diff options
Diffstat (limited to 'src/encoding/binary')
-rw-r--r-- | src/encoding/binary/binary.go | 737 | ||||
-rw-r--r-- | src/encoding/binary/binary_test.go | 731 | ||||
-rw-r--r-- | src/encoding/binary/example_test.go | 187 | ||||
-rw-r--r-- | src/encoding/binary/varint.go | 139 | ||||
-rw-r--r-- | src/encoding/binary/varint_test.go | 231 |
5 files changed, 2025 insertions, 0 deletions
diff --git a/src/encoding/binary/binary.go b/src/encoding/binary/binary.go new file mode 100644 index 0000000..a311499 --- /dev/null +++ b/src/encoding/binary/binary.go @@ -0,0 +1,737 @@ +// Copyright 2009 The Go Authors. All rights reserved. +// Use of this source code is governed by a BSD-style +// license that can be found in the LICENSE file. + +// Package binary implements simple translation between numbers and byte +// sequences and encoding and decoding of varints. +// +// Numbers are translated by reading and writing fixed-size values. +// A fixed-size value is either a fixed-size arithmetic +// type (bool, int8, uint8, int16, float32, complex64, ...) +// or an array or struct containing only fixed-size values. +// +// The varint functions encode and decode single integer values using +// a variable-length encoding; smaller values require fewer bytes. +// For a specification, see +// https://developers.google.com/protocol-buffers/docs/encoding. +// +// This package favors simplicity over efficiency. Clients that require +// high-performance serialization, especially for large data structures, +// should look at more advanced solutions such as the encoding/gob +// package or protocol buffers. +package binary + +import ( + "errors" + "io" + "math" + "reflect" + "sync" +) + +// A ByteOrder specifies how to convert byte sequences into +// 16-, 32-, or 64-bit unsigned integers. +type ByteOrder interface { + Uint16([]byte) uint16 + Uint32([]byte) uint32 + Uint64([]byte) uint64 + PutUint16([]byte, uint16) + PutUint32([]byte, uint32) + PutUint64([]byte, uint64) + String() string +} + +// LittleEndian is the little-endian implementation of ByteOrder. +var LittleEndian littleEndian + +// BigEndian is the big-endian implementation of ByteOrder. +var BigEndian bigEndian + +type littleEndian struct{} + +func (littleEndian) Uint16(b []byte) uint16 { + _ = b[1] // bounds check hint to compiler; see golang.org/issue/14808 + return uint16(b[0]) | uint16(b[1])<<8 +} + +func (littleEndian) PutUint16(b []byte, v uint16) { + _ = b[1] // early bounds check to guarantee safety of writes below + b[0] = byte(v) + b[1] = byte(v >> 8) +} + +func (littleEndian) Uint32(b []byte) uint32 { + _ = b[3] // bounds check hint to compiler; see golang.org/issue/14808 + return uint32(b[0]) | uint32(b[1])<<8 | uint32(b[2])<<16 | uint32(b[3])<<24 +} + +func (littleEndian) PutUint32(b []byte, v uint32) { + _ = b[3] // early bounds check to guarantee safety of writes below + b[0] = byte(v) + b[1] = byte(v >> 8) + b[2] = byte(v >> 16) + b[3] = byte(v >> 24) +} + +func (littleEndian) Uint64(b []byte) uint64 { + _ = b[7] // bounds check hint to compiler; see golang.org/issue/14808 + return uint64(b[0]) | uint64(b[1])<<8 | uint64(b[2])<<16 | uint64(b[3])<<24 | + uint64(b[4])<<32 | uint64(b[5])<<40 | uint64(b[6])<<48 | uint64(b[7])<<56 +} + +func (littleEndian) PutUint64(b []byte, v uint64) { + _ = b[7] // early bounds check to guarantee safety of writes below + b[0] = byte(v) + b[1] = byte(v >> 8) + b[2] = byte(v >> 16) + b[3] = byte(v >> 24) + b[4] = byte(v >> 32) + b[5] = byte(v >> 40) + b[6] = byte(v >> 48) + b[7] = byte(v >> 56) +} + +func (littleEndian) String() string { return "LittleEndian" } + +func (littleEndian) GoString() string { return "binary.LittleEndian" } + +type bigEndian struct{} + +func (bigEndian) Uint16(b []byte) uint16 { + _ = b[1] // bounds check hint to compiler; see golang.org/issue/14808 + return uint16(b[1]) | uint16(b[0])<<8 +} + +func (bigEndian) PutUint16(b []byte, v uint16) { + _ = b[1] // early bounds check to guarantee safety of writes below + b[0] = byte(v >> 8) + b[1] = byte(v) +} + +func (bigEndian) Uint32(b []byte) uint32 { + _ = b[3] // bounds check hint to compiler; see golang.org/issue/14808 + return uint32(b[3]) | uint32(b[2])<<8 | uint32(b[1])<<16 | uint32(b[0])<<24 +} + +func (bigEndian) PutUint32(b []byte, v uint32) { + _ = b[3] // early bounds check to guarantee safety of writes below + b[0] = byte(v >> 24) + b[1] = byte(v >> 16) + b[2] = byte(v >> 8) + b[3] = byte(v) +} + +func (bigEndian) Uint64(b []byte) uint64 { + _ = b[7] // bounds check hint to compiler; see golang.org/issue/14808 + return uint64(b[7]) | uint64(b[6])<<8 | uint64(b[5])<<16 | uint64(b[4])<<24 | + uint64(b[3])<<32 | uint64(b[2])<<40 | uint64(b[1])<<48 | uint64(b[0])<<56 +} + +func (bigEndian) PutUint64(b []byte, v uint64) { + _ = b[7] // early bounds check to guarantee safety of writes below + b[0] = byte(v >> 56) + b[1] = byte(v >> 48) + b[2] = byte(v >> 40) + b[3] = byte(v >> 32) + b[4] = byte(v >> 24) + b[5] = byte(v >> 16) + b[6] = byte(v >> 8) + b[7] = byte(v) +} + +func (bigEndian) String() string { return "BigEndian" } + +func (bigEndian) GoString() string { return "binary.BigEndian" } + +// Read reads structured binary data from r into data. +// Data must be a pointer to a fixed-size value or a slice +// of fixed-size values. +// Bytes read from r are decoded using the specified byte order +// and written to successive fields of the data. +// When decoding boolean values, a zero byte is decoded as false, and +// any other non-zero byte is decoded as true. +// When reading into structs, the field data for fields with +// blank (_) field names is skipped; i.e., blank field names +// may be used for padding. +// When reading into a struct, all non-blank fields must be exported +// or Read may panic. +// +// The error is EOF only if no bytes were read. +// If an EOF happens after reading some but not all the bytes, +// Read returns ErrUnexpectedEOF. +func Read(r io.Reader, order ByteOrder, data interface{}) error { + // Fast path for basic types and slices. + if n := intDataSize(data); n != 0 { + bs := make([]byte, n) + if _, err := io.ReadFull(r, bs); err != nil { + return err + } + switch data := data.