From 7080a8dd69912813ca4944feebfffab77ef92973 Mon Sep 17 00:00:00 2001
From: Priya Wadhwa <priyawadhwa@google.com>
Date: Mon, 27 Aug 2018 12:00:20 -0700
Subject: [PATCH] Add specific files from tar archives to list of snapshotted
 filesa I changed UnpackLocalTarArchive to return a list of files that were
 extracted, so that the list of snapshotted files for ADD is more accurate.
 Previously, we used to add all files in the extracted dir to be snapshotted,
 but this could result in preexisting files being snapshotted again.

---
 pkg/commands/add.go  | 10 +++-------
 pkg/util/fs_util.go  | 29 +++++++----------------------
 pkg/util/tar_util.go | 15 ++++++++-------
 3 files changed, 18 insertions(+), 36 deletions(-)

diff --git a/pkg/commands/add.go b/pkg/commands/add.go
index cf6ff7772..e0e317967 100644
--- a/pkg/commands/add.go
+++ b/pkg/commands/add.go
@@ -79,16 +79,12 @@ func (a *AddCommand) ExecuteCommand(config *v1.Config, buildArgs *dockerfile.Bui
 			a.snapshotFiles = append(a.snapshotFiles, urlDest)
 		} else if util.IsFileLocalTarArchive(fullPath) {
 			logrus.Infof("Unpacking local tar archive %s to %s", src, dest)
-			if err := util.UnpackLocalTarArchive(fullPath, dest); err != nil {
-				return err
-			}
-			// Add the unpacked files to the snapshotter
-			filesAdded, err := util.Files(dest)
+			extractedFiles, err := util.UnpackLocalTarArchive(fullPath, dest)
 			if err != nil {
 				return err
 			}
-			logrus.Debugf("Added %v from local tar archive %s", filesAdded, src)
-			a.snapshotFiles = append(a.snapshotFiles, filesAdded...)
+			logrus.Debugf("Added %v from local tar archive %s", extractedFiles, src)
+			a.snapshotFiles = append(a.snapshotFiles, extractedFiles...)
 		} else {
 			unresolvedSrcs = append(unresolvedSrcs, src)
 		}
diff --git a/pkg/util/fs_util.go b/pkg/util/fs_util.go
index 858312b24..67a972002 100644
--- a/pkg/util/fs_util.go
+++ b/pkg/util/fs_util.go
@@ -155,7 +155,9 @@ func ChildDirInWhitelist(path, directory string) bool {
 	return false
 }
 
-func unTar(r io.Reader, dest string) error {
+// unTar returns a list of files that have been extracted from the tar archive at r to the path at dest
+func unTar(r io.Reader, dest string) ([]string, error) {
+	var extractedFiles []string
 	tr := tar.NewReader(r)
 	for {
 		hdr, err := tr.Next()
@@ -163,13 +165,14 @@ func unTar(r io.Reader, dest string) error {
 			break
 		}
 		if err != nil {
-			return err
+			return nil, err
 		}
 		if err := extractFile(dest, hdr, tr); err != nil {
-			return err
+			return nil, err
 		}
+		extractedFiles = append(extractedFiles, dest)
 	}
-	return nil
+	return extractedFiles, nil
 }
 
 func extractFile(dest string, hdr *tar.Header, tr io.Reader) error {
@@ -349,24 +352,6 @@ func RelativeFiles(fp string, root string) ([]string, error) {
 	return files, err
 }
 
-// Files returns a list of all files rooted at root
-func Files(root string) ([]string, error) {
-	var files []string
-	logrus.Debugf("Getting files and contents at root %s", root)
-	err := filepath.Walk(root, func(path string, info os.FileInfo, err error) error {
-		whitelisted, err := CheckWhitelist(path)
-		if err != nil {
-			return err
-		}
-		if whitelisted {
-			return nil
-		}
-		files = append(files, path)
-		return err
-	})
-	return files, err
-}
-
 // ParentDirectories returns a list of paths to all parent directories
 // Ex. /some/temp/dir -> [/, /some, /some/temp, /some/temp/dir]
 func ParentDirectories(path string) []string {
diff --git a/pkg/util/tar_util.go b/pkg/util/tar_util.go
index f95574deb..898844784 100644
--- a/pkg/util/tar_util.go
+++ b/pkg/util/tar_util.go
@@ -137,17 +137,17 @@ func (t *Tar) checkHardlink(p string, i os.FileInfo) (bool, string) {
 }
 
 // UnpackLocalTarArchive unpacks the tar archive at path to the directory dest
-// Returns true if the path was actually unpacked
-func UnpackLocalTarArchive(path, dest string) error {
+// Returns the files extracted from the tar archive
+func UnpackLocalTarArchive(path, dest string) ([]string, error) {
 	// First, we need to check if the path is a local tar archive
 	if compressed, compressionLevel := fileIsCompressedTar(path); compressed {
 		file, err := os.Open(path)
 		if err != nil {
-			return err
+			return nil, err
 		}
 		defer file.Close()
 		if compressionLevel == archive.Gzip {
-			return UnpackCompressedTar(path, dest)
+			return nil, UnpackCompressedTar(path, dest)
 		} else if compressionLevel == archive.Bzip2 {
 			bzr := bzip2.NewReader(file)
 			return unTar(bzr, dest)
@@ -156,12 +156,12 @@ func UnpackLocalTarArchive(path, dest string) error {
 	if fileIsUncompressedTar(path) {
 		file, err := os.Open(path)
 		if err != nil {
-			return err
+			return nil, err
 		}
 		defer file.Close()
 		return unTar(file, dest)
 	}
-	return errors.New("path does not lead to local tar archive")
+	return nil, errors.New("path does not lead to local tar archive")
 }
 
 //IsFileLocalTarArchive returns true if the file is a local tar archive
@@ -223,5 +223,6 @@ func UnpackCompressedTar(path, dir string) error {
 		return err
 	}
 	defer gzr.Close()
-	return unTar(gzr, dir)
+	_, err = unTar(gzr, dir)
+	return err
 }