From 43a123c1ae6613b3efeed291fa552ecd909d3acf Mon Sep 17 00:00:00 2001 From: Daniel Baumann Date: Tue, 16 Apr 2024 21:23:18 +0200 Subject: Adding upstream version 1.20.14. Signed-off-by: Daniel Baumann --- src/encoding/gob/codec_test.go | 1595 ++++++++++++++++++++++++++++ src/encoding/gob/debug.go | 733 +++++++++++++ src/encoding/gob/dec_helpers.go | 544 ++++++++++ src/encoding/gob/decgen.go | 243 +++++ src/encoding/gob/decode.go | 1308 +++++++++++++++++++++++ src/encoding/gob/decoder.go | 237 +++++ src/encoding/gob/doc.go | 423 ++++++++ src/encoding/gob/dump.go | 29 + src/encoding/gob/enc_helpers.go | 414 ++++++++ src/encoding/gob/encgen.go | 220 ++++ src/encoding/gob/encode.go | 705 ++++++++++++ src/encoding/gob/encoder.go | 258 +++++ src/encoding/gob/encoder_test.go | 1280 ++++++++++++++++++++++ src/encoding/gob/error.go | 42 + src/encoding/gob/example_encdec_test.go | 61 ++ src/encoding/gob/example_interface_test.go | 81 ++ src/encoding/gob/example_test.go | 60 ++ src/encoding/gob/gobencdec_test.go | 822 ++++++++++++++ src/encoding/gob/timing_test.go | 328 ++++++ src/encoding/gob/type.go | 913 ++++++++++++++++ src/encoding/gob/type_test.go | 262 +++++ 21 files changed, 10558 insertions(+) create mode 100644 src/encoding/gob/codec_test.go create mode 100644 src/encoding/gob/debug.go create mode 100644 src/encoding/gob/dec_helpers.go create mode 100644 src/encoding/gob/decgen.go create mode 100644 src/encoding/gob/decode.go create mode 100644 src/encoding/gob/decoder.go create mode 100644 src/encoding/gob/doc.go create mode 100644 src/encoding/gob/dump.go create mode 100644 src/encoding/gob/enc_helpers.go create mode 100644 src/encoding/gob/encgen.go create mode 100644 src/encoding/gob/encode.go create mode 100644 src/encoding/gob/encoder.go create mode 100644 src/encoding/gob/encoder_test.go create mode 100644 src/encoding/gob/error.go create mode 100644 src/encoding/gob/example_encdec_test.go create mode 100644 src/encoding/gob/example_interface_test.go create mode 100644 src/encoding/gob/example_test.go create mode 100644 src/encoding/gob/gobencdec_test.go create mode 100644 src/encoding/gob/timing_test.go create mode 100644 src/encoding/gob/type.go create mode 100644 src/encoding/gob/type_test.go (limited to 'src/encoding/gob') diff --git a/src/encoding/gob/codec_test.go b/src/encoding/gob/codec_test.go new file mode 100644 index 0000000..54c356c --- /dev/null +++ b/src/encoding/gob/codec_test.go @@ -0,0 +1,1595 @@ +// 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 gob + +import ( + "bytes" + "errors" + "flag" + "math" + "math/rand" + "reflect" + "strings" + "testing" + "time" +) + +var doFuzzTests = flag.Bool("gob.fuzz", false, "run the fuzz tests, which are large and very slow") + +// Guarantee encoding format by comparing some encodings to hand-written values +type EncodeT struct { + x uint64 + b []byte +} + +var encodeT = []EncodeT{ + {0x00, []byte{0x00}}, + {0x0F, []byte{0x0F}}, + {0xFF, []byte{0xFF, 0xFF}}, + {0xFFFF, []byte{0xFE, 0xFF, 0xFF}}, + {0xFFFFFF, []byte{0xFD, 0xFF, 0xFF, 0xFF}}, + {0xFFFFFFFF, []byte{0xFC, 0xFF, 0xFF, 0xFF, 0xFF}}, + {0xFFFFFFFFFF, []byte{0xFB, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF}}, + {0xFFFFFFFFFFFF, []byte{0xFA, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF}}, + {0xFFFFFFFFFFFFFF, []byte{0xF9, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF}}, + {0xFFFFFFFFFFFFFFFF, []byte{0xF8, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF}}, + {0x1111, []byte{0xFE, 0x11, 0x11}}, + {0x1111111111111111, []byte{0xF8, 0x11, 0x11, 0x11, 0x11, 0x11, 0x11, 0x11, 0x11}}, + {0x8888888888888888, []byte{0xF8, 0x88, 0x88, 0x88, 0x88, 0x88, 0x88, 0x88, 0x88}}, + {1 << 63, []byte{0xF8, 0x80, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00}}, +} + +// testError is meant to be used as a deferred function to turn a panic(gobError) into a +// plain test.Error call. +func testError(t *testing.T) { + if e := recover(); e != nil { + t.Error(e.(gobError).err) // Will re-panic if not one of our errors, such as a runtime error. + } +} + +func newDecBuffer(data []byte) *decBuffer { + return &decBuffer{ + data: data, + } +} + +// Test basic encode/decode routines for unsigned integers +func TestUintCodec(t *testing.T) { + defer testError(t) + b := new(encBuffer) + encState := newEncoderState(b) + for _, tt := range encodeT { + b.Reset() + encState.encodeUint(tt.x) + if !bytes.Equal(tt.b, b.Bytes()) { + t.Errorf("encodeUint: %#x encode: expected % x got % x", tt.x, tt.b, b.Bytes()) + } + } + for u := uint64(0); ; u = (u + 1) * 7 { + b.Reset() + encState.encodeUint(u) + decState := newDecodeState(newDecBuffer(b.Bytes())) + v := decState.decodeUint() + if u != v { + t.Errorf("Encode/Decode: sent %#x received %#x", u, v) + } + if u&(1<<63) != 0 { + break + } + } +} + +func verifyInt(i int64, t *testing.T) { + defer testError(t) + var b = new(encBuffer) + encState := newEncoderState(b) + encState.encodeInt(i) + decState := newDecodeState(newDecBuffer(b.Bytes())) + j := decState.decodeInt() + if i != j { + t.Errorf("Encode/Decode: sent %#x received %#x", uint64(i), uint64(j)) + } +} + +// Test basic encode/decode routines for signed integers +func TestIntCodec(t *testing.T) { + for u := uint64(0); ; u = (u + 1) * 7 { + // Do positive and negative values + i := int64(u) + verifyInt(i, t) + verifyInt(-i, t) + verifyInt(^i, t) + if u&(1<<63) != 0 { + break + } + } + verifyInt(-1<<63, t) // a tricky case +} + +// The result of encoding a true boolean with field number 7 +var boolResult = []byte{0x07, 0x01} + +// The result of encoding a number 17 with field number 7 +var signedResult = []byte{0x07, 2 * 17} +var unsignedResult = []byte{0x07, 17} +var floatResult = []byte{0x07, 0xFE, 0x31, 0x40} + +// The result of encoding a number 17+19i with field number 7 +var complexResult = []byte{0x07, 0xFE, 0x31, 0x40, 0xFE, 0x33, 0x40} + +// The result of encoding "hello" with field number 7 +var bytesResult = []byte{0x07, 0x05, 'h', 'e', 'l', 'l', 'o'} + +func newDecodeState(buf *decBuffer) *decoderState { + d := new(decoderState) + d.b = buf + return d +} + +func newEncoderState(b *encBuffer) *encoderState { + b.Reset() + state := &encoderState{enc: nil, b: b} + state.fieldnum = -1 + return state +} + +// Test instruction execution for encoding. +// Do not run the machine yet; instead do individual instructions crafted by hand. +func TestScalarEncInstructions(t *testing.T) { + var b = new(encBuffer) + + // bool + { + var data bool = true + instr := &encInstr{encBool, 6, nil, 0} + state := newEncoderState(b) + instr.op(instr, state, reflect.ValueOf(data)) + if !bytes.Equal(boolResult, b.Bytes()) { + t.Errorf("bool enc instructions: expected % x got % x", boolResult, b.Bytes()) + } + } + + // int + { + b.Reset() + var data int = 17 + instr := &encInstr{encInt, 6, nil, 0} + state := newEncoderState(b) + instr.op(instr, state, reflect.ValueOf(data)) + if !bytes.Equal(signedResult, b.Bytes()) { + t.Errorf("int enc instructions: expected % x got % x", signedResult, b.Bytes()) + } + } + + // uint + { + b.Reset() + var data uint = 17 + instr := &encInstr{encUint, 6, nil, 0} + state := newEncoderState(b) + instr.op(instr, state, reflect.ValueOf(data)) + if !bytes.Equal(unsignedResult, b.Bytes()) { + t.Errorf("uint enc instructions: expected % x got % x", unsignedResult, b.Bytes()) + } + } + + // int8 + { + b.Reset() + var data int8 = 17 + instr := &encInstr{encInt, 6, nil, 0} + state := newEncoderState(b) + instr.op(instr, state, reflect.ValueOf(data)) + if !bytes.Equal(signedResult, b.Bytes()) { + t.Errorf("int8 enc instructions: expected % x got % x", signedResult, b.Bytes()) + } + } + + // uint8 + { + b.Reset() + var data uint8 = 17 + instr := &encInstr{encUint, 6, nil, 0} + state := newEncoderState(b) + instr.op(instr, state, reflect.ValueOf(data)) + if !bytes.Equal(unsignedResult, b.Bytes()) { + t.Errorf("uint8 enc instructions: expected % x got % x", unsignedResult, b.Bytes()) + } + } + + // int16 + { + b.Reset() + var data int16 = 17 + instr := &encInstr{encInt, 6, nil, 0} + state := newEncoderState(b) + instr.op(instr, state, reflect.ValueOf(data)) + if !bytes.Equal(signedResult, b.Bytes()) { + t.Errorf("int16 enc instructions: expected % x got % x", signedResult, b.Bytes()) + } + } + + // uint16 + { + b.Reset() + var data uint16 = 17 + instr := &encInstr{encUint, 6, nil, 0} + state := newEncoderState(b) + instr.op(instr, state, reflect.ValueOf(data)) + if !bytes.Equal(unsignedResult, b.Bytes()) { + t.Errorf("uint16 enc instructions: expected % x got % x", unsignedResult, b.Bytes()) + } + } + + // int32 + { + b.Reset() + var data int32 = 17 + instr := &encInstr{encInt, 6, nil, 0} + state := newEncoderState(b) + instr.op(instr, state, reflect.ValueOf(data)) + if !bytes.Equal(signedResult, b.Bytes()) { + t.Errorf("int32 enc instructions: expected % x got % x", signedResult, b.Bytes()) + } + } + + // uint32 + { + b.Reset() + var data uint32 = 17 + instr := &encInstr{encUint, 6, nil, 0} + state := newEncoderState(b) + instr.op(instr, state, reflect.ValueOf(data)) + if !bytes.Equal(unsignedResult, b.Bytes()) { + t.Errorf("uint32 enc instructions: expected % x got % x", unsignedResult, b.Bytes()) + } + } + + // int64 + { + b.Reset() + var data int64 = 17 + instr := &encInstr{encInt, 6, nil, 0} + state := newEncoderState(b) + instr.op(instr, state, reflect.ValueOf(data)) + if !bytes.Equal(signedResult, b.Bytes()) { + t.Errorf("int64 enc instructions: expected % x got % x", signedResult, b.Bytes()) + } + } + + // uint64 + { + b.Reset() + var data uint64 = 17 + instr := &encInstr{encUint, 6, nil, 0} + state := newEncoderState(b) + instr.op(instr, state, reflect.ValueOf(data)) + if !bytes.Equal(unsignedResult, b.Bytes()) { + t.Errorf("uint64 enc instructions: expected % x got % x", unsignedResult, b.Bytes()) + } + } + + // float32 + { + b.Reset() + var data float32 = 17 + instr := &encInstr{encFloat, 6, nil, 0} + state := newEncoderState(b) + instr.op(instr, state, reflect.ValueOf(data)) + if !bytes.Equal(floatResult, b.Bytes()) { + t.Errorf("float32 enc instructions: expected % x got % x", floatResult, b.Bytes()) + } + } + + // float64 + { + b.Reset() + var data float64 = 17 + instr := &encInstr{encFloat, 6, nil, 0} + state := newEncoderState(b) + instr.op(instr, state, reflect.ValueOf(data)) + if !bytes.Equal(floatResult, b.Bytes()) { + t.Errorf("float64 enc instructions: expected % x got % x", floatResult, b.Bytes()) + } + } + + // bytes == []uint8 + { + b.Reset() + data := []byte("hello") + instr := &encInstr{encUint8Array, 6, nil, 0} + state := newEncoderState(b) + instr.op(instr, state, reflect.ValueOf(data)) + if !bytes.Equal(bytesResult, b.Bytes()) { + t.Errorf("bytes enc instructions: expected % x got % x", bytesResult, b.Bytes()) + } + } + + // string + { + b.Reset() + var data string = "hello" + instr := &encInstr{encString, 6, nil, 0} + state := newEncoderState(b) + instr.op(instr, state, reflect.ValueOf(data)) + if !bytes.Equal(bytesResult, b.Bytes()) { + t.Errorf("string enc instructions: expected % x got % x", bytesResult, b.Bytes()) + } + } +} + +func execDec(instr *decInstr, state *decoderState, t *testing.T, value reflect.Value) { + defer testError(t) + v := int(state.decodeUint()) + if v+state.fieldnum != 6 { + t.Fatalf("decoding field number %d, got %d", 6, v+state.fieldnum) + } + instr.op(instr, state, value.Elem()) + state.fieldnum = 6 +} + +func newDecodeStateFromData(data []byte) *decoderState { + b := newDecBuffer(data) + state := newDecodeState(b) + state.fieldnum = -1 + return state +} + +// Test instruction execution for decoding. +// Do not run the machine yet; instead do individual instructions crafted by hand. +func TestScalarDecInstructions(t *testing.T) { + ovfl := errors.New("overflow") + + // bool + { + var data bool + instr := &decInstr{decBool, 6, nil, ovfl} + state := newDecodeStateFromData(boolResult) + execDec(instr, state, t, reflect.ValueOf(&data)) + if data != true { + t.Errorf("bool a = %v not true", data) + } + } + // int + { + var data int + instr := &decInstr{decOpTable[reflect.Int], 6, nil, ovfl} + state := newDecodeStateFromData(signedResult) + execDec(instr, state, t, reflect.ValueOf(&data)) + if data != 17 { + t.Errorf("int a = %v not 17", data) + } + } + + // uint + { + var data uint + instr := &decInstr{decOpTable[reflect.Uint], 6, nil, ovfl} + state := newDecodeStateFromData(unsignedResult) + execDec(instr, state, t, reflect.ValueOf(&data)) + if data != 17 { + t.Errorf("uint a = %v not 17", data) + } + } + + // int8 + { + var data int8 + instr := &decInstr{decInt8, 6, nil, ovfl} + state := newDecodeStateFromData(signedResult) + execDec(instr, state, t, reflect.ValueOf(&data)) + if data != 17 { + t.Errorf("int8 a = %v not 17", data) + } + } + + // uint8 + { + var data uint8 + instr := &decInstr{decUint8, 6, nil, ovfl} + state := newDecodeStateFromData(unsignedResult) + execDec(instr, state, t, reflect.ValueOf(&data)) + if data != 17 { + t.Errorf("uint8 a = %v not 17", data) + } + } + + // int16 + { + var data int16 + instr := &decInstr{decInt16, 6, nil, ovfl} + state := newDecodeStateFromData(signedResult) + execDec(instr, state, t, reflect.ValueOf(&data)) + if data != 17 { + t.Errorf("int16 a = %v not 17", data) + } + } + + // uint16 + { + var data uint16 + instr := &decInstr{decUint16, 6, nil, ovfl} + state := newDecodeStateFromData(unsignedResult) + execDec(instr, state, t, reflect.ValueOf(&data)) + if data != 17 { + t.Errorf("uint16 a = %v not 17", data) + } + } + + // int32 + { + var data int32 + instr := &decInstr{decInt32, 6, nil, ovfl} + state := newDecodeStateFromData(signedResult) + execDec(instr, state, t, reflect.ValueOf(&data)) + if data != 17 { + t.Errorf("int32 a = %v not 17", data) + } + } + + // uint32 + { + var data uint32 + instr := &decInstr{decUint32, 6, nil, ovfl} + state := newDecodeStateFromData(unsignedResult) + execDec(instr, state, t, reflect.ValueOf(&data)) + if data != 17 { + t.Errorf("uint32 a = %v not 17", data) + } + } + + // uintptr + { + var data uintptr + instr := &decInstr{decOpTable[reflect.Uintptr], 6, nil, ovfl} + state := newDecodeStateFromData(unsignedResult) + execDec(instr, state, t, reflect.ValueOf(&data)) + if data != 17 { + t.Errorf("uintptr a = %v not 17", data) + } + } + + // int64 + { + var data int64 + instr := &decInstr{decInt64, 6, nil, ovfl} + state := newDecodeStateFromData(signedResult) + execDec(instr, state, t, reflect.ValueOf(&data)) + if data != 17 { + t.Errorf("int64 a = %v not 17", data) + } + } + + // uint64 + { + var data uint64 + instr := &decInstr{decUint64, 6, nil, ovfl} + state := newDecodeStateFromData(unsignedResult) + execDec(instr, state, t, reflect.ValueOf(&data)) + if data != 17 { + t.Errorf("uint64 a = %v not 17", data) + } + } + + // float32 + { + var data float32 + instr := &decInstr{decFloat32, 6, nil, ovfl} + state := newDecodeStateFromData(floatResult) + execDec(instr, state, t, reflect.ValueOf(&data)) + if data != 17 { + t.Errorf("float32 a = %v not 17", data) + } + } + + // float64 + { + var data float64 + instr := &decInstr{decFloat64, 6, nil, ovfl} + state := newDecodeStateFromData(floatResult) + execDec(instr, state, t, reflect.ValueOf(&data)) + if data != 17 { + t.Errorf("float64 a = %v not 17", data) + } + } + + // complex64 + { + var data complex64 + instr := &decInstr{decOpTable[reflect.Complex64], 6, nil, ovfl} + state := newDecodeStateFromData(complexResult) + execDec(instr, state, t, reflect.ValueOf(&data)) + if data != 17+19i { + t.Errorf("complex a = %v not 17+19i", data) + } + } + + // complex128 + { + var data complex128 + instr := &decInstr{decOpTable[reflect.Complex128], 6, nil, ovfl} + state := newDecodeStateFromData(complexResult) + execDec(instr, state, t, reflect.ValueOf(&data)) + if data != 17+19i { + t.Errorf("complex a = %v not 17+19i", data) + } + } + + // bytes == []uint8 + { + var data []byte + instr := &decInstr{decUint8Slice, 6, nil, ovfl} + state := newDecodeStateFromData(bytesResult) + execDec(instr, state, t, reflect.ValueOf(&data)) + if string(data) != "hello" { + t.Errorf(`bytes a = %q not "hello"`, string(data)) + } + } + + // string + { + var data string + instr := &decInstr{decString, 6, nil, ovfl} + state := newDecodeStateFromData(bytesResult) + execDec(instr, state, t, reflect.ValueOf(&data)) + if data != "hello" { + t.Errorf(`bytes a = %q not "hello"`, data) + } + } +} + +func TestEndToEnd(t *testing.T) { + type T2 struct { + T string + } + type T3 struct { + X float64 + Z *int + } + type T1 struct { + A, B, C int + M map[string]*float64 + M2 map[int]T3 + Mstring map[string]string + Mintptr map[int]*int + Mcomp map[complex128]complex128 + Marr map[[2]string][2]*float64 + EmptyMap map[string]int // to check that we receive a non-nil map. + N *[3]float64 + Strs *[2]string + Int64s *[]int64 + RI complex64 + S string + Y []byte + T *T2 + } + pi := 3.14159 + e := 2.71828 + two := 2.0 + meaning := 42 + fingers := 5 + s1 := "string1" + s2 := "string2" + var comp1 complex128 = complex(1.0, 1.0) + var comp2 complex128 = complex(1.0, 1.0) + var arr1 [2]string + arr1[0] = s1 + arr1[1] = s2 + var arr2 [2]string + arr2[0] = s2 + arr2[1] = s1 + var floatArr1 [2]*float64 + floatArr1[0] = &pi + floatArr1[1] = &e + var floatArr2 [2]*float64 + floatArr2[0] = &e + floatArr2[1] = &two + t1 := &T1{ + A: 17, + B: 18, + C: -5, + M: map[string]*float64{"pi": &pi, "e": &e}, + M2: map[int]T3{4: {X: pi, Z: &meaning}, 10: {X: e, Z: &fingers}}, + Mstring: map[string]string{"pi": "3.14", "e": "2.71"}, + Mintptr: map[int]*int{meaning: &fingers, fingers: &meaning}, + Mcomp: map[complex128]complex128{comp1: comp2, comp2: comp1}, + Marr: map[[2]string][2]*float64{arr1: floatArr1, arr2: floatArr2}, + EmptyMap: make(map[string]int), + N: &[3]float64{1.5, 2.5, 3.5}, + Strs: &[2]string{s1, s2}, + Int64s: &[]int64{77, 89, 123412342134}, + RI: 17 - 23i, + S: "Now is the time", + Y: []byte("hello, sailor"), + T: &T2{"this is T2"}, + } + b := new(bytes.Buffer) + err := NewEncoder(b).Encode(t1) + if err != nil { + t.Error("encode:", err) + } + var _t1 T1 + err = NewDecoder(b).Decode(&_t1) + if err != nil { + t.Fatal("decode:", err) + } + if !reflect.DeepEqual(t1, &_t1) { + t.Errorf("encode expected %v got %v", *t1, _t1) + } + // Be absolutely sure the received map is non-nil. + if t1.EmptyMap == nil { + t.Errorf("nil map sent") + } + if _t1.EmptyMap == nil { + t.Errorf("nil map received") + } +} + +func TestOverflow(t *testing.T) { + type inputT struct { + Maxi int64 + Mini int64 + Maxu uint64 + Maxf float64 + Minf float64 + Maxc complex128 + Minc complex128 + } + var it inputT + var err error + b := new(bytes.Buffer) + enc := NewEncoder(b) + dec := NewDecoder(b) + + // int8 + b.Reset() + it = inputT{ + Maxi: math.MaxInt8 + 1, + } + type outi8 struct { + Maxi int8 + Mini int8 + } + var o1 outi8 + enc.Encode(it) + err = dec.Decode(&o1) + if err == nil || err.Error() != `value for "Maxi" out of range` { + t.Error("wrong overflow error for int8:", err) + } + it = inputT{ + Mini: math.MinInt8 - 1, + } + b.Reset() + enc.Encode(it) + err = dec.Decode(&o1) + if err == nil || err.Error() != `value for "Mini" out of range` { + t.Error("wrong underflow error for int8:", err) + } + + // int16 + b.Reset() + it = inputT{ + Maxi: math.MaxInt16 + 1, + } + type outi16 struct { + Maxi int16 + Mini int16 + } + var o2 outi16 + enc.Encode(it) + err = dec.Decode(&o2) + if err == nil || err.Error() != `value for "Maxi" out of range` { + t.Error("wrong overflow error for int16:", err) + } + it = inputT{ + Mini: math.MinInt16 - 1, + } + b.Reset() + enc.Encode(it) + err = dec.Decode(&o2) + if err == nil || err.Error() != `value for "Mini" out of range` { + t.Error("wrong underflow error for int16:", err) + } + + // int32 + b.Reset() + it = inputT{ + Maxi: math.MaxInt32 + 1, + } + type outi32 struct { + Maxi int32 + Mini int32 + } + var o3 outi32 + enc.Encode(it) + err = dec.Decode(&o3) + if err == nil || err.Error() != `value for "Maxi" out of range` { + t.Error("wrong overflow error for int32:", err) + } + it = inputT{ + Mini: math.MinInt32 - 1, + } + b.Reset() + enc.Encode(it) + err = dec.Decode(&o3) + if err == nil || err.Error() != `value for "Mini" out of range` { + t.Error("wrong underflow error for int32:", err) + } + + // uint8 + b.Reset() + it = inputT{ + Maxu: math.MaxUint8 + 1, + } + type outu8 struct { + Maxu uint8 + } + var o4 outu8 + enc.Encode(it) + err = dec.Decode(&o4) + if err == nil || err.Error() != `value for "Maxu" out of range` { + t.Error("wrong overflow error for uint8:", err) + } + + // uint16 + b.Reset() + it = inputT{ + Maxu: math.MaxUint16 + 1, + } + type outu16 struct { + Maxu uint16 + } + var o5 outu16 + enc.Encode(it) + err = dec.Decode(&o5) + if err == nil || err.Error() != `value for "Maxu" out of range` { + t.Error("wrong overflow error for uint16:", err) + } + + // uint32 + b.Reset() + it = inputT{ + Maxu: math.MaxUint32 + 1, + } + type outu32 struct { + Maxu uint32 + } + var o6 outu32 + enc.Encode(it) + err = dec.Decode(&o6) + if err == nil || err.Error() != `value for "Maxu" out of range` { + t.Error("wrong overflow error for uint32:", err) + } + + // float32 + b.Reset() + it = inputT{ + Maxf: math.MaxFloat32 * 2, + } + type outf32 struct { + Maxf float32 + Minf float32 + } + var o7 outf32 + enc.Encode(it) + err = dec.Decode(&o7) + if err == nil || err.Error() != `value for "Maxf" out of range` { + t.Error("wrong overflow error for float32:", err) + } + + // complex64 + b.Reset() + it = inputT{ + Maxc: complex(math.MaxFloat32*2, math.MaxFloat32*2), + } + type outc64 struct { + Maxc complex64 + Minc complex64 + } + var o8 outc64 + enc.Encode(it) + err = dec.Decode(&o8) + if err == nil || err.Error() != `value for "Maxc" out of range` { + t.Error("wrong overflow error for complex64:", err) + } +} + +func TestNesting(t *testing.T) { + type RT struct { + A string + Next *RT + } + rt := new(RT) + rt.A = "level1" + rt.Next = new(RT) + rt.Next.A = "level2" + b := new(bytes.Buffer) + NewEncoder(b).Encode(rt) + var drt RT + dec := NewDecoder(b) + err := dec.Decode(&drt) + if err != nil { + t.Fatal("decoder error:", err) + } + if drt.A != rt.A { + t.Errorf("nesting: encode expected %v got %v", *rt, drt) + } + if drt.Next == nil { + t.Errorf("nesting: recursion failed") + } + if drt.Next.A != rt.Next.A { + t.Errorf("nesting: encode expected %v got %v", *rt.Next, *drt.Next) + } +} + +// These three structures have the same data with different indirections +type T0 struct { + A int + B int + C int + D int +} +type T1 struct { + A int + B *int + C **int + D ***int +} +type T2 struct { + A ***int + B **int + C *int + D int +} + +func TestAutoIndirection(t *testing.T) { + // First transfer t1 into t0 + var t1 T1 + t1.A = 17 + t1.B = new(int) + *t1.B = 177 + t1.C = new(*int) + *t1.C = new(int) + **t1.C = 1777 + t1.D = new(**int) + *t1.D = new(*int) + **t1.D = new(int) + ***t1.D = 17777 + b := new(bytes.Buffer) + enc := NewEncoder(b) + enc.Encode(t1) + dec := NewDecoder(b) + var t0 T0 + dec.Decode(&t0) + if t0.A != 17 || t0.B != 177 || t0.C != 1777 || t0.D != 17777 { + t.Errorf("t1->t0: expected {17 177 1777 17777}; got %v", t0) + } + + // Now transfer t2 into t0 + var t2 T2 + t2.D = 17777 + t2.C = new(int) + *t2.C = 1777 + t2.B = new(*int) + *t2.B = new(int) + **t2.B = 177 + t2.A = new(**int) + *t2.A = new(*int) + **t2.A = new(int) + ***t2.A = 17 + b.Reset() + enc.Encode(t2) + t0 = T0{} + dec.Decode(&t0) + if t0.A != 17 || t0.B != 177 || t0.C != 1777 || t0.D != 17777 { + t.Errorf("t2->t0 expected {17 177 1777 17777}; got %v", t0) + } + + // Now transfer t0 into t1 + t0 = T0{17, 177, 1777, 17777} + b.Reset() + enc.Encode(t0) + t1 = T1{} + dec.Decode(&t1) + if t1.A != 17 || *t1.B != 177 || **t1.C != 1777 || ***t1.D != 17777 { + t.Errorf("t0->t1 expected {17 177 1777 17777}; got {%d %d %d %d}", t1.A, *t1.B, **t1.C, ***t1.D) + } + + // Now transfer t0 into t2 + b.Reset() + enc.Encode(t0) + t2 = T2{} + dec.Decode(&t2) + if ***t2.A != 17 || **t2.B != 177 || *t2.C != 1777 || t2.D != 17777 { + t.Errorf("t0->t2 expected {17 177 1777 17777}; got {%d %d %d %d}", ***t2.A, **t2.B, *t2.C, t2.D) + } + + // Now do t2 again but without pre-allocated pointers. + b.Reset() + enc.Encode(t0) + ***t2.A = 0 + **t2.B = 0 + *t2.C = 0 + t2.D = 0 + dec.Decode(&t2) + if ***t2.A != 17 || **t2.B != 177 || *t2.C != 1777 || t2.D != 17777 { + t.Errorf("t0->t2 expected {17 177 1777 17777}; got {%d %d %d %d}", ***t2.A, **t2.B, *t2.C, t2.D) + } +} + +type RT0 struct { + A int + B string + C float64 +} +type RT1 struct { + C float64 + B string + A int + NotSet string +} + +func TestReorderedFields(t *testing.T) { + var rt0 RT0 + rt0.A = 17 + rt0.B = "hello" + rt0.C = 3.14159 + b := new(bytes.Buffer) + NewEncoder(b).Encode(rt0) + dec := NewDecoder(b) + var rt1 RT1 + // Wire type is RT0, local type is RT1. + err := dec.Decode(&rt1) + if err != nil { + t.Fatal("decode error:", err) + } + if rt0.A != rt1.A || rt0.B != rt1.B || rt0.C != rt1.C { + t.Errorf("rt1->rt0: expected %v; got %v", rt0, rt1) + } +} + +// Like an RT0 but with fields we'll ignore on the decode side. +type IT0 struct { + A int64 + B string + Ignore_d []int + Ignore_e [3]float64 + Ignore_f bool + Ignore_g string + Ignore_h []byte + Ignore_i *RT1 + Ignore_m map[string]int + C float64 +} + +func TestIgnoredFields(t *testing.T) { + var it0 IT0 + it0.A = 17 + it0.B = "hello" + it0.C = 3.14159 + it0.Ignore_d = []int{1, 2, 3} + it0.Ignore_e[0] = 1.0 + it0.Ignore_e[1] = 2.0 + it0.Ignore_e[2] = 3.0 + it0.Ignore_f = true + it0.Ignore_g = "pay no attention" + it0.Ignore_h = []byte("to the curtain") + it0.Ignore_i = &RT1{3.1, "hi", 7, "hello"} + it0.Ignore_m = map[string]int{"one": 1, "two": 2} + + b := new(bytes.Buffer) + NewEncoder(b).Encode(it0) + dec := NewDecoder(b) + var rt1 RT1 + // Wire type is IT0, local type is RT1. + err := dec.Decode(&rt1) + if err != nil { + t.Error("error: ", err) + } + if int(it0.A) != rt1.A || it0.B != rt1.B || it0.C != rt1.C { + t.Errorf("rt0->rt1: expected %v; got %v", it0, rt1) + } +} + +func TestBadRecursiveType(t *testing.T) { + type Rec ***Rec + var rec Rec + b := new(bytes.Buffer) + err := NewEncoder(b).Encode(&rec) + if err == nil { + t.Error("expected error; got none") + } else if !strings.Contains(err.Error(), "recursive") { + t.Error("expected recursive type error; got", err) + } + // Can't test decode easily because we can't encode one, so we can't pass one to a Decoder. +} + +type Indirect struct { + A ***[3]int + S ***[]int + M ****map[string]int +} + +type Direct struct { + A [3]int + S []int + M map[string]int +} + +func TestIndirectSliceMapArray(t *testing.T) { + // Marshal indirect, unmarshal to direct. + i := new(Indirect) + i.A = new(**[3]int) + *i.A = new(*[3]int) + **i.A = new([3]int) + ***i.A = [3]int{1, 2, 3} + i.S = new(**[]int) + *i.S = new(*[]int) + **i.S = new([]int) + ***i.S = []int{4, 5, 6} + i.M = new(***map[string]int) + *i.M = new(**map[string]int) + **i.M = new(*map[string]int) + ***i.M = new(map[string]int) + ****i.M = map[string]int{"one": 1, "two": 2, "three": 3} + b := new(bytes.Buffer) + NewEncoder(b).Encode(i) + dec := NewDecoder(b) + var d Direct + err := dec.Decode(&d) + if err != nil { + t.Error("error: ", err) + } + if len(d.A) != 3 || d.A[0] != 1 || d.A[1] != 2 || d.A[2] != 3 { + t.Errorf("indirect to direct: d.A is %v not %v", d.A, ***i.A) + } + if len(d.S) != 3 || d.S[0] != 4 || d.S[1] != 5 || d.S[2] != 6 { + t.Errorf("indirect to direct: d.S is %v not %v", d.S, ***i.S) + } + if len(d.M) != 3 || d.M["one"] != 1 || d.M["two"] != 2 || d.M["three"] != 3 { + t.Errorf("indirect to direct: d.M is %v not %v", d.M, ***i.M) + } + // Marshal direct, unmarshal to indirect. + d.A = [3]int{11, 22, 33} + d.S = []int{44, 55, 66} + d.M = map[string]int{"four": 4, "five": 5, "six": 6} + i = new(Indirect) + b.Reset() + NewEncoder(b).Encode(d) + dec = NewDecoder(b) + err = dec.Decode(&i) + if err != nil { + t.Fatal("error: ", err) + } + if len(***i.A) != 3 || (***i.A)[0] != 11 || (***i.A)[1] != 22 || (***i.A)[2] != 33 { + t.Errorf("direct to indirect: ***i.A is %v not %v", ***i.A, d.A) + } + if len(***i.S) != 3 || (***i.S)[0] != 44 || (***i.S)[1] != 55 || (***i.S)[2] != 66 { + t.Errorf("direct to indirect: ***i.S is %v not %v", ***i.S, ***i.S) + } + if len(****i.M) != 3 || (****i.M)["four"] != 4 || (****i.M)["five"] != 5 || (****i.M)["six"] != 6 { + t.Errorf("direct to indirect: ****i.M is %v not %v", ****i.M, d.M) + } +} + +// An interface with several implementations +type Squarer interface { + Square() int +} + +type Int int + +func (i Int) Square() int { + return int(i * i) +} + +type Float float64 + +func (f Float) Square() int { + return int(f * f) +} + +type Vector []int + +func (v Vector) Square() int { + sum := 0 + for _, x := range v { + sum += x * x + } + return sum +} + +type Point struct { + X, Y int +} + +func (p Point) Square() int { + return p.X*p.X + p.Y*p.Y +} + +// A struct with interfaces in it. +type InterfaceItem struct { + I int + Sq1, Sq2, Sq3 Squarer + F float64 + Sq []Squarer +} + +// The same struct without interfaces +type NoInterfaceItem struct { + I int + F float64 +} + +func TestInterface(t *testing.T) { + iVal := Int(3) + fVal := Float(5) + // Sending a Vector will require that the receiver define a type in the middle of + // receiving the value for item2. + vVal := Vector{1, 2, 3} + b := new(bytes.Buffer) + item1 := &InterfaceItem{1, iVal, fVal, vVal, 11.5, []Squarer{iVal, fVal, nil, vVal}} + // Register the types. + Register(Int(0)) + Register(Float(0)) + Register(Vector{}) + err := NewEncoder(b).Encode(item1) + if err != nil { + t.Error("expected no encode error; got", err) + } + + item2 := InterfaceItem{} + err = NewDecoder(b).Decode(&item2) + if err != nil { + t.Fatal("decode:", err) + } + if item2.I != item1.I { + t.Error("normal int did not decode correctly") + } + if item2.Sq1 == nil || item2.Sq1.Square() != iVal.Square() { + t.Error("Int did not decode correctly") + } + if item2.Sq2 == nil || item2.Sq2.Square() != fVal.Square() { + t.Error("Float did not decode correctly") + } + if item2.Sq3 == nil || item2.Sq3.Square() != vVal.Square() { + t.Error("Vector did not decode correctly") + } + if item2.F != item1.F { + t.Error("normal float did not decode correctly") + } + // Now check that we received a slice of Squarers correctly, including a nil element + if len(item1.Sq) != len(item2.Sq) { + t.Fatalf("[]Squarer length wrong: got %d; expected %d", len(item2.Sq), len(item1.Sq)) + } + for i, v1 := range item1.Sq { + v2 := item2.Sq[i] + if v1 == nil || v2 == nil { + if v1 != nil || v2 != nil { + t.Errorf("item %d inconsistent nils", i) + } + } else if v1.Square() != v2.Square() { + t.Errorf("item %d inconsistent values: %v %v", i, v1, v2) + } + } +} + +// A struct with all basic types, stored in interfaces. +type BasicInterfaceItem struct { + Int, Int8, Int16, Int32, Int64 any + Uint, Uint8, Uint16, Uint32, Uint64 any + Float32, Float64 any + Complex64, Complex128 any + Bool any + String any + Bytes any +} + +func TestInterfaceBasic(t *testing.T) { + b := new(bytes.Buffer) + item1 := &BasicInterfaceItem{ + int(1), int8(1), int16(1), int32(1), int64(1), + uint(1), uint8(1), uint16(1), uint32(1), uint64(1), + float32(1), 1.0, + complex64(1i), complex128(1i), + true, + "hello", + []byte("sailor"), + } + err := NewEncoder(b).Encode(item1) + if err != nil { + t.Error("expected no encode error; got", err) + } + + item2 := &BasicInterfaceItem{} + err = NewDecoder(b).Decode(&item2) + if err != nil { + t.Fatal("decode:", err) + } + if !reflect.DeepEqual(item1, item2) { + t.Errorf("encode expected %v got %v", item1, item2) + } + // Hand check a couple for correct types. + if v, ok := item2.Bool.(bool); !ok || !v { + t.Error("boolean should be true") + } + if v, ok := item2.String.(string); !ok || v != item1.String.(string) { + t.Errorf("string should be %v is %v", item1.String, v) + } +} + +type String string + +type PtrInterfaceItem struct { + Str1 any // basic + Str2 any // derived +} + +// We'll send pointers; should receive values. +// Also check that we can register T but send *T. +func TestInterfacePointer(t *testing.T) { + b := new(bytes.Buffer) + str1 := "howdy" + str2 := String("kiddo") + item1 := &PtrInterfaceItem{ + &str1, + &str2, + } + // Register the type. + Register(str2) + err := NewEncoder(b).Encode(item1) + if err != nil { + t.Error("expected no encode error; got", err) + } + + item2 := &PtrInterfaceItem{} + err = NewDecoder(b).Decode(&item2) + if err != nil { + t.Fatal("decode:", err) + } + // Hand test for correct types and values. + if v, ok := item2.Str1.(string); !ok || v != str1 { + t.Errorf("basic string failed: %q should be %q", v, str1) + } + if v, ok := item2.Str2.(String); !ok || v != str2 { + t.Errorf("derived type String failed: %q should be %q", v, str2) + } +} + +func TestIgnoreInterface(t *testing.T) { + iVal := Int(3) + fVal := Float(5) + // Sending a Point will require that the receiver define a type in the middle of + // receiving the value for item2. + pVal := Point{2, 3} + b := new(bytes.Buffer) + item1 := &InterfaceItem{1, iVal, fVal, pVal, 11.5, nil} + // Register the types. + Register(Int(0)) + Register(Float(0)) + Register(Point{}) + err := NewEncoder(b).Encode(item1) + if err != nil { + t.Error("expected no encode error; got", err) + } + + item2 := NoInterfaceItem{} + err = NewDecoder(b).Decode(&item2) + if err != nil { + t.Fatal("decode:", err) + } + if item2.I != item1.I { + t.Error("normal int did not decode correctly") + } + if item2.F != item1.F { + t.Error("normal float did not decode correctly") + } +} + +type U struct { + A int + B string + c float64 + D uint +} + +func TestUnexportedFields(t *testing.T) { + var u0 U + u0.A = 17 + u0.B = "hello" + u0.c = 3.14159 + u0.D = 23 + b := new(bytes.Buffer) + NewEncoder(b).Encode(u0) + dec := NewDecoder(b) + var u1 U + u1.c = 1234. + err := dec.Decode(&u1) + if err != nil { + t.Fatal("decode error:", err) + } + if u0.A != u1.A || u0.B != u1.B || u0.D != u1.D { + t.Errorf("u1->u0: expected %v; got %v", u0, u1) + } + if u1.c != 1234. { + t.Error("u1.c modified") + } +} + +var singletons = []any{ + true, + 7, + uint(10), + 3.2, + "hello", + [3]int{11, 22, 33}, + []float32{0.5, 0.25, 0.125}, + map[string]int{"one": 1, "two": 2}, +} + +func TestDebugSingleton(t *testing.T) { + if debugFunc == nil { + return + } + b := new(bytes.Buffer) + // Accumulate a number of values and print them out all at once. + for _, x := range singletons { + err := NewEncoder(b).Encode(x) + if err != nil { + t.Fatal("encode:", err) + } + } + debugFunc(b) +} + +// A type that won't be defined in the gob until we send it in an interface value. +type OnTheFly struct { + A int +} + +type DT struct { + // X OnTheFly + A int + B string + C float64 + I any + J any + I_nil any + M map[string]int + T [3]int + S []string +} + +func newDT() DT { + var dt DT + dt.A = 17 + dt.B = "hello" + dt.C = 3.14159 + dt.I = 271828 + dt.J = OnTheFly{3} + dt.I_nil = nil + dt.M = map[string]int{"one": 1, "two": 2} + dt.T = [3]int{11, 22, 33} + dt.S = []string{"hi", "joe"} + return dt +} + +func TestDebugStruct(t *testing.T) { + if debugFunc == nil { + return + } + Register(OnTheFly{}) + dt := newDT() + b := new(bytes.Buffer) + err := NewEncoder(b).Encode(dt) + if err != nil { + t.Fatal("encode:", err) + } + debugBuffer := bytes.NewBuffer(b.Bytes()) + dt2 := &DT{} + err = NewDecoder(b).Decode(&dt2) + if err != nil { + t.Error("decode:", err) + } + debugFunc(debugBuffer) +} + +func encFuzzDec(rng *rand.Rand, in any) error { + buf := new(bytes.Buffer) + enc := NewEncoder(buf) + if err := enc.Encode(&in); err != nil { + return err + } + + b := buf.Bytes() + for i, bi := range b { + if rng.Intn(10) < 3 { + b[i] = bi + uint8(rng.Intn(256)) + } + } + + dec := NewDecoder(buf) + var e any + if err := dec.Decode(&e); err != nil { + return err + } + return nil +} + +// This does some "fuzz testing" by attempting to decode a sequence of random bytes. +func TestFuzz(t *testing.T) { + if !*doFuzzTests { + t.Skipf("disabled; run with -gob.fuzz to enable") + } + + // all possible inputs + input := []any{ + new(int), + new(float32), + new(float64), + new(complex128), + &ByteStruct{255}, + &ArrayStruct{}, + &StringStruct{"hello"}, + &GobTest1{0, &StringStruct{"hello"}}, + } + testFuzz(t, time.Now().UnixNano(), 100, input...) +} + +func TestFuzzRegressions(t *testing.T) { + if !