(type) { + case *bool: + *data = bs[0] != 0 + case *int8: + *data = int8(bs[0]) + case *uint8: + *data = bs[0] + case *int16: + *data = int16(order.Uint16(bs)) + case *uint16: + *data = order.Uint16(bs) + case *int32: + *data = int32(order.Uint32(bs)) + case *uint32: + *data = order.Uint32(bs) + case *int64: + *data = int64(order.Uint64(bs)) + case *uint64: + *data = order.Uint64(bs) + case *float32: + *data = math.Float32frombits(order.Uint32(bs)) + case *float64: + *data = math.Float64frombits(order.Uint64(bs)) + case []bool: + for i, x := range bs { // Easier to loop over the input for 8-bit values. + data[i] = x != 0 + } + case []int8: + for i, x := range bs { + data[i] = int8(x) + } + case []uint8: + copy(data, bs) + case []int16: + for i := range data { + data[i] = int16(order.Uint16(bs[2*i:])) + } + case []uint16: + for i := range data { + data[i] = order.Uint16(bs[2*i:]) + } + case []int32: + for i := range data { + data[i] = int32(order.Uint32(bs[4*i:])) + } + case []uint32: + for i := range data { + data[i] = order.Uint32(bs[4*i:]) + } + case []int64: + for i := range data { + data[i] = int64(order.Uint64(bs[8*i:])) + } + case []uint64: + for i := range data { + data[i] = order.Uint64(bs[8*i:]) + } + case []float32: + for i := range data { + data[i] = math.Float32frombits(order.Uint32(bs[4*i:])) + } + case []float64: + for i := range data { + data[i] = math.Float64frombits(order.Uint64(bs[8*i:])) + } + default: + n = 0 // fast path doesn't apply + } + if n != 0 { + return nil + } + } + + // Fallback to reflect-based decoding. + v := reflect.ValueOf(data) + size := -1 + switch v.Kind() { + case reflect.Ptr: + v = v.Elem() + size = dataSize(v) + case reflect.Slice: + size = dataSize(v) + } + if size < 0 { + return errors.New("binary.Read: invalid type " + reflect.TypeOf(data).String()) + } + d := &decoder{order: order, buf: make([]byte, size)} + if _, err := io.ReadFull(r, d.buf); err != nil { + return err + } + d.value(v) + return nil +} + +// Write writes the binary representation of data into w. +// Data must be a fixed-size value or a slice of fixed-size +// values, or a pointer to such data. +// Boolean values encode as one byte: 1 for true, and 0 for false. +// Bytes written to w are encoded using the specified byte order +// and read from successive fields of the data. +// When writing structs, zero values are written for fields +// with blank (_) field names. +func Write(w io.Writer, order ByteOrder, data interface{}) error { + // Fast path for basic types and slices. + if n := intDataSize(data); n != 0 { + bs := make([]byte, n) + switch v := data.(type) { + case *bool: + if *v { + bs[0] = 1 + } else { + bs[0] = 0 + } + case bool: + if v { + bs[0] = 1 + } else { + bs[0] = 0 + } + case []bool: + for i, x := range v { + if x { + bs[i] = 1 + } else { + bs[i] = 0 + } + } + case *int8: + bs[0] = byte(*v) + case int8: + bs[0] = byte(v) + case []int8: + for i, x := range v { + bs[i] = byte(x) + } + case *uint8: + bs[0] = *v + case uint8: + bs[0] = v + case []uint8: + bs = v + case *int16: + order.PutUint16(bs, uint16(*v)) + case int16: + order.PutUint16(bs, uint16(v)) + case []int16: + for i, x := range v { + order.PutUint16(bs[2*i:], uint16(x)) + } + case *uint16: + order.PutUint16(bs, *v) + case uint16: + order.PutUint16(bs, v) + case []uint16: + for i, x := range v { + order.PutUint16(bs[2*i:], x) + } + case *int32: + order.PutUint32(bs, uint32(*v)) + case int32: + order.PutUint32(bs, uint32(v)) + case []int32: + for i, x := range v { + order.PutUint32(bs[4*i:], uint32(x)) + } + case *uint32: + order.PutUint32(bs, *v) + case uint32: + order.PutUint32(bs, v) + case []uint32: + for i, x := range v { + order.PutUint32(bs[4*i:], x) + } + case *int64: + order.PutUint64(bs, uint64(*v)) + case int64: + order.PutUint64(bs, uint64(v)) + case []int64: + for i, x := range v { + order.PutUint64(bs[8*i:], uint64(x)) + } + case *uint64: + order.PutUint64(bs, *v) + case uint64: + order.PutUint64(bs, v) + case []uint64: + for i, x := range v { + order.PutUint64(bs[8*i:], x) + } + case *float32: + order.PutUint32(bs, math.Float32bits(*v)) + case float32: + order.PutUint32(bs, math.Float32bits(v)) + case []float32: + for i, x := range v { + order.PutUint32(bs[4*i:], math.Float32bits(x)) + } + case *float64: + order.PutUint64(bs, math.Float64bits(*v)) + case float64: + order.PutUint64(bs, math.Float64bits(v)) + case []float64: + for i, x := range v { + order.PutUint64(bs[8*i:], math.Float64bits(x)) + } + } + _, err := w.Write(bs) + return err + } + + // Fallback to reflect-based encoding. + v := reflect.Indirect(reflect.ValueOf(data)) + size := dataSize(v) + if size < 0 { + return errors.New("binary.Write: invalid type " + reflect.TypeOf(data).String()) + } + buf := make([]byte, size) + e := &encoder{order: order, buf: buf} + e.value(v) + _, err := w.Write(buf) + return err +} + +// Size returns how many bytes Write would generate to encode the value v, which +// must be a fixed-size value or a slice of fixed-size values, or a pointer to such data. +// If v is neither of these, Size returns -1. +func Size(v interface{}) int { + return dataSize(reflect.Indirect(reflect.ValueOf(v))) +} + +var structSize sync.Map // map[reflect.Type]int + +// dataSize returns the number of bytes the actual data represented by v occupies in memory. +// For compound structures, it sums the sizes of the elements. Thus, for instance, for a slice +// it returns the length of the slice times the element size and does not count the memory +// occupied by the header. If the type of v is not acceptable, dataSize returns -1. +func dataSize(v reflect.Value) int { + switch v.Kind() { + case reflect.Slice: + if s := sizeof(v.Type().Elem()); s >= 0 { + return s * v.Len() + } + return -1 + + case reflect.Struct: + t := v.Type() + if size, ok := structSize.Load(t); ok { + return size.