Commit | Line | Data |
---|---|---|
be60ffd0 | 1 | #!/usr/bin/env python3 |
7584f5c9 ERE |
2 | #------------------------------------------------------------------- |
3 | # tarfile.py | |
4 | #------------------------------------------------------------------- | |
5 | # Copyright (C) 2002 Lars Gustäbel <lars@gustaebel.de> | |
6 | # All rights reserved. | |
7 | # | |
8 | # Permission is hereby granted, free of charge, to any person | |
9 | # obtaining a copy of this software and associated documentation | |
10 | # files (the "Software"), to deal in the Software without | |
11 | # restriction, including without limitation the rights to use, | |
12 | # copy, modify, merge, publish, distribute, sublicense, and/or sell | |
13 | # copies of the Software, and to permit persons to whom the | |
14 | # Software is furnished to do so, subject to the following | |
15 | # conditions: | |
16 | # | |
17 | # The above copyright notice and this permission notice shall be | |
18 | # included in all copies or substantial portions of the Software. | |
19 | # | |
20 | # THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, | |
21 | # EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES | |
22 | # OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND | |
23 | # NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT | |
24 | # HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, | |
25 | # WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING | |
26 | # FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR | |
27 | # OTHER DEALINGS IN THE SOFTWARE. | |
28 | # | |
29 | """Read from and write to tar format archives. | |
30 | """ | |
31 | ||
32 | __version__ = "$Revision: 85213 $" | |
33 | # $Source$ | |
34 | ||
35 | version = "0.9.0" | |
36 | __author__ = "Lars Gustäbel (lars@gustaebel.de)" | |
37 | __date__ = "$Date$" | |
38 | __cvsid__ = "$Id$" | |
5fdff89f | 39 | __credits__ = "Gustavo Niemeyer, Niels Gustäbel, Richard Townsend, Eduardo Robles." |
7584f5c9 ERE |
40 | |
41 | #--------- | |
42 | # Imports | |
43 | #--------- | |
c7c736b6 | 44 | import binascii |
7584f5c9 ERE |
45 | import sys |
46 | import os | |
be60ffd0 | 47 | import io |
7584f5c9 ERE |
48 | import shutil |
49 | import stat | |
50 | import errno | |
51 | import time | |
52 | import struct | |
53 | import copy | |
54 | import re | |
55 | import operator | |
56 | ||
c7c736b6 PG |
57 | import traceback # XXX |
58 | ||
8ab8fac5 | 59 | from . import crypto |
6e812ad9 | 60 | |
7584f5c9 ERE |
61 | try: |
62 | import grp, pwd | |
63 | except ImportError: | |
64 | grp = pwd = None | |
65 | ||
be60ffd0 ERE |
66 | # os.symlink on Windows prior to 6.0 raises NotImplementedError |
67 | symlink_exception = (AttributeError, NotImplementedError) | |
68 | try: | |
69 | # OSError (winerror=1314) will be raised if the caller does not hold the | |
70 | # SeCreateSymbolicLinkPrivilege privilege | |
71 | symlink_exception += (OSError,) | |
72 | except NameError: | |
73 | pass | |
74 | ||
7584f5c9 ERE |
75 | # from tarfile import * |
76 | __all__ = ["TarFile", "TarInfo", "is_tarfile", "TarError"] | |
77 | ||
be60ffd0 ERE |
78 | from builtins import open as _open # Since 'open' is TarFile.open |
79 | ||
7584f5c9 ERE |
80 | #--------------------------------------------------------- |
81 | # tar constants | |
82 | #--------------------------------------------------------- | |
be60ffd0 | 83 | NUL = b"\0" # the null character |
7584f5c9 ERE |
84 | BLOCKSIZE = 512 # length of processing blocks |
85 | RECORDSIZE = BLOCKSIZE * 20 # length of records | |
be60ffd0 ERE |
86 | GNU_MAGIC = b"ustar \0" # magic gnu tar string |
87 | POSIX_MAGIC = b"ustar\x0000" # magic posix tar string | |
7584f5c9 ERE |
88 | |
89 | LENGTH_NAME = 100 # maximum length of a filename | |
90 | LENGTH_LINK = 100 # maximum length of a linkname | |
91 | LENGTH_PREFIX = 155 # maximum length of the prefix field | |
92 | ||
be60ffd0 ERE |
93 | REGTYPE = b"0" # regular file |
94 | AREGTYPE = b"\0" # regular file | |
95 | LNKTYPE = b"1" # link (inside tarfile) | |
96 | SYMTYPE = b"2" # symbolic link | |
97 | CHRTYPE = b"3" # character special device | |
98 | BLKTYPE = b"4" # block special device | |
99 | DIRTYPE = b"5" # directory | |
100 | FIFOTYPE = b"6" # fifo special device | |
101 | CONTTYPE = b"7" # contiguous file | |
102 | ||
103 | GNUTYPE_LONGNAME = b"L" # GNU tar longname | |
104 | GNUTYPE_LONGLINK = b"K" # GNU tar longlink | |
105 | GNUTYPE_SPARSE = b"S" # GNU tar sparse file | |
106 | GNUTYPE_MULTIVOL = b"M" # GNU tar continuation of a file that began on | |
68ddf955 | 107 | # another volume |
7584f5c9 | 108 | |
be60ffd0 ERE |
109 | XHDTYPE = b"x" # POSIX.1-2001 extended header |
110 | XGLTYPE = b"g" # POSIX.1-2001 global header | |
111 | SOLARIS_XHDTYPE = b"X" # Solaris extended header | |
7584f5c9 ERE |
112 | |
113 | USTAR_FORMAT = 0 # POSIX.1-1988 (ustar) format | |
114 | GNU_FORMAT = 1 # GNU tar format | |
115 | PAX_FORMAT = 2 # POSIX.1-2001 (pax) format | |
116 | DEFAULT_FORMAT = GNU_FORMAT | |
117 | ||
15a81fc0 | 118 | GZ_FMT_HEADER = b"<BBBBLBB" |
203cb25e | 119 | GZ_HEADER_SIZE = 10 # not including the name |
15a81fc0 PG |
120 | GZ_MAGIC = (0x1f, 0x8b) # 0o37, 0o213 |
121 | GZ_METHOD_DEFLATE = 0x08 # 0o10 | |
122 | GZ_FLAG_ORIG_NAME = 0x08 # 0o10, default in gzip | |
123 | GZ_DEFLATE_FLAGS = 0x00 # 0o00, never read (deflate.c) | |
124 | GZ_OS_CODE = 0x03 # 0o03, default in gzip (tailor.h) | |
d601d33b PG |
125 | GZ_MAGIC_BYTES = struct.pack ("<BB", GZ_MAGIC [0], GZ_MAGIC [1]) |
126 | GZ_MAGIC_DEFLATE = struct.pack ("<BBB", GZ_MAGIC [0], GZ_MAGIC [1], | |
127 | GZ_METHOD_DEFLATE) | |
15a81fc0 | 128 | |
7584f5c9 | 129 | #--------------------------------------------------------- |
d1c38f40 PG |
130 | # archive handling mode |
131 | #--------------------------------------------------------- | |
132 | ||
133 | ARCMODE_PLAIN = 0 | |
134 | ARCMODE_ENCRYPT = 1 << 0 | |
135 | ARCMODE_COMPRESS = 1 << 1 | |
136 | ARCMODE_CONCAT = 1 << 2 | |
137 | ||
138 | def arcmode_fmt (m): | |
139 | if m == ARCMODE_PLAIN: | |
140 | return "PLAIN" | |
141 | first = True | |
142 | ret = "[" | |
143 | def chkappend (b, s): | |
144 | nonlocal m | |
145 | nonlocal ret | |
146 | nonlocal first | |
147 | if m & b: | |
148 | if first is True: first = False | |
149 | else: ret += " |" | |
150 | ret += " " + s | |
151 | chkappend (ARCMODE_ENCRYPT, "ENCRYPT") | |
152 | chkappend (ARCMODE_COMPRESS, "COMPRESS") | |
153 | chkappend (ARCMODE_CONCAT, "CONCAT") | |
154 | return ret + " ]" | |
155 | ||
156 | ||
157 | def arcmode_set (concat=False, encryption=None, comptype=None, init=ARCMODE_PLAIN): | |
158 | ret = init | |
159 | if bool (concat) is True: | |
160 | ret |= ARCMODE_CONCAT | |
161 | if encryption is not None: | |
162 | ret |= ARCMODE_ENCRYPT | |
163 | if comptype == "gz": | |
164 | ret |= ARCMODE_COMPRESS | |
165 | return ret | |
166 | ||
167 | #--------------------------------------------------------- | |
7584f5c9 ERE |
168 | # tarfile constants |
169 | #--------------------------------------------------------- | |
170 | # File types that tarfile supports: | |
171 | SUPPORTED_TYPES = (REGTYPE, AREGTYPE, LNKTYPE, | |
172 | SYMTYPE, DIRTYPE, FIFOTYPE, | |
173 | CONTTYPE, CHRTYPE, BLKTYPE, | |
174 | GNUTYPE_LONGNAME, GNUTYPE_LONGLINK, | |
68ddf955 | 175 | GNUTYPE_SPARSE, GNUTYPE_MULTIVOL) |
7584f5c9 ERE |
176 | |
177 | # File types that will be treated as a regular file. | |
178 | REGULAR_TYPES = (REGTYPE, AREGTYPE, | |
179 | CONTTYPE, GNUTYPE_SPARSE) | |
180 | ||
181 | # File types that are part of the GNU tar format. | |
182 | GNU_TYPES = (GNUTYPE_LONGNAME, GNUTYPE_LONGLINK, | |
68ddf955 | 183 | GNUTYPE_SPARSE, GNUTYPE_MULTIVOL) |
7584f5c9 ERE |
184 | |
185 | # Fields from a pax header that override a TarInfo attribute. | |
186 | PAX_FIELDS = ("path", "linkpath", "size", "mtime", | |
187 | "uid", "gid", "uname", "gname") | |
188 | ||
be60ffd0 ERE |
189 | # Fields from a pax header that are affected by hdrcharset. |
190 | PAX_NAME_FIELDS = {"path", "linkpath", "uname", "gname"} | |
191 | ||
7584f5c9 ERE |
192 | # Fields in a pax header that are numbers, all other fields |
193 | # are treated as strings. | |
194 | PAX_NUMBER_FIELDS = { | |
195 | "atime": float, | |
196 | "ctime": float, | |
197 | "mtime": float, | |
198 | "uid": int, | |
199 | "gid": int, | |
200 | "size": int | |
201 | } | |
202 | ||
203 | #--------------------------------------------------------- | |
7584f5c9 ERE |
204 | # initialization |
205 | #--------------------------------------------------------- | |
be60ffd0 ERE |
206 | |
207 | if os.name in ("nt", "ce"): | |
208 | ENCODING = "utf-8" | |
209 | else: | |
210 | ENCODING = sys.getfilesystemencoding() | |
7584f5c9 ERE |
211 | |
212 | #--------------------------------------------------------- | |
213 | # Some useful functions | |
214 | #--------------------------------------------------------- | |
215 | ||
be60ffd0 ERE |
216 | def stn(s, length, encoding, errors): |
217 | """Convert a string to a null-terminated bytes object. | |
7584f5c9 | 218 | """ |
be60ffd0 | 219 | s = s.encode(encoding, errors) |
7584f5c9 ERE |
220 | return s[:length] + (length - len(s)) * NUL |
221 | ||
be60ffd0 ERE |
222 | def nts(s, encoding, errors): |
223 | """Convert a null-terminated bytes object to a string. | |
7584f5c9 | 224 | """ |
be60ffd0 ERE |
225 | p = s.find(b"\0") |
226 | if p != -1: | |
227 | s = s[:p] | |
228 | return s.decode(encoding, errors) | |
229 | ||
230 | def sbtn(s, length, encoding, errors): | |
231 | """Convert a string or a bunch of bytes to a null-terminated bytes object | |
232 | of specific size. | |
233 | """ | |
234 | if isinstance(s, str): | |
235 | s = s.encode(encoding, errors) | |
236 | return s[:length] + (length - len(s)) * NUL | |
7584f5c9 ERE |
237 | |
238 | def nti(s): | |
239 | """Convert a number field to a python number. | |
240 | """ | |
241 | # There are two possible encodings for a number field, see | |
242 | # itn() below. | |
be60ffd0 ERE |
243 | if s[0] in (0o200, 0o377): |
244 | n = 0 | |
245 | for i in range(len(s) - 1): | |
246 | n <<= 8 | |
247 | n += s[i + 1] | |
248 | if s[0] == 0o377: | |
249 | n = -(256 ** (len(s) - 1) - n) | |
250 | else: | |
7584f5c9 | 251 | try: |
be60ffd0 | 252 | n = int(nts(s, "ascii", "strict") or "0", 8) |
7584f5c9 ERE |
253 | except ValueError: |
254 | raise InvalidHeaderError("invalid header") | |
7584f5c9 ERE |
255 | return n |
256 | ||
257 | def itn(n, digits=8, format=DEFAULT_FORMAT): | |
258 | """Convert a python number to a number field. | |
259 | """ | |
260 | # POSIX 1003.1-1988 requires numbers to be encoded as a string of | |
261 | # octal digits followed by a null-byte, this allows values up to | |
262 | # (8**(digits-1))-1. GNU tar allows storing numbers greater than | |
be60ffd0 ERE |
263 | # that if necessary. A leading 0o200 or 0o377 byte indicate this |
264 | # particular encoding, the following digits-1 bytes are a big-endian | |
265 | # base-256 representation. This allows values up to (256**(digits-1))-1. | |
266 | # A 0o200 byte indicates a positive number, a 0o377 byte a negative | |
267 | # number. | |
7584f5c9 | 268 | if 0 <= n < 8 ** (digits - 1): |
8112b0ed | 269 | s = bytes("%0*o" % (digits - 1, int(n)), "ascii") + NUL |
be60ffd0 ERE |
270 | elif format == GNU_FORMAT and -256 ** (digits - 1) <= n < 256 ** (digits - 1): |
271 | if n >= 0: | |
272 | s = bytearray([0o200]) | |
273 | else: | |
274 | s = bytearray([0o377]) | |
275 | n = 256 ** digits + n | |
7584f5c9 | 276 | |
be60ffd0 ERE |
277 | for i in range(digits - 1): |
278 | s.insert(1, n & 0o377) | |
7584f5c9 | 279 | n >>= 8 |
7584f5c9 | 280 | else: |
be60ffd0 ERE |
281 | raise ValueError("overflow in number field") |
282 | ||
283 | return s | |
7584f5c9 ERE |
284 | |
285 | def calc_chksums(buf): | |
286 | """Calculate the checksum for a member's header by summing up all | |
287 | characters except for the chksum field which is treated as if | |
288 | it was filled with spaces. According to the GNU tar sources, | |
289 | some tars (Sun and NeXT) calculate chksum with signed char, | |
290 | which will be different if there are chars in the buffer with | |
291 | the high bit set. So we calculate two checksums, unsigned and | |
292 | signed. | |
293 | """ | |
be60ffd0 ERE |
294 | unsigned_chksum = 256 + sum(struct.unpack_from("148B8x356B", buf)) |
295 | signed_chksum = 256 + sum(struct.unpack_from("148b8x356b", buf)) | |
7584f5c9 ERE |
296 | return unsigned_chksum, signed_chksum |
297 | ||
298 | def copyfileobj(src, dst, length=None): | |
299 | """Copy length bytes from fileobj src to fileobj dst. | |
300 | If length is None, copy the entire content. | |
301 | """ | |
302 | if length == 0: | |
303 | return | |
304 | if length is None: | |
305 | shutil.copyfileobj(src, dst) | |
306 | return | |
307 | ||
308 | BUFSIZE = 16 * 1024 | |
309 | blocks, remainder = divmod(length, BUFSIZE) | |
be60ffd0 | 310 | for b in range(blocks): |
7584f5c9 | 311 | buf = src.read(BUFSIZE) |
c474439c | 312 | dst.write(buf) |
7584f5c9 | 313 | if len(buf) < BUFSIZE: |
be60ffd0 | 314 | raise OSError("end of file reached") |
7584f5c9 ERE |
315 | if remainder != 0: |
316 | buf = src.read(remainder) | |
c474439c | 317 | dst.write(buf) |
7584f5c9 | 318 | if len(buf) < remainder: |
be60ffd0 | 319 | raise OSError("end of file reached") |
c7c736b6 | 320 | |
7584f5c9 | 321 | |
7584f5c9 | 322 | def filemode(mode): |
be60ffd0 ERE |
323 | """Deprecated in this location; use stat.filemode.""" |
324 | import warnings | |
325 | warnings.warn("deprecated in favor of stat.filemode", | |
326 | DeprecationWarning, 2) | |
327 | return stat.filemode(mode) | |
7584f5c9 ERE |
328 | |
329 | class TarError(Exception): | |
330 | """Base exception.""" | |
331 | pass | |
332 | class ExtractError(TarError): | |
333 | """General exception for extract errors.""" | |
334 | pass | |
335 | class ReadError(TarError): | |
be60ffd0 | 336 | """Exception for unreadable tar archives.""" |
7584f5c9 ERE |
337 | pass |
338 | class CompressionError(TarError): | |
339 | """Exception for unavailable compression methods.""" | |
340 | pass | |
341 | class StreamError(TarError): | |
342 | """Exception for unsupported operations on stream-like TarFiles.""" | |
343 | pass | |
344 | class HeaderError(TarError): | |
345 | """Base exception for header errors.""" | |
346 | pass | |
347 | class EmptyHeaderError(HeaderError): | |
348 | """Exception for empty headers.""" | |
349 | pass | |
350 | class TruncatedHeaderError(HeaderError): | |
351 | """Exception for truncated headers.""" | |
352 | pass | |
353 | class EOFHeaderError(HeaderError): | |
354 | """Exception for end of file headers.""" | |
355 | pass | |
356 | class InvalidHeaderError(HeaderError): | |
357 | """Exception for invalid headers.""" | |
358 | pass | |
359 | class SubsequentHeaderError(HeaderError): | |
360 | """Exception for missing and invalid extended headers.""" | |
361 | pass | |
8ab8fac5 PG |
362 | class InvalidEncryptionError(TarError): |
363 | """Exception for undefined crypto modes and combinations.""" | |
364 | pass | |
e4e5d0b8 PG |
365 | class DecryptionError(TarError): |
366 | """Exception for error during decryption.""" | |
367 | pass | |
c7c736b6 | 368 | class EncryptionError(TarError): |
e93f83f1 | 369 | """Exception for error during encryption.""" |
c7c736b6 | 370 | pass |
e50fa574 PG |
371 | class EndOfFile(Exception): |
372 | """Signal end of file condition when they’re not an error.""" | |
7584f5c9 ERE |
373 | |
374 | #--------------------------- | |
375 | # internal stream interface | |
376 | #--------------------------- | |
377 | class _LowLevelFile: | |
378 | """Low-level file object. Supports reading and writing. | |
379 | It is used instead of a regular file object for streaming | |
380 | access. | |
381 | """ | |
382 | ||
383 | def __init__(self, name, mode): | |
ad4402e8 | 384 | _mode = { |
7584f5c9 | 385 | "r": os.O_RDONLY, |
c7c736b6 | 386 | "w": os.O_RDWR | os.O_CREAT | os.O_TRUNC, |
7584f5c9 ERE |
387 | }[mode] |
388 | if hasattr(os, "O_BINARY"): | |
f0287fb7 | 389 | _mode |= os.O_BINARY # pylint: disable=no-member |
be60ffd0 | 390 | self.fd = os.open(name, _mode, 0o666) |
ad4402e8 | 391 | self.offset = 0 |
7584f5c9 ERE |
392 | |
393 | def close(self): | |
394 | os.close(self.fd) | |
395 | ||
396 | def read(self, size): | |
ad4402e8 ERE |
397 | ret = os.read(self.fd, size) |
398 | self.offset += len(ret) | |
399 | return ret | |
7584f5c9 | 400 | |
867f75f7 PG |
401 | def write(self, s, pos=None): |
402 | if pos is not None: | |
403 | p0 = self.offset | |
404 | os.lseek (self.fd, pos, os.SEEK_SET) | |
405 | n = os.write(self.fd, s) | |
406 | if pos is None: | |
407 | self.offset += len(s) | |
408 | else: | |
409 | append = pos + n - p0 | |
410 | if append > 0: | |
411 | self.offset += append | |
412 | os.lseek (self.fd, p0, os.SEEK_SET) | |
7584f5c9 | 413 | |
ad4402e8 ERE |
414 | def tell(self): |
415 | return self.offset | |
416 | ||
c7c736b6 PG |
417 | def seek_set (self, pos): |
418 | os.lseek (self.fd, pos, os.SEEK_SET) | |
419 | self.offset = pos | |
420 | ||
8ab8fac5 | 421 | |
15a81fc0 PG |
