1PAX(1) BSD General Commands Manual PAX(1)
2
4 pax — read and write file archives and copy directory hierarchies
5
7 pax [-0cdOnvz] [-f archive] [-s replstr] ... [-U user] ... [-G group] ...
8 [-T [from_date] [,to_date]] ... [pattern ...]
9 pax -r [-cdiknuvzDOYZ] [-f archive] [-o options] ... [-p string] ...
10 [-s replstr] ... [-E limit] [-U user] ... [-G group] ... [-T
11 [from_date] [,to_date]] ... [pattern ...]
12 pax -w [-0dituvzHLOPX] [-b blocksize] [[-a] [-f archive]] [-x format]
13 [-s replstr] ... [-o options] ... [-U user] ... [-G group] ...
14 [-B bytes] [-T [from_date] [,to_date] [/[c][m]]] ... [file ...]
15 pax -r -w [-0diklntuvDHLOPXYZ] [-p string] ... [-s replstr] ... [-U user]
16 ... [-G group] ... [-T [from_date] [,to_date] [/[c][m]]] ...
17 [file ...] directory
18
20 pax will read, write, and list the members of an archive file, and will
21 copy directory hierarchies. pax operation is independent of the specific
22 archive format, and supports a wide variety of different archive formats.
23 A list of supported archive formats can be found under the description of
24 the -x option.
25
26 The presence of the -r and the -w options specifies which of the follow‐
27 ing functional modes pax will operate under: list, read, write, and copy.
28
29 <none> List. pax will write to standard output a table of contents of
30 the members of the archive file read from standard input, whose
31 pathnames match the specified patterns. The table of contents
32 contains one filename per line and is written using single line
33 buffering.
34
35 -r Read. pax extracts the members of the archive file read from the
36 standard input, with pathnames matching the specified patterns.
37 The archive format and blocking is automatically determined on
38 input. When an extracted file is a directory, the entire file
39 hierarchy rooted at that directory is extracted. All extracted
40 files are created relative to the current file hierarchy. The
41 setting of ownership, access and modification times, and file
42 mode of the extracted files are discussed in more detail under
43 the -p option.
44
45 -w Write. pax writes an archive containing the file operands to
46 standard output using the specified archive format. When no file
47 operands are specified, a list of files to copy with one per line
48 is read from standard input. When a file operand is also a
49 directory, the entire file hierarchy rooted at that directory
50 will be included.
51
52 -r -w Copy. pax copies the file operands to the destination directory.
53 When no file operands are specified, a list of files to copy with
54 one per line is read from the standard input. When a file oper‐
55 and is also a directory the entire file hierarchy rooted at that
56 directory will be included. The effect of the copy is as if the
57 copied files were written to an archive file and then subse‐
58 quently extracted, except that there may be hard links between
59 the original and the copied files (see the -l option below).
60
61 Warning: The destination directory must not be one of the file
62 operands or a member of a file hierarchy rooted at one of the
63 file operands. The result of a copy under these conditions is
64 unpredictable.
65
66 While processing a damaged archive during a read or list operation, pax
67 will attempt to recover from media defects and will search through the
68 archive to locate and process the largest number of archive members pos‐
69 sible (see the -E option for more details on error handling).
70
71 The directory operand specifies a destination directory pathname. If the
72 directory operand does not exist, or it is not writable by the user, or
73 it is not of type directory, pax will exit with a non-zero exit status.
74
75 The pattern operand is used to select one or more pathnames of archive
76 members. Archive members are selected using the pattern matching nota‐
77 tion described by fnmatch(3). When the pattern operand is not supplied,
78 all members of the archive will be selected. When a pattern matches a
79 directory, the entire file hierarchy rooted at that directory will be
80 selected. When a pattern operand does not select at least one archive
81 member, pax will write these pattern operands in a diagnostic message to
82 standard error and then exit with a non-zero exit status.
83
84 The file operand specifies the pathname of a file to be copied or
85 archived. When a file operand does not select at least one archive mem‐
86 ber, pax will write these file operand pathnames in a diagnostic message
87 to standard error and then exit with a non-zero exit status.