*doFuzzTests { + t.Skipf("disabled; run with -gob.fuzz to enable") + } + + // An instance triggering a type name of length ~102 GB. + testFuzz(t, 1328492090837718000, 100, new(float32)) + // An instance triggering a type name of 1.6 GB. + // Note: can take several minutes to run. + testFuzz(t, 1330522872628565000, 100, new(int)) +} + +func testFuzz(t *testing.T, seed int64, n int, input ...any) { + for _, e := range input { + t.Logf("seed=%d n=%d e=%T", seed, n, e) + rng := rand.New(rand.NewSource(seed)) + for i := 0; i < n; i++ { + encFuzzDec(rng, e) + } + } +} + +// TestFuzzOneByte tries to decode corrupted input sequences +// and checks that no panic occurs. +func TestFuzzOneByte(t *testing.T) { + if !*doFuzzTests { + t.Skipf("disabled; run with -gob.fuzz to enable") + } + + buf := new(strings.Builder) + Register(OnTheFly{}) + dt := newDT() + if err := NewEncoder(buf).Encode(dt); err != nil { + t.Fatal(err) + } + s := buf.String() + + indices := make([]int, 0, len(s)) + for i := 0; i < len(s); i++ { + switch i { + case 14, 167, 231, 265: // a slice length, corruptions are not handled yet. + continue + case 248: + // Large map size, which currently causes an out of memory panic. + // See golang.org/issue/24308 and golang.org/issue/20221. + continue + } + indices = append(indices, i) + } + if testing.Short() { + indices = []int{1, 111, 178} // known fixed panics + } + for _, i := range indices { + for j := 0; j < 256; j += 3 { + b := []byte(s) + b[i] ^= byte(j) + var e DT + func() { + defer func() { + if p := recover(); p != nil { + t.Errorf("crash for b[%d] ^= 0x%x", i, j) + panic(p) + } + }() + err := NewDecoder(bytes.NewReader(b)).Decode(&e) + _ = err + }() + } + } +} + +// Don't crash, just give error with invalid type id. +// Issue 9649. +func TestErrorInvalidTypeId(t *testing.T) { + data := []byte{0x01, 0x00, 0x01, 0x00} + d := NewDecoder(bytes.NewReader(data)) + // When running d.Decode(&foo) the first time the decoder stops + // after []byte{0x01, 0x00} and reports an errBadType. Running + // d.Decode(&foo) again on exactly the same input sequence should + // give another errBadType, but instead caused a panic because + // decoderMap wasn't cleaned up properly after the first error. + for i := 0; i < 2; i++ { + var foo struct{} + err := d.Decode(&foo) + if err != errBadType { + t.Fatalf("decode: expected %s, got %s", errBadType, err) + } + } +} + +type LargeSliceByte struct { + S []byte +} + +type LargeSliceInt8 struct { + S []int8 +} + +type StringPair struct { + A, B string +} + +type LargeSliceStruct struct { + S []StringPair +} + +func testEncodeDecode(t *testing.T, in, out any) { + t.Helper() + var b bytes.Buffer + err := NewEncoder(&b).Encode(in) + if err != nil { + t.Fatal("encode:", err) + } + err = NewDecoder(&b).Decode(out) + if err != nil { + t.Fatal("decode:", err) + } + if !reflect.DeepEqual(in, out) { + t.Errorf("output mismatch") + } +} + +func TestLargeSlice(t *testing.T) { + t.Run("byte", func(t *testing.T) { + t.Parallel() + s := make([]byte, 10<<21) + for i := range s { + s[i] = byte(i) + } + st := &LargeSliceByte{S: s} + rt := &LargeSliceByte{} + testEncodeDecode(t, st, rt) + }) + t.Run("int8", func(t *testing.T) { + t.Parallel() + s := make([]int8, 10<<21) + for i := range s { + s[i] = int8(i) + } + st := &LargeSliceInt8{S: s} + rt := &LargeSliceInt8{} + testEncodeDecode(t, st, rt) + }) + t.Run("struct", func(t *testing.T) { + t.Parallel() + s := make([]StringPair, 1<<21) + for i := range s { + s[i].A = string(rune(i)) + s[i].B = s[i].A + } + st := &LargeSliceStruct{S: s} + rt := &LargeSliceStruct{} + testEncodeDecode(t, st, rt) + }) +} diff --git a/src/encoding/gob/debug.go b/src/encoding/gob/debug.go new file mode 100644 index 0000000..dc572fc --- /dev/null +++ b/src/encoding/gob/debug.go @@ -0,0 +1,733 @@ +// 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. + +// Delete the next line to include in the gob package. +// +//go:build ignore + +package gob + +// This file is not normally included in the gob package. Used only for debugging the package itself. +// Except for reading uints, it is an implementation of a reader that is independent of +// the one implemented by Decoder. +// To enable the Debug function, delete the +build ignore line above and do +// go install + +import ( + "bytes" + "fmt" + "io" + "os" + "strings" + "sync" +) + +var dumpBytes = false // If true, print the remaining bytes in the input buffer at each item. + +// Init installs the debugging facility. If this file is not compiled in the +// package, the tests in codec_test.go are no-ops. +func init() { + debugFunc = Debug +} + +var ( + blanks = bytes.Repeat([]byte{' '}, 3*10) + empty = []byte(": \n") + tabs = strings.Repeat("\t", 100) +) + +// tab indents itself when printed. +type tab int + +func (t tab) String() string { + n := int(t) + if n > len(tabs) { + n = len(tabs) + } + return tabs[0:n] +} + +func (t tab) print() { + fmt.Fprint(os.Stderr, t) +} + +// A peekReader wraps an io.Reader, allowing one to peek ahead to see +// what's coming without stealing the data from the client of the Reader. +type peekReader struct { + r io.Reader + data []byte // read-ahead data +} + +// newPeekReader returns a peekReader that wraps r. +func newPeekReader(r io.Reader) *peekReader { + return &peekReader{r: r} +} + +// Read is the usual method. It will first take data that has been read ahead. +func (p *peekReader) Read(b []byte) (n int, err error) { + if len(p.data) == 0 { + return p.r.Read(b) + } + // Satisfy what's possible from the read-ahead data. + n = copy(b, p.data) + // Move data down to beginning of slice, to avoid endless growth + copy(p.data, p.data[n:]) + p.data = p.data[:len(p.data)-n] + return +} + +// peek returns as many bytes as possible from the unread +// portion of the stream, up to the length of b. +func (p *peekReader) peek(b []byte) (n int, err error) { + if len(p.data) > 0 { + n = copy(b, p.data) + if n == len(b) { + return + } + b = b[n:] + } + if len(b) == 0 { + return + } + m, e := io.ReadFull(p.r, b) + if m > 0 { + p.data = append(p.data, b[:m]...) + } + n += m + if e == io.ErrUnexpectedEOF { + // That means m > 0 but we reached EOF. If we got data + // we won't complain about not being able to peek enough. + if n > 0 { + e = nil + } else { + e = io.EOF + } + } + return n, e +} + +type debugger struct { + mutex sync.Mutex + remain int // the number of bytes known to remain in the input + remainingKnown bool // the value of 'remain' is valid + r *peekReader + wireType map[typeId]*wireType + tmp []byte // scratch space for decoding uints. +} + +// dump prints the next nBytes of the input. +// It arranges to print the output aligned from call to +// call, to make it easy to see what has been consumed. +func (deb *debugger) dump(format string, args ...any) { + if !dumpBytes { + return + } + fmt.Fprintf(os.Stderr, format+" ", args...) + if !deb.remainingKnown { + return + } + if deb.remain < 0 { + fmt.Fprintf(os.Stderr, "remaining byte count is negative! %d\n", deb.remain) + return + } + data := make([]byte, deb.remain) + n, _ := deb.r.peek(data) + if n == 0 { + os.Stderr.Write(empty) + return + } + b := new(bytes.Buffer) + fmt.Fprintf(b, "[%d]{\n", deb.remain) + // Blanks until first byte + lineLength := 0 + if n := len(data); n%10 != 0 { + lineLength = 10 - n%10 + fmt.Fprintf(b, "\t%s", blanks[:lineLength*3]) + } + // 10 bytes per line + for len(data) > 0 { + if lineLength == 0 { + fmt.Fprint(b, "\t") + } + m := 10 - lineLength + lineLength = 0 + if m > len(data) { + m = len(data) + } + fmt.Fprintf(b, "% x\n", data[:m]) + data = data[m:] + } + fmt.Fprint(b, "}\n") + os.Stderr.Write(b.Bytes()) +} + +// Debug prints a human-readable representation of the gob data read from r. +// It is a no-op unless debugging was enabled when the package was built. +func Debug(r io.Reader) { + err := debug(r) + if err != nil { + fmt.Fprintf(os.Stderr, "gob debug: %s\n", err) + } +} + +// debug implements Debug, but catches panics and returns +// them as errors to be printed by Debug. +func debug(r io.Reader) (err error) { + defer catchError(&err) + fmt.Fprintln(os.Stderr, "Start of debugging") + deb := &debugger{ + r: newPeekReader(r), + wireType: make(map[typeId]*wireType), + tmp: make([]byte, 16), + } + if b, ok := r.(*bytes.Buffer); ok { + deb.remain = b.Len() + deb.remainingKnown = true + } + deb.gobStream() + return +} + +// note that we've consumed some bytes +func (deb *debugger) consumed(n int) { + if deb.remainingKnown { + deb.remain -= n + } +} + +// int64 decodes and returns the next integer, which must be present. +// Don't call this if you could be at EOF. +func (deb *debugger) int64() int64 { + return toInt(deb.uint64()) +} + +// uint64 returns and decodes the next unsigned integer, which must be present. +// Don't call this if you could be at EOF. +// TODO: handle errors better. +func (deb *debugger) uint64() uint64 { + n, w, err := decodeUintReader(deb.r, deb.tmp) + if err != nil { + errorf("debug: read error: %s", err) + } + deb.consumed(w) + return n +} + +// GobStream: +// +// DelimitedMessage* (until EOF) +func (deb *debugger) gobStream() { + // Make sure we're single-threaded through here. + deb.mutex.Lock() + defer deb.mutex.Unlock() + + for deb.delimitedMessage(0) { + } +} + +// DelimitedMessage: +// +// uint(lengthOfMessage) Message +func (deb *debugger) delimitedMessage(indent tab) bool { + for { + n := deb.loadBlock(true) + if n < 0 { + return false + } + deb.dump("Delimited message of length %d", n) + deb.message(indent) + } + return true +} + +// loadBlock preps us to read a message +// of the length specified next in the input. It returns +// the length of the block. The argument tells whether +// an EOF is acceptable now. If it is and one is found, +// the return value is negative. +func (deb *debugger) loadBlock(eofOK bool) int { + n64, w, err := decodeUintReader(deb.r, deb.tmp) // deb.uint64 will error at EOF + if err != nil { + if eofOK && err == io.EOF { + return -1 + } + errorf("debug: unexpected error: %s", err) + } + deb.consumed(w) + n := int(n64) + if n < 0 { + errorf("huge value for message length: %d", n64) + } + return int(n) +} + +// Message: +// +// TypeSequence TypedValue +// +// TypeSequence +// +// (TypeDefinition DelimitedTypeDefinition*)? +// +// DelimitedTypeDefinition: +// +// uint(lengthOfTypeDefinition) TypeDefinition +// +// TypedValue: +// +// int(typeId) Value +func (deb *debugger) message(indent tab) bool { + for { + // Convert the uint64 to a signed integer typeId + uid := deb.int64() + id := typeId(uid) + deb.dump("type id=%d", id) + if id < 0 { + deb.typeDefinition(indent, -id) + n := deb.loadBlock(false) + deb.dump("Message of length %d", n) + continue + } else { + deb.value(indent, id) + break + } + } + return true +} + +// Helper methods to make it easy to scan a type descriptor. + +// common returns the CommonType at the input point. +func (deb *debugger) common() CommonType { + fieldNum := -1 + name := "" + id := typeId(0) + for { + delta := deb.delta(-1) + if delta == 0 { + break + } + fieldNum += delta + switch fieldNum { + case 0: + name = deb.string() + case 1: + // Id typeId + id = deb.typeId() + default: + errorf("corrupted CommonType, delta is %d fieldNum is %d", delta, fieldNum) + } + } + return CommonType{name, id} +} + +// uint returns the unsigned int at the input point, as a uint (not uint64). +func (deb *debugger) uint() uint { + return uint(deb.uint64()) +} + +// int returns the signed int at the input point, as an int (not int64). +func (deb *debugger) int() int { + return int(deb.int64()) +} + +// typeId returns the type id at the input point. +func (deb *debugger) typeId() typeId { + return typeId(deb.int64()) +} + +// string returns the string at the input point. +func (deb *debugger) string() string { + x := int(deb.uint64()) + b := make([]byte, x) + nb, _ := deb.r.Read(b) + if nb != x { + errorf("corrupted type") + } + deb.consumed(nb) + return string(b) +} + +// delta returns the field delta at the input point. The expect argument, +// if non-negative, identifies what the value should be. +func (deb *debugger) delta(expect int) int { + delta := int(deb.uint64()) + if delta < 0 || (expect >= 0 && delta != expect) { + errorf("decode: corrupted type: delta %d expected %d", delta, expect) + } + return delta +} + +// TypeDefinition: +// +// [int(-typeId) (already read)] encodingOfWireType +func (deb *debugger) typeDefinition(indent tab, id typeId) { + deb.dump("type definition for id %d", id) + // Encoding is of a wireType. Decode the structure as usual + fieldNum := -1 + wire := new(wireType) + // A wireType defines a single field. + delta := deb.delta(-1) + fieldNum += delta + switch fieldNum { + case 0: // array type, one field of {{Common}, elem, length} + // Field number 0 is CommonType + deb.delta(1) + com := deb.common() + // Field number 1 is type Id of elem + deb.delta(1) + id := deb.typeId() + // Field number 3 is length + deb.delta(1) + length := deb.int() + wire.ArrayT = &arrayType{com, id, length} + + case 1: // slice type, one field of {{Common}, elem} + // Field number 0 is CommonType + deb.delta(1) + com := deb.common() + // Field number 1 is type Id of elem + deb.delta(1) + id := deb.typeId() + wire.SliceT = &sliceType{com, id} + + case 2: // struct type, one field of {{Common}, []fieldType} + // Field number 0 is CommonType + deb.delta(1) + com := deb.common() + // Field number 1 is slice of FieldType + deb.delta(1) + numField := int(deb.uint()) + field := make([]*fieldType, numField) + for i := 0; i < numField; i++ { + field[i] = new(fieldType) + deb.delta(1) // field 0 of fieldType: name + field[i].Name = deb.string() + deb.delta(1) // field 1 of fieldType: id + field[i].Id = deb.typeId() + deb.delta(0) // end of fieldType + } + wire.StructT = &structType{com, field} + + case 3: // map type, one field of {{Common}, key, elem} + // Field number 0 is CommonType + deb.delta(1) + com := deb.common() + // Field number 1 is type Id of key + deb.delta(1) + keyId := deb.typeId() + // Field number 2 is type Id of elem + deb.delta(1) + elemId := deb.typeId() + wire.MapT = &mapType{com, keyId, elemId} + case 4: // GobEncoder type, one field of {{Common}} + // Field number 0 is CommonType + deb.delta(1) + com := deb.common() + wire.GobEncoderT = &gobEncoderType{com} + case 5: // BinaryMarshaler type, one field of {{Common}} + // Field number 0 is CommonType + deb.delta(1) + com := deb.common() + wire.BinaryMarshalerT = &gobEncoderType{com} + case 6: // TextMarshaler type, one field of {{Common}} + // Field number 0 is CommonType + deb.delta(1) + com := deb.common() + wire.TextMarshalerT = &gobEncoderType{com} + default: + errorf("bad field in type %d", fieldNum) + } + deb.printWireType(indent, wire) + deb.delta(0) // end inner type (arrayType, etc.) + deb.delta(0) // end wireType + // Remember we've seen this type. + deb.wireType[id] = wire +} + +// Value: +// +// SingletonValue | StructValue +func (deb *debugger) value(indent tab, id typeId) { + wire, ok := deb.wireType[id] + if ok && wire.StructT != nil { + deb.structValue(indent, id) + } else { + deb.singletonValue(indent, id) + } +} + +// SingletonValue: +// +// uint(0) FieldValue +func (deb *debugger) singletonValue(indent tab, id typeId) { + deb.dump("Singleton value") + // is it a builtin type? + wire := deb.wireType[id] + _, ok := builtinIdToType[id] + if !ok && wire == nil { + errorf("type id %d not defined", id) + } + m := deb.uint64() + if m != 0 { + errorf("expected zero; got %d", m) + } + deb.fieldValue(indent, id) +} + +// InterfaceValue: +// +// NilInterfaceValue | NonNilInterfaceValue +func (deb *debugger) interfaceValue(indent tab) { + deb.dump("Start of interface value") + if nameLen := deb.uint64(); nameLen == 0 { + deb.nilInterfaceValue(indent) + } else { + deb.nonNilInterfaceValue(indent, int(nameLen)) + } +} + +// NilInterfaceValue: +// +// uint(0) [already read] +func (deb *debugger) nilInterfaceValue(indent tab) int { + fmt.Fprintf(os.Stderr, "%snil interface\n", indent) + return 0 +} + +// NonNilInterfaceValue: +// +// ConcreteTypeName TypeSequence InterfaceContents +// +// ConcreteTypeName: +// +// uint(lengthOfName) [already read=n] name +// +// InterfaceContents: +// +// int(concreteTypeId) DelimitedValue +// +// DelimitedValue: +// +// uint(length) Value +func (deb *debugger) nonNilInterfaceValue(indent tab, nameLen int) { + // ConcreteTypeName + b := make([]byte, nameLen) + deb.r.Read(b) // TODO: CHECK THESE READS!! + deb.consumed(nameLen) + name := string(b) + + for { + id := deb.typeId() + if id < 0 { + deb.typeDefinition(indent, -id) + n := deb.loadBlock(false) + deb.dump("Nested message of length %d", n) + } else { + // DelimitedValue + x := deb.uint64() // in case we want to ignore the value; we don't. + fmt.Fprintf(os.Stderr, "%sinterface value, type %q id=%d; valueLength %d\n", indent, name, id, x) + deb.value(indent, id) + break + } + } +} + +// printCommonType prints a common type; used by printWireType. +func (deb *debugger) printCommonType(indent tab, kind string, common *CommonType) { + indent.print() + fmt.Fprintf(os.Stderr, "%s %q id=%d\n", kind, common.Name, common.Id) +} + +// printWireType prints the contents of a wireType. +func (deb *debugger) printWireType(indent tab, wire *wireType) { + fmt.Fprintf(os.Stderr, "%stype definition {\n", indent) + indent++ + switch { + case wire.ArrayT != nil: + deb.printCommonType(indent, "array", &wire.ArrayT.CommonType) + fmt.Fprintf(os.Stderr, "%slen %d\n", indent+1, wire.ArrayT.Len) + fmt.Fprintf(os.Stderr, "%selemid %d\n", indent+1, wire.ArrayT.Elem) + case wire.MapT != nil: + deb.printCommonType(indent, "map", &wire.MapT.CommonType) + fmt.Fprintf(os.Stderr, "%skey id=%d\n", indent+1, wire.MapT.Key) + fmt.Fprintf(os.Stderr, "%selem id=%d\n", indent+1, wire.MapT.Elem) + case wire.SliceT != nil: + deb.printCommonType(indent, "slice", &wire.SliceT.CommonType) + fmt.Fprintf(os.Stderr, "%selem id=%d\n", indent+1, wire.SliceT.Elem) + case wire.StructT != nil: + deb.printCommonType(indent, "struct", &wire.StructT.CommonType) + for i, field := range wire.StructT.Field { + fmt.Fprintf(os.Stderr, "%sfield %d:\t%s\tid=%d\n", indent+1, i, field.Name, field.Id) + } + case wire.GobEncoderT != nil: + deb.printCommonType(indent, "GobEncoder", &wire.GobEncoderT.CommonType) + } + indent-- + fmt.Fprintf(os.Stderr, "%s}\n", indent) +} + +// fieldValue prints a value of any type, such as a struct field. +// FieldValue: +// +// builtinValue | ArrayValue | MapValue | SliceValue | StructValue | InterfaceValue +func (deb *debugger) fieldValue(indent tab, id typeId) { + _, ok := builtinIdToType[id] + if ok { + if id == tInterface { + deb.interfaceValue(indent) + } else { + deb.printBuiltin(indent, id) + } + return + } + wire, ok := deb.wireType[id] + if !ok { + errorf("type id %d not defined", id) + } + switch { + case wire.ArrayT != nil: + deb.arrayValue(indent, wire) + case wire.MapT != nil: + deb.mapValue(indent, wire) + case wire.SliceT != nil: + deb.sliceValue(indent, wire) + case wire.StructT != nil: + deb.structValue(indent, id) + case wire.GobEncoderT != nil: + deb.gobEncoderValue(indent, id) + default: + panic("bad wire type for field") + } +} + +// printBuiltin prints a value not of a fundamental type, that is, +// one whose type is known to gobs at bootstrap time. +func (deb *debugger) printBuiltin(indent tab, id typeId) { + switch id { + case tBool: + x := deb.int64() + if x == 0 { + fmt.Fprintf(os.Stderr, "%sfalse\n", indent) + } else { + fmt.Fprintf(os.Stderr, "%strue\n", indent) + } + case tInt: + x := deb.int64() + fmt.Fprintf(os.Stderr, "%s%d\n", indent, x) + case tUint: + x := deb.uint64() + fmt.Fprintf(os.Stderr, "%s%d\n", indent, x) + case tFloat: + x := deb.uint64() + fmt.Fprintf(os.Stderr, "%s%g\n", indent, float64FromBits(x)) + case tComplex: + r := deb.uint64() + i := deb.uint64() + fmt.Fprintf(os.Stderr, "%s%g+%gi\n", indent, float64FromBits(r), float64FromBits(i)) + case tBytes: + x := int(deb.uint64()) + b := make([]byte, x) + deb.r.Read(b) + deb.consumed(x) + fmt.Fprintf(os.Stderr, "%s{% x}=%q\n", indent, b, b) + case tString: + x := int(deb.uint64()) + b := make([]byte, x) + deb.r.Read(b) + deb.consumed(x) + fmt.Fprintf(os.Stderr, "%s%q\n", indent, b) + default: + panic("unknown builtin") + } +} + +// ArrayValue: +// +// uint(n) FieldValue*n +func (deb *debugger) arrayValue(indent tab, wire *wireType) { + elemId := wire.ArrayT.Elem + u := deb.uint64() + length := int(u) + for i := 0; i < length; i++ { + deb.fieldValue(indent, elemId) + } + if length != wire.ArrayT.Len { + fmt.Fprintf(os.Stderr, "%s(wrong length for array: %d should be %d)\n", indent, length, wire.ArrayT.Len) + } +} + +// MapValue: +// +// uint(n) (FieldValue FieldValue)*n [n (key, value) pairs] +func (deb *debugger) mapValue(indent tab, wire *wireType) { + keyId := wire.MapT.Key + elemId := wire.MapT.Elem + u := deb.uint64() + length := int(u) + for i := 0; i < length; i++ { + deb.fieldValue(indent+1, keyId) + deb.fieldValue(indent+1, elemId) + } +} + +// SliceValue: +// +// uint(n) (n FieldValue) +func (deb *debugger) sliceValue(indent tab, wire *wireType) { + elemId := wire.SliceT.Elem + u := deb.uint64() + length := int(u) + deb.dump("Start of slice of length %d", length) + + for i := 0; i < length; i++ { + deb.fieldValue(indent, elemId) + } +} + +// StructValue: +// +// (uint(fieldDelta) FieldValue)* +func (deb *debugger) structValue(indent tab, id typeId) { + deb.dump("Start of struct value of %q id=%d\n<<\n", id.name(), id) + fmt.Fprintf(os.Stderr, "%s%s struct {\n", indent, id.name()) + wire, ok := deb.wireType[id] + if !ok { + errorf("type id %d not defined", id) + } + strct := wire.StructT + fieldNum := -1 + indent++ + for { + delta := deb.uint64() + if delta == 0 { // struct terminator is zero delta fieldnum + break + } + fieldNum += int(delta) + if fieldNum < 0 || fieldNum >= len(strct.Field) { + deb.dump("field number out of range: prevField=%d delta=%d", fieldNum-int(delta), delta) + break + } + fmt.Fprintf(os.Stderr, "%sfield %d:\t%s\n", indent, fieldNum, wire.StructT.Field[fieldNum].Name) + deb.fieldValue(indent+1, strct.Field[fieldNum].Id) + } + indent-- + fmt.Fprintf(os.Stderr, "%s} // end %s struct\n", indent, id.name()) + deb.dump(">> End of struct value of type %d %q", id, id.name()) +} + +// GobEncoderValue: +// +// uint(n) byte*n +func (deb *debugger) gobEncoderValue(indent tab, id typeId) { + len := deb.uint64() + deb.dump("GobEncoder value of %q id=%d, length %d\n", id.name(), id, len) + fmt.Fprintf(os.Stderr, "%s%s (implements GobEncoder)\n", indent, id.name()) + data := make([]byte, len) + _, err := deb.r.Read(data) + if err != nil { + errorf("gobEncoder data read: %s", err) + } + fmt.Fprintf(os.Stderr, "%s[% .2x]\n", indent+1, data) +} diff --git a/src/encoding/gob/dec_helpers.go b/src/encoding/gob/dec_helpers.go new file mode 100644 index 0000000..a09ac8f --- /dev/null +++ b/src/encoding/gob/dec_helpers.go @@ -0,0 +1,544 @@ +// Code generated by go run decgen.go -output dec_helpers.go; DO NOT EDIT. + +// Copyright 2014 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 gob + +import ( + "math" + "reflect" +) + +var decArrayHelper = map[reflect.Kind]decHelper{ + reflect.Bool: decBoolArray, + reflect.Complex64: decComplex64Array, + reflect.Complex128: decComplex128Array, + reflect.Float32: decFloat32Array, + reflect.Float64: decFloat64Array, + reflect.Int: decIntArray, + reflect.Int16: decInt16Array, + reflect.Int32: decInt32Array, + reflect.Int64: decInt64Array, + reflect.Int8: decInt8Array, + reflect.String: decStringArray, + reflect.Uint: decUintArray, + reflect.Uint16: decUint16Array, + reflect.Uint32: decUint32Array, + reflect.Uint64: decUint64Array, + reflect.Uintptr: decUintptrArray, +} + +var decSliceHelper = map[reflect.Kind]decHelper{ + reflect.Bool: decBoolSlice, + reflect.Complex64: decComplex64Slice, + reflect.Complex128: decComplex128Slice, + reflect.Float32: decFloat32Slice, + reflect.Float64: decFloat64Slice, + reflect.Int: decIntSlice, + reflect.Int16: decInt16Slice, + reflect.Int32: decInt32Slice, + reflect.Int64: decInt64Slice, + reflect.Int8: decInt8Slice, + reflect.String: decStringSlice, + reflect.Uint: decUintSlice, + reflect.Uint16: decUint16Slice, + reflect.Uint32: decUint32Slice, + reflect.Uint64: decUint64Slice, + reflect.Uintptr: decUintptrSlice, +} + +func decBoolArray(state *decoderState, v reflect.Value, length int, ovfl error) bool { + // Can only slice if it is addressable. + if !v.CanAddr() { + return false + } + return decBoolSlice(state, v.Slice(0, v.Len()), length, ovfl) +} + +func decBoolSlice(state *decoderState, v reflect.Value, length int, ovfl error) bool { + slice, ok := v.Interface().([]bool) + if !ok { + // It is kind bool but not type bool. TODO: We can handle this unsafely. + return false + } + for i := 0; i < length; i++ { + if state.b.Len() == 0 { + errorf("decoding bool array or slice: length exceeds input size (%d elements)", length) + } + if i >= len(slice) { + // This is a slice that we only partially allocated. + growSlice(v, &slice, length) + } + slice[i] = state.decodeUint() != 0 + } + return true +} + +func decComplex64Array(state *decoderState, v reflect.Value, length int, ovfl error) bool { + // Can only slice if it is addressable. + if !v.CanAddr() { + return false + } + return decComplex64Slice(state, v.Slice(0, v.Len()), length, ovfl) +} + +func decComplex64Slice(state *decoderState, v reflect.Value, length int, ovfl error) bool { + slice, ok := v.Interface().([]complex64) + if !ok { + // It is kind complex64 but not type complex64. TODO: We can handle this unsafely. + return false + } + for i := 0; i < length; i++ { + if state.b.Len() == 0 { + errorf("decoding complex64 array or slice: length exceeds input size (%d elements)", length) + } + if i >= len(slice) { + // This is a slice that we only partially allocated. + growSlice(v, &slice, length) + } + real := float32FromBits(state.decodeUint(), ovfl) + imag := float32FromBits(state.decodeUint(), ovfl) + slice[i] = complex(float32(real), float32(imag)) + } + return true +} + +func decComplex128Array(state *decoderState, v reflect.Value, length int, ovfl error) bool { + // Can only slice if it is addressable. + if !v.CanAddr() { + return false + } + return decComplex128Slice(state, v.Slice(0, v.Len()), length, ovfl) +} + +func decComplex128Slice(state *decoderState, v reflect.Value, length int, ovfl error) bool { + slice, ok := v.Interface().([]complex128) + if !ok { + // It is kind complex128 but not type complex128. TODO: We can handle this unsafely. + return false + } + for i := 0; i < length; i++ { + if state.b.Len() == 0 { + errorf("decoding complex128 array or slice: length exceeds input size (%d elements)", length) + } + if i >= len(slice) { + // This is a slice that we only partially allocated. + growSlice(v, &slice, length) + } + real := float64FromBits(state.decodeUint()) + imag := float64FromBits(state.decodeUint()) + slice[i] = complex(real, imag) + } + return true +} + +func decFloat32Array(state *decoderState, v reflect.Value, length int, ovfl error) bool { + // Can only slice if it is addressable. + if !v.CanAddr() { + return false + } + return decFloat32Slice(state, v.Slice(0, v.Len()), length, ovfl) +} + +func decFloat32Slice(state *decoderState, v reflect.Value, length int, ovfl error) bool { + slice, ok := v.Interface().([]float32) + if !ok { + // It is kind float32 but not type float32. TODO: We can handle this unsafely. + return false + } + for i := 0; i < length; i++ { + if state.b.Len() == 0 { + errorf("decoding float32 array or slice: length exceeds input size (%d elements)", length) + } + if i >= len(slice) { + // This is a slice that we only partially allocated. + growSlice(v, &slice, length) + } + slice[i] = float32(float32FromBits(state.decodeUint(), ovfl)) + } + return true +} + +func decFloat64Array(state *decoderState, v reflect.Value, length int, ovfl error) bool { + // Can only slice if it is addressable. + if !v.CanAddr() { + return false + } + return decFloat64Slice(state, v.Slice(0, v.Len()), length, ovfl) +} + +func decFloat64Slice(state *decoderState, v reflect.Value, length int, ovfl error) bool { + slice, ok := v.Interface().([]float64) + if !ok { + // It is kind float64 but not type float64. TODO: We can handle this unsafely. + return false + } + for i := 0; i < length; i++ { + if state.b.Len() == 0 { + errorf("decoding float64 array or slice: length exceeds input size (%d elements)", length) + } + if i >= len(slice) { + // This is a slice that we only partially allocated. + growSlice(v, &slice, length) + } + slice[i] = float64FromBits(state.decodeUint()) + } + return true +} + +func decIntArray(state *decoderState, v reflect.Value, length int, ovfl error) bool { + // Can only slice if it is addressable. + if !v.CanAddr() { + return false + } + return decIntSlice(state, v.Slice(0, v.Len()), length, ovfl) +} + +func decIntSlice(state *decoderState, v reflect.Value, length int, ovfl error) bool { + slice, ok := v.Interface().([]int) + if !ok { + // It is kind int but not type int. TODO: We can handle this unsafely. + return false + } + for i := 0; i < length; i++ { + if state.b.Len() == 0 { + errorf("decoding int array or slice: length exceeds input size (%d elements)", length) + } + if i >= len(slice) { + // This is a slice that we only partially allocated. + growSlice(v, &slice, length) + } + x := state.decodeInt() + // MinInt and MaxInt + if x < ^int64(^uint(0)>>1) || int64(^uint(0)>>1) < x { + error_(ovfl) + } + slice[i] = int(x) + } + return true +} + +func decInt16Array(state *decoderState, v reflect.Value, length int, ovfl error) bool { + // Can only slice if it is addressable. + if !v.CanAddr() { + return false + } + return decInt16Slice(state, v.Slice(0, v.Len()), length, ovfl) +} + +func decInt16Slice(state *decoderState, v reflect.Value, length int, ovfl error) bool { + slice, ok := v.Interface().([]int16) + if !ok { + // It is kind int16 but not type int16. TODO: We can handle this unsafely. + return false + } + for i := 0; i < length; i++ { + if state.b.Len() == 0 { + errorf("decoding int16 array or slice: length exceeds input size (%d elements)", length) + } + if i >= len(slice) { + // This is a slice that we only partially allocated. + growSlice(v, &slice, length) + } + x := state.decodeInt() + if x < math.MinInt16 || math.MaxInt16 < x { + error_(ovfl) + } + slice[i] = int16(x) + } + return true +} + +func decInt32Array(state *decoderState, v reflect.Value, length int, ovfl error) bool { + // Can only slice if it is addressable. + if !v.CanAddr() { + return false + } + return decInt32Slice(state, v.Slice(0, v.Len()), length, ovfl) +} + +func decInt32Slice(state *decoderState, v reflect.Value, length int, ovfl error) bool { + slice, ok := v.Interface().([]int32) + if !ok { + // It is kind int32 but not type int32. TODO: We can handle this unsafely. + return false + } + for i := 0; i < length; i++ { + if state.b.Len() == 0 { + errorf("decoding int32 array or slice: length exceeds input size (%d elements)", length) + } + if i >= len(slice) { + // This is a slice that we only partially allocated. + growSlice(v, &slice, length) + } + x := state.decodeInt() + if x < math.MinInt32 || math.MaxInt32 < x { + error_(ovfl) + } + slice[i] = int32(x) + } + return true +} + +func decInt64Array(state *decoderState, v reflect.Value, length int, ovfl error) bool { + // Can only slice if it is addressable. + if !v.CanAddr() { + return false + } + return decInt64Slice(state, v.Slice(0, v.Len()), length, ovfl) +} + +func decInt64Slice(state *decoderState, v reflect.Value, length int, ovfl error) bool { + slice, ok := v.Interface().([]int64) + if !ok { + // It is kind int64 but not type int64. TODO: We can handle this unsafely. + return false + } + for i := 0; i < length; i++ { + if state.b.Len() == 0 { + errorf("decoding int64 array or slice: length exceeds input size (%d elements)", length) + } + if i >= len(slice) { + // This is a slice that we only partially allocated. + growSlice(v, &slice, length) + } + slice[i] = state.decodeInt() + } + return true +} + +func decInt8Array(state *decoderState, v reflect.Value, length int, ovfl error) bool { + // Can only slice if it is addressable. + if !v.CanAddr() { + return false + } + return decInt8Slice(state, v.Slice(0, v.Len()), length, ovfl) +} + +func decInt8Slice(state *decoderState, v reflect.Value, length int, ovfl error) bool { + slice, ok := v.Interface().([]int8) + if !ok { + // It is kind int8 but not type int8. TODO: We can handle this unsafely. + return false + } + for i := 0; i < length; i++ { + if state.b.Len() == 0 { + errorf("decoding int8 array or slice: length exceeds input size (%d elements)", length) + } + if i >= len(slice) { + // This is a slice that we only partially allocated. + growSlice(v, &slice, length) + } + x := state.decodeInt() + if x < math.MinInt8 || math.MaxInt8 < x { + error_(ovfl) + } + slice[i] = int8(x) + } + return true +} + +func decStringArray(state *decoderState, v reflect.Value, length int, ovfl error) bool { + // Can only slice if it is addressable. + if !v.CanAddr() { + return false + } + return decStringSlice(state, v.Slice(0, v.Len()), length, ovfl) +} + +func decStringSlice(state *decoderState, v reflect.Value, length int, ovfl error) bool { + slice, ok := v.Interface().([]string) + if !ok { + // It is kind string but not type string. TODO: We can handle this unsafely. + return false + } + for i := 0; i < length; i++ { + if state.b.Len() == 0 { + errorf("decoding string array or slice: length exceeds input size (%d elements)", length) + } + u := state.decodeUint() + n := int(u) + if n < 0 || uint64(n) != u || n > state.b.Len() { + errorf("length of string exceeds input size (%d bytes)", u) + } + if n > state.b.Len() { + errorf("string data too long for buffer: %d", n) + } + // Read the data. + data := state.b.Bytes() + if len(data) < n { + errorf("invalid string length %d: exceeds input size %d", n, len(data)) + } + slice[i] = string(data[:n]) + state.b.Drop(n) + } + return true +} + +func decUintArray(state *decoderState, v reflect.Value, length int, ovfl error) bool { + // Can only slice if it is addressable. + if !v.CanAddr() { + return false + } + return decUintSlice(state, v.Slice(0, v.Len()), length, ovfl) +} + +func decUintSlice(state *decoderState, v reflect.Value, length int, ovfl error) bool { + slice, ok := v.Interface().([]uint) + if !ok { + // It is kind uint but not type uint. TODO: We can handle this unsafely. + return false + } + for i := 0; i < length; i++ { + if state.b.Len() == 0 { + errorf("decoding uint array or slice: length exceeds input size (%d elements)", length) + } + if i >= len(slice) { + // This is a slice that we only partially allocated. + growSlice(v, &slice, length) + } + x := state.decodeUint() + /*TODO if math.MaxUint32 < x { + error_(ovfl) + }*/ + slice[i] = uint(x) + } + return true +} + +func decUint16Array(state *decoderState, v reflect.Value, length int, ovfl error) bool { + // Can only slice if it is addressable. + if !v.CanAddr() { + return false + } + return decUint16Slice(state, v.Slice(0, v.Len()), length, ovfl) +} + +func decUint16Slice(state *decoderState, v reflect.Value, length int, ovfl error) bool { + slice, ok := v.Interface().([]uint16) + if !ok { + // It is kind uint16 but not type uint16. TODO: We can handle this unsafely. + return false + } + for i := 0; i < length; i++ { + if state.b.Len() == 0 { + errorf("decoding uint16 array or slice: length exceeds input size (%d elements)", length) + } + if i >= len(slice) { + // This is a slice that we only partially allocated. + growSlice(v, &slice, length) + } + x := state.decodeUint() + if math.MaxUint16 < x { + error_(ovfl) + } + slice[i] = uint16(x) + } + return true +} + +func decUint32Array(state *decoderState, v reflect.Value, length int, ovfl error) bool { + // Can only slice if it is addressable. + if !v.CanAddr() { + return false + } + return decUint32Slice(state, v.Slice(0, v.Len()), length, ovfl) +} + +func decUint32Slice(state *decoderState, v reflect.Value, length int, ovfl error) bool { + slice, ok := v.Interface().