cri-o/vendor/github.com/Microsoft/go-winio/wim/wim.go
Mrunal Patel 8e5b17cf13 Switch to github.com/golang/dep for vendoring
Signed-off-by: Mrunal Patel <mrunalp@gmail.com>
2017-01-31 16:45:59 -08:00

866 lines
21 KiB
Go

// Package wim implements a WIM file parser.
//
// WIM files are used to distribute Windows file system and container images.
// They are documented at https://msdn.microsoft.com/en-us/library/windows/desktop/dd861280.aspx.
package wim
import (
"bytes"
"crypto/sha1"
"encoding/binary"
"encoding/xml"
"errors"
"fmt"
"io"
"io/ioutil"
"strconv"
"sync"
"time"
"unicode/utf16"
)
// File attribute constants from Windows.
const (
FILE_ATTRIBUTE_READONLY = 0x00000001
FILE_ATTRIBUTE_HIDDEN = 0x00000002
FILE_ATTRIBUTE_SYSTEM = 0x00000004
FILE_ATTRIBUTE_DIRECTORY = 0x00000010
FILE_ATTRIBUTE_ARCHIVE = 0x00000020
FILE_ATTRIBUTE_DEVICE = 0x00000040
FILE_ATTRIBUTE_NORMAL = 0x00000080
FILE_ATTRIBUTE_TEMPORARY = 0x00000100
FILE_ATTRIBUTE_SPARSE_FILE = 0x00000200
FILE_ATTRIBUTE_REPARSE_POINT = 0x00000400
FILE_ATTRIBUTE_COMPRESSED = 0x00000800
FILE_ATTRIBUTE_OFFLINE = 0x00001000
FILE_ATTRIBUTE_NOT_CONTENT_INDEXED = 0x00002000
FILE_ATTRIBUTE_ENCRYPTED = 0x00004000
FILE_ATTRIBUTE_INTEGRITY_STREAM = 0x00008000
FILE_ATTRIBUTE_VIRTUAL = 0x00010000
FILE_ATTRIBUTE_NO_SCRUB_DATA = 0x00020000
FILE_ATTRIBUTE_EA = 0x00040000
)
// Windows processor architectures.
const (
PROCESSOR_ARCHITECTURE_INTEL = 0
PROCESSOR_ARCHITECTURE_MIPS = 1
PROCESSOR_ARCHITECTURE_ALPHA = 2
PROCESSOR_ARCHITECTURE_PPC = 3
PROCESSOR_ARCHITECTURE_SHX = 4
PROCESSOR_ARCHITECTURE_ARM = 5
PROCESSOR_ARCHITECTURE_IA64 = 6
PROCESSOR_ARCHITECTURE_ALPHA64 = 7
PROCESSOR_ARCHITECTURE_MSIL = 8
PROCESSOR_ARCHITECTURE_AMD64 = 9
PROCESSOR_ARCHITECTURE_IA32_ON_WIN64 = 10
PROCESSOR_ARCHITECTURE_NEUTRAL = 11
PROCESSOR_ARCHITECTURE_ARM64 = 12
)
var wimImageTag = [...]byte{'M', 'S', 'W', 'I', 'M', 0, 0, 0}
type guid struct {
Data1 uint32
Data2 uint16
Data3 uint16
Data4 [8]byte
}
func (g guid) String() string {
return fmt.Sprintf("%08x-%04x-%04x-%02x%02x-%02x%02x%02x%02x%02x%02x", g.Data1, g.Data2, g.Data3, g.Data4[0], g.Data4[1], g.Data4[2], g.Data4[3], g.Data4[4], g.Data4[5], g.Data4[6], g.Data4[7])
}
type resourceDescriptor struct {
FlagsAndCompressedSize uint64
Offset int64
OriginalSize int64
}
type resFlag byte
const (
resFlagFree resFlag = 1 << iota
resFlagMetadata
resFlagCompressed
resFlagSpanned
)
const validate = false
const supportedResFlags = resFlagMetadata | resFlagCompressed
func (r *resourceDescriptor) Flags() resFlag {
return resFlag(r.FlagsAndCompressedSize >> 56)
}
func (r *resourceDescriptor) CompressedSize() int64 {
return int64(r.FlagsAndCompressedSize & 0xffffffffffffff)
}
func (r *resourceDescriptor) String() string {
s := fmt.Sprintf("%d bytes at %d", r.CompressedSize(), r.Offset)
if r.Flags()&4 != 0 {
s += fmt.Sprintf(" (uncompresses to %d)", r.OriginalSize)
}
return s
}
// SHA1Hash contains the SHA1 hash of a file or stream.
