Add library function to strip out a single file from a deb package (#51)

go.work.sum

@@ -1,3 +1,5 @@
 github.com/cespare/xxhash v1.1.0 h1:a6HrQnmkObjyL+Gs60czilIUGqrzKutQD6XZog3p+ko=
+github.com/opencontainers/image-spec v1.1.0-rc2 h1:2zx/Stx4Wc5pIPDvIxHXvXtQFW/7XWJGmnM7r3wg034=
 github.com/opencontainers/image-spec v1.1.0-rc2/go.mod h1:3OVijpioIKYWTqjiG0zfF6wvoJ4fAXGbjdZuI2NgsRQ=
+github.com/russross/blackfriday v1.6.0 h1:KqfZb0pUVN2lYqZUYRddxF4OR8ZMURnJIG5Y3VRLtww=
 github.com/urfave/cli v1.22.2 h1:gsqYFH8bb9ekPA12kRo0hfjngWQjkJPlN9R0N78BoUo=

go/pkg/deb2layer/BUILD.bazel

@@ -13,4 +13,5 @@ go_test(
     srcs = ["deb_test.go"],
     embed = [":go_default_library"],
     deps = ["@com_github_stretchr_testify//assert:go_default_library"],
+    data = ["testdeb.deb"],
 )

go/pkg/deb2layer/deb.go

@@ -32,9 +32,12 @@ var (
 	// ErrDuplicateSection is returned when there is a duplicate deb package
 	// sections
 	ErrDuplicateSection = fmt.Errorf("duplicate section")
+
+	// ErrSectionNotFound is returned when an expected section is not found
+	ErrSectionNotFound = fmt.Errorf("section not found")
 )
 
-// DebToLayer convernts a deb package into a tarffile that can be used as a
+// DebToLayer convernts a deb package into a tar file that can be used as a
 // layer in an OCI image. Including adding the appropriate files for dpkg to
 // track the package.
 //
@@ -123,6 +126,53 @@ func DebToLayer(debReader io.Reader, writer io.Writer) error {
 	return nil
 }
 
+// CopyFileFromDeb copies a file from the debian file to a writer.
+func CopyFileFromDeb(filename string, debReader io.Reader, writer io.Writer) error {
+	arReader := ar.NewReader(debReader)
+	for {
+		header, err := arReader.Next()
+		if err == io.EOF {
+			break
+		}
+		if err != nil {
+			return err
+		}
+
+		// > System V ar uses a '/' character (0x2F) to mark the end of the filename;
+		// https://en.wikipedia.org/wiki/Ar_(Unix)
+		name := header.Name
+		if strings.HasSuffix(name, "/") {
+			name = name[:len(name)-1]
+		}
+
+		if strings.HasPrefix(name, debDataFilePrefix) {
+			reader, err := extToReader(getFullExt(name), arReader)
+			if err != nil {
+				return err
+			}
+
+			tarReader, ok := reader.(*tar.Reader)
+			if !ok {
+				return fmt.Errorf("%q section is not a tar file: %q (%T)", debDataFilePrefix, name, reader)
+			}
+
+			_, err = seekToTarFile(tarReader, filename)
+			if err != nil {
+				return err
+			}
+
+			_, err = io.Copy(writer, tarReader)
+			if err != nil {
+				return err
+			}
+
+			return nil
+		}
+	}
+
+	return ErrSectionNotFound
+}
+
 // seekToTarFile seeks forward to the tar entry whose header's name is
 // equivalent to the one provider. If found, the provided *tar.Reader will be
 // left at the start of the data, otherwise it will leave at io.EOF.

go/pkg/deb2layer/deb_test.go

@@ -11,16 +11,17 @@ import (
 	"github.com/stretchr/testify/assert"
 )
 
+// Rebuild deb with `dpkg-deb -Zgzip --build testdeb`
+
 var (
-	expectedFiles = []string{"/var/lib/dpkg/status.d/test", "./", "./deb_text.txt"}
+	expectedFiles = []string{"/var/lib/dpkg/status.d/testdeb", "./", "./deb_text.txt"}
 )
 
 // TestDebToLayer checks that the expected files are in the resulting layer
 // based on a deb file created by Bazel's rules_pkg
 func TestDebToLayer(t *testing.T) {
     // XXX Fix generating deb package
-    t.Skip()
-	f, err := os.Open("test-tar-deb.deb")
+	f, err := os.Open("testdeb.deb")
 	if err != nil {
 		t.Fatalf("%v", err)
 	}
@@ -52,3 +53,19 @@ func TestDebToLayer(t *testing.T) {
 
 	assert.Equal(t, expectedFiles, actualFiles)
 }
+
+func TestCopyFileFromDeb(t *testing.T) {
+	f, err := os.Open("testdeb.deb")
+	if err != nil {
+		t.Fatalf("%v", err)
+	}
+	defer f.Close()
+
+	var buf bytes.Buffer
+	err = CopyFileFromDeb("deb_text.txt", f, &buf)
+	if err != nil {
+		t.Fatalf("%v", err)
+	}
+
+	assert.Equal(t, "my-deb\n", buf.String())
+}

go/pkg/deb2layer/testdeb/DEBIAN/control

@@ -0,0 +1,5 @@
+Package: testdeb
+Version: 0.0.1
+Maintainer: DataDog
+Architecture: all
+Description: hello world