([]uint32) + if !ok { + // It is kind uint32 but not type uint32. TODO: We can handle this unsafely. + return false + } + for i := 0; i < length; i++ { + if state.b.Len() == 0 { + errorf("decoding uint32 array or slice: length exceeds input size (%d elements)", length) + } + if i >= len(slice) { + // This is a slice that we only partially allocated. + growSlice(v, &slice, length) + } + x := state.decodeUint() + if math.MaxUint32 < x { + error_(ovfl) + } + slice[i] = uint32(x) + } + return true +} + +func decUint64Array(state *decoderState, v reflect.Value, length int, ovfl error) bool { + // Can only slice if it is addressable. + if !v.CanAddr() { + return false + } + return decUint64Slice(state, v.Slice(0, v.Len()), length, ovfl) +} + +func decUint64Slice(state *decoderState, v reflect.Value, length int, ovfl error) bool { + slice, ok := v.Interface().([]uint64) + if !ok { + // It is kind uint64 but not type uint64. TODO: We can handle this unsafely. + return false + } + for i := 0; i < length; i++ { + if state.b.Len() == 0 { + errorf("decoding uint64 array or slice: length exceeds input size (%d elements)", length) + } + if i >= len(slice) { + // This is a slice that we only partially allocated. + growSlice(v, &slice, length) + } + slice[i] = state.decodeUint() + } + return true +} + +func decUintptrArray(state *decoderState, v reflect.Value, length int, ovfl error) bool { + // Can only slice if it is addressable. + if !v.CanAddr() { + return false + } + return decUintptrSlice(state, v.Slice(0, v.Len()), length, ovfl) +} + +func decUintptrSlice(state *decoderState, v reflect.Value, length int, ovfl error) bool { + slice, ok := v.Interface().([]uintptr) + if !ok { + // It is kind uintptr but not type uintptr. TODO: We can handle this unsafely. + return false + } + for i := 0; i < length; i++ { + if state.b.Len() == 0 { + errorf("decoding uintptr array or slice: length exceeds input size (%d elements)", length) + } + if i >= len(slice) { + // This is a slice that we only partially allocated. + growSlice(v, &slice, length) + } + x := state.decodeUint() + if uint64(^uintptr(0)) < x { + error_(ovfl) + } + slice[i] = uintptr(x) + } + return true +} + +// growSlice is called for a slice that we only partially allocated, +// to grow it up to length. +func growSlice[E any](v reflect.Value, ps *[]E, length int) { + var zero E + s := *ps + s = append(s, zero) + cp := cap(s) + if cp > length { + cp = length + } + s = s[:cp] + v.Set(reflect.ValueOf(s)) + *ps = s +} diff --git a/src/encoding/gob/decgen.go b/src/encoding/gob/decgen.go new file mode 100644 index 0000000..e40816e --- /dev/null +++ b/src/encoding/gob/decgen.go @@ -0,0 +1,243 @@ +// 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. + +//go:build ignore + +// encgen writes the helper functions for encoding. Intended to be +// used with go generate; see the invocation in encode.go. + +// TODO: We could do more by being unsafe. Add a -unsafe flag? + +package main + +import ( + "bytes" + "flag" + "fmt" + "go/format" + "log" + "os" +) + +var output = flag.String("output", "dec_helpers.go", "file name to write") + +type Type struct { + lower string + upper string + decoder string +} + +var types = []Type{ + { + "bool", + "Bool", + `slice[i] = state.decodeUint() != 0`, + }, + { + "complex64", + "Complex64", + `real := float32FromBits(state.decodeUint(), ovfl) + imag := float32FromBits(state.decodeUint(), ovfl) + slice[i] = complex(float32(real), float32(imag))`, + }, + { + "complex128", + "Complex128", + `real := float64FromBits(state.decodeUint()) + imag := float64FromBits(state.decodeUint()) + slice[i] = complex(real, imag)`, + }, + { + "float32", + "Float32", + `slice[i] = float32(float32FromBits(state.decodeUint(), ovfl))`, + }, + { + "float64", + "Float64", + `slice[i] = float64FromBits(state.decodeUint())`, + }, + { + "int", + "Int", + `x := state.decodeInt() + // MinInt and MaxInt + if x < ^int64(^uint(0)>>1) || int64(^uint(0)>>1) < x { + error_(ovfl) + } + slice[i] = int(x)`, + }, + { + "int16", + "Int16", + `x := state.decodeInt() + if x < math.MinInt16 || math.MaxInt16 < x { + error_(ovfl) + } + slice[i] = int16(x)`, + }, + { + "int32", + "Int32", + `x := state.decodeInt() + if x < math.MinInt32 || math.MaxInt32 < x { + error_(ovfl) + } + slice[i] = int32(x)`, + }, + { + "int64", + "Int64", + `slice[i] = state.decodeInt()`, + }, + { + "int8", + "Int8", + `x := state.decodeInt() + if x < math.MinInt8 || math.MaxInt8 < x { + error_(ovfl) + } + slice[i] = int8(x)`, + }, + { + "string", + "String", + `u := state.decodeUint() + n := int(u) + if n < 0 || uint64(n) != u || n > state.b.Len() { + errorf("length of string exceeds input size (%d bytes)", u) + } + if n > state.b.Len() { + errorf("string data too long for buffer: %d", n) + } + // Read the data. + data := state.b.Bytes() + if len(data) < n { + errorf("invalid string length %d: exceeds input size %d", n, len(data)) + } + slice[i] = string(data[:n]) + state.b.Drop(n)`, + }, + { + "uint", + "Uint", + `x := state.decodeUint() + /*TODO if math.MaxUint32 < x { + error_(ovfl) + }*/ + slice[i] = uint(x)`, + }, + { + "uint16", + "Uint16", + `x := state.decodeUint() + if math.MaxUint16 < x { + error_(ovfl) + } + slice[i] = uint16(x)`, + }, + { + "uint32", + "Uint32", + `x := state.decodeUint() + if math.MaxUint32 < x { + error_(ovfl) + } + slice[i] = uint32(x)`, + }, + { + "uint64", + "Uint64", + `slice[i] = state.decodeUint()`, + }, + { + "uintptr", + "Uintptr", + `x := state.decodeUint() + if uint64(^uintptr(0)) < x { + error_(ovfl) + } + slice[i] = uintptr(x)`, + }, + // uint8 Handled separately. +} + +func main() { + log.SetFlags(0) + log.SetPrefix("decgen: ") + flag.Parse() + if flag.NArg() != 0 { + log.Fatal("usage: decgen [--output filename]") + } + var b bytes.Buffer + fmt.Fprintf(&b, "// Code generated by go run decgen.go -output %s; DO NOT EDIT.\n", *output) + fmt.Fprint(&b, header) + printMaps(&b, "Array") + fmt.Fprint(&b, "\n") + printMaps(&b, "Slice") + for _, t := range types { + fmt.Fprintf(&b, arrayHelper, t.lower, t.upper) + fmt.Fprintf(&b, sliceHelper, t.lower, t.upper, t.decoder) + } + source, err := format.Source(b.Bytes()) + if err != nil { + log.Fatal("source format error:", err) + } + fd, err := os.Create(*output) + if err != nil { + log.Fatal(err) + } + if _, err := fd.Write(source); err != nil { + log.Fatal(err) + } +} + +func printMaps(b *bytes.Buffer, upperClass string) { + fmt.Fprintf(b, "var dec%sHelper = map[reflect.Kind]decHelper{\n", upperClass) + for _, t := range types { + fmt.Fprintf(b, "reflect.%s: dec%s%s,\n", t.upper, t.upper, upperClass) + } + fmt.Fprintf(b, "}\n") +} + +const header = ` +// Copyright 2014 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 gob + +import ( + "math" + "reflect" +) + +` + +const arrayHelper = ` +func dec%[2]sArray(state *decoderState, v reflect.Value, length int, ovfl error) bool { + // Can only slice if it is addressable. + if !v.CanAddr() { + return false + } + return dec%[2]sSlice(state, v.Slice(0, v.Len()), length, ovfl) +} +` + +const sliceHelper = ` +func dec%[2]sSlice(state *decoderState, v reflect.Value, length int, ovfl error) bool { + slice, ok := v.Interface().([]%[1]s) + if !ok { + // It is kind %[1]s but not type %[1]s. TODO: We can handle this unsafely. + return false + } + for i := 0; i < length; i++ { + if state.b.Len() == 0 { + errorf("decoding %[1]s array or slice: length exceeds input size (%%d elements)", length) + } + %[3]s + } + return true +} +` diff --git a/src/encoding/gob/decode.go b/src/encoding/gob/decode.go new file mode 100644 index 0000000..f46a391 --- /dev/null +++ b/src/encoding/gob/decode.go @@ -0,0 +1,1308 @@ +// 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. + +//go:generate go run decgen.go -output dec_helpers.go + +package gob + +import ( + "encoding" + "errors" + "internal/saferio" + "io" + "math" + "math/bits" + "reflect" +) + +var ( + errBadUint = errors.New("gob: encoded unsigned integer out of range") + errBadType = errors.New("gob: unknown type id or corrupted data") + errRange = errors.New("gob: bad data: field numbers out of bounds") +) + +type decHelper func(state *decoderState, v reflect.Value, length int, ovfl error) bool + +// decoderState is the execution state of an instance of the decoder. A new state +// is created for nested objects. +type decoderState struct { + dec *Decoder + // The buffer is stored with an extra indirection because it may be replaced + // if we load a type during decode (when reading an interface value). + b *decBuffer + fieldnum int // the last field number read. + next *decoderState // for free list +} + +// decBuffer is an extremely simple, fast implementation of a read-only byte buffer. +// It is initialized by calling Size and then copying the data into the slice returned by Bytes(). +type decBuffer struct { + data []byte + offset int // Read offset. +} + +func (d *decBuffer) Read(p []byte) (int, error) { + n := copy(p, d.data[d.offset:]) + if n == 0 && len(p) != 0 { + return 0, io.EOF + } + d.offset += n + return n, nil +} + +func (d *decBuffer) Drop(n int) { + if n > d.Len() { + panic("drop") + } + d.offset += n +} + +func (d *decBuffer) ReadByte() (byte, error) { + if d.offset >= len(d.data) { + return 0, io.EOF + } + c := d.data[d.offset] + d.offset++ + return c, nil +} + +func (d *decBuffer) Len() int { + return len(d.data) - d.offset +} + +func (d *decBuffer) Bytes() []byte { + return d.data[d.offset:] +} + +// SetBytes sets the buffer to the bytes, discarding any existing data. +func (d *decBuffer) SetBytes(data []byte) { + d.data = data + d.offset = 0 +} + +func (d *decBuffer) Reset() { + d.data = d.data[0:0] + d.offset = 0 +} + +// We pass the bytes.Buffer separately for easier testing of the infrastructure +// without requiring a full Decoder. +func (dec *Decoder) newDecoderState(buf *decBuffer) *decoderState { + d := dec.freeList + if d == nil { + d = new(decoderState) + d.dec = dec + } else { + dec.freeList = d.next + } + d.b = buf + return d +} + +func (dec *Decoder) freeDecoderState(d *decoderState) { + d.next = dec.freeList + dec.freeList = d +} + +func overflow(name string) error { + return errors.New(`value for "` + name + `" out of range`) +} + +// decodeUintReader reads an encoded unsigned integer from an io.Reader. +// Used only by the Decoder to read the message length. +func decodeUintReader(r io.Reader, buf []byte) (x uint64, width int, err error) { + width = 1 + n, err := io.ReadFull(r, buf[0:width]) + if n == 0 { + return + } + b := buf[0] + if b <= 0x7f { + return uint64(b), width, nil + } + n = -int(int8(b)) + if n > uint64Size { + err = errBadUint + return + } + width, err = io.ReadFull(r, buf[0:n]) + if err != nil { + if err == io.EOF { + err = io.ErrUnexpectedEOF + } + return + } + // Could check that the high byte is zero but it's not worth it. + for _, b := range buf[0:width] { + x = x<<8 | uint64(b) + } + width++ // +1 for length byte + return +} + +// decodeUint reads an encoded unsigned integer from state.r. +// Does not check for overflow. +func (state *decoderState) decodeUint() (x uint64) { + b, err := state.b.ReadByte() + if err != nil { + error_(err) + } + if b <= 0x7f { + return uint64(b) + } + n := -int(int8(b)) + if n > uint64Size { + error_(errBadUint) + } + buf := state.b.Bytes() + if len(buf) < n { + errorf("invalid uint data length %d: exceeds input size %d", n, len(buf)) + } + // Don't need to check error; it's safe to loop regardless. + // Could check that the high byte is zero but it's not worth it. + for _, b := range buf[0:n] { + x = x<<8 | uint64(b) + } + state.b.Drop(n) + return x +} + +// decodeInt reads an encoded signed integer from state.r. +// Does not check for overflow. +func (state *decoderState) decodeInt() int64 { + x := state.decodeUint() + if x&1 != 0 { + return ^int64(x >> 1) + } + return int64(x >> 1) +} + +// getLength decodes the next uint and makes sure it is a possible +// size for a data item that follows, which means it must fit in a +// non-negative int and fit in the buffer. +func (state *decoderState) getLength() (int, bool) { + n := int(state.decodeUint()) + if n < 0 || state.b.Len() < n || tooBig <= n { + return 0, false + } + return n, true +} + +// decOp is the signature of a decoding operator for a given type. +type decOp func(i *decInstr, state *decoderState, v reflect.Value) + +// The 'instructions' of the decoding machine +type decInstr struct { + op decOp + field int // field number of the wire type + index []int // field access indices for destination type + ovfl error // error message for overflow/underflow (for arrays, of the elements) +} + +// ignoreUint discards a uint value with no destination. +func ignoreUint(i *decInstr, state *decoderState, v reflect.Value) { + state.decodeUint() +} + +// ignoreTwoUints discards a uint value with no destination. It's used to skip +// complex values. +func ignoreTwoUints(i *decInstr, state *decoderState, v reflect.Value) { + state.decodeUint() + state.decodeUint() +} + +// Since the encoder writes no zeros, if we arrive at a decoder we have +// a value to extract and store. The field number has already been read +// (it's how we knew to call this decoder). +// Each decoder is responsible for handling any indirections associated +// with the data structure. If any pointer so reached is nil, allocation must +// be done. + +// decAlloc takes a value and returns a settable value that can +// be assigned to. If the value is a pointer, decAlloc guarantees it points to storage. +// The callers to the individual decoders are expected to have used decAlloc. +// The individual decoders don't need to it. +func decAlloc(v reflect.Value) reflect.Value { + for v.Kind() == reflect.Pointer { + if v.IsNil() { + v.Set(reflect.New(v.Type().Elem())) + } + v = v.Elem() + } + return v +} + +// decBool decodes a uint and stores it as a boolean in value. +func decBool(i *decInstr, state *decoderState, value reflect.Value) { + value.SetBool(state.decodeUint() != 0) +} + +// decInt8 decodes an integer and stores it as an int8 in value. +func decInt8(i *decInstr, state *decoderState, value reflect.Value) { + v := state.decodeInt() + if v < math.MinInt8 || math.MaxInt8 < v { + error_(i.ovfl) + } + value.SetInt(v) +} + +// decUint8 decodes an unsigned integer and stores it as a uint8 in value. +func decUint8(i *decInstr, state *decoderState, value reflect.Value) { + v := state.decodeUint() + if math.MaxUint8 < v { + error_(i.ovfl) + } + value.SetUint(v) +} + +// decInt16 decodes an integer and stores it as an int16 in value. +func decInt16(i *decInstr, state *decoderState, value reflect.Value) { + v := state.decodeInt() + if v < math.MinInt16 || math.MaxInt16 < v { + error_(i.ovfl) + } + value.SetInt(v) +} + +// decUint16 decodes an unsigned integer and stores it as a uint16 in value. +func decUint16(i *decInstr, state *decoderState, value reflect.Value) { + v := state.decodeUint() + if math.MaxUint16 < v { + error_(i.ovfl) + } + value.SetUint(v) +} + +// decInt32 decodes an integer and stores it as an int32 in value. +func decInt32(i *decInstr, state *decoderState, value reflect.Value) { + v := state.decodeInt() + if v < math.MinInt32 || math.MaxInt32 < v { + error_(i.ovfl) + } + value.SetInt(v) +} + +// decUint32 decodes an unsigned integer and stores it as a uint32 in value. +func decUint32(i *decInstr, state *decoderState, value reflect.Value) { + v := state.decodeUint() + if math.MaxUint32 < v { + error_(i.ovfl) + } + value.SetUint(v) +} + +// decInt64 decodes an integer and stores it as an int64 in value. +func decInt64(i *decInstr, state *decoderState, value reflect.Value) { + v := state.decodeInt() + value.SetInt(v) +} + +// decUint64 decodes an unsigned integer and stores it as a uint64 in value. +func decUint64(i *decInstr, state *decoderState, value reflect.Value) { + v := state.decodeUint() + value.SetUint(v) +} + +// Floating-point numbers are transmitted as uint64s holding the bits +// of the underlying representation. They are sent byte-reversed, with +// the exponent end coming out first, so integer floating point numbers +// (for example) transmit more compactly. This routine does the +// unswizzling. +func float64FromBits(u uint64) float64 { + v := bits.ReverseBytes64(u) + return math.Float64frombits(v) +} + +// float32FromBits decodes an unsigned integer, treats it as a 32-bit floating-point +// number, and returns it. It's a helper function for float32 and complex64. +// It returns a float64 because that's what reflection needs, but its return +// value is known to be accurately representable in a float32. +func float32FromBits(u uint64, ovfl error) float64 { + v := float64FromBits(u) + av := v + if av < 0 { + av = -av + } + // +Inf is OK in both 32- and 64-bit floats. Underflow is always OK. + if math.MaxFloat32 < av && av <= math.MaxFloat64 { + error_(ovfl) + } + return v +} + +// decFloat32 decodes an unsigned integer, treats it as a 32-bit floating-point +// number, and stores it in value. +func decFloat32(i *decInstr, state *decoderState, value reflect.Value) { + value.SetFloat(float32FromBits(state.decodeUint(), i.ovfl)) +} + +// decFloat64 decodes an unsigned integer, treats it as a 64-bit floating-point +// number, and stores it in value. +func decFloat64(i *decInstr, state *decoderState, value reflect.Value) { + value.SetFloat(float64FromBits(state.decodeUint())) +} + +// decComplex64 decodes a pair of unsigned integers, treats them as a +// pair of floating point numbers, and stores them as a complex64 in value. +// The real part comes first. +func decComplex64(i *decInstr, state *decoderState, value reflect.Value) { + real := float32FromBits(state.decodeUint(), i.ovfl) + imag := float32FromBits(state.decodeUint(), i.ovfl) + value.SetComplex(complex(real, imag)) +} + +// decComplex128 decodes a pair of unsigned integers, treats them as a +// pair of floating point numbers, and stores them as a complex128 in value. +// The real part comes first. +func decComplex128(i *decInstr, state *decoderState, value reflect.Value) { + real := float64FromBits(state.decodeUint()) + imag := float64FromBits(state.decodeUint()) + value.SetComplex(complex(real, imag)) +} + +// decUint8Slice decodes a byte slice and stores in value a slice header +// describing the data. +// uint8 slices are encoded as an unsigned count followed by the raw bytes. +func decUint8Slice(i *decInstr, state *decoderState, value reflect.Value) { + n, ok := state.getLength() + if !ok { + errorf("bad %s slice length: %d", value.Type(), n) + } + if value.Cap() < n { + safe := saferio.SliceCap((*byte)(nil), uint64(n)) + if safe < 0 { + errorf("%s slice too big: %d elements", value.Type(), n) + } + value.Set(reflect.MakeSlice(value.Type(), safe, safe)) + ln := safe + i := 0 + for i < n { + if i >= ln { + // We didn't allocate the entire slice, + // due to using saferio.SliceCap. + // Append a value to grow the slice. + // The slice is full, so this should + // bump up the capacity. + value.Set(reflect.Append(value, reflect.Zero(value.Type().Elem()))) + } + // Copy into s up to the capacity or n, + // whichever is less. + ln = value.Cap() + if ln > n { + ln = n + } + value.SetLen(ln) + sub := value.Slice(i, ln) + if _, err := state.b.Read(sub.Bytes()); err != nil { + errorf("error decoding []byte at %d: %s", err, i) + } + i = ln + } + } else { + value.SetLen(n) + if _, err := state.b.Read(value.Bytes()); err != nil { + errorf("error decoding []byte: %s", err) + } + } +} + +// decString decodes byte array and stores in value a string header +// describing the data. +// Strings are encoded as an unsigned count followed by the raw bytes. +func decString(i *decInstr, state *decoderState, value reflect.Value) { + n, ok := state.getLength() + if !ok { + errorf("bad %s slice length: %d", value.Type(), n) + } + // Read the data. + data := state.b.Bytes() + if len(data) < n { + errorf("invalid string length %d: exceeds input size %d", n, len(data)) + } + s := string(data[:n]) + state.b.Drop(n) + value.SetString(s) +} + +// ignoreUint8Array skips over the data for a byte slice value with no destination. +func ignoreUint8Array(i *decInstr, state *decoderState, value reflect.Value) { + n, ok := state.getLength() + if !ok { + errorf("slice length too large") + } + bn := state.b.Len() + if bn < n { + errorf("invalid slice length %d: exceeds input size %d", n, bn) + } + state.b.Drop(n) +} + +// Execution engine + +// The encoder engine is an array of instructions indexed by field number of the incoming +// decoder. It is executed with random access according to field number. +type decEngine struct { + instr []decInstr + numInstr int // the number of active instructions +} + +// decodeSingle decodes a top-level value that is not a struct and stores it in value. +// Such values are preceded by a zero, making them have the memory layout of a +// struct field (although with an illegal field number). +func (dec *Decoder) decodeSingle(engine *decEngine, value reflect.Value) { + state := dec.newDecoderState(&dec.buf) + defer dec.freeDecoderState(state) + state.fieldnum = singletonField + if state.decodeUint() != 0 { + errorf("decode: corrupted data: non-zero delta for singleton") + } + instr := &engine.instr[singletonField] + instr.op(instr, state, value) +} + +// decodeStruct decodes a top-level struct and stores it in value. +// Indir is for the value, not the type. At the time of the call it may +// differ from ut.indir, which was computed when the engine was built. +// This state cannot arise for decodeSingle, which is called directly +// from the user's value, not from the innards of an engine. +func (dec *Decoder) decodeStruct(engine *decEngine, value reflect.Value) { + state := dec.newDecoderState(&dec.buf) + defer dec.freeDecoderState(state) + state.fieldnum = -1 + for state.b.Len() > 0 { + delta := int(state.decodeUint()) + if delta < 0 { + errorf("decode: corrupted data: negative delta") + } + if delta == 0 { // struct terminator is zero delta fieldnum + break + } + if state.fieldnum >= len(engine.instr)-delta { // subtract to compare without overflow + error_(errRange) + } + fieldnum := state.fieldnum + delta + instr := &engine.instr[fieldnum] + var field reflect.Value + if instr.index != nil { + // Otherwise the field is unknown to us and instr.op is an ignore op. + field = value.FieldByIndex(instr.index) + if field.Kind() == reflect.Pointer { + field = decAlloc(field) + } + } + instr.op(instr, state, field) + state.fieldnum = fieldnum + } +} + +var noValue reflect.Value + +// ignoreStruct discards the data for a struct with no destination. +func (dec *Decoder) ignoreStruct(engine *decEngine) { + state := dec.newDecoderState(&dec.buf) + defer dec.freeDecoderState(state) + state.fieldnum = -1 + for state.b.Len() > 0 { + delta := int(state.decodeUint()) + if delta < 0 { + errorf("ignore decode: corrupted data: negative delta") + } + if delta == 0 { // struct terminator is zero delta fieldnum + break + } + fieldnum := state.fieldnum + delta + if fieldnum >= len(engine.instr) { + error_(errRange) + } + instr := &engine.instr[fieldnum] + instr.op(instr, state, noValue) + state.fieldnum = fieldnum + } +} + +// ignoreSingle discards the data for a top-level non-struct value with no +// destination. It's used when calling Decode with a nil value. +func (dec *Decoder) ignoreSingle(engine *decEngine) { + state := dec.newDecoderState(&dec.buf) + defer dec.freeDecoderState(state) + state.fieldnum = singletonField + delta := int(state.decodeUint()) + if delta != 0 { + errorf("decode: corrupted data: non-zero delta for singleton") + } + instr := &engine.instr[singletonField] + instr.op(instr, state, noValue) +} + +// decodeArrayHelper does the work for decoding arrays and slices. +func (dec *Decoder) decodeArrayHelper(state *decoderState, value reflect.Value, elemOp decOp, length int, ovfl error, helper decHelper) { + if helper != nil && helper(state, value, length, ovfl) { + return + } + instr := &decInstr{elemOp, 0, nil, ovfl} + isPtr := value.Type().Elem().Kind() == reflect.Pointer + ln := value.Len() + for i := 0; i < length; i++ { + if state.b.Len() == 0 { + errorf("decoding array or slice: length exceeds input size (%d elements)", length) + } + if i >= ln { + // This is a slice that we only partially allocated. + // Grow it using append, up to length. + value.Set(reflect.Append(value, reflect.Zero(value.Type().Elem()))) + cp := value.Cap() + if cp > length { + cp = length + } + value.SetLen(cp) + ln = cp + } + v := value.Index(i) + if isPtr { + v = decAlloc(v) + } + elemOp(instr, state, v) + } +} + +// decodeArray decodes an array and stores it in value. +// The length is an unsigned integer preceding the elements. Even though the length is redundant +// (it's part of the type), it's a useful check and is included in the encoding. +func (dec *Decoder) decodeArray(state *decoderState, value reflect.Value, elemOp decOp, length int, ovfl error, helper decHelper) { + if n := state.decodeUint(); n != uint64(length) { + errorf("length mismatch in decodeArray") + } + dec.decodeArrayHelper(state, value, elemOp, length, ovfl, helper) +} + +// decodeIntoValue is a helper for map decoding. +func decodeIntoValue(state *decoderState, op decOp, isPtr bool, value reflect.Value, instr *decInstr) reflect.Value { + v := value + if isPtr { + v = decAlloc(value) + } + + op(instr, state, v) + return value +} + +// decodeMap decodes a map and stores it in value. +// Maps are encoded as a length followed by key:value pairs. +// Because the internals of maps are not visible to us, we must +// use reflection rather than pointer magic. +func (dec *Decoder) decodeMap(mtyp reflect.Type, state *decoderState, value reflect.Value, keyOp, elemOp decOp, ovfl error) { + n := int(state.decodeUint()) + if value.IsNil() { + value.Set(reflect.MakeMapWithSize(mtyp, n)) + } + keyIsPtr := mtyp.Key().Kind() == reflect.Pointer + elemIsPtr := mtyp.Elem().Kind() == reflect.Pointer + keyInstr := &decInstr{keyOp, 0, nil, ovfl} + elemInstr := &decInstr{elemOp, 0, nil, ovfl} + keyP := reflect.New(mtyp.Key()) + keyZ := reflect.Zero(mtyp.Key()) + elemP := reflect.New(mtyp.Elem()) + elemZ := reflect.Zero(mtyp.Elem()) + for i := 0; i < n; i++ { + key := decodeIntoValue(state, keyOp, keyIsPtr, keyP.Elem(), keyInstr) + elem := decodeIntoValue(state, elemOp, elemIsPtr, elemP.Elem(), elemInstr) + value.SetMapIndex(key, elem) + keyP.Elem().Set(keyZ) + elemP.Elem().Set(elemZ) + } +} + +// ignoreArrayHelper does the work for discarding arrays and slices. +func (dec *Decoder) ignoreArrayHelper(state *decoderState, elemOp decOp, length int) { + instr := &decInstr{elemOp, 0, nil, errors.New("no error")} + for i := 0; i < length; i++ { + if state.b.Len() == 0 { + errorf("decoding array or slice: length exceeds input size (%d elements)", length) + } + elemOp(instr, state, noValue) + } +} + +// ignoreArray discards the data for an array value with no destination. +func (dec *Decoder) ignoreArray(state *decoderState, elemOp decOp, length int) { + if n := state.decodeUint(); n != uint64(length) { + errorf("length mismatch in ignoreArray") + } + dec.ignoreArrayHelper(state, elemOp, length) +} + +// ignoreMap discards the data for a map value with no destination. +func (dec *Decoder) ignoreMap(state *decoderState, keyOp, elemOp decOp) { + n := int(state.decodeUint()) + keyInstr := &decInstr{keyOp, 0, nil, errors.New("no error")} + elemInstr := &decInstr{elemOp, 0, nil, errors.New("no error")} + for i := 0; i < n; i++ { + keyOp(keyInstr, state, noValue) + elemOp(elemInstr, state, noValue) + } +} + +// decodeSlice decodes a slice and stores it in value. +// Slices are encoded as an unsigned length followed by the elements. +func (dec *Decoder) decodeSlice(state *decoderState, value reflect.Value, elemOp decOp, ovfl error, helper decHelper) { + u := state.decodeUint() + typ := value.Type() + size := uint64(typ.Elem().Size()) + nBytes := u * size + n := int(u) + // Take care with overflow in this calculation. + if n < 0 || uint64(n) != u || nBytes > tooBig || (size > 0 && nBytes/size != u) { + // We don't check n against buffer length here because if it's a slice + // of interfaces, there will be buffer reloads. + errorf("%s slice too big: %d elements of %d bytes", typ.Elem(), u, size) + } + if value.Cap() < n { + safe := saferio.SliceCap(reflect.Zero(reflect.PtrTo(typ.Elem())).Interface(), uint64(n)) + if safe < 0 { + errorf("%s slice too big: %d elements of %d bytes", typ.Elem(), u, size) + } + value.Set(reflect.MakeSlice(typ, safe, safe)) + } else { + value.SetLen(n) + } + dec.decodeArrayHelper(state, value, elemOp, n, ovfl, helper) +} + +// ignoreSlice skips over the data for a slice value with no destination. +func (dec *Decoder) ignoreSlice(state *decoderState, elemOp decOp) { + dec.ignoreArrayHelper(state, elemOp, int(state.decodeUint())) +} + +// decodeInterface decodes an interface value and stores it in value. +// Interfaces are encoded as the name of a concrete type followed by a value. +// If the name is empty, the value is nil and no value is sent. +func (dec *Decoder) decodeInterface(ityp reflect.Type, state *decoderState, value reflect.Value) { + // Read the name of the concrete type. + nr := state.decodeUint() + if nr > 1<<31 { // zero is permissible for anonymous types + errorf("invalid type name length %d", nr) + } + if nr > uint64(state.b.Len()) { + errorf("invalid type name length %d: exceeds input size", nr) + } + n := int(nr) + name := state.b.Bytes()[:n] + state.b.Drop(n) + // Allocate the destination interface value. + if len(name) == 0 { + // Copy the nil interface value to the target. + value.Set(reflect.Zero(value.Type())) + return + } + if len(name) > 1024 { + errorf("name too long (%d bytes): %.20q...", len(name), name) + } + // The concrete type must be registered. + typi, ok := nameToConcreteType.Load(string(name)) + if !ok { + errorf("name not registered for interface: %q", name) + } + typ := typi.(reflect.Type) + + // Read the type id of the concrete value. + concreteId := dec.decodeTypeSequence(true) + if concreteId < 0 { + error_(dec.err) + } + // Byte count of value is next; we don't care what it is (it's there + // in case we want to ignore the value by skipping it completely). + state.decodeUint() + // Read the concrete value. + v := allocValue(typ) + dec.decodeValue(concreteId, v) + if dec.err != nil { + error_(dec.err) + } + // Assign the concrete value to the interface. + // Tread carefully; it might not satisfy the interface. + if !typ.AssignableTo(ityp) { + errorf("%s is not assignable to type %s", typ, ityp) + } + // Copy the interface value to the target. + value.Set(v) +} + +// ignoreInterface discards the data for an interface value with no destination. +func (dec *Decoder) ignoreInterface(state *decoderState) { + // Read the name of the concrete type. + n, ok := state.getLength() + if !ok { + errorf("bad interface encoding: name too large for buffer") + } + bn := state.b.Len() + if bn < n { + errorf("invalid interface value length %d: exceeds input size %d", n, bn) + } + state.b.Drop(n) + id := dec.decodeTypeSequence(true) + if id < 0 { + error_(dec.err) + } + // At this point, the decoder buffer contains a delimited value. Just toss it. + n, ok = state.getLength() + if !ok { + errorf("bad interface encoding: data length too large for buffer") + } + state.b.Drop(n) +} + +// decodeGobDecoder decodes something implementing the GobDecoder interface. +// The data is encoded as a byte slice. +func (dec *Decoder) decodeGobDecoder(ut *userTypeInfo, state *decoderState, value reflect.Value) { + // Read the bytes for the value. + n, ok := state.getLength() + if !ok { + errorf("GobDecoder: length too large for buffer") + } + b := state.b.Bytes() + if len(b) < n { + errorf("GobDecoder: invalid data length %d: exceeds input size %d", n, len(b)) + } + b = b[:n] + state.b.Drop(n) + var err error + // We know it's one of these. + switch ut.externalDec { + case xGob: + err = value.Interface().(GobDecoder).GobDecode(b) + case xBinary: + err = value.Interface().(encoding.BinaryUnmarshaler).UnmarshalBinary(b) + case xText: + err = value.Interface().(encoding.TextUnmarshaler).UnmarshalText(b) + } + if err != nil { + error_(err) + } +} + +// ignoreGobDecoder discards the data for a GobDecoder value with no destination. +func (dec *Decoder) ignoreGobDecoder(state *decoderState) { + // Read the bytes for the value. + n, ok := state.getLength() + if !ok { + errorf("GobDecoder: length too large for buffer") + } + bn := state.b.Len() + if bn < n { + errorf("GobDecoder: invalid data length %d: exceeds input size %d", n, bn) + } + state.b.Drop(n) +} + +// Index by Go types. +var decOpTable = [...]decOp{ + reflect.Bool: decBool, + reflect.Int8: decInt8, + reflect.Int16: decInt16, + reflect.Int32: decInt32, + reflect.Int64: decInt64, + reflect.Uint8: decUint8, + reflect.Uint16: decUint16, + reflect.Uint32: decUint32, + reflect.Uint64: decUint64, + reflect.Float32: decFloat32, + reflect.Float64: decFloat64, + reflect.Complex64: decComplex64, + reflect.Complex128: decComplex128, + reflect.String: decString, +} + +// Indexed by gob types. tComplex will be added during type.init(). +var decIgnoreOpMap = map[typeId]decOp{ + tBool: ignoreUint, + tInt: ignoreUint, + tUint: ignoreUint, + tFloat: ignoreUint, + tBytes: ignoreUint8Array, + tString: ignoreUint8Array, + tComplex: ignoreTwoUints, +} + +// decOpFor returns the decoding op for the base type under rt and +// the indirection count to reach it. +func (dec *Decoder) decOpFor(wireId typeId, rt reflect.Type, name string, inProgress map[reflect.Type]*decOp) *decOp { + ut := userType(rt) + // If the type implements GobEncoder, we handle it without further processing. + if ut.externalDec != 0 { + return dec.gobDecodeOpFor(ut) + } + + // If this type is already in progress, it's a recursive type (e.g. map[string]*T). + // Return the pointer to the op we're already building. + if opPtr := inProgress[rt]; opPtr != nil { + return opPtr + } + typ := ut.base + var op decOp + k := typ.Kind() + if int(k) < len(decOpTable) { + op = decOpTable[k] + } + if op == nil { + inProgress[rt] = &op + // Special cases + switch t := typ; t.Kind() { + case reflect.Array: + name = "element of " + name + elemId := dec.wireType[wireId].ArrayT.Elem + elemOp := dec.decOpFor(elemId, t.Elem(), name, inProgress) + ovfl := overflow(name) + helper := decArrayHelper[t.Elem().Kind()] + op = func(i *decInstr, state *decoderState, value reflect.Value) { + state.dec.decodeArray(state, value, *elemOp, t.Len(), ovfl, helper) + } + + case reflect.Map: + keyId := dec.wireType[wireId].MapT.Key + elemId := dec.wireType[wireId].MapT.Elem + keyOp := dec.decOpFor(keyId, t.Key(), "key of "+name, inProgress) + elemOp := dec.decOpFor(elemId, t.Elem(), "element of "+name, inProgress) + ovfl := overflow(name) + op = func(i *decInstr, state *decoderState, value reflect.Value) { + state.dec.decodeMap(t, state, value, *keyOp, *elemOp, ovfl) + } + + case reflect.Slice: + name = "element of " + name + if t.Elem().Kind() == reflect.Uint8 { + op = decUint8Slice + break + } + var elemId typeId + if tt, ok := builtinIdToType[wireId]; ok { + elemId = tt.