type SHA1Hash [20]byte
type streamDescriptor struct {
resourceDescriptor
PartNumber uint16
RefCount uint32
Hash SHA1Hash
}
type hdrFlag uint32
const (
hdrFlagReserved hdrFlag = 1 << iota
hdrFlagCompressed
hdrFlagReadOnly
hdrFlagSpanned
hdrFlagResourceOnly
hdrFlagMetadataOnly
hdrFlagWriteInProgress
hdrFlagRpFix
)
const (
hdrFlagCompressReserved hdrFlag = 1 << (iota + 16)
hdrFlagCompressXpress
hdrFlagCompressLzx
)
const supportedHdrFlags = hdrFlagRpFix | hdrFlagReadOnly | hdrFlagCompressed | hdrFlagCompressLzx
type wimHeader struct {
ImageTag [8]byte
Size uint32
Version uint32
Flags hdrFlag
CompressionSize uint32
WIMGuid guid
PartNumber uint16
TotalParts uint16
ImageCount uint32
OffsetTable resourceDescriptor
XMLData resourceDescriptor
BootMetadata resourceDescriptor
BootIndex uint32
Padding uint32
Integrity resourceDescriptor
Unused [60]byte
}
type securityblockDisk struct {
TotalLength uint32
NumEntries uint32
}
const securityblockDiskSize = 8
type direntry struct {
Attributes uint32
SecurityID uint32
SubdirOffset int64
Unused1, Unused2 int64
CreationTime Filetime
LastAccessTime Filetime
LastWriteTime Filetime
Hash SHA1Hash
Padding uint32
ReparseHardLink int64
StreamCount uint16
ShortNameLength uint16
FileNameLength uint16
}
var direntrySize = int64(binary.Size(direntry{}) + 8) // includes an 8-byte length prefix
type streamentry struct {
Unused int64
Hash SHA1Hash
NameLength int16
}
var streamentrySize = int64(binary.Size(streamentry{}) + 8) // includes an 8-byte length prefix
// Filetime represents a Windows time.
type Filetime struct {
LowDateTime uint32
HighDateTime uint32
}
// Time returns the time as time.Time.
func (ft *Filetime) Time() time.Time {
// 100-nanosecond intervals since January 1, 1601
nsec := int64(ft.HighDateTime)<<32 + int64(ft.LowDateTime)
// change starting time to the Epoch (00:00:00 UTC, January 1, 1970)
nsec -= 116444736000000000
// convert into nanoseconds
nsec *= 100
return time.Unix(0, nsec)
}
// UnmarshalXML unmarshals the time from a WIM XML blob.
func (ft *Filetime) UnmarshalXML(d *xml.Decoder, start xml.StartElement) error {
type time struct {
Low string `xml:"LOWPART"`
High string `xml:"HIGHPART"`
}
var t time
err := d.DecodeElement(&t, &start)
if err != nil {
return err
}
low, err := strconv.ParseUint(t.Low, 0, 32)
if err != nil {
return err
}
high, err := strconv.ParseUint(t.High, 0, 32)
if err != nil {
return err
}
ft.LowDateTime = uint32(low)
ft.HighDateTime = uint32(high)
return nil
}
type info struct {
Image []ImageInfo `xml:"IMAGE"`
}
// ImageInfo contains information about the image.
type ImageInfo struct {
Name string `xml:"NAME"`
Index int `xml:"INDEX,attr"`
CreationTime Filetime `xml:"CREATIONTIME"`
ModTime Filetime `xml:"LASTMODIFICATIONTIME"`
Windows *WindowsInfo `xml:"WINDOWS"`
}
// WindowsInfo contains information about the Windows installation in the image.
type WindowsInfo struct {
Arch byte `xml:"ARCH"`
ProductName string `xml:"PRODUCTNAME"`
EditionID string `xml:"EDITIONID"`
InstallationType string `xml:"INSTALLATIONTYPE"`
ProductType string `xml:"PRODUCTTYPE"`
Languages []string `xml:"LANGUAGES>LANGUAGE"`
DefaultLanguage string `xml:"LANGUAGES>DEFAULT"`
Version Version `xml:"VERSION"`
SystemRoot string `xml:"SYSTEMROOT"`
}
// Version represents a Windows build version.
