| 1 |
747 |
jeremybenn |
// Copyright 2010 The Go Authors. All rights reserved.
|
| 2 |
|
|
// Use of this source code is governed by a BSD-style
|
| 3 |
|
|
// license that can be found in the LICENSE file.
|
| 4 |
|
|
|
| 5 |
|
|
package zip
|
| 6 |
|
|
|
| 7 |
|
|
import (
|
| 8 |
|
|
"bufio"
|
| 9 |
|
|
"compress/flate"
|
| 10 |
|
|
"encoding/binary"
|
| 11 |
|
|
"errors"
|
| 12 |
|
|
"hash"
|
| 13 |
|
|
"hash/crc32"
|
| 14 |
|
|
"io"
|
| 15 |
|
|
"io/ioutil"
|
| 16 |
|
|
"os"
|
| 17 |
|
|
)
|
| 18 |
|
|
|
| 19 |
|
|
var (
|
| 20 |
|
|
ErrFormat = errors.New("zip: not a valid zip file")
|
| 21 |
|
|
ErrAlgorithm = errors.New("zip: unsupported compression algorithm")
|
| 22 |
|
|
ErrChecksum = errors.New("zip: checksum error")
|
| 23 |
|
|
)
|
| 24 |
|
|
|
| 25 |
|
|
type Reader struct {
|
| 26 |
|
|
r io.ReaderAt
|
| 27 |
|
|
File []*File
|
| 28 |
|
|
Comment string
|
| 29 |
|
|
}
|
| 30 |
|
|
|
| 31 |
|
|
type ReadCloser struct {
|
| 32 |
|
|
f *os.File
|
| 33 |
|
|
Reader
|
| 34 |
|
|
}
|
| 35 |
|
|
|
| 36 |
|
|
type File struct {
|
| 37 |
|
|
FileHeader
|
| 38 |
|
|
zipr io.ReaderAt
|
| 39 |
|
|
zipsize int64
|
| 40 |
|
|
headerOffset int64
|
| 41 |
|
|
}
|
| 42 |
|
|
|
| 43 |
|
|
func (f *File) hasDataDescriptor() bool {
|
| 44 |
|
|
return f.Flags&0x8 != 0
|
| 45 |
|
|
}
|
| 46 |
|
|
|
| 47 |
|
|
// OpenReader will open the Zip file specified by name and return a ReadCloser.
|
| 48 |
|
|
func OpenReader(name string) (*ReadCloser, error) {
|
| 49 |
|
|
f, err := os.Open(name)
|
| 50 |
|
|
if err != nil {
|
| 51 |
|
|
return nil, err
|
| 52 |
|
|
}
|
| 53 |
|
|
fi, err := f.Stat()
|
| 54 |
|
|
if err != nil {
|
| 55 |
|
|
f.Close()
|
| 56 |
|
|
return nil, err
|
| 57 |
|
|
}
|
| 58 |
|
|
r := new(ReadCloser)
|
| 59 |
|
|
if err := r.init(f, fi.Size()); err != nil {
|
| 60 |
|
|
f.Close()
|
| 61 |
|
|
return nil, err
|
| 62 |
|
|
}
|
| 63 |
|
|
r.f = f
|
| 64 |
|
|
return r, nil
|
| 65 |
|
|
}
|
| 66 |
|
|
|
| 67 |
|
|
// NewReader returns a new Reader reading from r, which is assumed to
|
| 68 |
|
|
// have the given size in bytes.
|
| 69 |
|
|
func NewReader(r io.ReaderAt, size int64) (*Reader, error) {
|
| 70 |
|
|
zr := new(Reader)
|
| 71 |
|
|
if err := zr.init(r, size); err != nil {
|
| 72 |
|
|
return nil, err
|
| 73 |
|
|
}
|
| 74 |
|
|
return zr, nil
|
| 75 |
|
|
}
|
| 76 |
|
|
|
| 77 |
|
|
func (z *Reader) init(r io.ReaderAt, size int64) error {
|
| 78 |
|
|
end, err := readDirectoryEnd(r, size)
|
| 79 |
|
|
if err != nil {
|
| 80 |
|
|
return err
|
| 81 |
|
|
}
|
| 82 |
|
|
z.r = r
|
| 83 |
|
|
z.File = make([]*File, 0, end.directoryRecords)
|
| 84 |
|
|
z.Comment = end.comment
|
| 85 |
|
|
rs := io.NewSectionReader(r, 0, size)
|
| 86 |
|
|
if _, err = rs.Seek(int64(end.directoryOffset), os.SEEK_SET); err != nil {
|
| 87 |
|
|
return err
|
| 88 |
|
|
}
|
| 89 |
|
|
buf := bufio.NewReader(rs)
|
| 90 |
|
|
|
| 91 |
|
|
// The count of files inside a zip is truncated to fit in a uint16.