(*sliceType).Elem + } else { + elemId = dec.wireType[wireId].SliceT.Elem + } + elemOp := dec.decOpFor(elemId, t.Elem(), name, inProgress) + ovfl := overflow(name) + helper := decSliceHelper[t.Elem().Kind()] + op = func(i *decInstr, state *decoderState, value reflect.Value) { + state.dec.decodeSlice(state, value, *elemOp, ovfl, helper) + } + + case reflect.Struct: + // Generate a closure that calls out to the engine for the nested type. + ut := userType(typ) + enginePtr, err := dec.getDecEnginePtr(wireId, ut) + if err != nil { + error_(err) + } + op = func(i *decInstr, state *decoderState, value reflect.Value) { + // indirect through enginePtr to delay evaluation for recursive structs. + dec.decodeStruct(*enginePtr, value) + } + case reflect.Interface: + op = func(i *decInstr, state *decoderState, value reflect.Value) { + state.dec.decodeInterface(t, state, value) + } + } + } + if op == nil { + errorf("decode can't handle type %s", rt) + } + return &op +} + +var maxIgnoreNestingDepth = 10000 + +// decIgnoreOpFor returns the decoding op for a field that has no destination. +func (dec *Decoder) decIgnoreOpFor(wireId typeId, inProgress map[typeId]*decOp, depth int) *decOp { + if depth > maxIgnoreNestingDepth { + error_(errors.New("invalid nesting depth")) + } + // If this type is already in progress, it's a recursive type (e.g. map[string]*T). + // Return the pointer to the op we're already building. + if opPtr := inProgress[wireId]; opPtr != nil { + return opPtr + } + op, ok := decIgnoreOpMap[wireId] + if !ok { + inProgress[wireId] = &op + if wireId == tInterface { + // Special case because it's a method: the ignored item might + // define types and we need to record their state in the decoder. + op = func(i *decInstr, state *decoderState, value reflect.Value) { + state.dec.ignoreInterface(state) + } + return &op + } + // Special cases + wire := dec.wireType[wireId] + switch { + case wire == nil: + errorf("bad data: undefined type %s", wireId.string()) + case wire.ArrayT != nil: + elemId := wire.ArrayT.Elem + elemOp := dec.decIgnoreOpFor(elemId, inProgress, depth+1) + op = func(i *decInstr, state *decoderState, value reflect.Value) { + state.dec.ignoreArray(state, *elemOp, wire.ArrayT.Len) + } + + case wire.MapT != nil: + keyId := dec.wireType[wireId].MapT.Key + elemId := dec.wireType[wireId].MapT.Elem + keyOp := dec.decIgnoreOpFor(keyId, inProgress, depth+1) + elemOp := dec.decIgnoreOpFor(elemId, inProgress, depth+1) + op = func(i *decInstr, state *decoderState, value reflect.Value) { + state.dec.ignoreMap(state, *keyOp, *elemOp) + } + + case wire.SliceT != nil: + elemId := wire.SliceT.Elem + elemOp := dec.decIgnoreOpFor(elemId, inProgress, depth+1) + op = func(i *decInstr, state *decoderState, value reflect.Value) { + state.dec.ignoreSlice(state, *elemOp) + } + + case wire.StructT != nil: + // Generate a closure that calls out to the engine for the nested type. + enginePtr, err := dec.getIgnoreEnginePtr(wireId) + if err != nil { + error_(err) + } + op = func(i *decInstr, state *decoderState, value reflect.Value) { + // indirect through enginePtr to delay evaluation for recursive structs + state.dec.ignoreStruct(*enginePtr) + } + + case wire.GobEncoderT != nil, wire.BinaryMarshalerT != nil, wire.TextMarshalerT != nil: + op = func(i *decInstr, state *decoderState, value reflect.Value) { + state.dec.ignoreGobDecoder(state) + } + } + } + if op == nil { + errorf("bad data: ignore can't handle type %s", wireId.string()) + } + return &op +} + +// gobDecodeOpFor returns the op for a type that is known to implement +// GobDecoder. +func (dec *Decoder) gobDecodeOpFor(ut *userTypeInfo) *decOp { + rcvrType := ut.user + if ut.decIndir == -1 { + rcvrType = reflect.PointerTo(rcvrType) + } else if ut.decIndir > 0 { + for i := int8(0); i < ut.decIndir; i++ { + rcvrType = rcvrType.Elem() + } + } + var op decOp + op = func(i *decInstr, state *decoderState, value reflect.Value) { + // We now have the base type. We need its address if the receiver is a pointer. + if value.Kind() != reflect.Pointer && rcvrType.Kind() == reflect.Pointer { + value = value.Addr() + } + state.dec.decodeGobDecoder(ut, state, value) + } + return &op +} + +// compatibleType asks: Are these two gob Types compatible? +// Answers the question for basic types, arrays, maps and slices, plus +// GobEncoder/Decoder pairs. +// Structs are considered ok; fields will be checked later. +func (dec *Decoder) compatibleType(fr reflect.Type, fw typeId, inProgress map[reflect.Type]typeId) bool { + if rhs, ok := inProgress[fr]; ok { + return rhs == fw + } + inProgress[fr] = fw + ut := userType(fr) + wire, ok := dec.wireType[fw] + // If wire was encoded with an encoding method, fr must have that method. + // And if not, it must not. + // At most one of the booleans in ut is set. + // We could possibly relax this constraint in the future in order to + // choose the decoding method using the data in the wireType. + // The parentheses look odd but are correct. + if (ut.externalDec == xGob) != (ok && wire.GobEncoderT != nil) || + (ut.externalDec == xBinary) != (ok && wire.BinaryMarshalerT != nil) || + (ut.externalDec == xText) != (ok && wire.TextMarshalerT != nil) { + return false + } + if ut.externalDec != 0 { // This test trumps all others. + return true + } + switch t := ut.base; t.Kind() { + default: + // chan, etc: cannot handle. + return false + case reflect.Bool: + return fw == tBool + case reflect.Int, reflect.Int8, reflect.Int16, reflect.Int32, reflect.Int64: + return fw == tInt + case reflect.Uint, reflect.Uint8, reflect.Uint16, reflect.Uint32, reflect.Uint64, reflect.Uintptr: + return fw == tUint + case reflect.Float32, reflect.Float64: + return fw == tFloat + case reflect.Complex64, reflect.Complex128: + return fw == tComplex + case reflect.String: + return fw == tString + case reflect.Interface: + return fw == tInterface + case reflect.Array: + if !ok || wire.ArrayT == nil { + return false + } + array := wire.ArrayT + return t.Len() == array.Len && dec.compatibleType(t.Elem(), array.Elem, inProgress) + case reflect.Map: + if !ok || wire.MapT == nil { + return false + } + MapType := wire.MapT + return dec.compatibleType(t.Key(), MapType.Key, inProgress) && dec.compatibleType(t.Elem(), MapType.Elem, inProgress) + case reflect.Slice: + // Is it an array of bytes? + if t.Elem().Kind() == reflect.Uint8 { + return fw == tBytes + } + // Extract and compare element types. + var sw *sliceType + if tt, ok := builtinIdToType[fw]; ok { + sw, _ = tt.(*sliceType) + } else if wire != nil { + sw = wire.SliceT + } + elem := userType(t.Elem()).base + return sw != nil && dec.compatibleType(elem, sw.Elem, inProgress) + case reflect.Struct: + return true + } +} + +// typeString returns a human-readable description of the type identified by remoteId. +func (dec *Decoder) typeString(remoteId typeId) string { + typeLock.Lock() + defer typeLock.Unlock() + if t := idToType[remoteId]; t != nil { + // globally known type. + return t.string() + } + return dec.wireType[remoteId].string() +} + +// compileSingle compiles the decoder engine for a non-struct top-level value, including +// GobDecoders. +func (dec *Decoder) compileSingle(remoteId typeId, ut *userTypeInfo) (engine *decEngine, err error) { + rt := ut.user + engine = new(decEngine) + engine.instr = make([]decInstr, 1) // one item + name := rt.String() // best we can do + if !dec.compatibleType(rt, remoteId, make(map[reflect.Type]typeId)) { + remoteType := dec.typeString(remoteId) + // Common confusing case: local interface type, remote concrete type. + if ut.base.Kind() == reflect.Interface && remoteId != tInterface { + return nil, errors.New("gob: local interface type " + name + " can only be decoded from remote interface type; received concrete type " + remoteType) + } + return nil, errors.New("gob: decoding into local type " + name + ", received remote type " + remoteType) + } + op := dec.decOpFor(remoteId, rt, name, make(map[reflect.Type]*decOp)) + ovfl := errors.New(`value for "` + name + `" out of range`) + engine.instr[singletonField] = decInstr{*op, singletonField, nil, ovfl} + engine.numInstr = 1 + return +} + +// compileIgnoreSingle compiles the decoder engine for a non-struct top-level value that will be discarded. +func (dec *Decoder) compileIgnoreSingle(remoteId typeId) *decEngine { + engine := new(decEngine) + engine.instr = make([]decInstr, 1) // one item + op := dec.decIgnoreOpFor(remoteId, make(map[typeId]*decOp), 0) + ovfl := overflow(dec.typeString(remoteId)) + engine.instr[0] = decInstr{*op, 0, nil, ovfl} + engine.numInstr = 1 + return engine +} + +// compileDec compiles the decoder engine for a value. If the value is not a struct, +// it calls out to compileSingle. +func (dec *Decoder) compileDec(remoteId typeId, ut *userTypeInfo) (engine *decEngine, err error) { + defer catchError(&err) + rt := ut.base + srt := rt + if srt.Kind() != reflect.Struct || ut.externalDec != 0 { + return dec.compileSingle(remoteId, ut) + } + var wireStruct *structType + // Builtin types can come from global pool; the rest must be defined by the decoder. + // Also we know we're decoding a struct now, so the client must have sent one. + if t, ok := builtinIdToType[remoteId]; ok { + wireStruct, _ = t.(*structType) + } else { + wire := dec.wireType[remoteId] + if wire == nil { + error_(errBadType) + } + wireStruct = wire.StructT + } + if wireStruct == nil { + errorf("type mismatch in decoder: want struct type %s; got non-struct", rt) + } + engine = new(decEngine) + engine.instr = make([]decInstr, len(wireStruct.Field)) + seen := make(map[reflect.Type]*decOp) + // Loop over the fields of the wire type. + for fieldnum := 0; fieldnum < len(wireStruct.Field); fieldnum++ { + wireField := wireStruct.Field[fieldnum] + if wireField.Name == "" { + errorf("empty name for remote field of type %s", wireStruct.Name) + } + ovfl := overflow(wireField.Name) + // Find the field of the local type with the same name. + localField, present := srt.FieldByName(wireField.Name) + // TODO(r): anonymous names + if !present || !isExported(wireField.Name) { + op := dec.decIgnoreOpFor(wireField.Id, make(map[typeId]*decOp), 0) + engine.instr[fieldnum] = decInstr{*op, fieldnum, nil, ovfl} + continue + } + if !dec.compatibleType(localField.Type, wireField.Id, make(map[reflect.Type]typeId)) { + errorf("wrong type (%s) for received field %s.%s", localField.Type, wireStruct.Name, wireField.Name) + } + op := dec.decOpFor(wireField.Id, localField.Type, localField.Name, seen) + engine.instr[fieldnum] = decInstr{*op, fieldnum, localField.Index, ovfl} + engine.numInstr++ + } + return +} + +// getDecEnginePtr returns the engine for the specified type. +func (dec *Decoder) getDecEnginePtr(remoteId typeId, ut *userTypeInfo) (enginePtr **decEngine, err error) { + rt := ut.user + decoderMap, ok := dec.decoderCache[rt] + if !ok { + decoderMap = make(map[typeId]**decEngine) + dec.decoderCache[rt] = decoderMap + } + if enginePtr, ok = decoderMap[remoteId]; !ok { + // To handle recursive types, mark this engine as underway before compiling. + enginePtr = new(*decEngine) + decoderMap[remoteId] = enginePtr + *enginePtr, err = dec.compileDec(remoteId, ut) + if err != nil { + delete(decoderMap, remoteId) + } + } + return +} + +// emptyStruct is the type we compile into when ignoring a struct value. +type emptyStruct struct{} + +var emptyStructType = reflect.TypeOf(emptyStruct{}) + +// getIgnoreEnginePtr returns the engine for the specified type when the value is to be discarded. +func (dec *Decoder) getIgnoreEnginePtr(wireId typeId) (enginePtr **decEngine, err error) { + var ok bool + if enginePtr, ok = dec.ignorerCache[wireId]; !ok { + // To handle recursive types, mark this engine as underway before compiling. + enginePtr = new(*decEngine) + dec.ignorerCache[wireId] = enginePtr + wire := dec.wireType[wireId] + if wire != nil && wire.StructT != nil { + *enginePtr, err = dec.compileDec(wireId, userType(emptyStructType)) + } else { + *enginePtr = dec.compileIgnoreSingle(wireId) + } + if err != nil { + delete(dec.ignorerCache, wireId) + } + } + return +} + +// decodeValue decodes the data stream representing a value and stores it in value. +func (dec *Decoder) decodeValue(wireId typeId, value reflect.Value) { + defer catchError(&dec.err) + // If the value is nil, it means we should just ignore this item. + if !value.IsValid() { + dec.decodeIgnoredValue(wireId) + return + } + // Dereference down to the underlying type. + ut := userType(value.Type()) + base := ut.base + var enginePtr **decEngine + enginePtr, dec.err = dec.getDecEnginePtr(wireId, ut) + if dec.err != nil { + return + } + value = decAlloc(value) + engine := *enginePtr + if st := base; st.Kind() == reflect.Struct && ut.externalDec == 0 { + wt := dec.wireType[wireId] + if engine.numInstr == 0 && st.NumField() > 0 && + wt != nil && len(wt.StructT.Field) > 0 { + name := base.Name() + errorf("type mismatch: no fields matched compiling decoder for %s", name) + } + dec.decodeStruct(engine, value) + } else { + dec.decodeSingle(engine, value) + } +} + +// decodeIgnoredValue decodes the data stream representing a value of the specified type and discards it. +func (dec *Decoder) decodeIgnoredValue(wireId typeId) { + var enginePtr **decEngine + enginePtr, dec.err = dec.getIgnoreEnginePtr(wireId) + if dec.err != nil { + return + } + wire := dec.wireType[wireId] + if wire != nil && wire.StructT != nil { + dec.ignoreStruct(*enginePtr) + } else { + dec.ignoreSingle(*enginePtr) + } +} + +const ( + intBits = 32 << (^uint(0) >> 63) + uintptrBits = 32 << (^uintptr(0) >> 63) +) + +func init() { + var iop, uop decOp + switch intBits { + case 32: + iop = decInt32 + uop = decUint32 + case 64: + iop = decInt64 + uop = decUint64 + default: + panic("gob: unknown size of int/uint") + } + decOpTable[reflect.Int] = iop + decOpTable[reflect.Uint] = uop + + // Finally uintptr + switch uintptrBits { + case 32: + uop = decUint32 + case 64: + uop = decUint64 + default: + panic("gob: unknown size of uintptr") + } + decOpTable[reflect.Uintptr] = uop +} + +// Gob depends on being able to take the address +// of zeroed Values it creates, so use this wrapper instead +// of the standard reflect.Zero. +// Each call allocates once. +func allocValue(t reflect.Type) reflect.Value { + return reflect.New(t).Elem() +} diff --git a/src/encoding/gob/decoder.go b/src/encoding/gob/decoder.go new file mode 100644 index 0000000..5b77adc --- /dev/null +++ b/src/encoding/gob/decoder.go @@ -0,0 +1,237 @@ +// 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 gob + +import ( + "bufio" + "errors" + "internal/saferio" + "io" + "reflect" + "sync" +) + +// tooBig provides a sanity check for sizes; used in several places. Upper limit +// of is 1GB on 32-bit systems, 8GB on 64-bit, allowing room to grow a little +// without overflow. +const tooBig = (1 << 30) << (^uint(0) >> 62) + +// A Decoder manages the receipt of type and data information read from the +// remote side of a connection. It is safe for concurrent use by multiple +// goroutines. +// +// The Decoder does only basic sanity checking on decoded input sizes, +// and its limits are not configurable. Take caution when decoding gob data +// from untrusted sources. +type Decoder struct { + mutex sync.Mutex // each item must be received atomically + r io.Reader // source of the data + buf decBuffer // buffer for more efficient i/o from r + wireType map[typeId]*wireType // map from remote ID to local description + decoderCache map[reflect.Type]map[typeId]**decEngine // cache of compiled engines + ignorerCache map[typeId]**decEngine // ditto for ignored objects + freeList *decoderState // list of free decoderStates; avoids reallocation + countBuf []byte // used for decoding integers while parsing messages + err error +} + +// NewDecoder returns a new decoder that reads from the io.Reader. +// If r does not also implement io.ByteReader, it will be wrapped in a +// bufio.Reader. +func NewDecoder(r io.Reader) *Decoder { + dec := new(Decoder) + // We use the ability to read bytes as a plausible surrogate for buffering. + if _, ok := r.(io.ByteReader); !ok { + r = bufio.NewReader(r) + } + dec.r = r + dec.wireType = make(map[typeId]*wireType) + dec.decoderCache = make(map[reflect.Type]map[typeId]**decEngine) + dec.ignorerCache = make(map[typeId]**decEngine) + dec.countBuf = make([]byte, 9) // counts may be uint64s (unlikely!), require 9 bytes + + return dec +} + +// recvType loads the definition of a type. +func (dec *Decoder) recvType(id typeId) { + // Have we already seen this type? That's an error + if id < firstUserId || dec.wireType[id] != nil { + dec.err = errors.New("gob: duplicate type received") + return + } + + // Type: + wire := new(wireType) + dec.decodeValue(tWireType, reflect.ValueOf(wire)) + if dec.err != nil { + return + } + // Remember we've seen this type. + dec.wireType[id] = wire +} + +var errBadCount = errors.New("invalid message length") + +// recvMessage reads the next count-delimited item from the input. It is the converse +// of Encoder.writeMessage. It returns false on EOF or other error reading the message. +func (dec *Decoder) recvMessage() bool { + // Read a count. + nbytes, _, err := decodeUintReader(dec.r, dec.countBuf) + if err != nil { + dec.err = err + return false + } + if nbytes >= tooBig { + dec.err = errBadCount + return false + } + dec.readMessage(int(nbytes)) + return dec.err == nil +} + +// readMessage reads the next nbytes bytes from the input. +func (dec *Decoder) readMessage(nbytes int) { + if dec.buf.Len() != 0 { + // The buffer should always be empty now. + panic("non-empty decoder buffer") + } + // Read the data + var buf []byte + buf, dec.err = saferio.ReadData(dec.r, uint64(nbytes)) + dec.buf.SetBytes(buf) + if dec.err == io.EOF { + dec.err = io.ErrUnexpectedEOF + } +} + +// toInt turns an encoded uint64 into an int, according to the marshaling rules. +func toInt(x uint64) int64 { + i := int64(x >> 1) + if x&1 != 0 { + i = ^i + } + return i +} + +func (dec *Decoder) nextInt() int64 { + n, _, err := decodeUintReader(&dec.buf, dec.countBuf) + if err != nil { + dec.err = err + } + return toInt(n) +} + +func (dec *Decoder) nextUint() uint64 { + n, _, err := decodeUintReader(&dec.buf, dec.countBuf) + if err != nil { + dec.err = err + } + return n +} + +// decodeTypeSequence parses: +// TypeSequence +// +// (TypeDefinition DelimitedTypeDefinition*)? +// +// and returns the type id of the next value. It returns -1 at +// EOF. Upon return, the remainder of dec.buf is the value to be +// decoded. If this is an interface value, it can be ignored by +// resetting that buffer. +func (dec *Decoder) decodeTypeSequence(isInterface bool) typeId { + firstMessage := true + for dec.err == nil { + if dec.buf.Len() == 0 { + if !dec.recvMessage() { + // We can only return io.EOF if the input was empty. + // If we read one or more type spec messages, + // require a data item message to follow. + // If we hit an EOF before that, then give ErrUnexpectedEOF. + if !firstMessage && dec.err == io.EOF { + dec.err = io.ErrUnexpectedEOF + } + break + } + } + // Receive a type id. + id := typeId(dec.nextInt()) + if id >= 0 { + // Value follows. + return id + } + // Type definition for (-id) follows. + dec.recvType(-id) + if dec.err != nil { + break + } + // When decoding an interface, after a type there may be a + // DelimitedValue still in the buffer. Skip its count. + // (Alternatively, the buffer is empty and the byte count + // will be absorbed by recvMessage.) + if dec.buf.Len() > 0 { + if !isInterface { + dec.err = errors.New("extra data in buffer") + break + } + dec.nextUint() + } + firstMessage = false + } + return -1 +} + +// Decode reads the next value from the input stream and stores +// it in the data represented by the empty interface value. +// If e is nil, the value will be discarded. Otherwise, +// the value underlying e must be a pointer to the +// correct type for the next data item received. +// If the input is at EOF, Decode returns io.EOF and +// does not modify e. +func (dec *Decoder) Decode(e any) error { + if e == nil { + return dec.DecodeValue(reflect.Value{}) + } + value := reflect.ValueOf(e) + // If e represents a value as opposed to a pointer, the answer won't + // get back to the caller. Make sure it's a pointer. + if value.Type().Kind() != reflect.Pointer { + dec.err = errors.New("gob: attempt to decode into a non-pointer") + return dec.err + } + return dec.DecodeValue(value) +} + +// DecodeValue reads the next value from the input stream. +// If v is the zero reflect.Value (v.Kind() == Invalid), DecodeValue discards the value. +// Otherwise, it stores the value into v. In that case, v must represent +// a non-nil pointer to data or be an assignable reflect.Value (v.CanSet()) +// If the input is at EOF, DecodeValue returns io.EOF and +// does not modify v. +func (dec *Decoder) DecodeValue(v reflect.Value) error { + if v.IsValid() { + if v.Kind() == reflect.Pointer && !v.IsNil() { + // That's okay, we'll store through the pointer. + } else if !v.CanSet() { + return errors.New("gob: DecodeValue of unassignable value") + } + } + // Make sure we're single-threaded through here. + dec.mutex.Lock() + defer dec.mutex.Unlock() + + dec.buf.Reset() // In case data lingers from previous invocation. + dec.err = nil + id := dec.decodeTypeSequence(false) + if dec.err == nil { + dec.decodeValue(id, v) + } + return dec.err +} + +// If debug.go is compiled into the program, debugFunc prints a human-readable +// representation of the gob data read from r by calling that file's Debug function. +// Otherwise it is nil. +var debugFunc func(io.Reader) diff --git a/src/encoding/gob/doc.go b/src/encoding/gob/doc.go new file mode 100644 index 0000000..15473f1 --- /dev/null +++ b/src/encoding/gob/doc.go @@ -0,0 +1,423 @@ +// 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 gob manages streams of gobs - binary values exchanged between an +Encoder (transmitter) and a Decoder (receiver). A typical use is transporting +arguments and results of remote procedure calls (RPCs) such as those provided by +package "net/rpc". + +The implementation compiles a custom codec for each data type in the stream and +is most efficient when a single Encoder is used to transmit a stream of values, +amortizing the cost of compilation. + +# Basics + +A stream of gobs is self-describing. Each data item in the stream is preceded by +a specification of its type, expressed in terms of a small set of predefined +types. Pointers are not transmitted, but the things they point to are +transmitted; that is, the values are flattened. Nil pointers are not permitted, +as they have no value. Recursive types work fine, but +recursive values (data with cycles) are problematic. This may change. + +To use gobs, create an Encoder and present it with a series of data items as +values or addresses that can be dereferenced to values. The Encoder makes sure +all type information is sent before it is needed. At the receive side, a +Decoder retrieves values from the encoded stream and unpacks them into local +variables. + +# Types and Values + +The source and destination values/types need not correspond exactly. For structs, +fields (identified by name) that are in the source but absent from the receiving +variable will be ignored. Fields that are in the receiving variable but missing +from the transmitted type or value will be ignored in the destination. If a field +with the same name is present in both, their types must be compatible. Both the +receiver and transmitter will do all necessary indirection and dereferencing to +convert between gobs and actual Go values. For instance, a gob type that is +schematically, + + struct { A, B int } + +can be sent from or received into any of these Go types: + + struct { A, B int } // the same + *struct { A, B int } // extra indirection of the struct + struct { *A, **B int } // extra indirection of the fields + struct { A, B int64 } // different concrete value type; see below + +It may also be received into any of these: + + struct { A, B int } // the same + struct { B, A int } // ordering doesn't matter; matching is by name + struct { A, B, C int } // extra field (C) ignored + struct { B int } // missing field (A) ignored; data will be dropped + struct { B, C int } // missing field (A) ignored; extra field (C) ignored. + +Attempting to receive into these types will draw a decode error: + + struct { A int; B uint } // change of signedness for B + struct { A int; B float } // change of type for B + struct { } // no field names in common + struct { C, D int } // no field names in common + +Integers are transmitted two ways: arbitrary precision signed integers or +arbitrary precision unsigned integers. There is no int8, int16 etc. +discrimination in the gob format; there are only signed and unsigned integers. As +described below, the transmitter sends the value in a variable-length encoding; +the receiver accepts the value and stores it in the destination variable. +Floating-point numbers are always sent using IEEE-754 64-bit precision (see +below). + +Signed integers may be received into any signed integer variable: int, int16, etc.; +unsigned integers may be received into any unsigned integer variable; and floating +point values may be received into any floating point variable. However, +the destination variable must be able to represent the value or the decode +operation will fail. + +Structs, arrays and slices are also supported. Structs encode and decode only +exported fields. Strings and arrays of bytes are supported with a special, +efficient representation (see below). When a slice is decoded, if the existing +slice has capacity the slice will be extended in place; if not, a new array is +allocated. Regardless, the length of the resulting slice reports the number of +elements decoded. + +In general, if allocation is required, the decoder will allocate memory. If not, +it will update the destination variables with values read from the stream. It does +not initialize them first, so if the destination is a compound value such as a +map, struct, or slice, the decoded values will be merged elementwise into the +existing variables. + +Functions and channels will not be sent in a gob. Attempting to encode such a value +at the top level will fail. A struct field of chan or func type is treated exactly +like an unexported field and is ignored. + +Gob can encode a value of any type implementing the GobEncoder or +encoding.BinaryMarshaler interfaces by calling the corresponding method, +in that order of preference. + +Gob can decode a value of any type implementing the GobDecoder or +encoding.BinaryUnmarshaler interfaces by calling the corresponding method, +again in that order of preference. + +# Encoding Details + +This section documents the encoding, details that are not important for most +users. Details are presented bottom-up. + +An unsigned integer is sent one of two ways. If it is less than 128, it is sent +as a byte with that value. Otherwise it is sent as a minimal-length big-endian +(high byte first) byte stream holding the value, preceded by one byte holding the +byte count, negated. Thus 0 is transmitted as (00), 7 is transmitted as (07) and +256 is transmitted as (FE 01 00). + +A boolean is encoded within an unsigned integer: 0 for false, 1 for true. + +A signed integer, i, is encoded within an unsigned integer, u. Within u, bits 1 +upward contain the value; bit 0 says whether they should be complemented upon +receipt. The encode algorithm looks like this: + + var u uint + if i < 0 { + u = (^uint(i) << 1) | 1 // complement i, bit 0 is 1 + } else { + u = (uint(i) << 1) // do not complement i, bit 0 is 0 + } + encodeUnsigned(u) + +The low bit is therefore analogous to a sign bit, but making it the complement bit +instead guarantees that the largest negative integer is not a special case. For +example, -129=^128=(^256>>1) encodes as (FE 01 01). + +Floating-point numbers are always sent as a representation of a float64 value. +That value is converted to a uint64 using math.Float64bits. The uint64 is then +byte-reversed and sent as a regular unsigned integer. The byte-reversal means the +exponent and high-precision part of the mantissa go first. Since the low bits are +often zero, this can save encoding bytes. For instance, 17.0 is encoded in only +three bytes (FE 31 40). + +Strings and slices of bytes are sent as an unsigned count followed by that many +uninterpreted bytes of the value. + +All other slices and arrays are sent as an unsigned count followed by that many +elements using the standard gob encoding for their type, recursively. + +Maps are sent as an unsigned count followed by that many key, element +pairs. Empty but non-nil maps are sent, so if the receiver has not allocated +one already, one will always be allocated on receipt unless the transmitted map +is nil and not at the top level. + +In slices and arrays, as well as maps, all elements, even zero-valued elements, +are transmitted, even if all the elements are zero. + +Structs are sent as a sequence of (field number, field value) pairs. The field +value is sent using the standard gob encoding for its type, recursively. If a +field has the zero value for its type (except for arrays; see above), it is omitted +from the transmission. The field number is defined by the type of the encoded +struct: the first field of the encoded type is field 0, the second is field 1, +etc. When encoding a value, the field numbers are delta encoded for efficiency +and the fields are always sent in order of increasing field number; the deltas are +therefore unsigned. The initialization for the delta encoding sets the field +number to -1, so an unsigned integer field 0 with value 7 is transmitted as unsigned +delta = 1, unsigned value = 7 or (01 07). Finally, after all the fields have been +sent a terminating mark denotes the end of the struct. That mark is a delta=0 +value, which has representation (00). + +Interface types are not checked for compatibility; all interface types are +treated, for transmission, as members of a single "interface" type, analogous to +int or []byte - in effect they're all treated as interface{}. Interface values +are transmitted as a string identifying the concrete type being sent (a name +that must be pre-defined by calling Register), followed by a byte count of the +length of the following data (so the value can be skipped if it cannot be +stored), followed by the usual encoding of concrete (dynamic) value stored in +the interface value. (A nil interface value is identified by the empty string +and transmits no value.) Upon receipt, the decoder verifies that the unpacked +concrete item satisfies the interface of the receiving variable. + +If a value is passed to Encode and the type is not a struct (or pointer to struct, +etc.), for simplicity of processing it is represented as a struct of one field. +The only visible effect of this is to encode a zero byte after the value, just as +after the last field of an encoded struct, so that the decode algorithm knows when +the top-level value is complete. + +The representation of types is described below. When a type is defined on a given +connection between an Encoder and Decoder, it is assigned a signed integer type +id. When Encoder.Encode(v) is called, it makes sure there is an id assigned for +the type of v and all its elements and then it sends the pair (typeid, encoded-v) +where typeid is the type id of the encoded type of v and encoded-v is the gob +encoding of the value v. + +To define a type, the encoder chooses an unused, positive type id and sends the +pair (-type id, encoded-type) where encoded-type is the gob encoding of a wireType +description, constructed from these types: + + type wireType struct { + ArrayT *ArrayType + SliceT *SliceType + StructT *StructType + MapT *MapType + GobEncoderT *gobEncoderType + BinaryMarshalerT *gobEncoderType + TextMarshalerT *gobEncoderType + + } + type arrayType struct { + CommonType + Elem typeId + Len int + } + type CommonType struct { + Name string // the name of the struct type + Id int // the id of the type, repeated so it's inside the type + } + type sliceType struct { + CommonType + Elem typeId + } + type structType struct { + CommonType + Field []*fieldType // the fields of the struct. + } + type fieldType struct { + Name string // the name of the field. + Id int // the type id of the field, which must be already defined + } + type mapType struct { + CommonType + Key typeId + Elem typeId + } + type gobEncoderType struct { + CommonType + } + +If there are nested type ids, the types for all inner type ids must be defined +before the top-level type id is used to describe an encoded-v. + +For simplicity in setup, the connection is defined to understand these types a +priori, as well as the basic gob types int, uint, etc. Their ids are: + + bool 1 + int 2 + uint 3 + float 4 + []byte 5 + string 6 + complex 7 + interface 8 + // gap for reserved ids. + WireType 16 + ArrayType 17 + CommonType 18 + SliceType 19 + StructType 20 + FieldType 21 + // 22 is slice of fieldType. + MapType 23 + +Finally, each message created by a call to Encode is preceded by an encoded +unsigned integer count of the number of bytes remaining in the message. After +the initial type name, interface values are wrapped the same way; in effect, the +interface value acts like a recursive invocation of Encode. + +In summary, a gob stream looks like + + (byteCount (-type id, encoding of a wireType)* (type id, encoding of a value))* + +where * signifies zero or more repetitions and the type id of a value must +be predefined or be defined before the value in the stream. + +Compatibility: Any future changes to the package will endeavor to maintain +compatibility with streams encoded using previous versions. That is, any released +version of this package should be able to decode data written with any previously +released version, subject to issues such as security fixes. See the Go compatibility +document for background: https://golang.org/doc/go1compat + +See "Gobs of data" for a design discussion of the gob wire format: +https://blog.golang.org/gobs-of-data + +# Security + +This package is not designed to be hardened against adversarial inputs, and is +outside the scope of https://go.dev/security/policy. In particular, the Decoder +does only basic sanity checking on decoded input sizes, and its limits are not +configurable. Care should be taken when decoding gob data from untrusted +sources, which may consume significant resources. +*/ +package gob + +/* +Grammar: + +Tokens starting with a lower case letter are terminals; int(n) +and uint(n) represent the signed/unsigned encodings of the value n. + +GobStream: + DelimitedMessage* +DelimitedMessage: + uint(lengthOfMessage) Message +Message: + TypeSequence TypedValue +TypeSequence + (TypeDefinition DelimitedTypeDefinition*)? +DelimitedTypeDefinition: + uint(lengthOfTypeDefinition) TypeDefinition +TypedValue: + int(typeId) Value +TypeDefinition: + int(-typeId) encodingOfWireType +Value: + SingletonValue | StructValue +SingletonValue: + uint(0) FieldValue +FieldValue: + builtinValue | ArrayValue | MapValue | SliceValue | StructValue | InterfaceValue +InterfaceValue: + NilInterfaceValue | NonNilInterfaceValue +NilInterfaceValue: + uint(0) +NonNilInterfaceValue: + ConcreteTypeName TypeSequence InterfaceContents +ConcreteTypeName: + uint(lengthOfName) [already read=n] name +InterfaceContents: + int(concreteTypeId) DelimitedValue +DelimitedValue: + uint(length) Value +ArrayValue: + uint(n) FieldValue*n [n elements] +MapValue: + uint(n) (FieldValue FieldValue)*n [n (key, value) pairs] +SliceValue: + uint(n) FieldValue*n [n elements] +StructValue: + (uint(fieldDelta) FieldValue)* +*/ + +/* +For implementers and the curious, here is an encoded example. Given + type Point struct {X, Y int} +and the value + p := Point{22, 33} +the bytes transmitted that encode p will be: + 1f ff 81 03 01 01 05 50 6f 69 6e 74 01 ff 82 00 + 01 02 01 01 58 01 04 00 01 01 59 01 04 00 00 00 + 07 ff 82 01 2c 01 42 00 +They are determined as follows. + +Since this is the first transmission of type Point, the type descriptor +for Point itself must be sent before the value. This is the first type +we've sent on this Encoder, so it has type id 65 (0 through 64 are +reserved). + + 1f // This item (a type descriptor) is 31 bytes long. + ff 81 // The negative of the id for the type we're defining, -65. + // This is one byte (indicated by FF = -1) followed by + // ^-65<<1 | 1. The low 1 bit signals to complement the + // rest upon receipt. + + // Now we send a type descriptor, which is itself a struct (wireType). + // The type of wireType itself is known (it's built in, as is the type of + // all its components), so we just need to send a *value* of type wireType + // that represents type "Point". + // Here starts the encoding of that value. + // Set the field number implicitly to -1; this is done at the beginning + // of every struct, including nested structs. + 03 // Add 3 to field number; now 2 (wireType.structType; this is a struct). + // structType starts with an embedded CommonType, which appears + // as a regular structure here too. + 01 // add 1 to field number (now 0); start of embedded CommonType. + 01 // add 1 to field number (now 0, the name of the type) + 05 // string is (unsigned) 5 bytes long + 50 6f 69 6e 74 // wireType.structType.CommonType.name = "Point" + 01 // add 1 to field number (now 1, the id of the type) + ff 82 // wireType.structType.CommonType._id = 65 + 00 // end of embedded wiretype.structType.CommonType struct + 01 // add 1 to field number (now 1, the field array in wireType.structType) + 02 // There are two fields in the type (len(structType.field)) + 01 // Start of first field structure; add 1 to get field number 0: field[0].name + 01 // 1 byte + 58 // structType.field[0].name = "X" + 01 // Add 1 to get field number 1: field[0].id + 04 // structType.field[0].typeId is 2 (signed int). + 00 // End of structType.field[0]; start structType.field[1]; set field number to -1. + 01 // Add 1 to get field number 0: field[1].name + 01 // 1 byte + 59 // structType.field[1].name = "Y" + 01 // Add 1 to get field number 1: field[1].id + 04 // struct.Type.field[1].typeId is 2 (signed int). + 00 // End of structType.field[1]; end of structType.field. + 00 // end of wireType.structType structure + 00 // end of wireType structure + +Now we can send the Point value. Again the field number resets to -1: + + 07 // this value is 7 bytes long + ff 82 // the type number, 65 (1 byte (-FF) followed by 65<<1) + 01 // add one to field number, yielding field 0 + 2c // encoding of signed "22" (0x2c = 44 = 22<<1); Point.x = 22 + 01 // add one to field number, yielding field 1 + 42 // encoding of signed "33" (0x42 = 66 = 33<<1); Point.y = 33 + 00 // end of structure + +The type encoding is long and fairly intricate but we send it only once. +If p is transmitted a second time, the type is already known so the +output will be just: + + 07 ff 82 01 2c 01 42 00 + +A single non-struct value at top level is transmitted like a field with +delta tag 0. For instance, a signed integer with value 3 presented as +the argument to Encode will emit: + + 03 04 00 06 + +Which represents: + + 03 // this value is 3 bytes long + 04 // the type number, 2, represents an integer + 00 // tag delta 0 + 06 // value 3 + +*/ diff --git a/src/encoding/gob/dump.go b/src/encoding/gob/dump.go new file mode 100644 index 0000000..f4b1beb --- /dev/null +++ b/src/encoding/gob/dump.go @@ -0,0 +1,29 @@ +// 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. + +//go:build ignore + +package main + +// Need to compile package gob with debug.go to build this program. +// See comments in debug.go for how to do this. + +import ( + "encoding/gob" + "fmt" + "os" +) + +func main() { + var err error + file := os.Stdin + if len(os.Args) > 1 { + file, err = os.Open(os.Args[1]) + if err != nil { + fmt.Fprintf(os.Stderr, "dump: %s\n", err) + os.Exit(1) + } + } + gob.Debug(file) +} diff --git a/src/encoding/gob/enc_helpers.go b/src/encoding/gob/enc_helpers.go new file mode 100644 index 0000000..c3b4ca8 --- /dev/null +++ b/src/encoding/gob/enc_helpers.go @@ -0,0 +1,414 @@ +// Code generated by go run encgen.go -output enc_helpers.go; DO NOT EDIT. + +// Copyright 2014 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 gob + +import ( + "reflect" +) + +var encArrayHelper = map[reflect.Kind]encHelper{ + reflect.Bool: encBoolArray, + reflect.Complex64: encComplex64Array, + reflect.Complex128: encComplex128Array, + reflect.Float32: encFloat32Array, + reflect.Float64: encFloat64Array, + reflect.Int: encIntArray, + reflect.Int16: encInt16Array, + reflect.Int32: encInt32Array, + reflect.Int64: encInt64Array, + reflect.Int8: encInt8Array, + reflect.String: encStringArray, + reflect.Uint: encUintArray, + reflect.Uint16: encUint16Array, + reflect.Uint32: encUint32Array, + reflect.Uint64: encUint64Array, + reflect.Uintptr: encUintptrArray, +} + +var encSliceHelper = map[reflect.Kind]encHelper{ + reflect.Bool: encBoolSlice, + reflect.Complex64: encComplex64Slice, + reflect.Complex128: encComplex128Slice, + reflect.Float32: encFloat32Slice, + reflect.Float64: encFloat64Slice, + reflect.Int: encIntSlice, + reflect.Int16: encInt16Slice, + reflect.Int32: encInt32Slice, + reflect.Int64: encInt64Slice, + reflect.Int8: encInt8Slice, + reflect.String: encStringSlice, + reflect.Uint: encUintSlice, + reflect.Uint16: encUint16Slice, + reflect.Uint32: encUint32Slice, + reflect.Uint64: encUint64Slice, + reflect.Uintptr: encUintptrSlice, +} + +func encBoolArray(state *encoderState, v reflect.Value) bool { + // Can only slice if it is addressable. + if !v.CanAddr() { + return false + } + return encBoolSlice(state, v.Slice(0, v.Len())) +} + +func encBoolSlice(state *encoderState, v reflect.Value) bool { + slice, ok := v.Interface().