88
89 The options are as follows:
90
91 -r Read an archive file from standard input and extract the speci‐
92 fied files. If any intermediate directories are needed in order
93 to extract an archive member, these directories will be created
94 as if mkdir(2) was called with the bitwise inclusive OR of
95 S_IRWXU, S_IRWXG, and S_IRWXO as the mode argument. When the
96 selected archive format supports the specification of linked
97 files and these files cannot be linked while the archive is being
98 extracted, pax will write a diagnostic message to standard error
99 and exit with a non-zero exit status at the completion of opera‐
100 tion.
101
102 -w Write files to the standard output in the specified archive for‐
103 mat. When no file operands are specified, standard input is read
104 for a list of pathnames with one per line without any leading or
105 trailing ⟨blanks⟩.
106
107 -a Append files to the end of an archive that was previously writ‐
108 ten. If an archive format is not specified with a -x option, the
109 format currently being used in the archive will be selected. Any
110 attempt to append to an archive in a format different from the
111 format already used in the archive will cause pax to exit immedi‐
112 ately with a non-zero exit status. The blocking size used in the
113 archive volume where writing starts will continue to be used for
114 the remainder of that archive volume.
115
116 Warning: Many storage devices are not able to support the opera‐
117 tions necessary to perform an append operation. Any attempt to
118 append to an archive stored on such a device may damage the ar‐
119 chive or have other unpredictable results. Tape drives in par‐
120 ticular are more likely to not support an append operation. An
121 archive stored in a regular file system file or on a disk device
122 will usually support an append operation.
123
124 -0 Use the NUL (‘\0’) character as a pathname terminator, instead of
125 newline (‘\n’). This applies only to the pathnames read from
126 standard input in the write and copy modes, and to the pathnames
127 written to standard output in list mode. This option is expected
128 to be used in concert with the -print0 function in find(1) or the
129 -0 flag in xargs(1).
130
131 -b blocksize
132 When writing an archive, block the output at a positive decimal
133 integer number of bytes per write to the archive file. The
134 blocksize must be a multiple of 512 bytes with a maximum of 64512
135 bytes. Archives larger than 32256 bytes violate the POSIX stan‐
136 dard and will not be portable to all systems. A blocksize can
137 end with ‘k’ or ‘b’ to specify multiplication by 1024 (1K) or
138 512, respectively. A pair of blocksizes can be separated by ‘x’
139 to indicate a product. A specific archive device may impose
140 additional restrictions on the size of blocking it will support.
141 When blocking is not specified, the default blocksize is depen‐
142 dent on the specific archive format being used (see the -x
143 option).
144
145 -c Match all file or archive members except those specified by the
146 pattern and file operands.
147
148 -d Cause files of type directory being copied or archived, or ar‐
149 chive members of type directory being extracted, to match only
150 the directory file or archive member and not the file hierarchy
151 rooted at the directory.
152
153 -f archive
154 Specify archive as the pathname of the input or output archive,
155 overriding the default standard input (for list and read) or
156 standard output (for write). A single archive may span multiple
157 files and different archive devices. When required, pax will
158 prompt for the pathname of the file or device of the next volume
159 in the archive.
160
161 -i Interactively rename files or archive members. For each archive
162 member matching a pattern operand or each file matching a file
163 operand, pax will prompt to /dev/tty giving the name of the file,
164 its file mode, and its modification time. pax will then read a
165 line from /dev/tty. If this line is blank, the file or archive
166 member is skipped. If this line consists of a single period, the
167 file or archive member is processed with no modification to its
168 name. Otherwise, its name is replaced with the contents of the
169 line. pax will immediately exit with a non-zero exit status if
170 EOF is encountered when reading a response or if /dev/tty cannot
171 be opened for reading and writing.
172
173 -k Do not overwrite existing files.
174
175 -l (The lowercase letter “ell.”) Link files. In the copy mode (-r
176 -w), hard links are made between the source and destination file
177 hierarchies whenever possible.
