status-go/vendor/github.com/multiformats/go-multihash/multihash.go

323 lines
8.9 KiB
Go

// Package multihash is the Go implementation of
// https://github.com/multiformats/multihash, or self-describing
// hashes.
package multihash
import (
"encoding/hex"
"errors"
"fmt"
"math"
b58 "github.com/mr-tron/base58/base58"
"github.com/multiformats/go-varint"
)
// errors
var (
ErrUnknownCode = errors.New("unknown multihash code")
ErrTooShort = errors.New("multihash too short. must be >= 2 bytes")
ErrTooLong = errors.New("multihash too long. must be < 129 bytes")
ErrLenNotSupported = errors.New("multihash does not yet support digests longer than 127 bytes")
ErrInvalidMultihash = errors.New("input isn't valid multihash")
ErrVarintBufferShort = errors.New("uvarint: buffer too small")
ErrVarintTooLong = errors.New("uvarint: varint too big (max 64bit)")
)
// ErrInconsistentLen is returned when a decoded multihash has an inconsistent length
type ErrInconsistentLen struct {
dm DecodedMultihash
lengthFound int
}
func (e ErrInconsistentLen) Error() string {
return fmt.Sprintf("multihash length inconsistent: expected %d; got %d", e.dm.Length, e.lengthFound)
}
// constants
const (
IDENTITY = 0x00
// Deprecated: use IDENTITY
ID = IDENTITY
SHA1 = 0x11
SHA2_256 = 0x12
SHA2_512 = 0x13
SHA3_224 = 0x17
SHA3_256 = 0x16
SHA3_384 = 0x15
SHA3_512 = 0x14
SHA3 = SHA3_512
KECCAK_224 = 0x1A
KECCAK_256 = 0x1B
KECCAK_384 = 0x1C
KECCAK_512 = 0x1D
BLAKE3 = 0x1E
SHAKE_128 = 0x18
SHAKE_256 = 0x19
BLAKE2B_MIN = 0xb201
BLAKE2B_MAX = 0xb240
BLAKE2S_MIN = 0xb241
BLAKE2S_MAX = 0xb260
MD5 = 0xd5
DBL_SHA2_256 = 0x56
MURMUR3X64_64 = 0x22
// Deprecated: use MURMUR3X64_64
MURMUR3 = MURMUR3X64_64
SHA2_256_TRUNC254_PADDED = 0x1012
X11 = 0x1100
POSEIDON_BLS12_381_A1_FC1 = 0xb401
)
func init() {
// Add blake2b (64 codes)
for c := uint64(BLAKE2B_MIN); c <= BLAKE2B_MAX; c++ {
n := c - BLAKE2B_MIN + 1
name := fmt.Sprintf("blake2b-%d", n*8)
Names[name] = c
Codes[c] = name
}
// Add blake2s (32 codes)
for c := uint64(BLAKE2S_MIN); c <= BLAKE2S_MAX; c++ {
n := c - BLAKE2S_MIN + 1
name := fmt.Sprintf("blake2s-%d", n*8)
Names[name] = c
Codes[c] = name
}
}
// Names maps the name of a hash to the code
var Names = map[string]uint64{
"identity": IDENTITY,
"sha1": SHA1,
"sha2-256": SHA2_256,
"sha2-512": SHA2_512,
"sha3": SHA3_512,
"sha3-224": SHA3_224,
"sha3-256": SHA3_256,
"sha3-384": SHA3_384,
"sha3-512": SHA3_512,
"dbl-sha2-256": DBL_SHA2_256,
"murmur3-x64-64": MURMUR3X64_64,
"keccak-224": KECCAK_224,
"keccak-256": KECCAK_256,
"keccak-384": KECCAK_384,
"keccak-512": KECCAK_512,
"blake3": BLAKE3,
"shake-128": SHAKE_128,
"shake-256": SHAKE_256,
"sha2-256-trunc254-padded": SHA2_256_TRUNC254_PADDED,
"x11": X11,
"md5": MD5,
"poseidon-bls12_381-a2-fc1": POSEIDON_BLS12_381_A1_FC1,
}
// Codes maps a hash code to it's name
var Codes = map[uint64]string{
IDENTITY: "identity",
SHA1: "sha1",
SHA2_256: "sha2-256",
SHA2_512: "sha2-512",
SHA3_224: "sha3-224",
SHA3_256: "sha3-256",
SHA3_384: "sha3-384",
SHA3_512: "sha3-512",
DBL_SHA2_256: "dbl-sha2-256",
MURMUR3X64_64: "murmur3-x64-64",
KECCAK_224: "keccak-224",
KECCAK_256: "keccak-256",
KECCAK_384: "keccak-384",
KECCAK_512: "keccak-512",
BLAKE3: "blake3",
SHAKE_128: "shake-128",
SHAKE_256: "shake-256",
SHA2_256_TRUNC254_PADDED: "sha2-256-trunc254-padded",
X11: "x11",
POSEIDON_BLS12_381_A1_FC1: "poseidon-bls12_381-a2-fc1",
MD5: "md5",
}
// reads a varint from buf and returns bytes read.
func uvarint(buf []byte) (uint64, []byte, error) {
n, c, err := varint.FromUvarint(buf)
if err != nil {
return n, buf, err
}
if c == 0 {
return n, buf, ErrVarintBufferShort
} else if c < 0 {
return n, buf[-c:], ErrVarintTooLong
} else {
return n, buf[c:], nil
}
}
// DecodedMultihash represents a parsed multihash and allows
// easy access to the different parts of a multihash.
type DecodedMultihash struct {
Code uint64
Name string
Length int // Length is just int as it is type of len() opearator
Digest []byte // Digest holds the raw multihash bytes
}
// Multihash is byte slice with the following form:
// <hash function code><digest size><hash function output>.