([]bool) + if !ok { + // It is kind bool but not type bool. TODO: We can handle this unsafely. + return false + } + for _, x := range slice { + if x != false || state.sendZero { + if x { + state.encodeUint(1) + } else { + state.encodeUint(0) + } + } + } + return true +} + +func encComplex64Array(state *encoderState, v reflect.Value) bool { + // Can only slice if it is addressable. + if !v.CanAddr() { + return false + } + return encComplex64Slice(state, v.Slice(0, v.Len())) +} + +func encComplex64Slice(state *encoderState, v reflect.Value) bool { + slice, ok := v.Interface().([]complex64) + if !ok { + // It is kind complex64 but not type complex64. TODO: We can handle this unsafely. + return false + } + for _, x := range slice { + if x != 0+0i || state.sendZero { + rpart := floatBits(float64(real(x))) + ipart := floatBits(float64(imag(x))) + state.encodeUint(rpart) + state.encodeUint(ipart) + } + } + return true +} + +func encComplex128Array(state *encoderState, v reflect.Value) bool { + // Can only slice if it is addressable. + if !v.CanAddr() { + return false + } + return encComplex128Slice(state, v.Slice(0, v.Len())) +} + +func encComplex128Slice(state *encoderState, v reflect.Value) bool { + slice, ok := v.Interface().([]complex128) + if !ok { + // It is kind complex128 but not type complex128. TODO: We can handle this unsafely. + return false + } + for _, x := range slice { + if x != 0+0i || state.sendZero { + rpart := floatBits(real(x)) + ipart := floatBits(imag(x)) + state.encodeUint(rpart) + state.encodeUint(ipart) + } + } + return true +} + +func encFloat32Array(state *encoderState, v reflect.Value) bool { + // Can only slice if it is addressable. + if !v.CanAddr() { + return false + } + return encFloat32Slice(state, v.Slice(0, v.Len())) +} + +func encFloat32Slice(state *encoderState, v reflect.Value) bool { + slice, ok := v.Interface().([]float32) + if !ok { + // It is kind float32 but not type float32. TODO: We can handle this unsafely. + return false + } + for _, x := range slice { + if x != 0 || state.sendZero { + bits := floatBits(float64(x)) + state.encodeUint(bits) + } + } + return true +} + +func encFloat64Array(state *encoderState, v reflect.Value) bool { + // Can only slice if it is addressable. + if !v.CanAddr() { + return false + } + return encFloat64Slice(state, v.Slice(0, v.Len())) +} + +func encFloat64Slice(state *encoderState, v reflect.Value) bool { + slice, ok := v.Interface().([]float64) + if !ok { + // It is kind float64 but not type float64. TODO: We can handle this unsafely. + return false + } + for _, x := range slice { + if x != 0 || state.sendZero { + bits := floatBits(x) + state.encodeUint(bits) + } + } + return true +} + +func encIntArray(state *encoderState, v reflect.Value) bool { + // Can only slice if it is addressable. + if !v.CanAddr() { + return false + } + return encIntSlice(state, v.Slice(0, v.Len())) +} + +func encIntSlice(state *encoderState, v reflect.Value) bool { + slice, ok := v.Interface().([]int) + if !ok { + // It is kind int but not type int. TODO: We can handle this unsafely. + return false + } + for _, x := range slice { + if x != 0 || state.sendZero { + state.encodeInt(int64(x)) + } + } + return true +} + +func encInt16Array(state *encoderState, v reflect.Value) bool { + // Can only slice if it is addressable. + if !v.CanAddr() { + return false + } + return encInt16Slice(state, v.Slice(0, v.Len())) +} + +func encInt16Slice(state *encoderState, v reflect.Value) bool { + slice, ok := v.Interface().([]int16) + if !ok { + // It is kind int16 but not type int16. TODO: We can handle this unsafely. + return false + } + for _, x := range slice { + if x != 0 || state.sendZero { + state.encodeInt(int64(x)) + } + } + return true +} + +func encInt32Array(state *encoderState, v reflect.Value) bool { + // Can only slice if it is addressable. + if !v.CanAddr() { + return false + } + return encInt32Slice(state, v.Slice(0, v.Len())) +} + +func encInt32Slice(state *encoderState, v reflect.Value) bool { + slice, ok := v.Interface().([]int32) + if !ok { + // It is kind int32 but not type int32. TODO: We can handle this unsafely. + return false + } + for _, x := range slice { + if x != 0 || state.sendZero { + state.encodeInt(int64(x)) + } + } + return true +} + +func encInt64Array(state *encoderState, v reflect.Value) bool { + // Can only slice if it is addressable. + if !v.CanAddr() { + return false + } + return encInt64Slice(state, v.Slice(0, v.Len())) +} + +func encInt64Slice(state *encoderState, v reflect.Value) bool { + slice, ok := v.Interface().([]int64) + if !ok { + // It is kind int64 but not type int64. TODO: We can handle this unsafely. + return false + } + for _, x := range slice { + if x != 0 || state.sendZero { + state.encodeInt(x) + } + } + return true +} + +func encInt8Array(state *encoderState, v reflect.Value) bool { + // Can only slice if it is addressable. + if !v.CanAddr() { + return false + } + return encInt8Slice(state, v.Slice(0, v.Len())) +} + +func encInt8Slice(state *encoderState, v reflect.Value) bool { + slice, ok := v.Interface().([]int8) + if !ok { + // It is kind int8 but not type int8. TODO: We can handle this unsafely. + return false + } + for _, x := range slice { + if x != 0 || state.sendZero { + state.encodeInt(int64(x)) + } + } + return true +} + +func encStringArray(state *encoderState, v reflect.Value) bool { + // Can only slice if it is addressable. + if !v.CanAddr() { + return false + } + return encStringSlice(state, v.Slice(0, v.Len())) +} + +func encStringSlice(state *encoderState, v reflect.Value) bool { + slice, ok := v.Interface().([]string) + if !ok { + // It is kind string but not type string. TODO: We can handle this unsafely. + return false + } + for _, x := range slice { + if x != "" || state.sendZero { + state.encodeUint(uint64(len(x))) + state.b.WriteString(x) + } + } + return true +} + +func encUintArray(state *encoderState, v reflect.Value) bool { + // Can only slice if it is addressable. + if !v.CanAddr() { + return false + } + return encUintSlice(state, v.Slice(0, v.Len())) +} + +func encUintSlice(state *encoderState, v reflect.Value) bool { + slice, ok := v.Interface().([]uint) + if !ok { + // It is kind uint but not type uint. TODO: We can handle this unsafely. + return false + } + for _, x := range slice { + if x != 0 || state.sendZero { + state.encodeUint(uint64(x)) + } + } + return true +} + +func encUint16Array(state *encoderState, v reflect.Value) bool { + // Can only slice if it is addressable. + if !v.CanAddr() { + return false + } + return encUint16Slice(state, v.Slice(0, v.Len())) +} + +func encUint16Slice(state *encoderState, v reflect.Value) bool { + slice, ok := v.Interface().([]uint16) + if !ok { + // It is kind uint16 but not type uint16. TODO: We can handle this unsafely. + return false + } + for _, x := range slice { + if x != 0 || state.sendZero { + state.encodeUint(uint64(x)) + } + } + return true +} + +func encUint32Array(state *encoderState, v reflect.Value) bool { + // Can only slice if it is addressable. + if !v.CanAddr() { + return false + } + return encUint32Slice(state, v.Slice(0, v.Len())) +} + +func encUint32Slice(state *encoderState, v reflect.Value) bool { + slice, ok := v.Interface().([]uint32) + if !ok { + // It is kind uint32 but not type uint32. TODO: We can handle this unsafely. + return false + } + for _, x := range slice { + if x != 0 || state.sendZero { + state.encodeUint(uint64(x)) + } + } + return true +} + +func encUint64Array(state *encoderState, v reflect.Value) bool { + // Can only slice if it is addressable. + if !v.CanAddr() { + return false + } + return encUint64Slice(state, v.Slice(0, v.Len())) +} + +func encUint64Slice(state *encoderState, v reflect.Value) bool { + slice, ok := v.Interface().([]uint64) + if !ok { + // It is kind uint64 but not type uint64. TODO: We can handle this unsafely. + return false + } + for _, x := range slice { + if x != 0 || state.sendZero { + state.encodeUint(x) + } + } + return true +} + +func encUintptrArray(state *encoderState, v reflect.Value) bool { + // Can only slice if it is addressable. + if !v.CanAddr() { + return false + } + return encUintptrSlice(state, v.Slice(0, v.Len())) +} + +func encUintptrSlice(state *encoderState, v reflect.Value) bool { + slice, ok := v.Interface().([]uintptr) + if !ok { + // It is kind uintptr but not type uintptr. TODO: We can handle this unsafely. + return false + } + for _, x := range slice { + if x != 0 || state.sendZero { + state.encodeUint(uint64(x)) + } + } + return true +} diff --git a/src/encoding/gob/encgen.go b/src/encoding/gob/encgen.go new file mode 100644 index 0000000..e5f6878 --- /dev/null +++ b/src/encoding/gob/encgen.go @@ -0,0 +1,220 @@ +// 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. + +//go:build ignore + +// encgen writes the helper functions for encoding. Intended to be +// used with go generate; see the invocation in encode.go. + +// TODO: We could do more by being unsafe. Add a -unsafe flag? + +package main + +import ( + "bytes" + "flag" + "fmt" + "go/format" + "log" + "os" +) + +var output = flag.String("output", "enc_helpers.go", "file name to write") + +type Type struct { + lower string + upper string + zero string + encoder string +} + +var types = []Type{ + { + "bool", + "Bool", + "false", + `if x { + state.encodeUint(1) + } else { + state.encodeUint(0) + }`, + }, + { + "complex64", + "Complex64", + "0+0i", + `rpart := floatBits(float64(real(x))) + ipart := floatBits(float64(imag(x))) + state.encodeUint(rpart) + state.encodeUint(ipart)`, + }, + { + "complex128", + "Complex128", + "0+0i", + `rpart := floatBits(real(x)) + ipart := floatBits(imag(x)) + state.encodeUint(rpart) + state.encodeUint(ipart)`, + }, + { + "float32", + "Float32", + "0", + `bits := floatBits(float64(x)) + state.encodeUint(bits)`, + }, + { + "float64", + "Float64", + "0", + `bits := floatBits(x) + state.encodeUint(bits)`, + }, + { + "int", + "Int", + "0", + `state.encodeInt(int64(x))`, + }, + { + "int16", + "Int16", + "0", + `state.encodeInt(int64(x))`, + }, + { + "int32", + "Int32", + "0", + `state.encodeInt(int64(x))`, + }, + { + "int64", + "Int64", + "0", + `state.encodeInt(x)`, + }, + { + "int8", + "Int8", + "0", + `state.encodeInt(int64(x))`, + }, + { + "string", + "String", + `""`, + `state.encodeUint(uint64(len(x))) + state.b.WriteString(x)`, + }, + { + "uint", + "Uint", + "0", + `state.encodeUint(uint64(x))`, + }, + { + "uint16", + "Uint16", + "0", + `state.encodeUint(uint64(x))`, + }, + { + "uint32", + "Uint32", + "0", + `state.encodeUint(uint64(x))`, + }, + { + "uint64", + "Uint64", + "0", + `state.encodeUint(x)`, + }, + { + "uintptr", + "Uintptr", + "0", + `state.encodeUint(uint64(x))`, + }, + // uint8 Handled separately. +} + +func main() { + log.SetFlags(0) + log.SetPrefix("encgen: ") + flag.Parse() + if flag.NArg() != 0 { + log.Fatal("usage: encgen [--output filename]") + } + var b bytes.Buffer + fmt.Fprintf(&b, "// Code generated by go run encgen.go -output %s; DO NOT EDIT.\n", *output) + fmt.Fprint(&b, header) + printMaps(&b, "Array") + fmt.Fprint(&b, "\n") + printMaps(&b, "Slice") + for _, t := range types { + fmt.Fprintf(&b, arrayHelper, t.lower, t.upper) + fmt.Fprintf(&b, sliceHelper, t.lower, t.upper, t.zero, t.encoder) + } + source, err := format.Source(b.Bytes()) + if err != nil { + log.Fatal("source format error:", err) + } + fd, err := os.Create(*output) + if err != nil { + log.Fatal(err) + } + if _, err := fd.Write(source); err != nil { + log.Fatal(err) + } +} + +func printMaps(b *bytes.Buffer, upperClass string) { + fmt.Fprintf(b, "var enc%sHelper = map[reflect.Kind]encHelper{\n", upperClass) + for _, t := range types { + fmt.Fprintf(b, "reflect.%s: enc%s%s,\n", t.upper, t.upper, upperClass) + } + fmt.Fprintf(b, "}\n") +} + +const header = ` +// Copyright 2014 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 gob + +import ( + "reflect" +) + +` + +const arrayHelper = ` +func enc%[2]sArray(state *encoderState, v reflect.Value) bool { + // Can only slice if it is addressable. + if !v.CanAddr() { + return false + } + return enc%[2]sSlice(state, v.Slice(0, v.Len())) +} +` + +const sliceHelper = ` +func enc%[2]sSlice(state *encoderState, v reflect.Value) bool { + slice, ok := v.Interface().([]%[1]s) + if !ok { + // It is kind %[1]s but not type %[1]s. TODO: We can handle this unsafely. + return false + } + for _, x := range slice { + if x != %[3]s || state.sendZero { + %[4]s + } + } + return true +} +` diff --git a/src/encoding/gob/encode.go b/src/encoding/gob/encode.go new file mode 100644 index 0000000..3843034 --- /dev/null +++ b/src/encoding/gob/encode.go @@ -0,0 +1,705 @@ +// 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. + +//go:generate go run encgen.go -output enc_helpers.go + +package gob + +import ( + "encoding" + "encoding/binary" + "math" + "math/bits" + "reflect" + "sync" +) + +const uint64Size = 8 + +type encHelper func(state *encoderState, v reflect.Value) bool + +// encoderState is the global execution state of an instance of the encoder. +// Field numbers are delta encoded and always increase. The field +// number is initialized to -1 so 0 comes out as delta(1). A delta of +// 0 terminates the structure. +type encoderState struct { + enc *Encoder + b *encBuffer + sendZero bool // encoding an array element or map key/value pair; send zero values + fieldnum int // the last field number written. + buf [1 + uint64Size]byte // buffer used by the encoder; here to avoid allocation. + next *encoderState // for free list +} + +// encBuffer is an extremely simple, fast implementation of a write-only byte buffer. +// It never returns a non-nil error, but Write returns an error value so it matches io.Writer. +type encBuffer struct { + data []byte + scratch [64]byte +} + +var encBufferPool = sync.Pool{ + New: func() any { + e := new(encBuffer) + e.data = e.scratch[0:0] + return e + }, +} + +func (e *encBuffer) writeByte(c byte) { + e.data = append(e.data, c) +} + +func (e *encBuffer) Write(p []byte) (int, error) { + e.data = append(e.data, p...) + return len(p), nil +} + +func (e *encBuffer) WriteString(s string) { + e.data = append(e.data, s...) +} + +func (e *encBuffer) Len() int { + return len(e.data) +} + +func (e *encBuffer) Bytes() []byte { + return e.data +} + +func (e *encBuffer) Reset() { + if len(e.data) >= tooBig { + e.data = e.scratch[0:0] + } else { + e.data = e.data[0:0] + } +} + +func (enc *Encoder) newEncoderState(b *encBuffer) *encoderState { + e := enc.freeList + if e == nil { + e = new(encoderState) + e.enc = enc + } else { + enc.freeList = e.next + } + e.sendZero = false + e.fieldnum = 0 + e.b = b + if len(b.data) == 0 { + b.data = b.scratch[0:0] + } + return e +} + +func (enc *Encoder) freeEncoderState(e *encoderState) { + e.next = enc.freeList + enc.freeList = e +} + +// Unsigned integers have a two-state encoding. If the number is less +// than 128 (0 through 0x7F), its value is written directly. +// Otherwise the value is written in big-endian byte order preceded +// by the byte length, negated. + +// encodeUint writes an encoded unsigned integer to state.b. +func (state *encoderState) encodeUint(x uint64) { + if x <= 0x7F { + state.b.writeByte(uint8(x)) + return + } + + binary.BigEndian.PutUint64(state.buf[1:], x) + bc := bits.LeadingZeros64(x) >> 3 // 8 - bytelen(x) + state.buf[bc] = uint8(bc - uint64Size) // and then we subtract 8 to get -bytelen(x) + + state.b.Write(state.buf[bc : uint64Size+1]) +} + +// encodeInt writes an encoded signed integer to state.w. +// The low bit of the encoding says whether to bit complement the (other bits of the) +// uint to recover the int. +func (state *encoderState) encodeInt(i int64) { + var x uint64 + if i < 0 { + x = uint64(^i<<1) | 1 + } else { + x = uint64(i << 1) + } + state.encodeUint(x) +} + +// encOp is the signature of an encoding operator for a given type. +type encOp func(i *encInstr, state *encoderState, v reflect.Value) + +// The 'instructions' of the encoding machine +type encInstr struct { + op encOp + field int // field number in input + index []int // struct index + indir int // how many pointer indirections to reach the value in the struct +} + +// update emits a field number and updates the state to record its value for delta encoding. +// If the instruction pointer is nil, it does nothing +func (state *encoderState) update(instr *encInstr) { + if instr != nil { + state.encodeUint(uint64(instr.field - state.fieldnum)) + state.fieldnum = instr.field + } +} + +// Each encoder for a composite is responsible for handling any +// indirections associated with the elements of the data structure. +// If any pointer so reached is nil, no bytes are written. If the +// data item is zero, no bytes are written. Single values - ints, +// strings etc. - are indirected before calling their encoders. +// Otherwise, the output (for a scalar) is the field number, as an +// encoded integer, followed by the field data in its appropriate +// format. + +// encIndirect dereferences pv indir times and returns the result. +func encIndirect(pv reflect.Value, indir int) reflect.Value { + for ; indir > 0; indir-- { + if pv.IsNil() { + break + } + pv = pv.Elem() + } + return pv +} + +// encBool encodes the bool referenced by v as an unsigned 0 or 1. +func encBool(i *encInstr, state *encoderState, v reflect.Value) { + b := v.Bool() + if b || state.sendZero { + state.update(i) + if b { + state.encodeUint(1) + } else { + state.encodeUint(0) + } + } +} + +// encInt encodes the signed integer (int int8 int16 int32 int64) referenced by v. +func encInt(i *encInstr, state *encoderState, v reflect.Value) { + value := v.Int() + if value != 0 || state.sendZero { + state.update(i) + state.encodeInt(value) + } +} + +// encUint encodes the unsigned integer (uint uint8 uint16 uint32 uint64 uintptr) referenced by v. +func encUint(i *encInstr, state *encoderState, v reflect.Value) { + value := v.Uint() + if value != 0 || state.sendZero { + state.update(i) + state.encodeUint(value) + } +} + +// floatBits returns a uint64 holding the bits of a floating-point number. +// Floating-point numbers are transmitted as uint64s holding the bits +// of the underlying representation. They are sent byte-reversed, with +// the exponent end coming out first, so integer floating point numbers +// (for example) transmit more compactly. This routine does the +// swizzling. +func floatBits(f float64) uint64 { + u := math.Float64bits(f) + return bits.ReverseBytes64(u) +} + +// encFloat encodes the floating point value (float32 float64) referenced by v. +func encFloat(i *encInstr, state *encoderState, v reflect.Value) { + f := v.Float() + if f != 0 || state.sendZero { + bits := floatBits(f) + state.update(i) + state.encodeUint(bits) + } +} + +// encComplex encodes the complex value (complex64 complex128) referenced by v. +// Complex numbers are just a pair of floating-point numbers, real part first. +func encComplex(i *encInstr, state *encoderState, v reflect.Value) { + c := v.Complex() + if c != 0+0i || state.sendZero { + rpart := floatBits(real(c)) + ipart := floatBits(imag(c)) + state.update(i) + state.encodeUint(rpart) + state.encodeUint(ipart) + } +} + +// encUint8Array encodes the byte array referenced by v. +// Byte arrays are encoded as an unsigned count followed by the raw bytes. +func encUint8Array(i *encInstr, state *encoderState, v reflect.Value) { + b := v.Bytes() + if len(b) > 0 || state.sendZero { + state.update(i) + state.encodeUint(uint64(len(b))) + state.b.Write(b) + } +} + +// encString encodes the string referenced by v. +// Strings are encoded as an unsigned count followed by the raw bytes. +func encString(i *encInstr, state *encoderState, v reflect.Value) { + s := v.String() + if len(s) > 0 || state.sendZero { + state.update(i) + state.encodeUint(uint64(len(s))) + state.b.WriteString(s) + } +} + +// encStructTerminator encodes the end of an encoded struct +// as delta field number of 0. +func encStructTerminator(i *encInstr, state *encoderState, v reflect.Value) { + state.encodeUint(0) +} + +// Execution engine + +// encEngine an array of instructions indexed by field number of the encoding +// data, typically a struct. It is executed top to bottom, walking the struct. +type encEngine struct { + instr []encInstr +} + +const singletonField = 0 + +// valid reports whether the value is valid and a non-nil pointer. +// (Slices, maps, and chans take care of themselves.) +func valid(v reflect.Value) bool { + switch v.Kind() { + case reflect.Invalid: + return false + case reflect.Pointer: + return !v.IsNil() + } + return true +} + +// encodeSingle encodes a single top-level non-struct value. +func (enc *Encoder) encodeSingle(b *encBuffer, engine *encEngine, value reflect.Value) { + state := enc.newEncoderState(b) + defer enc.freeEncoderState(state) + state.fieldnum = singletonField + // There is no surrounding struct to frame the transmission, so we must + // generate data even if the item is zero. To do this, set sendZero. + state.sendZero = true + instr := &engine.instr[singletonField] + if instr.indir > 0 { + value = encIndirect(value, instr.indir) + } + if valid(value) { + instr.op(instr, state, value) + } +} + +// encodeStruct encodes a single struct value. +func (enc *Encoder) encodeStruct(b *encBuffer, engine *encEngine, value reflect.Value) { + if !valid(value) { + return + } + state := enc.newEncoderState(b) + defer enc.freeEncoderState(state) + state.fieldnum = -1 + for i := 0; i < len(engine.instr); i++ { + instr := &engine.instr[i] + if i >= value.NumField() { + // encStructTerminator + instr.op(instr, state, reflect.Value{}) + break + } + field := value.FieldByIndex(instr.index) + if instr.indir > 0 { + field = encIndirect(field, instr.indir) + // TODO: Is field guaranteed valid? If so we could avoid this check. + if !valid(field) { + continue + } + } + instr.op(instr, state, field) + } +} + +// encodeArray encodes an array. +func (enc *Encoder) encodeArray(b *encBuffer, value reflect.Value, op encOp, elemIndir int, length int, helper encHelper) { + state := enc.newEncoderState(b) + defer enc.freeEncoderState(state) + state.fieldnum = -1 + state.sendZero = true + state.encodeUint(uint64(length)) + if helper != nil && helper(state, value) { + return + } + for i := 0; i < length; i++ { + elem := value.Index(i) + if elemIndir > 0 { + elem = encIndirect(elem, elemIndir) + // TODO: Is elem guaranteed valid? If so we could avoid this check. + if !valid(elem) { + errorf("encodeArray: nil element") + } + } + op(nil, state, elem) + } +} + +// encodeReflectValue is a helper for maps. It encodes the value v. +func encodeReflectValue(state *encoderState, v reflect.Value, op encOp, indir int) { + for i := 0; i < indir && v.IsValid(); i++ { + v = reflect.Indirect(v) + } + if !v.IsValid() { + errorf("encodeReflectValue: nil element") + } + op(nil, state, v) +} + +// encodeMap encodes a map as unsigned count followed by key:value pairs. +func (enc *Encoder) encodeMap(b *encBuffer, mv reflect.Value, keyOp, elemOp encOp, keyIndir, elemIndir int) { + state := enc.newEncoderState(b) + state.fieldnum = -1 + state.sendZero = true + state.encodeUint(uint64(mv.Len())) + mi := mv.MapRange() + for mi.Next() { + encodeReflectValue(state, mi.Key(), keyOp, keyIndir) + encodeReflectValue(state, mi.Value(), elemOp, elemIndir) + } + enc.freeEncoderState(state) +} + +// encodeInterface encodes the interface value iv. +// To send an interface, we send a string identifying the concrete type, followed +// by the type identifier (which might require defining that type right now), followed +// by the concrete value. A nil value gets sent as the empty string for the name, +// followed by no value. +func (enc *Encoder) encodeInterface(b *encBuffer, iv reflect.Value) { + // Gobs can encode nil interface values but not typed interface + // values holding nil pointers, since nil pointers point to no value. + elem := iv.Elem() + if elem.Kind() == reflect.Pointer && elem.IsNil() { + errorf("gob: cannot encode nil pointer of type %s inside interface", iv.Elem().Type()) + } + state := enc.newEncoderState(b) + state.fieldnum = -1 + state.sendZero = true + if iv.IsNil() { + state.encodeUint(0) + return + } + + ut := userType(iv.Elem().Type()) + namei, ok := concreteTypeToName.Load(ut.base) + if !ok { + errorf("type not registered for interface: %s", ut.base) + } + name := namei.(string) + + // Send the name. + state.encodeUint(uint64(len(name))) + state.b.WriteString(name) + // Define the type id if necessary. + enc.sendTypeDescriptor(enc.writer(), state, ut) + // Send the type id. + enc.sendTypeId(state, ut) + // Encode the value into a new buffer. Any nested type definitions + // should be written to b, before the encoded value. + enc.pushWriter(b) + data := encBufferPool.Get().(*encBuffer) + data.Write(spaceForLength) + enc.encode(data, elem, ut) + if enc.err != nil { + error_(enc.err) + } + enc.popWriter() + enc.writeMessage(b, data) + data.Reset() + encBufferPool.Put(data) + if enc.err != nil { + error_(enc.err) + } + enc.freeEncoderState(state) +} + +// isZero reports whether the value is the zero of its type. +func isZero(val reflect.Value) bool { + switch val.Kind() { + case reflect.Array: + for i := 0; i < val.Len(); i++ { + if !isZero(val.Index(i)) { + return false + } + } + return true + case reflect.Map, reflect.Slice, reflect.String: + return val.Len() == 0 + case reflect.Bool: + return !val.Bool() + case reflect.Complex64, reflect.Complex128: + return val.Complex() == 0 + case reflect.Chan, reflect.Func, reflect.Interface, reflect.Pointer: + return val.IsNil() + case reflect.Int, reflect.Int8, reflect.Int16, reflect.Int32, reflect.Int64: + return val.Int() == 0 + case reflect.Float32, reflect.Float64: + return val.Float() == 0 + case reflect.Uint, reflect.Uint8, reflect.Uint16, reflect.Uint32, reflect.Uint64, reflect.Uintptr: + return val.Uint() == 0 + case reflect.Struct: + for i := 0; i < val.NumField(); i++ { + if !isZero(val.Field(i)) { + return false + } + } + return true + } + panic("unknown type in isZero " + val.Type().String()) +} + +// encGobEncoder encodes a value that implements the GobEncoder interface. +// The data is sent as a byte array. +func (enc *Encoder) encodeGobEncoder(b *encBuffer, ut *userTypeInfo, v reflect.Value) { + // TODO: should we catch panics from the called method? + + var data []byte + var err error + // We know it's one of these. + switch ut.externalEnc { + case xGob: + data, err = v.Interface().(GobEncoder).GobEncode() + case xBinary: + data, err = v.Interface().(encoding.BinaryMarshaler).MarshalBinary() + case xText: + data, err = v.Interface().