178
179 -n Select the first archive member that matches each pattern oper‐
180 and. No more than one archive member is matched for each
181 pattern. When members of type directory are matched, the file
182 hierarchy rooted at that directory is also matched (unless -d is
183 also specified).
184
185 -o options
186 Information to modify the algorithm for extracting or writing ar‐
187 chive files which is specific to the archive format specified by
188 -x. In general, options take the form: name=value.
189
190 -p string
191 Specify one or more file characteristic options (privileges).
192 The string option-argument is a string specifying file character‐
193 istics to be retained or discarded on extraction. The string
194 consists of the specification characters a, e, m, o, and p. Mul‐
195 tiple characteristics can be concatenated within the same string
196 and multiple -p options can be specified. The meaning of the
197 specification characters are as follows:
198
199 a Do not preserve file access times. By default, file access
200 times are preserved whenever possible.
201
202 e ‘Preserve everything’, the user ID, group ID, file mode bits,
203 file access time, and file modification time. This is
204 intended to be used by root, someone with all the appropriate
205 privileges, in order to preserve all aspects of the files as
206 they are recorded in the archive. The e flag is the sum of
207 the o and p flags.
208
209 m Do not preserve file modification times. By default, file
210 modification times are preserved whenever possible.
211
212 o Preserve the user ID and group ID.
213
214 p ‘Preserve’ the file mode bits. This is intended to be used
215 by a user with regular privileges who wants to preserve all
216 aspects of the file other than the ownership. The file times
217 are preserved by default, but two other flags are offered to
218 disable this and use the time of extraction instead.
219
220 In the preceding list, ‘preserve’ indicates that an attribute
221 stored in the archive is given to the extracted file, subject to
222 the permissions of the invoking process. Otherwise the attribute
223 of the extracted file is determined as part of the normal file
224 creation action. If neither the e nor the o specification char‐
225 acter is specified, or the user ID and group ID are not preserved
226 for any reason, pax will not set the S_ISUID (setuid) and S_ISGID
227 (setgid) bits of the file mode. If the preservation of any of
228 these items fails for any reason, pax will write a diagnostic
229 message to standard error. Failure to preserve these items will
230 affect the final exit status, but will not cause the extracted
231 file to be deleted. If the file characteristic letters in any of
232 the string option-arguments are duplicated or conflict with each
233 other, the one(s) given last will take precedence. For example,
234 if
235 -p eme
236 is specified, file modification times are still preserved.
237
238 -s replstr
239 Modify the file or archive member names specified by the pattern
240 or file operands according to the substitution expression
241 replstr, using the syntax of the ed(1) utility regular expres‐
242 sions. The format of these regular expressions are:
243 /old/new/[gp]
244 As in ed(1), old is a basic regular expression and new can con‐
245 tain an ampersand (‘&’), ‘\n’ (where n is a digit) back-refer‐
246 ences, or subexpression matching. The old string may also con‐
247 tain newline characters. Any non-null character can be used as a
248 delimiter (‘/’ is shown here). Multiple -s expressions can be
249 specified. The expressions are applied in the order they are
250 specified on the command line, terminating with the first suc‐
251 cessful substitution. The optional trailing g continues to apply
252 the substitution expression to the pathname substring which
253 starts with the first character following the end of the last
254 successful substitution. The first unsuccessful substitution
255 stops the operation of the g option. The optional trailing p
256 will cause the final result of a successful substitution to be
257 written to standard error in the following format:
258 <original pathname> >> <new pathname>
259 File or archive member names that substitute to the empty string
260 are not selected and will be skipped.
261
262 -t Reset the access times of any file or directory read or accessed
263 by pax to be the same as they were before being read or accessed
264 by pax.
265
266 -u Ignore files that are older (having a less recent file modifica‐
267 tion time) than a pre-existing file or archive member with the
268 same name. During read, an archive member with the same name as
269 a file in the file system will be extracted if the archive member
270 is newer than the file. During write, a file system member with
271 the same name as an archive member will be written to the archive
272 if it is newer than the archive member. During copy, the file in
273 the destination hierarchy is replaced by the file in the source
274 hierarchy or by a link to the file in the source hierarchy if the
275 file in the source hierarchy is newer.