422 | def gz_header (name=None): |
423 | timestamp = int(time.time()) | |
424 | flags = 0x0 | |
425 | ||
426 | if name is None: | |
427 | name = b"" | |
428 | else: | |
429 | flags |= GZ_FLAG_ORIG_NAME | |
430 | if type(name) is str: | |
431 | name = name.encode("iso-8859-1", "replace") | |
6e99d23a PG |
432 | if name.endswith(b".pdtcrypt"): |
433 | name = name[:-9] | |
15a81fc0 PG |
434 | if name.endswith(b".gz"): |
435 | name = name[:-3] | |
436 | # RFC1952 says we must use ISO-8859-1 for the FNAME field. | |
437 | name += NUL | |
438 | ||
439 | hdr = struct.pack (GZ_FMT_HEADER, | |
440 | GZ_MAGIC [0], GZ_MAGIC [1], | |
441 | GZ_METHOD_DEFLATE, flags, | |
442 | timestamp, | |
443 | GZ_DEFLATE_FLAGS, GZ_OS_CODE) | |
444 | ||
445 | return hdr + name | |
446 | ||
d601d33b | 447 | |
7584f5c9 ERE |
448 | class _Stream: |
449 | """Class that serves as an adapter between TarFile and | |
450 | a stream-like object. The stream-like object only | |
451 | needs to have a read() or write() method and is accessed | |
452 | blockwise. Use of gzip or bzip2 compression is possible. | |
453 | A stream-like object could be for example: sys.stdin, | |
454 | sys.stdout, a socket, a tape device etc. | |
455 | ||
3031b7ae PG |
456 | _Stream is intended to be used only internally but is |
457 | nevertherless used externally by Deltatar. | |
458 | ||
459 | When encrypting, the ``enccounter`` will be used for | |
460 | initializing the first cryptographic context. When | |
461 | decrypting, its value will be compared to the decrypted | |
462 | object. Decryption fails if the value does not match. | |
463 | In effect, this means that a ``_Stream`` whose ctor was | |
464 | passed ``enccounter`` can only be used to encrypt or | |
465 | decrypt a single object. | |
7584f5c9 ERE |
466 | """ |
467 | ||
c7c736b6 | 468 | remainder = -1 # track size in encrypted entries |
e93f83f1 | 469 | tolerant = False |
c7c736b6 | 470 | |
6e812ad9 | 471 | def __init__(self, name, mode, comptype, fileobj, bufsize, |
d1c38f40 | 472 | concat=False, encryption=None, enccounter=None, |
e93f83f1 | 473 | compresslevel=9, tolerant=False): |
7584f5c9 ERE |
474 | """Construct a _Stream object. |
475 | """ | |
d1c38f40 | 476 | self.arcmode = arcmode_set (concat, encryption, comptype) |
e93f83f1 | 477 | self.tolerant = tolerant |
d1c38f40 | 478 | |
7584f5c9 ERE |
479 | self._extfileobj = True |
480 | if fileobj is None: | |
481 | fileobj = _LowLevelFile(name, mode) | |
482 | self._extfileobj = False | |
483 | ||
484 | if comptype == '*': | |
485 | # Enable transparent compression detection for the | |
486 | # stream interface | |
487 | fileobj = _StreamProxy(fileobj) | |
488 | comptype = fileobj.getcomptype() | |
d1c38f40 PG |
489 | if comptype == '': |
490 | comptype = "tar" | |
7584f5c9 | 491 | |
3031b7ae PG |
492 | self.enccounter = None |
493 | if self.arcmode & ARCMODE_ENCRYPT: | |
494 | self.enccounter = enccounter | |
495 | ||
7584f5c9 ERE |
496 | self.name = name or "" |
497 | self.mode = mode | |
498 | self.comptype = comptype | |
53732900 | 499 | self.cmp = None |
7584f5c9 ERE |
500 | self.fileobj = fileobj |
501 | self.bufsize = bufsize | |
be60ffd0 ERE |
502 | self.buf = b"" |
503 | self.pos = 0 | |
504 | self.concat_pos = 0 | |
7584f5c9 | 505 | self.closed = False |
be60ffd0 | 506 | self.flags = 0 |
be60ffd0 | 507 | self.last_block_offset = 0 |
e4e5d0b8 | 508 | self.dbuf = b"" # ??? |
46c03c02 | 509 | self.exception = None # communicate decompression failure |
2b82f50c | 510 | self.compresslevel = compresslevel |
784175ba | 511 | self.bytes_written = 0 |
c7c736b6 | 512 | # crypto parameters |
2ae46844 | 513 | self.encryption = encryption |
c7c736b6 | 514 | self.lasthdr = None |
7584f5c9 | 515 | |
be60ffd0 ERE |
516 | try: |
517 | if comptype == "gz": | |
518 | try: | |
519 | import zlib | |
520 | except ImportError: | |
521 | raise CompressionError("zlib module is not available") | |
522 | self.zlib = zlib | |
bec34b42 PG |
523 | if mode == "r": |
524 | self.exception = zlib.error | |
8ae983c4 | 525 | self._init_read_gz() |
bec34b42 | 526 | elif mode == "w": |
d1c38f40 PG |
527 | if not (self.arcmode & ARCMODE_CONCAT): |
528 | if self.arcmode & ARCMODE_ENCRYPT: | |
3031b7ae | 529 | self._init_write_encrypt (name) |
a0873dcc | 530 | self._init_write_gz () |
c2ffe2ec | 531 | self.crc = zlib.crc32(b"") & 0xFFFFffff |
7584f5c9 | 532 | |
be60ffd0 | 533 | elif comptype == "bz2": |
d1c38f40 | 534 | if self.arcmode & ARCMODE_ENCRYPT: |
2ae46844 | 535 | raise InvalidEncryptionError("encryption not available for " |
d1c38f40 | 536 | "compression “%s”" % comptype) |
be60ffd0 ERE |
537 | try: |
538 | import bz2 | |
539 | except ImportError: | |
540 | raise CompressionError("bz2 module is not available") | |
541 | if mode == "r": | |
542 | self.dbuf = b"" | |
543 | self.cmp = bz2.BZ2Decompressor() | |
544 | self.exception = OSError | |
545 | else: | |
546 | self.cmp = bz2.BZ2Compressor() | |
7584f5c9 | 547 | |
be60ffd0 | 548 | elif comptype == 'xz': |
d1c38f40 | 549 | if self.arcmode & ARCMODE_ENCRYPT: |
2ae46844 | 550 | raise InvalidEncryptionError("encryption not available for " |
d1c38f40 | 551 | "compression “%s”" % comptype) |
c7c736b6 PG |
552 | try: |
553 | import lzma | |
554 | except ImportError: | |
555 | raise CompressionError("lzma module is not available") | |
556 | if mode == "r": | |
557 | self.dbuf = b"" | |
558 | self.cmp = lzma.LZMADecompressor() | |
559 | self.exception = lzma.LZMAError | |
560 | else: | |
561 | self.cmp = lzma.LZMACompressor() | |
562 | ||
6de9444a | 563 | elif comptype == "tar": |
d1c38f40 | 564 | if not (self.arcmode & ARCMODE_CONCAT) \ |
6de9444a | 565 | and mode == "w" \ |
d1c38f40 | 566 | and self.arcmode & ARCMODE_ENCRYPT: |
3031b7ae | 567 | self._init_write_encrypt (name) |
6de9444a PG |
568 | |
569 | else: | |
d1c38f40 | 570 | if self.arcmode & ARCMODE_ENCRYPT: |
2ae46844 | 571 | raise InvalidEncryptionError("encryption not available for " |
d1c38f40 | 572 | "compression “%s”" % comptype) |
c7c736b6 | 573 | raise CompressionError("unknown compression type %r" % comptype) |
be60ffd0 | 574 | |
200d4866 | 575 | except: |
be60ffd0 ERE |
576 | if not self._extfileobj: |
577 | self.fileobj.close() | |
578 | self.closed = True | |
579 | raise | |
ac5e4184 | 580 | |
7584f5c9 ERE |
581 | def __del__(self): |
582 | if hasattr(self, "closed") and not self.closed: | |
fac2cfe1 PG |
583 | try: |
584 | self.close() | |
585 | except crypto.InternalError: | |
586 | # context already finalized due to abort but close() tried | |
587 | # to use it | |
588 | pass | |
7584f5c9 | 589 | |
c7c736b6 | 590 | |
d1c38f40 PG |
591 | def next (self, name): |
592 | if self.arcmode & ARCMODE_COMPRESS: | |
593 | if getattr (self, "cmp", None) is not None: | |
594 | self._finalize_write_gz () | |
0349168a PG |
595 | self.__sync() |
596 | if self.arcmode & ~(ARCMODE_ENCRYPT | ARCMODE_COMPRESS): | |
597 | self.last_block_offset = self.fileobj.tell() | |
d1c38f40 PG |
598 | if self.arcmode & ARCMODE_ENCRYPT: |
599 | self._finalize_write_encrypt () | |
600 | self._init_write_encrypt (name, set_last_block_offset=True) | |
601 | if self.arcmode & ARCMODE_COMPRESS: | |
602 | self._init_write_gz (set_last_block_offset = | |
0349168a | 603 | not (self.arcmode & ARCMODE_ENCRYPT)) |
d1c38f40 PG |
604 | return self.last_block_offset |
605 | ||
606 | ||
607 | def next_volume (self, name): | |
608 | # with non-concat modes, this is taken care by the _Stream | |
609 | # ctor as invoked by the newvol handler | |
610 | if self.arcmode & ARCMODE_COMPRESS: | |
611 | if getattr (self, "cmp", None) is not None: | |
612 | # e. g. compressed PAX header written | |
613 | self._finalize_write_gz () | |
614 | if self.arcmode & ARCMODE_ENCRYPT: | |
615 | self._init_write_encrypt (name) | |
616 | if self.arcmode & ARCMODE_COMPRESS: | |
617 | self._init_write_gz () | |
618 | ||
c7c736b6 | 619 | |
d1c38f40 PG |
620 | def _init_write_encrypt (self, entry=None, set_last_block_offset=False): |
621 | """ | |
622 | Save position for delayed write of header; fill the header location | |
623 | with dummy bytes. | |
624 | """ | |
625 | # first thing, proclaim new object to the encryption context | |
626 | # secondly, assemble the header with the updated parameters | |
627 | # and commit it directly to the underlying stream, bypassing the | |
628 | # encryption layer in .__write(). | |
629 | dummyhdr = self.encryption.next (entry, counter=self.enccounter) | |
630 | if dummyhdr is None: | |
631 | raise EncryptionError ("Crypto.next(): bad dummy header") # XXX | |
632 | self.lasthdr = self.fileobj.tell() | |
633 | self.__write_to_file(dummyhdr) | |
634 | if set_last_block_offset is True: | |
635 | self.last_block_offset = self.lasthdr | |
c7c736b6 PG |
636 | |
637 | ||
638 | def _finalize_write_encrypt (self): | |
639 | """ | |
640 | Seek back to header position, read dummy bytes, finalize crypto | |
641 | obtaining the actual header, write header, seek back to current | |
642 | position. | |
963d0db4 PG |
643 | |
644 | Returns the list of IV fixed parts as used during encryption. | |
c7c736b6 | 645 | """ |
d1c38f40 | 646 | if self.lasthdr is not None: |
c7c736b6 PG |
647 | pos0 = self.fileobj.tell () |
648 | self.fileobj.seek_set (self.lasthdr) | |
dd47d6a2 | 649 | dummy = self.fileobj.read (crypto.PDTCRYPT_HDR_SIZE) |
c7c736b6 PG |
650 | pos1 = self.fileobj.tell () |
651 | dpos = pos1 - self.lasthdr | |
dd47d6a2 | 652 | assert dpos == crypto.PDTCRYPT_HDR_SIZE |
c7c736b6 | 653 | self.fileobj.seek_set (pos0) |
c8c72fe1 | 654 | data, hdr, _ = self.encryption.done (dummy) |
5f38bff6 | 655 | self.__write_to_file(hdr, pos=self.lasthdr) |
c7c736b6 PG |
656 | self.__write_to_file(data) # append remainder of data |
657 | self.lasthdr = -1 | |
658 | ||
659 | ||
57db1546 PG |
660 | def _finalize_write_gz (self): |
661 | if self.cmp is not None: | |
662 | chunk = self.buf + self.cmp.flush() | |
663 | if chunk: | |
664 | if self.comptype == "gz": | |
665 | # The native zlib crc is an unsigned 32-bit integer, but | |
666 | # the Python wrapper implicitly casts that to a signed C | |
667 | # long. So, on a 32-bit box self.crc may "look negative", | |
668 | # while the same crc on a 64-bit box may "look positive". | |
669 | # To avoid irksome warnings from the `struct` module, force | |
670 | # it to look positive on all boxes. | |
671 | chunk += struct.pack("<L", self.crc & 0xffffffff) | |
672 | chunk += struct.pack("<L", self.concat_pos & 0xffffFFFF) | |
673 | self.__enc_write (chunk) | |
15a81fc0 | 674 | self.buf = b"" |
57db1546 PG |
675 | |
676 | ||
a0873dcc | 677 | def _init_write_gz (self, set_last_block_offset=False): |
5fdff89f ERE |
678 | ''' |
679 | Add a new gzip block, closing last one | |
680 | ''' | |
be60ffd0 | 681 | self.concat_pos = 0 |
c2ffe2ec | 682 | self.crc = self.zlib.crc32(b"") & 0xFFFFffff |
6de9444a | 683 | first = self.cmp is None |
2b82f50c ERE |
684 | self.cmp = self.zlib.compressobj(self.compresslevel, |
685 | self.zlib.DEFLATED, | |
686 | -self.zlib.MAX_WBITS, | |
687 | self.zlib.DEF_MEM_LEVEL, | |
688 | 0) | |
6e812ad9 DGM |
689 | |
690 | # if aes, we encrypt after compression | |
6de9444a | 691 | if set_last_block_offset is True: |
ad4402e8 | 692 | self.last_block_offset = self.fileobj.tell() |
6e812ad9 | 693 | |
15a81fc0 | 694 | self.__write(gz_header (self.name if first is True else None)) |
5fdff89f | 695 | |
ac5e4184 | 696 | |
7584f5c9 ERE |
697 | def write(self, s): |
698 | """Write string s to the stream. | |
699 | """ | |
700 | if self.comptype == "gz": | |
c2ffe2ec | 701 | self.crc = self.zlib.crc32(s, self.crc) & 0xFFFFffff |
7584f5c9 | 702 | self.pos += len(s) |
5fdff89f | 703 | self.concat_pos += len(s) |
53732900 | 704 | if self.cmp is not None: |
7584f5c9 ERE |
705 | s = self.cmp.compress(s) |
706 | self.__write(s) | |
707 | ||
c7c736b6 | 708 | def __sync(self): |
cb7a3911 | 709 | """Write what’s left in the buffer to the stream.""" |
c7c736b6 PG |
710 | self.__write (b"") # → len (buf) <= bufsiz |
711 | self.__enc_write (self.buf) | |
712 | self.buf = b"" | |
713 | ||
7584f5c9 | 714 | def __write(self, s): |
548bb8d5 CH |
715 | """Writes (and encodes) string s to the stream blockwise |
716 | ||
717 | will wait with encoding/writing until block is complete | |
7584f5c9 ERE |
718 | """ |
719 | self.buf += s | |
720 | while len(self.buf) > self.bufsize: | |
6e812ad9 | 721 | self.__enc_write(self.buf[:self.bufsize]) |
7584f5c9 ERE |
722 | self.buf = self.buf[self.bufsize:] |
723 | ||
867f75f7 | 724 | |
5f38bff6 | 725 | def __write_to_file(self, s, pos=None): |
6e812ad9 | 726 | ''' |
5f38bff6 | 727 | Writes directly to the fileobj; updates self.bytes_written. If “pos” is |
cb7a3911 | 728 | given, the stream will seek to that position first and back afterwards, |
5f38bff6 | 729 | and the total of bytes written is not updated. |
6e812ad9 | 730 | ''' |
867f75f7 | 731 | self.fileobj.write(s, pos) |
5f38bff6 PG |
732 | if pos is None: |
733 | self.bytes_written += len(s) | |
867f75f7 | 734 | |
6e812ad9 DGM |
735 | |
736 | def __enc_write(self, s): | |
cb7a3911 PG |
737 | """ |
738 | If encryption is active, the string s is encrypted before being written | |
739 | to the file. | |
740 | """ | |
741 | if len (s) == 0: | |
742 | return | |
d1c38f40 | 743 | if self.arcmode & ARCMODE_ENCRYPT: |
cb7a3911 PG |
744 | buf = s |
745 | while len (buf) > 0: | |
746 | n, ct = self.encryption.process(buf) | |
747 | self.__write_to_file(ct) | |
748 | buf = buf [n:] | |
749 | if len (buf) > 0: | |
750 | # The entire plaintext was not consumed: The size limit | |
751 | # for encrypted objects was reached. Transparently create | |
752 | # a new encrypted object and continue processing the input. | |
753 | self._finalize_write_encrypt () | |
754 | self._init_write_encrypt () | |
755 | else: | |
756 | self.__write_to_file(s) | |
757 | ||
6e812ad9 | 758 | |
784175ba CH |
759 | def estim_file_size(self): |
760 | """ estimates size of file if closing it now | |
761 | ||
762 | The result may differ greatly from the amount of data sent to write() | |
763 | due to compression, encryption and buffering. | |
764 | ||
765 | In tests the result (before calling close()) was up to 12k smaller than | |
766 | the final file size if compression is being used because zlib/bz2 | |
767 | compressors do not allow inspection of their buffered data :-( | |
768 | ||
ba5a449e CH |
769 | Still, we add what close() would add: 8 bytes for gz checksum, one |
770 | encryption block size if encryption is used and the size of our own | |
771 | buffer | |
784175ba CH |
772 | """ |
773 | if self.closed: | |
774 | return self.bytes_written | |
775 | ||
776 | result = self.bytes_written | |
777 | if self.buf: | |
778 | result += len(self.buf) | |
779 | if self.comptype == 'gz': | |
ba5a449e | 780 | result += 8 # 2 longs = 8 byte (no extra info written for bzip2) |
784175ba CH |
781 | return result |
782 | ||
5fdff89f | 783 | def close(self, close_fileobj=True): |
7584f5c9 ERE |
784 | """Close the _Stream object. No operation should be |
785 | done on it afterwards. | |
786 | """ | |
963d0db4 | 787 | |
7584f5c9 ERE |
788 | if self.closed: |
789 | return | |
790 | ||
963d0db4 | 791 | if close_fileobj is True: |
a0873dcc | 792 | |
ae3d0f2a | 793 | if self.mode == "w": |
d1c38f40 | 794 | if self.arcmode & ARCMODE_COMPRESS: |
a0873dcc | 795 | self._finalize_write_gz () |
ae3d0f2a | 796 | # end of Tar archive marker (two empty blocks) was written |
267bc643 PG |
797 | # finalize encryption last; no writes may be performed after |
798 | # this point | |
cb7a3911 | 799 | self.__sync () |
d1c38f40 PG |
800 | if self.arcmode & ARCMODE_ENCRYPT: |
801 | self._finalize_write_encrypt () | |
267bc643 | 802 | |
963d0db4 PG |
803 | if not self._extfileobj: |
804 | self.fileobj.close() | |
805 | else: | |
806 | # read the zlib crc and length and check them | |
807 | if self.mode == "r" and self.comptype == "gz": | |
808 | read_crc = self.__read(4) | |
809 | read_length = self.__read(4) | |
810 | calculated_crc = self.crc | |
811 | if struct.unpack("<L", read_crc)[0] != calculated_crc: | |
812 | raise CompressionError("bad gzip crc") | |
7584f5c9 ERE |
813 | self.closed = True |
814 | ||
54128a00 | 815 | |
7584f5c9 ERE |
816 | def _init_read_gz(self): |
817 | """Initialize for reading a gzip compressed fileobj. | |
818 | """ | |
819 | self.cmp = self.zlib.decompressobj(-self.zlib.MAX_WBITS) | |
7584f5c9 | 820 | |
85737f48 | 821 | read2 = self.__read(2) |
e50fa574 PG |
822 | if read2 == b"": |
823 | raise EndOfFile ("_init_read_gz(): read returned zero bytes at pos " | |
824 | "%d" % self.fileobj.tell()) | |
825 | # taken from gzip.GzipFile with some alterations | |
d601d33b | 826 | if read2 != GZ_MAGIC_BYTES: |
7584f5c9 | 827 | raise ReadError("not a gzip file") |
85737f48 ERE |
828 | |
829 | read1 = self.__read(1) | |
be60ffd0 | 830 | if read1 != b"\010": |
7584f5c9 ERE |
831 | raise CompressionError("unsupported compression method") |
832 | ||
85737f48 | 833 | self.flags = flag = ord(self.__read(1)) |
7584f5c9 ERE |
834 | self.__read(6) |
835 | ||
836 | if flag & 4: | |