type Version struct {
Major int `xml:"MAJOR"`
Minor int `xml:"MINOR"`
Build int `xml:"BUILD"`
SPBuild int `xml:"SPBUILD"`
SPLevel int `xml:"SPLEVEL"`
}
// ParseError is returned when the WIM cannot be parsed.
type ParseError struct {
Oper string
Path string
Err error
}
func (e *ParseError) Error() string {
if e.Path == "" {
return "WIM parse error at " + e.Oper + ": " + e.Err.Error()
}
return fmt.Sprintf("WIM parse error: %s %s: %s", e.Oper, e.Path, e.Err.Error())
}
// Reader provides functions to read a WIM file.
type Reader struct {
hdr wimHeader
r io.ReaderAt
fileData map[SHA1Hash]resourceDescriptor
XMLInfo string // The XML information about the WIM.
Image []*Image // The WIM's images.
}
// Image represents an image within a WIM file.
type Image struct {
wim *Reader
offset resourceDescriptor
sds [][]byte
rootOffset int64
r io.ReadCloser
curOffset int64
m sync.Mutex
ImageInfo
}
// StreamHeader contains alternate data stream metadata.
type StreamHeader struct {
Name string
Hash SHA1Hash
Size int64
}
// Stream represents an alternate data stream or reparse point data stream.
type Stream struct {
StreamHeader
wim *Reader
offset resourceDescriptor
}
// FileHeader contains file metadata.
type FileHeader struct {
Name string
ShortName string
Attributes uint32
SecurityDescriptor []byte
CreationTime Filetime
LastAccessTime Filetime
LastWriteTime Filetime
Hash SHA1Hash
Size int64
LinkID int64
ReparseTag uint32
ReparseReserved uint32
}
// File represents a file or directory in a WIM image.
type File struct {
FileHeader
Streams []*Stream
offset resourceDescriptor
img *Image
subdirOffset int64
}
// NewReader returns a Reader that can be used to read WIM file data.
func NewReader(f io.ReaderAt) (*Reader, error) {
r := &Reader{r: f}
section := io.NewSectionReader(f, 0, 0xffff)
err := binary.Read(section, binary.LittleEndian, &r.hdr)
if err != nil {
return nil, err
}
if r.hdr.ImageTag != wimImageTag {
return nil, &ParseError{Oper: "image tag", Err: errors.New("not a WIM file")}
}
if r.hdr.Flags&^supportedHdrFlags != 0 {
return nil, fmt.Errorf("unsupported WIM flags %x", r.hdr.Flags&^supportedHdrFlags)
}
if r.hdr.CompressionSize != 0x8000 {
return nil, fmt.Errorf("unsupported compression size %d", r.hdr.CompressionSize)
}
if r.hdr.TotalParts != 1 {
return nil, errors.New("multi-part WIM not supported")
}
fileData, images, err := r.readOffsetTable(&r.hdr.OffsetTable)
if err != nil {
return nil, err
}
xmlinfo, err := r.readXML()
if err != nil {
return nil, err
}
var info info
err = xml.Unmarshal([]byte(xmlinfo), &info)
if err != nil {
return nil, &ParseError{Oper: "XML info", Err: err}
}
for i, img := range images {
for _, imgInfo := range info.Image {
if imgInfo.Index == i+1 {
img.ImageInfo = imgInfo
break
}
}
}
r.fileData = fileData
r.Image = images
r.XMLInfo = xmlinfo
return r, nil
}
// Close releases resources associated with the Reader.
