Version 1.34

Related discussion in the Fedora pull-request:
https://src.fedoraproject.org/rpms/tar/pull-request/8

* tests/capabs_raw01.at: Newer systems (currently e.g. Fedora 34)
print getcap output in format CAP=VAL, not CAP+VAL.

.gitignore

@@ -8,6 +8,7 @@
 .deps
 .emacs*
 .libs
+.gdbinit
 ABOUT-NLS
 ChangeLog
 INSTALL
@@ -23,13 +24,9 @@ config.h.in
 config.log
 config.status
 configure
-gnu/*.h
-gnu/*/
-gnu/.gitignore
-gnu/charset.alias
-gnulib
+gnu
 libtool
 m4
-paxutils
 rmt
 stamp-h1
+TAGS

.gitmodules

@@ -0,0 +1,6 @@
+[submodule "gnulib"]
+	path = gnulib
+	url = git://git.sv.gnu.org/gnulib.git
+[submodule "paxutils"]
+	path = paxutils
+	url = git://git.sv.gnu.org/paxutils.git

ChangeLog.1

@@ -3186,7 +3186,7 @@ along with this program.  If not, see <http://www.gnu.org/licenses/>.
 	Reported by Jeffrey Goldberg.
 
 	* rmt.h (_remdev): A filename is not remote if the colon is
-	preceeded by a slash, to take care of `/:/' which is a shorthand
+	preceded by a slash, to take care of `/:/' which is a shorthand
 	for `/.../<CELL-NAME>/fs' on OSF's Distributing Computing
 	Environment (DCE) and Distributed File System (DFS).
 	Reported by Travis L. Priest.

Makefile.am

@@ -1,7 +1,6 @@
 # Main Makefile for GNU tar.
 
-# Copyright 1994-1997, 1999-2001, 2003, 2007, 2009, 2013-2014 Free
-# Software Foundation, Inc.
+# Copyright 1994-2021 Free Software Foundation, Inc.
 
 # This file is part of GNU tar.
 
@@ -26,7 +25,7 @@ dist-hook:
 	$(MAKE) changelog_dir=$(distdir) ChangeLog
 	-rm -f $(distdir).cpio
 	find $(distdir) | cpio -Hcrc -o | \
-          GZIP=$(GZIP_ENV) gzip -c > $(distdir).cpio.gz
+          eval GZIP= gzip $(GZIP_ENV) -c > $(distdir).cpio.gz
 
