swap format readseekers for readers (#2515)

Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>

syft/format/cyclonedxjson/decoder.go

@@ -10,6 +10,7 @@ import (
 	"github.com/anchore/syft/internal/log"
 	"github.com/anchore/syft/syft/format/common/cyclonedxhelpers"
 	"github.com/anchore/syft/syft/format/internal/cyclonedxutil"
+	"github.com/anchore/syft/syft/format/internal/stream"
 	"github.com/anchore/syft/syft/sbom"
 )
 
@@ -25,10 +26,12 @@ func NewFormatDecoder() sbom.FormatDecoder {
 	}
 }
 
-func (d decoder) Decode(reader io.ReadSeeker) (*sbom.SBOM, sbom.FormatID, string, error) {
+func (d decoder) Decode(r io.Reader) (*sbom.SBOM, sbom.FormatID, string, error) {
-	if reader == nil {
+	reader, err := stream.SeekableReader(r)
-		return nil, "", "", fmt.Errorf("no SBOM bytes provided")
+	if err != nil {
+		return nil, "", "", err
 	}
+
 	id, version := d.Identify(reader)
 	if id != ID {
 		return nil, "", "", fmt.Errorf("not a cyclonedx json document")
@@ -50,10 +53,12 @@ func (d decoder) Decode(reader io.ReadSeeker) (*sbom.SBOM, sbom.FormatID, string
 	return s, id, version, nil
 }
 
-func (d decoder) Identify(reader io.ReadSeeker) (sbom.FormatID, string) {
+func (d decoder) Identify(r io.Reader) (sbom.FormatID, string) {
-	if reader == nil {
+	reader, err := stream.SeekableReader(r)
+	if err != nil {
 		return "", ""
 	}
+
 	if _, err := reader.Seek(0, io.SeekStart); err != nil {
 		log.Debugf("unable to seek to start of CycloneDX JSON SBOM: %+v", err)
 		return "", ""
@@ -68,8 +73,7 @@ func (d decoder) Identify(reader io.ReadSeeker) (sbom.FormatID, string) {
 	dec := json.NewDecoder(reader)
 
 	var doc Document
-	err := dec.Decode(&doc)
+	if err = dec.Decode(&doc); err != nil {
-	if err != nil {
 		// maybe not json? maybe not valid? doesn't matter, we won't process it.
 		return "", ""
 	}

syft/format/cyclonedxxml/decoder.go

@@ -11,6 +11,7 @@ import (
 	"github.com/anchore/syft/internal/log"
 	"github.com/anchore/syft/syft/format/common/cyclonedxhelpers"
 	"github.com/anchore/syft/syft/format/internal/cyclonedxutil"
+	"github.com/anchore/syft/syft/format/internal/stream"
 	"github.com/anchore/syft/syft/sbom"
 )
 
@@ -26,9 +27,10 @@ func NewFormatDecoder() sbom.FormatDecoder {
 	}
 }
 
-func (d decoder) Decode(reader io.ReadSeeker) (*sbom.SBOM, sbom.FormatID, string, error) {
+func (d decoder) Decode(r io.Reader) (*sbom.SBOM, sbom.FormatID, string, error) {
-	if reader == nil {
+	reader, err := stream.SeekableReader(r)
-		return nil, "", "", fmt.Errorf("no SBOM bytes provided")
+	if err != nil {
+		return nil, "", "", err
 	}
 
 	id, version := d.Identify(reader)
@@ -52,8 +54,9 @@ func (d decoder) Decode(reader io.ReadSeeker) (*sbom.SBOM, sbom.FormatID, string
 	return s, id, version, nil
 }
 
-func (d decoder) Identify(reader io.ReadSeeker) (sbom.FormatID, string) {
+func (d decoder) Identify(r io.Reader) (sbom.FormatID, string) {
-	if reader == nil {
+	reader, err := stream.SeekableReader(r)
+	if err != nil {
 		return "", ""
 	}
 
@@ -69,8 +72,7 @@ func (d decoder) Identify(reader io.ReadSeeker) (sbom.FormatID, string) {
 	dec := xml.NewDecoder(reader)
 
 	var doc Document
-	err := dec.Decode(&doc)
+	if err = dec.Decode(&doc); err != nil {
-	if err != nil {
 		// maybe not xml? maybe not valid? doesn't matter, we won't process it.
 		return "", ""
 	}

syft/format/decoders.go

@@ -42,7 +42,7 @@ func NewDecoderCollection(decoders ...sbom.FormatDecoder) sbom.FormatDecoder {
 }
 