(int) + } + size := sizeof(t) + structSize.Store(t, size) + return size + + default: + return sizeof(v.Type()) + } +} + +// sizeof returns the size >= 0 of variables for the given type or -1 if the type is not acceptable. +func sizeof(t reflect.Type) int { + switch t.Kind() { + case reflect.Array: + if s := sizeof(t.Elem()); s >= 0 { + return s * t.Len() + } + + case reflect.Struct: + sum := 0 + for i, n := 0, t.NumField(); i < n; i++ { + s := sizeof(t.Field(i).Type) + if s < 0 { + return -1 + } + sum += s + } + return sum + + case reflect.Bool, + reflect.Uint8, reflect.Uint16, reflect.Uint32, reflect.Uint64, + reflect.Int8, reflect.Int16, reflect.Int32, reflect.Int64, + reflect.Float32, reflect.Float64, reflect.Complex64, reflect.Complex128: + return int(t.Size()) + } + + return -1 +} + +type coder struct { + order ByteOrder + buf []byte + offset int +} + +type decoder coder +type encoder coder + +func (d *decoder) bool() bool { + x := d.buf[d.offset] + d.offset++ + return x != 0 +} + +func (e *encoder) bool(x bool) { + if x { + e.buf[e.offset] = 1 + } else { + e.buf[e.offset] = 0 + } + e.offset++ +} + +func (d *decoder) uint8() uint8 { + x := d.buf[d.offset] + d.offset++ + return x +} + +func (e *encoder) uint8(x uint8) { + e.buf[e.offset] = x + e.offset++ +} + +func (d *decoder) uint16() uint16 { + x := d.order.Uint16(d.buf[d.offset : d.offset+2]) + d.offset += 2 + return x +} + +func (e *encoder) uint16(x uint16) { + e.order.PutUint16(e.buf[e.offset:e.offset+2], x) + e.offset += 2 +} + +func (d *decoder) uint32() uint32 { + x := d.order.Uint32(d.buf[d.offset : d.offset+4]) + d.offset += 4 + return x +} + +func (e *encoder) uint32(x uint32) { + e.order.PutUint32(e.buf[e.offset:e.offset+4], x) + e.offset += 4 +} + +func (d *decoder) uint64() uint64 { + x := d.order.Uint64(d.buf[d.offset : d.offset+8]) + d.offset += 8 + return x +} + +func (e *encoder) uint64(x uint64) { + e.order.PutUint64(e.buf[e.offset:e.offset+8], x) + e.offset += 8 +} + +func (d *decoder) int8() int8 { return int8(d.uint8()) } + +func (e *encoder) int8(x int8) { e.uint8(uint8(x)) } + +func (d *decoder) int16() int16 { return int16(d.uint16()) } + +func (e *encoder) int16(x int16) { e.uint16(uint16(x)) } + +func (d *decoder) int32() int32 { return int32(d.uint32()) } + +func (e *encoder) int32(x int32) { e.uint32(uint32(x)) } + +func (d *decoder) int64() int64 { return int64(d.uint64()) } + +func (e *encoder) int64(x int64) { e.uint64(uint64(x)) } + +func (d *decoder) value(v reflect.Value) { + switch v.Kind() { + case reflect.Array: + l := v.Len() + for i := 0; i < l; i++ { + d.value(v.Index(i)) + } + + case reflect.Struct: + t := v.Type() + l := v.NumField() + for i := 0; i < l; i++ { + // Note: Calling v.CanSet() below is an optimization. + // It would be sufficient to check the field name, + // but creating the StructField info for each field is + // costly (run "go test -bench=ReadStruct" and compare + // results when making changes to this code). + if v := v.Field(i); v.CanSet() || t.Field(i).Name != "_" { + d.value(v) + } else { + d.skip(v) + } + } + + case reflect.Slice: + l := v.Len() + for i := 0; i < l; i++ { + d.value(v.Index(i)) + } + + case reflect.Bool: + v.SetBool(d.bool()) + + case reflect.Int8: + v.SetInt(int64(d.int8())) + case reflect.Int16: + v.SetInt(int64(d.int16())) + case reflect.Int32: + v.SetInt(int64(d.int32())) + case reflect.Int64: + v.SetInt(d.int64()) + + case reflect.Uint8: + v.SetUint(uint64(d.uint8())) + case reflect.Uint16: + v.SetUint(uint64(d.uint16())) + case reflect.Uint32: + v.SetUint(uint64(d.uint32())) + case reflect.Uint64: + v.SetUint(d.uint64()) + + case reflect.Float32: + v.SetFloat(float64(math.Float32frombits(d.uint32()))) + case reflect.Float64: + v.SetFloat(math.Float64frombits(d.uint64())) + + case reflect.Complex64: + v.SetComplex(complex( + float64(math.Float32frombits(d.uint32())), + float64(math.Float32frombits(d.uint32())), + )) + case reflect.Complex128: + v.SetComplex(complex( + math.Float64frombits(d.uint64()), + math.Float64frombits(d.uint64()), + )) + } +} + +func (e *encoder) value(v reflect.Value) { + switch v.Kind() { + case reflect.Array: + l := v.Len() + for i := 0; i < l; i++ { + e.value(v.Index(i)) + } + + case reflect.Struct: + t := v.Type() + l := v.NumField() + for i := 0; i < l; i++ { + // see comment for corresponding code in decoder.value() + if v := v.Field(i); v.CanSet() || t.Field(i).Name != "_" { + e.value(v) + } else { + e.skip(v) + } + } + + case reflect.Slice: + l := v.Len() + for i := 0; i < l; i++ { + e.value(v.Index(i)) + } + + case reflect.Bool: + e.bool(v.Bool()) + + case reflect.Int, reflect.Int8, reflect.Int16, reflect.Int32, reflect.Int64: + switch v.Type().Kind() { + case reflect.Int8: + e.int8(int8(v.Int())) + case reflect.Int16: + e.int16(int16(v.Int())) + case reflect.Int32: + e.int32(int32(v.Int())) + case reflect.Int64: + e.int64(v.Int()) + } + + case reflect.Uint, reflect.Uint8, reflect.Uint16, reflect.Uint32, reflect.Uint64, reflect.Uintptr: + switch v.Type().Kind() { + case reflect.Uint8: + e.uint8(uint8(v.Uint())) + case reflect.Uint16: + e.uint16(uint16(v.Uint())) + case reflect.Uint32: + e.uint32(uint32(v.Uint())) + case reflect.Uint64: + e.uint64(v.Uint()) + } + + case reflect.Float32, reflect.Float64: + switch v.Type().Kind() { + case reflect.Float32: + e.uint32(math.Float32bits(float32(v.Float()))) + case reflect.Float64: + e.uint64(math.Float64bits(v.Float())) + } + + case reflect.Complex64, reflect.Complex128: + switch v.Type().Kind() { + case reflect.Complex64: + x := v.Complex() + e.uint32(math.Float32bits(float32(real(x)))) + e.uint32(math.Float32bits(float32(imag(x)))) + case reflect.Complex128: + x := v.Complex() + e.uint64(math.Float64bits(real(x))) + e.uint64(math.Float64bits(imag(x))) + } + } +} + +func (d *decoder) skip(v reflect.Value) { + d.offset += dataSize(v) +} + +func (e *encoder) skip(v reflect.Value) { + n := dataSize(v) + zero := e.buf[e.offset : e.offset+n] + for i := range zero { + zero[i] = 0 + } + e.offset += n +} + +// intDataSize returns the size of the data required to represent the data when encoded. +// It returns zero if the type cannot be implemented by the fast path in Read or Write. +func intDataSize(data interface{}) int { + switch data := data.