837 | xlen = ord(self.__read(1)) + 256 * ord(self.__read(1)) | |
838 | self.read(xlen) | |
839 | if flag & 8: | |
840 | while True: | |
841 | s = self.__read(1) | |
842 | if not s or s == NUL: | |
843 | break | |
844 | if flag & 16: | |
845 | while True: | |
846 | s = self.__read(1) | |
847 | if not s or s == NUL: | |
848 | break | |
849 | if flag & 2: | |
850 | self.__read(2) | |
851 | ||
c7c736b6 PG |
852 | def _init_read_encrypt (self): |
853 | """Initialize encryption for next entry in archive. Read a header and | |
854 | notify the crypto context.""" | |
d1c38f40 | 855 | if self.arcmode & ARCMODE_ENCRYPT: |
6e99d23a | 856 | lasthdr = self.fileobj.tell () |
15d3eefd PG |
857 | try: |
858 | hdr = crypto.hdr_read_stream (self.fileobj) | |
8a8ac469 PG |
859 | except crypto.EndOfFile: |
860 | return False | |
6e99d23a | 861 | except crypto.InvalidHeader as exn: |
c7c736b6 | 862 | raise DecryptionError ("Crypto.hdr_read_stream(): error “%s” " |
6e99d23a PG |
863 | "processing %r at pos %d" |
864 | % (exn, self.fileobj, lasthdr)) \ | |
ae3d0f2a | 865 | from exn |
3031b7ae PG |
866 | if self.enccounter is not None: |
867 | # enforce that the iv counter in the header matches an | |
868 | # explicitly requested one | |
869 | iv = crypto.hdr_iv_counter (hdr) | |
870 | if iv != self.enccounter: | |
871 | raise DecryptionError ("expected IV counter %d, got %d" | |
872 | % (self.enccounter, iv)) | |
6e99d23a | 873 | self.lasthdr = lasthdr |
c7c736b6 | 874 | self.remainder = hdr ["ctsize"] # distance to next header |
1ed44e7b PG |
875 | try: |
876 | self.encryption.next (hdr) | |
877 | except crypto.InvalidParameter as exn: | |
878 | raise DecryptionError ("Crypto.next(): error “%s” " | |
879 | "processing %r at pos %d" | |
880 | % (exn, self.fileobj, lasthdr)) \ | |
881 | from exn | |
8a8ac469 PG |
882 | |
883 | return True | |
c7c736b6 PG |
884 | |
885 | ||
8de91f4f PG |
886 | def _read_encrypt (self, buf): |
887 | """ | |
888 | Demote a program error to a decryption error in tolerant mode. This | |
889 | allows recovery from corrupted headers and invalid data. | |
890 | """ | |
891 | try: | |
892 | return self.encryption.process (buf) | |
893 | except RuntimeError as exn: | |
894 | if self.tolerant is True: | |
895 | raise DecryptionError (exn) | |
896 | raise | |
897 | ||
898 | ||
c7c736b6 PG |
899 | def _finalize_read_encrypt (self): |
900 | """ | |
901 | Finalize decryption. | |
902 | """ | |
d1c38f40 PG |
903 | if self.arcmode & ARCMODE_ENCRYPT \ |
904 | and self.lasthdr is not None : | |
c7c736b6 PG |
905 | assert self.remainder >= 0 |
906 | if self.remainder > 0: | |
c7c736b6 | 907 | self.remainder = 0 |
b0078f26 PG |
908 | try: |
909 | data = self.encryption.done () | |
910 | except crypto.InvalidGCMTag as exn: | |
911 | raise DecryptionError ("decryption failed: %s" % exn) | |
c7c736b6 PG |
912 | return data |
913 | ||
914 | ||
7584f5c9 ERE |
915 | def tell(self): |
916 | """Return the stream's file pointer position. | |
917 | """ | |
918 | return self.pos | |
919 | ||
920 | def seek(self, pos=0): | |
921 | """Set the stream's file pointer to pos. Negative seeking | |
922 | is forbidden. | |
923 | """ | |
924 | if pos - self.pos >= 0: | |
925 | blocks, remainder = divmod(pos - self.pos, self.bufsize) | |
be60ffd0 | 926 | for i in range(blocks): |
7584f5c9 ERE |
927 | self.read(self.bufsize) |
928 | self.read(remainder) | |
929 | else: | |
930 | raise StreamError("seeking backwards is not allowed") | |
931 | return self.pos | |
932 | ||
933 | def read(self, size=None): | |
934 | """Return the next size number of bytes from the stream. | |
935 | If size is not defined, return all bytes of the stream | |
936 | up to EOF. | |
937 | """ | |
938 | if size is None: | |
939 | t = [] | |
940 | while True: | |
941 | buf = self._read(self.bufsize) | |
942 | if not buf: | |
943 | break | |
944 | t.append(buf) | |
9dc7ac5c | 945 | buf = b"".join(t) |
7584f5c9 ERE |
946 | else: |
947 | buf = self._read(size) | |
948 | self.pos += len(buf) | |
949 | return buf | |
950 | ||
3a7e1a50 ERE |
951 | def readline(self): |
952 | """Reads just one line, new line character included | |
953 | """ | |
f0fd5e3a | 954 | # if \n in dbuf, no read neads to be done |
be60ffd0 ERE |
955 | if b'\n' in self.dbuf: |
956 | pos = self.dbuf.index(b'\n') + 1 | |
f0fd5e3a ERE |
957 | ret = self.dbuf[:pos] |
958 | self.dbuf = self.dbuf[pos:] | |
959 | return ret | |
960 | ||
1215b602 | 961 | buf = [] |
3a7e1a50 ERE |
962 | while True: |
963 | chunk = self._read(self.bufsize) | |
964 | ||
f0fd5e3a | 965 | # nothing more to read, so return the buffer |
3a7e1a50 | 966 | if not chunk: |
be60ffd0 | 967 | return b''.join(buf) |
3a7e1a50 ERE |
968 | |
969 | buf.append(chunk) | |
f0fd5e3a ERE |
970 | |
971 | # if \n found, return the new line | |
be60ffd0 ERE |
972 | if b'\n' in chunk: |
973 | dbuf = b''.join(buf) | |
974 | pos = dbuf.index(b'\n') + 1 | |
1215b602 | 975 | self.dbuf = dbuf[pos:] + self.dbuf |
3a7e1a50 ERE |
976 | return dbuf[:pos] |
977 | ||
7584f5c9 ERE |
978 | def _read(self, size): |
979 | """Return size bytes from the stream. | |
980 | """ | |
7584f5c9 ERE |
981 | c = len(self.dbuf) |
982 | t = [self.dbuf] | |
e4e5d0b8 | 983 | |
7584f5c9 | 984 | while c < size: |
867f75f7 | 985 | buf = self.__read(self.bufsize) |
7584f5c9 ERE |
986 | if not buf: |
987 | break | |
3a7e1a50 | 988 | |
53732900 | 989 | if self.cmp is not None: |
85737f48 | 990 | try: |
3a7e1a50 | 991 | buf = self.cmp.decompress(buf) |
54128a00 PG |
992 | except self.exception as exn: |
993 | raise ReadError("invalid compressed data (%r)" % exn) | |
be60ffd0 | 994 | except Exception as e: |
04fb06f4 DGM |
995 | # happens at the end of the file |
996 | # _init_read_gz failed in the previous iteration so | |
e4e5d0b8 | 997 | # self.cmp.decompress fails here |
d1c38f40 | 998 | if self.arcmode & ARCMODE_CONCAT: |
be60ffd0 ERE |
999 | pass |
1000 | else: | |
1001 | raise ReadError("invalid compressed data") | |
d1c38f40 | 1002 | if self.arcmode & ARCMODE_COMPRESS and hasattr(self, "crc"): |
c2ffe2ec | 1003 | self.crc = self.zlib.crc32(buf, self.crc) & 0xFFFFffff |
d1c38f40 PG |
1004 | if self.arcmode & ARCMODE_CONCAT \ |
1005 | and len(self.cmp.unused_data) != 0: | |
3a7e1a50 ERE |
1006 | self.buf = self.cmp.unused_data + self.buf |
1007 | self.close(close_fileobj=False) | |
1008 | try: | |
1009 | self._init_read_gz() | |
8de91f4f PG |
1010 | except DecryptionError: |
1011 | if self.tolerant is True: | |
1012 | # return whatever data was processed successfully | |
1013 | if len (buf) > 0: | |
1014 | t.append (buf) | |
1015 | if len (t) > 0: | |
1016 | break | |
1017 | raise | |
e50fa574 | 1018 | except EndOfFile: |
3a7e1a50 ERE |
1019 | # happens at the end of the file |
1020 | pass | |
c2ffe2ec | 1021 | self.crc = self.zlib.crc32(b"") & 0xFFFFffff |
3a7e1a50 | 1022 | self.closed = False |
7584f5c9 ERE |
1023 | t.append(buf) |
1024 | c += len(buf) | |
be60ffd0 | 1025 | t = b"".join(t) |
7584f5c9 ERE |
1026 | self.dbuf = t[size:] |
1027 | return t[:size] | |
1028 | ||
e4e5d0b8 | 1029 | |
7584f5c9 | 1030 | def __read(self, size): |
ef3b4499 PG |
1031 | """ |
1032 | Return size bytes from stream. If internal buffer is empty, read | |
1033 | another block from the stream. | |
1034 | ||
1035 | The function returns up to size bytes of data. When an error occurs | |
1036 | during decryption, everything until the end of the last successfully | |
1037 | finalized object is returned. | |
7584f5c9 ERE |
1038 | """ |
1039 | c = len(self.buf) | |
8de91f4f | 1040 | t = [self.buf] if c > 0 else [] |
1ed44e7b | 1041 | good_crypto = len (t) |
8de91f4f | 1042 | |
7584f5c9 | 1043 | while c < size: |
c7c736b6 | 1044 | todo = size |
8de91f4f PG |
1045 | try: |
1046 | if self.arcmode & ARCMODE_ENCRYPT: | |
1047 | if self.remainder <= 0: | |
1048 | # prepare next object | |
044585c6 PG |
1049 | if self._init_read_encrypt () is False: # EOF |
1050 | buf = None | |
1051 | break # while | |
8de91f4f PG |
1052 | |
1053 | # only read up to the end of the encrypted object | |
1054 | todo = min (size, self.remainder) | |
1055 | buf = self.fileobj.read(todo) | |
1056 | if self.arcmode & ARCMODE_ENCRYPT: | |
1057 | # decrypt the thing | |
1058 | buf = self._read_encrypt (buf) | |
1059 | if todo == self.remainder: | |
1060 | # at the end of a crypto object; finalization will fail if | |
1061 | # the GCM tag does not match | |
ef3b4499 | 1062 | trailing = self._finalize_read_encrypt () |
8de91f4f PG |
1063 | good_crypto = len (t) + 1 |
1064 | if len (trailing) > 0: | |
1065 | buf += trailing | |
1066 | self.remainder = 0 | |
1067 | else: | |
1068 | self.remainder -= todo | |
1069 | except DecryptionError: | |
1070 | if self.tolerant is False: | |
1071 | raise | |
1072 | self.encryption.drop () | |
1073 | if good_crypto == 0: | |
1074 | raise | |
1075 | # this may occur at any of the three crypto operations above. | |
1076 | # some objects did validate; discard all data after it; next | |
1077 | # call will start with the bad object and error out immediately | |
1078 | self.buf = b"".join (t [good_crypto:]) | |
1079 | return b"".join (t [:good_crypto]) | |
c7c736b6 PG |
1080 | |
1081 | if not buf: ## XXX stream terminated prematurely; this should be an error | |
7584f5c9 | 1082 | break |
c7c736b6 | 1083 | |
7584f5c9 ERE |
1084 | t.append(buf) |
1085 | c += len(buf) | |
be60ffd0 | 1086 | t = b"".join(t) |
7584f5c9 | 1087 | self.buf = t[size:] |
fb27c6e8 | 1088 | |
7584f5c9 | 1089 | return t[:size] |
7d372216 | 1090 | |
7584f5c9 ERE |
1091 | |
1092 | class _StreamProxy(object): | |
1093 | """Small proxy class that enables transparent compression | |
1094 | detection for the Stream interface (mode 'r|*'). | |
1095 | """ | |
1096 | ||
1097 | def __init__(self, fileobj): | |
1098 | self.fileobj = fileobj | |
1099 | self.buf = self.fileobj.read(BLOCKSIZE) | |
1100 | ||
f0287fb7 | 1101 | def read(self, size): # pylint: disable=method-hidden |
7584f5c9 ERE |
1102 | self.read = self.fileobj.read |
1103 | return self.buf | |
1104 | ||
1105 | def getcomptype(self): | |
d601d33b | 1106 | if self.buf.startswith(GZ_MAGIC_DEFLATE): |
7584f5c9 | 1107 | return "gz" |
be60ffd0 | 1108 | elif self.buf[0:3] == b"BZh" and self.buf[4:10] == b"1AY&SY": |
7584f5c9 | 1109 | return "bz2" |
be60ffd0 ERE |
1110 | elif self.buf.startswith((b"\x5d\x00\x00\x80", b"\xfd7zXZ")): |
1111 | return "xz" | |
1112 | else: | |
1113 | return "tar" | |
7584f5c9 ERE |
1114 | |
1115 | def close(self): | |
1116 | self.fileobj.close() | |
1117 | # class StreamProxy | |
1118 | ||
7584f5c9 ERE |
1119 | #------------------------ |
1120 | # Extraction file object | |
1121 | #------------------------ | |
1122 | class _FileInFile(object): | |
1123 | """A thin wrapper around an existing file object that | |
1124 | provides a part of its data as an individual file | |
1125 | object. | |
1126 | """ | |
1127 | ||
be60ffd0 | 1128 | def __init__(self, fileobj, offset, size, blockinfo=None): |
7584f5c9 ERE |
1129 | self.fileobj = fileobj |
1130 | self.offset = offset | |
1131 | self.size = size | |
7584f5c9 | 1132 | self.position = 0 |
be60ffd0 ERE |
1133 | self.name = getattr(fileobj, "name", None) |
1134 | self.closed = False | |
1135 | ||
1136 | if blockinfo is None: | |
1137 | blockinfo = [(0, size)] | |
1138 | ||
1139 | # Construct a map with data and zero blocks. | |
1140 | self.map_index = 0 | |
1141 | self.map = [] | |
1142 | lastpos = 0 | |
1143 | realpos = self.offset | |
1144 | for offset, size in blockinfo: | |
1145 | if offset > lastpos: | |
1146 | self.map.append((False, lastpos, offset, None)) | |
1147 | self.map.append((True, offset, offset + size, realpos)) | |
1148 | realpos += size | |
1149 | lastpos = offset + size | |
1150 | if lastpos < self.size: | |
1151 | self.map.append((False, lastpos, self.size, None)) | |
1152 | ||
1153 | def flush(self): | |
1154 | pass | |
1155 | ||
1156 | def readable(self): | |
1157 | return True | |
1158 | ||
1159 | def writable(self): | |
1160 | return False | |
1161 | ||
1162 | def seekable(self): | |
1163 | return self.fileobj.seekable() | |
7584f5c9 ERE |
1164 | |
1165 | def tell(self): | |
1166 | """Return the current file position. | |
1167 | """ | |
1168 | return self.position | |
1169 | ||
be60ffd0 | 1170 | def seek(self, position, whence=io.SEEK_SET): |
7584f5c9 ERE |
1171 | """Seek to a position in the file. |
1172 | """ | |
be60ffd0 ERE |
1173 | if whence == io.SEEK_SET: |
1174 | self.position = min(max(position, 0), self.size) | |
1175 | elif whence == io.SEEK_CUR: | |
1176 | if position < 0: | |
1177 | self.position = max(self.position + position, 0) | |
1178 | else: | |
1179 | self.position = min(self.position + position, self.size) | |
1180 | elif whence == io.SEEK_END: | |
1181 | self.position = max(min(self.size + position, self.size), 0) | |
1182 | else: | |
1183 | raise ValueError("Invalid argument") | |
1184 | return self.position | |
7584f5c9 ERE |
1185 | |
1186 | def read(self, size=None): | |
1187 | """Read data from the file. | |
1188 | """ | |
1189 | if size is None: | |
1190 | size = self.size - self.position | |
1191 | else: | |
1192 | size = min(size, self.size - self.position) | |
1193 | ||
be60ffd0 | 1194 | buf = b"" |
7584f5c9 | 1195 | while size > 0: |
7584f5c9 | 1196 | while True: |
be60ffd0 ERE |
1197 | data, start, stop, offset = self.map[self.map_index] |
1198 | if start <= self.position < stop: | |
7584f5c9 | 1199 | break |
be60ffd0 ERE |
1200 | else: |
1201 | self.map_index += 1 | |
1202 | if self.map_index == len(self.map): | |
1203 | self.map_index = 0 | |
1204 | length = min(size, stop - self.position) | |
1205 | if data: | |
1206 | self.fileobj.seek(offset + (self.position - start)) | |
1207 | buf += self.fileobj.read(length) | |
7584f5c9 | 1208 | else: |
be60ffd0 ERE |
1209 | buf += NUL * length |
1210 | size -= length | |
1211 | self.position += length | |
1212 | return buf | |
7584f5c9 | 1213 | |
be60ffd0 ERE |
1214 | def readinto(self, b): |
1215 | buf = self.read(len(b)) | |
1216 | b[:len(buf)] = buf | |
1217 | return len(buf) | |
7584f5c9 ERE |
1218 | |
1219 | def close(self): | |
7584f5c9 | 1220 | self.closed = True |
be60ffd0 | 1221 | #class _FileInFile |
7584f5c9 | 1222 | |
be60ffd0 ERE |
1223 | |
1224 | class ExFileObject(io.BufferedReader): | |
1225 | ||
1226 | def __init__(self, tarfile, tarinfo): | |
1227 | fileobj = _FileInFile(tarfile.fileobj, tarinfo.offset_data, | |
1228 | tarinfo.size, tarinfo.sparse) | |
1229 | super().__init__(fileobj) | |
7584f5c9 ERE |
1230 | #class ExFileObject |
1231 | ||
1232 | #------------------ | |
1233 | # Exported Classes | |
1234 | #------------------ | |
1235 | class TarInfo(object): | |
1236 | """Informational class which holds the details about an | |
1237 | archive member given by a tar header block. | |
1238 | TarInfo objects are returned by TarFile.getmember(), | |
1239 | TarFile.getmembers() and TarFile.gettarinfo() and are | |
1240 | usually created internally. | |
1241 | """ | |
1242 | ||
be60ffd0 ERE |
1243 | __slots__ = ("name", "mode", "uid", "gid", "size", "mtime", |
1244 | "chksum", "type", "linkname", "uname", "gname", | |
1245 | "devmajor", "devminor", "volume_offset", | |
1246 | "offset", "offset_data", "pax_headers", "sparse", | |
1247 | "tarfile", "_sparse_structs", "_link_target") | |
1248 | ||
7584f5c9 ERE |
1249 | def __init__(self, name=""): |
1250 | """Construct a TarInfo object. name is the optional name | |
1251 | of the member. | |
1252 | """ | |
1253 | self.name = name # member name | |
be60ffd0 | 1254 | self.mode = 0o644 # file permissions |
7584f5c9 ERE |
1255 | self.uid = 0 # user id |
1256 | self.gid = 0 # group id | |
1257 | self.size = 0 # file size | |
1258 | self.mtime = 0 # modification time | |
1259 | self.chksum = 0 # header checksum | |
1260 | self.type = REGTYPE # member type | |
1261 | self.linkname = "" # link name | |
1262 | self.uname = "" # user name | |
1263 | self.gname = "" # group name | |
1264 | self.devmajor = 0 # device major number | |
1265 | self.devminor = 0 # device minor number | |
1266 | ||
1267 | self.offset = 0 # the tar header starts here | |
1268 | self.offset_data = 0 # the file's data starts here | |
0eb5048f ERE |
1269 | self.volume_offset = 0 # the file's data corresponds with the data |
1270 | # starting at this position | |
7584f5c9 | 1271 | |
be60ffd0 | 1272 | self.sparse = None # sparse member information |
7584f5c9 ERE |
1273 | self.pax_headers = {} # pax header information |
1274 | ||
1275 | # In pax headers the "name" and "linkname" field are called | |
1276 | # "path" and "linkpath". | |
1277 | def _getpath(self): | |
1278 | return self.name | |
1279 | def _setpath(self, name): | |
1280 | self.name = name | |
1281 | path = property(_getpath, _setpath) | |
1282 | ||
1283 | def _getlinkpath(self): | |
1284 | return self.linkname | |
1285 | def _setlinkpath(self, linkname): | |
1286 | self.linkname = linkname | |
1287 | linkpath = property(_getlinkpath, _setlinkpath) | |
1288 | ||
1289 | def __repr__(self): | |
1290 | return "<%s %r at %#x>" % (self.__class__.__name__,self.name,id(self)) | |
1291 | ||
