container: implement header based serialization, also add SPEC test vectors

pkg/container/SPEC.md

@@ -53,13 +53,13 @@ The following base encoding combination are REQUIRED to be supported:
 The CBOR bytes MUST be prepended by a single byte header to indicate the selection combination of base encoding and compression. This header value MUST be set according to the following table:
 
 | Header as hex | Header as ASCII | Base encoding           | Compression    |
-|---------------|-----------------|--------------------------|----------------|
+|---------------|-----------------|-------------------------|----------------|
 | 0x40          | @               | raw bytes               | no compression | 
 | 0x42          | B               | base64 std padding      | no compression | 
-| 0x43          | C               | base64 url               | no compression | 
+| 0x43          | C               | base64 url (no padding) | no compression | 
 | 0x4D          | M               | raw bytes               | gzip           | 
 | 0x4F          | O               | base64 std padding      | gzip           | 
-| 0x50          | P               | base64 url               | gzip           | 
+| 0x50          | P               | base64 url (no padding) | gzip           | 
 
 # 3 FAQ
 

pkg/container/packaging.go

@@ -0,0 +1,118 @@
+package container
+
+import (
+	"compress/gzip"
+	"encoding/base64"
+	"errors"
+	"fmt"
+	"io"
+)
+
+const containerVersionTag = "ctn-v1"
+
+type header byte
+
+const (
+	headerRawBytes             = header(0x40)
+	headerBase64StdPadding     = header(0x42)
+	headerBase64URL            = header(0x43)
+	headerRawBytesGzip         = header(0x4D)
+	headerBase64StdPaddingGzip = header(0x4F)
+	headerBase64URLGzip        = header(0x50)
+)
+
+func (h header) encoder(w io.Writer) *payloadWriter {
+	res := &payloadWriter{rawWriter: w, writer: w, header: h}
+
+	switch h {
+	case headerBase64StdPadding, headerBase64StdPaddingGzip:
+		b64Writer := base64.NewEncoder(base64.StdEncoding, res.writer)
+		res.writer = b64Writer
+		res.closers = append([]io.Closer{b64Writer}, res.closers...)
+	case headerBase64URL, headerBase64URLGzip:
+		b64Writer := base64.NewEncoder(base64.RawURLEncoding, res.writer)
+		res.writer = b64Writer
+		res.closers = append([]io.Closer{b64Writer}, res.closers...)
+	}
+
+	switch h {
+	case headerRawBytesGzip, headerBase64StdPaddingGzip, headerBase64URLGzip:
+		gzipWriter := gzip.NewWriter(res.writer)
+		res.writer = gzipWriter
+		res.closers = append([]io.Closer{gzipWriter}, res.closers...)
+	}
+
+	return res
+}
+
+func decodePayload(r io.Reader) (io.Reader, error) {
+	headerBuf := make([]byte, 1)
+	_, err := r.Read(headerBuf)
+	if err != nil {
+		return nil, err
+	}
+	h := header(headerBuf[0])
+
+	switch h {
+	case headerRawBytes,
+		headerBase64StdPadding,
+		headerBase64URL,
+		headerRawBytesGzip,
+		headerBase64StdPaddingGzip,
+		headerBase64URLGzip:
+	default:
+		return nil, fmt.Errorf("unknown container header")
+	}
+
+	switch h {
+	case headerBase64StdPadding, headerBase64StdPaddingGzip:
+		r = base64.NewDecoder(base64.StdEncoding, r)
+	case headerBase64URL, headerBase64URLGzip:
+		r = base64.NewDecoder(base64.RawURLEncoding, r)
+	}
+
+	switch h {
+	case headerRawBytesGzip, headerBase64StdPaddingGzip, headerBase64URLGzip:
+		gzipReader, err := gzip.NewReader(r)
+		if err != nil {
+			return nil, err
+		}
+		r = gzipReader
+	}
+
+	return r, nil
+}
+
+var _ io.WriteCloser = &payloadWriter{}
+
+// payloadWriter is tasked with two things:
+// - prepend the header byte
+// - call Close() on all the underlying io.Writer
+type payloadWriter struct {
+	rawWriter   io.Writer
+	writer      io.Writer
+	header      header
+	headerWrote bool
+	closers     []io.Closer
+}
+
+func (w *payloadWriter) Write(p []byte) (n int, err error) {
+	if !w.headerWrote {
+		_, err := w.rawWriter.Write([]byte{byte(w.header)})
+		if err != nil {
+			return 0, err
+		}
+		w.headerWrote = true
+	}
+	return w.writer.Write(p)
+}
+
+func (w *payloadWriter) Close() error {
+	var errs error
+	for _, closer := range w.closers {
+		if err := closer.Close(); err != nil {
+			errs = errors.Join(errs, err)
+		}
+	}
+	return errs
+}