(type) { + case bool, int8, uint8, *bool, *int8, *uint8: + return 1 + case []bool: + return len(data) + case []int8: + return len(data) + case []uint8: + return len(data) + case int16, uint16, *int16, *uint16: + return 2 + case []int16: + return 2 * len(data) + case []uint16: + return 2 * len(data) + case int32, uint32, *int32, *uint32: + return 4 + case []int32: + return 4 * len(data) + case []uint32: + return 4 * len(data) + case int64, uint64, *int64, *uint64: + return 8 + case []int64: + return 8 * len(data) + case []uint64: + return 8 * len(data) + case float32, *float32: + return 4 + case float64, *float64: + return 8 + case []float32: + return 4 * len(data) + case []float64: + return 8 * len(data) + } + return 0 +} diff --git a/src/encoding/binary/binary_test.go b/src/encoding/binary/binary_test.go new file mode 100644 index 0000000..83af89e --- /dev/null +++ b/src/encoding/binary/binary_test.go @@ -0,0 +1,731 @@ +// Copyright 2009 The Go Authors. All rights reserved. +// Use of this source code is governed by a BSD-style +// license that can be found in the LICENSE file. + +package binary + +import ( + "bytes" + "fmt" + "io" + "math" + "reflect" + "strings" + "sync" + "testing" +) + +type Struct struct { + Int8 int8 + Int16 int16 + Int32 int32 + Int64 int64 + Uint8 uint8 + Uint16 uint16 + Uint32 uint32 + Uint64 uint64 + Float32 float32 + Float64 float64 + Complex64 complex64 + Complex128 complex128 + Array [4]uint8 + Bool bool + BoolArray [4]bool +} + +type T struct { + Int int + Uint uint + Uintptr uintptr + Array [4]int +} + +var s = Struct{ + 0x01, + 0x0203, + 0x04050607, + 0x08090a0b0c0d0e0f, + 0x10, + 0x1112, + 0x13141516, + 0x1718191a1b1c1d1e, + + math.Float32frombits(0x1f202122), + math.Float64frombits(0x232425262728292a), + complex( + math.Float32frombits(0x2b2c2d2e), + math.Float32frombits(0x2f303132), + ), + complex( + math.Float64frombits(0x333435363738393a), + math.Float64frombits(0x3b3c3d3e3f404142), + ), + + [4]uint8{0x43, 0x44, 0x45, 0x46}, + + true, + [4]bool{true, false, true, false}, +} + +var big = []byte{ + 1, + 2, 3, + 4, 5, 6, 7, + 8, 9, 10, 11, 12, 13, 14, 15, + 16, + 17, 18, + 19, 20, 21, 22, + 23, 24, 25, 26, 27, 28, 29, 30, + + 31, 32, 33, 34, + 35, 36, 37, 38, 39, 40, 41, 42, + 43, 44, 45, 46, 47, 48, 49, 50, + 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, + + 67, 68, 69, 70, + + 1, + 1, 0, 1, 0, +} + +var little = []byte{ + 1, + 3, 2, + 7, 6, 5, 4, + 15, 14, 13, 12, 11, 10, 9, 8, + 16, + 18, 17, + 22, 21, 20, 19, + 30, 29, 28, 27, 26, 25, 24, 23, + + 34, 33, 32, 31, + 42, 41, 40, 39, 38, 37, 36, 35, + 46, 45, 44, 43, 50, 49, 48, 47, + 58, 57, 56, 55, 54, 53, 52, 51, 66, 65, 64, 63, 62, 61, 60, 59, + + 67, 68, 69, 70, + + 1, + 1, 0, 1, 0, +} + +var src = []byte{1, 2, 3, 4, 5, 6, 7, 8} +var res = []int32{0x01020304, 0x05060708} +var putbuf = []byte{0, 0, 0, 0, 0, 0, 0, 0} + +func checkResult(t *testing.T, dir string, order ByteOrder, err error, have, want interface{}) { + if err != nil { + t.Errorf("%v %v: %v", dir, order, err) + return + } + if !reflect.DeepEqual(have, want) { + t.Errorf("%v %v:\n\thave %+v\n\twant %+v", dir, order, have, want) + } +} + +func testRead(t *testing.T, order ByteOrder, b []byte, s1 interface{}) { + var s2 Struct + err := Read(bytes.NewReader(b), order, &s2) + checkResult(t, "Read", order, err, s2, s1) +} + +func testWrite(t *testing.T, order ByteOrder, b []byte, s1 interface{}) { + buf := new(bytes.Buffer) + err := Write(buf, order, s1) + checkResult(t, "Write", order, err, buf.Bytes(), b) +} + +func TestLittleEndianRead(t *testing.T) { testRead(t, LittleEndian, little, s) } +func TestLittleEndianWrite(t *testing.T) { testWrite(t, LittleEndian, little, s) } +func TestLittleEndianPtrWrite(t *testing.T) { testWrite(t, LittleEndian, little, &s) } + +func TestBigEndianRead(t *testing.T) { testRead(t, BigEndian, big, s) } +func TestBigEndianWrite(t *testing.T) { testWrite(t, BigEndian, big, s) } +func TestBigEndianPtrWrite(t *testing.T) { testWrite(t, BigEndian, big, &s) } + +func TestReadSlice(t *testing.T) { + slice := make([]int32, 2) + err := Read(bytes.NewReader(src), BigEndian, slice) + checkResult(t, "ReadSlice", BigEndian, err, slice, res) +} + +func TestWriteSlice(t *testing.T) { + buf := new(bytes.Buffer) + err := Write(buf, BigEndian, res) + checkResult(t, "WriteSlice", BigEndian, err, buf.Bytes(), src) +} + +func TestReadBool(t *testing.T) { + var res bool + var err error + err = Read(bytes.NewReader([]byte{0}), BigEndian, &res) + checkResult(t, "ReadBool", BigEndian, err, res, false) + res = false + err = Read(bytes.NewReader([]byte{1}), BigEndian, &res) + checkResult(t, "ReadBool", BigEndian, err, res, true) + res = false + err = Read(bytes.NewReader([]byte{2}), BigEndian, &res) + checkResult(t, "ReadBool", BigEndian, err, res, true) +} + +func TestReadBoolSlice(t *testing.T) { + slice := make([]bool, 4) + err := Read(bytes.NewReader([]byte{0, 1, 2, 255}), BigEndian, slice) + checkResult(t, "ReadBoolSlice", BigEndian, err, slice, []bool{false, true, true, true}) +} + +// Addresses of arrays are easier to manipulate with reflection than are slices. +var intArrays = []interface{}{ + &[100]int8{}, + &[100]int16{}, + &[100]int32{}, + &[100]int64{}, + &[100]uint8{}, + &[100]uint16{}, + &[100]uint32{}, + &[100]uint64{}, +} + +func TestSliceRoundTrip(t *testing.T) { + buf := new(bytes.Buffer) + for _, array := range intArrays { + src := reflect.ValueOf(array).Elem() + unsigned := false + switch src.Index(0).Kind() { + case reflect.Uint8, reflect.Uint16, reflect.Uint32, reflect.Uint64: + unsigned = true + } + for i := 0; i < src.Len(); i++ { + if unsigned { + src.Index(i).SetUint(uint64(i * 0x07654321)) + } else { + src.Index(i).SetInt(int64(i * 0x07654321)) + } + } + buf.Reset() + srcSlice := src.Slice(0, src.Len()) + err := Write(buf, BigEndian, srcSlice.Interface()) + if err != nil { + t.Fatal(err) + } + dst := reflect.New(src.Type()).Elem() + dstSlice := dst.Slice(0, dst.Len()) + err = Read(buf, BigEndian, dstSlice.Interface()) + if err != nil { + t.Fatal(err) + } + if !reflect.DeepEqual(src.Interface(), dst.Interface()) { + t.Fatal(src) + } + } +} + +func TestWriteT(t *testing.T) { + buf := new(bytes.Buffer) + ts := T{} + if err := Write(buf, BigEndian, ts); err == nil { + t.Errorf("WriteT: have err == nil, want non-nil") + } + + tv := reflect.Indirect(reflect.ValueOf(ts)) + for i, n := 0, tv.NumField(); i < n; i++ { + typ := tv.Field(i).Type().String() + if typ == "[4]int" { + typ = "int" // the problem is int, not the [4] + } + if err := Write(buf, BigEndian, tv.Field(i).Interface()); err == nil { + t.Errorf("WriteT.