|
| 92 |
|
|
// Gloss over this by reading headers until we encounter
|
| 93 |
|
|
// a bad one, and then only report a ErrFormat or UnexpectedEOF if
|
| 94 |
|
|
// the file count modulo 65536 is incorrect.
|
| 95 |
|
|
for {
|
| 96 |
|
|
f := &File{zipr: r, zipsize: size}
|
| 97 |
|
|
err = readDirectoryHeader(f, buf)
|
| 98 |
|
|
if err == ErrFormat || err == io.ErrUnexpectedEOF {
|
| 99 |
|
|
break
|
| 100 |
|
|
}
|
| 101 |
|
|
if err != nil {
|
| 102 |
|
|
return err
|
| 103 |
|
|
}
|
| 104 |
|
|
z.File = append(z.File, f)
|
| 105 |
|
|
}
|
| 106 |
|
|
if uint16(len(z.File)) != end.directoryRecords {
|
| 107 |
|
|
// Return the readDirectoryHeader error if we read
|
| 108 |
|
|
// the wrong number of directory entries.
|
| 109 |
|
|
return err
|
| 110 |
|
|
}
|
| 111 |
|
|
return nil
|
| 112 |
|
|
}
|
| 113 |
|
|
|
| 114 |
|
|
// Close closes the Zip file, rendering it unusable for I/O.
|
| 115 |
|
|
func (rc *ReadCloser) Close() error {
|
| 116 |
|
|
return rc.f.Close()
|
| 117 |
|
|
}
|
| 118 |
|
|
|
| 119 |
|
|
// Open returns a ReadCloser that provides access to the File's contents.
|
| 120 |
|
|
// Multiple files may be read concurrently.
|
| 121 |
|
|
func (f *File) Open() (rc io.ReadCloser, err error) {
|
| 122 |
|
|
bodyOffset, err := f.findBodyOffset()
|
| 123 |
|
|
if err != nil {
|
| 124 |
|
|
return
|
| 125 |
|
|
}
|
| 126 |
|
|
size := int64(f.CompressedSize)
|
| 127 |
|
|
if size == 0 && f.hasDataDescriptor() {
|
| 128 |
|
|
// permit SectionReader to see the rest of the file
|
| 129 |
|
|
size = f.zipsize - (f.headerOffset + bodyOffset)
|
| 130 |
|
|
}
|
| 131 |
|
|
r := io.NewSectionReader(f.zipr, f.headerOffset+bodyOffset, size)
|
| 132 |
|
|
switch f.Method {
|
| 133 |
|
|
case Store: // (no compression)
|
| 134 |
|
|
rc = ioutil.NopCloser(r)
|
| 135 |
|
|
case Deflate:
|
| 136 |
|
|
rc = flate.NewReader(r)
|
| 137 |
|
|
default:
|
| 138 |
|
|
err = ErrAlgorithm
|
| 139 |
|
|
}
|
| 140 |
|
|
if rc != nil {
|
| 141 |
|
|
rc = &checksumReader{rc, crc32.NewIEEE(), f, r}
|
| 142 |
|
|
}
|
| 143 |
|
|
return
|
| 144 |
|
|
}
|
| 145 |
|
|
|
| 146 |
|
|
type checksumReader struct {
|
| 147 |
|
|
rc io.ReadCloser
|
| 148 |
|
|
hash hash.Hash32
|
| 149 |
|
|
f *File
|
| 150 |
|
|
zipr io.Reader // for reading the data descriptor
|
| 151 |
|
|
}
|
| 152 |
|
|
|
| 153 |
|
|
func (r *checksumReader) Read(b []byte) (n int, err error) {
|
| 154 |
|
|
n, err = r.rc.Read(b)
|
| 155 |
|
|
r.hash.Write(b[:n])
|
| 156 |
|
|
if err != io.EOF {
|
| 157 |
|
|
return
|
| 158 |
|
|
}
|
| 159 |
|
|
if r.f.hasDataDescriptor() {
|
| 160 |
|
|
if err = readDataDescriptor(r.zipr, r.f); err != nil {
|
| 161 |
|
|
return
|
| 162 |
|
|
}
|
| 163 |
|
|
}
|
| 164 |
|
|
if r.hash.Sum32() != r.f.CRC32 {
|
| 165 |
|
|
err = ErrChecksum
|
| 166 |
|
|
}
|
| 167 |
|
|
return
|
| 168 |
|
|
}
|
| 169 |
|
|
|
| 170 |
|
|
func (r *checksumReader) Close() error { return r.rc.Close() }
|
| 171 |
|
|
|
| 172 |
|
|
func readFileHeader(f *File, r io.Reader) error {
|
| 173 |
|
|