be60ffd0 | 1292 | def get_info(self, encoding=None, errors=None): |
7584f5c9 ERE |
1293 | """Return the TarInfo's attributes as a dictionary. |
1294 | """ | |
1295 | info = { | |
1296 | "name": self.name, | |
be60ffd0 | 1297 | "mode": self.mode & 0o7777, |
7584f5c9 ERE |
1298 | "uid": self.uid, |
1299 | "gid": self.gid, | |
1300 | "size": self.size, | |
1301 | "mtime": self.mtime, | |
1302 | "chksum": self.chksum, | |
1303 | "type": self.type, | |
1304 | "linkname": self.linkname, | |
1305 | "uname": self.uname, | |
1306 | "gname": self.gname, | |
1307 | "devmajor": self.devmajor, | |
36a315a0 | 1308 | "devminor": self.devminor, |
0eb5048f ERE |
1309 | "offset_data": self.offset_data, |
1310 | "volume_offset": self.volume_offset | |
7584f5c9 ERE |
1311 | } |
1312 | ||
1313 | if info["type"] == DIRTYPE and not info["name"].endswith("/"): | |
1314 | info["name"] += "/" | |
1315 | ||
7584f5c9 ERE |
1316 | return info |
1317 | ||
be60ffd0 ERE |
1318 | def tobuf(self, format=DEFAULT_FORMAT, encoding=ENCODING, |
1319 | errors="surrogateescape"): | |
7584f5c9 ERE |
1320 | """Return a tar header as a string of 512 byte blocks. |
1321 | """ | |
1322 | info = self.get_info(encoding, errors) | |
1323 | ||
1324 | if format == USTAR_FORMAT: | |
be60ffd0 | 1325 | return self.create_ustar_header(info, encoding, errors) |
7584f5c9 | 1326 | elif format == GNU_FORMAT: |
be60ffd0 | 1327 | return self.create_gnu_header(info, encoding, errors) |
7584f5c9 ERE |
1328 | elif format == PAX_FORMAT: |
1329 | return self.create_pax_header(info, encoding, errors) | |
1330 | else: | |
1331 | raise ValueError("invalid format") | |
1332 | ||
be60ffd0 | 1333 | def create_ustar_header(self, info, encoding, errors): |
7584f5c9 ERE |
1334 | """Return the object as a ustar header block. |
1335 | """ | |
1336 | info["magic"] = POSIX_MAGIC | |
1337 | ||
1338 | if len(info["linkname"]) > LENGTH_LINK: | |
1339 | raise ValueError("linkname is too long") | |
1340 | ||
1341 | if len(info["name"]) > LENGTH_NAME: | |
1342 | info["prefix"], info["name"] = self._posix_split_name(info["name"]) | |
1343 | ||
be60ffd0 | 1344 | return self._create_header(info, USTAR_FORMAT, encoding, errors) |
7584f5c9 | 1345 | |
be60ffd0 | 1346 | def create_gnu_header(self, info, encoding, errors): |
7584f5c9 ERE |
1347 | """Return the object as a GNU header block sequence. |
1348 | """ | |
1349 | info["magic"] = GNU_MAGIC | |
1350 | ||
2f854e77 ERE |
1351 | if self.ismultivol(): |
1352 | prefix = [ | |
1353 | itn(info.get("atime", 0), 12, GNU_FORMAT), | |
1354 | itn(info.get("ctime", 0), 12, GNU_FORMAT), | |
0eb5048f | 1355 | itn(self.volume_offset, 12, GNU_FORMAT), |
2f854e77 ERE |
1356 | itn(0, 119, GNU_FORMAT), # stuff unused in this tar implementation, set to zero |
1357 | ] | |
be60ffd0 | 1358 | info['prefix'] = b"".join(prefix) |
0eb5048f | 1359 | info['size'] = info['size'] - self.volume_offset |
2f854e77 | 1360 | |
be60ffd0 | 1361 | buf = b"" |
7584f5c9 | 1362 | if len(info["linkname"]) > LENGTH_LINK: |
be60ffd0 ERE |
1363 | buf += self._create_gnu_long_header(info["linkname"], |
1364 | GNUTYPE_LONGLINK, encoding, errors) | |
7584f5c9 ERE |
1365 | |
1366 | if len(info["name"]) > LENGTH_NAME: | |
be60ffd0 ERE |
1367 | buf += self._create_gnu_long_header(info["name"], GNUTYPE_LONGNAME, |
1368 | encoding, errors) | |
7584f5c9 | 1369 | |
be60ffd0 | 1370 | return buf + self._create_header(info, GNU_FORMAT, encoding, errors) |
7584f5c9 ERE |
1371 | |
1372 | def create_pax_header(self, info, encoding, errors): | |
1373 | """Return the object as a ustar header block. If it cannot be | |
1374 | represented this way, prepend a pax extended header sequence | |
1375 | with supplement information. | |
1376 | """ | |
1377 | info["magic"] = POSIX_MAGIC | |
1378 | pax_headers = self.pax_headers.copy() | |
c04e0751 ERE |
1379 | if self.ismultivol(): |
1380 | info['size'] = info['size'] - self.volume_offset | |
7584f5c9 ERE |
1381 | |
1382 | # Test string fields for values that exceed the field length or cannot | |
1383 | # be represented in ASCII encoding. | |
1384 | for name, hname, length in ( | |
36a315a0 ERE |
1385 | ("name", "path", LENGTH_NAME), |
1386 | ("linkname", "linkpath", LENGTH_LINK), | |
1387 | ("uname", "uname", 32), | |
1388 | ("gname", "gname", 32)): | |
7584f5c9 ERE |
1389 | |
1390 | if hname in pax_headers: | |
1391 | # The pax header has priority. | |
1392 | continue | |
1393 | ||
7584f5c9 ERE |
1394 | # Try to encode the string as ASCII. |
1395 | try: | |
be60ffd0 | 1396 | info[name].encode("ascii", "strict") |
7584f5c9 | 1397 | except UnicodeEncodeError: |
be60ffd0 | 1398 | pax_headers[hname] = info[name] |
7584f5c9 ERE |
1399 | continue |
1400 | ||
1401 | if len(info[name]) > length: | |
be60ffd0 | 1402 | pax_headers[hname] = info[name] |
7584f5c9 ERE |
1403 | |
1404 | # Test number fields for values that exceed the field limit or values | |
1405 | # that like to be stored as float. | |
1406 | for name, digits in (("uid", 8), ("gid", 8), ("size", 12), ("mtime", 12)): | |
1407 | if name in pax_headers: | |
1408 | # The pax header has priority. Avoid overflow. | |
1409 | info[name] = 0 | |
1410 | continue | |
1411 | ||
1412 | val = info[name] | |
1413 | if not 0 <= val < 8 ** (digits - 1) or isinstance(val, float): | |
be60ffd0 | 1414 | pax_headers[name] = str(val) |
7584f5c9 ERE |
1415 | info[name] = 0 |
1416 | ||
1417 | # Create a pax extended header if necessary. | |
1418 | if pax_headers: | |
be60ffd0 | 1419 | buf = self._create_pax_generic_header(pax_headers, XHDTYPE, encoding) |
7584f5c9 | 1420 | else: |
be60ffd0 | 1421 | buf = b"" |
7584f5c9 | 1422 | |
be60ffd0 | 1423 | return buf + self._create_header(info, USTAR_FORMAT, "ascii", "replace") |
7584f5c9 ERE |
1424 | |
1425 | @classmethod | |
1426 | def create_pax_global_header(cls, pax_headers): | |
1427 | """Return the object as a pax global header block sequence. | |
1428 | """ | |
be60ffd0 | 1429 | return cls._create_pax_generic_header(pax_headers, XGLTYPE, "utf-8") |
7584f5c9 ERE |
1430 | |
1431 | def _posix_split_name(self, name): | |
1432 | """Split a name longer than 100 chars into a prefix | |
1433 | and a name part. | |
1434 | """ | |
1435 | prefix = name[:LENGTH_PREFIX + 1] | |
1436 | while prefix and prefix[-1] != "/": | |
1437 | prefix = prefix[:-1] | |
1438 | ||
1439 | name = name[len(prefix):] | |
1440 | prefix = prefix[:-1] | |
1441 | ||
1442 | if not prefix or len(name) > LENGTH_NAME: | |
1443 | raise ValueError("name is too long") | |
1444 | return prefix, name | |
1445 | ||
1446 | @staticmethod | |
be60ffd0 | 1447 | def _create_header(info, format, encoding, errors): |
7584f5c9 ERE |
1448 | """Return a header block. info is a dictionary with file |
1449 | information, format must be one of the *_FORMAT constants. | |
1450 | """ | |
1451 | parts = [ | |
be60ffd0 ERE |
1452 | stn(info.get("name", ""), 100, encoding, errors), |
1453 | itn(info.get("mode", 0) & 0o7777, 8, format), | |
7584f5c9 ERE |
1454 | itn(info.get("uid", 0), 8, format), |
1455 | itn(info.get("gid", 0), 8, format), | |
1456 | itn(info.get("size", 0), 12, format), | |
1457 | itn(info.get("mtime", 0), 12, format), | |
be60ffd0 | 1458 | b" ", # checksum field |
2f854e77 | 1459 | info.get("type", REGTYPE), |
be60ffd0 ERE |
1460 | stn(info.get("linkname", ""), 100, encoding, errors), |
1461 | info.get("magic", POSIX_MAGIC), | |
1462 | stn(info.get("uname", ""), 32, encoding, errors), | |
1463 | stn(info.get("gname", ""), 32, encoding, errors), | |
7584f5c9 ERE |
1464 | itn(info.get("devmajor", 0), 8, format), |
1465 | itn(info.get("devminor", 0), 8, format), | |
be60ffd0 | 1466 | sbtn(info.get("prefix", ""), 155, encoding, errors) |
7584f5c9 ERE |
1467 | ] |
1468 | ||
be60ffd0 | 1469 | buf = struct.pack("%ds" % BLOCKSIZE, b"".join(parts)) |
7584f5c9 | 1470 | chksum = calc_chksums(buf[-BLOCKSIZE:])[0] |
be60ffd0 | 1471 | buf = buf[:-364] + bytes("%06o\0" % chksum, "ascii") + buf[-357:] |
7584f5c9 ERE |
1472 | return buf |
1473 | ||
1474 | @staticmethod | |
1475 | def _create_payload(payload): | |
1476 | """Return the string payload filled with zero bytes | |
1477 | up to the next 512 byte border. | |
1478 | """ | |
1479 | blocks, remainder = divmod(len(payload), BLOCKSIZE) | |
1480 | if remainder > 0: | |
1481 | payload += (BLOCKSIZE - remainder) * NUL | |
1482 | return payload | |
1483 | ||
1484 | @classmethod | |
be60ffd0 | 1485 | def _create_gnu_long_header(cls, name, type, encoding, errors): |
7584f5c9 ERE |
1486 | """Return a GNUTYPE_LONGNAME or GNUTYPE_LONGLINK sequence |
1487 | for name. | |
1488 | """ | |
be60ffd0 | 1489 | name = name.encode(encoding, errors) + NUL |
7584f5c9 ERE |
1490 | |
1491 | info = {} | |
1492 | info["name"] = "././@LongLink" | |
1493 | info["type"] = type | |
1494 | info["size"] = len(name) | |
1495 | info["magic"] = GNU_MAGIC | |
1496 | ||
1497 | # create extended header + name blocks. | |
be60ffd0 | 1498 | return cls._create_header(info, USTAR_FORMAT, encoding, errors) + \ |
7584f5c9 ERE |
1499 | cls._create_payload(name) |
1500 | ||
1501 | @classmethod | |
be60ffd0 ERE |
1502 | def _create_pax_generic_header(cls, pax_headers, type, encoding): |
1503 | """Return a POSIX.1-2008 extended or global header sequence | |
7584f5c9 | 1504 | that contains a list of keyword, value pairs. The values |
be60ffd0 | 1505 | must be strings. |
7584f5c9 | 1506 | """ |
be60ffd0 ERE |
1507 | # Check if one of the fields contains surrogate characters and thereby |
1508 | # forces hdrcharset=BINARY, see _proc_pax() for more information. | |
1509 | binary = False | |
1510 | for keyword, value in pax_headers.items(): | |
1511 | try: | |
1512 | value.encode("utf-8", "strict") | |
1513 | except UnicodeEncodeError: | |
1514 | binary = True | |
1515 | break | |
1516 | ||
1517 | records = b"" | |
1518 | if binary: | |
1519 | # Put the hdrcharset field at the beginning of the header. | |
1520 | records += b"21 hdrcharset=BINARY\n" | |
1521 | ||
1522 | for keyword, value in pax_headers.items(): | |
1523 | keyword = keyword.encode("utf-8") | |
1524 | if binary: | |
1525 | # Try to restore the original byte representation of `value'. | |
1526 | # Needless to say, that the encoding must match the string. | |
1527 | value = value.encode(encoding, "surrogateescape") | |
1528 | else: | |
1529 | value = value.encode("utf-8") | |
1530 | ||
7584f5c9 ERE |
1531 | l = len(keyword) + len(value) + 3 # ' ' + '=' + '\n' |
1532 | n = p = 0 | |
1533 | while True: | |
1534 | n = l + len(str(p)) | |
1535 | if n == p: | |
1536 | break | |
1537 | p = n | |
be60ffd0 | 1538 | records += bytes(str(p), "ascii") + b" " + keyword + b"=" + value + b"\n" |
7584f5c9 ERE |
1539 | |
1540 | # We use a hardcoded "././@PaxHeader" name like star does | |
1541 | # instead of the one that POSIX recommends. | |
1542 | info = {} | |
1543 | info["name"] = "././@PaxHeader" | |
1544 | info["type"] = type | |
1545 | info["size"] = len(records) | |
1546 | info["magic"] = POSIX_MAGIC | |
1547 | ||
1548 | # Create pax header + record blocks. | |
be60ffd0 | 1549 | return cls._create_header(info, USTAR_FORMAT, "ascii", "replace") + \ |
7584f5c9 ERE |
1550 | cls._create_payload(records) |
1551 | ||
1552 | @classmethod | |
be60ffd0 ERE |
1553 | def frombuf(cls, buf, encoding, errors): |
1554 | """Construct a TarInfo object from a 512 byte bytes object. | |
7584f5c9 ERE |
1555 | """ |
1556 | if len(buf) == 0: | |
1557 | raise EmptyHeaderError("empty header") | |
1558 | if len(buf) != BLOCKSIZE: | |
1559 | raise TruncatedHeaderError("truncated header") | |
1560 | if buf.count(NUL) == BLOCKSIZE: | |
1561 | raise EOFHeaderError("end of file header") | |
1562 | ||
1563 | chksum = nti(buf[148:156]) | |
1564 | if chksum not in calc_chksums(buf): | |
1565 | raise InvalidHeaderError("bad checksum") | |
1566 | ||
1567 | obj = cls() | |
be60ffd0 | 1568 | obj.name = nts(buf[0:100], encoding, errors) |
7584f5c9 ERE |
1569 | obj.mode = nti(buf[100:108]) |
1570 | obj.uid = nti(buf[108:116]) | |
1571 | obj.gid = nti(buf[116:124]) | |
1572 | obj.size = nti(buf[124:136]) | |
1573 | obj.mtime = nti(buf[136:148]) | |
1574 | obj.chksum = chksum | |
1575 | obj.type = buf[156:157] | |
be60ffd0 ERE |
1576 | obj.linkname = nts(buf[157:257], encoding, errors) |
1577 | obj.uname = nts(buf[265:297], encoding, errors) | |
1578 | obj.gname = nts(buf[297:329], encoding, errors) | |
7584f5c9 ERE |
1579 | obj.devmajor = nti(buf[329:337]) |
1580 | obj.devminor = nti(buf[337:345]) | |
be60ffd0 ERE |
1581 | prefix = nts(buf[345:500], encoding, errors) |
1582 | ||
1583 | # The old GNU sparse format occupies some of the unused | |
1584 | # space in the buffer for up to 4 sparse structures. | |
1585 | # Save the them for later processing in _proc_sparse(). | |
1586 | if obj.type == GNUTYPE_SPARSE: | |
1587 | pos = 386 | |
1588 | structs = [] | |
1589 | for i in range(4): | |
1590 | try: | |
1591 | offset = nti(buf[pos:pos + 12]) | |
1592 | numbytes = nti(buf[pos + 12:pos + 24]) | |
1593 | except ValueError: | |
1594 | break | |
1595 | structs.append((offset, numbytes)) | |
1596 | pos += 24 | |
1597 | isextended = bool(buf[482]) | |
1598 | origsize = nti(buf[483:495]) | |
1599 | obj._sparse_structs = (structs, isextended, origsize) | |
7584f5c9 ERE |
1600 | |
1601 | # Old V7 tar format represents a directory as a regular | |
1602 | # file with a trailing slash. | |
1603 | if obj.type == AREGTYPE and obj.name.endswith("/"): | |
1604 | obj.type = DIRTYPE | |
1605 | ||
1606 | # Remove redundant slashes from directories. | |
1607 | if obj.isdir(): | |
1608 | obj.name = obj.name.rstrip("/") | |
1609 | ||
1610 | # Reconstruct a ustar longname. | |
1611 | if prefix and obj.type not in GNU_TYPES: | |
1612 | obj.name = prefix + "/" + obj.name | |
c474439c ERE |
1613 | else: |
1614 | obj.offset_data = nti(buf[369:381]) | |
7584f5c9 ERE |
1615 | return obj |
1616 | ||
1617 | @classmethod | |
1618 | def fromtarfile(cls, tarfile): | |
1619 | """Return the next TarInfo object from TarFile object | |
1620 | tarfile. | |
1621 | """ | |
1622 | buf = tarfile.fileobj.read(BLOCKSIZE) | |
be60ffd0 | 1623 | obj = cls.frombuf(buf, tarfile.encoding, tarfile.errors) |
7584f5c9 ERE |
1624 | obj.offset = tarfile.fileobj.tell() - BLOCKSIZE |
1625 | return obj._proc_member(tarfile) | |
1626 | ||
1627 | #-------------------------------------------------------------------------- | |
1628 | # The following are methods that are called depending on the type of a | |
1629 | # member. The entry point is _proc_member() which can be overridden in a | |
1630 | # subclass to add custom _proc_*() methods. A _proc_*() method MUST | |
1631 | # implement the following | |
1632 | # operations: | |
1633 | # 1. Set self.offset_data to the position where the data blocks begin, | |
1634 | # if there is data that follows. | |
1635 | # 2. Set tarfile.offset to the position where the next member's header will | |
1636 | # begin. | |
1637 | # 3. Return self or another valid TarInfo object. | |
1638 | def _proc_member(self, tarfile): | |
1639 | """Choose the right processing method depending on | |
1640 | the type and call it. | |
1641 | """ | |
1642 | if self.type in (GNUTYPE_LONGNAME, GNUTYPE_LONGLINK): | |
1643 | return self._proc_gnulong(tarfile) | |
1644 | elif self.type == GNUTYPE_SPARSE: | |
1645 | return self._proc_sparse(tarfile) | |
1646 | elif self.type in (XHDTYPE, XGLTYPE, SOLARIS_XHDTYPE): | |
1647 | return self._proc_pax(tarfile) | |
1648 | else: | |
1649 | return self._proc_builtin(tarfile) | |
1650 | ||
1651 | def _proc_builtin(self, tarfile): | |
1652 | """Process a builtin type or an unknown type which | |
1653 | will be treated as a regular file. | |
1654 | """ | |
1655 | self.offset_data = tarfile.fileobj.tell() | |
1656 | offset = self.offset_data | |
00c34a12 | 1657 | if self.isreg() or self.ismultivol() or self.type not in SUPPORTED_TYPES: |
7584f5c9 ERE |
1658 | # Skip the following data blocks. |
1659 | offset += self._block(self.size) | |
1660 | tarfile.offset = offset | |
1661 | ||
1662 | # Patch the TarInfo object with saved global | |
1663 | # header information. | |
1664 | self._apply_pax_info(tarfile.pax_headers, tarfile.encoding, tarfile.errors) | |
1665 | ||
1666 | return self | |
1667 | ||
1668 | def _proc_gnulong(self, tarfile): | |
1669 | """Process the blocks that hold a GNU longname | |
1670 | or longlink member. | |
1671 | """ | |
1672 | buf = tarfile.fileobj.read(self._block(self.size)) | |
1673 | ||
1674 | # Fetch the next header and process it. | |
1675 | try: | |
1676 | next = self.fromtarfile(tarfile) | |
1677 | except HeaderError: | |
1678 | raise SubsequentHeaderError("missing or bad subsequent header") | |
1679 | ||
1680 | # Patch the TarInfo object from the next header with | |
1681 | # the longname information. | |
1682 | next.offset = self.offset | |
1683 | if self.type == GNUTYPE_LONGNAME: | |
be60ffd0 | 1684 | next.name = nts(buf, tarfile.encoding, tarfile.errors) |
7584f5c9 | 1685 | elif self.type == GNUTYPE_LONGLINK: |
be60ffd0 | 1686 | next.linkname = nts(buf, tarfile.encoding, tarfile.errors) |
7584f5c9 ERE |
1687 | |
1688 | return next | |
1689 | ||
1690 | def _proc_sparse(self, tarfile): | |