%v: have err == nil, want non-nil", tv.Field(i).Type()) + } else if !strings.Contains(err.Error(), typ) { + t.Errorf("WriteT: have err == %q, want it to mention %s", err, typ) + } + } +} + +type BlankFields struct { + A uint32 + _ int32 + B float64 + _ [4]int16 + C byte + _ [7]byte + _ struct { + f [8]float32 + } +} + +type BlankFieldsProbe struct { + A uint32 + P0 int32 + B float64 + P1 [4]int16 + C byte + P2 [7]byte + P3 struct { + F [8]float32 + } +} + +func TestBlankFields(t *testing.T) { + buf := new(bytes.Buffer) + b1 := BlankFields{A: 1234567890, B: 2.718281828, C: 42} + if err := Write(buf, LittleEndian, &b1); err != nil { + t.Error(err) + } + + // zero values must have been written for blank fields + var p BlankFieldsProbe + if err := Read(buf, LittleEndian, &p); err != nil { + t.Error(err) + } + + // quick test: only check first value of slices + if p.P0 != 0 || p.P1[0] != 0 || p.P2[0] != 0 || p.P3.F[0] != 0 { + t.Errorf("non-zero values for originally blank fields: %#v", p) + } + + // write p and see if we can probe only some fields + if err := Write(buf, LittleEndian, &p); err != nil { + t.Error(err) + } + + // read should ignore blank fields in b2 + var b2 BlankFields + if err := Read(buf, LittleEndian, &b2); err != nil { + t.Error(err) + } + if b1.A != b2.A || b1.B != b2.B || b1.C != b2.C { + t.Errorf("%#v != %#v", b1, b2) + } +} + +func TestSizeStructCache(t *testing.T) { + // Reset the cache, otherwise multiple test runs fail. + structSize = sync.Map{} + + count := func() int { + var i int + structSize.Range(func(_, _ interface{}) bool { + i++ + return true + }) + return i + } + + var total int + added := func() int { + delta := count() - total + total += delta + return delta + } + + type foo struct { + A uint32 + } + + type bar struct { + A Struct + B foo + C Struct + } + + testcases := []struct { + val interface{} + want int + }{ + {new(foo), 1}, + {new(bar), 1}, + {new(bar), 0}, + {new(struct{ A Struct }), 1}, + {new(struct{ A Struct }), 0}, + } + + for _, tc := range testcases { + if Size(tc.val) == -1 { + t.Fatalf("Can't get the size of %T", tc.val) + } + + if n := added(); n != tc.want { + t.Errorf("Sizing %T added %d entries to the cache, want %d", tc.val, n, tc.want) + } + } +} + +// An attempt to read into a struct with an unexported field will +// panic. This is probably not the best choice, but at this point +// anything else would be an API change. + +type Unexported struct { + a int32 +} + +func TestUnexportedRead(t *testing.T) { + var buf bytes.Buffer + u1 := Unexported{a: 1} + if err := Write(&buf, LittleEndian, &u1); err != nil { + t.Fatal(err) + } + + defer func() { + if recover() == nil { + t.Fatal("did not panic") + } + }() + var u2 Unexported + Read(&buf, LittleEndian, &u2) +} + +func TestReadErrorMsg(t *testing.T) { + var buf bytes.Buffer + read := func(data interface{}) { + err := Read(&buf, LittleEndian, data) + want := "binary.Read: invalid type " + reflect.TypeOf(data).String() + if err == nil { + t.Errorf("%T: got no error; want %q", data, want) + return + } + if got := err.Error(); got != want { + t.Errorf("%T: got %q; want %q", data, got, want) + } + } + read(0) + s := new(struct{}) + read(&s) + p := &s + read(&p) +} + +func TestReadTruncated(t *testing.T) { + const data = "0123456789abcdef" + + var b1 = make([]int32, 4) + var b2 struct { + A, B, C, D byte + E int32 + F float64 + } + + for i := 0; i <= len(data); i++ { + var errWant error + switch i { + case 0: + errWant = io.EOF + case len(data): + errWant = nil + default: + errWant = io.ErrUnexpectedEOF + } + + if err := Read(strings.NewReader(data[:i]), LittleEndian, &b1); err != errWant { + t.Errorf("Read(%d) with slice: got %v, want %v", i, err, errWant) + } + if err := Read(strings.NewReader(data[:i]), LittleEndian, &b2); err != errWant { + t.Errorf("Read(%d) with struct: got %v, want %v", i, err, errWant) + } + } +} + +func testUint64SmallSliceLengthPanics() (panicked bool) { + defer func() { + panicked = recover() != nil + }() + b := [8]byte{1, 2, 3, 4, 5, 6, 7, 8} + LittleEndian.Uint64(b[:4]) + return false +} + +func testPutUint64SmallSliceLengthPanics() (panicked bool) { + defer func() { + panicked = recover() != nil + }() + b := [8]byte{} + LittleEndian.PutUint64(b[:4], 0x0102030405060708) + return false +} + +func TestEarlyBoundsChecks(t *testing.T) { + if testUint64SmallSliceLengthPanics() != true { + t.Errorf("binary.LittleEndian.Uint64 expected to panic for small slices, but didn't") + } + if testPutUint64SmallSliceLengthPanics() != true { + t.Errorf("binary.LittleEndian.PutUint64 expected to panic for small slices, but didn't") + } +} + +func TestReadInvalidDestination(t *testing.T) { + testReadInvalidDestination(t, BigEndian) + testReadInvalidDestination(t, LittleEndian) +} + +func testReadInvalidDestination(t *testing.T, order ByteOrder) { + destinations := []interface{}{ + int8(0), + int16(0), + int32(0), + int64(0), + + uint8(0), + uint16(0), + uint32(0), + uint64(0), + + bool(false), + } + + for _, dst := range destinations { + err := Read(bytes.NewReader([]byte{1, 2, 3, 4, 5, 6, 7, 8}), order, dst) + want := fmt.Sprintf("binary.Read: invalid type %T", dst) + if err == nil || err.Error() != want { + t.Fatalf("for type %T: got %q; want %q", dst, err, want) + } + } +} + +type byteSliceReader struct { + remain []byte +} + +func (br *byteSliceReader) Read(p []byte) (int, error) { + n := copy(p, br.remain) + br.remain = br.remain[n:] + return n, nil +} + +func BenchmarkReadSlice1000Int32s(b *testing.B) { + bsr := &byteSliceReader{} + slice := make([]int32, 1000) + buf := make([]byte, len(slice)*4) + b.SetBytes(int64(len(buf))) + b.ResetTimer() + for i := 0; i < b.N; i++ { + bsr.remain = buf + Read(bsr, BigEndian, slice) + } +} + +func BenchmarkReadStruct(b *testing.B) { + bsr := &byteSliceReader{} + var buf bytes.Buffer + Write(&buf, BigEndian, &s) + b.SetBytes(int64(dataSize(reflect.ValueOf(s)))) + t := s + b.ResetTimer() + for i := 0; i < b.N; i++ { + bsr.remain = buf.Bytes() + Read(bsr, BigEndian, &t) + } + b.StopTimer() + if b.N > 0 && !reflect.DeepEqual(s, t) { + b.Fatalf("struct doesn't match:\ngot %v;\nwant %v", t, s) + } +} + +func BenchmarkWriteStruct(b *testing.B) { + b.SetBytes(int64(Size(&s))) + b.ResetTimer() + for i := 0; i < b.N; i++ { + Write(io.Discard, BigEndian, &s) + } +} + +func BenchmarkReadInts(b *testing.B) { + var ls Struct + bsr := &byteSliceReader{} + var r io.Reader = bsr + b.SetBytes(2 * (1 + 2 + 4 + 8)) + b.ResetTimer() + for i := 0; i < b.N; i++ { + bsr.remain = big + Read(r, BigEndian, &ls.Int8) + Read(r, BigEndian, &ls.Int16) + Read(r, BigEndian, &ls.Int32) + Read(r, BigEndian, &ls.Int64) + Read(r, BigEndian, &ls.Uint8) + Read(r, BigEndian, &ls.Uint16) + Read(r, BigEndian, &ls.Uint32) + Read(r, BigEndian, &ls.Uint64) + } + b.StopTimer() + want := s + want.Float32 = 0 + want.Float64 = 0 + want.Complex64 = 0 + want.Complex128 = 0 + want.Array = [4]uint8{0, 0, 0, 0} + want.Bool = false + want.BoolArray = [4]bool{false, false, false, false} + if b.N > 0 && !reflect.DeepEqual(ls, want) { + b.Fatalf("struct doesn't match:\ngot %v;\nwant %v", ls, want) + } +} + +func BenchmarkWriteInts(b *testing.B) { + buf := new(bytes.Buffer) + var w io.Writer = buf + b.SetBytes(2 * (1 + 2 + 4 + 8)) + b.ResetTimer() + for i := 0; i < b.N; i++ { + buf.Reset() + Write(w, BigEndian, s.Int8) + Write(w, BigEndian, s.Int16) + Write(w, BigEndian, s.Int32) + Write(w, BigEndian, s.Int64) + Write(w, BigEndian, s.Uint8) + Write(w, BigEndian, s.Uint16) + Write(w, BigEndian, s.Uint32) + Write(w, BigEndian, s.Uint64) + } + b.StopTimer() + if b.N > 0 && !bytes.Equal(buf.Bytes(), big[:30]) { + b.Fatalf("first half doesn't match: %x %x", buf.Bytes(), big[:30]) + } +} + +func BenchmarkWriteSlice1000Int32s(b *testing.B) { + slice := make([]int32, 1000) + buf := new(bytes.Buffer) + var w io.Writer = buf + b.SetBytes(4 * 1000) + b.ResetTimer() + for i := 0; i < b.N; i++ { + buf.Reset() + Write(w, BigEndian, slice) + } + b.StopTimer() +} + +func BenchmarkPutUint16(b *testing.B) { + b.SetBytes(2) + for i := 0; i < b.N; i++ { + BigEndian.PutUint16(putbuf[:], uint16(i)) + } +} + +func BenchmarkPutUint32(b *testing.B) { + b.SetBytes(4) + for i := 0; i < b.N; i++ { + BigEndian.PutUint32(putbuf[:], uint32(i)) + } +} + +func BenchmarkPutUint64(b *testing.B) { + b.SetBytes(8) + for i := 0; i < b.N; i++ { + BigEndian.PutUint64(putbuf[:], uint64(i)) + } +} +func BenchmarkLittleEndianPutUint16(b *testing.B) { + b.SetBytes(2) + for i := 0; i < b.N; i++ { + LittleEndian.PutUint16(putbuf[:], uint16(i)) + } +} + +func BenchmarkLittleEndianPutUint32(b *testing.B) { + b.SetBytes(4) + for i := 0; i < b.N; i++ { + LittleEndian.PutUint32(putbuf[:], uint32(i)) + } +} + +func BenchmarkLittleEndianPutUint64(b *testing.B) { + b.SetBytes(8) + for i := 0; i < b.N; i++ { + LittleEndian.PutUint64(putbuf[:], uint64(i)) + } +} + +func BenchmarkReadFloats(b *testing.B) { + var ls Struct + bsr := &byteSliceReader{} + var r io.Reader = bsr + b.SetBytes(4 + 8) + b.ResetTimer() + for i := 0; i < b.N; i++ { + bsr.remain = big[30:] + Read(r, BigEndian, &ls.Float32) + Read(r, BigEndian, &ls.Float64) + } + b.StopTimer() + want := s + want.Int8 = 0 + want.Int16 = 0 + want.Int32 = 0 + want.Int64 = 0 + want.Uint8 = 0 + want.Uint16 = 0 + want.Uint32 = 0 + want.Uint64 = 0 + want.Complex64 = 0 + want.Complex128 = 0 + want.Array = [4]uint8{0, 0, 0, 0} + want.Bool = false + want.BoolArray = [4]bool{false, false, false, false} + if b.N > 0 && !reflect.DeepEqual(ls, want) { + b.Fatalf("struct doesn't match:\ngot %v;\nwant %v", ls, want) + } +} + +func BenchmarkWriteFloats(b *testing.B) { + buf := new(bytes.Buffer) + var w io.Writer = buf + b.SetBytes(4 + 8) + b.ResetTimer() + for i := 0; i < b.N; i++ { + buf.Reset() + Write(w, BigEndian, s.Float32) + Write(w, BigEndian, s.Float64) + } + b.StopTimer() + if b.N > 0 && !bytes.Equal(buf.Bytes(), big[30:30+4+8]) { + b.Fatalf("first half doesn't match: %x %x", buf.Bytes(), big[30:30+4+8]) + } +} + +func BenchmarkReadSlice1000Float32s(b *testing.B) { + bsr := &byteSliceReader{} + slice := make([]float32, 1000) + buf := make([]byte, len(slice)*4) + b.SetBytes(int64(len(buf))) + b.ResetTimer() + for i := 0; i < b.N; i++ { + bsr.remain = buf + Read(bsr, BigEndian, slice) + } +} + +func BenchmarkWriteSlice1000Float32s(b *testing.B) { + slice := make([]float32, 1000) + buf := new(bytes.Buffer) + var w io.Writer = buf + b.SetBytes(4 * 1000) + b.ResetTimer() + for i := 0; i < b.N; i++ { + buf.Reset() + Write(w, BigEndian, slice) + } + b.StopTimer() +} + +func BenchmarkReadSlice1000Uint8s(b *testing.B) { + bsr := &byteSliceReader{} + slice := make([]uint8, 1000) + buf := make([]byte, len(slice)) + b.SetBytes(int64(len(buf))) + b.ResetTimer() + for i := 0; i < b.N; i++ { + bsr.remain = buf + Read(bsr, BigEndian, slice) + } +} + +func BenchmarkWriteSlice1000Uint8s(b *testing.B) { + slice := make([]uint8, 1000) + buf := new(bytes.Buffer) + var w io.Writer = buf + b.SetBytes(1000) + b.ResetTimer() + for i := 0; i < b.N; i++ { + buf.Reset() + Write(w, BigEndian, slice) + } +} diff --git a/src/encoding/binary/example_test.go b/src/encoding/binary/example_test.go new file mode 100644 index 0000000..b994b89 --- /dev/null +++ b/src/encoding/binary/example_test.go @@ -0,0 +1,187 @@ +// Copyright 2011 The Go Authors. All rights reserved. +// Use of this source code is governed by a BSD-style +// license that can be found in the LICENSE file. + +package binary_test + +import ( + "bytes" + "encoding/binary" + "fmt" + "math" +) + +func ExampleWrite() { + buf := new(bytes.Buffer) + var pi float64 = math.Pi + err := binary.Write(buf, binary.LittleEndian, pi) + if err != nil { + fmt.Println("binary.Write failed:", err) + } + fmt.Printf("% x", buf.Bytes()) + // Output: 18 2d 44 54 fb 21 09 40 +} + +func ExampleWrite_multi() { + buf := new(bytes.Buffer) + var data = []interface{}{ + uint16(61374), + int8(-54), + uint8(254), + } + for _, v := range data { + err := binary.Write(buf, binary.LittleEndian, v) + if err != nil { + fmt.Println("binary.Write