fix: properly read mimetype

Signed-off-by: Keith Zantow <kzantow@gmail.com>

syft/source/filesource/file_source.go

@@ -70,8 +70,14 @@ func New(cfg Config) (source.Source, error) {
 		return nil, fmt.Errorf("given path is a directory: %q", cfg.Path)
 	}
 
+	mimeType := stereoFile.MIMEType(f)
+
 	var digests []file.Digest
 	if len(cfg.DigestAlgorithms) > 0 {
+		// reading the MIME type moved past the start of the file
+		if _, err := f.Seek(0, io.SeekStart); err != nil {
+			return nil, fmt.Errorf("unable to seek file=%q: %w", cfg.Path, err)
+		}
 		digests, err = intFile.NewDigestsFromFile(context.TODO(), f, cfg.DigestAlgorithms)
 		if err != nil {
 			return nil, fmt.Errorf("unable to calculate digests for file=%q: %w", cfg.Path, err)
@@ -98,7 +104,7 @@ func New(cfg Config) (source.Source, error) {
 		analysisPath:     analysisPath,
 		digestForVersion: versionDigest,
 		digests:          digests,
-		mimeType:         stereoFile.MIMEType(f),
+		mimeType:         mimeType,
 	}, nil
 }
 

syft/source/filesource/file_source_test.go

@@ -4,6 +4,7 @@ import (
 	"archive/zip"
 	"bytes"
 	"context"
+	"crypto"
 	"io"
 	"os"
 	"os/exec"
@@ -243,6 +244,22 @@ func TestNewFromFile_RarExtension(t *testing.T) {
 	}
 }
 
+func TestNewFromFile_MIMETypeWithDigests(t *testing.T) {
+	// the digests read the whole file, which must not leave the MIME type to be read from its end
+	archivePath := filepath.Join(t.TempDir(), "example.zip")
+	require.NoError(t, os.WriteFile(archivePath, zipBytes(t, [][2]string{{"a.txt", "a"}}), 0o600))
+
+	src, err := New(Config{Path: archivePath, DigestAlgorithms: []crypto.Hash{crypto.SHA256}})
+	require.NoError(t, err)
+	t.Cleanup(func() {
+		require.NoError(t, src.Close())
+	})
+
+	m := src.Describe().Metadata.(source.FileMetadata)
+	assert.Equal(t, "application/zip", m.MIMEType)
+	assert.Len(t, m.Digests, 1)
+}
+
 func TestNewFromFile_ArchivePathsAreRootedAtTheArchive(t *testing.T) {
 	// an extracted archive is scanned like a directory: its root is the empty path, and the temporary directory it
 	// was extracted to does not appear