pkg/container/reader.go

@@ -2,7 +2,6 @@ package container
 
 import (
 	"bytes"
-	"encoding/base64"
 	"errors"
 	"fmt"
 	"io"
@@ -25,6 +24,72 @@ var ErrMultipleInvocations = fmt.Errorf("multiple invocations")
 // Reader is a token container reader. It exposes the tokens conveniently decoded.
 type Reader map[cid.Cid]token.Token
 
+// FromBytes decodes a container from a []byte
+func FromBytes(data []byte) (Reader, error) {
+	return FromReader(bytes.NewReader(data))
+}
+
+// FromString decodes a container from a string
+func FromString(s string) (Reader, error) {
+	return FromReader(strings.NewReader(s))
+}
+
+// FromReader decodes a container from an io.Reader.
+func FromReader(r io.Reader) (Reader, error) {
+	payload, err := decodePayload(r)
+	if err != nil {
+		return nil, err
+	}
+
+	n, err := ipld.DecodeStreaming(payload, cbor.Decode)
+	if err != nil {
+		return nil, err
+	}
+	if n.Kind() != datamodel.Kind_Map {
+		return nil, fmt.Errorf("invalid container format: expected map")
+	}
+	if n.Length() != 1 {
+		return nil, fmt.Errorf("invalid container format: expected single version key")
+	}
+
+	// get the first (and only) key-value pair
+	it := n.MapIterator()
+	key, tokensNode, err := it.Next()
+	if err != nil {
+		return nil, err
+	}
+
+	version, err := key.AsString()
+	if err != nil {
+		return nil, fmt.Errorf("invalid container format: version must be string")
+	}
+	if version != containerVersionTag {
+		return nil, fmt.Errorf("unsupported container version: %s", version)
+	}
+
+	if tokensNode.Kind() != datamodel.Kind_List {
+		return nil, fmt.Errorf("invalid container format: tokens must be a list")
+	}
+
+	ctn := make(Reader, tokensNode.Length())
+	it2 := tokensNode.ListIterator()
+	for !it2.Done() {
+		_, val, err := it2.Next()
+		if err != nil {
+			return nil, err
+		}
+		data, err := val.AsBytes()
+		if err != nil {
+			return nil, err
+		}
+		err = ctn.addToken(data)
+		if err != nil {
+			return nil, err
+		}
+	}
+	return ctn, nil
+}
+
 // GetToken returns an arbitrary decoded token, from its CID.
 // If not found, ErrNotFound is returned.
 func (ctn Reader) GetToken(cid cid.Cid) (token.Token, error) {
@@ -65,7 +130,7 @@ func (ctn Reader) GetAllDelegations() iter.Seq2[cid.Cid, *delegation.Token] {
 
 // GetInvocation returns a single invocation.Token.
 // If none are found, ErrNotFound is returned.
-// If more than one invocation exist, ErrMultipleInvocations is returned.
+// If more than one invocation exists, ErrMultipleInvocations is returned.
 func (ctn Reader) GetInvocation() (*invocation.Token, error) {
 	var res *invocation.Token
 	for _, t := range ctn {
@@ -95,110 +160,6 @@ func (ctn Reader) GetAllInvocations() iter.Seq2[cid.Cid, *invocation.Token] {
 	}
 }
 