1691 | """Process a GNU sparse header plus extra headers. | |
1692 | """ | |
be60ffd0 ERE |
1693 | # We already collected some sparse structures in frombuf(). |
1694 | structs, isextended, origsize = self._sparse_structs | |
1695 | del self._sparse_structs | |
1696 | ||
1697 | # Collect sparse structures from extended header blocks. | |
1698 | while isextended: | |
7584f5c9 ERE |
1699 | buf = tarfile.fileobj.read(BLOCKSIZE) |
1700 | pos = 0 | |
be60ffd0 | 1701 | for i in range(21): |
7584f5c9 ERE |
1702 | try: |
1703 | offset = nti(buf[pos:pos + 12]) | |
1704 | numbytes = nti(buf[pos + 12:pos + 24]) | |
1705 | except ValueError: | |
1706 | break | |
be60ffd0 ERE |
1707 | if offset and numbytes: |
1708 | structs.append((offset, numbytes)) | |
7584f5c9 | 1709 | pos += 24 |
be60ffd0 ERE |
1710 | isextended = bool(buf[504]) |
1711 | self.sparse = structs | |
7584f5c9 ERE |
1712 | |
1713 | self.offset_data = tarfile.fileobj.tell() | |
1714 | tarfile.offset = self.offset_data + self._block(self.size) | |
1715 | self.size = origsize | |
7584f5c9 ERE |
1716 | return self |
1717 | ||
1718 | def _proc_pax(self, tarfile): | |
1719 | """Process an extended or global header as described in | |
be60ffd0 | 1720 | POSIX.1-2008. |
7584f5c9 ERE |
1721 | """ |
1722 | # Read the header information. | |
1723 | buf = tarfile.fileobj.read(self._block(self.size)) | |
1724 | ||
1725 | # A pax header stores supplemental information for either | |
1726 | # the following file (extended) or all following files | |
1727 | # (global). | |
1728 | if self.type == XGLTYPE: | |
1729 | pax_headers = tarfile.pax_headers | |
1730 | else: | |
1731 | pax_headers = tarfile.pax_headers.copy() | |
1732 | ||
be60ffd0 ERE |
1733 | # Check if the pax header contains a hdrcharset field. This tells us |
1734 | # the encoding of the path, linkpath, uname and gname fields. Normally, | |
1735 | # these fields are UTF-8 encoded but since POSIX.1-2008 tar | |
1736 | # implementations are allowed to store them as raw binary strings if | |
1737 | # the translation to UTF-8 fails. | |
1738 | match = re.search(br"\d+ hdrcharset=([^\n]+)\n", buf) | |
1739 | if match is not None: | |
1740 | pax_headers["hdrcharset"] = match.group(1).decode("utf-8") | |
1741 | ||
1742 | # For the time being, we don't care about anything other than "BINARY". | |
1743 | # The only other value that is currently allowed by the standard is | |
1744 | # "ISO-IR 10646 2000 UTF-8" in other words UTF-8. | |
1745 | hdrcharset = pax_headers.get("hdrcharset") | |
1746 | if hdrcharset == "BINARY": | |
1747 | encoding = tarfile.encoding | |
1748 | else: | |
1749 | encoding = "utf-8" | |
1750 | ||
7584f5c9 ERE |
1751 | # Parse pax header information. A record looks like that: |
1752 | # "%d %s=%s\n" % (length, keyword, value). length is the size | |
1753 | # of the complete record including the length field itself and | |
1754 | # the newline. keyword and value are both UTF-8 encoded strings. | |
be60ffd0 | 1755 | regex = re.compile(br"(\d+) ([^=]+)=") |
7584f5c9 ERE |
1756 | pos = 0 |
1757 | while True: | |
1758 | match = regex.match(buf, pos) | |
1759 | if not match: | |
1760 | break | |
1761 | ||
1762 | length, keyword = match.groups() | |
1763 | length = int(length) | |
1764 | value = buf[match.end(2) + 1:match.start(1) + length - 1] | |
1765 | ||
be60ffd0 ERE |
1766 | # Normally, we could just use "utf-8" as the encoding and "strict" |
1767 | # as the error handler, but we better not take the risk. For | |
1768 | # example, GNU tar <= 1.23 is known to store filenames it cannot | |
1769 | # translate to UTF-8 as raw strings (unfortunately without a | |
1770 | # hdrcharset=BINARY header). | |
1771 | # We first try the strict standard encoding, and if that fails we | |
1772 | # fall back on the user's encoding and error handler. | |
1773 | keyword = self._decode_pax_field(keyword, "utf-8", "utf-8", | |
1774 | tarfile.errors) | |
1775 | if keyword in PAX_NAME_FIELDS: | |
1776 | value = self._decode_pax_field(value, encoding, tarfile.encoding, | |
1777 | tarfile.errors) | |
1778 | else: | |
1779 | value = self._decode_pax_field(value, "utf-8", "utf-8", | |
1780 | tarfile.errors) | |
7584f5c9 ERE |
1781 | |
1782 | pax_headers[keyword] = value | |
1783 | pos += length | |
1784 | ||
36a315a0 | 1785 | |
7584f5c9 ERE |
1786 | # Fetch the next header. |
1787 | try: | |
1788 | next = self.fromtarfile(tarfile) | |
1789 | except HeaderError: | |
1790 | raise SubsequentHeaderError("missing or bad subsequent header") | |
1791 | ||
be60ffd0 ERE |
1792 | # Process GNU sparse information. |
1793 | if "GNU.sparse.map" in pax_headers: | |
1794 | # GNU extended sparse format version 0.1. | |
1795 | self._proc_gnusparse_01(next, pax_headers) | |
1796 | ||
1797 | elif "GNU.sparse.size" in pax_headers: | |
1798 | # GNU extended sparse format version 0.0. | |
1799 | self._proc_gnusparse_00(next, pax_headers, buf) | |
1800 | ||
1801 | elif pax_headers.get("GNU.sparse.major") == "1" and pax_headers.get("GNU.sparse.minor") == "0": | |
1802 | # GNU extended sparse format version 1.0. | |
1803 | self._proc_gnusparse_10(next, pax_headers, tarfile) | |
1804 | ||
7584f5c9 ERE |
1805 | if self.type in (XHDTYPE, SOLARIS_XHDTYPE): |
1806 | # Patch the TarInfo object with the extended header info. | |
1807 | next._apply_pax_info(pax_headers, tarfile.encoding, tarfile.errors) | |
1808 | next.offset = self.offset | |
1809 | ||
1810 | if "size" in pax_headers: | |
1811 | # If the extended header replaces the size field, | |
1812 | # we need to recalculate the offset where the next | |
1813 | # header starts. | |
1814 | offset = next.offset_data | |
1815 | if next.isreg() or next.type not in SUPPORTED_TYPES: | |
1816 | offset += next._block(next.size) | |
1817 | tarfile.offset = offset | |
1818 | ||
c04e0751 ERE |
1819 | if next is not None: |
1820 | if "GNU.volume.filename" in pax_headers: | |
1821 | if pax_headers["GNU.volume.filename"] == next.name: | |
1822 | if "GNU.volume.size" in pax_headers: | |
1823 | next.size = int(pax_headers["GNU.volume.size"]) | |
1824 | if "GNU.volume.offset" in pax_headers: | |
1825 | next.volume_offset = int(pax_headers["GNU.volume.offset"]) | |
1826 | ||
1827 | for key in pax_headers.keys(): | |
1828 | if key.startswith("GNU.volume"): | |
1829 | del tarfile.pax_headers[key] | |
0eb5048f | 1830 | |
7584f5c9 ERE |
1831 | return next |
1832 | ||
be60ffd0 ERE |
1833 | def _proc_gnusparse_00(self, next, pax_headers, buf): |
1834 | """Process a GNU tar extended sparse header, version 0.0. | |
7584f5c9 | 1835 | """ |
be60ffd0 ERE |
1836 | offsets = [] |
1837 | for match in re.finditer(br"\d+ GNU.sparse.offset=(\d+)\n", buf): | |
1838 | offsets.append(int(match.group(1))) | |
1839 | numbytes = [] | |
1840 | for match in re.finditer(br"\d+ GNU.sparse.numbytes=(\d+)\n", buf): | |
1841 | numbytes.append(int(match.group(1))) | |
1842 | next.sparse = list(zip(offsets, numbytes)) | |
7584f5c9 | 1843 | |
be60ffd0 ERE |
1844 | def _proc_gnusparse_01(self, next, pax_headers): |
1845 | """Process a GNU tar extended sparse header, version 0.1. | |
1846 | """ | |
1847 | sparse = [int(x) for x in pax_headers["GNU.sparse.map"].split(",")] | |
1848 | next.sparse = list(zip(sparse[::2], sparse[1::2])) | |
7584f5c9 | 1849 | |
be60ffd0 ERE |
1850 | def _proc_gnusparse_10(self, next, pax_headers, tarfile): |
1851 | """Process a GNU tar extended sparse header, version 1.0. | |
1852 | """ | |
1853 | fields = None | |
1854 | sparse = [] | |
1855 | buf = tarfile.fileobj.read(BLOCKSIZE) | |
1856 | fields, buf = buf.split(b"\n", 1) | |
1857 | fields = int(fields) | |
1858 | while len(sparse) < fields * 2: | |
1859 | if b"\n" not in buf: | |
1860 | buf += tarfile.fileobj.read(BLOCKSIZE) | |
1861 | number, buf = buf.split(b"\n", 1) | |
1862 | sparse.append(int(number)) | |
1863 | next.offset_data = tarfile.fileobj.tell() | |
1864 | next.sparse = list(zip(sparse[::2], sparse[1::2])) | |
7584f5c9 | 1865 | |
be60ffd0 ERE |
1866 | def _apply_pax_info(self, pax_headers, encoding, errors): |
1867 | """Replace fields with supplemental information from a previous | |
1868 | pax extended or global header. | |
1869 | """ | |
1870 | for keyword, value in pax_headers.items(): | |
1871 | if keyword == "GNU.sparse.name": | |
1872 | setattr(self, "path", value) | |
1873 | elif keyword == "GNU.sparse.size": | |
1874 | setattr(self, "size", int(value)) | |
1875 | elif keyword == "GNU.sparse.realsize": | |
1876 | setattr(self, "size", int(value)) | |
1877 | elif keyword in PAX_FIELDS: | |
1878 | if keyword in PAX_NUMBER_FIELDS: | |
1879 | try: | |
1880 | value = PAX_NUMBER_FIELDS[keyword](value) | |
1881 | except ValueError: | |
1882 | value = 0 | |
1883 | if keyword == "path": | |
f0287fb7 | 1884 | value = value.rstrip("/") # pylint: disable=no-member |
be60ffd0 | 1885 | setattr(self, keyword, value) |
7584f5c9 ERE |
1886 | |
1887 | self.pax_headers = pax_headers.copy() | |
1888 | ||
be60ffd0 ERE |
1889 | def _decode_pax_field(self, value, encoding, fallback_encoding, fallback_errors): |
1890 | """Decode a single field from a pax record. | |
1891 | """ | |
1892 | try: | |
1893 | return value.decode(encoding, "strict") | |
1894 | except UnicodeDecodeError: | |
1895 | return value.decode(fallback_encoding, fallback_errors) | |
1896 | ||
7584f5c9 ERE |
1897 | def _block(self, count): |
1898 | """Round up a byte count by BLOCKSIZE and return it, | |
1899 | e.g. _block(834) => 1024. | |
1900 | """ | |
1901 | blocks, remainder = divmod(count, BLOCKSIZE) | |
1902 | if remainder: | |
1903 | blocks += 1 | |
1904 | return blocks * BLOCKSIZE | |
1905 | ||
1906 | def isreg(self): | |
1907 | return self.type in REGULAR_TYPES | |
1908 | def isfile(self): | |
1909 | return self.isreg() | |
1910 | def isdir(self): | |
1911 | return self.type == DIRTYPE | |
1912 | def issym(self): | |
1913 | return self.type == SYMTYPE | |
1914 | def islnk(self): | |
1915 | return self.type == LNKTYPE | |
1916 | def ischr(self): | |
1917 | return self.type == CHRTYPE | |
1918 | def isblk(self): | |
1919 | return self.type == BLKTYPE | |
1920 | def isfifo(self): | |
1921 | return self.type == FIFOTYPE | |
1922 | def issparse(self): | |
be60ffd0 | 1923 | return self.sparse is not None |
7584f5c9 ERE |
1924 | def isdev(self): |
1925 | return self.type in (CHRTYPE, BLKTYPE, FIFOTYPE) | |
68ddf955 | 1926 | def ismultivol(self): |
c04e0751 ERE |
1927 | return self.type == GNUTYPE_MULTIVOL or self.volume_offset > 0 or\ |
1928 | "GNU.volume.offset" in self.pax_headers | |
7584f5c9 ERE |
1929 | # class TarInfo |
1930 | ||
1931 | class TarFile(object): | |
1932 | """The TarFile Class provides an interface to tar archives. | |
1933 | """ | |
1934 | ||
1935 | debug = 0 # May be set from 0 (no msgs) to 3 (all msgs) | |
1936 | ||
1937 | dereference = False # If true, add content of linked file to the | |
1938 | # tar file, else the link. | |
1939 | ||
1940 | ignore_zeros = False # If true, skips empty or invalid blocks and | |
1941 | # continues processing. | |
1942 | ||
83f2d71e | 1943 | max_volume_size = None # If different from None, establishes maximum |
68ddf955 ERE |
1944 | # size of tar volumes |
1945 | ||
1946 | new_volume_handler = None # function handler to be executed before when | |
1947 | # a new volume is needed | |
1948 | ||
1949 | volume_number = 0 # current volume number, used for multi volume | |
1950 | # support | |
1951 | ||
7584f5c9 ERE |
1952 | errorlevel = 1 # If 0, fatal errors only appear in debug |
1953 | # messages (if debug >= 0). If > 0, errors | |
1954 | # are passed to the caller as exceptions. | |
1955 | ||
1956 | format = DEFAULT_FORMAT # The format to use when creating an archive. | |
1957 | ||
1958 | encoding = ENCODING # Encoding for 8-bit character strings. | |
1959 | ||
1960 | errors = None # Error handler for unicode conversion. | |
1961 | ||
1962 | tarinfo = TarInfo # The default TarInfo class to use. | |
1963 | ||
be60ffd0 | 1964 | fileobject = ExFileObject # The file-object for extractfile(). |
7584f5c9 | 1965 | |
d1c38f40 PG |
1966 | arcmode = ARCMODE_PLAIN # Object processing mode (“concat”, encryption, |
1967 | # compression) | |
5fdff89f | 1968 | |
ea625b04 ERE |
1969 | save_to_members = True # If new members are saved. This can be disabled |
1970 | # if you manage lots of files and don't want | |
1971 | # to have high memory usage | |
1972 | ||
9ef1fb87 TJ |
1973 | cache_uid2user = {} # cache to avoid getpwuid calls. It always parses /etc/passwd. |
1974 | cache_gid2group = {} # same cache for groups | |
1975 | ||
7584f5c9 ERE |
1976 | def __init__(self, name=None, mode="r", fileobj=None, format=None, |
1977 | tarinfo=None, dereference=None, ignore_zeros=None, encoding=None, | |
be60ffd0 | 1978 | errors="surrogateescape", pax_headers=None, debug=None, |
548bb8d5 | 1979 | errorlevel=None, max_volume_size=None, new_volume_handler=None, |
d1c38f40 | 1980 | concat=False, nacl=None, |
c7c736b6 | 1981 | save_to_members=True): |
7584f5c9 ERE |
1982 | """Open an (uncompressed) tar archive `name'. `mode' is either 'r' to |
1983 | read from an existing archive, 'a' to append data to an existing | |
1984 | file or 'w' to create a new file overwriting an existing one. `mode' | |
1985 | defaults to 'r'. | |
1986 | If `fileobj' is given, it is used for reading or writing data. If it | |
1987 | can be determined, `mode' is overridden by `fileobj's mode. | |
1988 | `fileobj' is not closed, when TarFile is closed. | |
1989 | """ | |
1990 | if len(mode) > 1 or mode not in "raw": | |
1991 | raise ValueError("mode must be 'r', 'a' or 'w'") | |
1992 | self.mode = mode | |
d1c38f40 | 1993 | self.arcmode = arcmode_set (concat) |
c7c736b6 | 1994 | self.nacl = nacl |
7584f5c9 ERE |
1995 | self._mode = {"r": "rb", "a": "r+b", "w": "wb"}[mode] |
1996 | ||
1997 | if not fileobj: | |
1998 | if self.mode == "a" and not os.path.exists(name): | |
1999 | # Create nonexistent files in append mode. | |
2000 | self.mode = "w" | |
2001 | self._mode = "wb" | |
2002 | fileobj = bltn_open(name, self._mode) | |
2003 | self._extfileobj = False | |
2004 | else: | |
2005 | if name is None and hasattr(fileobj, "name"): | |
2006 | name = fileobj.name | |
d5361dac | 2007 | # when fileobj is a gzip.GzipFile, fileobj.mode is an int (not valid for us) |
be60ffd0 | 2008 | if hasattr(fileobj, "mode") and isinstance(fileobj.mode, str): |
7584f5c9 ERE |
2009 | self._mode = fileobj.mode |
2010 | self._extfileobj = True | |
be60ffd0 | 2011 | self.name = os.path.abspath(name) if name else None |
2f854e77 | 2012 | self.base_name = self.name = os.path.abspath(name) if name else None |
7584f5c9 ERE |
2013 | self.fileobj = fileobj |
2014 | ||
2015 | # Init attributes. | |
2016 | if format is not None: | |
2017 | self.format = format | |
2018 | if tarinfo is not None: | |
2019 | self.tarinfo = tarinfo | |
2020 | if dereference is not None: | |
2021 | self.dereference = dereference | |
2022 | if ignore_zeros is not None: | |
2023 | self.ignore_zeros = ignore_zeros | |
2024 | if encoding is not None: | |
2025 | self.encoding = encoding | |
2026 | ||
be60ffd0 | 2027 | self.errors = errors |
7584f5c9 ERE |
2028 | |
2029 | if pax_headers is not None and self.format == PAX_FORMAT: | |
2030 | self.pax_headers = pax_headers | |
2031 | else: | |
2032 | self.pax_headers = {} | |
2033 | ||
2034 | if debug is not None: | |
2035 | self.debug = debug | |
2036 | if errorlevel is not None: | |
2037 | self.errorlevel = errorlevel | |
2038 | ||
2039 | # Init datastructures. | |
ae48acc8 | 2040 | if max_volume_size and max_volume_size < 3*BLOCKSIZE: |
0c818a18 | 2041 | raise ValueError("max_volume_size needs to be at least %d" % (3*BLOCKSIZE)) |
ae48acc8 ERE |
2042 | if max_volume_size and not callable(new_volume_handler): |
2043 | raise ValueError("new_volume_handler needs to be set and be callable for multivolume support") | |
5ab3f8f9 CH |
2044 | if max_volume_size: |
2045 | self.max_volume_size = int(max_volume_size) | |
2046 | else: | |
2047 | self.max_volume_size = None | |
ae48acc8 | 2048 | |
ea625b04 | 2049 | self.save_to_members = save_to_members |
68ddf955 | 2050 | self.new_volume_handler = new_volume_handler |
7584f5c9 ERE |
2051 | self.closed = False |
2052 | self.members = [] # list of members as TarInfo objects | |
2053 | self._loaded = False # flag if all members have been read | |
2054 | self.offset = self.fileobj.tell() | |
2055 | # current position in the archive file | |
2056 | self.inodes = {} # dictionary caching the inodes of | |
2057 | # archive members already added | |
2058 | ||
2059 | try: | |
2060 | if self.mode == "r": | |
2061 | self.firstmember = None | |
2062 | self.firstmember = self.next() | |
2063 | ||
2064 | if self.mode == "a": | |
2065 | # Move to the end of the archive, | |
2066 | # before the first empty block. | |
2067 | while True: | |
2068 | self.fileobj.seek(self.offset) | |
2069 | try: | |
2070 | tarinfo = self.tarinfo.fromtarfile(self) | |
2071 | self.members.append(tarinfo) | |
2072 | except EOFHeaderError: | |
2073 | self.fileobj.seek(self.offset) | |
2074 | break | |
be60ffd0 | 2075 | except HeaderError as e: |