 distclean-local:
 	-rm -f $(distdir).cpio.gz

NEWS

@@ -1,5 +1,234 @@
-GNU tar NEWS - User visible changes. 2014-07-27
+GNU tar NEWS - User visible changes. 2021-02-13
 Please send GNU tar bug reports to <bug-tar@gnu.org>
+
+version 1.34 - Sergey Poznyakoff, 2021-02-13
+
+* Fix extraction over pipe (savannah bug #60002)
+
+* Fix memory leak in read_header (savannah bug #59897)
+
+* Fix extraction when . and .. are unreadable
+
+See https://lists.gnu.org/archive/html/bug-tar/2021-01/msg00012.html
+
+* Gracefully handle duplicate symlinks when extracting
+
+See https://lists.gnu.org/archive/html/bug-tar/2021-01/msg00026.html
+
+* Re-initialize supplementary groups when switching to user privileges
+
+version 1.33 - Sergey Poznyakoff, 2021-01-07
+
+* POSIX extended format headers do not include PID by default
+    
+The intent is to make binary-equivalent PAX archives easy to create.  If
+POSIXLY_CORRECT is set, the POSIX standard default is used, which embeds
+the pid.
+
+* --delay-directory-restore works for archives with reversed member ordering
+
+* Fix extraction of a symbolic link hardlinked to another symbolic link
+
+* Wildcards in exclude-vcs-ignore mode don't match slash
+
+* Fix the --no-overwrite-dir option
+    
+Given this option, previous versions of tar failed to preserve
+permissions of empty directories and to create files under directories
+owned by the current user that did not have the S_IWUSR bit set.
+
+* Fix handling of chained renames in incremental backups
+
+* Link counting works for file names supplied with -T
+
+* Accept only position-sensitive (file-selection) options in file list files.
+    
+Using such options as -f, -z, etc. is senseless in a file list file and
+bypasses option consistency checks in decode_options.  Therefore, 
+only options related to file selection (a.k.a position-sensitive options)
+are allowed in file list files.
+
+
+version 1.32 - Sergey Poznyakoff, 2019-02-23
+
+* Fix the use of --checkpoint without explicit --checkpoint-action
+
+* Fix extraction with the -U option
+
+See http://lists.gnu.org/archive/html/bug-tar/2019-01/msg00015.html,
+for details
+
+* Fix iconv usage on BSD-based systems
+
+* Fix possible NULL dereference (savannah bug #55369)
+
+* Improve the testsuite
+
+
+version 1.31 - Sergey Poznyakoff, 2019-01-02
+
+* Fix heap-buffer-overrun with --one-top-level.
+Bug introduced with the addition of that option in 1.28.
+
+* Support for zstd compression
+
+New option '--zstd' instructs tar to use zstd as compression program.
+When listing, extractng and comparing, zstd compressed archives are
+recognized automatically.
+When '-a' option is in effect, zstd compression is selected if the
+destination archive name ends in '.zst' or '.tzst'.
+
+* The -K option interacts properly with member names given in the command line
+
+Names of members to extract can be specified along with the "-K NAME"
+option. In this case, tar will extract NAME and those of named members
+that appear in the archive after it, which is consistent with the
+semantics of the option.
+
+Previous versions of tar extracted NAME, those of named members that
+appeared before it, and everything after it.
+
+* Fix CVE-2018-20482
+
+When creating archives with the --sparse option, previous versions of
+tar would loop endlessly if a sparse file had been truncated while
+being archived.
+
+
+version 1.30 - Sergey Poznyakoff, 2017-12-17
+
+* Member names containing '..' components are now skipped when extracting.
+
+This fixes tar's behavior to match its documentation, and is a bit
+safer when extracting untrusted archives over old files (an unsafe
+practice that the tar manual has long recommended against).
+
+* Report erroneous use of position-sensitive options.
+
+During archive creation or update, tar keeps track of positional
+options (see the manual, subsection 3.4.4 "Position-Sensitive
+Options"), and reports those that had no effect.  For example, when
+invoked as
+
+   tar -cf a.tar . --exclude '*.o'
+
+tar will create the archive, but will exit with status 2, having
+issued the following error message
+
+   tar: The following options were used after non-optional
+   arguments in archive create or update mode.  These options are
+   positional and affect only arguments that follow them.  Please,
+   rearrange them properly.
+   tar: --exclude '*.o' has no effect
+   tar: Exiting with failure status due to previous errors
+
+* --numeric-owner now affects private headers too.
+
+This helps the output of 'tar' to be more deterministic.
+
+* Fixed the --delay-directory-restore option
+
+In some cases tar would restore the directory permissions too early,
+causing subsequent link extractions in that directory to fail.
+
+* The --warnings=failed-read option
+
+This new warning control option suppresses warning messages about
+unreadable files and directories. It has effect only if used together
+with the --ignore-failed-read option.
+
+* The --warnings=none option now suppresses all warnings
+
+This includes warnings about unreadable files produced when
+--ignore-failed-read is in effect. To output these, use
+--warnings=none --warnings=no-failed-read.
+
+* Fix reporting of hardlink mismatches during compare
+
+Tar reported incorrect target file name in the 'Not linked to'
+diagnostic message.
+
+
+version 1.29 - Sergey Poznyakoff, 2016-05-16
+
+* New options: --verbatim-files-from, --no-verbatim-files-from
+
+The --verbatim-files-from option instructs tar to treat each line read
+from a file list as a file name, even if it starts with a dash.
+
+File lists are supplied with the --files-from (-T) option.  By
+default, each line read from a file list is first stripped off the
+leading and trailing whitespace and, if the result begins with a dash,
+it is treated as tar command line option.
+
+Use the --verbatim-files-from option to disable this special handling.
+This facilitates the use of tar with file lists created automatically
+(e.g. by find(1) command).
+
+This option affects all --files-from options that occur after it in
+the command line.  Its effect is reverted by the
+--no-verbatim-files-from option.
+
+* --null option reads file names verbatim
+
+The --null option implies --verbatim-files-from.  I.e. each line
+read from null-delimited file lists is treated as a file name.
+
+This restores the documented behavior, which was broken in version
+1.27.
+
+* New options: --owner-map=FILE and --group-map=FILE
+
+These two options provide fine-grained control over what user/group
+names (or IDs) should be mapped when adding files to archive.
+
+For both options, FILE is a plain text file with user or group
+mappings.  Empty lines are ignored.  Comments are introduced with
+# sign (unless quoted) and extend to the end of the corresponding
+line.  Each non-empty line defines translation for a single UID (GID).
+It must consist of two fields, delimited by any amount of whitespace:
+
+     OLDNAME NEWNAME[:NEWID]
+
+OLDNAME is either a valid user (group) name or a ID prefixed with +.  Unless
+NEWID is supplied, NEWNAME must also be either a valid name or a
++ID.  Otherwise, both NEWNAME and NEWID need not be listed in the
+system user database.
+
+* New option --clamp-mtime
+
+The new --clamp-mtime option changes the behavior of --mtime to only
+use the time specified if the file mtime is newer than the given time.
+The --clamp-mtime option can only be used together with	--mtime.
+
+Typical use case is to make builds reproducible: to loose less
+information, it's better to keep the original date of an archive,
+except for files modified during the build process. In that case, using
+reference (and thus reproducible) timestamps for the latter is good
+enough.
+
+See <https://wiki.debian.org/ReproducibleBuilds> for more information.
+
+* Deprecated --preserve option removed
+
+* Sparse file detection
+
+Tar now uses SEEK_DATA/SEEK_HOLE on systems that support it.  This
+allows for considerable speed-up in sparse-file detection.
+
+New option --hole-detection is provided, that allows the user to
+select the algorithm used for hole detection.  Available arguments
+are:
+
+  --hole-detection=seek
+     Use lseek(2) SEEK_DATA and SEEK_HOLE "whence" parameters.
+
+  --hole-detection=raw
+     Scan entire file before storing it to determine where holes
+     are located.
+
+The default is to use "seek" whenever possible, and fall back to
+"raw" otherwise.
 
 
 version 1.28, 2014-07-28
@@ -21,7 +250,7 @@ sign and the specifier letter.
 	       deleted, correspondingly.
   %{FMT}t   -  Current local time using FMT as strftime(3) format.
                If {FMT} is omitted, use %c.
-  %{N}*     -  Pad output with spaces to the Nth column, or to the 
+  %{N}*     -  Pad output with spaces to the Nth column, or to the
                current screen width, if {N} is not given.
   %c        -  A shortcut for "%{%Y-%m-%d %H:%M:%S}t: %ds, %{read,wrote}T%*\r"
 
@@ -60,7 +289,7 @@ speed up archivation.
   --exclude-vcs-ignores   Read exclude tags from VCS ignore files,
                           where such files exist.  Supported VCS's
                           are: CVS, Git, Bazaar, Mercurial.
-			  
+
 
 * Tar refuses to read input from and write output to a tty device.
 