(encoding.TextMarshaler).MarshalText() + } + if err != nil { + error_(err) + } + state := enc.newEncoderState(b) + state.fieldnum = -1 + state.encodeUint(uint64(len(data))) + state.b.Write(data) + enc.freeEncoderState(state) +} + +var encOpTable = [...]encOp{ + reflect.Bool: encBool, + reflect.Int: encInt, + reflect.Int8: encInt, + reflect.Int16: encInt, + reflect.Int32: encInt, + reflect.Int64: encInt, + reflect.Uint: encUint, + reflect.Uint8: encUint, + reflect.Uint16: encUint, + reflect.Uint32: encUint, + reflect.Uint64: encUint, + reflect.Uintptr: encUint, + reflect.Float32: encFloat, + reflect.Float64: encFloat, + reflect.Complex64: encComplex, + reflect.Complex128: encComplex, + reflect.String: encString, +} + +// encOpFor returns (a pointer to) the encoding op for the base type under rt and +// the indirection count to reach it. +func encOpFor(rt reflect.Type, inProgress map[reflect.Type]*encOp, building map[*typeInfo]bool) (*encOp, int) { + ut := userType(rt) + // If the type implements GobEncoder, we handle it without further processing. + if ut.externalEnc != 0 { + return gobEncodeOpFor(ut) + } + // If this type is already in progress, it's a recursive type (e.g. map[string]*T). + // Return the pointer to the op we're already building. + if opPtr := inProgress[rt]; opPtr != nil { + return opPtr, ut.indir + } + typ := ut.base + indir := ut.indir + k := typ.Kind() + var op encOp + if int(k) < len(encOpTable) { + op = encOpTable[k] + } + if op == nil { + inProgress[rt] = &op + // Special cases + switch t := typ; t.Kind() { + case reflect.Slice: + if t.Elem().Kind() == reflect.Uint8 { + op = encUint8Array + break + } + // Slices have a header; we decode it to find the underlying array. + elemOp, elemIndir := encOpFor(t.Elem(), inProgress, building) + helper := encSliceHelper[t.Elem().Kind()] + op = func(i *encInstr, state *encoderState, slice reflect.Value) { + if !state.sendZero && slice.Len() == 0 { + return + } + state.update(i) + state.enc.encodeArray(state.b, slice, *elemOp, elemIndir, slice.Len(), helper) + } + case reflect.Array: + // True arrays have size in the type. + elemOp, elemIndir := encOpFor(t.Elem(), inProgress, building) + helper := encArrayHelper[t.Elem().Kind()] + op = func(i *encInstr, state *encoderState, array reflect.Value) { + state.update(i) + state.enc.encodeArray(state.b, array, *elemOp, elemIndir, array.Len(), helper) + } + case reflect.Map: + keyOp, keyIndir := encOpFor(t.Key(), inProgress, building) + elemOp, elemIndir := encOpFor(t.Elem(), inProgress, building) + op = func(i *encInstr, state *encoderState, mv reflect.Value) { + // We send zero-length (but non-nil) maps because the + // receiver might want to use the map. (Maps don't use append.) + if !state.sendZero && mv.IsNil() { + return + } + state.update(i) + state.enc.encodeMap(state.b, mv, *keyOp, *elemOp, keyIndir, elemIndir) + } + case reflect.Struct: + // Generate a closure that calls out to the engine for the nested type. + getEncEngine(userType(typ), building) + info := mustGetTypeInfo(typ) + op = func(i *encInstr, state *encoderState, sv reflect.Value) { + state.update(i) + // indirect through info to delay evaluation for recursive structs + enc := info.encoder.Load() + state.enc.encodeStruct(state.b, enc, sv) + } + case reflect.Interface: + op = func(i *encInstr, state *encoderState, iv reflect.Value) { + if !state.sendZero && (!iv.IsValid() || iv.IsNil()) { + return + } + state.update(i) + state.enc.encodeInterface(state.b, iv) + } + } + } + if op == nil { + errorf("can't happen: encode type %s", rt) + } + return &op, indir +} + +// gobEncodeOpFor returns the op for a type that is known to implement GobEncoder. +func gobEncodeOpFor(ut *userTypeInfo) (*encOp, int) { + rt := ut.user + if ut.encIndir == -1 { + rt = reflect.PointerTo(rt) + } else if ut.encIndir > 0 { + for i := int8(0); i < ut.encIndir; i++ { + rt = rt.Elem() + } + } + var op encOp + op = func(i *encInstr, state *encoderState, v reflect.Value) { + if ut.encIndir == -1 { + // Need to climb up one level to turn value into pointer. + if !v.CanAddr() { + errorf("unaddressable value of type %s", rt) + } + v = v.Addr() + } + if !state.sendZero && isZero(v) { + return + } + state.update(i) + state.enc.encodeGobEncoder(state.b, ut, v) + } + return &op, int(ut.encIndir) // encIndir: op will get called with p == address of receiver. +} + +// compileEnc returns the engine to compile the type. +func compileEnc(ut *userTypeInfo, building map[*typeInfo]bool) *encEngine { + srt := ut.base + engine := new(encEngine) + seen := make(map[reflect.Type]*encOp) + rt := ut.base + if ut.externalEnc != 0 { + rt = ut.user + } + if ut.externalEnc == 0 && srt.Kind() == reflect.Struct { + for fieldNum, wireFieldNum := 0, 0; fieldNum < srt.NumField(); fieldNum++ { + f := srt.Field(fieldNum) + if !isSent(&f) { + continue + } + op, indir := encOpFor(f.Type, seen, building) + engine.instr = append(engine.instr, encInstr{*op, wireFieldNum, f.Index, indir}) + wireFieldNum++ + } + if srt.NumField() > 0 && len(engine.instr) == 0 { + errorf("type %s has no exported fields", rt) + } + engine.instr = append(engine.instr, encInstr{encStructTerminator, 0, nil, 0}) + } else { + engine.instr = make([]encInstr, 1) + op, indir := encOpFor(rt, seen, building) + engine.instr[0] = encInstr{*op, singletonField, nil, indir} + } + return engine +} + +// getEncEngine returns the engine to compile the type. +func getEncEngine(ut *userTypeInfo, building map[*typeInfo]bool) *encEngine { + info, err := getTypeInfo(ut) + if err != nil { + error_(err) + } + enc := info.encoder.Load() + if enc == nil { + enc = buildEncEngine(info, ut, building) + } + return enc +} + +func buildEncEngine(info *typeInfo, ut *userTypeInfo, building map[*typeInfo]bool) *encEngine { + // Check for recursive types. + if building != nil && building[info] { + return nil + } + info.encInit.Lock() + defer info.encInit.Unlock() + enc := info.encoder.Load() + if enc == nil { + if building == nil { + building = make(map[*typeInfo]bool) + } + building[info] = true + enc = compileEnc(ut, building) + info.encoder.Store(enc) + } + return enc +} + +func (enc *Encoder) encode(b *encBuffer, value reflect.Value, ut *userTypeInfo) { + defer catchError(&enc.err) + engine := getEncEngine(ut, nil) + indir := ut.indir + if ut.externalEnc != 0 { + indir = int(ut.encIndir) + } + for i := 0; i < indir; i++ { + value = reflect.Indirect(value) + } + if ut.externalEnc == 0 && value.Type().Kind() == reflect.Struct { + enc.encodeStruct(b, engine, value) + } else { + enc.encodeSingle(b, engine, value) + } +} diff --git a/src/encoding/gob/encoder.go b/src/encoding/gob/encoder.go new file mode 100644 index 0000000..5a80e6c --- /dev/null +++ b/src/encoding/gob/encoder.go @@ -0,0 +1,258 @@ +// 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 gob + +import ( + "errors" + "io" + "reflect" + "sync" +) + +// An Encoder manages the transmission of type and data information to the +// other side of a connection. It is safe for concurrent use by multiple +// goroutines. +type Encoder struct { + mutex sync.Mutex // each item must be sent atomically + w []io.Writer // where to send the data + sent map[reflect.Type]typeId // which types we've already sent + countState *encoderState // stage for writing counts + freeList *encoderState // list of free encoderStates; avoids reallocation + byteBuf encBuffer // buffer for top-level encoderState + err error +} + +// Before we encode a message, we reserve space at the head of the +// buffer in which to encode its length. This means we can use the +// buffer to assemble the message without another allocation. +const maxLength = 9 // Maximum size of an encoded length. +var spaceForLength = make([]byte, maxLength) + +// NewEncoder returns a new encoder that will transmit on the io.Writer. +func NewEncoder(w io.Writer) *Encoder { + enc := new(Encoder) + enc.w = []io.Writer{w} + enc.sent = make(map[reflect.Type]typeId) + enc.countState = enc.newEncoderState(new(encBuffer)) + return enc +} + +// writer() returns the innermost writer the encoder is using +func (enc *Encoder) writer() io.Writer { + return enc.w[len(enc.w)-1] +} + +// pushWriter adds a writer to the encoder. +func (enc *Encoder) pushWriter(w io.Writer) { + enc.w = append(enc.w, w) +} + +// popWriter pops the innermost writer. +func (enc *Encoder) popWriter() { + enc.w = enc.w[0 : len(enc.w)-1] +} + +func (enc *Encoder) setError(err error) { + if enc.err == nil { // remember the first. + enc.err = err + } +} + +// writeMessage sends the data item preceded by a unsigned count of its length. +func (enc *Encoder) writeMessage(w io.Writer, b *encBuffer) { + // Space has been reserved for the length at the head of the message. + // This is a little dirty: we grab the slice from the bytes.Buffer and massage + // it by hand. + message := b.Bytes() + messageLen := len(message) - maxLength + // Length cannot be bigger than the decoder can handle. + if messageLen >= tooBig { + enc.setError(errors.New("gob: encoder: message too big")) + return + } + // Encode the length. + enc.countState.b.Reset() + enc.countState.encodeUint(uint64(messageLen)) + // Copy the length to be a prefix of the message. + offset := maxLength - enc.countState.b.Len() + copy(message[offset:], enc.countState.b.Bytes()) + // Write the data. + _, err := w.Write(message[offset:]) + // Drain the buffer and restore the space at the front for the count of the next message. + b.Reset() + b.Write(spaceForLength) + if err != nil { + enc.setError(err) + } +} + +// sendActualType sends the requested type, without further investigation, unless +// it's been sent before. +func (enc *Encoder) sendActualType(w io.Writer, state *encoderState, ut *userTypeInfo, actual reflect.Type) (sent bool) { + if _, alreadySent := enc.sent[actual]; alreadySent { + return false + } + info, err := getTypeInfo(ut) + if err != nil { + enc.setError(err) + return + } + // Send the pair (-id, type) + // Id: + state.encodeInt(-int64(info.id)) + // Type: + enc.encode(state.b, reflect.ValueOf(info.wire), wireTypeUserInfo) + enc.writeMessage(w, state.b) + if enc.err != nil { + return + } + + // Remember we've sent this type, both what the user gave us and the base type. + enc.sent[ut.base] = info.id + if ut.user != ut.base { + enc.sent[ut.user] = info.id + } + // Now send the inner types + switch st := actual; st.Kind() { + case reflect.Struct: + for i := 0; i < st.NumField(); i++ { + if isExported(st.Field(i).Name) { + enc.sendType(w, state, st.Field(i).Type) + } + } + case reflect.Array, reflect.Slice: + enc.sendType(w, state, st.Elem()) + case reflect.Map: + enc.sendType(w, state, st.Key()) + enc.sendType(w, state, st.Elem()) + } + return true +} + +// sendType sends the type info to the other side, if necessary. +func (enc *Encoder) sendType(w io.Writer, state *encoderState, origt reflect.Type) (sent bool) { + ut := userType(origt) + if ut.externalEnc != 0 { + // The rules are different: regardless of the underlying type's representation, + // we need to tell the other side that the base type is a GobEncoder. + return enc.sendActualType(w, state, ut, ut.base) + } + + // It's a concrete value, so drill down to the base type. + switch rt := ut.base; rt.Kind() { + default: + // Basic types and interfaces do not need to be described. + return + case reflect.Slice: + // If it's []uint8, don't send; it's considered basic. + if rt.Elem().Kind() == reflect.Uint8 { + return + } + // Otherwise we do send. + break + case reflect.Array: + // arrays must be sent so we know their lengths and element types. + break + case reflect.Map: + // maps must be sent so we know their lengths and key/value types. + break + case reflect.Struct: + // structs must be sent so we know their fields. + break + case reflect.Chan, reflect.Func: + // If we get here, it's a field of a struct; ignore it. + return + } + + return enc.sendActualType(w, state, ut, ut.base) +} + +// Encode transmits the data item represented by the empty interface value, +// guaranteeing that all necessary type information has been transmitted first. +// Passing a nil pointer to Encoder will panic, as they cannot be transmitted by gob. +func (enc *Encoder) Encode(e any) error { + return enc.EncodeValue(reflect.ValueOf(e)) +} + +// sendTypeDescriptor makes sure the remote side knows about this type. +// It will send a descriptor if this is the first time the type has been +// sent. +func (enc *Encoder) sendTypeDescriptor(w io.Writer, state *encoderState, ut *userTypeInfo) { + // Make sure the type is known to the other side. + // First, have we already sent this type? + rt := ut.base + if ut.externalEnc != 0 { + rt = ut.user + } + if _, alreadySent := enc.sent[rt]; !alreadySent { + // No, so send it. + sent := enc.sendType(w, state, rt) + if enc.err != nil { + return + } + // If the type info has still not been transmitted, it means we have + // a singleton basic type (int, []byte etc.) at top level. We don't + // need to send the type info but we do need to update enc.sent. + if !sent { + info, err := getTypeInfo(ut) + if err != nil { + enc.setError(err) + return + } + enc.sent[rt] = info.id + } + } +} + +// sendTypeId sends the id, which must have already been defined. +func (enc *Encoder) sendTypeId(state *encoderState, ut *userTypeInfo) { + // Identify the type of this top-level value. + state.encodeInt(int64(enc.sent[ut.base])) +} + +// EncodeValue transmits the data item represented by the reflection value, +// guaranteeing that all necessary type information has been transmitted first. +// Passing a nil pointer to EncodeValue will panic, as they cannot be transmitted by gob. +func (enc *Encoder) EncodeValue(value reflect.Value) error { + if value.Kind() == reflect.Invalid { + return errors.New("gob: cannot encode nil value") + } + if value.Kind() == reflect.Pointer && value.IsNil() { + panic("gob: cannot encode nil pointer of type " + value.Type().String()) + } + + // Make sure we're single-threaded through here, so multiple + // goroutines can share an encoder. + enc.mutex.Lock() + defer enc.mutex.Unlock() + + // Remove any nested writers remaining due to previous errors. + enc.w = enc.w[0:1] + + ut, err := validUserType(value.Type()) + if err != nil { + return err + } + + enc.err = nil + enc.byteBuf.Reset() + enc.byteBuf.Write(spaceForLength) + state := enc.newEncoderState(&enc.byteBuf) + + enc.sendTypeDescriptor(enc.writer(), state, ut) + enc.sendTypeId(state, ut) + if enc.err != nil { + return enc.err + } + + // Encode the object. + enc.encode(state.b, value, ut) + if enc.err == nil { + enc.writeMessage(enc.writer(), state.b) + } + + enc.freeEncoderState(state) + return enc.err +} diff --git a/src/encoding/gob/encoder_test.go b/src/encoding/gob/encoder_test.go new file mode 100644 index 0000000..484be43 --- /dev/null +++ b/src/encoding/gob/encoder_test.go @@ -0,0 +1,1280 @@ +// 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 gob + +import ( + "bytes" + "encoding/hex" + "fmt" + "io" + "math" + "reflect" + "sort" + "strings" + "testing" +) + +// Test basic operations in a safe manner. +func TestBasicEncoderDecoder(t *testing.T) { + var values = []any{ + true, + int(123), + int8(123), + int16(-12345), + int32(123456), + int64(-1234567), + uint(123), + uint8(123), + uint16(12345), + uint32(123456), + uint64(1234567), + uintptr(12345678), + float32(1.2345), + float64(1.2345678), + complex64(1.2345 + 2.3456i), + complex128(1.2345678 + 2.3456789i), + []byte("hello"), + string("hello"), + } + for _, value := range values { + b := new(bytes.Buffer) + enc := NewEncoder(b) + err := enc.Encode(value) + if err != nil { + t.Error("encoder fail:", err) + } + dec := NewDecoder(b) + result := reflect.New(reflect.TypeOf(value)) + err = dec.Decode(result.Interface()) + if err != nil { + t.Fatalf("error decoding %T: %v:", reflect.TypeOf(value), err) + } + if !reflect.DeepEqual(value, result.Elem().Interface()) { + t.Fatalf("%T: expected %v got %v", value, value, result.Elem().Interface()) + } + } +} + +func TestEncodeIntSlice(t *testing.T) { + + s8 := []int8{1, 5, 12, 22, 35, 51, 70, 92, 117} + s16 := []int16{145, 176, 210, 247, 287, 330, 376, 425, 477} + s32 := []int32{532, 590, 651, 715, 782, 852, 925, 1001, 1080} + s64 := []int64{1162, 1247, 1335, 1426, 1520, 1617, 1717, 1820, 1926} + + t.Run("int8", func(t *testing.T) { + var sink bytes.Buffer + enc := NewEncoder(&sink) + enc.Encode(s8) + + dec := NewDecoder(&sink) + res := make([]int8, 9) + dec.Decode(&res) + + if !reflect.DeepEqual(s8, res) { + t.Fatalf("EncodeIntSlice: expected %v, got %v", s8, res) + } + }) + + t.Run("int16", func(t *testing.T) { + var sink bytes.Buffer + enc := NewEncoder(&sink) + enc.Encode(s16) + + dec := NewDecoder(&sink) + res := make([]int16, 9) + dec.Decode(&res) + + if !reflect.DeepEqual(s16, res) { + t.Fatalf("EncodeIntSlice: expected %v, got %v", s16, res) + } + }) + + t.Run("int32", func(t *testing.T) { + var sink bytes.Buffer + enc := NewEncoder(&sink) + enc.Encode(s32) + + dec := NewDecoder(&sink) + res := make([]int32, 9) + dec.Decode(&res) + + if !reflect.DeepEqual(s32, res) { + t.Fatalf("EncodeIntSlice: expected %v, got %v", s32, res) + } + }) + + t.Run("int64", func(t *testing.T) { + var sink bytes.Buffer + enc := NewEncoder(&sink) + enc.Encode(s64) + + dec := NewDecoder(&sink) + res := make([]int64, 9) + dec.Decode(&res) + + if !reflect.DeepEqual(s64, res) { + t.Fatalf("EncodeIntSlice: expected %v, got %v", s64, res) + } + }) + +} + +type ET0 struct { + A int + B string +} + +type ET2 struct { + X string +} + +type ET1 struct { + A int + Et2 *ET2 + Next *ET1 +} + +// Like ET1 but with a different name for a field +type ET3 struct { + A int + Et2 *ET2 + DifferentNext *ET1 +} + +// Like ET1 but with a different type for a field +type ET4 struct { + A int + Et2 float64 + Next int +} + +func TestEncoderDecoder(t *testing.T) { + b := new(bytes.Buffer) + enc := NewEncoder(b) + et0 := new(ET0) + et0.A = 7 + et0.B = "gobs of fun" + err := enc.Encode(et0) + if err != nil { + t.Error("encoder fail:", err) + } + //fmt.Printf("% x %q\n", b, b) + //Debug(b) + dec := NewDecoder(b) + newEt0 := new(ET0) + err = dec.Decode(newEt0) + if err != nil { + t.Fatal("error decoding ET0:", err) + } + + if !reflect.DeepEqual(et0, newEt0) { + t.Fatalf("invalid data for et0: expected %+v; got %+v", *et0, *newEt0) + } + if b.Len() != 0 { + t.Error("not at eof;", b.Len(), "bytes left") + } + // t.FailNow() + + b = new(bytes.Buffer) + enc = NewEncoder(b) + et1 := new(ET1) + et1.A = 7 + et1.Et2 = new(ET2) + err = enc.Encode(et1) + if err != nil { + t.Error("encoder fail:", err) + } + dec = NewDecoder(b) + newEt1 := new(ET1) + err = dec.Decode(newEt1) + if err != nil { + t.Fatal("error decoding ET1:", err) + } + + if !reflect.DeepEqual(et1, newEt1) { + t.Fatalf("invalid data for et1: expected %+v; got %+v", *et1, *newEt1) + } + if b.Len() != 0 { + t.Error("not at eof;", b.Len(), "bytes left") + } + + enc.Encode(et1) + newEt1 = new(ET1) + err = dec.Decode(newEt1) + if err != nil { + t.Fatal("round 2: error decoding ET1:", err) + } + if !reflect.DeepEqual(et1, newEt1) { + t.Fatalf("round 2: invalid data for et1: expected %+v; got %+v", *et1, *newEt1) + } + if b.Len() != 0 { + t.Error("round 2: not at eof;", b.Len(), "bytes left") + } + + // Now test with a running encoder/decoder pair that we recognize a type mismatch. + err = enc.Encode(et1) + if err != nil { + t.Error("round 3: encoder fail:", err) + } + newEt2 := new(ET2) + err = dec.Decode(newEt2) + if err == nil { + t.Fatal("round 3: expected `bad type' error decoding ET2") + } +} + +// Run one value through the encoder/decoder, but use the wrong type. +// Input is always an ET1; we compare it to whatever is under 'e'. +func badTypeCheck(e any, shouldFail bool, msg string, t *testing.T) { + b := new(bytes.Buffer) + enc := NewEncoder(b) + et1 := new(ET1) + et1.A = 7 + et1.Et2 = new(ET2) + err := enc.Encode(et1) + if err != nil { + t.Error("encoder fail:", err) + } + dec := NewDecoder(b) + err = dec.Decode(e) + if shouldFail && err == nil { + t.Error("expected error for", msg) + } + if !shouldFail && err != nil { + t.Error("unexpected error for", msg, err) + } +} + +// Test that we recognize a bad type the first time. +func TestWrongTypeDecoder(t *testing.T) { + badTypeCheck(new(ET2), true, "no fields in common", t) + badTypeCheck(new(ET3), false, "different name of field", t) + badTypeCheck(new(ET4), true, "different type of field", t) +} + +// Types not supported at top level by the Encoder. +var unsupportedValues = []any{ + make(chan int), + func(a int) bool { return true }, +} + +func TestUnsupported(t *testing.T) { + var b bytes.Buffer + enc := NewEncoder(&b) + for _, v := range unsupportedValues { + err := enc.Encode(v) + if err == nil { + t.Errorf("expected error for %T; got none", v) + } + } +} + +func encAndDec(in, out any) error { + b := new(bytes.Buffer) + enc := NewEncoder(b) + err := enc.Encode(in) + if err != nil { + return err + } + dec := NewDecoder(b) + err = dec.Decode(out) + if err != nil { + return err + } + return nil +} + +func TestTypeToPtrType(t *testing.T) { + // Encode a T, decode a *T + type Type0 struct { + A int + } + t0 := Type0{7} + t0p := new(Type0) + if err := encAndDec(t0, t0p); err != nil { + t.Error(err) + } +} + +func TestPtrTypeToType(t *testing.T) { + // Encode a *T, decode a T + type Type1 struct { + A uint + } + t1p := &Type1{17} + var t1 Type1 + if err := encAndDec(t1, t1p); err != nil { + t.Error(err) + } +} + +func TestTypeToPtrPtrPtrPtrType(t *testing.T) { + type Type2 struct { + A ****float64 + } + t2 := Type2{} + t2.A = new(***float64) + *t2.A = new(**float64) + **t2.A = new(*float64) + ***t2.A = new(float64) + ****t2.A = 27.4 + t2pppp := new(***Type2) + if err := encAndDec(t2, t2pppp); err != nil { + t.Fatal(err) + } + if ****(****t2pppp).A != ****t2.A { + t.Errorf("wrong value after decode: %g not %g", ****(****t2pppp).A, ****t2.A) + } +} + +func TestSlice(t *testing.T) { + type Type3 struct { + A []string + } + t3p := &Type3{[]string{"hello", "world"}} + var t3 Type3 + if err := encAndDec(t3, t3p); err != nil { + t.Error(err) + } +} + +func TestValueError(t *testing.T) { + // Encode a *T, decode a T + type Type4 struct { + A int + } + t4p := &Type4{3} + var t4 Type4 // note: not a pointer. + if err := encAndDec(t4p, t4); err == nil || !strings.Contains(err.Error(), "pointer") { + t.Error("expected error about pointer; got", err) + } +} + +func TestArray(t *testing.T) { + type Type5 struct { + A [3]string + B [3]byte + } + type Type6 struct { + A [2]string // can't hold t5.a + } + t5 := Type5{[3]string{"hello", ",", "world"}, [3]byte{1, 2, 3}} + var t5p Type5 + if err := encAndDec(t5, &t5p); err != nil { + t.Error(err) + } + var t6 Type6 + if err := encAndDec(t5, &t6); err == nil { + t.Error("should fail with mismatched array sizes") + } +} + +func TestRecursiveMapType(t *testing.T) { + type recursiveMap map[string]recursiveMap + r1 := recursiveMap{"A": recursiveMap{"B": nil, "C": nil}, "D": nil} + r2 := make(recursiveMap) + if err := encAndDec(r1, &r2); err != nil { + t.Error(err) + } +} + +func TestRecursiveSliceType(t *testing.T) { + type recursiveSlice []recursiveSlice + r1 := recursiveSlice{0: recursiveSlice{0: nil}, 1: nil} + r2 := make(recursiveSlice, 0) + if err := encAndDec(r1, &r2); err != nil { + t.Error(err) + } +} + +// Regression test for bug: must send zero values inside arrays +func TestDefaultsInArray(t *testing.T) { + type Type7 struct { + B []bool + I []int + S []string + F []float64 + } + t7 := Type7{ + []bool{false, false, true}, + []int{0, 0, 1}, + []string{"hi", "", "there"}, + []float64{0, 0, 1}, + } + var t7p Type7 + if err := encAndDec(t7, &t7p); err != nil { + t.Error(err) + } +} + +var testInt int +var testFloat32 float32 +var testString string +var testSlice []string +var testMap map[string]int +var testArray [7]int + +type SingleTest struct { + in any + out any + err string +} + +var singleTests = []SingleTest{ + {17, &testInt, ""}, + {float32(17.5), &testFloat32, ""}, + {"bike shed", &testString, ""}, + {[]string{"bike", "shed", "paint", "color"}, &testSlice, ""}, + {map[string]int{"seven": 7, "twelve": 12}, &testMap, ""}, + {[7]int{4, 55, 0, 0, 0, 0, 0}, &testArray, ""}, // case that once triggered a bug + {[7]int{4, 55, 1, 44, 22, 66, 1234}, &testArray, ""}, + + // Decode errors + {172, &testFloat32, "type"}, +} + +func TestSingletons(t *testing.T) { + b := new(bytes.Buffer) + enc := NewEncoder(b) + dec := NewDecoder(b) + for _, test := range singleTests { + b.Reset() + err := enc.Encode(test.in) + if err != nil { + t.Errorf("error encoding %v: %s", test.in, err) + continue + } + err = dec.Decode(test.out) + switch { + case err != nil && test.err == "": + t.Errorf("error decoding %v: %s", test.in, err) + continue + case err == nil && test.err != "": + t.Errorf("expected error decoding %v: %s", test.in, test.err) + continue + case err != nil && test.err != "": + if !strings.Contains(err.Error(), test.err) { + t.Errorf("wrong error decoding %v: wanted %s, got %v", test.in, test.err, err) + } + continue + } + // Get rid of the pointer in the rhs + val := reflect.ValueOf(test.out).Elem().Interface() + if !reflect.DeepEqual(test.in, val) { + t.Errorf("decoding singleton: expected %v got %v", test.in, val) + } + } +} + +func TestStructNonStruct(t *testing.T) { + type Struct struct { + A string + } + type NonStruct string + s := Struct{"hello"} + var sp Struct + if err := encAndDec(s, &sp); err != nil { + t.Error(err) + } + var ns NonStruct + if err := encAndDec(s, &ns); err == nil { + t.Error("should get error for struct/non-struct") + } else if !strings.Contains(err.Error(), "type") { + t.Error("for struct/non-struct expected type error; got", err) + } + // Now try the other way + var nsp NonStruct + if err := encAndDec(ns, &nsp); err != nil { + t.Error(err) + } + if err := encAndDec(ns, &s); err == nil { + t.Error("should get error for non-struct/struct") + } else if !strings.Contains(err.Error(), "type") { + t.Error("for non-struct/struct expected type error; got", err) + } +} + +type interfaceIndirectTestI interface { + F() bool +} + +type interfaceIndirectTestT struct{} + +func (this *interfaceIndirectTestT) F() bool { + return true +} + +// A version of a bug reported on golang-nuts. Also tests top-level +// slice of interfaces. The issue was registering *T caused T to be +// stored as the concrete type. +func TestInterfaceIndirect(t *testing.T) { + Register(&interfaceIndirectTestT{}) + b := new(bytes.Buffer) + w := []interfaceIndirectTestI{&interfaceIndirectTestT{}} + err := NewEncoder(b).Encode(w) + if err != nil { + t.Fatal("encode error:", err) + } + + var r []interfaceIndirectTestI + err = NewDecoder(b).Decode(&r) + if err != nil { + t.Fatal("decode error:", err) + } +} + +// Now follow various tests that decode into things that can't represent the +// encoded value, all of which should be legal. + +// Also, when the ignored object contains an interface value, it may define +// types. Make sure that skipping the value still defines the types by using +// the encoder/decoder pair to send a value afterwards. If an interface +// is sent, its type in the test is always NewType0, so this checks that the +// encoder and decoder don't skew with respect to type definitions. + +type Struct0 struct { + I any +} + +type NewType0 struct { + S string +} + +type ignoreTest struct { + in, out any +} + +var ignoreTests = []ignoreTest{ + // Decode normal struct into an empty struct + {&struct{ A int }{23}, &struct{}{}}, + // Decode normal struct into a nil. + {&struct{ A int }{23}, nil}, + // Decode singleton string into a nil. + {"hello, world", nil}, + // Decode singleton slice into a nil. + {[]int{1, 2, 3, 4}, nil}, + // Decode struct containing an interface into a nil. + {&Struct0{&NewType0{"value0"}}, nil}, + // Decode singleton slice of interfaces into a nil. + {[]any{"hi", &NewType0{"value1"}, 23}, nil}, +} + +func TestDecodeIntoNothing(t *testing.T) { + Register(new(NewType0)) + for i, test := range ignoreTests { + b := new(bytes.Buffer) + enc := NewEncoder(b) + err := enc.Encode(test.in) + if err != nil { + t.Errorf("%d: encode error %s:", i, err) + continue + } + dec := NewDecoder(b) + err = dec.Decode(test.out) + if err != nil { + t.Errorf("%d: decode error: %s", i, err) + continue + } + // Now see if the encoder and decoder are in a consistent state. + str := fmt.Sprintf("Value %d", i) + err = enc.Encode(&NewType0{str}) + if err != nil { + t.Fatalf("%d: NewType0 encode error: %s", i, err) + } + ns := new(NewType0) + err = dec.Decode(ns) + if err != nil { + t.Fatalf("%d: NewType0 decode error: %s", i, err) + } + if ns.S != str { + t.Fatalf("%d: expected %q got %q", i, str, ns.S) + } + } +} + +func TestIgnoreRecursiveType(t *testing.T) { + // It's hard to build a self-contained test for this because + // we can't build compatible types in one package with + // different items so something is ignored. Here is + // some data that represents, according to debug.go: + // type definition { + // slice "recursiveSlice" id=106 + // elem id=106 + // } + data := []byte{ + 0x1d, 0xff, 0xd3, 0x02, 0x01, 0x01, 0x0e, 0x72, + 0x65, 0x63, 0x75, 0x72, 0x73, 0x69, 0x76, 0x65, + 0x53, 0x6c, 0x69, 0x63, 0x65, 0x01, 0xff, 0xd4, + 0x00, 0x01, 0xff, 0xd4, 0x00, 0x00, 0x07, 0xff, + 0xd4, 0x00, 0x02, 0x01, 0x00, 0x00, + } + dec := NewDecoder(bytes.NewReader(data)) + // Issue 10415: This caused infinite recursion. + err := dec.Decode(nil) + if err != nil { + t.Fatal(err) + } +} + +// Another bug from golang-nuts, involving nested interfaces. +type Bug0Outer struct { + Bug0Field any +} + +type Bug0Inner struct { + A int +} + +func TestNestedInterfaces(t *testing.T) { + var buf bytes.Buffer + e := NewEncoder(&buf) + d := NewDecoder(&buf) + Register(new(Bug0Outer)) + Register(new(Bug0Inner)) + f := &Bug0Outer{&Bug0Outer{&Bug0Inner{7}}} + var v any = f + err := e.Encode(&v) + if err != nil { + t.Fatal("Encode:", err) + } + err = d.Decode(&v) + if err != nil { + t.Fatal("Decode:", err) + } + // Make sure it decoded correctly. + outer1, ok := v.(*Bug0Outer) + if !ok { + t.Fatalf("v not Bug0Outer: %T", v) + } + outer2, ok := outer1.Bug0Field.(*Bug0Outer) + if !ok { + t.Fatalf("v.Bug0Field not Bug0Outer: %T", outer1.Bug0Field) + } + inner, ok := outer2.Bug0Field.(*Bug0Inner) + if !ok { + t.Fatalf("v.Bug0Field.Bug0Field not Bug0Inner: %T", outer2.Bug0Field) + } + if inner.A != 7 { + t.Fatalf("final value %d; expected %d", inner.A, 7) + } +} + +// The bugs keep coming. We forgot to send map subtypes before the map. + +type Bug1Elem struct { + Name string + Id int +} + +type Bug1StructMap map[string]Bug1Elem + +func TestMapBug1(t *testing.T) { + in := make(Bug1StructMap) + in["val1"] = Bug1Elem{"elem1", 1} + in["val2"] = Bug1Elem{"elem2", 2} + + b := new(bytes.Buffer) + enc := NewEncoder(b) + err := enc.Encode(in) + if err != nil { + t.Fatal("encode:", err) + } + dec := NewDecoder(b) + out := make(Bug1StructMap) + err = dec.Decode(&out) + if err != nil { + t.Fatal("decode:", err) + } + if !reflect.DeepEqual(in, out) { + t.Errorf("mismatch: %v %v", in, out) + } +} + +func TestGobMapInterfaceEncode(t *testing.T) { + m := map[string]any{ + "up": uintptr(0), + "i0": []int{-1}, + "i1": []int8{-1}, + "i2": []int16{-1}, + "i3": []int32{-1}, + "i4": []int64{-1}, + "u0": []uint{1}, + "u1": []uint8{1}, + "u2": []uint16{1}, + "u3": []uint32{1}, + "u4": []uint64{1}, + "f0": []float32{1}, + "f1": []float64{1}, + "c0": []complex64{complex(2, -2)}, + "c1": []complex128{complex(2, float64(-2))}, + "us": []uintptr{0}, + "bo": []bool{false}, + "st": []string{"s"}, + } + enc := NewEncoder(new(bytes.Buffer)) + err := enc.Encode(m) + if err != nil { + t.Errorf("encode map: %s", err) + } +} + +func TestSliceReusesMemory(t *testing.T) { + buf := new(bytes.Buffer) + // Bytes + { + x := []byte("abcd") + enc := NewEncoder(buf) + err := enc.Encode(x) + if err != nil { + t.Errorf("bytes: encode: %s", err) + } + // Decode into y, which is big enough. + y := []byte("ABCDE") + addr := &y[0] + dec := NewDecoder(buf) + err = dec.Decode(&y) + if err != nil { + t.Fatal("bytes: decode:", err) + } + if !bytes.Equal(x, y) { + t.Errorf("bytes: expected %q got %q\n", x, y) + } + if addr != &y[0] { + t.Errorf("bytes: unnecessary reallocation") + } + } + // general slice + { + x := []rune("abcd") + enc := NewEncoder(buf) + err := enc.Encode(x) + if err != nil { + t.Errorf("ints: encode: %s", err) + } + // Decode into y, which is big enough. + y := []rune("ABCDE") + addr := &y[0] + dec := NewDecoder(buf) + err = dec.Decode(&y) + if err != nil { + t.Fatal("ints: decode:", err) + } + if !reflect.DeepEqual(x, y) { + t.Errorf("ints: expected %q got %q\n", x, y) + } + if addr != &y[0] { + t.Errorf("ints: unnecessary reallocation") + } + } +} + +// Used to crash: negative count in recvMessage. +func TestBadCount(t *testing.T) { + b := []byte{0xfb, 0xa5, 0x82, 0x2f, 0xca, 0x1} + if err := NewDecoder(bytes.NewReader(b)).Decode(nil); err == nil { + t.Error("expected error from bad count") + } else if err.Error() != errBadCount.Error() { + t.Error("expected bad count error; got", err) + } +} + +// Verify that sequential Decoders built on a single input will +// succeed if the input implements ReadByte and there is no +// type information in the stream. +func TestSequentialDecoder(t *testing.T) { + b := new(bytes.Buffer) + enc := NewEncoder(b) + const count = 10 + for i := 0; i < count; i++ { + s := fmt.Sprintf("%d", i) + if err := enc.Encode(s); err != nil { + t.Error("encoder fail:", err) + } + } + for i := 0; i < count; i++ { + dec := NewDecoder(b) + var s string + if err := dec.Decode(&s); err != nil { + t.Fatal("decoder fail:", err) + } + if s != fmt.Sprintf("%d", i) { + t.Fatalf("decode expected %d got %s", i, s) + } + } +} + +// Should be able to have unrepresentable fields (chan, func, *chan etc.); we just ignore them. +type Bug2 struct { + A int + C chan int + CP *chan int + F func() + FPP **func() +} + +func TestChanFuncIgnored(t *testing.T) { + c := make(chan int) + f := func() {} + fp := &f + b0 := Bug2{23, c, &c, f, &fp} + var buf bytes.Buffer + enc := NewEncoder(&buf) + if err := enc.Encode(b0); err != nil { + t.Fatal("error encoding:", err) + } + var b1 Bug2 + err := NewDecoder(&buf).Decode(&b1) + if err != nil { + t.Fatal("decode:", err) + } + if b1.A != b0.A { + t.Fatalf("got %d want %d", b1.A, b0.A) + } + if b1.C != nil || b1.CP != nil || b1.F != nil || b1.FPP != nil { + t.Fatal("unexpected value for chan or func") + } +} + +func TestSliceIncompatibility(t *testing.T) { + var in = []byte{1, 2, 3} + var out []int + if err := encAndDec(in, &out); err == nil { + t.Error("expected compatibility error") + } +} + +// Mutually recursive slices of structs caused problems. +type Bug3 struct { + Num int + Children []*Bug3 +} + +func TestGobPtrSlices(t *testing.T) { + in := []*Bug3{ + {1, nil}, + {2, nil}, + } + b := new(bytes.Buffer) + err := NewEncoder(b).Encode(&in) + if err != nil { + t.Fatal("encode:", err) + } + + var out []*Bug3 + err = NewDecoder(b).Decode(&out) + if err != nil { + t.Fatal("decode:", err) + } + if !reflect.DeepEqual(in, out) { + t.Fatalf("got %v; wanted %v", out, in) + } +} + +// getDecEnginePtr cached engine for ut.base instead of ut.user so we passed +// a *map and then tried to reuse its engine to decode the inner map. +func TestPtrToMapOfMap(t *testing.T) { + Register(make(map[string]any)) + subdata := make(map[string]any) + subdata["bar"] = "baz" + data := make(map[string]any) + data["foo"] = subdata + + b := new(bytes.Buffer) + err := NewEncoder(b).Encode(data) + if err != nil { + t.Fatal("encode:", err) + } + var newData map[string]any + err = NewDecoder(b).Decode(&newData) + if err != nil { + t.Fatal("decode:", err) + } + if !reflect.DeepEqual(data, newData) { + t.Fatalf("expected %v got %v", data, newData) + } +} + +// Test that untyped nils generate an error, not a panic. +// See Issue 16204. +func TestCatchInvalidNilValue(t *testing.T) { + encodeErr, panicErr := encodeAndRecover(nil) + if panicErr != nil { + t.Fatalf("panicErr=%v, should not panic encoding untyped nil", panicErr) + } + if encodeErr == nil { + t.Errorf("got err=nil, want non-nil error when encoding untyped nil value") + } else if !strings.Contains(encodeErr.Error(), "nil value") { + t.Errorf("expected 'nil value' error; got err=%v", encodeErr) + } +} + +// A top-level nil pointer generates a panic with a helpful string-valued message. +func TestTopLevelNilPointer(t *testing.T) { + var ip *int + encodeErr, panicErr := encodeAndRecover(ip) + if encodeErr != nil { + t.Fatal("error in encode:", encodeErr) + } + if panicErr == nil { + t.Fatal("top-level nil pointer did not panic") + } + errMsg := panicErr.Error() + if !strings.Contains(errMsg, "nil pointer") { + t.Fatal("expected nil pointer error, got:", errMsg) + } +} + +func encodeAndRecover(value any) (encodeErr, panicErr error) { + defer func() { + e := recover() + if e != nil { + switch err := e.