7584f5c9 ERE |
2076 | raise ReadError(str(e)) |
2077 | ||
2078 | if self.mode in "aw": | |
2079 | self._loaded = True | |
2080 | ||
2081 | if self.pax_headers: | |
2082 | buf = self.tarinfo.create_pax_global_header(self.pax_headers.copy()) | |
2083 | self.fileobj.write(buf) | |
2084 | self.offset += len(buf) | |
2085 | except: | |
2086 | if not self._extfileobj: | |
2087 | self.fileobj.close() | |
2088 | self.closed = True | |
2089 | raise | |
2090 | ||
7584f5c9 ERE |
2091 | #-------------------------------------------------------------------------- |
2092 | # Below are the classmethods which act as alternate constructors to the | |
2093 | # TarFile class. The open() method is the only one that is needed for | |
2094 | # public use; it is the "super"-constructor and is able to select an | |
2095 | # adequate "sub"-constructor for a particular compression using the mapping | |
2096 | # from OPEN_METH. | |
2097 | # | |
2098 | # This concept allows one to subclass TarFile without losing the comfort of | |
2099 | # the super-constructor. A sub-constructor is registered and made available | |
2100 | # by adding it to the mapping in OPEN_METH. | |
2101 | ||
2102 | @classmethod | |
2b82f50c | 2103 | def open(cls, name=None, mode="r", fileobj=None, bufsize=RECORDSIZE, |
044585c6 | 2104 | encryption=None, compresslevel=9, tolerant=False, **kwargs): |
7584f5c9 ERE |
2105 | """Open a tar archive for reading, writing or appending. Return |
2106 | an appropriate TarFile class. | |
2107 | ||
2108 | mode: | |
2109 | 'r' or 'r:*' open for reading with transparent compression | |
2110 | 'r:' open for reading exclusively uncompressed | |
2111 | 'r:gz' open for reading with gzip compression | |
2112 | 'r:bz2' open for reading with bzip2 compression | |
be60ffd0 | 2113 | 'r:xz' open for reading with lzma compression |
7584f5c9 ERE |
2114 | 'a' or 'a:' open for appending, creating the file if necessary |
2115 | 'w' or 'w:' open for writing without compression | |
2116 | 'w:gz' open for writing with gzip compression | |
2117 | 'w:bz2' open for writing with bzip2 compression | |
be60ffd0 | 2118 | 'w:xz' open for writing with lzma compression |
7584f5c9 ERE |
2119 | |
2120 | 'r|*' open a stream of tar blocks with transparent compression | |
2121 | 'r|' open an uncompressed stream of tar blocks for reading | |
2122 | 'r|gz' open a gzip compressed stream of tar blocks | |
2123 | 'r|bz2' open a bzip2 compressed stream of tar blocks | |
be60ffd0 | 2124 | 'r|xz' open an lzma compressed stream of tar blocks |
7584f5c9 ERE |
2125 | 'w|' open an uncompressed stream for writing |
2126 | 'w|gz' open a gzip compressed stream for writing | |
2127 | 'w|bz2' open a bzip2 compressed stream for writing | |
be60ffd0 | 2128 | 'w|xz' open an lzma compressed stream for writing |
85737f48 ERE |
2129 | |
2130 | 'r#gz' open a stream of gzip compressed tar blocks for reading | |
2131 | 'w#gz' open a stream of gzip compressed tar blocks for writing | |
7584f5c9 | 2132 | """ |
7584f5c9 ERE |
2133 | if not name and not fileobj: |
2134 | raise ValueError("nothing to open") | |
2135 | ||
2136 | if mode in ("r", "r:*"): | |
2137 | # Find out which *open() is appropriate for opening the file. | |
2138 | for comptype in cls.OPEN_METH: | |
2139 | func = getattr(cls, cls.OPEN_METH[comptype]) | |
2140 | if fileobj is not None: | |
2141 | saved_pos = fileobj.tell() | |
2142 | try: | |
2143 | return func(name, "r", fileobj, **kwargs) | |
be60ffd0 | 2144 | except (ReadError, CompressionError) as e: |
c7c736b6 | 2145 | # usually nothing exceptional but sometimes is |
7584f5c9 ERE |
2146 | if fileobj is not None: |
2147 | fileobj.seek(saved_pos) | |
2148 | continue | |
2149 | raise ReadError("file could not be opened successfully") | |
2150 | ||
2151 | elif ":" in mode: | |
2152 | filemode, comptype = mode.split(":", 1) | |
2153 | filemode = filemode or "r" | |
2154 | comptype = comptype or "tar" | |
2155 | ||
2156 | # Select the *open() function according to | |
2157 | # given compression. | |
2158 | if comptype in cls.OPEN_METH: | |
2159 | func = getattr(cls, cls.OPEN_METH[comptype]) | |
2160 | else: | |
2161 | raise CompressionError("unknown compression type %r" % comptype) | |
e05f0440 TJ |
2162 | |
2163 | # Pass on compression level for gzip / bzip2. | |
2164 | if comptype == 'gz' or comptype == 'bz2': | |
2165 | kwargs['compresslevel'] = compresslevel | |
2166 | ||
7a2b9329 CH |
2167 | if 'max_volume_size' in kwargs: |
2168 | if comptype != 'tar' and filemode in 'wa' \ | |
2169 | and kwargs['max_volume_size']: | |
2170 | import warnings | |
2171 | warnings.warn('Only the first volume will be compressed ' | |
2172 | 'for modes with "w:"!') | |
2173 | ||
e05f0440 | 2174 | return func(name, filemode, fileobj, **kwargs) |
7584f5c9 ERE |
2175 | |
2176 | elif "|" in mode: | |
2177 | filemode, comptype = mode.split("|", 1) | |
2178 | filemode = filemode or "r" | |
2179 | comptype = comptype or "tar" | |
2180 | ||
2181 | if filemode not in "rw": | |
2182 | raise ValueError("mode must be 'r' or 'w'") | |
2183 | ||
2184 | t = cls(name, filemode, | |
2b82f50c ERE |
2185 | _Stream(name, filemode, comptype, fileobj, bufsize, |
2186 | compresslevel=compresslevel), | |
7584f5c9 ERE |
2187 | **kwargs) |
2188 | t._extfileobj = False | |
2189 | return t | |
2190 | ||
5fdff89f ERE |
2191 | elif "#" in mode: |
2192 | filemode, comptype = mode.split("#", 1) | |
2193 | filemode = filemode or "r" | |
5fdff89f ERE |
2194 | |
2195 | if filemode not in "rw": | |
5faea0e1 PG |
2196 | raise ValueError ("mode %s not compatible with concat " |
2197 | "archive; must be 'r' or 'w'" % mode) | |
5fdff89f | 2198 | |
be60ffd0 | 2199 | stream = _Stream(name, filemode, comptype, fileobj, bufsize, |
d1c38f40 | 2200 | concat=True, encryption=encryption, |
044585c6 | 2201 | compresslevel=compresslevel, tolerant=tolerant) |
d1c38f40 | 2202 | kwargs ["concat"] = True |
be60ffd0 ERE |
2203 | try: |
2204 | t = cls(name, filemode, stream, **kwargs) | |
c7c736b6 | 2205 | except: # XXX except what? |
be60ffd0 | 2206 | stream.close() |
c7c736b6 | 2207 | raise # XXX raise what? |
5fdff89f ERE |
2208 | t._extfileobj = False |
2209 | return t | |
2210 | ||
7584f5c9 ERE |
2211 | elif mode in "aw": |
2212 | return cls.taropen(name, mode, fileobj, **kwargs) | |
2213 | ||
133d30da | 2214 | raise ValueError("undiscernible mode %r" % mode) |
7584f5c9 ERE |
2215 | |
2216 | @classmethod | |
2217 | def taropen(cls, name, mode="r", fileobj=None, **kwargs): | |
2218 | """Open uncompressed tar archive name for reading or writing. | |
2219 | """ | |
2220 | if len(mode) > 1 or mode not in "raw": | |
2221 | raise ValueError("mode must be 'r', 'a' or 'w'") | |
2222 | return cls(name, mode, fileobj, **kwargs) | |
2223 | ||
2224 | @classmethod | |
2225 | def gzopen(cls, name, mode="r", fileobj=None, compresslevel=9, **kwargs): | |
2226 | """Open gzip compressed tar archive name for reading or writing. | |
2227 | Appending is not allowed. | |
2228 | """ | |
2229 | if len(mode) > 1 or mode not in "rw": | |
2230 | raise ValueError("mode must be 'r' or 'w'") | |
2231 | ||
2232 | try: | |
2233 | import gzip | |
2234 | gzip.GzipFile | |
2235 | except (ImportError, AttributeError): | |
2236 | raise CompressionError("gzip module is not available") | |
2237 | ||
be60ffd0 | 2238 | extfileobj = fileobj is not None |
7584f5c9 | 2239 | try: |
be60ffd0 ERE |
2240 | fileobj = gzip.GzipFile(name, mode + "b", compresslevel, fileobj) |
2241 | t = cls.taropen(name, mode, fileobj, **kwargs) | |
2242 | except OSError: | |
2243 | if not extfileobj and fileobj is not None: | |
2244 | fileobj.close() | |
2245 | if fileobj is None: | |
2246 | raise | |
7584f5c9 | 2247 | raise ReadError("not a gzip file") |
be60ffd0 ERE |
2248 | except: |
2249 | if not extfileobj and fileobj is not None: | |
2250 | fileobj.close() | |
2251 | raise | |
2252 | t._extfileobj = extfileobj | |
7584f5c9 ERE |
2253 | return t |
2254 | ||
2255 | @classmethod | |
2256 | def bz2open(cls, name, mode="r", fileobj=None, compresslevel=9, **kwargs): | |
2257 | """Open bzip2 compressed tar archive name for reading or writing. | |
2258 | Appending is not allowed. | |
2259 | """ | |
2260 | if len(mode) > 1 or mode not in "rw": | |
2261 | raise ValueError("mode must be 'r' or 'w'.") | |
2262 | ||
2263 | try: | |
2264 | import bz2 | |
2265 | except ImportError: | |
2266 | raise CompressionError("bz2 module is not available") | |
2267 | ||
be60ffd0 ERE |
2268 | fileobj = bz2.BZ2File(fileobj or name, mode, |
2269 | compresslevel=compresslevel) | |
7584f5c9 ERE |
2270 | |
2271 | try: | |
2272 | t = cls.taropen(name, mode, fileobj, **kwargs) | |
be60ffd0 ERE |
2273 | except (OSError, EOFError): |
2274 | fileobj.close() | |
7584f5c9 ERE |
2275 | raise ReadError("not a bzip2 file") |
2276 | t._extfileobj = False | |
2277 | return t | |
2278 | ||
be60ffd0 ERE |
2279 | @classmethod |
2280 | def xzopen(cls, name, mode="r", fileobj=None, preset=None, **kwargs): | |
2281 | """Open lzma compressed tar archive name for reading or writing. | |
2282 | Appending is not allowed. | |
2283 | """ | |
2284 | if mode not in ("r", "w"): | |
2285 | raise ValueError("mode must be 'r' or 'w'") | |
2286 | ||
2287 | try: | |
2288 | import lzma | |
2289 | except ImportError: | |
2290 | raise CompressionError("lzma module is not available") | |
2291 | ||
2292 | fileobj = lzma.LZMAFile(fileobj or name, mode, preset=preset) | |
2293 | ||
2294 | try: | |
2295 | t = cls.taropen(name, mode, fileobj, **kwargs) | |
2296 | except (lzma.LZMAError, EOFError): | |
2297 | fileobj.close() | |
2298 | raise ReadError("not an lzma file") | |
2299 | t._extfileobj = False | |
2300 | return t | |
2301 | ||
7584f5c9 ERE |
2302 | # All *open() methods are registered here. |
2303 | OPEN_METH = { | |
2304 | "tar": "taropen", # uncompressed tar | |
2305 | "gz": "gzopen", # gzip compressed tar | |
be60ffd0 ERE |
2306 | "bz2": "bz2open", # bzip2 compressed tar |
2307 | "xz": "xzopen" # lzma compressed tar | |
7584f5c9 ERE |
2308 | } |
2309 | ||
2310 | #-------------------------------------------------------------------------- | |
2311 | # The public methods which TarFile provides: | |
2312 | ||
2313 | def close(self): | |
2314 | """Close the TarFile. In write-mode, two finishing zero blocks are | |
fd2f01f2 PG |
2315 | appended to the archive. A special case are empty archives which are |
2316 | initialized accordingly so the two mandatory blocks of zeros are | |
2317 | written abiding by the requested encryption and compression settings. | |
7584f5c9 ERE |
2318 | """ |
2319 | if self.closed: | |
2320 | return | |
2321 | ||
2322 | if self.mode in "aw": | |
fd2f01f2 PG |
2323 | if self.arcmode & ARCMODE_CONCAT and self.fileobj.tell () == 0: |
2324 | self.fileobj.next ("") | |
7584f5c9 ERE |
2325 | self.fileobj.write(NUL * (BLOCKSIZE * 2)) |
2326 | self.offset += (BLOCKSIZE * 2) | |
2327 | # fill up the end with zero-blocks | |
2328 | # (like option -b20 for tar does) | |
2329 | blocks, remainder = divmod(self.offset, RECORDSIZE) | |
2330 | if remainder > 0: | |
2331 | self.fileobj.write(NUL * (RECORDSIZE - remainder)) | |
7584f5c9 ERE |
2332 | if not self._extfileobj: |
2333 | self.fileobj.close() | |
2334 | self.closed = True | |
2335 | ||
2336 | def getmember(self, name): | |
2337 | """Return a TarInfo object for member `name'. If `name' can not be | |
2338 | found in the archive, KeyError is raised. If a member occurs more | |
2339 | than once in the archive, its last occurrence is assumed to be the | |
2340 | most up-to-date version. | |
2341 | """ | |
2342 | tarinfo = self._getmember(name) | |
2343 | if tarinfo is None: | |
2344 | raise KeyError("filename %r not found" % name) | |
2345 | return tarinfo | |
2346 | ||
2347 | def getmembers(self): | |
2348 | """Return the members of the archive as a list of TarInfo objects. The | |
2349 | list has the same order as the members in the archive. | |
2350 | """ | |
2351 | self._check() | |
2352 | if not self._loaded: # if we want to obtain a list of | |
2353 | self._load() # all members, we first have to | |
2354 | # scan the whole archive. | |
2355 | return self.members | |
2356 | ||
ad4402e8 ERE |
2357 | def get_last_member_offset(self): |
2358 | """Return the last member offset. Usually this is self.fileobj.tell(), | |
2359 | but when there's encryption or concat compression going on it's more | |
2360 | complicated than that. | |
2361 | """ | |
b8fc2f5d | 2362 | return self.last_block_offset |
ad4402e8 | 2363 | |
7584f5c9 ERE |
2364 | def getnames(self): |
2365 | """Return the members of the archive as a list of their names. It has | |
2366 | the same order as the list returned by getmembers(). | |
2367 | """ | |
2368 | return [tarinfo.name for tarinfo in self.getmembers()] | |
2369 | ||
2370 | def gettarinfo(self, name=None, arcname=None, fileobj=None): | |
2371 | """Create a TarInfo object for either the file `name' or the file | |
2372 | object `fileobj' (using os.fstat on its file descriptor). You can | |
2373 | modify some of the TarInfo's attributes before you add it using | |
2374 | addfile(). If given, `arcname' specifies an alternative name for the | |
2375 | file in the archive. | |
2376 | """ | |
2377 | self._check("aw") | |
2378 | ||
2379 | # When fileobj is given, replace name by | |
2380 | # fileobj's real name. | |
2381 | if fileobj is not None: | |
2382 | name = fileobj.name | |
2383 | ||
2384 | # Building the name of the member in the archive. | |
2385 | # Backward slashes are converted to forward slashes, | |
2386 | # Absolute paths are turned to relative paths. | |
2387 | if arcname is None: | |
2388 | arcname = name | |
2389 | drv, arcname = os.path.splitdrive(arcname) | |
be60ffd0 | 2390 | arcname = arcname.replace(os.sep, "/") |
7584f5c9 ERE |
2391 | arcname = arcname.lstrip("/") |
2392 | ||
2393 | # Now, fill the TarInfo object with | |
2394 | # information specific for the file. | |
2395 | tarinfo = self.tarinfo() | |
2396 | tarinfo.tarfile = self | |
2397 | ||
2398 | # Use os.stat or os.lstat, depending on platform | |
2399 | # and if symlinks shall be resolved. | |
2400 | if fileobj is None: | |
2401 | if hasattr(os, "lstat") and not self.dereference: | |
2402 | statres = os.lstat(name) | |
2403 | else: | |
2404 | statres = os.stat(name) | |
2405 | else: | |
2406 | statres = os.fstat(fileobj.fileno()) | |
2407 | linkname = "" | |
2408 | ||
2409 | stmd = statres.st_mode | |
2410 | if stat.S_ISREG(stmd): | |
2411 | inode = (statres.st_ino, statres.st_dev) | |
2412 | if not self.dereference and statres.st_nlink > 1 and \ | |
2413 | inode in self.inodes and arcname != self.inodes[inode]: | |
2414 | # Is it a hardlink to an already | |
2415 | # archived file? | |
2416 | type = LNKTYPE | |
2417 | linkname = self.inodes[inode] | |
2418 | else: | |
2419 | # The inode is added only if its valid. | |
2420 | # For win32 it is always 0. | |
2421 | type = REGTYPE | |
6f422b65 | 2422 | if inode[0] and self.save_to_members: |
7584f5c9 ERE |
2423 | self.inodes[inode] = arcname |
2424 | elif stat.S_ISDIR(stmd): | |
2425 | type = DIRTYPE | |
2426 | elif stat.S_ISFIFO(stmd): | |
2427 | type = FIFOTYPE | |
2428 | elif stat.S_ISLNK(stmd): | |
2429 | type = SYMTYPE | |
2430 | linkname = os.readlink(name) | |
2431 | elif stat.S_ISCHR(stmd): | |
2432 | type = CHRTYPE | |
2433 | elif stat.S_ISBLK(stmd): | |
2434 | type = BLKTYPE | |
2435 | else: | |
2436 | return None | |
2437 | ||
2438 | # Fill the TarInfo object with all | |
2439 | # information we can get. | |
2440 | tarinfo.name = arcname | |
2441 | tarinfo.mode = stmd | |
2442 | tarinfo.uid = statres.st_uid | |
2443 | tarinfo.gid = statres.st_gid | |
2444 | if type == REGTYPE: | |
2445 | tarinfo.size = statres.st_size | |
2446 | else: | |
be60ffd0 | 2447 | tarinfo.size = 0 |
7584f5c9 ERE |
2448 | tarinfo.mtime = statres.st_mtime |
2449 | tarinfo.type = type | |
2450 | tarinfo.linkname = linkname | |
2451 | if pwd: | |
9ef1fb87 TJ |
2452 | if tarinfo.uid in self.cache_uid2user: |
2453 | tarinfo.uname = self.cache_uid2user[tarinfo.uid] | |
2454 | else: | |
2455 | try: | |
2456 | tarinfo.uname = pwd.getpwuid(tarinfo.uid)[0] | |
2457 | self.cache_uid2user[tarinfo.uid] = tarinfo.uname | |
2458 | except KeyError: | |
2459 | # remember user does not exist: | |
2460 | # same default value as in tarinfo class | |
2461 | self.cache_uid2user[tarinfo.uid] = "" | |
7584f5c9 | 2462 | if grp: |
9ef1fb87 TJ |
2463 | if tarinfo.gid in self.cache_gid2group: |
2464 | tarinfo.gname = self.cache_gid2group[tarinfo.gid] | |
2465 | else: | |
2466 | try: | |
2467 | tarinfo.gname = grp.getgrgid(tarinfo.gid)[0] | |
2468 | self.cache_gid2group[tarinfo.gid] = tarinfo.gname | |
2469 | except KeyError: | |
2470 | # remember group does not exist: | |
2471 | # same default value as in tarinfo class | |
2472 | self.cache_gid2group[tarinfo.gid] = "" | |
7584f5c9 ERE |
2473 | |
2474 | if type in (CHRTYPE, BLKTYPE): | |
2475 | if hasattr(os, "major") and hasattr(os, "minor"): | |
2476 | tarinfo.devmajor = os.major(statres.st_rdev) | |
2477 | tarinfo.devminor = os.minor(statres.st_rdev) | |
2478 | return tarinfo | |
2479 | ||