@@ -69,7 +298,7 @@ speed up archivation.
 This release includes official tar(1) and rmt(8) manpages.
 Distribution maintainers are kindly asked to use these instead of the
 home-made pages they have been providing so far.
-  
+
 
 version 1.27.1 - Sergey Poznyakoff, 2013-11-17
 
@@ -162,7 +391,7 @@ tar discovers that the corresponding file name already exists and is a
 symbolic link, it first unlinks the entry, and then extracts the directory.
 
 This option disables this behavior and instructs tar to follow
-symlinks to directories when extracting from the archive.  
+symlinks to directories when extracting from the archive.
 
 It is mainly intended to provide compatibility with the Slackware
 installation scripts.
@@ -369,7 +598,7 @@ Modification times in ustar header blocks of extended headers
 are set to mtimes of the corresponding archive members.  This
 can be overridden by the
 
-  --pax-opion='exthdr.mtime=STRING'
+  --pax-option='exthdr.mtime=STRING'
 
 command line option.  The STRING is either number of seconds since
 the Epoch or a "Time reference" (see below).
@@ -379,7 +608,7 @@ headers are set to the time when tar was invoked.
 
 This can be overridden by the
 
-  --pax-opion='globexthdr.mtime=STRING'
+  --pax-option='globexthdr.mtime=STRING'
 
 command line option.  The STRING is either number of seconds since
 the Epoch or a "Time reference" (see below).
@@ -1505,7 +1734,7 @@ Versions 1.07 back to 1.00 by Jay Fenlason.
 
 
 
-Copyright 1994-2001, 2003-2010, 2013-2014 Free Software Foundation, Inc.
+Copyright 1994-2021 Free Software Foundation, Inc.
 
 This file is part of GNU tar.
 

README

@@ -23,7 +23,6 @@ GNU 'tar' is derived from John Gilmore's public domain 'tar'.
 See file 'ABOUT-NLS' for how to customize this program to your language.
 See file 'COPYING' for copying conditions.
 See file 'INSTALL' for compilation and installation instructions.
-See file 'PORTS' for various ports of GNU tar to non-Unix systems.
 See file 'NEWS' for a list of major changes in the current release.
 See file 'THANKS' for a list of contributors.
 
@@ -222,8 +221,7 @@ and share your findings by writing to <bug-tar@gnu.org>.
 
 * Copying
 
-Copyright 1990-1992, 1994, 1997-2001, 2003-2004, 2007, 2012-2014 Free
-Software Foundation, Inc.
+Copyright 1990-2021 Free Software Foundation, Inc.
 
 This file is part of GNU tar.
 

README-alpha

@@ -23,47 +23,19 @@ suggest using test version 1.3.5 (or later, if one becomes available).
 Valgrind <http://valgrind.org/> is also highly recommended, if
 Valgrind supports your architecture.
 
-Before building the package, run "bootstrap".  It obtains various
-additional files from the CVS repository and the Translation Project
-site and prepares the source directory for building.
+Before building the package, run "bootstrap".  It will obtain gnulib
+and paxutils files from their Git repositories on Savannah.  Then, it will
+fetch the po files from tar page at Translation Project, and, finally, it
+will start autoconfiguration process.  Simply running bootstrap without
+arguments should do in most cases.
 
-When run without arguments, bootstrap will try to obtain gnulib and
-paxutils files from their corresponding CVS repositories on Savannah
-using anonymous SSH access. Then, it will fetch the po files from tar
-page at Translation Project, and, finally, it will start autoconfiguration
-process. Simply running it without arguments should do in most cases.
-Several options allow to control the behavior of bootstrap:
-
- --gnulib-srcdir=DIRNAME      Specify the local directory where gnulib
-                              sources reside. Use this if you already
-                              have gnulib sources on your machine, and
-                              do not want to waste your bandwidth dowloading
-                              them again.
-
- --paxutils-srcdir=DIRNAME    Specify the local directory where paxutils
-                              sources reside. Use this if you already
-                              have paxutils sources on your machine, and
-                              do not want to waste your bandwidth dowloading
-                              them again.
-
- --cvs-auth=METHOD            Set the CVS access method used for downloading
-                              gnulib files. METHOD is one of the keywords
-                              accepted by cvs -d option (see info cvs
-                              repository).
-
- --cvs-user=USERNAME          Set the CVS username to be used when accessing
-                              the gnulib repository.
-
- --no-po                      Do not download po files.
-
-Notice also that when using CVS authentication method "ext", bootstrap
-will set the variable CVS_RSH to "ssh", unless it is already set to
-some other value.
+Bootstrap reads its configuration from file bootstrap.conf located on the
+top of tar source tree.  Several options are provided that modify its
+behavior.  Run 'bootstrap --help' for a list.
 
 
 
-Copyright 2001, 2003-2005, 2007, 2013-2014 Free Software Foundation,
-Inc.
+Copyright 2001-2021 Free Software Foundation, Inc.
 
 This file is part of GNU tar.
 