func (r *Reader) Close() error {
for _, img := range r.Image {
img.reset()
}
return nil
}
func (r *Reader) resourceReader(hdr *resourceDescriptor) (io.ReadCloser, error) {
return r.resourceReaderWithOffset(hdr, 0)
}
func (r *Reader) resourceReaderWithOffset(hdr *resourceDescriptor, offset int64) (io.ReadCloser, error) {
var sr io.ReadCloser
section := io.NewSectionReader(r.r, hdr.Offset, hdr.CompressedSize())
if hdr.Flags()&resFlagCompressed == 0 {
section.Seek(offset, 0)
sr = ioutil.NopCloser(section)
} else {
cr, err := newCompressedReader(section, hdr.OriginalSize, offset)
if err != nil {
return nil, err
}
sr = cr
}
return sr, nil
}
func (r *Reader) readResource(hdr *resourceDescriptor) ([]byte, error) {
rsrc, err := r.resourceReader(hdr)
if err != nil {
return nil, err
}
defer rsrc.Close()
return ioutil.ReadAll(rsrc)
}
func (r *Reader) readXML() (string, error) {
if r.hdr.XMLData.CompressedSize() == 0 {
return "", nil
}
rsrc, err := r.resourceReader(&r.hdr.XMLData)
if err != nil {
return "", err
}
defer rsrc.Close()
XMLData := make([]uint16, r.hdr.XMLData.OriginalSize/2)
err = binary.Read(rsrc, binary.LittleEndian, XMLData)
if err != nil {
return "", &ParseError{Oper: "XML data", Err: err}
}
// The BOM will always indicate little-endian UTF-16.
if XMLData[0] != 0xfeff {
return "", &ParseError{Oper: "XML data", Err: errors.New("invalid BOM")}
}
return string(utf16.Decode(XMLData[1:])), nil
}
func (r *Reader) readOffsetTable(res *resourceDescriptor) (map[SHA1Hash]resourceDescriptor, []*Image, error) {
fileData := make(map[SHA1Hash]resourceDescriptor)
var images []*Image
offsetTable, err := r.readResource(res)
if err != nil {
return nil, nil, &ParseError{Oper: "offset table", Err: err}
}
br := bytes.NewReader(offsetTable)
for i := 0; ; i++ {
var res streamDescriptor
err := binary.Read(br, binary.LittleEndian, &res)
if err == io.EOF {
break
}
if err != nil {
return nil, nil, &ParseError{Oper: "offset table", Err: err}
}
if res.Flags()&^supportedResFlags != 0 {
return nil, nil, &ParseError{Oper: "offset table", Err: errors.New("unsupported resource flag")}
}
// Validation for ad-hoc testing
if validate {
sec, err := r.resourceReader(&res.resourceDescriptor)
if err != nil {
panic(fmt.Sprint(i, err))
}
hash := sha1.New()
_, err = io.Copy(hash, sec)
sec.Close()
if err != nil {
panic(fmt.Sprint(i, err))
}
var cmphash SHA1Hash
copy(cmphash[:], hash.Sum(nil))
if cmphash != res.Hash {
panic(fmt.Sprint(i, "hash mismatch"))
}
}
if res.Flags()&resFlagMetadata != 0 {
image := &Image{
wim: r,
offset: res.resourceDescriptor,
}
images = append(images, image)
} else {
fileData[res.Hash] = res.resourceDescriptor
}
}
if len(images) != int(r.hdr.ImageCount) {
return nil, nil, &ParseError{Oper: "offset table", Err: errors.New("mismatched image count")}
}
return fileData, images, nil
}
func (r *Reader) readSecurityDescriptors(rsrc io.Reader) (sds [][]byte, n int64, err error) {
var secBlock securityblockDisk
err = binary.Read(rsrc, binary.LittleEndian, &secBlock)
if err != nil {
err = &ParseError{Oper: "security table", Err: err}
return
}
n += securityblockDiskSize
secSizes := make([]int64, secBlock.NumEntries)
err = binary.Read(rsrc, binary.LittleEndian, &secSizes)
if err != nil {
err = &ParseError{Oper: "security table sizes", Err: err}
return
}
n += int64(secBlock.NumEntries * 8)
sds = make([][]byte, secBlock.NumEntries)
for i, size := range secSizes {
sd := make([]byte, size&0xffffffff)
_, err = io.ReadFull(rsrc, sd)
if err != nil {
err = &ParseError{Oper: "security descriptor", Err: err}
return
}
n += int64(len(sd))
sds[i] = sd
}
secsize := int64((secBlock.TotalLength + 7) &^ 7)
if n > secsize {
err = &ParseError{Oper: "security descriptor", Err: errors.New("security descriptor table too small")}
return
}
_, err = io.CopyN(ioutil.Discard, rsrc, secsize-n)
if err != nil {
return
}
n = secsize
return
}
// Open parses the image and returns the root directory.