 // Decode takes a set of bytes and attempts to decode it into an SBOM relative to the decoders in the collection.
-func (c *DecoderCollection) Decode(reader io.ReadSeeker) (*sbom.SBOM, sbom.FormatID, string, error) {
+func (c *DecoderCollection) Decode(reader io.Reader) (*sbom.SBOM, sbom.FormatID, string, error) {
 	if reader == nil {
 		return nil, "", "", fmt.Errorf("no SBOM bytes provided")
 	}
@@ -67,7 +67,7 @@ func (c *DecoderCollection) Decode(reader io.ReadSeeker) (*sbom.SBOM, sbom.Forma
 }
 
 // Identify takes a set of bytes and attempts to identify the format of the SBOM relative to the decoders in the collection.
-func (c *DecoderCollection) Identify(reader io.ReadSeeker) (sbom.FormatID, string) {
+func (c *DecoderCollection) Identify(reader io.Reader) (sbom.FormatID, string) {
 	if reader == nil {
 		return "", ""
 	}
@@ -81,11 +81,11 @@ func (c *DecoderCollection) Identify(reader io.ReadSeeker) (sbom.FormatID, strin
 }
 
 // Identify takes a set of bytes and attempts to identify the format of the SBOM.
-func Identify(reader io.ReadSeeker) (sbom.FormatID, string) {
+func Identify(reader io.Reader) (sbom.FormatID, string) {
 	return staticDecoders.Identify(reader)
 }
 
 // Decode takes a set of bytes and attempts to decode it into an SBOM.
-func Decode(reader io.ReadSeeker) (*sbom.SBOM, sbom.FormatID, string, error) {
+func Decode(reader io.Reader) (*sbom.SBOM, sbom.FormatID, string, error) {
 	return staticDecoders.Decode(reader)
 }

syft/format/internal/cyclonedxutil/decoder.go

@@ -5,6 +5,8 @@ import (
 	"io"
 
 	"github.com/CycloneDX/cyclonedx-go"
+
+	"github.com/anchore/syft/syft/format/internal/stream"
 )
 
 type Decoder struct {
@@ -17,14 +19,20 @@ func NewDecoder(format cyclonedx.BOMFileFormat) Decoder {
 	}
 }
 
-func (d Decoder) Decode(reader io.ReadSeeker) (*cyclonedx.BOM, error) {
+func (d Decoder) Decode(r io.Reader) (*cyclonedx.BOM, error) {
+	reader, err := stream.SeekableReader(r)
+	if err != nil {
+		return nil, err
+	}
+
 	doc := &cyclonedx.BOM{
 		Components: &[]cyclonedx.Component{},
 	}
 	if _, err := reader.Seek(0, io.SeekStart); err != nil {
 		return nil, fmt.Errorf("unable to seek to start of CycloneDX SBOM: %w", err)
 	}
-	err := cyclonedx.NewBOMDecoder(reader, d.format).Decode(doc)
+
+	err = cyclonedx.NewBOMDecoder(reader, d.format).Decode(doc)
 	if err != nil {
 		return nil, err
 	}

syft/format/internal/stream/seekable_reader.go

@@ -0,0 +1,24 @@
+package stream
+
+import (
+	"bytes"
+	"fmt"
+	"io"
+)
+
+func SeekableReader(reader io.Reader) (io.ReadSeeker, error) {
+	if reader == nil {
+		return nil, fmt.Errorf("no bytes provided")
+	}
+
+	if r, ok := reader.(io.ReadSeeker); ok {
+		return r, nil
+	}
+
+	content, err := io.ReadAll(reader)
+	if err != nil {
+		return nil, err
+	}
+
+	return bytes.NewReader(content), nil
+}