var b [fileHeaderLen]byte
|
| 174 |
|
|
if _, err := io.ReadFull(r, b[:]); err != nil {
|
| 175 |
|
|
return err
|
| 176 |
|
|
}
|
| 177 |
|
|
c := binary.LittleEndian
|
| 178 |
|
|
if sig := c.Uint32(b[:4]); sig != fileHeaderSignature {
|
| 179 |
|
|
return ErrFormat
|
| 180 |
|
|
}
|
| 181 |
|
|
f.ReaderVersion = c.Uint16(b[4:6])
|
| 182 |
|
|
f.Flags = c.Uint16(b[6:8])
|
| 183 |
|
|
f.Method = c.Uint16(b[8:10])
|
| 184 |
|
|
f.ModifiedTime = c.Uint16(b[10:12])
|
| 185 |
|
|
f.ModifiedDate = c.Uint16(b[12:14])
|
| 186 |
|
|
f.CRC32 = c.Uint32(b[14:18])
|
| 187 |
|
|
f.CompressedSize = c.Uint32(b[18:22])
|
| 188 |
|
|
f.UncompressedSize = c.Uint32(b[22:26])
|
| 189 |
|
|
filenameLen := int(c.Uint16(b[26:28]))
|
| 190 |
|
|
extraLen := int(c.Uint16(b[28:30]))
|
| 191 |
|
|
d := make([]byte, filenameLen+extraLen)
|
| 192 |
|
|
if _, err := io.ReadFull(r, d); err != nil {
|
| 193 |
|
|
return err
|
| 194 |
|
|
}
|
| 195 |
|
|
f.Name = string(d[:filenameLen])
|
| 196 |
|
|
f.Extra = d[filenameLen:]
|
| 197 |
|
|
return nil
|
| 198 |
|
|
}
|
| 199 |
|
|
|
| 200 |
|
|
// findBodyOffset does the minimum work to verify the file has a header
|
| 201 |
|
|
// and returns the file body offset.
|
| 202 |
|
|
func (f *File) findBodyOffset() (int64, error) {
|
| 203 |
|
|
r := io.NewSectionReader(f.zipr, f.headerOffset, f.zipsize-f.headerOffset)
|
| 204 |
|
|
var b [fileHeaderLen]byte
|
| 205 |
|
|
if _, err := io.ReadFull(r, b[:]); err != nil {
|
| 206 |
|
|
return 0, err
|
| 207 |
|
|
}
|
| 208 |
|
|
c := binary.LittleEndian
|
| 209 |
|
|
if sig := c.Uint32(b[:4]); sig != fileHeaderSignature {
|
| 210 |
|
|
return 0, ErrFormat
|
| 211 |
|
|
}
|
| 212 |
|
|
filenameLen := int(c.Uint16(b[26:28]))
|
| 213 |
|
|
extraLen := int(c.Uint16(b[28:30]))
|
| 214 |
|
|
return int64(fileHeaderLen + filenameLen + extraLen), nil
|
| 215 |
|
|
}
|
| 216 |
|
|
|
| 217 |
|
|
// readDirectoryHeader attempts to read a directory header from r.
|
| 218 |
|
|
// It returns io.ErrUnexpectedEOF if it cannot read a complete header,
|
| 219 |
|
|
// and ErrFormat if it doesn't find a valid header signature.
|
| 220 |
|
|
func readDirectoryHeader(f *File, r io.Reader) error {
|
| 221 |
|
|
var b [directoryHeaderLen]byte
|
| 222 |
|
|
if _, err := io.ReadFull(r, b[:]); err != nil {
|
| 223 |
|
|
return err
|
| 224 |
|
|
}
|
| 225 |
|
|
c := binary.LittleEndian
|
| 226 |
|
|
if sig := c.Uint32(b[:4]); sig != directoryHeaderSignature {
|
| 227 |
|
|
return ErrFormat
|
| 228 |
|
|
}
|
| 229 |
|
|
f.CreatorVersion = c.Uint16(b[4:6])
|
| 230 |
|
|
f.ReaderVersion = c.Uint16(b[6:8])
|
| 231 |
|
|
f.Flags = c.Uint16(b[8:10])
|
| 232 |
|
|
f.Method = c.Uint16(b[10:12])
|
| 233 |
|
|
f.ModifiedTime = c.Uint16(b[12:14])
|
| 234 |
|
|
f.ModifiedDate = c.Uint16(b[14:16])
|
| 235 |
|
|
f.CRC32 = c.Uint32(b[16:20])
|
| 236 |
|
|
f.CompressedSize = c.Uint32(b[20:24])
|
| 237 |
|
|
f.UncompressedSize = c.Uint32(b[24:28])
|
| 238 |
|
|
filenameLen := int(c.Uint16(b[28:30]))
|
| 239 |
|
|
extraLen := int(c.Uint16(b[30:32]))
|
