tarfs: handle tars without trailer

Spotted some layers in the wild without the two-block trailer. This changes the
tar parsing code to allow EOFs at block boundaries.

Signed-off-by: Hank Donnay <[email protected]>

pkg/tarfs/parse.go

@@ -58,14 +58,18 @@ Scan:
 		off := blk * blockSz
 		n, err := r.ReadAt(b, off)
 		switch {
-		case n != blockSz:
+		case errors.Is(err, nil) && n != blockSz:
 			// Should be impossible with a well-formed archive, so raise an
-			// error.
-			return nil, parseErr("short read at offset: %d", off)
-		case errors.Is(err, io.EOF) && !zeroes:
+			// error. Should also be impossible with a conforming [io.ReaderAt].
+			return nil, parseErr("short read at offset: %d (got: %d, want: %d)", off, n, blockSz)
+		case errors.Is(err, nil): // OK
+		case errors.Is(err, io.EOF) && n == 0:
 			// This is an early EOF: in a well-formed archive, ReadAt should
 			// only return EOF on the second block of zeroes.
-		case errors.Is(err, nil): // OK
+			//
+			// Handle this case because some layers in the wild are a single
+			// directory block, with no trailer.
+			break Scan
 		case errors.Is(err, io.EOF) && zeroes:
 			break Scan
 		default:

pkg/tarfs/tarfs_test.go

@@ -359,3 +359,23 @@ func TestSymlinks(t *testing.T) {
 		}
 	}))
 }
+
+func TestKnownLayers(t *testing.T) {
+	ents, err := os.ReadDir(`testdata/known`)
+	if err != nil {
+		t.Fatal(err)
+	}
+	for _, ent := range ents {
+		n := ent.Name()
+		t.Run(n, func(t *testing.T) {
+			f, err := os.Open(filepath.Join(`testdata/known`, n))
+			if err != nil {
+				t.Fatal(err)
+			}
+			defer f.Close()
+			if _, err := New(f); err != nil {
+				t.Fatal(err)
+			}
+		})
+	}
+}

pkg/tarfs/testdata/.gitignore

@@ -1,2 +1,2 @@
-*
-!.gitignore
+*.layer
+fstest.tar