-// FromCbor decodes a DAG-CBOR encoded container.
-func FromCbor(data []byte) (Reader, error) {
-	return FromCborReader(bytes.NewReader(data))
-}
-
-// FromCborReader is the same as FromCbor, but with an io.Reader.
-func FromCborReader(r io.Reader) (Reader, error) {
-	n, err := ipld.DecodeStreaming(r, cbor.Decode)
-	if err != nil {
-		return nil, err
-	}
-	if n.Kind() != datamodel.Kind_Map {
-		return nil, fmt.Errorf("invalid container format: expected map")
-	}
-	if n.Length() != 1 {
-		return nil, fmt.Errorf("invalid container format: expected single version key")
-	}
-
-	// get the first (and only) key-value pair
-	it := n.MapIterator()
-	key, tokensNode, err := it.Next()
-	if err != nil {
-		return nil, err
-	}
-
-	version, err := key.AsString()
-	if err != nil {
-		return nil, fmt.Errorf("invalid container format: version must be string")
-	}
-	if version != currentContainerVersion {
-		return nil, fmt.Errorf("unsupported container version: %s", version)
-	}
-
-	if tokensNode.Kind() != datamodel.Kind_List {
-		return nil, fmt.Errorf("invalid container format: tokens must be a list")
-	}
-
-	ctn := make(Reader, tokensNode.Length())
-	it2 := tokensNode.ListIterator()
-	for !it2.Done() {
-		_, val, err := it2.Next()
-		if err != nil {
-			return nil, err
-		}
-		data, err := val.AsBytes()
-		if err != nil {
-			return nil, err
-		}
-		err = ctn.addToken(data)
-		if err != nil {
-			return nil, err
-		}
-	}
-	return ctn, nil
-}
-
-// FromCborBase64 decodes a base64 DAG-CBOR encoded container.
-func FromCborBase64(data string) (Reader, error) {
-	return FromCborBase64Reader(strings.NewReader(data))
-}
-
-// FromCborBase64Reader is the same as FromCborBase64, but with an io.Reader.
-func FromCborBase64Reader(r io.Reader) (Reader, error) {
-	return FromCborReader(base64.NewDecoder(base64.StdEncoding, r))
-}
-
-// FromCar decodes a CAR file encoded container.
-func FromCar(data []byte) (Reader, error) {
-	return FromCarReader(bytes.NewReader(data))
-}
-
-// FromCarReader is the same as FromCar, but with an io.Reader.
-func FromCarReader(r io.Reader) (Reader, error) {
-	_, it, err := readCar(r)
-	if err != nil {
-		return nil, err
-	}
-
-	ctn := make(Reader)
-
-	for block, err := range it {
-		if err != nil {
-			return nil, err
-		}
-
-		err = ctn.addToken(block.data)
-		if err != nil {
-			return nil, err
-		}
-	}
-
-	return ctn, nil
-}
-
-// FromCarBase64 decodes a base64 CAR file encoded container.
-func FromCarBase64(data string) (Reader, error) {
-	return FromCarReader(strings.NewReader(data))
-}
-
-// FromCarBase64Reader is the same as FromCarBase64, but with an io.Reader.
-func FromCarBase64Reader(r io.Reader) (Reader, error) {
-	return FromCarReader(base64.NewDecoder(base64.StdEncoding, r))
-}
-
 func (ctn Reader) addToken(data []byte) error {
 	tkn, c, err := token.FromSealed(data)
 	if err != nil {