func (img *Image) Open() (*File, error) {
if img.sds == nil {
rsrc, err := img.wim.resourceReaderWithOffset(&img.offset, img.rootOffset)
if err != nil {
return nil, err
}
sds, n, err := img.wim.readSecurityDescriptors(rsrc)
if err != nil {
rsrc.Close()
return nil, err
}
img.sds = sds
img.r = rsrc
img.rootOffset = n
img.curOffset = n
}
f, err := img.readdir(img.rootOffset)
if err != nil {
return nil, err
}
if len(f) != 1 {
return nil, &ParseError{Oper: "root directory", Err: errors.New("expected exactly 1 root directory entry")}
}
return f[0], err
}
func (img *Image) reset() {
if img.r != nil {
img.r.Close()
img.r = nil
}
img.curOffset = -1
}
func (img *Image) readdir(offset int64) ([]*File, error) {
img.m.Lock()
defer img.m.Unlock()
if offset < img.curOffset || offset > img.curOffset+chunkSize {
// Reset to seek backward or to seek forward very far.
img.reset()
}
if img.r == nil {
rsrc, err := img.wim.resourceReaderWithOffset(&img.offset, offset)
if err != nil {
return nil, err
}
img.r = rsrc
img.curOffset = offset
}
if offset > img.curOffset {
_, err := io.CopyN(ioutil.Discard, img.r, offset-img.curOffset)
if err != nil {
img.reset()
if err == io.EOF {
err = io.ErrUnexpectedEOF
}
return nil, err
}
}
var entries []*File
for {
e, n, err := img.readNextEntry(img.r)
img.curOffset += n
if err == io.EOF {
break
}
if err != nil {
img.reset()
return nil, err
}
entries = append(entries, e)
}
return entries, nil
}
func (img *Image) readNextEntry(r io.Reader) (*File, int64, error) {
var length int64
err := binary.Read(r, binary.LittleEndian, &length)
if err != nil {
return nil, 0, &ParseError{Oper: "directory length check", Err: err}
}
if length == 0 {
return nil, 8, io.EOF
}
left := length
if left < direntrySize {
return nil, 0, &ParseError{Oper: "directory entry", Err: errors.New("size too short")}
}
var dentry direntry
err = binary.Read(r, binary.LittleEndian, &dentry)
if err != nil {
return nil, 0, &ParseError{Oper: "directory entry", Err: err}
}
left -= direntrySize
namesLen := int64(dentry.FileNameLength + 2 + dentry.ShortNameLength)
if left < namesLen {
return nil, 0, &ParseError{Oper: "directory entry", Err: errors.New("size too short for names")}
}
names := make([]uint16, namesLen/2)
err = binary.Read(r, binary.LittleEndian, names)
if err != nil {
return nil, 0, &ParseError{Oper: "file name", Err: err}
}
left -= namesLen
var name, shortName string
if dentry.FileNameLength > 0 {
name = string(utf16.Decode(names[:dentry.FileNameLength/2]))
}
if dentry.ShortNameLength > 0 {
shortName = string(utf16.Decode(names[dentry.FileNameLength/2+1:]))
}
var offset resourceDescriptor
zerohash := SHA1Hash{}
if dentry.Hash != zerohash {
var ok bool
offset, ok = img.wim.fileData[dentry.Hash]
if !ok {
return nil, 0, &ParseError{Oper: "directory entry", Path: name, Err: fmt.Errorf("could not find file data matching hash %#v", dentry)}
}
}
f := &File{
FileHeader: FileHeader{
Attributes: dentry.Attributes,
CreationTime: dentry.CreationTime,
LastAccessTime: dentry.LastAccessTime,
LastWriteTime: dentry.LastWriteTime,
Hash: dentry.Hash,
Size: offset.OriginalSize,
Name: name,
ShortName: shortName,
},
offset: offset,
img: img,
subdirOffset: dentry.SubdirOffset,
}
isDir := false
if dentry.Attributes&FILE_ATTRIBUTE_REPARSE_POINT == 0 {
f.LinkID = dentry.ReparseHardLink
if dentry.Attributes&FILE_ATTRIBUTE_DIRECTORY != 0 {
isDir = true
}
} else {
f.ReparseTag = uint32(dentry.ReparseHardLink)
f.ReparseReserved = uint32(dentry.ReparseHardLink >> 32)
}
if isDir && f.subdirOffset == 0 {
return nil, 0, &ParseError{Oper: "directory entry", Path: name, Err: errors.New("no subdirectory data for directory")}
} else if !isDir && f.subdirOffset != 0 {
return nil, 0, &ParseError{Oper: "directory entry", Path: name, Err: errors.New("unexpected subdirectory data for non-directory")}
}
if dentry.SecurityID != 0xffffffff {
f.SecurityDescriptor = img.sds[dentry.SecurityID]
}
_, err = io.CopyN(ioutil.Discard, r, left)
if err != nil {
if err == io.EOF {
err = io.ErrUnexpectedEOF
}
return nil, 0, err
}
if dentry.StreamCount > 0 {
var streams []*Stream
for i := uint16(0); i < dentry.StreamCount; i++ {
s, n, err := img.readNextStream(r)
length += n
if err != nil {
return nil, 0, err
}
// The first unnamed stream should be treated as the file stream.