README-hacking

@@ -12,7 +12,6 @@ version.
 - Autoconf <http://www.gnu.org/software/autoconf/>
 - M4 <http://www.gnu.org/software/m4/>
 - Texinfo <http://www.gnu.org/software/texinfo>
-- Gnulib <http://www.gnu.org/software/gnulib>
 - Git <http://git.or.cz>
 
 * Bootstrapping
@@ -33,23 +32,14 @@ Once done, proceed as described in the file README (section
 INSTALLATION).
 
 Normally you will have to run bootstrap only once. However, if you
-intend to hack on GNU tar, you might need to run it again later. In
-this case, you will probably want to save some time and bandwidth by
-avoiding downloading the same files again. If so, create in GNU tar
-root directory the file named '.bootstrap' with the following
-contents:
-
-   --gnulib-srcdir=$HOME/gnulib
-
-Replace '$HOME/gnulib' with the actual directory where the Gnulib
-sources reside.
-
-For more information about 'bootstrap', run 'bootstrap --help'.
+intend to hack on GNU tar, you might need to run it again later.
+There are lots of options that you may find useful in this case.
+See 'bootstrap --help' for a detailed list.
 
 
 * Copyright information
 
-Copyright 2007-2009, 2013-2014 Free Software Foundation, Inc.
+Copyright 2007-2021 Free Software Foundation, Inc.
 
 This file is part of GNU tar.
 

THANKS

@@ -8,6 +8,7 @@ list of these people.  Help me keep it complete and exempt of errors.
 See various ChangeLogs for a detailed description of contributions.
 
 Aage Robeck		aagero@ifi.uio.no
+Adam Borowski		kilobyte@angband.pl
 Adye, TJ (Tim)		T.J.Adye@rl.ac.uk
 Akiko Matsushita	matusita@sra.co.jp
 Alan Bawden		Alan@lcs.mit.edu
@@ -78,6 +79,7 @@ Cesar Romani		romani@ifm.uni-hamburg.de
 Chad Hurwitz		churritz@cts.com
 Chance Reschke		creschke@usra.edu
 Charles Fu		ccwf@klab.caltech.edu
+Charles McGarvey        chazmcgarvey@brokenzipper.com
 Charles Lopes		Charles.Lopes@infm.ulst.ac.uk
 Charles M. Hannum	mycroft@gnu.org
 Chip Salzenberg		tct!chip
@@ -93,6 +95,7 @@ Christian Laubscher	christian.laubscher@tiscalinet.ch
 Christian T. Dum	ctd@mpe-garching.mpg.de
 Christian von Roques	roques@pond.sub.org
 Christian Wetzel	wetzel@phoenix-pacs.de
+Christian Weisgerber	naddy@mips.inka.de
 Christoph Litauer	litauer@mailhost.uni-koblenz.de
 Christophe Colle	colle@krtkg1.rug.ac.be
 Christophe Kalt		Christophe.Kalt@kbcfp.com
@@ -107,6 +110,7 @@ Conrad Hughes		chughes@maths.tcd.ie
 Constantin Belous	const@cris.net
 Coranth Gryphon		gryphon@bur.visidyne.com
 Cyril Strejc		strejc@unicontrols.cz
+Dagobert Michelsen	dam@opencsw.org
 Dale R. Worley		worley@world.std.com
 Dale Wiles		wiles@geordi.calspan.com
 Dan Bloch		dan@transarc.com
@@ -114,6 +118,7 @@ Dan Drake               dan@dandrake.org
 Dan Reish		dreish@izzy.net
 Daniel Hagerty		hag@gnu.org
 Daniel Quinlan		quinlan@pathname.com
+Daniel Kahn Gillmor	dkg@fifthhorseman.net
 Daniel R. Guilderson	d.guilderson@ma30.bull.com
 Daniel S. Barclay	daniel@compass-da.com
 Daniel Trinkle		trinkle@cs.purdue.edu
@@ -135,11 +140,13 @@ David Nugent		davidn@blaze.net.au
 David Shaw		david.shaw@alcatel.com.au
 David Steiner		dsteiner@ispa.uni-osnabrueck.de
 David Taylor		taylor@think.com
+Dawid			dpc@dpc.pw
 Dean Gaudet		dgaudet@watdragon.uwaterloo.ca
 Demizu Noritoshi	nori-d@is.aist-nara.ac.jp
 Denis Excoffier         denis.excoffier@free.fr
 Denis Fortin		fortin@acm.org
 Dennis Pixton		dennis@math.binghamton.edu
+Derek Terveer           dt@hawkmoon.mn.org
 Dick Streefland		dicks@tasking.nl
 Dietmar Braun		dietmar@highway.bertelsmann.de
 Dimitri Bougoulias	opus@hol.gr
@@ -172,6 +179,7 @@ Erik D. Frederick	edf@deckard.mc.duke.edu
 Esa Karell		karell@cs.helsinki.fi
 Ezra Peisach		epeisach@mit.edu
 Fabio d'Alessi		cars@civ.bio.unipd.it
+Flavio Poletti          polettix@gmail.com
 Frank Heckenbach	frank@g-n-u.de
 Frank Koenen		koenfr@lidp.com
 Franz-Werner Gergen	gergen@edvulx.mpi-stuttgart.mpg.de
@@ -202,6 +210,7 @@ Helmut Waitzmann	Helmut.Waitzmann@web.de
 Henrik Bakman		hb@csd.uu.se
 Hernan Prieto Schmidt	hernan@pea.usp.br
 Hiroyuki Bessho		bsh@grotto.iijnet.or.jp
+Holger Levsen		holger@layer-acht.org
 Holger Teutsch		holger@hotbso.rhein-main.de
 Hugh Secker-Walker	hugh@ear.mit.edu
 Hunyue Yau		hunyue.yau@picksys.com
@@ -240,6 +249,7 @@ Jeffrey Goldberg	J.Goldberg@cranfield.ac.uk
 Jeffrey Mark Siskind	Qobi@emba.uvm.edu
 Jeffrey W. Parker	jwpkr@mcs.com
 Jens Henrik Jensen	recjhl@mediator.uni-c.dk
+Jérémy Bobbio		lunar@debian.org
 Jim Blandy		jimb@totoro.cs.oberlin.edu
 Jim Clausing		jac@postbox.acs.ohio-state.edu
 Jim Farrell		jwf@platinum.com

TODO

@@ -45,7 +45,7 @@ Suggestions for improving GNU tar.
 