pkg/container/serial_test.go

@@ -23,13 +23,21 @@ import (
 func TestContainerRoundTrip(t *testing.T) {
 	for _, tc := range []struct {
 		name           string
-		writer func(ctn Writer, w io.Writer) error
+		expectedHeader header
-		reader func(io.Reader) (Reader, error)
+		writer         any
 	}{
-		{"car", Writer.ToCarWriter, FromCarReader},
+		{"Bytes", headerRawBytes, Writer.ToBytes},
-		{"carBase64", Writer.ToCarBase64Writer, FromCarBase64Reader},
+		{"BytesWriter", headerRawBytes, Writer.ToBytesWriter},
-		{"cbor", Writer.ToCborWriter, FromCborReader},
+		{"BytesGzipped", headerRawBytesGzip, Writer.ToBytesGzipped},
-		{"cborBase64", Writer.ToCborBase64Writer, FromCborBase64Reader},
+		{"BytesGzippedWriter", headerRawBytesGzip, Writer.ToBytesGzippedWriter},
+		{"Base64StdPadding", headerBase64StdPadding, Writer.ToBase64StdPadding},
+		{"Base64StdPaddingWriter", headerBase64StdPadding, Writer.ToBase64StdPaddingWriter},
+		{"Base64StdPaddingGzipped", headerBase64StdPaddingGzip, Writer.ToBase64StdPaddingGzipped},
+		{"Base64StdPaddingGzippedWriter", headerBase64StdPaddingGzip, Writer.ToBase64StdPaddingGzippedWriter},
+		{"Base64URL", headerBase64URL, Writer.ToBase64URL},
+		{"Base64URLWriter", headerBase64URL, Writer.ToBase64URLWriter},
+		{"Base64URLGzip", headerBase64URLGzip, Writer.ToBase64URLGzip},
+		{"Base64URLGzipWriter", headerBase64URLGzip, Writer.ToBase64URLGzipWriter},
 	} {
 		t.Run(tc.name, func(t *testing.T) {
 			tokens := make(map[cid.Cid]*delegation.Token)
@@ -44,17 +52,49 @@ func TestContainerRoundTrip(t *testing.T) {
 				dataSize += len(data)
 			}
 
+			var reader Reader
+			var serialLen int
+
+			switch fn := tc.writer.(type) {
+			case func(ctn Writer, w io.Writer) error:
 				buf := bytes.NewBuffer(nil)
+				err := fn(writer, buf)
+				require.NoError(t, err)
+				serialLen = buf.Len()
 
-			err := tc.writer(writer, buf)
+				h, err := buf.ReadByte()
+				require.NoError(t, err)
+				require.Equal(t, byte(tc.expectedHeader), h)
+				err = buf.UnreadByte()
 				require.NoError(t, err)
 
-			t.Logf("data size %d", dataSize)
+				reader, err = FromReader(bytes.NewReader(buf.Bytes()))
-			t.Logf("container overhead: %d%%, %d bytes", int(float32(buf.Len()-dataSize)/float32(dataSize)*100.0), buf.Len()-dataSize)
-
-			reader, err := tc.reader(bytes.NewReader(buf.Bytes()))
 				require.NoError(t, err)
 
+			case func(ctn Writer) ([]byte, error):
+				b, err := fn(writer)
+				require.NoError(t, err)
+				serialLen = len(b)
+
+				require.Equal(t, byte(tc.expectedHeader), b[0])
+
+				reader, err = FromBytes(b)
+				require.NoError(t, err)
+
+			case func(ctn Writer) (string, error):
+				s, err := fn(writer)
+				require.NoError(t, err)
+				serialLen = len(s)
+
+				require.Equal(t, byte(tc.expectedHeader), s[0])
+
+				reader, err = FromString(s)
+				require.NoError(t, err)
+			}
+
+			t.Logf("data size %d, container size %d, overhead: %d%%, %d bytes",
+				dataSize, serialLen, int(float32(serialLen-dataSize)/float32(dataSize)*100.0), serialLen-dataSize)
+
 			for c, dlg := range tokens {
 				tknRead, err := reader.GetToken(c)
 				require.NoError(t, err)
@@ -98,10 +138,12 @@ func BenchmarkContainerSerialisation(b *testing.B) {
 		writer func(ctn Writer, w io.Writer) error
 		reader func(io.Reader) (Reader, error)
 	}{
-		{"car", Writer.ToCarWriter, FromCarReader},
+		{"Bytes", Writer.ToBytesWriter, FromReader},
-		{"carBase64", Writer.ToCarBase64Writer, FromCarBase64Reader},
+		{"BytesGzipped", Writer.ToBytesGzippedWriter, FromReader},
-		{"cbor", Writer.ToCborWriter, FromCborReader},
+		{"Base64StdPadding", Writer.ToBase64StdPaddingWriter, FromReader},
-		{"cborBase64", Writer.ToCborBase64Writer, FromCborBase64Reader},
+		{"Base64StdPaddingGzipped", Writer.ToBase64StdPaddingGzippedWriter, FromReader},
+		{"Base64URL", Writer.ToBase64URLWriter, FromReader},
+		{"Base64URLGzip", Writer.ToBase64URLGzipWriter, FromReader},
 	} {
 		writer := NewWriter()
 