// See the spec for more information.
type Multihash []byte
// HexString returns the hex-encoded representation of a multihash.
func (m Multihash) HexString() string {
return hex.EncodeToString([]byte(m))
}
// String is an alias to HexString().
func (m Multihash) String() string {
return m.HexString()
}
// FromHexString parses a hex-encoded multihash.
func FromHexString(s string) (Multihash, error) {
b, err := hex.DecodeString(s)
if err != nil {
return Multihash{}, err
}
return Cast(b)
}
// B58String returns the B58-encoded representation of a multihash.
func (m Multihash) B58String() string {
return b58.Encode([]byte(m))
}
// FromB58String parses a B58-encoded multihash.
func FromB58String(s string) (m Multihash, err error) {
b, err := b58.Decode(s)
if err != nil {
return Multihash{}, ErrInvalidMultihash
}
return Cast(b)
}
// Cast casts a buffer onto a multihash, and returns an error
// if it does not work.
func Cast(buf []byte) (Multihash, error) {
_, err := Decode(buf)
if err != nil {
return Multihash{}, err
}
return Multihash(buf), nil
}
// Decode parses multihash bytes into a DecodedMultihash.
func Decode(buf []byte) (*DecodedMultihash, error) {
// outline decode allowing the &dm expression to be inlined into the caller.
// This moves the heap allocation into the caller and if the caller doesn't
// leak dm the compiler will use a stack allocation instead.
// If you do not outline this &dm always heap allocate since the pointer is
// returned which cause a heap allocation because Decode's stack frame is
// about to disapear.
dm, err := decode(buf)
if err != nil {
return nil, err
}
return &dm, nil
}
func decode(buf []byte) (dm DecodedMultihash, err error) {
rlen, code, hdig, err := readMultihashFromBuf(buf)
if err != nil {
return DecodedMultihash{}, err
}
dm = DecodedMultihash{
Code: code,
Name: Codes[code],
Length: len(hdig),
Digest: hdig,
}
if len(buf) != rlen {
return dm, ErrInconsistentLen{dm, rlen}
}
return dm, nil
}
// Encode a hash digest along with the specified function code.
// Note: the length is derived from the length of the digest itself.
//
// The error return is legacy; it is always nil.
func Encode(buf []byte, code uint64) ([]byte, error) {
// FUTURE: this function always causes heap allocs... but when used, this value is almost always going to be appended to another buffer (either as part of CID creation, or etc) -- should this whole function be rethought and alternatives offered?
newBuf := make([]byte, varint.UvarintSize(code)+varint.UvarintSize(uint64(len(buf)))+len(buf))
n := varint.PutUvarint(newBuf, code)
n += varint.PutUvarint(newBuf[n:], uint64(len(buf)))
copy(newBuf[n:], buf)
return newBuf, nil
}
// EncodeName is like Encode() but providing a string name
// instead of a numeric code. See Names for allowed values.
func EncodeName(buf []byte, name string) ([]byte, error) {
return Encode(buf, Names[name])
}
// readMultihashFromBuf reads a multihash from the given buffer, returning the
// individual pieces of the multihash.
// Note: the returned digest is a slice over the passed in data and should be
// copied if the buffer will be reused
func readMultihashFromBuf(buf []byte) (int, uint64, []byte, error) {
initBufLength := len(buf)
if initBufLength < 2 {
return 0, 0, nil, ErrTooShort
}
var err error
var code, length uint64
code, buf, err = uvarint(buf)
if err != nil {
return 0, 0, nil, err
}
length, buf, err = uvarint(buf)
if err != nil {
return 0, 0, nil, err
}
if length > math.MaxInt32 {
return 0, 0, nil, errors.New("digest too long, supporting only <= 2^31-1")
}
if int(length) > len(buf) {
return 0, 0, nil, errors.New("length greater than remaining number of bytes in buffer")
}
// rlen is the advertised size of the CID
rlen := (initBufLength - len(buf)) + int(length)
return rlen, code, buf[:length], nil
}
// MHFromBytes reads a multihash from the given byte buffer, returning the
// number of bytes read as well as the multihash
func MHFromBytes(buf []byte) (int, Multihash, error) {
nr, _, _, err := readMultihashFromBuf(buf)
if err != nil {
return 0, nil, err
}
return nr, Multihash(buf[:nr]), nil
}