 * Copyright notice
 
-Copyright 2003-2004, 2007, 2013-2014 Free Software Foundation, Inc.
+Copyright 2003-2021 Free Software Foundation, Inc.
 
 This file is part of GNU tar.
 

acinclude.m4

@@ -1,6 +1,6 @@
 dnl Special Autoconf macros for GNU tar         -*- autoconf -*-
 
-dnl Copyright 2009, 2013-2014 Free Software Foundation, Inc.
+dnl Copyright 2009-2021 Free Software Foundation, Inc.
 dnl
 dnl This file is part of GNU tar.
 dnl
@@ -40,37 +40,23 @@ AC_DEFUN([TAR_HEADERS_ATTR_XATTR_H],
   # First check for <sys/xattr.h>
   AC_CHECK_HEADERS([sys/xattr.h])
   AM_CONDITIONAL([TAR_COND_XATTR_H],[test "$ac_cv_header_sys_xattr_h" = yes])
-  AM_CONDITIONAL([TAR_LIB_ATTR],[false])
-  if test "$ac_cv_header_sys_xattr_h" = yes; then
-    AC_CHECK_FUNCS(getxattr  fgetxattr  lgetxattr \
-                   setxattr  fsetxattr  lsetxattr \
-                   listxattr flistxattr llistxattr,
-        # only when functions are present
-        AC_DEFINE([HAVE_SYS_XATTR_H], [1],
-                    [define to 1 if we have <sys/xattr.h> header])
-        if test "$with_xattrs" != no; then
-          AC_DEFINE([HAVE_XATTRS],,[Define when we have working linux xattrs.])
-        fi
-    )
-  fi
-
-  # If <sys/xattr.h> is not found, then check for <attr/xattr.h>
   if test "$ac_cv_header_sys_xattr_h" != yes; then
     AC_CHECK_HEADERS([attr/xattr.h])
     AM_CONDITIONAL([TAR_COND_XATTR_H],[test "$ac_cv_header_attr_xattr_h" = yes])
-    AC_CHECK_LIB([attr],[fgetxattr])
-    AM_CONDITIONAL([TAR_LIB_ATTR],[test "$ac_cv_lib_attr_fgetxattr" = yes])
-    if test "$ac_cv_header_attr_xattr_h" = yes; then
-      AC_CHECK_FUNCS(getxattr  fgetxattr  lgetxattr \
-                     setxattr  fsetxattr  lsetxattr \
-                     listxattr flistxattr llistxattr,
-          # only when functions are present
-          AC_DEFINE([HAVE_ATTR_XATTR_H], [1],
-                      [define to 1 if we have <attr/xattr.h> header])
-          if test "$with_xattrs" != no; then
-            AC_DEFINE([HAVE_XATTRS],,[Define when we have working linux xattrs.])
-          fi
-      )
+  fi
+
+  if test "$with_xattrs" != no; then
+    for i in getxattr  fgetxattr  lgetxattr \
+             setxattr  fsetxattr  lsetxattr \
+             listxattr flistxattr llistxattr
+    do
+      AC_SEARCH_LIBS($i, attr)
+      eval found=\$ac_cv_search_$i
+      test "$found" = "no" && break
+    done
+
+    if test "$found" != no; then
+      AC_DEFINE([HAVE_XATTRS],,[Define when we have working linux xattrs.])
     fi
   fi
 ])

bootstrap.conf

@@ -1,6 +1,6 @@
 # Bootstrap configuration for GNU tar.
 
-# Copyright 2006-2009, 2013-2014 Free Software Foundation, Inc.
+# Copyright 2006-2021 Free Software Foundation, Inc.
 
 # This file is part of GNU tar.
 
@@ -26,10 +26,6 @@ avoided_gnulib_modules='
 	--avoid=lock
 '
 
-# gnulib modules used by this package.
-gnulib_modules="$avoided_gnulib_modules
-`grep '^[^#]' gnulib.modules`
-"
 