@@ -184,7 +226,7 @@ func FuzzContainerRead(f *testing.F) {
 			_, c, data := randToken()
 			writer.AddSealed(c, data)
 		}
-		data, err := writer.ToCbor()
+		data, err := writer.ToBytes()
 		require.NoError(f, err)
 
 		f.Add(data)
@@ -194,7 +236,7 @@ func FuzzContainerRead(f *testing.F) {
 		start := time.Now()
 
 		// search for panics
-		_, _ = FromCbor(data)
+		_, _ = FromBytes(data)
 
 		if time.Since(start) > 100*time.Millisecond {
 			panic("too long")

pkg/container/writer.go

@@ -2,7 +2,6 @@ package container
 
 import (
 	"bytes"
-	"encoding/base64"
 	"io"
 
 	"github.com/ipfs/go-cid"
@@ -25,22 +24,92 @@ func (ctn Writer) AddSealed(cid cid.Cid, data []byte) {
 	ctn[cid] = data
 }
 
-const currentContainerVersion = "ctn-v1"
+// ToBytes encode the container into raw bytes.
+func (ctn Writer) ToBytes() ([]byte, error) {
+	return ctn.toBytes(headerRawBytes)
+}
 
-// ToCbor encode the container into a CBOR binary format.
+// ToBytesWriter is the same as ToBytes, but with an io.Writer.
-func (ctn Writer) ToCbor() ([]byte, error) {
+func (ctn Writer) ToBytesWriter(w io.Writer) error {
+	return ctn.toWriter(headerRawBytes, w)
+}
+
+// ToBytesGzipped encode the container into gzipped bytes.
+func (ctn Writer) ToBytesGzipped() ([]byte, error) {
+	return ctn.toBytes(headerRawBytesGzip)
+}
+
+// ToBytesGzippedWriter is the same as ToBytesGzipped, but with an io.Writer.
+func (ctn Writer) ToBytesGzippedWriter(w io.Writer) error {
+	return ctn.toWriter(headerRawBytesGzip, w)
+}
+
+// ToBase64StdPadding encode the container into a base64 string, with standard encoding and padding.
+func (ctn Writer) ToBase64StdPadding() (string, error) {
+	return ctn.toString(headerBase64StdPadding)
+}
+
+// ToBase64StdPaddingWriter is the same as ToBase64StdPadding, but with an io.Writer.
+func (ctn Writer) ToBase64StdPaddingWriter(w io.Writer) error {
+	return ctn.toWriter(headerBase64StdPadding, w)
+}
+
+// ToBase64StdPaddingGzipped encode the container into a pre-gzipped base64 string, with standard encoding and padding.
+func (ctn Writer) ToBase64StdPaddingGzipped() (string, error) {
+	return ctn.toString(headerBase64StdPaddingGzip)
+}
+
+// ToBase64StdPaddingGzippedWriter is the same as ToBase64StdPaddingGzipped, but with an io.Writer.
+func (ctn Writer) ToBase64StdPaddingGzippedWriter(w io.Writer) error {
+	return ctn.toWriter(headerBase64StdPaddingGzip, w)
+}
+
+// ToBase64URL encode the container into base64 string, with URL-safe encoding and no padding.
+func (ctn Writer) ToBase64URL() (string, error) {
+	return ctn.toString(headerBase64URL)
+}
+
+// ToBase64URLWriter is the same as ToBase64URL, but with an io.Writer.
+func (ctn Writer) ToBase64URLWriter(w io.Writer) error {
+	return ctn.toWriter(headerBase64URL, w)
+}
+
+// ToBase64URL encode the container into pre-gzipped base64 string, with URL-safe encoding and no padding.
+func (ctn Writer) ToBase64URLGzip() (string, error) {
+	return ctn.toString(headerBase64URLGzip)
+}
+
+// ToBase64URLWriter is the same as ToBase64URL, but with an io.Writer.
+func (ctn Writer) ToBase64URLGzipWriter(w io.Writer) error {
+	return ctn.toWriter(headerBase64URLGzip, w)
+}
+
+func (ctn Writer) toBytes(header header) ([]byte, error) {
 	var buf bytes.Buffer
-	err := ctn.ToCborWriter(&buf)
+	err := ctn.toWriter(header, &buf)
 	if err != nil {
 		return nil, err
 	}
 	return buf.Bytes(), nil
 }
 