syft/format/internal/stream/seekable_reader_test.go

@@ -0,0 +1,101 @@
+package stream
+
+import (
+	"bytes"
+	"io"
+	"reflect"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+func TestSeekableReader(t *testing.T) {
+	tests := []struct {
+		name    string
+		input   io.Reader
+		assert  func(io.Reader, io.ReadSeeker)
+		wantErr require.ErrorAssertionFunc
+	}{
+		{
+			name:    "nil reader",
+			input:   nil,
+			wantErr: require.Error,
+		},
+		{
+			name:  "empty reader",
+			input: bytes.NewBuffer([]byte{}), // does not implement io.Seeker (but does implement io.Reader)
+			assert: func(input io.Reader, got io.ReadSeeker) {
+				impl, ok := got.(*bytes.Reader) // implements bytes.Reader
+				require.True(t, ok)
+				_, err := impl.Seek(0, io.SeekStart)
+				require.NoError(t, err)
+				content, err := io.ReadAll(impl)
+				require.NoError(t, err)
+				require.Equal(t, []byte{}, content)
+			},
+		},
+		{
+			name:  "empty read seeker",
+			input: bytes.NewReader([]byte{}), // implements io.ReadSeeker
+			assert: func(input io.Reader, got io.ReadSeeker) {
+				impl, ok := got.(*bytes.Reader)
+				require.True(t, ok)
+				_, err := impl.Seek(0, io.SeekStart)
+				require.NoError(t, err)
+				content, err := io.ReadAll(impl)
+				require.NoError(t, err)
+				require.Equal(t, []byte{}, content)
+
+				// assert this is the same read seeker (reflect tt.input pointer is the same as the impl pointer
+				inputImpl, ok := input.(*bytes.Reader)
+				require.True(t, ok)
+				assert.Equal(t, reflect.ValueOf(inputImpl).Pointer(), reflect.ValueOf(impl).Pointer())
+			},
+		},
+		{
+			name:  "non-empty read seeker",
+			input: bytes.NewReader([]byte("hello world!")), // implements io.ReadSeeker
+			assert: func(input io.Reader, got io.ReadSeeker) {
+				impl, ok := got.(*bytes.Reader)
+				require.True(t, ok)
+				_, err := impl.Seek(0, io.SeekStart)
+				require.NoError(t, err)
+				content, err := io.ReadAll(impl)
+				require.NoError(t, err)
+				require.Equal(t, []byte("hello world!"), content)
+
+				// assert this is the same read seeker (reflect tt.input pointer is the same as the impl pointer
+				inputImpl, ok := input.(*bytes.Reader)
+				require.True(t, ok)
+				assert.Equal(t, reflect.ValueOf(inputImpl).Pointer(), reflect.ValueOf(impl).Pointer())
+			},
+		},
+		{
+			name:  "non-empty reader",
+			input: bytes.NewBufferString("hello world!"), // does not implement io.Seeker (but does implement io.Reader)
+			assert: func(input io.Reader, got io.ReadSeeker) {
+				impl, ok := got.(*bytes.Reader)
+				require.True(t, ok)
+				_, err := impl.Seek(0, io.SeekStart)
+				require.NoError(t, err)
+				content, err := io.ReadAll(impl)
+				require.NoError(t, err)
+				require.Equal(t, []byte("hello world!"), content)
+			},
+		},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			if tt.wantErr == nil {
+				tt.wantErr = require.NoError
+			}
+			got, err := SeekableReader(tt.input)
+			tt.wantErr(t, err)
+			if err != nil {
+				return
+			}
+			tt.assert(tt.input, got)
+		})
+	}
+}

syft/format/spdxjson/decoder.go

@@ -10,6 +10,7 @@ import (
 
 	"github.com/anchore/syft/internal/log"
 	"github.com/anchore/syft/syft/format/common/spdxhelpers"
+	"github.com/anchore/syft/syft/format/internal/stream"
 	"github.com/anchore/syft/syft/sbom"
 )
 
@@ -22,9 +23,10 @@ func NewFormatDecoder() sbom.FormatDecoder {
 	return decoder{}
 }
 
-func (d decoder) Decode(reader io.ReadSeeker) (*sbom.SBOM, sbom.FormatID, string, error) {
+func (d decoder) Decode(r io.Reader) (*sbom.SBOM, sbom.FormatID, string, error) {
-	if reader == nil {
+	reader, err := stream.SeekableReader(r)
-		return nil, "", "", fmt.Errorf("no SBOM bytes provided")
+	if err != nil {
+		return nil, "", "", err
 	}
 
 	// since spdx lib will always return the latest version of the document, we need to identify the version
@@ -54,8 +56,9 @@ func (d decoder) Decode(reader io.ReadSeeker) (*sbom.SBOM, sbom.FormatID, string
 	return s, id, version, nil
 }
 
-func (d decoder) Identify(reader io.ReadSeeker) (sbom.FormatID, string) {
+func (d decoder) Identify(r io.Reader) (sbom.FormatID, string) {
-	if reader == nil {
+	reader, err := stream.SeekableReader(r)
+	if err != nil {
 		return "", ""
 	}
 
@@ -75,8 +78,7 @@ func (d decoder) Identify(reader io.ReadSeeker) (sbom.FormatID, string) {
 	dec := json.NewDecoder(reader)
 
 	var doc Document
-	err := dec.Decode(&doc)
+	if err = dec.Decode(&doc); err != nil {
-	if err != nil {
 		// maybe not json? maybe not valid? doesn't matter, we won't process it.
 		return "", ""
 	}

syft/format/spdxtagvalue/decoder.go

@@ -10,6 +10,7 @@ import (
 
 	"github.com/anchore/syft/internal/log"
 	"github.com/anchore/syft/syft/format/common/spdxhelpers"
+	"github.com/anchore/syft/syft/format/internal/stream"
 	"github.com/anchore/syft/syft/sbom"
 )
 