 # Additional xgettext options to use.  Use "\\\newline" to break lines.
 XGETTEXT_OPTIONS=$XGETTEXT_OPTIONS'\\\
@@ -71,3 +67,60 @@ test -f ChangeLog || cat > ChangeLog <<EOT
 This file is a placeholder. It will be replaced with the actual ChangeLog
 by make dist.  Run make ChangeLog if you wish to create it earlier.
 EOT
+
+git submodule init
+git submodule update
+PAXUTILS=paxutils
+
+# gnulib modules used by this package.
+gnulib_modules="$avoided_gnulib_modules
+`grep -h '^[^#]' gnulib.modules $PAXUTILS/gnulib.modules`
+"
+
+# copy_files srcdir dstdir
+copy_files() {
+  for file in `cat $1/DISTFILES`
+  do
+    case $file in
+    "#*")  continue;;
+    esac
+    dst=`echo $file | sed 's^.*/^^'`
+    if [ $# -eq 3 ]; then
+      case $dst in
+      ${3}*) ;;
+      *) dst=${3}$dst;;
+      esac
+    fi
+    if [ "$2" = '.' ]; then
+      ln -sf $1/$file $2
+    else
+      symlink_to_dir "$1" "$file" "$2/$dst" || exit
+    fi
+# FIXME    ignorefile $2 $dst
+  done
+}
+
+# Import from paxutils
+copy_files ${PAXUTILS} .
+copy_files ${PAXUTILS}/am m4
+
+echo "$0: Creating m4/paxutils.m4"
+(echo "# This file is generated automatically. Please, do not edit."
+ echo "#"
+ echo "AC_DEFUN([${package}_PAXUTILS],["
+ cat ${PAXUTILS}/am/DISTFILES | sed '/^#/d;s/\(.*\)\.m4/pu_\1/' | tr a-z A-Z
+ echo "])") > ./m4/paxutils.m4
+#FIXME ignorefile m4 paxutils.m4
+
+if [ -d rmt ]; then
+   :
+else
+   mkdir rmt
+fi
+
+for dir in doc rmt lib tests
+do
+	copy_files ${PAXUTILS}/$dir $dir
+done
+
+copy_files ${PAXUTILS}/paxlib lib pax

configure.ac

@@ -1,6 +1,6 @@
 # Configure template for GNU tar.  -*- autoconf -*-
 
-# Copyright 1991, 1994-2010, 2013-2014 Free Software Foundation, Inc.
+# Copyright 1991, 1994-2010, 2013-2021 Free Software Foundation, Inc.
 
 # This file is part of GNU tar.
 
@@ -17,12 +17,12 @@
 # You should have received a copy of the GNU General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 
-AC_INIT([GNU tar], [1.28], [bug-tar@gnu.org])
+AC_INIT([GNU tar], [1.34], [bug-tar@gnu.org])
 AC_CONFIG_SRCDIR([src/tar.c])
 AC_CONFIG_AUX_DIR([build-aux])
 AC_CONFIG_HEADERS([config.h])
-AC_PREREQ([2.63])
-AM_INIT_AUTOMAKE([1.11 gnits tar-ustar dist-bzip2 dist-xz std-options silent-rules])
+AC_PREREQ([2.64])
+AM_INIT_AUTOMAKE([1.15 gnits tar-ustar dist-bzip2 dist-xz std-options silent-rules])
 
 # Enable silent rules by default:
 AM_SILENT_RULES([yes])
@@ -102,6 +102,18 @@ gt_TYPE_SSIZE_T
 
 # gnulib modules
 gl_INIT
+
+if test $ac_cv_lib_error_at_line = no; then
+  # This means that the error() function is not present in libc, so
+  # the one from gnulib will be used instead. This function precedes
+  # error messages it prints with the program name as returned by getprogname()
+  # call, instead of using the name set by set_program_name.
+  # Install workaround.
+  AC_DEFINE([ENABLE_ERROR_PRINT_PROGNAME],[1],
+            [Enable the use of error_print_progname to print program name with error messages.
+	     See comment to function tar_print_progname in src/tar.c])
+fi  
+
 # paxutils modules
 tar_PAXUTILS
 
@@ -150,6 +162,7 @@ if test "$gl_gcc_warnings" = yes; then
   nw="$nw -Winline"                 # It's OK to not inline.
   nw="$nw -Wstrict-overflow"	    # It's OK to optimize strictly.
   nw="$nw -Wsuggest-attribute=pure" # Too many warnings for now.
+  nw="$nw -Wstack-protector"
 
   gl_MANYWARN_ALL_GCC([ws])
   gl_MANYWARN_COMPLEMENT([ws], [$ws], [$nw])
@@ -160,9 +173,10 @@ if test "$gl_gcc_warnings" = yes; then
   gl_WARN_ADD([-Wno-type-limits])      # It's OK to optimize based on types.
   gl_WARN_ADD([-Wno-unused-parameter]) # Too many warnings for now
   gl_WARN_ADD([-Wno-format-nonliteral])
-
+  
   gl_WARN_ADD([-fdiagnostics-show-option])
   gl_WARN_ADD([-funit-at-a-time])
+  
 
   AC_SUBST([WARN_CFLAGS])
 
@@ -204,8 +218,7 @@ fi
 
 TAR_HEADERS_ATTR_XATTR_H
 
-AC_CHECK_FUNCS_ONCE([fchmod fchown fsync lstat mkfifo readlink symlink grantpt])
-AM_CONDITIONAL([TAR_COND_GRANTPT], [test $ac_cv_func_grantpt = yes])
+AC_CHECK_FUNCS_ONCE([fchmod fchown fsync lstat mkfifo readlink symlink])
 
 AC_CHECK_DECLS([getgrgid],,, [#include <grp.h>])
 AC_CHECK_DECLS([getpwuid],,, [#include <pwd.h>])
@@ -249,6 +262,7 @@ TAR_COMPR_PROGRAM(lzip)
 TAR_COMPR_PROGRAM(lzma)
 TAR_COMPR_PROGRAM(lzop)
 TAR_COMPR_PROGRAM(xz)
+TAR_COMPR_PROGRAM(zstd)
 
 AC_MSG_CHECKING(for default archive format)
 

directory

@@ -1,5 +1,5 @@
 %%comments:
-Copyright 2004, 2013-2014 Free Software Foundation, Inc.
+Copyright 2004-2021 Free Software Foundation, Inc.
 