276
277 -v During a list operation, produce a verbose table of contents
278 using the format of the ls(1) utility with the -l option. For
279 pathnames representing a hard link to a previous member of the
280 archive, the output has the format:
281 <ls -l listing> == <link name>
282 For pathnames representing a symbolic link, the output has the
283 format:
284 <ls -l listing> => <link name>
285 Where <ls -l listing> is the output format specified by the ls(1)
286 utility when used with the -l option. Otherwise for all the
287 other operational modes (read, write, and copy), pathnames are
288 written and flushed to standard error without a trailing newline
289 as soon as processing begins on that file or archive member. The
290 trailing newline is not buffered and is written only after the
291 file has been read or written.
292
293 -x format
294 Specify the output archive format, with the default format being
295 ustar. pax currently supports the following formats:
296
297 cpio The extended cpio interchange format specified in the
298 IEEE Std 1003.2 (“POSIX.2”) standard. The default
299 blocksize for this format is 5120 bytes. Inode and
300 device information about a file (used for detecting file
301 hard links by this format) which may be truncated by
302 this format is detected by pax and is repaired.
303
304 bcpio The old binary cpio format. The default blocksize for
305 this format is 5120 bytes. This format is not very por‐
306 table and should not be used when other formats are
307 available. Inode and device information about a file
308 (used for detecting file hard links by this format)
309 which may be truncated by this format is detected by pax
310 and is repaired.
311
312 sv4cpio The System V release 4 cpio. The default blocksize for
313 this format is 5120 bytes. Inode and device information
314 about a file (used for detecting file hard links by this
315 format) which may be truncated by this format is
316 detected by pax and is repaired.
317
318 sv4crc The System V release 4 cpio with file crc checksums.
319 The default blocksize for this format is 5120 bytes.
320 Inode and device information about a file (used for
321 detecting file hard links by this format) which may be
322 truncated by this format is detected by pax and is
323 repaired.
324
325 tar The old BSD tar format as found in BSD4.3. The default
326 blocksize for this format is 10240 bytes. Pathnames
327 stored by this format must be 100 characters or less in
328 length (including the trailing character, which means
329 that filenames can have a maximum length of 99 charac‐
330 ters). Only regular files, hard links, soft links, and
331 directories will be archived (other file system types
332 are not supported). For backwards compatibility with
333 even older tar formats, a -o option can be used when
334 writing an archive to omit the storage of directories.
335 This option takes the form:
336 -o write_opt=nodir
337
338 ustar The extended tar interchange format specified in the
339 IEEE Std 1003.2 (“POSIX.2”) standard. The default
340 blocksize for this format is 10240 bytes. Filenames
341 stored by this format must be 100 characters or less in
342 length (including the trailing character, which means
343 that filenames can have a maximum length of 99 charac‐
344 ters). Pathnames (directorynames + filenames) stored by
345 this format must be 250 characters or less in length.
346
347 pax will detect and report any file that it is unable to store or
348 extract as the result of any specific archive format restric‐
349 tions. The individual archive formats may impose additional
350 restrictions on use. Typical archive format restrictions include
351 (but are not limited to): file pathname length, file size, link
352 pathname length, and the type of the file.
353
354 -z Use gzip(1) to compress (decompress) the archive while writing
355 (reading). Incompatible with -a.
356
357 -B bytes
358 Limit the number of bytes written to a single archive volume to
359 bytes. The bytes limit can end with ‘m’, ‘k’, or ‘b’ to specify
360 multiplication by 1048576 (1M), 1024 (1K) or 512, respectively.
361 A pair of bytes limits can be separated by ‘x’ to indicate a
362 product.
363
364 Warning: Only use this option when writing an archive to a device
365 which supports an end of file read condition based on last (or
366 largest) write offset (such as a regular file or a tape drive).
367 The use of this option with a floppy or hard disk is not recom‐
368 mended.
369
370 -D This option is the same as the -u option, except that the file
371 inode change time is checked instead of the file modification
372 time. The file inode change time can be used to select files
373 whose inode information (e.g., UID, GID, etc.) is newer than a
374 copy of the file in the destination directory.