2480 | def list(self, verbose=True): | |
2481 | """Print a table of contents to sys.stdout. If `verbose' is False, only | |
2482 | the names of the members are printed. If it is True, an `ls -l'-like | |
2483 | output is produced. | |
2484 | """ | |
2485 | self._check() | |
2486 | ||
2487 | for tarinfo in self: | |
2488 | if verbose: | |
be60ffd0 ERE |
2489 | print(stat.filemode(tarinfo.mode), end=' ') |
2490 | print("%s/%s" % (tarinfo.uname or tarinfo.uid, | |
2491 | tarinfo.gname or tarinfo.gid), end=' ') | |
7584f5c9 | 2492 | if tarinfo.ischr() or tarinfo.isblk(): |
be60ffd0 ERE |
2493 | print("%10s" % ("%d,%d" \ |
2494 | % (tarinfo.devmajor, tarinfo.devminor)), end=' ') | |
7584f5c9 | 2495 | else: |
be60ffd0 ERE |
2496 | print("%10d" % tarinfo.size, end=' ') |
2497 | print("%d-%02d-%02d %02d:%02d:%02d" \ | |
2498 | % time.localtime(tarinfo.mtime)[:6], end=' ') | |
7584f5c9 | 2499 | |
be60ffd0 | 2500 | print(tarinfo.name + ("/" if tarinfo.isdir() else ""), end=' ') |
7584f5c9 ERE |
2501 | |
2502 | if verbose: | |
2503 | if tarinfo.issym(): | |
be60ffd0 | 2504 | print("->", tarinfo.linkname, end=' ') |
7584f5c9 | 2505 | if tarinfo.islnk(): |
be60ffd0 ERE |
2506 | print("link to", tarinfo.linkname, end=' ') |
2507 | print() | |
7584f5c9 | 2508 | |
be60ffd0 | 2509 | def add(self, name, arcname=None, recursive=True, exclude=None, *, filter=None): |
7584f5c9 ERE |
2510 | """Add the file `name' to the archive. `name' may be any type of file |
2511 | (directory, fifo, symbolic link, etc.). If given, `arcname' | |
2512 | specifies an alternative name for the file in the archive. | |
2513 | Directories are added recursively by default. This can be avoided by | |
2514 | setting `recursive' to False. `exclude' is a function that should | |
2515 | return True for each filename to be excluded. `filter' is a function | |
2516 | that expects a TarInfo object argument and returns the changed | |
2517 | TarInfo object, if it returns None the TarInfo object will be | |
2518 | excluded from the archive. | |
2519 | """ | |
2520 | self._check("aw") | |
2521 | ||
2522 | if arcname is None: | |
2523 | arcname = name | |
2524 | ||
2525 | # Exclude pathnames. | |
2526 | if exclude is not None: | |
2527 | import warnings | |
2528 | warnings.warn("use the filter argument instead", | |
2529 | DeprecationWarning, 2) | |
2530 | if exclude(name): | |
2531 | self._dbg(2, "tarfile: Excluded %r" % name) | |
2532 | return | |
2533 | ||
2534 | # Skip if somebody tries to archive the archive... | |
2535 | if self.name is not None and os.path.abspath(name) == self.name: | |
2536 | self._dbg(2, "tarfile: Skipped %r" % name) | |
2537 | return | |
2538 | ||
2539 | self._dbg(1, name) | |
2540 | ||
2541 | # Create a TarInfo object from the file. | |
2542 | tarinfo = self.gettarinfo(name, arcname) | |
2543 | ||
2544 | if tarinfo is None: | |
2545 | self._dbg(1, "tarfile: Unsupported type %r" % name) | |
2546 | return | |
2547 | ||
2548 | # Change or exclude the TarInfo object. | |
2549 | if filter is not None: | |
2550 | tarinfo = filter(tarinfo) | |
2551 | if tarinfo is None: | |
2552 | self._dbg(2, "tarfile: Excluded %r" % name) | |
2553 | return | |
2554 | ||
2555 | # Append the tar header and data to the archive. | |
2556 | if tarinfo.isreg(): | |
2557 | with bltn_open(name, "rb") as f: | |
2558 | self.addfile(tarinfo, f) | |
2559 | ||
2560 | elif tarinfo.isdir(): | |
2561 | self.addfile(tarinfo) | |
2562 | if recursive: | |
2563 | for f in os.listdir(name): | |
2564 | self.add(os.path.join(name, f), os.path.join(arcname, f), | |
be60ffd0 | 2565 | recursive, exclude, filter=filter) |
7584f5c9 ERE |
2566 | |
2567 | else: | |
2568 | self.addfile(tarinfo) | |
2569 | ||
defc9a22 | 2570 | def _size_left_file(self): |
be60ffd0 | 2571 | """Calculates size left in a volume with a maximum volume size. |
ba5a449e | 2572 | |
be60ffd0 | 2573 | Assumes self.max_volume_size is set. |
ba5a449e | 2574 | If using compression through a _Stream, use _size_left_stream instead |
be60ffd0 | 2575 | """ |
ba5a449e | 2576 | # left-over size = max_size - offset - 2 zero-blocks written in close |
ae48acc8 ERE |
2577 | size_left = self.max_volume_size - 2*BLOCKSIZE - self.offset |
2578 | # limit size left to a discrete number of blocks, because we won't | |
be60ffd0 | 2579 | # write only half a block when writting the end of a volume |
ae48acc8 | 2580 | # and filling with zeros |
defc9a22 CH |
2581 | return BLOCKSIZE * (size_left // BLOCKSIZE) |
2582 | ||
2583 | def _size_left_stream(self): | |
ba5a449e CH |
2584 | """ Calculates size left in a volume if using comression/encryption |
2585 | ||
2586 | Assumes self.max_volume_size is set and self.fileobj is a _Stream | |
2587 | (otherwise use _size_left_file) | |
2588 | """ | |
2589 | # left-over size = max_size - bytes written - 2 zero-blocks (close) | |
defc9a22 CH |
2590 | size_left = self.max_volume_size - self.fileobj.estim_file_size() \ |
2591 | - 2*BLOCKSIZE | |
2592 | return BLOCKSIZE * (size_left // BLOCKSIZE) | |
ae48acc8 | 2593 | |
7584f5c9 ERE |
2594 | def addfile(self, tarinfo, fileobj=None): |
2595 | """Add the TarInfo object `tarinfo' to the archive. If `fileobj' is | |
2596 | given, tarinfo.size bytes are read from it and added to the archive. | |
2597 | You can create TarInfo objects using gettarinfo(). | |
2598 | On Windows platforms, `fileobj' should always be opened with mode | |
2599 | 'rb' to avoid irritation about the file size. | |
2600 | """ | |
2601 | self._check("aw") | |
2602 | ||
2603 | tarinfo = copy.copy(tarinfo) | |
cbf55ffb | 2604 | |
d1c38f40 PG |
2605 | if self.arcmode & ARCMODE_CONCAT: |
2606 | self.last_block_offset = self.fileobj.next (tarinfo.name) | |
11684b1d ERE |
2607 | else: |
2608 | self.last_block_offset = self.fileobj.tell() | |
7584f5c9 ERE |
2609 | |
2610 | buf = tarinfo.tobuf(self.format, self.encoding, self.errors) | |
2611 | self.fileobj.write(buf) | |
2612 | self.offset += len(buf) | |
2613 | ||
ae9c8de2 CH |
2614 | if self.max_volume_size: |
2615 | if isinstance(self.fileobj, _Stream): | |
2616 | _size_left = self._size_left_stream | |
2617 | else: | |
2618 | _size_left = self._size_left_file | |
2619 | else: | |
2620 | _size_left = lambda: tarinfo.size | |
68ddf955 | 2621 | |
29c354ac PG |
2622 | # If there's no data to follow, finish |
2623 | if not fileobj: | |
29c354ac PG |
2624 | if self.save_to_members: |
2625 | self.members.append(tarinfo) | |
2626 | return | |
2627 | ||
2628 | target_size_left = _size_left() | |
2629 | source_size_left = tarinfo.size | |
2630 | assert tarinfo.volume_offset == 0 | |
2631 | ||
2632 | # we only split volumes in the middle of a file, that means we have | |
2633 | # to write at least one block | |
2634 | if target_size_left < BLOCKSIZE: | |
2635 | target_size_left = BLOCKSIZE | |
2636 | ||
ae9c8de2 CH |
2637 | # loop over multiple volumes |
2638 | while source_size_left > 0: | |
ae48acc8 | 2639 | |
ae9c8de2 CH |
2640 | # Write as much data as possble from source into target. |
2641 | # When compressing data, we cannot easily predict how much data we | |
2642 | # can write until target_size_left == 0 --> need to iterate | |
2643 | size_can_write = min(target_size_left, source_size_left) | |
c04e0751 | 2644 | |
ae9c8de2 CH |
2645 | while size_can_write > 0: |
2646 | copyfileobj(fileobj, self.fileobj, size_can_write) | |
2647 | self.offset += size_can_write | |
2648 | source_size_left -= size_can_write | |
2649 | target_size_left = _size_left() | |
2650 | size_can_write = min(target_size_left, source_size_left) | |
68ddf955 | 2651 | |
ae9c8de2 CH |
2652 | # now target_size_left == 0 or source_size_left == 0 |
2653 | ||
2654 | # if there is data left to write, we need to create a new volume | |
2655 | if source_size_left > 0: | |
5f38bff6 PG |
2656 | # Only finalize the crypto entry here if we’re continuing with |
2657 | # another one; otherwise, the encryption must include the block | |
2658 | # padding below. | |
2f854e77 | 2659 | tarinfo.type = GNUTYPE_MULTIVOL |
68ddf955 ERE |
2660 | |
2661 | if not self.new_volume_handler or\ | |
2662 | not callable(self.new_volume_handler): | |
c04e0751 | 2663 | raise Exception("We need to create a new volume and you " |
ae9c8de2 | 2664 | "didn't supply a new_volume_handler") |
68ddf955 | 2665 | |
54128a00 | 2666 | |
68ddf955 ERE |
2667 | # the new volume handler should do everything needed to |
2668 | # start working in a new volume. usually, the handler calls | |
2669 | # to self.open_volume | |
2f854e77 | 2670 | self.volume_number += 1 |
0eb5048f | 2671 | |
ae9c8de2 | 2672 | # set to be used by open_volume, because in the case of a PAX |
0eb5048f ERE |
2673 | # tar it needs to write information about the volume and offset |
2674 | # in the global header | |
ae9c8de2 | 2675 | tarinfo.volume_offset = tarinfo.size - source_size_left |
0eb5048f | 2676 | self.volume_tarinfo = tarinfo |
ae9c8de2 | 2677 | |
a0873dcc PG |
2678 | # the “new_volume_handler” is supposed to call .close() on the |
2679 | # “fileobj” _Stream | |
2f854e77 ERE |
2680 | self.new_volume_handler(self, self.base_name, self.volume_number) |
2681 | ||
0eb5048f ERE |
2682 | self.volume_tarinfo = None |
2683 | ||
d1c38f40 PG |
2684 | if self.arcmode & ARCMODE_CONCAT: |
2685 | self.fileobj.next_volume (tarinfo.name) | |
5f38bff6 | 2686 | |
2f854e77 ERE |
2687 | # write new volume header |
2688 | buf = tarinfo.tobuf(self.format, self.encoding, self.errors) | |
2f854e77 | 2689 | self.fileobj.write(buf) |
ae9c8de2 CH |
2690 | self.offset += len(buf) |
2691 | ||
2692 | # adjust variables; open_volume should have reset self.offset | |
2693 | # --> _size_left should be big again | |
2694 | target_size_left = _size_left() | |
2695 | size_can_write = min(target_size_left, source_size_left) | |
e0da4709 | 2696 | self._dbg(3, 'new volume') |
ae9c8de2 CH |
2697 | |
2698 | # now, all data has been written. We may have to fill up the rest of | |
2699 | # the block in target with 0s | |
2700 | remainder = (tarinfo.size - tarinfo.volume_offset) % BLOCKSIZE | |
2701 | if remainder > 0: | |
2702 | self.fileobj.write(NUL * (BLOCKSIZE - remainder)) | |
2703 | self.offset += BLOCKSIZE - remainder | |
7584f5c9 | 2704 | |
ea625b04 ERE |
2705 | if self.save_to_members: |
2706 | self.members.append(tarinfo) | |
7584f5c9 | 2707 | |
170c6c52 | 2708 | def open_volume(self, name="", fileobj=None, encryption=None): |
68ddf955 | 2709 | ''' |
0eb5048f | 2710 | Called by the user to change this tar file to point to a new volume. |
68ddf955 ERE |
2711 | ''' |
2712 | # open the file using either fileobj or name | |
2713 | if not fileobj: | |
2714 | if self.mode == "a" and not os.path.exists(name): | |
2715 | # Create nonexistent files in append mode. | |
2716 | self.mode = "w" | |
2717 | self._mode = "wb" | |
68ddf955 | 2718 | self._extfileobj = False |
26fa5ad5 ERE |
2719 | |
2720 | if isinstance(self.fileobj, _Stream): | |
e0da4709 | 2721 | self._dbg(3, 'open_volume: create a _Stream') |
26fa5ad5 ERE |
2722 | fileobj = _Stream(name=name, |
2723 | mode=self.fileobj.mode, | |
2724 | comptype=self.fileobj.comptype, | |
2725 | fileobj=None, | |
2726 | bufsize=self.fileobj.bufsize, | |
cea130ec | 2727 | encryption=encryption or self.fileobj.encryption, |
d1c38f40 | 2728 | concat=self.fileobj.arcmode & ARCMODE_CONCAT) |
26fa5ad5 | 2729 | else: |
7a2b9329 | 2730 | # here, we lose information about compression/encryption! |
e0da4709 | 2731 | self._dbg(3, 'open_volume: builtin open') |
26fa5ad5 | 2732 | fileobj = bltn_open(name, self._mode) |
68ddf955 ERE |
2733 | else: |
2734 | if name is None and hasattr(fileobj, "name"): | |
2735 | name = fileobj.name | |
2736 | if hasattr(fileobj, "mode"): | |
2737 | self._mode = fileobj.mode | |
2738 | self._extfileobj = True | |
1027433a | 2739 | self._dbg(3, 'open_volume: using external fileobj {}', fileobj) |
68ddf955 ERE |
2740 | self.name = os.path.abspath(name) if name else None |
2741 | self.fileobj = fileobj | |
2742 | ||
2743 | # init data structures | |
2744 | self.closed = False | |
2745 | self.members = [] # list of members as TarInfo objects | |
2746 | self._loaded = False # flag if all members have been read | |
2747 | self.offset = self.fileobj.tell() | |
2748 | # current position in the archive file | |
2749 | self.inodes = {} # dictionary caching the inodes of | |
2750 | # archive members already added | |
2751 | ||
2752 | try: | |
2753 | if self.mode == "r": | |
2754 | self.firstmember = None | |
2755 | self.firstmember = self.next() | |
2756 | ||
2757 | if self.mode == "a": | |
2758 | # Move to the end of the archive, | |
2759 | # before the first empty block. | |
2760 | while True: | |
2761 | self.fileobj.seek(self.offset) | |
2762 | try: | |
2763 | tarinfo = self.tarinfo.fromtarfile(self) | |
2764 | self.members.append(tarinfo) | |
2765 | except EOFHeaderError: | |
2766 | self.fileobj.seek(self.offset) | |
2767 | break | |
be60ffd0 | 2768 | except HeaderError as e: |
68ddf955 ERE |
2769 | raise ReadError(str(e)) |
2770 | ||
2771 | if self.mode in "aw": | |
2772 | self._loaded = True | |
2773 | ||
c04e0751 ERE |
2774 | if self.format == PAX_FORMAT: |
2775 | volume_info = { | |
be60ffd0 ERE |
2776 | "GNU.volume.filename": str(self.volume_tarinfo.name), |
2777 | "GNU.volume.size": str(self.volume_tarinfo.size - self.volume_tarinfo.volume_offset), | |
2778 | "GNU.volume.offset": str(self.volume_tarinfo.volume_offset), | |
c04e0751 | 2779 | } |
0eb5048f | 2780 | |
c04e0751 ERE |
2781 | self.pax_headers.update(volume_info) |
2782 | ||
a0873dcc PG |
2783 | if isinstance(self.fileobj, _Stream): |
2784 | self.fileobj._init_write_gz () | |
c04e0751 ERE |
2785 | buf = self.tarinfo.create_pax_global_header(volume_info.copy()) |
2786 | self.fileobj.write(buf) | |
2787 | self.offset += len(buf) | |
54128a00 | 2788 | except Exception as exn: |
68ddf955 ERE |
2789 | if not self._extfileobj: |
2790 | self.fileobj.close() | |
2791 | self.closed = True | |
2792 | raise | |
2793 | ||
e5f5681b | 2794 | def extractall(self, path=".", members=None, filter=None): |
7584f5c9 ERE |
2795 | """Extract all members from the archive to the current working |
2796 | directory and set owner, modification time and permissions on | |
2797 | directories afterwards. `path' specifies a different directory | |
2798 | to extract to. `members' is optional and must be a subset of the | |
2799 | list returned by getmembers(). | |
2800 | """ | |
2801 | directories = [] | |
2802 | ||
2803 | if members is None: | |
2804 | members = self | |
2805 | ||
2806 | for tarinfo in members: | |
c474439c ERE |
2807 | if self.volume_number > 0 and tarinfo.ismultivol(): |
2808 | continue | |
2809 | ||
974408b5 | 2810 | if filter and not filter(tarinfo): |
e5f5681b ERE |
2811 | continue |
2812 | ||
7584f5c9 ERE |
2813 | if tarinfo.isdir(): |
2814 | # Extract directories with a safe mode. | |
2815 | directories.append(tarinfo) | |
2816 | tarinfo = copy.copy(tarinfo) | |
be60ffd0 ERE |
2817 | tarinfo.mode = 0o0700 |
2818 | # Do not set_attrs directories, as we will do that further down | |
2819 | self.extract(tarinfo, path, set_attrs=not tarinfo.isdir()) | |
7584f5c9 ERE |
2820 | |
2821 | # Reverse sort directories. | |
be60ffd0 | 2822 | directories.sort(key=lambda a: a.name) |
7584f5c9 ERE |
2823 | directories.reverse() |
2824 | ||
2825 | # Set correct owner, mtime and filemode on directories. | |
2826 | for tarinfo in directories: | |
2827 | dirpath = os.path.join(path, tarinfo.name) | |
2828 | try: | |
2829 | self.chown(tarinfo, dirpath) | |
2830 | self.utime(tarinfo, dirpath) | |
2831 | self.chmod(tarinfo, dirpath) | |
be60ffd0 | 2832 | except ExtractError as e: |
7584f5c9 ERE |
2833 | if self.errorlevel > 1: |
2834 | raise | |
2835 | else: | |
2836 | self._dbg(1, "tarfile: %s" % e) | |
2837 | ||
786addd6 | 2838 | def extract(self, member, path="", set_attrs=True, symlink_cb=None): |
7584f5c9 ERE |
2839 | """Extract a member from the archive to the current working directory, |
2840 | using its full name. Its file information is extracted as accurately | |
2841 | as possible. `member' may be a filename or a TarInfo object. You can | |
be60ffd0 ERE |
2842 | specify a different directory using `path'. File attributes (owner, |
2843 | mtime, mode) are set unless `set_attrs' is False. | |
786addd6 PG |
2844 | ``symlink_cb`` is a hook accepting a function that is passed the |
2845 | ``member``, ``path``, and ``set_attrs`` arguments if the tarinfo for | |
2846 | ``member`` indicates a symlink in which case only the callback | |
9b13f5c4 PG |
2847 | passed will be applied, skipping the actual extraction. In case the |
2848 | callback is invoked, its return value is passed on to the caller. | |
7584f5c9 ERE |
2849 | """ |
2850 | self._check("r") | |
2851 | ||
be60ffd0 | 2852 | if isinstance(member, str): |
7584f5c9 ERE |
2853 | tarinfo = self.getmember(member) |
2854 | else: | |
2855 | tarinfo = member | |
2856 | ||
2857 | # Prepare the link target for makelink(). | |
2858 | if tarinfo.islnk(): | |
2859 | tarinfo._link_target = os.path.join(path, tarinfo.linkname) | |
2860 | ||
9b13f5c4 | 2861 | if symlink_cb is not None and tarinfo.issym(): |
