ints {
|
|
|
|
ints[i] = math.BigPow(2, int64(i))
|
|
|
|
}
|
|
|
|
enc, err := EncodeToBytes(ints)
|
|
|
|
if err != nil {
|
|
|
|
b.Fatal(err)
|
|
|
|
}
|
|
|
|
b.SetBytes(int64(len(enc)))
|
|
|
|
b.ReportAllocs()
|
|
|
|
b.ResetTimer()
|
|
|
|
|
|
|
|
var out []*big.Int
|
|
|
|
for i := 0; i < b.N; i++ {
|
|
|
|
if err := DecodeBytes(enc, &out); err != nil {
|
|
|
|
b.Fatal(err)
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2023-03-17 13:51:55 +03:00
|
|
|
func BenchmarkDecodeU256Ints(b *testing.B) {
|
|
|
|
ints := make([]*uint256.Int, 200)
|
|
|
|
for i := range ints {
|
|
|
|
ints[i], _ = uint256.FromBig(math.BigPow(2, int64(i)))
|
|
|
|
}
|
|
|
|
enc, err := EncodeToBytes(ints)
|
|
|
|
if err != nil {
|
|
|
|
b.Fatal(err)
|
|
|
|
}
|
|
|
|
b.SetBytes(int64(len(enc)))
|
|
|
|
b.ReportAllocs()
|
|
|
|
b.ResetTimer()
|
|
|
|
|
|
|
|
var out []*uint256.Int
|
|
|
|
for i := 0; i < b.N; i++ {
|
|
|
|
if err := DecodeBytes(enc, &out); err != nil {
|
|
|
|
b.Fatal(err)
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2015-01-15 12:12:39 +02:00
|
|
|
func encodeTestSlice(n uint) []byte {
|
|
|
|
s := make([]uint, n)
|
|
|
|
for i := uint(0); i < n; i++ {
|
2014-11-13 21:31:48 +02:00
|
|
|
s[i] = i
|
|
|
|
}
|
2015-01-15 12:12:39 +02:00
|
|
|
b, err := EncodeToBytes(s)
|
|
|
|
if err != nil {
|
|
|
|
panic(fmt.Sprintf("encode error: %v", err))
|
|
|
|
}
|
|
|
|
return b
|
2014-11-13 21:31:48 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
func unhex(str string) []byte {
|
2022-05-09 13:13:23 +03:00
|
|
|
b, err := hex.DecodeString(strings.ReplaceAll(str, " ", ""))
|
2014-11-13 21:31:48 +02:00
|
|
|
if err != nil {
|
|
|
|
panic(fmt.Sprintf("invalid hex string: %q", str))
|
|
|
|
}
|
|
|
|
return b
|
|
|
|
}
|