375
376 -E limit
377 Limit the number of consecutive read faults while trying to read
378 a flawed archive to limit. With a positive limit, pax will
379 attempt to recover from an archive read error and will continue
380 processing starting with the next file stored in the archive. A
381 limit of 0 will cause pax to stop operation after the first read
382 error is detected on an archive volume. A limit of NONE will
383 cause pax to attempt to recover from read errors forever. The
384 default limit is a small positive number of retries.
385
386 Warning: Using this option with NONE should be used with extreme
387 caution as pax may get stuck in an infinite loop on a very badly
388 flawed archive.
389
390 -G group
391 Select a file based on its group name, or when starting with a #,
392 a numeric gid. A ‘\’ can be used to escape the #. Multiple -G
393 options may be supplied and checking stops with the first match.
394
395 -H Follow only command-line symbolic links while performing a physi‐
396 cal file system traversal.
397
398 -L Follow all symbolic links to perform a logical file system tra‐
399 versal.
400
401 -O Force the archive to be one volume. If a volume ends prema‐
402 turely, pax will not prompt for a new volume. This option can be
403 useful for automated tasks where error recovery cannot be per‐
404 formed by a human.
405
406 -P Do not follow symbolic links, perform a physical file system tra‐
407 versal. This is the default mode.
408
409 -T [from_date][,to_date][/[c][m]]
410 Allow files to be selected based on a file modification or inode
411 change time falling within a specified time range of from_date to
412 to_date (the dates are inclusive). If only a from_date is sup‐
413 plied, all files with a modification or inode change time equal
414 to or younger are selected. If only a to_date is supplied, all
415 files with a modification or inode change time equal to or older
416 will be selected. When the from_date is equal to the to_date,
417 only files with a modification or inode change time of exactly
418 that time will be selected.
419
420 When pax is in the write or copy mode, the optional trailing
421 field [c][m] can be used to determine which file time (inode
422 change, file modification or both) are used in the comparison.
423 If neither is specified, the default is to use file modification
424 time only. The m specifies the comparison of file modification
425 time (the time when the file was last written). The c specifies
426 the comparison of inode change time (the time when the file inode
427 was last changed; e.g., a change of owner, group, mode, etc).
428 When c and m are both specified, then the modification and inode
429 change times are both compared. The inode change time comparison
430 is useful in selecting files whose attributes were recently
431 changed or selecting files which were recently created and had
432 their modification time reset to an older time (as what happens
433 when a file is extracted from an archive and the modification
434 time is preserved). Time comparisons using both file times is
435 useful when pax is used to create a time based incremental ar‐
436 chive (only files that were changed during a specified time range
437 will be archived).
438
439 A time range is made up of six different fields and each field
440 must contain two digits. The format is:
441 [[[[[cc]yy]mm]dd]HH]MM[.SS]
442 Where cc is the first two digits of the year (the century), yy is
443 the last two digits of the year, the first mm is the month (from
444 01 to 12), dd is the day of the month (from 01 to 31), HH is the
445 hour of the day (from 00 to 23), MM is the minute (from 00 to
446 59), and SS is the seconds (from 00 to 59). The minute field MM
447 is required, while the other fields are optional and must be
448 added in the following order:
449 HH, dd, mm, yy, cc.
450 The SS field may be added independently of the other fields.
451 Time ranges are relative to the current time, so
452 -T 1234/cm
453 would select all files with a modification or inode change time
454 of 12:34 PM today or later. Multiple -T time range can be sup‐
455 plied and checking stops with the first match.
456
457 -U user
458 Select a file based on its user name, or when starting with a #,
459 a numeric UID. A ‘\’ can be used to escape the #. Multiple -U
460 options may be supplied and checking stops with the first match.
461
462 -X When traversing the file hierarchy specified by a pathname, do
463 not descend into directories that have a different device ID.
464 See the st_dev field as described in stat(2) for more information
465 about device IDs.
466
467 -Y This option is the same as the -D option, except that the inode
468 change time is checked using the pathname created after all the
469 file name modifications have completed.