83f5fd71 | 2862 | return symlink_cb(member, path, set_attrs) |
786addd6 | 2863 | |
7584f5c9 | 2864 | try: |
be60ffd0 ERE |
2865 | self._extract_member(tarinfo, os.path.join(path, tarinfo.name), |
2866 | set_attrs=set_attrs) | |
2867 | except EnvironmentError as e: | |
7584f5c9 ERE |
2868 | if self.errorlevel > 0: |
2869 | raise | |
2870 | else: | |
2871 | if e.filename is None: | |
2872 | self._dbg(1, "tarfile: %s" % e.strerror) | |
2873 | else: | |
2874 | self._dbg(1, "tarfile: %s %r" % (e.strerror, e.filename)) | |
be60ffd0 | 2875 | except ExtractError as e: |
7584f5c9 ERE |
2876 | if self.errorlevel > 1: |
2877 | raise | |
2878 | else: | |
2879 | self._dbg(1, "tarfile: %s" % e) | |
2880 | ||
2881 | def extractfile(self, member): | |
2882 | """Extract a member from the archive as a file object. `member' may be | |
be60ffd0 ERE |
2883 | a filename or a TarInfo object. If `member' is a regular file or a |
2884 | link, an io.BufferedReader object is returned. Otherwise, None is | |
2885 | returned. | |
7584f5c9 ERE |
2886 | """ |
2887 | self._check("r") | |
2888 | ||
be60ffd0 | 2889 | if isinstance(member, str): |
7584f5c9 ERE |
2890 | tarinfo = self.getmember(member) |
2891 | else: | |
2892 | tarinfo = member | |
2893 | ||
be60ffd0 ERE |
2894 | if tarinfo.isreg() or tarinfo.ismultivol() or\ |
2895 | tarinfo.type not in SUPPORTED_TYPES: | |
7584f5c9 ERE |
2896 | # If a member's type is unknown, it is treated as a |
2897 | # regular file. | |
2898 | return self.fileobject(self, tarinfo) | |
2899 | ||
2900 | elif tarinfo.islnk() or tarinfo.issym(): | |
2901 | if isinstance(self.fileobj, _Stream): | |
2902 | # A small but ugly workaround for the case that someone tries | |
2903 | # to extract a (sym)link as a file-object from a non-seekable | |
2904 | # stream of tar blocks. | |
2905 | raise StreamError("cannot extract (sym)link as file object") | |
2906 | else: | |
2907 | # A (sym)link's file object is its target's file object. | |
2908 | return self.extractfile(self._find_link_target(tarinfo)) | |
2909 | else: | |
2910 | # If there's no data associated with the member (directory, chrdev, | |
2911 | # blkdev, etc.), return None instead of a file object. | |
2912 | return None | |
2913 | ||
be60ffd0 | 2914 | def _extract_member(self, tarinfo, targetpath, set_attrs=True): |
7584f5c9 ERE |
2915 | """Extract the TarInfo object tarinfo to a physical |
2916 | file called targetpath. | |
2917 | """ | |
2918 | # Fetch the TarInfo object for the given name | |
2919 | # and build the destination pathname, replacing | |
2920 | # forward slashes to platform specific separators. | |
2921 | targetpath = targetpath.rstrip("/") | |
2922 | targetpath = targetpath.replace("/", os.sep) | |
2923 | ||
2924 | # Create all upper directories. | |
2925 | upperdirs = os.path.dirname(targetpath) | |
2926 | if upperdirs and not os.path.exists(upperdirs): | |
2927 | # Create directories that are not part of the archive with | |
2928 | # default permissions. | |
2929 | os.makedirs(upperdirs) | |
2930 | ||
2931 | if tarinfo.islnk() or tarinfo.issym(): | |
2932 | self._dbg(1, "%s -> %s" % (tarinfo.name, tarinfo.linkname)) | |
2933 | else: | |
2934 | self._dbg(1, tarinfo.name) | |
2935 | ||
2936 | if tarinfo.isreg(): | |
2937 | self.makefile(tarinfo, targetpath) | |
2938 | elif tarinfo.isdir(): | |
2939 | self.makedir(tarinfo, targetpath) | |
2940 | elif tarinfo.isfifo(): | |
2941 | self.makefifo(tarinfo, targetpath) | |
2942 | elif tarinfo.ischr() or tarinfo.isblk(): | |
2943 | self.makedev(tarinfo, targetpath) | |
2944 | elif tarinfo.islnk() or tarinfo.issym(): | |
2945 | self.makelink(tarinfo, targetpath) | |
2946 | elif tarinfo.type not in SUPPORTED_TYPES: | |
2947 | self.makeunknown(tarinfo, targetpath) | |
2948 | else: | |
2949 | self.makefile(tarinfo, targetpath) | |
2950 | ||
be60ffd0 ERE |
2951 | if set_attrs: |
2952 | self.chown(tarinfo, targetpath) | |
2953 | if not tarinfo.issym(): | |
2954 | self.chmod(tarinfo, targetpath) | |
2955 | self.utime(tarinfo, targetpath) | |
7584f5c9 ERE |
2956 | |
2957 | #-------------------------------------------------------------------------- | |
2958 | # Below are the different file methods. They are called via | |
2959 | # _extract_member() when extract() is called. They can be replaced in a | |
2960 | # subclass to implement other functionality. | |
2961 | ||
2962 | def makedir(self, tarinfo, targetpath): | |
2963 | """Make a directory called targetpath. | |
2964 | """ | |
2965 | try: | |
2966 | # Use a safe mode for the directory, the real mode is set | |
2967 | # later in _extract_member(). | |
be60ffd0 ERE |
2968 | os.mkdir(targetpath, 0o0700) |
2969 | except FileExistsError: | |
2970 | pass | |
7584f5c9 ERE |
2971 | |
2972 | def makefile(self, tarinfo, targetpath): | |
2973 | """Make a file called targetpath. | |
2974 | """ | |
be60ffd0 ERE |
2975 | source = self.fileobj |
2976 | source.seek(tarinfo.offset_data) | |
c7c736b6 | 2977 | decrypt = False |
c474439c ERE |
2978 | iterate = True |
2979 | target = bltn_open(targetpath, "wb") | |
2980 | ||
be60ffd0 ERE |
2981 | if tarinfo.sparse is not None: |
2982 | try: | |
2983 | for offset, size in tarinfo.sparse: | |
2984 | target.seek(offset) | |
2985 | copyfileobj(source, target, size) | |
2986 | target.seek(tarinfo.size) | |
2987 | target.truncate() | |
2988 | finally: | |
2989 | target.close() | |
2990 | return | |
2991 | ||
c474439c ERE |
2992 | while iterate: |
2993 | iterate = False | |
2994 | try: | |
2995 | copyfileobj(source, target, tarinfo.size) | |
aa828cd1 | 2996 | except OSError: |
c474439c ERE |
2997 | source.close() |
2998 | # only if we are extracting a multivolume this can be treated | |
2999 | if not self.new_volume_handler: | |
3000 | target.close() | |
3001 | raise Exception("We need to read a new volume and you" | |
3002 | " didn't supply a new_volume_handler") | |
3003 | ||
3004 | # the new volume handler should do everything needed to | |
3005 | # start working in a new volume. usually, the handler calls | |
3006 | # to self.open_volume | |
3007 | self.volume_number += 1 | |
3008 | self.new_volume_handler(self, self.base_name, self.volume_number) | |
be60ffd0 ERE |
3009 | tarinfo = self.firstmember |
3010 | source = self.fileobj | |
c474439c | 3011 | iterate = True |
c474439c ERE |
3012 | target.close() |
3013 | ||
7584f5c9 ERE |
3014 | |
3015 | def makeunknown(self, tarinfo, targetpath): | |
3016 | """Make a file from a TarInfo object with an unknown type | |
3017 | at targetpath. | |
3018 | """ | |
3019 | self.makefile(tarinfo, targetpath) | |
3020 | self._dbg(1, "tarfile: Unknown file type %r, " \ | |
3021 | "extracted as regular file." % tarinfo.type) | |
3022 | ||
3023 | def makefifo(self, tarinfo, targetpath): | |
3024 | """Make a fifo called targetpath. | |
3025 | """ | |
3026 | if hasattr(os, "mkfifo"): | |
3027 | os.mkfifo(targetpath) | |
3028 | else: | |
3029 | raise ExtractError("fifo not supported by system") | |
3030 | ||
3031 | def makedev(self, tarinfo, targetpath): | |
3032 | """Make a character or block device called targetpath. | |
3033 | """ | |
3034 | if not hasattr(os, "mknod") or not hasattr(os, "makedev"): | |
3035 | raise ExtractError("special devices not supported by system") | |
3036 | ||
3037 | mode = tarinfo.mode | |
3038 | if tarinfo.isblk(): | |
3039 | mode |= stat.S_IFBLK | |
3040 | else: | |
3041 | mode |= stat.S_IFCHR | |
3042 | ||
3043 | os.mknod(targetpath, mode, | |
3044 | os.makedev(tarinfo.devmajor, tarinfo.devminor)) | |
3045 | ||
3046 | def makelink(self, tarinfo, targetpath): | |
3047 | """Make a (symbolic) link called targetpath. If it cannot be created | |
3048 | (platform limitation), we try to make a copy of the referenced file | |
3049 | instead of a link. | |
3050 | """ | |
be60ffd0 | 3051 | try: |
7584f5c9 ERE |
3052 | # For systems that support symbolic and hard links. |
3053 | if tarinfo.issym(): | |
7584f5c9 ERE |
3054 | os.symlink(tarinfo.linkname, targetpath) |
3055 | else: | |
3056 | # See extract(). | |
3057 | if os.path.exists(tarinfo._link_target): | |
7584f5c9 ERE |
3058 | os.link(tarinfo._link_target, targetpath) |
3059 | else: | |
be60ffd0 ERE |
3060 | self._extract_member(self._find_link_target(tarinfo), |
3061 | targetpath) | |
3062 | except symlink_exception: | |
7584f5c9 | 3063 | try: |
be60ffd0 ERE |
3064 | self._extract_member(self._find_link_target(tarinfo), |
3065 | targetpath) | |
7584f5c9 ERE |
3066 | except KeyError: |
3067 | raise ExtractError("unable to resolve link inside archive") | |
3068 | ||
3069 | def chown(self, tarinfo, targetpath): | |
3070 | """Set owner of targetpath according to tarinfo. | |
3071 | """ | |
3072 | if pwd and hasattr(os, "geteuid") and os.geteuid() == 0: | |
3073 | # We have to be root to do so. | |
3074 | try: | |
3075 | g = grp.getgrnam(tarinfo.gname)[2] | |
3076 | except KeyError: | |
3077 | g = tarinfo.gid | |
3078 | try: | |
3079 | u = pwd.getpwnam(tarinfo.uname)[2] | |
3080 | except KeyError: | |
3081 | u = tarinfo.uid | |
3082 | try: | |
3083 | if tarinfo.issym() and hasattr(os, "lchown"): | |
3084 | os.lchown(targetpath, u, g) | |
3085 | else: | |
be60ffd0 ERE |
3086 | os.chown(targetpath, u, g) |
3087 | except OSError as e: | |
7584f5c9 ERE |
3088 | raise ExtractError("could not change owner") |
3089 | ||
3090 | def chmod(self, tarinfo, targetpath): | |
3091 | """Set file permissions of targetpath according to tarinfo. | |
3092 | """ | |
3093 | if hasattr(os, 'chmod'): | |
3094 | try: | |
3095 | os.chmod(targetpath, tarinfo.mode) | |
be60ffd0 | 3096 | except OSError as e: |
7584f5c9 ERE |
3097 | raise ExtractError("could not change mode") |
3098 | ||
3099 | def utime(self, tarinfo, targetpath): | |
3100 | """Set modification time of targetpath according to tarinfo. | |
3101 | """ | |
3102 | if not hasattr(os, 'utime'): | |
3103 | return | |
3104 | try: | |
3105 | os.utime(targetpath, (tarinfo.mtime, tarinfo.mtime)) | |
be60ffd0 | 3106 | except OSError as e: |
7584f5c9 ERE |
3107 | raise ExtractError("could not change modification time") |
3108 | ||
3109 | #-------------------------------------------------------------------------- | |
3110 | def next(self): | |
3111 | """Return the next member of the archive as a TarInfo object, when | |
3112 | TarFile is opened for reading. Return None if there is no more | |
3113 | available. | |
3114 | """ | |
3115 | self._check("ra") | |
3116 | if self.firstmember is not None: | |
3117 | m = self.firstmember | |
3118 | self.firstmember = None | |
3119 | return m | |
3120 | ||
be60ffd0 ERE |
3121 | # Read the next block. |
3122 | self.fileobj.seek(self.offset) | |
7584f5c9 ERE |
3123 | tarinfo = None |
3124 | while True: | |
3125 | try: | |
3126 | tarinfo = self.tarinfo.fromtarfile(self) | |
be60ffd0 | 3127 | except EOFHeaderError as e: |
7584f5c9 ERE |
3128 | if self.ignore_zeros: |
3129 | self._dbg(2, "0x%X: %s" % (self.offset, e)) | |
3130 | self.offset += BLOCKSIZE | |
3131 | continue | |
be60ffd0 | 3132 | except InvalidHeaderError as e: |
7584f5c9 ERE |
3133 | if self.ignore_zeros: |
3134 | self._dbg(2, "0x%X: %s" % (self.offset, e)) | |
3135 | self.offset += BLOCKSIZE | |
3136 | continue | |
3137 | elif self.offset == 0: | |
3138 | raise ReadError(str(e)) | |
3139 | except EmptyHeaderError: | |
3140 | if self.offset == 0: | |
3141 | raise ReadError("empty file") | |
be60ffd0 | 3142 | except TruncatedHeaderError as e: |
7584f5c9 ERE |
3143 | if self.offset == 0: |
3144 | raise ReadError(str(e)) | |
be60ffd0 | 3145 | except SubsequentHeaderError as e: |
7584f5c9 ERE |
3146 | raise ReadError(str(e)) |
3147 | break | |
3148 | ||
3149 | if tarinfo is not None: | |
ea625b04 ERE |
3150 | if self.save_to_members: |
3151 | self.members.append(tarinfo) | |
7584f5c9 ERE |
3152 | else: |
3153 | self._loaded = True | |
3154 | ||
3155 | return tarinfo | |
3156 | ||
3157 | #-------------------------------------------------------------------------- | |
3158 | # Little helper methods: | |
3159 | ||
3160 | def _getmember(self, name, tarinfo=None, normalize=False): | |
3161 | """Find an archive member by name from bottom to top. | |
3162 | If tarinfo is given, it is used as the starting point. | |
3163 | """ | |
3164 | # Ensure that all members have been loaded. | |
3165 | members = self.getmembers() | |
3166 | ||
3167 | # Limit the member search list up to tarinfo. | |
3168 | if tarinfo is not None: | |
3169 | members = members[:members.index(tarinfo)] | |
3170 | ||
3171 | if normalize: | |
3172 | name = os.path.normpath(name) | |
3173 | ||
3174 | for member in reversed(members): | |
3175 | if normalize: | |
3176 | member_name = os.path.normpath(member.name) | |
3177 | else: | |
3178 | member_name = member.name | |
3179 | ||
3180 | if name == member_name: | |
3181 | return member | |
3182 | ||
3183 | def _load(self): | |
3184 | """Read through the entire archive file and look for readable | |
3185 | members. | |
3186 | """ | |
3187 | while True: | |
3188 | tarinfo = self.next() | |
3189 | if tarinfo is None: | |
3190 | break | |
3191 | self._loaded = True | |
3192 | ||
3193 | def _check(self, mode=None): | |
3194 | """Check if TarFile is still open, and if the operation's mode | |
3195 | corresponds to TarFile's mode. | |
3196 | """ | |
3197 | if self.closed: | |
be60ffd0 | 3198 | raise OSError("%s is closed" % self.__class__.__name__) |
7584f5c9 | 3199 | if mode is not None and self.mode not in mode: |
be60ffd0 | 3200 | raise OSError("bad operation for mode %r" % self.mode) |
7584f5c9 ERE |
3201 | |
3202 | def _find_link_target(self, tarinfo): | |
3203 | """Find the target member of a symlink or hardlink member in the | |
3204 | archive. | |
3205 | """ | |
3206 | if tarinfo.issym(): | |
3207 | # Always search the entire archive. | |
3208 | linkname = "/".join(filter(None, (os.path.dirname(tarinfo.name), tarinfo.linkname))) | |
3209 | limit = None | |
3210 | else: | |
3211 | # Search the archive before the link, because a hard link is | |
3212 | # just a reference to an already archived file. | |
3213 | linkname = tarinfo.linkname | |
3214 | limit = tarinfo | |
3215 | ||
3216 | member = self._getmember(linkname, tarinfo=limit, normalize=True) | |
3217 | if member is None: | |
3218 | raise KeyError("linkname %r not found" % linkname) | |
3219 | return member | |
3220 | ||
3221 | def __iter__(self): | |
3222 | """Provide an iterator object. | |
3223 | """ | |
3224 | if self._loaded: | |
3225 | return iter(self.members) | |
3226 | else: | |
3227 | return TarIter(self) | |
3228 | ||
1027433a | 3229 | def _dbg(self, level, msg, *args): |
7584f5c9 ERE |
3230 | """Write debugging output to sys.stderr. |
3231 | """ | |
3232 | if level <= self.debug: | |
1027433a | 3233 | print(msg.format(*args), file=sys.stderr) |
7584f5c9 ERE |
3234 | |
3235 | def __enter__(self): | |
3236 | self._check() | |
3237 | return self | |
3238 | ||
3239 | def __exit__(self, type, value, traceback): | |
3240 | if type is None: | |
3241 | self.close() | |
3242 | else: | |
3243 | # An exception occurred. We must not call close() because | |
3244 | # it would try to write end-of-archive blocks and padding. | |
3245 | if not self._extfileobj: | |
3246 | self.fileobj.close() | |
3247 | self.closed = True | |
3248 | # class TarFile | |
3249 | ||
3250 | class TarIter: | |
3251 | """Iterator Class. | |
3252 | ||
3253 | for tarinfo in TarFile(...): | |
3254 | suite... | |
3255 | """ | |
3256 | ||
3257 | def __init__(self, tarfile): | |
3258 | """Construct a TarIter object. | |
3259 | """ | |
3260 | self.tarfile = tarfile | |
3261 | self.index = 0 | |
3262 | def __iter__(self): | |
3263 | """Return iterator object. | |
3264 | """ | |
3265 | return self | |
be60ffd0 | 3266 | def __next__(self): |
7584f5c9 ERE |
3267 | """Return the next item using TarFile's next() method. |
3268 | When all members have been read, set TarFile as _loaded. | |
3269 | """ | |
3270 | # Fix for SF #1100429: Under rare circumstances it can | |
3271 | # happen that getmembers() is called during iteration, | |
3272 | # which will cause TarIter to stop prematurely. | |
3273 | ||
3274 | if self.index == 0 and self.tarfile.firstmember is not None: | |
3275 | tarinfo = self.tarfile.next() | |
3276 | elif self.index < len(self.tarfile.members): | |
3277 | tarinfo = self.tarfile.members[self.index] | |
3278 | elif not self.tarfile._loaded: | |
3279 | tarinfo = self.tarfile.next() | |
3280 | if not tarinfo: | |
3281 | self.tarfile._loaded = True | |
3282 | raise StopIteration | |
3283 | else: | |
3284 | raise StopIteration | |
3285 | self.index += 1 | |
fb27c6e8 | 3286 | |
7584f5c9 ERE |
3287 | return tarinfo |
3288 | ||
7584f5c9 ERE |
3289 | |
3290 | #-------------------- | |
3291 | # exported functions | |
3292 | #-------------------- | |
3293 | def is_tarfile(name): | |
3294 | """Return True if name points to a tar archive that we | |
3295 | are able to handle, else return False. | |
3296 | """ | |
3297 | try: | |
3298 | t = open(name) | |
3299 | t.close() | |
3300 | return True | |
3301 | except TarError: | |
3302 | return False | |
3303 | ||
3304 | bltn_open = open | |
3305 | open = TarFile.open |