-// ToCborWriter is the same as ToCbor, but with an io.Writer.
+func (ctn Writer) toString(header header) (string, error) {
-func (ctn Writer) ToCborWriter(w io.Writer) error {
+	var buf bytes.Buffer
+	err := ctn.toWriter(header, &buf)
+	if err != nil {
+		return "", err
+	}
+	return buf.String(), nil
+}
+
+func (ctn Writer) toWriter(header header, w io.Writer) (err error) {
+	encoder := header.encoder(w)
+
+	defer func() {
+		err = encoder.Close()
+	}()
 	node, err := qp.BuildMap(basicnode.Prototype.Any, 1, func(ma datamodel.MapAssembler) {
-		qp.MapEntry(ma, currentContainerVersion, qp.List(int64(len(ctn)), func(la datamodel.ListAssembler) {
+		qp.MapEntry(ma, containerVersionTag, qp.List(int64(len(ctn)), func(la datamodel.ListAssembler) {
 			for _, data := range ctn {
 				qp.ListEntry(la, qp.Bytes(data))
 			}
@@ -49,60 +118,6 @@ func (ctn Writer) ToCborWriter(w io.Writer) error {
 	if err != nil {
 		return err
 	}
-	return ipld.EncodeStreaming(w, node, cbor.Encode)
-}
 
-// ToCborBase64 encode the container into a base64 encoded CBOR binary format.
+	return ipld.EncodeStreaming(encoder, node, cbor.Encode)
-func (ctn Writer) ToCborBase64() (string, error) {
-	var buf bytes.Buffer
-	err := ctn.ToCborBase64Writer(&buf)
-	if err != nil {
-		return "", err
-	}
-	return buf.String(), nil
-}
-
-// ToCborBase64Writer is the same as ToCborBase64, but with an io.Writer.
-func (ctn Writer) ToCborBase64Writer(w io.Writer) error {
-	w2 := base64.NewEncoder(base64.StdEncoding, w)
-	defer w2.Close()
-	return ctn.ToCborWriter(w2)
-}
-
-// ToCar encode the container into a CAR file.
-func (ctn Writer) ToCar() ([]byte, error) {
-	var buf bytes.Buffer
-	err := ctn.ToCarWriter(&buf)
-	if err != nil {
-		return nil, err
-	}
-	return buf.Bytes(), nil
-}
-
-// ToCarWriter is the same as ToCar, but with an io.Writer.
-func (ctn Writer) ToCarWriter(w io.Writer) error {
-	return writeCar(w, nil, func(yield func(carBlock, error) bool) {
-		for c, data := range ctn {
-			if !yield(carBlock{c: c, data: data}, nil) {
-				return
-			}
-		}
-	})
-}
-
-// ToCarBase64 encode the container into a base64 encoded CAR file.
-func (ctn Writer) ToCarBase64() (string, error) {
-	var buf bytes.Buffer
-	err := ctn.ToCarBase64Writer(&buf)
-	if err != nil {
-		return "", err
-	}
-	return buf.String(), nil
-}
-
-// ToCarBase64Writer is the same as ToCarBase64, but with an io.Writer.
-func (ctn Writer) ToCarBase64Writer(w io.Writer) error {
-	w2 := base64.NewEncoder(base64.StdEncoding, w)
-	defer w2.Close()
-	return ctn.ToCarWriter(w2)
 }