feat: implement gc command

cmd/crane/cmd/pull.go

@@ -30,8 +30,8 @@ import (
 // NewCmdPull creates a new cobra.Command for the pull subcommand.
 func NewCmdPull(options *[]crane.Option) *cobra.Command {
 	var (
-		cachePath, format string
-		annotateRef       bool
+		cachePath, format  string
+		annotateRef, prune bool
 	)
 
 	cmd := &cobra.Command{
@@ -124,6 +124,13 @@ func NewCmdPull(options *[]crane.Option) *cobra.Command {
 						return err
 					}
 				}
+
+				if prune {
+					if err := p.GarbageCollect(); err != nil {
+						return err
+					}
+				}
+
 			default:
 				return fmt.Errorf("unexpected --format: %q (valid values are: tarball, legacy, and oci)", format)
 			}
@@ -133,6 +140,7 @@ func NewCmdPull(options *[]crane.Option) *cobra.Command {
 	cmd.Flags().StringVarP(&cachePath, "cache_path", "c", "", "Path to cache image layers")
 	cmd.Flags().StringVar(&format, "format", "tarball", fmt.Sprintf("Format in which to save images (%q, %q, or %q)", "tarball", "legacy", "oci"))
 	cmd.Flags().BoolVar(&annotateRef, "annotate-ref", false, "Preserves image reference used to pull as an annotation when used with --format=oci")
-
+	cmd.Flags().BoolVar(&prune, "prune", false, "Removes orphan blobs from the oci-layout after pull")
+	cmd.Flags().MarkHidden("prune")
 	return cmd
 }

pkg/v1/layout/write.go

@@ -20,6 +20,7 @@
 	"errors"
 	"fmt"
 	"io"
+	"io/fs"
 	"os"
 	"path/filepath"
 
@@ -37,6 +38,100 @@
     "imageLayoutVersion": "1.0.0"
 }`
 
+// GarbageCollect removes unreferenced blobs from the oci-layout
+func (l Path) GarbageCollect() error {
+	idx, err := l.ImageIndex()
+	if err != nil {
+		return err
+	}
+	blobsToKeep := map[string]bool{}
+	if err := l.garbageCollectImageIndex(idx, blobsToKeep); err != nil {
+		return err
+	}
+	blobsDir := l.path("blobs")
+
+	if err := filepath.WalkDir(blobsDir, func(path string, d fs.DirEntry, err error) error {
+		if d.IsDir() {
+			return nil
+		}
+
+		rel, err := filepath.Rel(blobsDir, path)
+		if err != nil {
+			return err
+		}
+		if ok := blobsToKeep[rel]; !ok {
+			if err := os.Remove(path); err != nil {
+				return err
+			}
+		}
+		return nil
+	}); err != nil {
+		return err
+	}
+
+	return nil
+}
+
+func (l Path) garbageCollectImageIndex(index v1.ImageIndex, blobsToKeep map[string]bool) error {
+	idxm, err := index.IndexManifest()
+	if err != nil {
+		return err
+	}
+	if h, err := index.Digest(); err != nil {
+		return err
+	} else {
+		blobsToKeep[fmt.Sprintf("%s/%s", h.Algorithm, h.Hex)] = true
+	}
+	for _, descriptor := range idxm.Manifests {
+		if descriptor.MediaType.IsImage() {
+			img, err := index.Image(descriptor.Digest)
+			if err != nil {
+				return err
+			}
+			if err := l.garbageCollectImage(img, blobsToKeep); err != nil {
+				return err
+			}
+		} else if descriptor.MediaType.IsIndex() {
+			idx, err := index.ImageIndex(descriptor.Digest)
+			if err != nil {
+				return err
+			}
+			if err := l.garbageCollectImageIndex(idx, blobsToKeep); err != nil {
+				return err
+			}
+		}
+	}
+	return nil
+}
+
+func (l Path) garbageCollectImage(image v1.Image, blobsToKeep map[string]bool) error {
+
+	if h, err := image.Digest(); err != nil {
+		return err
+	} else {
+		blobsToKeep[fmt.Sprintf("%s/%s", h.Algorithm, h.Hex)] = true
+	}
+
+	if h, err := image.ConfigName(); err != nil {
+		return err
+	} else {
+		blobsToKeep[fmt.Sprintf("%s/%s", h.Algorithm, h.Hex)] = true
+	}
+
+	ls, err := image.Layers()
+	if err != nil {
+		return err
+	}
+	for _, l := range ls {
+		if h, err := l.Digest(); err != nil {
+			return err
+		} else {
+			blobsToKeep[fmt.Sprintf("%s/%s", h.Algorithm, h.Hex)] = true
+		}
+	}
+	return nil
+}
+
 // AppendImage writes a v1.Image to the Path and updates
 // the index.json to reference it.
 func (l Path) AppendImage(img v1.Image, options ...Option) error {