backup/scanner: Fix total size for overlapping targets

Before, the scanner would could files twice if they were included in the
list of backup targets twice, e.g. `restic backup foo foo/bar` would
could the file `foo/bar` twice.

This commit uses the tree structure from the archiver to run the
scanner, so both parts see the same files.

internal/archiver/scanner.go

@@ -6,6 +6,7 @@ import (
 	"path/filepath"
 	"sort"
 
+	"github.com/restic/restic/internal/debug"
 	"github.com/restic/restic/internal/fs"
 )
 
@@ -37,27 +38,63 @@ type ScanStats struct {
 	Bytes               uint64
 }
 
-// Scan traverses the targets. The function Result is called for each new item
-// found, the complete result is also returned by Scan.
-func (s *Scanner) Scan(ctx context.Context, targets []string) error {
-	var stats ScanStats
-	for _, target := range targets {
-		abstarget, err := s.FS.Abs(target)
+func (s *Scanner) scanTree(ctx context.Context, stats ScanStats, tree Tree) (ScanStats, error) {
+	// traverse the path in the file system for all leaf nodes
+	if tree.Leaf() {
+		abstarget, err := s.FS.Abs(tree.Path)
 		if err != nil {
-			return err
+			return ScanStats{}, err
 		}
 
 		stats, err = s.scan(ctx, stats, abstarget)
 		if err != nil {
-			return err
+			return ScanStats{}, err
+		}
+
+		return stats, nil
+	}
+
+	// otherwise recurse into the nodes in a deterministic order
+	for _, name := range tree.NodeNames() {
+		var err error
+		stats, err = s.scanTree(ctx, stats, tree.Nodes[name])
+		if err != nil {
+			return ScanStats{}, err
 		}
 
 		if ctx.Err() != nil {
-			return nil
+			return stats, nil
 		}
 	}
 
+	return stats, nil
+}
+
+// Scan traverses the targets. The function Result is called for each new item
+// found, the complete result is also returned by Scan.
+func (s *Scanner) Scan(ctx context.Context, targets []string) error {
+	debug.Log("start scan for %v", targets)
+
+	cleanTargets, err := resolveRelativeTargets(s.FS, targets)
+	if err != nil {
+		return err
+	}
+
+	debug.Log("clean targets %v", cleanTargets)
+
+	// we're using the same tree representation as the archiver does
+	tree, err := NewTree(s.FS, cleanTargets)
+	if err != nil {
+		return err
+	}
+
+	stats, err := s.scanTree(ctx, ScanStats{}, *tree)
+	if err != nil {
+		return err
+	}
+
 	s.Result("", stats)
+	debug.Log("result: %+v", stats)
 	return nil
 }
 

internal/archiver/scanner_test.go

@@ -40,8 +40,7 @@ func TestScanner(t *testing.T) {
 				filepath.FromSlash("work/subdir/other"):   {Files: 5, Bytes: 60},
 				filepath.FromSlash("work/subdir"):         {Files: 5, Dirs: 1, Bytes: 60},
 				filepath.FromSlash("work"):                {Files: 5, Dirs: 2, Bytes: 60},
-				filepath.FromSlash("."):                   {Files: 5, Dirs: 3, Bytes: 60},
-				filepath.FromSlash(""):                    {Files: 5, Dirs: 3, Bytes: 60},
+				filepath.FromSlash(""):                    {Files: 5, Dirs: 2, Bytes: 60},
 			},
 		},
 		{
@@ -72,8 +71,7 @@ func TestScanner(t *testing.T) {
 				filepath.FromSlash("work/subdir/bar.txt"): {Files: 2, Bytes: 30},
 				filepath.FromSlash("work/subdir"):         {Files: 2, Dirs: 1, Bytes: 30},
 				filepath.FromSlash("work"):                {Files: 2, Dirs: 2, Bytes: 30},
-				filepath.FromSlash("."):                   {Files: 2, Dirs: 3, Bytes: 30},
-				filepath.FromSlash(""):                    {Files: 2, Dirs: 3, Bytes: 30},
+				filepath.FromSlash(""):                    {Files: 2, Dirs: 2, Bytes: 30},
 			},
 		},
 	}
@@ -152,7 +150,7 @@ func TestScannerError(t *testing.T) {
 					},
 				},
 			},
-			result: ScanStats{Files: 5, Dirs: 3, Bytes: 60},
+			result: ScanStats{Files: 5, Dirs: 2, Bytes: 60},
 		},
 		{
 			name: "unreadable-dir",
@@ -168,7 +166,7 @@ func TestScannerError(t *testing.T) {
 					},
 				},
 			},
-			result: ScanStats{Files: 3, Dirs: 2, Bytes: 28},
+			result: ScanStats{Files: 3, Dirs: 1, Bytes: 28},
 			prepare: func(t testing.TB) {
 				err := os.Chmod(filepath.Join("work", "subdir"), 0000)
 				if err != nil {
@@ -191,7 +189,7 @@ func TestScannerError(t *testing.T) {
 				"foo":   TestFile{Content: "foo"},
 				"other": TestFile{Content: "other"},
 			},
-			result: ScanStats{Files: 3, Dirs: 1, Bytes: 11},
+			result: ScanStats{Files: 3, Dirs: 0, Bytes: 11},
 			resFn: func(t testing.TB, item string, s ScanStats) {
 				if item == "bar" {
 					err := os.Remove("foo")
@@ -289,7 +287,7 @@ func TestScannerCancel(t *testing.T) {
 		"other": TestFile{Content: "other"},
 	}
 
-	result := ScanStats{Files: 2, Dirs: 1, Bytes: 6}
+	result := ScanStats{Files: 2, Dirs: 0, Bytes: 6}
 
 	ctx, cancel := context.WithCancel(context.Background())
 	defer cancel()