if i == 0 && s.Name == "" {
f.Hash = s.Hash
f.Size = s.Size
f.offset = s.offset
} else if s.Name != "" {
streams = append(streams, s)
}
}
f.Streams = streams
}
if dentry.Attributes&FILE_ATTRIBUTE_REPARSE_POINT != 0 && f.Size == 0 {
return nil, 0, &ParseError{Oper: "directory entry", Path: name, Err: errors.New("reparse point is missing reparse stream")}
}
return f, length, nil
}
func (img *Image) readNextStream(r io.Reader) (*Stream, int64, error) {
var length int64
err := binary.Read(r, binary.LittleEndian, &length)
if err != nil {
if err == io.EOF {
err = io.ErrUnexpectedEOF
}
return nil, 0, &ParseError{Oper: "stream length check", Err: err}
}
left := length
if left < streamentrySize {
return nil, 0, &ParseError{Oper: "stream entry", Err: errors.New("size too short")}
}
var sentry streamentry
err = binary.Read(r, binary.LittleEndian, &sentry)
if err != nil {
return nil, 0, &ParseError{Oper: "stream entry", Err: err}
}
left -= streamentrySize
if left < int64(sentry.NameLength) {
return nil, 0, &ParseError{Oper: "stream entry", Err: errors.New("size too short for name")}
}
names := make([]uint16, sentry.NameLength/2)
err = binary.Read(r, binary.LittleEndian, names)
if err != nil {
return nil, 0, &ParseError{Oper: "file name", Err: err}
}
left -= int64(sentry.NameLength)
name := string(utf16.Decode(names))
var offset resourceDescriptor
if sentry.Hash != (SHA1Hash{}) {
var ok bool
offset, ok = img.wim.fileData[sentry.Hash]
if !ok {
return nil, 0, &ParseError{Oper: "stream entry", Path: name, Err: fmt.Errorf("could not find file data matching hash %v", sentry.Hash)}
}
}
s := &Stream{
StreamHeader: StreamHeader{
Hash: sentry.Hash,
Size: offset.OriginalSize,
Name: name,
},
wim: img.wim,
offset: offset,
}
_, err = io.CopyN(ioutil.Discard, r, left)
if err != nil {
if err == io.EOF {
err = io.ErrUnexpectedEOF
}
return nil, 0, err
}
return s, length, nil
}
// Open returns an io.ReadCloser that can be used to read the stream's contents.
func (s *Stream) Open() (io.ReadCloser, error) {
return s.wim.resourceReader(&s.offset)
}
// Open returns an io.ReadCloser that can be used to read the file's contents.
func (f *File) Open() (io.ReadCloser, error) {
return f.img.wim.resourceReader(&f.offset)
}
// Readdir reads the directory entries.
func (f *File) Readdir() ([]*File, error) {
if !f.IsDir() {
return nil, errors.New("not a directory")
}
return f.img.readdir(f.subdirOffset)
}
// IsDir returns whether the given file is a directory. It returns false when it
// is a directory reparse point.
func (f *FileHeader) IsDir() bool {
return f.Attributes&(FILE_ATTRIBUTE_DIRECTORY|FILE_ATTRIBUTE_REPARSE_POINT) == FILE_ATTRIBUTE_DIRECTORY
}