(type) { + case error: + panicErr = err + default: + panicErr = fmt.Errorf("%v", err) + } + } + }() + + encodeErr = NewEncoder(io.Discard).Encode(value) + return +} + +func TestNilPointerPanics(t *testing.T) { + var ( + nilStringPtr *string + intMap = make(map[int]int) + intMapPtr = &intMap + nilIntMapPtr *map[int]int + zero int + nilBoolChannel chan bool + nilBoolChannelPtr *chan bool + nilStringSlice []string + stringSlice = make([]string, 1) + nilStringSlicePtr *[]string + ) + + testCases := []struct { + value any + mustPanic bool + }{ + {nilStringPtr, true}, + {intMap, false}, + {intMapPtr, false}, + {nilIntMapPtr, true}, + {zero, false}, + {nilStringSlice, false}, + {stringSlice, false}, + {nilStringSlicePtr, true}, + {nilBoolChannel, false}, + {nilBoolChannelPtr, true}, + } + + for _, tt := range testCases { + _, panicErr := encodeAndRecover(tt.value) + if tt.mustPanic { + if panicErr == nil { + t.Errorf("expected panic with input %#v, did not panic", tt.value) + } + continue + } + if panicErr != nil { + t.Fatalf("expected no panic with input %#v, got panic=%v", tt.value, panicErr) + } + } +} + +func TestNilPointerInsideInterface(t *testing.T) { + var ip *int + si := struct { + I any + }{ + I: ip, + } + buf := new(bytes.Buffer) + err := NewEncoder(buf).Encode(si) + if err == nil { + t.Fatal("expected error, got none") + } + errMsg := err.Error() + if !strings.Contains(errMsg, "nil pointer") || !strings.Contains(errMsg, "interface") { + t.Fatal("expected error about nil pointer and interface, got:", errMsg) + } +} + +type Bug4Public struct { + Name string + Secret Bug4Secret +} + +type Bug4Secret struct { + a int // error: no exported fields. +} + +// Test that a failed compilation doesn't leave around an executable encoder. +// Issue 3723. +func TestMutipleEncodingsOfBadType(t *testing.T) { + x := Bug4Public{ + Name: "name", + Secret: Bug4Secret{1}, + } + buf := new(bytes.Buffer) + enc := NewEncoder(buf) + err := enc.Encode(x) + if err == nil { + t.Fatal("first encoding: expected error") + } + buf.Reset() + enc = NewEncoder(buf) + err = enc.Encode(x) + if err == nil { + t.Fatal("second encoding: expected error") + } + if !strings.Contains(err.Error(), "no exported fields") { + t.Errorf("expected error about no exported fields; got %v", err) + } +} + +// There was an error check comparing the length of the input with the +// length of the slice being decoded. It was wrong because the next +// thing in the input might be a type definition, which would lead to +// an incorrect length check. This test reproduces the corner case. + +type Z struct { +} + +func Test29ElementSlice(t *testing.T) { + Register(Z{}) + src := make([]any, 100) // Size needs to be bigger than size of type definition. + for i := range src { + src[i] = Z{} + } + buf := new(bytes.Buffer) + err := NewEncoder(buf).Encode(src) + if err != nil { + t.Fatalf("encode: %v", err) + return + } + + var dst []any + err = NewDecoder(buf).Decode(&dst) + if err != nil { + t.Errorf("decode: %v", err) + return + } +} + +// Don't crash, just give error when allocating a huge slice. +// Issue 8084. +func TestErrorForHugeSlice(t *testing.T) { + // Encode an int slice. + buf := new(bytes.Buffer) + slice := []int{1, 1, 1, 1, 1, 1, 1, 1, 1, 1} + err := NewEncoder(buf).Encode(slice) + if err != nil { + t.Fatal("encode:", err) + } + // Reach into the buffer and smash the count to make the encoded slice very long. + buf.Bytes()[buf.Len()-len(slice)-1] = 0xfa + // Decode and see error. + err = NewDecoder(buf).Decode(&slice) + if err == nil { + t.Fatal("decode: no error") + } + if !strings.Contains(err.Error(), "slice too big") { + t.Fatalf("decode: expected slice too big error, got %s", err.Error()) + } +} + +type badDataTest struct { + input string // The input encoded as a hex string. + error string // A substring of the error that should result. + data any // What to decode into. +} + +var badDataTests = []badDataTest{ + {"", "EOF", nil}, + {"7F6869", "unexpected EOF", nil}, + {"036e6f77206973207468652074696d6520666f7220616c6c20676f6f64206d656e", "unknown type id", new(ET2)}, + {"0424666f6f", "field numbers out of bounds", new(ET2)}, // Issue 6323. + {"05100028557b02027f8302", "interface encoding", nil}, // Issue 10270. + // Issue 10273. + {"130a00fb5dad0bf8ff020263e70002fa28020202a89859", "slice length too large", nil}, + {"0f1000fb285d003316020735ff023a65c5", "interface encoding", nil}, + {"03fffb0616fffc00f902ff02ff03bf005d02885802a311a8120228022c028ee7", "GobDecoder", nil}, + // Issue 10491. + {"10fe010f020102fe01100001fe010e000016fe010d030102fe010e00010101015801fe01100000000bfe011000f85555555555555555", "exceeds input size", nil}, +} + +// TestBadData tests that various problems caused by malformed input +// are caught as errors and do not cause panics. +func TestBadData(t *testing.T) { + for i, test := range badDataTests { + data, err := hex.DecodeString(test.input) + if err != nil { + t.Fatalf("#%d: hex error: %s", i, err) + } + d := NewDecoder(bytes.NewReader(data)) + err = d.Decode(test.data) + if err == nil { + t.Errorf("decode: no error") + continue + } + if !strings.Contains(err.Error(), test.error) { + t.Errorf("#%d: decode: expected %q error, got %s", i, test.error, err.Error()) + } + } +} + +func TestDecodeErrorMultipleTypes(t *testing.T) { + type Test struct { + A string + B int + } + var b bytes.Buffer + NewEncoder(&b).Encode(Test{"one", 1}) + + var result, result2 Test + dec := NewDecoder(&b) + err := dec.Decode(&result) + if err != nil { + t.Errorf("decode: unexpected error %v", err) + } + + b.Reset() + NewEncoder(&b).Encode(Test{"two", 2}) + err = dec.Decode(&result2) + if err == nil { + t.Errorf("decode: expected duplicate type error, got nil") + } else if !strings.Contains(err.Error(), "duplicate type") { + t.Errorf("decode: expected duplicate type error, got %s", err.Error()) + } +} + +// Issue 24075 +func TestMarshalFloatMap(t *testing.T) { + nan1 := math.NaN() + nan2 := math.Float64frombits(math.Float64bits(nan1) ^ 1) // A different NaN in the same class. + + in := map[float64]string{ + nan1: "a", + nan1: "b", + nan2: "c", + } + + var b bytes.Buffer + enc := NewEncoder(&b) + if err := enc.Encode(in); err != nil { + t.Errorf("Encode : %v", err) + } + + out := map[float64]string{} + dec := NewDecoder(&b) + if err := dec.Decode(&out); err != nil { + t.Fatalf("Decode : %v", err) + } + + type mapEntry struct { + keyBits uint64 + value string + } + readMap := func(m map[float64]string) (entries []mapEntry) { + for k, v := range m { + entries = append(entries, mapEntry{math.Float64bits(k), v}) + } + sort.Slice(entries, func(i, j int) bool { + ei, ej := entries[i], entries[j] + if ei.keyBits != ej.keyBits { + return ei.keyBits < ej.keyBits + } + return ei.value < ej.value + }) + return entries + } + + got := readMap(out) + want := readMap(in) + if !reflect.DeepEqual(got, want) { + t.Fatalf("\nEncode: %v\nDecode: %v", want, got) + } +} + +func TestDecodePartial(t *testing.T) { + type T struct { + X []int + Y string + } + + var buf bytes.Buffer + t1 := T{X: []int{1, 2, 3}, Y: "foo"} + t2 := T{X: []int{4, 5, 6}, Y: "bar"} + enc := NewEncoder(&buf) + + t1start := 0 + if err := enc.Encode(&t1); err != nil { + t.Fatal(err) + } + + t2start := buf.Len() + if err := enc.Encode(&t2); err != nil { + t.Fatal(err) + } + + data := buf.Bytes() + for i := 0; i <= len(data); i++ { + bufr := bytes.NewReader(data[:i]) + + // Decode both values, stopping at the first error. + var t1b, t2b T + dec := NewDecoder(bufr) + var err error + err = dec.Decode(&t1b) + if err == nil { + err = dec.Decode(&t2b) + } + + switch i { + case t1start, t2start: + // Either the first or the second Decode calls had zero input. + if err != io.EOF { + t.Errorf("%d/%d: expected io.EOF: %v", i, len(data), err) + } + case len(data): + // We reached the end of the entire input. + if err != nil { + t.Errorf("%d/%d: unexpected error: %v", i, len(data), err) + } + if !reflect.DeepEqual(t1b, t1) { + t.Fatalf("t1 value mismatch: got %v, want %v", t1b, t1) + } + if !reflect.DeepEqual(t2b, t2) { + t.Fatalf("t2 value mismatch: got %v, want %v", t2b, t2) + } + default: + // In between, we must see io.ErrUnexpectedEOF. + // The decoder used to erroneously return io.EOF in some cases here, + // such as if the input was cut off right after some type specs, + // but before any value was actually transmitted. + if err != io.ErrUnexpectedEOF { + t.Errorf("%d/%d: expected io.ErrUnexpectedEOF: %v", i, len(data), err) + } + } + } +} + +func TestDecoderOverflow(t *testing.T) { + // Issue 55337. + dec := NewDecoder(bytes.NewReader([]byte{ + 0x12, 0xff, 0xff, 0x2, 0x2, 0x20, 0x0, 0xf8, 0x7f, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x20, 0x20, 0x20, 0x20, 0x20, + })) + var r interface{} + err := dec.Decode(r) + if err == nil { + t.Fatalf("expected an error") + } +} diff --git a/src/encoding/gob/error.go b/src/encoding/gob/error.go new file mode 100644 index 0000000..9c614e3 --- /dev/null +++ b/src/encoding/gob/error.go @@ -0,0 +1,42 @@ +// 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 gob + +import "fmt" + +// Errors in decoding and encoding are handled using panic and recover. +// Panics caused by user error (that is, everything except run-time panics +// such as "index out of bounds" errors) do not leave the file that caused +// them, but are instead turned into plain error returns. Encoding and +// decoding functions and methods that do not return an error either use +// panic to report an error or are guaranteed error-free. + +// A gobError is used to distinguish errors (panics) generated in this package. +type gobError struct { + err error +} + +// errorf is like error_ but takes Printf-style arguments to construct an error. +// It always prefixes the message with "gob: ". +func errorf(format string, args ...any) { + error_(fmt.Errorf("gob: "+format, args...)) +} + +// error_ wraps the argument error and uses it as the argument to panic. +func error_(err error) { + panic(gobError{err}) +} + +// catchError is meant to be used as a deferred function to turn a panic(gobError) into a +// plain error. It overwrites the error return of the function that deferred its call. +func catchError(err *error) { + if e := recover(); e != nil { + ge, ok := e.(gobError) + if !ok { + panic(e) + } + *err = ge.err + } +} diff --git a/src/encoding/gob/example_encdec_test.go b/src/encoding/gob/example_encdec_test.go new file mode 100644 index 0000000..e45ad4c --- /dev/null +++ b/src/encoding/gob/example_encdec_test.go @@ -0,0 +1,61 @@ +// Copyright 2013 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 gob_test + +import ( + "bytes" + "encoding/gob" + "fmt" + "log" +) + +// The Vector type has unexported fields, which the package cannot access. +// We therefore write a BinaryMarshal/BinaryUnmarshal method pair to allow us +// to send and receive the type with the gob package. These interfaces are +// defined in the "encoding" package. +// We could equivalently use the locally defined GobEncode/GobDecoder +// interfaces. +type Vector struct { + x, y, z int +} + +func (v Vector) MarshalBinary() ([]byte, error) { + // A simple encoding: plain text. + var b bytes.Buffer + fmt.Fprintln(&b, v.x, v.y, v.z) + return b.Bytes(), nil +} + +// UnmarshalBinary modifies the receiver so it must take a pointer receiver. +func (v *Vector) UnmarshalBinary(data []byte) error { + // A simple encoding: plain text. + b := bytes.NewBuffer(data) + _, err := fmt.Fscanln(b, &v.x, &v.y, &v.z) + return err +} + +// This example transmits a value that implements the custom encoding and decoding methods. +func Example_encodeDecode() { + var network bytes.Buffer // Stand-in for the network. + + // Create an encoder and send a value. + enc := gob.NewEncoder(&network) + err := enc.Encode(Vector{3, 4, 5}) + if err != nil { + log.Fatal("encode:", err) + } + + // Create a decoder and receive a value. + dec := gob.NewDecoder(&network) + var v Vector + err = dec.Decode(&v) + if err != nil { + log.Fatal("decode:", err) + } + fmt.Println(v) + + // Output: + // {3 4 5} +} diff --git a/src/encoding/gob/example_interface_test.go b/src/encoding/gob/example_interface_test.go new file mode 100644 index 0000000..cf5ba38 --- /dev/null +++ b/src/encoding/gob/example_interface_test.go @@ -0,0 +1,81 @@ +// Copyright 2013 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 gob_test + +import ( + "bytes" + "encoding/gob" + "fmt" + "log" + "math" +) + +type Point struct { + X, Y int +} + +func (p Point) Hypotenuse() float64 { + return math.Hypot(float64(p.X), float64(p.Y)) +} + +type Pythagoras interface { + Hypotenuse() float64 +} + +// This example shows how to encode an interface value. The key +// distinction from regular types is to register the concrete type that +// implements the interface. +func Example_interface() { + var network bytes.Buffer // Stand-in for the network. + + // We must register the concrete type for the encoder and decoder (which would + // normally be on a separate machine from the encoder). On each end, this tells the + // engine which concrete type is being sent that implements the interface. + gob.Register(Point{}) + + // Create an encoder and send some values. + enc := gob.NewEncoder(&network) + for i := 1; i <= 3; i++ { + interfaceEncode(enc, Point{3 * i, 4 * i}) + } + + // Create a decoder and receive some values. + dec := gob.NewDecoder(&network) + for i := 1; i <= 3; i++ { + result := interfaceDecode(dec) + fmt.Println(result.Hypotenuse()) + } + + // Output: + // 5 + // 10 + // 15 +} + +// interfaceEncode encodes the interface value into the encoder. +func interfaceEncode(enc *gob.Encoder, p Pythagoras) { + // The encode will fail unless the concrete type has been + // registered. We registered it in the calling function. + + // Pass pointer to interface so Encode sees (and hence sends) a value of + // interface type. If we passed p directly it would see the concrete type instead. + // See the blog post, "The Laws of Reflection" for background. + err := enc.Encode(&p) + if err != nil { + log.Fatal("encode:", err) + } +} + +// interfaceDecode decodes the next interface value from the stream and returns it. +func interfaceDecode(dec *gob.Decoder) Pythagoras { + // The decode will fail unless the concrete type on the wire has been + // registered. We registered it in the calling function. + var p Pythagoras + err := dec.Decode(&p) + if err != nil { + log.Fatal("decode:", err) + } + return p +} diff --git a/src/encoding/gob/example_test.go b/src/encoding/gob/example_test.go new file mode 100644 index 0000000..16b7123 --- /dev/null +++ b/src/encoding/gob/example_test.go @@ -0,0 +1,60 @@ +// Copyright 2013 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 gob_test + +import ( + "bytes" + "encoding/gob" + "fmt" + "log" +) + +type P struct { + X, Y, Z int + Name string +} + +type Q struct { + X, Y *int32 + Name string +} + +// This example shows the basic usage of the package: Create an encoder, +// transmit some values, receive them with a decoder. +func Example_basic() { + // Initialize the encoder and decoder. Normally enc and dec would be + // bound to network connections and the encoder and decoder would + // run in different processes. + var network bytes.Buffer // Stand-in for a network connection + enc := gob.NewEncoder(&network) // Will write to network. + dec := gob.NewDecoder(&network) // Will read from network. + + // Encode (send) some values. + err := enc.Encode(P{3, 4, 5, "Pythagoras"}) + if err != nil { + log.Fatal("encode error:", err) + } + err = enc.Encode(P{1782, 1841, 1922, "Treehouse"}) + if err != nil { + log.Fatal("encode error:", err) + } + + // Decode (receive) and print the values. + var q Q + err = dec.Decode(&q) + if err != nil { + log.Fatal("decode error 1:", err) + } + fmt.Printf("%q: {%d, %d}\n", q.Name, *q.X, *q.Y) + err = dec.Decode(&q) + if err != nil { + log.Fatal("decode error 2:", err) + } + fmt.Printf("%q: {%d, %d}\n", q.Name, *q.X, *q.Y) + + // Output: + // "Pythagoras": {3, 4} + // "Treehouse": {1782, 1841} +} diff --git a/src/encoding/gob/gobencdec_test.go b/src/encoding/gob/gobencdec_test.go new file mode 100644 index 0000000..6fefd36 --- /dev/null +++ b/src/encoding/gob/gobencdec_test.go @@ -0,0 +1,822 @@ +// 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. + +// This file contains tests of the GobEncoder/GobDecoder support. + +package gob + +import ( + "bytes" + "errors" + "fmt" + "io" + "net" + "reflect" + "strings" + "testing" + "time" +) + +// Types that implement the GobEncoder/Decoder interfaces. + +type ByteStruct struct { + a byte // not an exported field +} + +type StringStruct struct { + s string // not an exported field +} + +type ArrayStruct struct { + a [8192]byte // not an exported field +} + +type Gobber int + +type ValueGobber string // encodes with a value, decodes with a pointer. + +type BinaryGobber int + +type BinaryValueGobber string + +type TextGobber int + +type TextValueGobber string + +// The relevant methods + +func (g *ByteStruct) GobEncode() ([]byte, error) { + b := make([]byte, 3) + b[0] = g.a + b[1] = g.a + 1 + b[2] = g.a + 2 + return b, nil +} + +func (g *ByteStruct) GobDecode(data []byte) error { + if g == nil { + return errors.New("NIL RECEIVER") + } + // Expect N sequential-valued bytes. + if len(data) == 0 { + return io.EOF + } + g.a = data[0] + for i, c := range data { + if c != g.a+byte(i) { + return errors.New("invalid data sequence") + } + } + return nil +} + +func (g *StringStruct) GobEncode() ([]byte, error) { + return []byte(g.s), nil +} + +func (g *StringStruct) GobDecode(data []byte) error { + // Expect N sequential-valued bytes. + if len(data) == 0 { + return io.EOF + } + a := data[0] + for i, c := range data { + if c != a+byte(i) { + return errors.New("invalid data sequence") + } + } + g.s = string(data) + return nil +} + +func (a *ArrayStruct) GobEncode() ([]byte, error) { + return a.a[:], nil +} + +func (a *ArrayStruct) GobDecode(data []byte) error { + if len(data) != len(a.a) { + return errors.New("wrong length in array decode") + } + copy(a.a[:], data) + return nil +} + +func (g *Gobber) GobEncode() ([]byte, error) { + return []byte(fmt.Sprintf("VALUE=%d", *g)), nil +} + +func (g *Gobber) GobDecode(data []byte) error { + _, err := fmt.Sscanf(string(data), "VALUE=%d", (*int)(g)) + return err +} + +func (g *BinaryGobber) MarshalBinary() ([]byte, error) { + return []byte(fmt.Sprintf("VALUE=%d", *g)), nil +} + +func (g *BinaryGobber) UnmarshalBinary(data []byte) error { + _, err := fmt.Sscanf(string(data), "VALUE=%d", (*int)(g)) + return err +} + +func (g *TextGobber) MarshalText() ([]byte, error) { + return []byte(fmt.Sprintf("VALUE=%d", *g)), nil +} + +func (g *TextGobber) UnmarshalText(data []byte) error { + _, err := fmt.Sscanf(string(data), "VALUE=%d", (*int)(g)) + return err +} + +func (v ValueGobber) GobEncode() ([]byte, error) { + return []byte(fmt.Sprintf("VALUE=%s", v)), nil +} + +func (v *ValueGobber) GobDecode(data []byte) error { + _, err := fmt.Sscanf(string(data), "VALUE=%s", (*string)(v)) + return err +} + +func (v BinaryValueGobber) MarshalBinary() ([]byte, error) { + return []byte(fmt.Sprintf("VALUE=%s", v)), nil +} + +func (v *BinaryValueGobber) UnmarshalBinary(data []byte) error { + _, err := fmt.Sscanf(string(data), "VALUE=%s", (*string)(v)) + return err +} + +func (v TextValueGobber) MarshalText() ([]byte, error) { + return []byte(fmt.Sprintf("VALUE=%s", v)), nil +} + +func (v *TextValueGobber) UnmarshalText(data []byte) error { + _, err := fmt.Sscanf(string(data), "VALUE=%s", (*string)(v)) + return err +} + +// Structs that include GobEncodable fields. + +type GobTest0 struct { + X int // guarantee we have something in common with GobTest* + G *ByteStruct +} + +type GobTest1 struct { + X int // guarantee we have something in common with GobTest* + G *StringStruct +} + +type GobTest2 struct { + X int // guarantee we have something in common with GobTest* + G string // not a GobEncoder - should give us errors +} + +type GobTest3 struct { + X int // guarantee we have something in common with GobTest* + G *Gobber + B *BinaryGobber + T *TextGobber +} + +type GobTest4 struct { + X int // guarantee we have something in common with GobTest* + V ValueGobber + BV BinaryValueGobber + TV TextValueGobber +} + +type GobTest5 struct { + X int // guarantee we have something in common with GobTest* + V *ValueGobber + BV *BinaryValueGobber + TV *TextValueGobber +} + +type GobTest6 struct { + X int // guarantee we have something in common with GobTest* + V ValueGobber + W *ValueGobber + BV BinaryValueGobber + BW *BinaryValueGobber + TV TextValueGobber + TW *TextValueGobber +} + +type GobTest7 struct { + X int // guarantee we have something in common with GobTest* + V *ValueGobber + W ValueGobber + BV *BinaryValueGobber + BW BinaryValueGobber + TV *TextValueGobber + TW TextValueGobber +} + +type GobTestIgnoreEncoder struct { + X int // guarantee we have something in common with GobTest* +} + +type GobTestValueEncDec struct { + X int // guarantee we have something in common with GobTest* + G StringStruct // not a pointer. +} + +type GobTestIndirectEncDec struct { + X int // guarantee we have something in common with GobTest* + G ***StringStruct // indirections to the receiver. +} + +type GobTestArrayEncDec struct { + X int // guarantee we have something in common with GobTest* + A ArrayStruct // not a pointer. +} + +type GobTestIndirectArrayEncDec struct { + X int // guarantee we have something in common with GobTest* + A ***ArrayStruct // indirections to a large receiver. +} + +func TestGobEncoderField(t *testing.T) { + b := new(bytes.Buffer) + // First a field that's a structure. + enc := NewEncoder(b) + err := enc.Encode(GobTest0{17, &ByteStruct{'A'}}) + if err != nil { + t.Fatal("encode error:", err) + } + dec := NewDecoder(b) + x := new(GobTest0) + err = dec.Decode(x) + if err != nil { + t.Fatal("decode error:", err) + } + if x.G.a != 'A' { + t.Errorf("expected 'A' got %c", x.G.a) + } + // Now a field that's not a structure. + b.Reset() + gobber := Gobber(23) + bgobber := BinaryGobber(24) + tgobber := TextGobber(25) + err = enc.Encode(GobTest3{17, &gobber, &bgobber, &tgobber}) + if err != nil { + t.Fatal("encode error:", err) + } + y := new(GobTest3) + err = dec.Decode(y) + if err != nil { + t.Fatal("decode error:", err) + } + if *y.G != 23 || *y.B != 24 || *y.T != 25 { + t.Errorf("expected '23 got %d", *y.G) + } +} + +// Even though the field is a value, we can still take its address +// and should be able to call the methods. +func TestGobEncoderValueField(t *testing.T) { + b := new(bytes.Buffer) + // First a field that's a structure. + enc := NewEncoder(b) + err := enc.Encode(&GobTestValueEncDec{17, StringStruct{"HIJKL"}}) + if err != nil { + t.Fatal("encode error:", err) + } + dec := NewDecoder(b) + x := new(GobTestValueEncDec) + err = dec.Decode(x) + if err != nil { + t.Fatal("decode error:", err) + } + if x.G.s != "HIJKL" { + t.Errorf("expected `HIJKL` got %s", x.G.s) + } +} + +// GobEncode/Decode should work even if the value is +// more indirect than the receiver. +func TestGobEncoderIndirectField(t *testing.T) { + b := new(bytes.Buffer) + // First a field that's a structure. + enc := NewEncoder(b) + s := &StringStruct{"HIJKL"} + sp := &s + err := enc.Encode(GobTestIndirectEncDec{17, &sp}) + if err != nil { + t.Fatal("encode error:", err) + } + dec := NewDecoder(b) + x := new(GobTestIndirectEncDec) + err = dec.Decode(x) + if err != nil { + t.Fatal("decode error:", err) + } + if (***x.G).s != "HIJKL" { + t.Errorf("expected `HIJKL` got %s", (***x.G).s) + } +} + +// Test with a large field with methods. +func TestGobEncoderArrayField(t *testing.T) { + b := new(bytes.Buffer) + enc := NewEncoder(b) + var a GobTestArrayEncDec + a.X = 17 + for i := range a.A.a { + a.A.a[i] = byte(i) + } + err := enc.Encode(&a) + if err != nil { + t.Fatal("encode error:", err) + } + dec := NewDecoder(b) + x := new(GobTestArrayEncDec) + err = dec.Decode(x) + if err != nil { + t.Fatal("decode error:", err) + } + for i, v := range x.A.a { + if v != byte(i) { + t.Errorf("expected %x got %x", byte(i), v) + break + } + } +} + +// Test an indirection to a large field with methods. +func TestGobEncoderIndirectArrayField(t *testing.T) { + b := new(bytes.Buffer) + enc := NewEncoder(b) + var a GobTestIndirectArrayEncDec + a.X = 17 + var array ArrayStruct + ap := &array + app := &ap + a.A = &app + for i := range array.a { + array.a[i] = byte(i) + } + err := enc.Encode(a) + if err != nil { + t.Fatal("encode error:", err) + } + dec := NewDecoder(b) + x := new(GobTestIndirectArrayEncDec) + err = dec.Decode(x) + if err != nil { + t.Fatal("decode error:", err) + } + for i, v := range (***x.A).a { + if v != byte(i) { + t.Errorf("expected %x got %x", byte(i), v) + break + } + } +} + +// As long as the fields have the same name and implement the +// interface, we can cross-connect them. Not sure it's useful +// and may even be bad but it works and it's hard to prevent +// without exposing the contents of the object, which would +// defeat the purpose. +func TestGobEncoderFieldsOfDifferentType(t *testing.T) { + // first, string in field to byte in field + b := new(bytes.Buffer) + enc := NewEncoder(b) + err := enc.Encode(GobTest1{17, &StringStruct{"ABC"}}) + if err != nil { + t.Fatal("encode error:", err) + } + dec := NewDecoder(b) + x := new(GobTest0) + err = dec.Decode(x) + if err != nil { + t.Fatal("decode error:", err) + } + if x.G.a != 'A' { + t.Errorf("expected 'A' got %c", x.G.a) + } + // now the other direction, byte in field to string in field + b.Reset() + err = enc.Encode(GobTest0{17, &ByteStruct{'X'}}) + if err != nil { + t.Fatal("encode error:", err) + } + y := new(GobTest1) + err = dec.Decode(y) + if err != nil { + t.Fatal("decode error:", err) + } + if y.G.s != "XYZ" { + t.Fatalf("expected `XYZ` got %q", y.G.s) + } +} + +// Test that we can encode a value and decode into a pointer. +func TestGobEncoderValueEncoder(t *testing.T) { + // first, string in field to byte in field + b := new(bytes.Buffer) + enc := NewEncoder(b) + err := enc.Encode(GobTest4{17, ValueGobber("hello"), BinaryValueGobber("Καλημέρα"), TextValueGobber("こんにちは")}) + if err != nil { + t.Fatal("encode error:", err) + } + dec := NewDecoder(b) + x := new(GobTest5) + err = dec.Decode(x) + if err != nil { + t.Fatal("decode error:", err) + } + if *x.V != "hello" || *x.BV != "Καλημέρα" || *x.TV != "こんにちは" { + t.Errorf("expected `hello` got %s", *x.V) + } +} + +// Test that we can use a value then a pointer type of a GobEncoder +// in the same encoded value. Bug 4647. +func TestGobEncoderValueThenPointer(t *testing.T) { + v := ValueGobber("forty-two") + w := ValueGobber("six-by-nine") + bv := BinaryValueGobber("1nanocentury") + bw := BinaryValueGobber("πseconds") + tv := TextValueGobber("gravitationalacceleration") + tw := TextValueGobber("π²ft/s²") + + // this was a bug: encoding a GobEncoder by value before a GobEncoder + // pointer would cause duplicate type definitions to be sent. + + b := new(bytes.Buffer) + enc := NewEncoder(b) + if err := enc.Encode(GobTest6{42, v, &w, bv, &bw, tv, &tw}); err != nil { + t.Fatal("encode error:", err) + } + dec := NewDecoder(b) + x := new(GobTest6) + if err := dec.Decode(x); err != nil { + t.Fatal("decode error:", err) + } + + if got, want := x.V, v; got != want { + t.Errorf("v = %q, want %q", got, want) + } + if got, want := x.W, w; got == nil { + t.Errorf("w = nil, want %q", want) + } else if *got != want { + t.Errorf("w = %q, want %q", *got, want) + } + + if got, want := x.BV, bv; got != want { + t.Errorf("bv = %q, want %q", got, want) + } + if got, want := x.BW, bw; got == nil { + t.Errorf("bw = nil, want %q", want) + } else if *got != want { + t.Errorf("bw = %q, want %q", *got, want) + } + + if got, want := x.TV, tv; got != want { + t.Errorf("tv = %q, want %q", got, want) + } + if got, want := x.TW, tw; got == nil { + t.Errorf("tw = nil, want %q", want) + } else if *got != want { + t.Errorf("tw = %q, want %q", *got, want) + } +} + +// Test that we can use a pointer then a value type of a GobEncoder +// in the same encoded value. +func TestGobEncoderPointerThenValue(t *testing.T) { + v := ValueGobber("forty-two") + w := ValueGobber("six-by-nine") + bv := BinaryValueGobber("1nanocentury") + bw := BinaryValueGobber("πseconds") + tv := TextValueGobber("gravitationalacceleration") + tw := TextValueGobber("π²ft/s²") + + b := new(bytes.Buffer) + enc := NewEncoder(b) + if err := enc.Encode(GobTest7{42, &v, w, &bv, bw, &tv, tw}); err != nil { + t.Fatal("encode error:", err) + } + dec := NewDecoder(b) + x := new(GobTest7) + if err := dec.Decode(x); err != nil { + t.Fatal("decode error:", err) + } + + if got, want := x.V, v; got == nil { + t.Errorf("v = nil, want %q", want) + } else if *got != want { + t.Errorf("v = %q, want %q", *got, want) + } + if got, want := x.W, w; got != want { + t.Errorf("w = %q, want %q", got, want) + } + + if got, want := x.BV, bv; got == nil { + t.Errorf("bv = nil, want %q", want) + } else if *got != want { + t.Errorf("bv = %q, want %q", *got, want) + } + if got, want := x.BW, bw; got != want { + t.Errorf("bw = %q, want %q", got, want) + } + + if got, want := x.TV, tv; got == nil { + t.Errorf("tv = nil, want %q", want) + } else if *got != want { + t.Errorf("tv = %q, want %q", *got, want) + } + if got, want := x.TW, tw; got != want { + t.Errorf("tw = %q, want %q", got, want) + } +} + +func TestGobEncoderFieldTypeError(t *testing.T) { + // GobEncoder to non-decoder: error + b := new(bytes.Buffer) + enc := NewEncoder(b) + err := enc.Encode(GobTest1{17, &StringStruct{"ABC"}}) + if err != nil { + t.Fatal("encode error:", err) + } + dec := NewDecoder(b) + x := &GobTest2{} + err = dec.Decode(x) + if err == nil { + t.Fatal("expected decode error for mismatched fields (encoder to non-decoder)") + } + if !strings.Contains(err.Error(), "type") { + t.Fatal("expected type error; got", err) + } + // Non-encoder to GobDecoder: error + b.Reset() + err = enc.Encode(GobTest2{17, "ABC"}) + if err != nil { + t.Fatal("encode error:", err) + } + y := &GobTest1{} + err = dec.Decode(y) + if err == nil { + t.Fatal("expected decode error for mismatched fields (non-encoder to decoder)") + } + if !strings.Contains(err.Error(), "type") { + t.Fatal("expected type error; got", err) + } +} + +// Even though ByteStruct is a struct, it's treated as a singleton at the top level. +func TestGobEncoderStructSingleton(t *testing.T) { + b := new(bytes.Buffer) + enc := NewEncoder(b) + err := enc.Encode(&ByteStruct{'A'}) + if err != nil { + t.Fatal("encode error:", err) + } + dec := NewDecoder(b) + x := new(ByteStruct) + err = dec.Decode(x) + if err != nil { + t.Fatal("decode error:", err) + } + if x.a != 'A' { + t.Errorf("expected 'A' got %c", x.a) + } +} + +func TestGobEncoderNonStructSingleton(t *testing.T) { + b := new(bytes.Buffer) + enc := NewEncoder(b) + var g Gobber = 1234 + err := enc.Encode(&g) + if err != nil { + t.Fatal("encode error:", err) + } + dec := NewDecoder(b) + var x Gobber + err = dec.Decode(&x) + if err != nil { + t.Fatal("decode error:", err) + } + if x != 1234 { + t.Errorf("expected 1234 got %d", x) + } +} + +func TestGobEncoderIgnoreStructField(t *testing.T) { + b := new(bytes.Buffer) + // First a field that's a structure. + enc := NewEncoder(b) + err := enc.Encode(GobTest0{17, &ByteStruct{'A'}}) + if err != nil { + t.Fatal("encode error:", err) + } + dec := NewDecoder(b) + x := new(GobTestIgnoreEncoder) + err = dec.Decode(x) + if err != nil { + t.Fatal("decode error:", err) + } + if x.X != 17 { + t.Errorf("expected 17 got %c", x.X) + } +} + +func TestGobEncoderIgnoreNonStructField(t *testing.T) { + b := new(bytes.Buffer) + // First a field that's a structure. + enc := NewEncoder(b) + gobber := Gobber(23) + bgobber := BinaryGobber(24) + tgobber := TextGobber(25) + err := enc.Encode(GobTest3{17, &gobber, &bgobber, &tgobber}) + if err != nil { + t.Fatal("encode error:", err) + } + dec := NewDecoder(b) + x := new(GobTestIgnoreEncoder) + err = dec.Decode(x) + if err != nil { + t.Fatal("decode error:", err) + } + if x.X != 17 { + t.Errorf("expected 17 got %c", x.X) + } +} + +func TestGobEncoderIgnoreNilEncoder(t *testing.T) { + b := new(bytes.Buffer) + // First a field that's a structure. + enc := NewEncoder(b) + err := enc.Encode(GobTest0{X: 18}) // G is nil + if err != nil { + t.Fatal("encode error:", err) + } + dec := NewDecoder(b) + x := new(GobTest0) + err = dec.Decode(x) + if err != nil { + t.Fatal("decode error:", err) + } + if x.X != 18 { + t.Errorf("expected x.X = 18, got %v", x.X) + } + if x.G != nil { + t.Errorf("expected x.G = nil, got %v", x.G) + } +} + +type gobDecoderBug0 struct { + foo, bar string +} + +func (br *gobDecoderBug0) String() string { + return br.foo + "-" + br.bar +} + +func (br *gobDecoderBug0) GobEncode() ([]byte, error) { + return []byte(br.String()), nil +} + +func (br *gobDecoderBug0) GobDecode(b []byte) error { + br.foo = "foo" + br.bar = "bar" + return nil +} + +// This was a bug: the receiver has a different indirection level +// than the variable. +func TestGobEncoderExtraIndirect(t *testing.T) { + gdb := &gobDecoderBug0{"foo", "bar"} + buf := new(bytes.Buffer) + e := NewEncoder(buf) + if err := e.Encode(gdb); err != nil { + t.Fatalf("encode: %v", err) + } + d := NewDecoder(buf) + var got *gobDecoderBug0 + if err := d.Decode(&got); err != nil { + t.Fatalf("decode: %v", err) + } + if got.foo != gdb.foo || got.bar != gdb.bar { + t.Errorf("got = %q, want %q", got, gdb) + } +} + +// Another bug: this caused a crash with the new Go1 Time type. +// We throw in a gob-encoding array, to test another case of isZero, +// and a struct containing a nil interface, to test a third. +type isZeroBug struct { + T time.Time + S string + I int + A isZeroBugArray + F isZeroBugInterface +} + +type isZeroBugArray [2]uint8 + +// Receiver is value, not pointer, to test isZero of array. +func (a isZeroBugArray) GobEncode() (b []byte, e error) { + b = append(b, a[:]...) + return b, nil +} + +func (a *isZeroBugArray) GobDecode(data []byte) error { + if len(data) != len(a) { + return io.EOF + } + a[0] = data[0] + a[1] = data[1] + return nil +} + +type isZeroBugInterface struct { + I any +} + +func (i isZeroBugInterface) GobEncode() (b []byte, e error) { + return []byte{}, nil +} + +func (i *isZeroBugInterface) GobDecode(data []byte) error { + return nil +} + +func TestGobEncodeIsZero(t *testing.T) { + x := isZeroBug{time.Unix(1e9, 0), "hello", -55, isZeroBugArray{1, 2}, isZeroBugInterface{}} + b := new(bytes.Buffer) + enc := NewEncoder(b) + err := enc.Encode(x) + if err != nil { + t.Fatal("encode:", err) + } + var y isZeroBug + dec := NewDecoder(b) + err = dec.Decode(&y) + if err != nil { + t.Fatal("decode:", err) + } + if x != y { + t.Fatalf("%v != %v", x, y) + } +} + +func TestGobEncodePtrError(t *testing.T) { + var err error + b := new(bytes.Buffer) + enc := NewEncoder(b) + err = enc.Encode(&err) + if err != nil { + t.Fatal("encode:", err) + } + dec := NewDecoder(b) + err2 := fmt.Errorf("foo") + err = dec.Decode(&err2) + if err != nil { + t.Fatal("decode:", err) + } + if err2 != nil { + t.Fatalf("expected nil, got %v", err2) + } +} + +func TestNetIP(t *testing.T) { + // Encoding of net.IP{1,2,3,4} in Go 1.1. + enc := []byte{0x07, 0x0a, 0x00, 0x04, 0x01, 0x02, 0x03, 0x04} + + var ip net.IP + err := NewDecoder(bytes.NewReader(enc)).Decode(&ip) + if err != nil { + t.Fatalf("decode: %v", err) + } + if ip.String() != "1.2.3.4" { + t.Errorf("decoded to %v, want 1.2.3.4", ip.String()) + } +} + +func TestIgnoreDepthLimit(t *testing.T) { + // We don't test the actual depth limit because it requires building an + // extremely large message, which takes quite a while. + oldNestingDepth := maxIgnoreNestingDepth + maxIgnoreNestingDepth = 100 + defer func() { maxIgnoreNestingDepth = oldNestingDepth }() + b := new(bytes.Buffer) + enc := NewEncoder(b) + typ := reflect.TypeOf(int(0)) + nested := reflect.ArrayOf(1, typ) + for i := 0; i < 100; i++ { + nested = reflect.ArrayOf(1, nested) + } + badStruct := reflect.New(reflect.StructOf([]reflect.StructField{{Name: "F", Type: nested}})) + enc.Encode(badStruct.Interface()) + dec := NewDecoder(b) + var output struct{ Hello int } + expectedErr := "invalid nesting depth" + if err := dec.Decode(&output); err == nil || err.Error() != expectedErr { + t.Errorf("Decode didn't fail with depth limit of 100: want %q, got %q", expectedErr, err) + } +} diff --git a/src/encoding/gob/timing_test.go b/src/encoding/gob/timing_test.go new file mode 100644 index 0000000..bdee39c --- /dev/null +++ b/src/encoding/gob/timing_test.go @@ -0,0 +1,328 @@ +// 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 gob + +import ( + "bytes" + "io" + "os" + "reflect" + "runtime" + "testing" +) + +type Bench struct { + A int + B float64 + C string + D []byte +} + +func benchmarkEndToEnd(b *testing.B, ctor func() any, pipe func() (r io.Reader, w io.Writer, err error)) { + b.RunParallel(func(pb *testing.PB) { + r, w, err := pipe() + if err != nil { + b.Fatal("can't get pipe:", err) + } + v := ctor() + enc := NewEncoder(w) + dec := NewDecoder(r) + for pb.Next() { + if err := enc.Encode(v); err != nil { + b.Fatal("encode error:", err) + } + if err := dec.Decode(v); err != nil { + b.Fatal("decode error:", err) + } + } + }) +} + +func BenchmarkEndToEndPipe(b *testing.B) { + benchmarkEndToEnd(b, func() any { + return &Bench{7, 3.2, "now is the time", bytes.Repeat([]byte("for all good men"), 100)} + }, func() (r io.Reader, w io.Writer, err error) { + r, w, err = os.Pipe() + return + }) +} + +func BenchmarkEndToEndByteBuffer(b *testing.B) { + benchmarkEndToEnd(b, func() any { + return &Bench{7, 3.2, "now is the time", bytes.Repeat([]byte("for all good men"), 100)} + }, func() (r io.Reader, w io.Writer, err error) { + var buf bytes.Buffer + return &buf, &buf, nil + }) +} + +func BenchmarkEndToEndSliceByteBuffer(b *testing.B) { + benchmarkEndToEnd(b, func() any { + v := &Bench{7, 3.2, "now is the time", nil} + Register(v) + arr := make([]any, 100) + for i := range arr { + arr[i] = v + } + return &arr + }, func() (r io.Reader, w io.Writer, err error) { + var buf bytes.Buffer + return &buf, &buf, nil + }) +} + +func TestCountEncodeMallocs(t *testing.T) { + if testing.Short() { + t.Skip("skipping malloc count in short mode") + } + if runtime.GOMAXPROCS(0) > 1 { + t.Skip("skipping; GOMAXPROCS>1") + } + + const N = 1000 + + var buf bytes.Buffer + enc := NewEncoder(&buf) + bench := &Bench{7, 3.2, "now is the time", []byte("for all good men")} + + allocs := testing.AllocsPerRun(N, func() { + err := enc.Encode(bench) + if err != nil { + t.Fatal("encode:", err) + } + }) + if allocs != 0 { + t.Fatalf("mallocs per encode of type Bench: %v; wanted 0\n", allocs) + } +} + +func TestCountDecodeMallocs(t *testing.T) { + if testing.Short() { + t.Skip("skipping malloc count in short mode") + } + if runtime.GOMAXPROCS(0) > 1 { + t.Skip("skipping; GOMAXPROCS>1") + } + + const N = 1000 + + var buf bytes.Buffer + enc := NewEncoder(&buf) + bench := &Bench{7, 3.2, "now is the time", []byte("for all good men")} + + // Fill the buffer with enough to decode + testing.AllocsPerRun(N, func() { + err := enc.Encode(bench) + if err != nil { + t.Fatal("encode:", err) + } + }) + + dec := NewDecoder(&buf) + allocs := testing.AllocsPerRun(N, func() { + *bench = Bench{} + err := dec.Decode(&bench) + if err != nil { + t.Fatal("decode:", err) + } + }) + if allocs != 3 { + t.Fatalf("mallocs per decode of type Bench: %v; wanted 3\n", allocs) + } +} + +func benchmarkEncodeSlice(b *testing.B, a any) { + b.ResetTimer() + b.RunParallel(func(pb *testing.PB) { + var buf bytes.Buffer + enc := NewEncoder(&buf) + + for pb.Next() { + buf.Reset() + err := enc.Encode(a) + if err != nil { + b.Fatal(err) + } + } + }) +} + +func BenchmarkEncodeComplex128Slice(b *testing.B) { + a := make([]complex128, 1000) + for i := range a { + a[i] = 1.2 + 3.4i + } + benchmarkEncodeSlice(b, a) +} + +func BenchmarkEncodeFloat64Slice(b *testing.B) { + a := make([]float64, 1000) + for i := range a { + a[i] = 1.23e4 + } + benchmarkEncodeSlice(b, a) +} + +func BenchmarkEncodeInt32Slice(b *testing.B) { + a := make([]int32, 1000) + for i := range a { + a[i] = int32(i * 100) + } + benchmarkEncodeSlice(b, a) +} + +func BenchmarkEncodeStringSlice(b *testing.B) { + a := make([]string, 1000) + for i := range a { + a[i] = "now is the time" + } + benchmarkEncodeSlice(b, a) +} + +func BenchmarkEncodeInterfaceSlice(b *testing.B) { + a := make([]any, 1000) + for i := range a { + a[i] = "now is the time" + } + benchmarkEncodeSlice(b, a) +} + +// benchmarkBuf is a read buffer we can reset +type benchmarkBuf struct { + offset int + data []byte +} + +func (b *benchmarkBuf) Read(p []byte) (n int, err error) { + n = copy(p, b.data[b.offset:]) + if n == 0 { + return 0, io.EOF + } + b.offset += n + return +} + +func (b *benchmarkBuf) ReadByte() (c byte, err error) { + if b.offset >= len(b.data) { + return 0, io.EOF + } + c = b.data[b.offset] + b.offset++ + return +} + +func (b *benchmarkBuf) reset() { + b.offset = 0 +} + +func benchmarkDecodeSlice(b *testing.B, a any) { + var buf bytes.Buffer + enc := NewEncoder(&buf) + err := enc.Encode(a) + if err != nil { + b.Fatal(err) + } + + ra := reflect.ValueOf(a) + rt := ra.Type() + b.ResetTimer() + + b.RunParallel(func(pb *testing.PB) { + // TODO(#19025): Move per-thread allocation before ResetTimer. + rp := reflect.New(rt) + rp.Elem().Set(reflect.MakeSlice(rt, ra.Len(), ra.Cap())) + p := rp.Interface() + + bbuf := benchmarkBuf{data: buf.Bytes()} + + for pb.Next() { + bbuf.reset() + dec := NewDecoder(&bbuf) + err := dec.Decode(p) + if err != nil { + b.Fatal(err) + } + } + }) +} + +func BenchmarkDecodeComplex128Slice(b *testing.B) { + a := make([]complex128, 1000) + for i := range a { + a[i] = 1.2 + 3.4i + } + benchmarkDecodeSlice(b, a) +} + +func BenchmarkDecodeFloat64Slice(b *testing.B) { + a := make([]float64, 1000) + for i := range a { + a[i] = 1.23e4 + } + benchmarkDecodeSlice(b, a) +} + +func BenchmarkDecodeInt32Slice(b *testing.B) { + a := make([]int32, 1000) + for i := range a { + a[i] = 1234 + } + benchmarkDecodeSlice(b, a) +} + +func BenchmarkDecodeStringSlice(b *testing.B) { + a := make([]string, 1000) + for i := range a { + a[i] = "now is the time" + } + benchmarkDecodeSlice(b, a) +} +func BenchmarkDecodeStringsSlice(b *testing.B) { + a := make([][]string, 1000) + for i := range a { + a[i] = []string{"now is the time"} + } + benchmarkDecodeSlice(b, a) +} +func BenchmarkDecodeBytesSlice(b *testing.B) { + a := make([][]byte, 1000) + for i := range a { + a[i] = []byte("now is the time") + } + benchmarkDecodeSlice(b, a) +} + +func BenchmarkDecodeInterfaceSlice(b *testing.B) { + a := make([]any, 1000) + for i := range a { + a[i] = "now is the time" + } + benchmarkDecodeSlice(b, a) +} + +func BenchmarkDecodeMap(b *testing.B) { + count := 1000 + m := make(map[int]int, count) + for i := 0; i < count; i++ { + m[i] = i + } + var buf bytes.Buffer + enc := NewEncoder(&buf) + err := enc.Encode(m) + if err != nil { + b.Fatal(err) + } + bbuf := benchmarkBuf{data: buf.Bytes()} + b.ResetTimer() + for i := 0; i < b.N; i++ { + var rm map[int]int + bbuf.reset() + dec := NewDecoder(&bbuf) + err := dec.Decode(&rm) + if err != nil { + b.Fatal(i, err) + } + } +} diff --git a/src/encoding/gob/type.go b/src/encoding/gob/type.go new file mode 100644 index 0000000..3114cb0 --- /dev/null +++ b/src/encoding/gob/type.go @@ -0,0 +1,913 @@ +// 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 gob + +import ( + "encoding" + "errors" + "fmt" + "os" + "reflect" + "sync" + "sync/atomic" + "unicode" + "unicode/utf8" +) + +// userTypeInfo stores the information associated with a type the user has handed +// to the package. It's computed once and stored in a map keyed by reflection +// type. +type userTypeInfo struct { + user reflect.Type // the type the user handed us + base reflect.Type // the base type after all indirections + indir int // number of indirections to reach the base type + externalEnc int // xGob, xBinary, or xText + externalDec int // xGob, xBinary or xText + encIndir int8 // number of indirections to reach the receiver type; may be negative + decIndir int8 // number of indirections to reach the receiver type; may be negative +} + +// externalEncoding bits +const ( + xGob = 1 + iota // GobEncoder or GobDecoder + xBinary // encoding.BinaryMarshaler or encoding.BinaryUnmarshaler + xText // encoding.TextMarshaler or encoding.TextUnmarshaler +) + +var userTypeCache sync.Map // map[reflect.Type]*userTypeInfo + +// validUserType returns, and saves, the information associated with user-provided type rt. +// If the user type is not valid, err will be non-nil. To be used when the error handler +// is not set up. +func validUserType(rt reflect.Type) (*userTypeInfo, error) { + if ui, ok := userTypeCache.Load(rt); ok { + return ui.