| 240 |
|
|
commentLen := int(c.Uint16(b[32:34]))
|
| 241 |
|
|
// startDiskNumber := c.Uint16(b[34:36]) // Unused
|
| 242 |
|
|
// internalAttributes := c.Uint16(b[36:38]) // Unused
|
| 243 |
|
|
f.ExternalAttrs = c.Uint32(b[38:42])
|
| 244 |
|
|
f.headerOffset = int64(c.Uint32(b[42:46]))
|
| 245 |
|
|
d := make([]byte, filenameLen+extraLen+commentLen)
|
| 246 |
|
|
if _, err := io.ReadFull(r, d); err != nil {
|
| 247 |
|
|
return err
|
| 248 |
|
|
}
|
| 249 |
|
|
f.Name = string(d[:filenameLen])
|
| 250 |
|
|
f.Extra = d[filenameLen : filenameLen+extraLen]
|
| 251 |
|
|
f.Comment = string(d[filenameLen+extraLen:])
|
| 252 |
|
|
return nil
|
| 253 |
|
|
}
|
| 254 |
|
|
|
| 255 |
|
|
func readDataDescriptor(r io.Reader, f *File) error {
|
| 256 |
|
|
var b [dataDescriptorLen]byte
|
| 257 |
|
|
if _, err := io.ReadFull(r, b[:]); err != nil {
|
| 258 |
|
|
return err
|
| 259 |
|
|
}
|
| 260 |
|
|
c := binary.LittleEndian
|
| 261 |
|
|
f.CRC32 = c.Uint32(b[:4])
|
| 262 |
|
|
f.CompressedSize = c.Uint32(b[4:8])
|
| 263 |
|
|
f.UncompressedSize = c.Uint32(b[8:12])
|
| 264 |
|
|
return nil
|
| 265 |
|
|
}
|
| 266 |
|
|
|
| 267 |
|
|
func readDirectoryEnd(r io.ReaderAt, size int64) (dir *directoryEnd, err error) {
|
| 268 |
|
|
// look for directoryEndSignature in the last 1k, then in the last 65k
|
| 269 |
|
|
var b []byte
|
| 270 |
|
|
for i, bLen := range []int64{1024, 65 * 1024} {
|
| 271 |
|
|
if bLen > size {
|
| 272 |
|
|
bLen = size
|
| 273 |
|
|
}
|
| 274 |
|
|
b = make([]byte, int(bLen))
|
| 275 |
|
|
if _, err := r.ReadAt(b, size-bLen); err != nil && err != io.EOF {
|
| 276 |
|
|
return nil, err
|
| 277 |
|
|
}
|
| 278 |
|
|
if p := findSignatureInBlock(b); p >= 0 {
|
| 279 |
|
|
b = b[p:]
|
| 280 |
|
|
break
|
| 281 |
|
|
}
|
| 282 |
|
|
if i == 1 || bLen == size {
|
| 283 |
|
|
return nil, ErrFormat
|
| 284 |
|
|
}
|
| 285 |
|
|
}
|
| 286 |
|
|
|
| 287 |
|
|
// read header into struct
|
| 288 |
|
|
c := binary.LittleEndian
|
| 289 |
|
|
d := new(directoryEnd)
|
| 290 |
|
|
d.diskNbr = c.Uint16(b[4:6])
|
| 291 |
|
|
d.dirDiskNbr = c.Uint16(b[6:8])
|
| 292 |
|
|
d.dirRecordsThisDisk = c.Uint16(b[8:10])
|
| 293 |
|
|
d.directoryRecords = c.Uint16(b[10:12])
|
| 294 |
|
|
d.directorySize = c.Uint32(b[12:16])
|
| 295 |
|
|
d.directoryOffset = c.Uint32(b[16:20])
|
| 296 |
|
|
d.commentLen = c.Uint16(b[20:22])
|
| 297 |
|
|
d.comment = string(b[22 : 22+int(d.commentLen)])
|
| 298 |
|
|
return d, nil
|
| 299 |
|
|
}
|
| 300 |
|
|
|
| 301 |
|
|
func findSignatureInBlock(b []byte) int {
|
| 302 |
|
|
for i := len(b) - directoryEndLen; i >= 0; i-- {
|
| 303 |
|
|
// defined from directoryEndSignature in struct.go
|
| 304 |
|
|
if b[i] == 'P' && b[i+1] == 'K' && b[i+2] == 0x05 && b[i+3] == 0x06 {
|
| 305 |
|
|
// n is length of comment
|
| 306 |
|
|
n := int(b[i+directoryEndLen-2]) | int(b[i+directoryEndLen-1])<<8
|
| 307 |
|
|
if n+directoryEndLen+i == len(b) {
|
| 308 |
|
|
return i
|
| 309 |
|
|
}
|
| 310 |
|
|
}
|
| 311 |
|
|
}
|
| 312 |
|
|
return -1
|
| 313 |
|
|
}
|