 Permission is granted to copy, distribute and/or modify this document
 under the terms of the GNU Free Documentation License, Version 1.3 or

doc/.gitignore

@@ -1,7 +1,6 @@
 genfile.texi
 header.texi
 manual
-parse-datetime.texi
 stamp-vti
 tar.aux
 tar.cp
@@ -24,3 +23,5 @@ tar.toc
 tar.tp
 tar.vr
 version.texi
+/parse-datetime.texi
+/rmt.8

doc/Makefile.am

@@ -1,7 +1,6 @@
 # Makefile for GNU tar documentation.
 
-# Copyright 1994-1997, 1999-2001, 2003, 2006-2007, 2013-2014 Free
-# Software Foundation, Inc.
+# Copyright 1994-2021 Free Software Foundation, Inc.
 
 # This file is part of GNU tar.
 
@@ -23,11 +22,11 @@ tar_TEXINFOS = \
  dumpdir.texi\
  tar-snapshot-edit.texi\
  fdl.texi\
- freemanuals.texi\
  genfile.texi\
  header.texi\
  intern.texi\
  parse-datetime.texi\
+ recipes.texi\
  rendition.texi\
  snapshot.texi\
  sparse.texi\

doc/dumpdir.texi

@@ -1,5 +1,5 @@
 @c This is part of the paxutils manual.
-@c Copyright (C) 2006-2007, 2014 Free Software Foundation, Inc.
+@c Copyright (C) 2006--2021 Free Software Foundation, Inc.
 @c Written by Sergey Poznyakoff
 @c This file is distributed under GFDL 1.1 or any later version
 @c published by the Free Software Foundation.

doc/fdl.texi

@@ -5,8 +5,7 @@
 @c hence no sectioning command or @node.
 
 @display
-Copyright @copyright{} 2000-2002, 2007-2008, 2014 Free Software
-Foundation, Inc.
+Copyright @copyright{} 2000--2021 Free Software Foundation, Inc.
 @uref{http://fsf.org/}
 
 Everyone is permitted to copy and distribute verbatim copies

doc/freemanuals.texi

@@ -1,89 +0,0 @@
-@cindex free documentation
-
-The biggest deficiency in the free software community today is not in
-the software---it is the lack of good free documentation that we can
-include with the free software.  Many of our most important
-programs do not come with free reference manuals and free introductory
-texts.  Documentation is an essential part of any software package;
-when an important free software package does not come with a free
-manual and a free tutorial, that is a major gap.  We have many such
-gaps today.
-
-Consider Perl, for instance.  The tutorial manuals that people
-normally use are non-free.  How did this come about?  Because the
-authors of those manuals published them with restrictive terms---no
-copying, no modification, source files not available---which exclude
-them from the free software world.
-
-That wasn't the first time this sort of thing happened, and it was far
-from the last.  Many times we have heard a GNU user eagerly describe a
-manual that he is writing, his intended contribution to the community,
-only to learn that he had ruined everything by signing a publication
-contract to make it non-free.
-
-Free documentation, like free software, is a matter of freedom, not
-price.  The problem with the non-free manual is not that publishers
-charge a price for printed copies---that in itself is fine.  (The Free
-Software Foundation sells printed copies of manuals, too.)  The
-problem is the restrictions on the use of the manual.  Free manuals
-are available in source code form, and give you permission to copy and
-modify.  Non-free manuals do not allow this.
-
-The criteria of freedom for a free manual are roughly the same as for
-free software.  Redistribution (including the normal kinds of
-commercial redistribution) must be permitted, so that the manual can
-accompany every copy of the program, both on-line and on paper.
-
-Permission for modification of the technical content is crucial too.
-When people modify the software, adding or changing features, if they
-are conscientious they will change the manual too---so they can
-provide accurate and clear documentation for the modified program.  A
-manual that leaves you no choice but to write a new manual to document
-a changed version of the program is not really available to our
-community.
-
-Some kinds of limits on the way modification is handled are
-acceptable.  For example, requirements to preserve the original
-author's copyright notice, the distribution terms, or the list of
-authors, are ok.  It is also no problem to require modified versions
-to include notice that they were modified.  Even entire sections that
-may not be deleted or changed are acceptable, as long as they deal
-with nontechnical topics (like this one).  These kinds of restrictions
-are acceptable because they don't obstruct the community's normal use
-of the manual.
-
-However, it must be possible to modify all the @emph{technical}
-content of the manual, and then distribute the result in all the usual
-media, through all the usual channels.  Otherwise, the restrictions
-obstruct the use of the manual, it is not free, and we need another
-manual to replace it.
-
-Please spread the word about this issue.  Our community continues to
-lose manuals to proprietary publishing.  If we spread the word that
-free software needs free reference manuals and free tutorials, perhaps
-the next person who wants to contribute by writing documentation will
-realize, before it is too late, that only free manuals contribute to
-the free software community.
-
-If you are writing documentation, please insist on publishing it under
-the GNU Free Documentation License or another free documentation
-license.  Remember that this decision requires your approval---you
-don't have to let the publisher decide.  Some commercial publishers
-will use a free license if you insist, but they will not propose the
-option; it is up to you to raise the issue and say firmly that this is
-what you want.  If the publisher you are dealing with refuses, please
-try other publishers.  If you're not sure whether a proposed license
-is free, write to @email{licensing@@gnu.org}.
-
-You can encourage commercial publishers to sell more free, copylefted
-manuals and tutorials by buying them, and particularly by buying
-copies from the publishers that paid for their writing or for major
-improvements.  Meanwhile, try to avoid buying non-free documentation
-at all.  Check the distribution terms of a manual before you buy it,
-and insist that whoever seeks your business must respect your freedom.
-Check the history of the book, and try reward the publishers that have
-paid or pay the authors to work on it.
-
-The Free Software Foundation maintains a list of free documentation
-published by other publishers, at
-@url{http://www.fsf.org/doc/other-free-books.html}.