failed:", err) + } + } + fmt.Printf("%x", buf.Bytes()) + // Output: beefcafe +} + +func ExampleRead() { + var pi float64 + b := []byte{0x18, 0x2d, 0x44, 0x54, 0xfb, 0x21, 0x09, 0x40} + buf := bytes.NewReader(b) + err := binary.Read(buf, binary.LittleEndian, &pi) + if err != nil { + fmt.Println("binary.Read failed:", err) + } + fmt.Print(pi) + // Output: 3.141592653589793 +} + +func ExampleRead_multi() { + b := []byte{0x18, 0x2d, 0x44, 0x54, 0xfb, 0x21, 0x09, 0x40, 0xff, 0x01, 0x02, 0x03, 0xbe, 0xef} + r := bytes.NewReader(b) + + var data struct { + PI float64 + Uate uint8 + Mine [3]byte + Too uint16 + } + + if err := binary.Read(r, binary.LittleEndian, &data); err != nil { + fmt.Println("binary.Read failed:", err) + } + + fmt.Println(data.PI) + fmt.Println(data.Uate) + fmt.Printf("% x\n", data.Mine) + fmt.Println(data.Too) + // Output: + // 3.141592653589793 + // 255 + // 01 02 03 + // 61374 +} + +func ExampleByteOrder_put() { + b := make([]byte, 4) + binary.LittleEndian.PutUint16(b[0:], 0x03e8) + binary.LittleEndian.PutUint16(b[2:], 0x07d0) + fmt.Printf("% x\n", b) + // Output: + // e8 03 d0 07 +} + +func ExampleByteOrder_get() { + b := []byte{0xe8, 0x03, 0xd0, 0x07} + x1 := binary.LittleEndian.Uint16(b[0:]) + x2 := binary.LittleEndian.Uint16(b[2:]) + fmt.Printf("%#04x %#04x\n", x1, x2) + // Output: + // 0x03e8 0x07d0 +} + +func ExamplePutUvarint() { + buf := make([]byte, binary.MaxVarintLen64) + + for _, x := range []uint64{1, 2, 127, 128, 255, 256} { + n := binary.PutUvarint(buf, x) + fmt.Printf("%x\n", buf[:n]) + } + // Output: + // 01 + // 02 + // 7f + // 8001 + // ff01 + // 8002 +} + +func ExamplePutVarint() { + buf := make([]byte, binary.MaxVarintLen64) + + for _, x := range []int64{-65, -64, -2, -1, 0, 1, 2, 63, 64} { + n := binary.PutVarint(buf, x) + fmt.Printf("%x\n", buf[:n]) + } + // Output: + // 8101 + // 7f + // 03 + // 01 + // 00 + // 02 + // 04 + // 7e + // 8001 +} + +func ExampleUvarint() { + inputs := [][]byte{ + {0x01}, + {0x02}, + {0x7f}, + {0x80, 0x01}, + {0xff, 0x01}, + {0x80, 0x02}, + } + for _, b := range inputs { + x, n := binary.Uvarint(b) + if n != len(b) { + fmt.Println("Uvarint did not consume all of in") + } + fmt.Println(x) + } + // Output: + // 1 + // 2 + // 127 + // 128 + // 255 + // 256 +} + +func ExampleVarint() { + inputs := [][]byte{ + {0x81, 0x01}, + {0x7f}, + {0x03}, + {0x01}, + {0x00}, + {0x02}, + {0x04}, + {0x7e}, + {0x80, 0x01}, + } + for _, b := range inputs { + x, n := binary.Varint(b) + if n != len(b) { + fmt.Println("Varint did not consume all of in") + } + fmt.Println(x) + } + // Output: + // -65 + // -64 + // -2 + // -1 + // 0 + // 1 + // 2 + // 63 + // 64 +} diff --git a/src/encoding/binary/varint.go b/src/encoding/binary/varint.go new file mode 100644 index 0000000..8fe20b5 --- /dev/null +++ b/src/encoding/binary/varint.go @@ -0,0 +1,139 @@ +// Copyright 2011 The Go Authors. All rights reserved. +// Use of this source code is governed by a BSD-style +// license that can be found in the LICENSE file. + +package binary + +// This file implements "varint" encoding of 64-bit integers. +// The encoding is: +// - unsigned integers are serialized 7 bits at a time, starting with the +// least significant bits +// - the most significant bit (msb) in each output byte indicates if there +// is a continuation byte (msb = 1) +// - signed integers are mapped to unsigned integers using "zig-zag" +// encoding: Positive values x are written as 2*x + 0, negative values +// are written as 2*(^x) + 1; that is, negative numbers are complemented +// and whether to complement is encoded in bit 0. +// +// Design note: +// At most 10 bytes are needed for 64-bit values. The encoding could +// be more dense: a full 64-bit value needs an extra byte just to hold bit 63. +// Instead, the msb of the previous byte could be used to hold bit 63 since we +// know there can't be more than 64 bits. This is a trivial improvement and +// would reduce the maximum encoding length to 9 bytes. However, it breaks the +// invariant that the msb is always the "continuation bit" and thus makes the +// format incompatible with a varint encoding for larger numbers (say 128-bit). + +import ( + "errors" + "io" +) + +// MaxVarintLenN is the maximum length of a varint-encoded N-bit integer. +const ( + MaxVarintLen16 = 3 + MaxVarintLen32 = 5 + MaxVarintLen64 = 10 +) + +// PutUvarint encodes a uint64 into buf and returns the number of bytes written. +// If the buffer is too small, PutUvarint will panic. +func PutUvarint(buf []byte, x uint64) int { + i := 0 + for x >= 0x80 { + buf[i] = byte(x) | 0x80 + x >>= 7 + i++ + } + buf[i] = byte(x) + return i + 1 +} + +// Uvarint decodes a uint64 from buf and returns that value and the +// number of bytes read (> 0). If an error occurred, the value is 0 +// and the number of bytes n is <= 0 meaning: +// +// n == 0: buf too small +// n < 0: value larger than 64 bits (overflow) +// and -n is the number of bytes read +// +func Uvarint(buf []byte) (uint64, int) { + var x uint64 + var s uint + for i, b := range buf { + if i == MaxVarintLen64 { + // Catch byte reads past MaxVarintLen64. + // See issue https://golang.org/issues/41185 + return 0, -(i + 1) // overflow + } + if b < 0x80 { + if i == MaxVarintLen64-1 && b > 1 { + return 0, -(i + 1) // overflow + } + return x | uint64(b)<<s, i + 1 + } + x |= uint64(b&0x7f) << s + s += 7 + } + return 0, 0 +} + +// PutVarint encodes an int64 into buf and returns the number of bytes written. +// If the buffer is too small, PutVarint will panic. +func PutVarint(buf []byte, x int64) int { + ux := uint64(x) << 1 + if x < 0 { + ux = ^ux + } + return PutUvarint(buf, ux) +} + +// Varint decodes an int64 from buf and returns that value and the +// number of bytes read (> 0). If an error occurred, the value is 0 +// and the number of bytes n is <= 0 with the following meaning: +// +// n == 0: buf too small +// n < 0: value larger than 64 bits (overflow) +// and -n is the number of bytes read +// +func Varint(buf []byte) (int64, int) { + ux, n := Uvarint(buf) // ok to continue in presence of error + x := int64(ux >> 1) + if ux&1 != 0 { + x = ^x + } + return x, n +} + +var overflow = errors.New("binary: varint overflows a 64-bit integer") + +// ReadUvarint reads an encoded unsigned integer from r and returns it as a uint64. +func ReadUvarint(r io.ByteReader) (uint64, error) { + var x uint64 + var s uint + for i := 0; i < MaxVarintLen64; i++ { + b, err := r.ReadByte() + if err != nil { + return x, err + } + if b < 0x80 { + if i == MaxVarintLen64-1 && b > 1 { + return x, overflow + } + return x | uint64(b)<<s, nil + } + x |= uint64(b&0x7f) << s + s += 7 + } + return x, overflow +} + +// ReadVarint reads an encoded signed integer from r and returns it as an int64. +func ReadVarint(r io.ByteReader) (int64, error) { + ux, err := ReadUvarint(r) // ok to continue in presence of error + x := int64(ux >> 1) + if ux&1 != 0 { + x = ^x + } + return x, err +} diff --git a/src/encoding/binary/varint_test.go b/src/encoding/binary/varint_test.go new file mode 100644 index 0000000..d025a67 --- /dev/null +++ b/src/encoding/binary/varint_test.go @@ -0,0 +1,231 @@ +// Copyright 2011 The Go Authors. All rights reserved. +// Use of this source code is governed by a BSD-style +// license that can be found in the LICENSE file. + +package binary + +import ( + "bytes" + "io" + "math" + "testing" +) + +func testConstant(t *testing.T, w uint, max int) { + buf := make([]byte, MaxVarintLen64) + n := PutUvarint(buf, 1<<w-1) + if n != max { + t.Errorf("MaxVarintLen%d = %d; want %d", w, max, n) + } +} + +func TestConstants(t *testing.T) { + testConstant(t, 16, MaxVarintLen16) + testConstant(t, 32, MaxVarintLen32) + testConstant(t, 64, MaxVarintLen64) +} + +func testVarint(t *testing.T, x int64) { + buf := make([]byte, MaxVarintLen64) + n := PutVarint(buf, x) + y, m := Varint(buf[0:n]) + if x != y { + t.Errorf("Varint(%d): got %d", x, y) + } + if n != m { + t.Errorf("Varint(%d): got n = %d; want %d", x, m, n) + } + + y, err := ReadVarint(bytes.NewReader(buf)) + if err != nil { + t.Errorf("ReadVarint(%d): %s", x, err) + } + if x != y { + t.Errorf("ReadVarint(%d): got %d", x, y) + } +} + +func testUvarint(t *testing.T, x uint64) { + buf := make([]byte, MaxVarintLen64) + n := PutUvarint(buf, x) + y, m := Uvarint(buf[0:n]) + if x != y { + t.Errorf("Uvarint(%d): got %d", x, y) + } + if n != m { + t.Errorf("Uvarint(%d): got n = %d; want %d", x, m, n) + } + + y, err := ReadUvarint(bytes.NewReader(buf)) + if err != nil { + t.Errorf("ReadUvarint(%d): %s", x, err) + } + if x != y { + t.Errorf("ReadUvarint(%d): got %d", x, y) + } +} + +var tests = []int64{ + -1 << 63, + -1<<63 + 1, + -1, + 0, + 1, + 2, + 10, + 20, + 63, + 64, + 65, + 127, + 128, + 129, + 255, + 256, + 257, + 1<<63 - 1, +} + +func TestVarint(t *testing.T) { + for _, x := range tests { + testVarint(t, x) + testVarint(t, -x) + } + for x := int64(0x7); x != 0; x <<= 1 { + testVarint(t, x) + testVarint(t, -x) + } +} + +func TestUvarint(t *testing.T) { + for _, x := range tests { + testUvarint(t, uint64(x)) + } + for x := uint64(0x7); x != 0; x <<= 1 { + testUvarint(t, x) + } +} + +func TestBufferTooSmall(t *testing.T) { + buf := []byte{0x80, 0x80, 0x80, 0x80} + for i := 0; i <= len(buf); i++ { + buf := buf[0:i] + x, n := Uvarint(buf) + if x != 0 || n != 0 { + t.Errorf("Uvarint(%v): got x = %d, n = %d", buf, x, n) + } + + x, err := ReadUvarint(bytes.NewReader(buf)) + if x != 0 || err != io.EOF { + t.Errorf("ReadUvarint(%v): got x = %d, err = %s", buf, x, err) + } + } +} + +// Ensure that we catch overflows of bytes going past MaxVarintLen64. +// See issue https://golang.org/issues/41185 +func TestBufferTooBigWithOverflow(t *testing.T) { + tests := []struct { + in []byte + name string + wantN int + wantValue uint64 + }{ + { + name: "invalid: 1000 bytes", + in: func() []byte { + b := make([]byte, 1000) + for i := range b { + b[i] = 0xff + } + b[999] = 0 + return b + }(), + wantN: -11, + wantValue: 0, + }, + { + name: "valid: math.MaxUint64-40", + in: []byte{0xd7, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x01}, + wantValue: math.MaxUint64 - 40, + wantN: 10, + }, + { + name: "invalid: with more than MaxVarintLen64 bytes", + in: []byte{0xd7, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x01}, + wantN: -11, + wantValue: 0, + }, + { + name: "invalid: 10th byte", + in: []byte{0xd7, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x7f}, + wantN: -10, + wantValue: 0, + }, + } + + for _, tt := range tests { + tt := tt + t.Run(tt.name, func(t *testing.T) { + value, n := Uvarint(tt.in) + if g, w := n, tt.wantN; g != w { + t.Errorf("bytes returned=%d, want=%d", g, w) + } + if g, w := value, tt.wantValue; g != w { + t.Errorf("value=%d, want=%d", g, w) + } + }) + } +} + +func testOverflow(t *testing.T, buf []byte, x0 uint64, n0 int, err0 error) { + x, n := Uvarint(buf) + if x != 0 || n != n0 { + t.Errorf("Uvarint(% X): got x = %d, n = %d; want 0, %d", buf, x, n, n0) + } + + r := bytes.NewReader(buf) + len := r.Len() + x, err := ReadUvarint(r) + if x != x0 || err != err0 { + t.Errorf("ReadUvarint(%v): got x = %d, err = %s; want %d, %s", buf, x, err, x0, err0) + } + if read := len - r.Len(); read > MaxVarintLen64 { + t.Errorf("ReadUvarint(%v): read more than MaxVarintLen64 bytes, got %d", buf, read) + } +} + +func TestOverflow(t *testing.T) { + testOverflow(t, []byte{0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x2}, 0, -10, overflow) + testOverflow(t, []byte{0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x1, 0, 0}, 0, -11, overflow) + testOverflow(t, []byte{0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF}, 1<<64-1, -11, overflow) // 11 bytes, should overflow +} + +func TestNonCanonicalZero(t *testing.T) { + buf := []byte{0x80, 0x80, 0x80, 0} + x, n := Uvarint(buf) + if x != 0 || n != 4 { + t.Errorf("Uvarint(%v): got x = %d, n = %d; want 0, 4", buf, x, n) + + } +} + +func BenchmarkPutUvarint32(b *testing.B) { + buf := make([]byte, MaxVarintLen32) + b.SetBytes(4) + for i := 0; i < b.N; i++ { + for j := uint(0); j < MaxVarintLen32; j++ { + PutUvarint(buf, 1<<(j*7)) + } + } +} + +func BenchmarkPutUvarint64(b *testing.B) { + buf := make([]byte, MaxVarintLen64) + b.SetBytes(8) + for i := 0; i < b.N; i++ { + for j := uint(0); j < MaxVarintLen64; j++ { + PutUvarint(buf, 1<<(j*7)) + } + } +} |