470
471 -Z This option is the same as the -u option, except that the modifi‐
472 cation time is checked using the pathname created after all the
473 file name modifications have completed.
474
475 The options that operate on the names of files or archive members (-c,
476 -i, -n, -s, -u, -v, -D, -G, -T, -U, -Y, and -Z) interact as follows.
477
478 When extracting files during a read operation, archive members are
479 ‘selected’, based only on the user specified pattern operands as modified
480 by the -c, -n, -u, -D, -G, -T, -U options. Then any -s and -i options
481 will modify in that order, the names of these selected files. Then the
482 -Y and -Z options will be applied based on the final pathname. Finally,
483 the -v option will write the names resulting from these modifications.
484
485 When archiving files during a write operation, or copying files during a
486 copy operation, archive members are ‘selected’, based only on the user
487 specified pathnames as modified by the -n, -u, -D, -G, -T, and -U options
488 (the -D option only applies during a copy operation). Then any -s and -i
489 options will modify in that order, the names of these selected files.
490 Then during a copy operation the -Y and the -Z options will be applied
491 based on the final pathname. Finally, the -v option will write the names
492 resulting from these modifications.
493
494 When one or both of the -u or -D options are specified along with the -n
495 option, a file is not considered selected unless it is newer than the
496 file to which it is compared.
497
499 TMPDIR Path in which to store temporary files.
500
502 $ pax -w -f /dev/rst0 .
503
504 Copies the contents of the current directory to the device /dev/rst0.
505
506 $ pax -v -f filename
507
508 Gives the verbose table of contents for an archive stored in filename.
509
510 $ mkdir newdir; cd olddir; pax -rw . newdir
511
512 This sequence of commands will copy the entire olddir directory hierarchy
513 to newdir.
514
515 $ pax -r -s ',^//*usr//*,,' -f a.pax
516
517 Reads the archive a.pax, with all files rooted in /usr into the archive
518 extracted relative to the current directory.
519
520 $ pax -rw -i . dest_dir
521
522 Can be used to interactively select the files to copy from the current
523 directory to dest_dir.
524
525 $ pax -r -pe -U root -G bin -f a.pax
526
527 Extract all files from the archive a.pax which are owned by root with
528 group bin and preserve all file permissions.
529
530 $ pax -r -w -v -Y -Z home /backup
531
532 Update (and list) only those files in the destination directory /backup
533 which are older (less recent inode change or file modification times)
534 than files with the same name found in the source file tree home.
535
537 pax will exit with one of the following values:
538
539 0 All files were processed successfully.
540
541 1 An error occurred.
542
543 Whenever pax cannot create a file or a link when reading an archive or
544 cannot find a file when writing an archive, or cannot preserve the user
545 ID, group ID, or file mode when the -p option is specified, a diagnostic
546 message is written to standard error and a non-zero exit status will be
547 returned, but processing will continue. In the case where pax cannot
548 create a link to a file, pax will not create a second copy of the file.
549
550 If the extraction of a file from an archive is prematurely terminated by
551 a signal or error, pax may have only partially extracted a file the user
552 wanted. Additionally, the file modes of extracted files and directories
553 may have incorrect file bits, and the modification and access times may
554 be wrong.
555
556 If the creation of an archive is prematurely terminated by a signal or
557 error, pax may have only partially created the archive which may violate
558 the specific archive format specification.
559
560 If while doing a copy, pax detects a file is about to overwrite itself,
561 the file is not copied, a diagnostic message is written to standard error
562 and when pax completes it will exit with a non-zero exit status.
563
565 cpio(1), tar(1)
566
568 The pax utility is a superset of the IEEE Std 1003.2 (“POSIX.2”) stan‐
569 dard. The options -B, -D, -E, -G, -H, -L, -O, -P, -T, -U, -Y, -Z, the
570 archive formats bcpio, sv4cpio, sv4crc, tar, and the flawed archive han‐
571 dling during list and read operations are extensions to the POSIX stan‐
572 dard.
573
575 Keith Muller at the University of California, San Diego.
576
577BSD April 18, 1994 BSD