@@ -22,9 +23,10 @@ func NewFormatDecoder() sbom.FormatDecoder {
 	return decoder{}
 }
 
-func (d decoder) Decode(reader io.ReadSeeker) (*sbom.SBOM, sbom.FormatID, string, error) {
+func (d decoder) Decode(r io.Reader) (*sbom.SBOM, sbom.FormatID, string, error) {
-	if reader == nil {
+	reader, err := stream.SeekableReader(r)
-		return nil, "", "", fmt.Errorf("no SBOM bytes provided")
+	if err != nil {
+		return nil, "", "", err
 	}
 
 	// since spdx lib will always return the latest version of the document, we need to identify the version
@@ -54,8 +56,9 @@ func (d decoder) Decode(reader io.ReadSeeker) (*sbom.SBOM, sbom.FormatID, string
 	return s, id, version, nil
 }
 
-func (d decoder) Identify(reader io.ReadSeeker) (sbom.FormatID, string) {
+func (d decoder) Identify(r io.Reader) (sbom.FormatID, string) {
-	if reader == nil {
+	reader, err := stream.SeekableReader(r)
+	if err != nil {
 		return "", ""
 	}
 

syft/format/syftjson/decoder.go

@@ -10,6 +10,7 @@ import (
 
 	"github.com/anchore/syft/internal"
 	"github.com/anchore/syft/internal/log"
+	"github.com/anchore/syft/syft/format/internal/stream"
 	"github.com/anchore/syft/syft/format/syftjson/model"
 	"github.com/anchore/syft/syft/sbom"
 )
@@ -22,9 +23,10 @@ func NewFormatDecoder() sbom.FormatDecoder {
 	return decoder{}
 }
 
-func (d decoder) Decode(reader io.ReadSeeker) (*sbom.SBOM, sbom.FormatID, string, error) {
+func (d decoder) Decode(r io.Reader) (*sbom.SBOM, sbom.FormatID, string, error) {
-	if reader == nil {
+	reader, err := stream.SeekableReader(r)
-		return nil, "", "", fmt.Errorf("no SBOM bytes provided")
+	if err != nil {
+		return nil, "", "", err
 	}
 
 	id, version := d.Identify(reader)
@@ -39,8 +41,7 @@ func (d decoder) Decode(reader io.ReadSeeker) (*sbom.SBOM, sbom.FormatID, string
 
 	dec := json.NewDecoder(reader)
 
-	err := dec.Decode(&doc)
+	if err = dec.Decode(&doc); err != nil {
-	if err != nil {
 		return nil, "", "", fmt.Errorf("unable to decode syft-json document: %w", err)
 	}
 
@@ -51,8 +52,9 @@ func (d decoder) Decode(reader io.ReadSeeker) (*sbom.SBOM, sbom.FormatID, string
 	return toSyftModel(doc), ID, doc.Schema.Version, nil
 }
 
-func (d decoder) Identify(reader io.ReadSeeker) (sbom.FormatID, string) {
+func (d decoder) Identify(r io.Reader) (sbom.FormatID, string) {
-	if reader == nil {
+	reader, err := stream.SeekableReader(r)
+	if err != nil {
 		return "", ""
 	}
 
@@ -68,8 +70,7 @@ func (d decoder) Identify(reader io.ReadSeeker) (sbom.FormatID, string) {
 	dec := json.NewDecoder(reader)
 
 	var doc Document
-	err := dec.Decode(&doc)
+	if err = dec.Decode(&doc); err != nil {
-	if err != nil {
 		// maybe not json? maybe not valid? doesn't matter, we won't process it.
 		return "", ""
 	}

syft/sbom/format.go

@@ -23,9 +23,9 @@ type FormatEncoder interface {
 type FormatDecoder interface {
 	// Decode will return an SBOM from the given reader. If the bytes are not a valid SBOM for the given format
 	// then an error will be returned.
-	Decode(io.ReadSeeker) (*SBOM, FormatID, string, error)
+	Decode(io.Reader) (*SBOM, FormatID, string, error)
 
 	// Identify will return the format ID and version for the given reader. Note: this does not validate the
 	// full SBOM, only pulls the minimal information necessary to identify the format.
-	Identify(io.ReadSeeker) (FormatID, string)
+	Identify(io.Reader) (FormatID, string)
 }