(*userTypeInfo), nil + } + + // Construct a new userTypeInfo and atomically add it to the userTypeCache. + // If we lose the race, we'll waste a little CPU and create a little garbage + // but return the existing value anyway. + + ut := new(userTypeInfo) + ut.base = rt + ut.user = rt + // A type that is just a cycle of pointers (such as type T *T) cannot + // be represented in gobs, which need some concrete data. We use a + // cycle detection algorithm from Knuth, Vol 2, Section 3.1, Ex 6, + // pp 539-540. As we step through indirections, run another type at + // half speed. If they meet up, there's a cycle. + slowpoke := ut.base // walks half as fast as ut.base + for { + pt := ut.base + if pt.Kind() != reflect.Pointer { + break + } + ut.base = pt.Elem() + if ut.base == slowpoke { // ut.base lapped slowpoke + // recursive pointer type. + return nil, errors.New("can't represent recursive pointer type " + ut.base.String()) + } + if ut.indir%2 == 0 { + slowpoke = slowpoke.Elem() + } + ut.indir++ + } + + if ok, indir := implementsInterface(ut.user, gobEncoderInterfaceType); ok { + ut.externalEnc, ut.encIndir = xGob, indir + } else if ok, indir := implementsInterface(ut.user, binaryMarshalerInterfaceType); ok { + ut.externalEnc, ut.encIndir = xBinary, indir + } + + // NOTE(rsc): Would like to allow MarshalText here, but results in incompatibility + // with older encodings for net.IP. See golang.org/issue/6760. + // } else if ok, indir := implementsInterface(ut.user, textMarshalerInterfaceType); ok { + // ut.externalEnc, ut.encIndir = xText, indir + // } + + if ok, indir := implementsInterface(ut.user, gobDecoderInterfaceType); ok { + ut.externalDec, ut.decIndir = xGob, indir + } else if ok, indir := implementsInterface(ut.user, binaryUnmarshalerInterfaceType); ok { + ut.externalDec, ut.decIndir = xBinary, indir + } + + // See note above. + // } else if ok, indir := implementsInterface(ut.user, textUnmarshalerInterfaceType); ok { + // ut.externalDec, ut.decIndir = xText, indir + // } + + ui, _ := userTypeCache.LoadOrStore(rt, ut) + return ui.(*userTypeInfo), nil +} + +var ( + gobEncoderInterfaceType = reflect.TypeOf((*GobEncoder)(nil)).Elem() + gobDecoderInterfaceType = reflect.TypeOf((*GobDecoder)(nil)).Elem() + binaryMarshalerInterfaceType = reflect.TypeOf((*encoding.BinaryMarshaler)(nil)).Elem() + binaryUnmarshalerInterfaceType = reflect.TypeOf((*encoding.BinaryUnmarshaler)(nil)).Elem() + textMarshalerInterfaceType = reflect.TypeOf((*encoding.TextMarshaler)(nil)).Elem() + textUnmarshalerInterfaceType = reflect.TypeOf((*encoding.TextUnmarshaler)(nil)).Elem() +) + +// implementsInterface reports whether the type implements the +// gobEncoder/gobDecoder interface. +// It also returns the number of indirections required to get to the +// implementation. +func implementsInterface(typ, gobEncDecType reflect.Type) (success bool, indir int8) { + if typ == nil { + return + } + rt := typ + // The type might be a pointer and we need to keep + // dereferencing to the base type until we find an implementation. + for { + if rt.Implements(gobEncDecType) { + return true, indir + } + if p := rt; p.Kind() == reflect.Pointer { + indir++ + if indir > 100 { // insane number of indirections + return false, 0 + } + rt = p.Elem() + continue + } + break + } + // No luck yet, but if this is a base type (non-pointer), the pointer might satisfy. + if typ.Kind() != reflect.Pointer { + // Not a pointer, but does the pointer work? + if reflect.PointerTo(typ).Implements(gobEncDecType) { + return true, -1 + } + } + return false, 0 +} + +// userType returns, and saves, the information associated with user-provided type rt. +// If the user type is not valid, it calls error. +func userType(rt reflect.Type) *userTypeInfo { + ut, err := validUserType(rt) + if err != nil { + error_(err) + } + return ut +} + +// A typeId represents a gob Type as an integer that can be passed on the wire. +// Internally, typeIds are used as keys to a map to recover the underlying type info. +type typeId int32 + +var nextId typeId // incremented for each new type we build +var typeLock sync.Mutex // set while building a type +const firstUserId = 64 // lowest id number granted to user + +type gobType interface { + id() typeId + setId(id typeId) + name() string + string() string // not public; only for debugging + safeString(seen map[typeId]bool) string +} + +var types = make(map[reflect.Type]gobType) +var idToType = make(map[typeId]gobType) +var builtinIdToType map[typeId]gobType // set in init() after builtins are established + +func setTypeId(typ gobType) { + // When building recursive types, someone may get there before us. + if typ.id() != 0 { + return + } + nextId++ + typ.setId(nextId) + idToType[nextId] = typ +} + +func (t typeId) gobType() gobType { + if t == 0 { + return nil + } + return idToType[t] +} + +// string returns the string representation of the type associated with the typeId. +func (t typeId) string() string { + if t.gobType() == nil { + return "" + } + return t.gobType().string() +} + +// Name returns the name of the type associated with the typeId. +func (t typeId) name() string { + if t.gobType() == nil { + return "" + } + return t.gobType().name() +} + +// CommonType holds elements of all types. +// It is a historical artifact, kept for binary compatibility and exported +// only for the benefit of the package's encoding of type descriptors. It is +// not intended for direct use by clients. +type CommonType struct { + Name string + Id typeId +} + +func (t *CommonType) id() typeId { return t.Id } + +func (t *CommonType) setId(id typeId) { t.Id = id } + +func (t *CommonType) string() string { return t.Name } + +func (t *CommonType) safeString(seen map[typeId]bool) string { + return t.Name +} + +func (t *CommonType) name() string { return t.Name } + +// Create and check predefined types +// The string for tBytes is "bytes" not "[]byte" to signify its specialness. + +var ( + // Primordial types, needed during initialization. + // Always passed as pointers so the interface{} type + // goes through without losing its interfaceness. + tBool = bootstrapType("bool", (*bool)(nil), 1) + tInt = bootstrapType("int", (*int)(nil), 2) + tUint = bootstrapType("uint", (*uint)(nil), 3) + tFloat = bootstrapType("float", (*float64)(nil), 4) + tBytes = bootstrapType("bytes", (*[]byte)(nil), 5) + tString = bootstrapType("string", (*string)(nil), 6) + tComplex = bootstrapType("complex", (*complex128)(nil), 7) + tInterface = bootstrapType("interface", (*any)(nil), 8) + // Reserve some Ids for compatible expansion + tReserved7 = bootstrapType("_reserved1", (*struct{ r7 int })(nil), 9) + tReserved6 = bootstrapType("_reserved1", (*struct{ r6 int })(nil), 10) + tReserved5 = bootstrapType("_reserved1", (*struct{ r5 int })(nil), 11) + tReserved4 = bootstrapType("_reserved1", (*struct{ r4 int })(nil), 12) + tReserved3 = bootstrapType("_reserved1", (*struct{ r3 int })(nil), 13) + tReserved2 = bootstrapType("_reserved1", (*struct{ r2 int })(nil), 14) + tReserved1 = bootstrapType("_reserved1", (*struct{ r1 int })(nil), 15) +) + +// Predefined because it's needed by the Decoder +var tWireType = mustGetTypeInfo(reflect.TypeOf(wireType{})).id +var wireTypeUserInfo *userTypeInfo // userTypeInfo of (*wireType) + +func init() { + // Some magic numbers to make sure there are no surprises. + checkId(16, tWireType) + checkId(17, mustGetTypeInfo(reflect.TypeOf(arrayType{})).id) + checkId(18, mustGetTypeInfo(reflect.TypeOf(CommonType{})).id) + checkId(19, mustGetTypeInfo(reflect.TypeOf(sliceType{})).id) + checkId(20, mustGetTypeInfo(reflect.TypeOf(structType{})).id) + checkId(21, mustGetTypeInfo(reflect.TypeOf(fieldType{})).id) + checkId(23, mustGetTypeInfo(reflect.TypeOf(mapType{})).id) + + builtinIdToType = make(map[typeId]gobType) + for k, v := range idToType { + builtinIdToType[k] = v + } + + // Move the id space upwards to allow for growth in the predefined world + // without breaking existing files. + if nextId > firstUserId { + panic(fmt.Sprintln("nextId too large:", nextId)) + } + nextId = firstUserId + registerBasics() + wireTypeUserInfo = userType(reflect.TypeOf((*wireType)(nil))) +} + +// Array type +type arrayType struct { + CommonType + Elem typeId + Len int +} + +func newArrayType(name string) *arrayType { + a := &arrayType{CommonType{Name: name}, 0, 0} + return a +} + +func (a *arrayType) init(elem gobType, len int) { + // Set our type id before evaluating the element's, in case it's our own. + setTypeId(a) + a.Elem = elem.id() + a.Len = len +} + +func (a *arrayType) safeString(seen map[typeId]bool) string { + if seen[a.Id] { + return a.Name + } + seen[a.Id] = true + return fmt.Sprintf("[%d]%s", a.Len, a.Elem.gobType().safeString(seen)) +} + +func (a *arrayType) string() string { return a.safeString(make(map[typeId]bool)) } + +// GobEncoder type (something that implements the GobEncoder interface) +type gobEncoderType struct { + CommonType +} + +func newGobEncoderType(name string) *gobEncoderType { + g := &gobEncoderType{CommonType{Name: name}} + setTypeId(g) + return g +} + +func (g *gobEncoderType) safeString(seen map[typeId]bool) string { + return g.Name +} + +func (g *gobEncoderType) string() string { return g.Name } + +// Map type +type mapType struct { + CommonType + Key typeId + Elem typeId +} + +func newMapType(name string) *mapType { + m := &mapType{CommonType{Name: name}, 0, 0} + return m +} + +func (m *mapType) init(key, elem gobType) { + // Set our type id before evaluating the element's, in case it's our own. + setTypeId(m) + m.Key = key.id() + m.Elem = elem.id() +} + +func (m *mapType) safeString(seen map[typeId]bool) string { + if seen[m.Id] { + return m.Name + } + seen[m.Id] = true + key := m.Key.gobType().safeString(seen) + elem := m.Elem.gobType().safeString(seen) + return fmt.Sprintf("map[%s]%s", key, elem) +} + +func (m *mapType) string() string { return m.safeString(make(map[typeId]bool)) } + +// Slice type +type sliceType struct { + CommonType + Elem typeId +} + +func newSliceType(name string) *sliceType { + s := &sliceType{CommonType{Name: name}, 0} + return s +} + +func (s *sliceType) init(elem gobType) { + // Set our type id before evaluating the element's, in case it's our own. + setTypeId(s) + // See the comments about ids in newTypeObject. Only slices and + // structs have mutual recursion. + if elem.id() == 0 { + setTypeId(elem) + } + s.Elem = elem.id() +} + +func (s *sliceType) safeString(seen map[typeId]bool) string { + if seen[s.Id] { + return s.Name + } + seen[s.Id] = true + return fmt.Sprintf("[]%s", s.Elem.gobType().safeString(seen)) +} + +func (s *sliceType) string() string { return s.safeString(make(map[typeId]bool)) } + +// Struct type +type fieldType struct { + Name string + Id typeId +} + +type structType struct { + CommonType + Field []*fieldType +} + +func (s *structType) safeString(seen map[typeId]bool) string { + if s == nil { + return "" + } + if _, ok := seen[s.Id]; ok { + return s.Name + } + seen[s.Id] = true + str := s.Name + " = struct { " + for _, f := range s.Field { + str += fmt.Sprintf("%s %s; ", f.Name, f.Id.gobType().safeString(seen)) + } + str += "}" + return str +} + +func (s *structType) string() string { return s.safeString(make(map[typeId]bool)) } + +func newStructType(name string) *structType { + s := &structType{CommonType{Name: name}, nil} + // For historical reasons we set the id here rather than init. + // See the comment in newTypeObject for details. + setTypeId(s) + return s +} + +// newTypeObject allocates a gobType for the reflection type rt. +// Unless ut represents a GobEncoder, rt should be the base type +// of ut. +// This is only called from the encoding side. The decoding side +// works through typeIds and userTypeInfos alone. +func newTypeObject(name string, ut *userTypeInfo, rt reflect.Type) (gobType, error) { + // Does this type implement GobEncoder? + if ut.externalEnc != 0 { + return newGobEncoderType(name), nil + } + var err error + var type0, type1 gobType + defer func() { + if err != nil { + delete(types, rt) + } + }() + // Install the top-level type before the subtypes (e.g. struct before + // fields) so recursive types can be constructed safely. + switch t := rt; t.Kind() { + // All basic types are easy: they are predefined. + case reflect.Bool: + return tBool.gobType(), nil + + case reflect.Int, reflect.Int8, reflect.Int16, reflect.Int32, reflect.Int64: + return tInt.gobType(), nil + + case reflect.Uint, reflect.Uint8, reflect.Uint16, reflect.Uint32, reflect.Uint64, reflect.Uintptr: + return tUint.gobType(), nil + + case reflect.Float32, reflect.Float64: + return tFloat.gobType(), nil + + case reflect.Complex64, reflect.Complex128: + return tComplex.gobType(), nil + + case reflect.String: + return tString.gobType(), nil + + case reflect.Interface: + return tInterface.gobType(), nil + + case reflect.Array: + at := newArrayType(name) + types[rt] = at + type0, err = getBaseType("", t.Elem()) + if err != nil { + return nil, err + } + // Historical aside: + // For arrays, maps, and slices, we set the type id after the elements + // are constructed. This is to retain the order of type id allocation after + // a fix made to handle recursive types, which changed the order in + // which types are built. Delaying the setting in this way preserves + // type ids while allowing recursive types to be described. Structs, + // done below, were already handling recursion correctly so they + // assign the top-level id before those of the field. + at.init(type0, t.Len()) + return at, nil + + case reflect.Map: + mt := newMapType(name) + types[rt] = mt + type0, err = getBaseType("", t.Key()) + if err != nil { + return nil, err + } + type1, err = getBaseType("", t.Elem()) + if err != nil { + return nil, err + } + mt.init(type0, type1) + return mt, nil + + case reflect.Slice: + // []byte == []uint8 is a special case + if t.Elem().Kind() == reflect.Uint8 { + return tBytes.gobType(), nil + } + st := newSliceType(name) + types[rt] = st + type0, err = getBaseType(t.Elem().Name(), t.Elem()) + if err != nil { + return nil, err + } + st.init(type0) + return st, nil + + case reflect.Struct: + st := newStructType(name) + types[rt] = st + idToType[st.id()] = st + for i := 0; i < t.NumField(); i++ { + f := t.Field(i) + if !isSent(&f) { + continue + } + typ := userType(f.Type).base + tname := typ.Name() + if tname == "" { + t := userType(f.Type).base + tname = t.String() + } + gt, err := getBaseType(tname, f.Type) + if err != nil { + return nil, err + } + // Some mutually recursive types can cause us to be here while + // still defining the element. Fix the element type id here. + // We could do this more neatly by setting the id at the start of + // building every type, but that would break binary compatibility. + if gt.id() == 0 { + setTypeId(gt) + } + st.Field = append(st.Field, &fieldType{f.Name, gt.id()}) + } + return st, nil + + default: + return nil, errors.New("gob NewTypeObject can't handle type: " + rt.String()) + } +} + +// isExported reports whether this is an exported - upper case - name. +func isExported(name string) bool { + rune, _ := utf8.DecodeRuneInString(name) + return unicode.IsUpper(rune) +} + +// isSent reports whether this struct field is to be transmitted. +// It will be transmitted only if it is exported and not a chan or func field +// or pointer to chan or func. +func isSent(field *reflect.StructField) bool { + if !isExported(field.Name) { + return false + } + // If the field is a chan or func or pointer thereto, don't send it. + // That is, treat it like an unexported field. + typ := field.Type + for typ.Kind() == reflect.Pointer { + typ = typ.Elem() + } + if typ.Kind() == reflect.Chan || typ.Kind() == reflect.Func { + return false + } + return true +} + +// getBaseType returns the Gob type describing the given reflect.Type's base type. +// typeLock must be held. +func getBaseType(name string, rt reflect.Type) (gobType, error) { + ut := userType(rt) + return getType(name, ut, ut.base) +} + +// getType returns the Gob type describing the given reflect.Type. +// Should be called only when handling GobEncoders/Decoders, +// which may be pointers. All other types are handled through the +// base type, never a pointer. +// typeLock must be held. +func getType(name string, ut *userTypeInfo, rt reflect.Type) (gobType, error) { + typ, present := types[rt] + if present { + return typ, nil + } + typ, err := newTypeObject(name, ut, rt) + if err == nil { + types[rt] = typ + } + return typ, err +} + +func checkId(want, got typeId) { + if want != got { + fmt.Fprintf(os.Stderr, "checkId: %d should be %d\n", int(got), int(want)) + panic("bootstrap type wrong id: " + got.name() + " " + got.string() + " not " + want.string()) + } +} + +// used for building the basic types; called only from init(). the incoming +// interface always refers to a pointer. +func bootstrapType(name string, e any, expect typeId) typeId { + rt := reflect.TypeOf(e).Elem() + _, present := types[rt] + if present { + panic("bootstrap type already present: " + name + ", " + rt.String()) + } + typ := &CommonType{Name: name} + types[rt] = typ + setTypeId(typ) + checkId(expect, nextId) + userType(rt) // might as well cache it now + return nextId +} + +// Representation of the information we send and receive about this type. +// Each value we send is preceded by its type definition: an encoded int. +// However, the very first time we send the value, we first send the pair +// (-id, wireType). +// For bootstrapping purposes, we assume that the recipient knows how +// to decode a wireType; it is exactly the wireType struct here, interpreted +// using the gob rules for sending a structure, except that we assume the +// ids for wireType and structType etc. are known. The relevant pieces +// are built in encode.go's init() function. +// To maintain binary compatibility, if you extend this type, always put +// the new fields last. +type wireType struct { + ArrayT *arrayType + SliceT *sliceType + StructT *structType + MapT *mapType + GobEncoderT *gobEncoderType + BinaryMarshalerT *gobEncoderType + TextMarshalerT *gobEncoderType +} + +func (w *wireType) string() string { + const unknown = "unknown type" + if w == nil { + return unknown + } + switch { + case w.ArrayT != nil: + return w.ArrayT.Name + case w.SliceT != nil: + return w.SliceT.Name + case w.StructT != nil: + return w.StructT.Name + case w.MapT != nil: + return w.MapT.Name + case w.GobEncoderT != nil: + return w.GobEncoderT.Name + case w.BinaryMarshalerT != nil: + return w.BinaryMarshalerT.Name + case w.TextMarshalerT != nil: + return w.TextMarshalerT.Name + } + return unknown +} + +type typeInfo struct { + id typeId + encInit sync.Mutex // protects creation of encoder + encoder atomic.Pointer[encEngine] + wire *wireType +} + +// typeInfoMap is an atomic pointer to map[reflect.Type]*typeInfo. +// It's updated copy-on-write. Readers just do an atomic load +// to get the current version of the map. Writers make a full copy of +// the map and atomically update the pointer to point to the new map. +// Under heavy read contention, this is significantly faster than a map +// protected by a mutex. +var typeInfoMap atomic.Value + +func lookupTypeInfo(rt reflect.Type) *typeInfo { + m, _ := typeInfoMap.Load().(map[reflect.Type]*typeInfo) + return m[rt] +} + +func getTypeInfo(ut *userTypeInfo) (*typeInfo, error) { + rt := ut.base + if ut.externalEnc != 0 { + // We want the user type, not the base type. + rt = ut.user + } + if info := lookupTypeInfo(rt); info != nil { + return info, nil + } + return buildTypeInfo(ut, rt) +} + +// buildTypeInfo constructs the type information for the type +// and stores it in the type info map. +func buildTypeInfo(ut *userTypeInfo, rt reflect.Type) (*typeInfo, error) { + typeLock.Lock() + defer typeLock.Unlock() + + if info := lookupTypeInfo(rt); info != nil { + return info, nil + } + + gt, err := getBaseType(rt.Name(), rt) + if err != nil { + return nil, err + } + info := &typeInfo{id: gt.id()} + + if ut.externalEnc != 0 { + userType, err := getType(rt.Name(), ut, rt) + if err != nil { + return nil, err + } + gt := userType.id().gobType().(*gobEncoderType) + switch ut.externalEnc { + case xGob: + info.wire = &wireType{GobEncoderT: gt} + case xBinary: + info.wire = &wireType{BinaryMarshalerT: gt} + case xText: + info.wire = &wireType{TextMarshalerT: gt} + } + rt = ut.user + } else { + t := info.id.gobType() + switch typ := rt; typ.Kind() { + case reflect.Array: + info.wire = &wireType{ArrayT: t.(*arrayType)} + case reflect.Map: + info.wire = &wireType{MapT: t.(*mapType)} + case reflect.Slice: + // []byte == []uint8 is a special case handled separately + if typ.Elem().Kind() != reflect.Uint8 { + info.wire = &wireType{SliceT: t.(*sliceType)} + } + case reflect.Struct: + info.wire = &wireType{StructT: t.(*structType)} + } + } + + // Create new map with old contents plus new entry. + newm := make(map[reflect.Type]*typeInfo) + m, _ := typeInfoMap.Load().(map[reflect.Type]*typeInfo) + for k, v := range m { + newm[k] = v + } + newm[rt] = info + typeInfoMap.Store(newm) + return info, nil +} + +// Called only when a panic is acceptable and unexpected. +func mustGetTypeInfo(rt reflect.Type) *typeInfo { + t, err := getTypeInfo(userType(rt)) + if err != nil { + panic("getTypeInfo: " + err.Error()) + } + return t +} + +// GobEncoder is the interface describing data that provides its own +// representation for encoding values for transmission to a GobDecoder. +// A type that implements GobEncoder and GobDecoder has complete +// control over the representation of its data and may therefore +// contain things such as private fields, channels, and functions, +// which are not usually transmissible in gob streams. +// +// Note: Since gobs can be stored permanently, it is good design +// to guarantee the encoding used by a GobEncoder is stable as the +// software evolves. For instance, it might make sense for GobEncode +// to include a version number in the encoding. +type GobEncoder interface { + // GobEncode returns a byte slice representing the encoding of the + // receiver for transmission to a GobDecoder, usually of the same + // concrete type. + GobEncode() ([]byte, error) +} + +// GobDecoder is the interface describing data that provides its own +// routine for decoding transmitted values sent by a GobEncoder. +type GobDecoder interface { + // GobDecode overwrites the receiver, which must be a pointer, + // with the value represented by the byte slice, which was written + // by GobEncode, usually for the same concrete type. + GobDecode([]byte) error +} + +var ( + nameToConcreteType sync.Map // map[string]reflect.Type + concreteTypeToName sync.Map // map[reflect.Type]string +) + +// RegisterName is like Register but uses the provided name rather than the +// type's default. +func RegisterName(name string, value any) { + if name == "" { + // reserved for nil + panic("attempt to register empty name") + } + + ut := userType(reflect.TypeOf(value)) + + // Check for incompatible duplicates. The name must refer to the + // same user type, and vice versa. + + // Store the name and type provided by the user.... + if t, dup := nameToConcreteType.LoadOrStore(name, reflect.TypeOf(value)); dup && t != ut.user { + panic(fmt.Sprintf("gob: registering duplicate types for %q: %s != %s", name, t, ut.user)) + } + + // but the flattened type in the type table, since that's what decode needs. + if n, dup := concreteTypeToName.LoadOrStore(ut.base, name); dup && n != name { + nameToConcreteType.Delete(name) + panic(fmt.Sprintf("gob: registering duplicate names for %s: %q != %q", ut.user, n, name)) + } +} + +// Register records a type, identified by a value for that type, under its +// internal type name. That name will identify the concrete type of a value +// sent or received as an interface variable. Only types that will be +// transferred as implementations of interface values need to be registered. +// Expecting to be used only during initialization, it panics if the mapping +// between types and names is not a bijection. +func Register(value any) { + // Default to printed representation for unnamed types + rt := reflect.TypeOf(value) + name := rt.String() + + // But for named types (or pointers to them), qualify with import path (but see inner comment). + // Dereference one pointer looking for a named type. + star := "" + if rt.Name() == "" { + if pt := rt; pt.Kind() == reflect.Pointer { + star = "*" + // NOTE: The following line should be rt = pt.Elem() to implement + // what the comment above claims, but fixing it would break compatibility + // with existing gobs. + // + // Given package p imported as "full/p" with these definitions: + // package p + // type T1 struct { ... } + // this table shows the intended and actual strings used by gob to + // name the types: + // + // Type Correct string Actual string + // + // T1 full/p.T1 full/p.T1 + // *T1 *full/p.T1 *p.T1 + // + // The missing full path cannot be fixed without breaking existing gob decoders. + rt = pt + } + } + if rt.Name() != "" { + if rt.PkgPath() == "" { + name = star + rt.Name() + } else { + name = star + rt.PkgPath() + "." + rt.Name() + } + } + + RegisterName(name, value) +} + +func registerBasics() { + Register(int(0)) + Register(int8(0)) + Register(int16(0)) + Register(int32(0)) + Register(int64(0)) + Register(uint(0)) + Register(uint8(0)) + Register(uint16(0)) + Register(uint32(0)) + Register(uint64(0)) + Register(float32(0)) + Register(float64(0)) + Register(complex64(0i)) + Register(complex128(0i)) + Register(uintptr(0)) + Register(false) + Register("") + Register([]byte(nil)) + Register([]int(nil)) + Register([]int8(nil)) + Register([]int16(nil)) + Register([]int32(nil)) + Register([]int64(nil)) + Register([]uint(nil)) + Register([]uint8(nil)) + Register([]uint16(nil)) + Register([]uint32(nil)) + Register([]uint64(nil)) + Register([]float32(nil)) + Register([]float64(nil)) + Register([]complex64(nil)) + Register([]complex128(nil)) + Register([]uintptr(nil)) + Register([]bool(nil)) + Register([]string(nil)) +} diff --git a/src/encoding/gob/type_test.go b/src/encoding/gob/type_test.go new file mode 100644 index 0000000..f5f8db8 --- /dev/null +++ b/src/encoding/gob/type_test.go @@ -0,0 +1,262 @@ +// 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 gob + +import ( + "bytes" + "reflect" + "sync" + "testing" +) + +type typeT struct { + id typeId + str string +} + +var basicTypes = []typeT{ + {tBool, "bool"}, + {tInt, "int"}, + {tUint, "uint"}, + {tFloat, "float"}, + {tBytes, "bytes"}, + {tString, "string"}, +} + +func getTypeUnlocked(name string, rt reflect.Type) gobType { + typeLock.Lock() + defer typeLock.Unlock() + t, err := getBaseType(name, rt) + if err != nil { + panic("getTypeUnlocked: " + err.Error()) + } + return t +} + +// Sanity checks +func TestBasic(t *testing.T) { + for _, tt := range basicTypes { + if tt.id.string() != tt.str { + t.Errorf("checkType: expected %q got %s", tt.str, tt.id.string()) + } + if tt.id == 0 { + t.Errorf("id for %q is zero", tt.str) + } + } +} + +// Reregister some basic types to check registration is idempotent. +func TestReregistration(t *testing.T) { + newtyp := getTypeUnlocked("int", reflect.TypeOf(int(0))) + if newtyp != tInt.gobType() { + t.Errorf("reregistration of %s got new type", newtyp.string()) + } + newtyp = getTypeUnlocked("uint", reflect.TypeOf(uint(0))) + if newtyp != tUint.gobType() { + t.Errorf("reregistration of %s got new type", newtyp.string()) + } + newtyp = getTypeUnlocked("string", reflect.TypeOf("hello")) + if newtyp != tString.gobType() { + t.Errorf("reregistration of %s got new type", newtyp.string()) + } +} + +func TestArrayType(t *testing.T) { + var a3 [3]int + a3int := getTypeUnlocked("foo", reflect.TypeOf(a3)) + newa3int := getTypeUnlocked("bar", reflect.TypeOf(a3)) + if a3int != newa3int { + t.Errorf("second registration of [3]int creates new type") + } + var a4 [4]int + a4int := getTypeUnlocked("goo", reflect.TypeOf(a4)) + if a3int == a4int { + t.Errorf("registration of [3]int creates same type as [4]int") + } + var b3 [3]bool + a3bool := getTypeUnlocked("", reflect.TypeOf(b3)) + if a3int == a3bool { + t.Errorf("registration of [3]bool creates same type as [3]int") + } + str := a3bool.string() + expected := "[3]bool" + if str != expected { + t.Errorf("array printed as %q; expected %q", str, expected) + } +} + +func TestSliceType(t *testing.T) { + var s []int + sint := getTypeUnlocked("slice", reflect.TypeOf(s)) + var news []int + newsint := getTypeUnlocked("slice1", reflect.TypeOf(news)) + if sint != newsint { + t.Errorf("second registration of []int creates new type") + } + var b []bool + sbool := getTypeUnlocked("", reflect.TypeOf(b)) + if sbool == sint { + t.Errorf("registration of []bool creates same type as []int") + } + str := sbool.string() + expected := "[]bool" + if str != expected { + t.Errorf("slice printed as %q; expected %q", str, expected) + } +} + +func TestMapType(t *testing.T) { + var m map[string]int + mapStringInt := getTypeUnlocked("map", reflect.TypeOf(m)) + var newm map[string]int + newMapStringInt := getTypeUnlocked("map1", reflect.TypeOf(newm)) + if mapStringInt != newMapStringInt { + t.Errorf("second registration of map[string]int creates new type") + } + var b map[string]bool + mapStringBool := getTypeUnlocked("", reflect.TypeOf(b)) + if mapStringBool == mapStringInt { + t.Errorf("registration of map[string]bool creates same type as map[string]int") + } + str := mapStringBool.string() + expected := "map[string]bool" + if str != expected { + t.Errorf("map printed as %q; expected %q", str, expected) + } +} + +type Bar struct { + X string +} + +// This structure has pointers and refers to itself, making it a good test case. +type Foo struct { + A int + B int32 // will become int + C string + D []byte + E *float64 // will become float64 + F ****float64 // will become float64 + G *Bar + H *Bar // should not interpolate the definition of Bar again + I *Foo // will not explode +} + +func TestStructType(t *testing.T) { + sstruct := getTypeUnlocked("Foo", reflect.TypeOf(Foo{})) + str := sstruct.string() + // If we can print it correctly, we built it correctly. + expected := "Foo = struct { A int; B int; C string; D bytes; E float; F float; G Bar = struct { X string; }; H Bar; I Foo; }" + if str != expected { + t.Errorf("struct printed as %q; expected %q", str, expected) + } +} + +// Should be OK to register the same type multiple times, as long as they're +// at the same level of indirection. +func TestRegistration(t *testing.T) { + type T struct{ a int } + Register(new(T)) + Register(new(T)) +} + +type N1 struct{} +type N2 struct{} + +// See comment in type.go/Register. +func TestRegistrationNaming(t *testing.T) { + testCases := []struct { + t any + name string + }{ + {&N1{}, "*gob.N1"}, + {N2{}, "encoding/gob.N2"}, + } + + for _, tc := range testCases { + Register(tc.t) + + tct := reflect.TypeOf(tc.t) + ct, _ := nameToConcreteType.Load(tc.name) + if ct != tct { + t.Errorf("nameToConcreteType[%q] = %v, want %v", tc.name, ct, tct) + } + // concreteTypeToName is keyed off the base type. + if tct.Kind() == reflect.Pointer { + tct = tct.Elem() + } + if n, _ := concreteTypeToName.Load(tct); n != tc.name { + t.Errorf("concreteTypeToName[%v] got %v, want %v", tct, n, tc.name) + } + } +} + +func TestStressParallel(t *testing.T) { + type T2 struct{ A int } + c := make(chan bool) + const N = 10 + for i := 0; i < N; i++ { + go func() { + p := new(T2) + Register(p) + b := new(bytes.Buffer) + enc := NewEncoder(b) + err := enc.Encode(p) + if err != nil { + t.Error("encoder fail:", err) + } + dec := NewDecoder(b) + err = dec.Decode(p) + if err != nil { + t.Error("decoder fail:", err) + } + c <- true + }() + } + for i := 0; i < N; i++ { + <-c + } +} + +// Issue 23328. Note that this test name is known to cmd/dist/test.go. +func TestTypeRace(t *testing.T) { + c := make(chan bool) + var wg sync.WaitGroup + for i := 0; i < 2; i++ { + wg.Add(1) + go func(i int) { + defer wg.Done() + var buf bytes.Buffer + enc := NewEncoder(&buf) + dec := NewDecoder(&buf) + var x any + switch i { + case 0: + x = &N1{} + case 1: + x = &N2{} + default: + t.Errorf("bad i %d", i) + return + } + m := make(map[string]string) + <-c + if err := enc.Encode(x); err != nil { + t.Error(err) + return + } + if err := enc.Encode(x); err != nil { + t.Error(err) + return + } + if err := dec.Decode(&m); err == nil { + t.Error("decode unexpectedly succeeded") + return + } + }(i) + } + close(c) + wg.Wait() +} -- cgit v1.2.3