doc/gendocs_template

@@ -106,7 +106,7 @@ Please send broken links and other corrections (or suggestions) to
 </p>
 
 <p>
-Copyright 2004, 2013-2014 Free Software Foundation, Inc.,
+Copyright 2004-2021 Free Software Foundation, Inc.,
 51 Franklin Street, Fifth Floor, Boston, MA 02111, USA
 <br />
 Verbatim copying and distribution of this entire article is

doc/intern.texi

@@ -1,5 +1,5 @@
 @c This is part of the paxutils manual.
-@c Copyright (C) 2006, 2014 Free Software Foundation, Inc.
+@c Copyright (C) 2006--2021 Free Software Foundation, Inc.
 @c This file is distributed under GFDL 1.1 or any later version
 @c published by the Free Software Foundation.
 
@@ -87,6 +87,8 @@ The @code{name}, @code{linkname}, @code{magic}, @code{uname}, and
 @code{gname} are null-terminated character strings.  All other fields
 are zero-filled octal numbers in ASCII.  Each numeric field of width
 @var{w} contains @var{w} minus 1 digits, and a null.
+(In the extended @acronym{GNU} format, the numeric fields can take
+other forms.)
 
 The @code{name} field is the file name of the file, with directory names
 (if any) preceding the file name, separated by slashes.
@@ -112,14 +114,12 @@ be ignored.
 The @code{size} field is the size of the file in bytes; linked files
 are archived with this field specified as zero.
 
-The @code{mtime} field is the data modification time of the file at
-the time it was archived.  It is the ASCII representation of the octal
-value of the last time the file's contents were modified, represented
-as an integer number of
+The @code{mtime} field represents the data modification time of the file at
+the time it was archived.  It represents the integer number of
 seconds since January 1, 1970, 00:00 Coordinated Universal Time.
 
-The @code{chksum} field is the ASCII representation of the octal value
-of the simple sum of all bytes in the header block.  Each 8-bit
+The @code{chksum} field represents
+the simple sum of all bytes in the header block.  Each 8-bit
 byte in the header is added to an unsigned integer, initialized to
 zero, the precision of which shall be no less than seventeen bits.
 When calculating the checksum, the @code{chksum} field is treated as
@@ -310,6 +310,18 @@ of an archive should have this type.
 
 @end table
 
+For fields containing numbers or timestamps that are out of range for
+the basic format, the @acronym{GNU} format uses a base-256
+representation instead of an ASCII octal number.  If the leading byte
+is 0xff (255), all the bytes of the field (including the leading byte)
+are concatenated in big-endian order, with the result being a negative
+number expressed in two's complement form.  If the leading byte is
+0x80 (128), the non-leading bytes of the field are concatenated in
+big-endian order, with the result being a positive number expressed in
+binary form.  Leading bytes other than 0xff, 0x80 and ASCII octal
+digits are reserved for future use, as are base-256 representations of
+values that would be in range for the basic format.
+
 You may have trouble reading a @acronym{GNU} format archive on a
 non-@acronym{GNU} system if the options @option{--incremental} (@option{-G}),
 @option{--multi-volume} (@option{-M}), @option{--sparse} (@option{-S}), or @option{--label=@var{archive-label}} (@option{-V @var{archive-label}}) were

doc/mastermenu.el

@@ -1,6 +1,6 @@
 ;;; mastermenu.el --- Redefinition of texinfo-master-menu-list
 
-;; Copyright 2006-2007, 2013-2014 Free Software Foundation, Inc.
+;; Copyright 2006-2021 Free Software Foundation, Inc.
 
 ;; Author: Sergey Poznyakoff
 ;; Maintainer: bug-tar@gnu.org

doc/recipes.texi

@@ -0,0 +1,134 @@
+@c This is part of the GNU tar manual.
+@c Copyright (C) 2017--2021 Free Software Foundation, Inc.
+@c This file is distributed under GFDL 1.3 or any later version
+@c published by the Free Software Foundation.
+
+This appendix provides several recipes for performing common tasks
+using @GNUTAR{}.
+
+@menu
+* copy directory hierarchy::
+* intermediate directories::
+@end menu
+
+@node copy directory hierarchy
+@appendixsec Copying directory hierarchies
+
+This is a traditional way to copy a directory hierarchy preserving
+the dates, modes, owners and link-structure of all the files therein.
+It was used back when the @command{cp} command lacked the @option{-a}
+option:
+
+@smallexample
+$ @kbd{(cd sourcedir; tar -cf - .) | (cd targetdir; tar -xf -)}
+@end smallexample
+
+@noindent
+You can avoid subshells by using @option{-C} option:
+
+@smallexample
+$ @kbd{tar -C sourcedir -cf - . | tar -C targetdir -xf -}
+@end smallexample
+
+@noindent
+The same command using long option forms:
+
+@smallexample
+@group
+$ @kbd{(cd sourcedir; tar --create --file=- . ) \
+       | (cd targetdir; tar --extract --file=-)}
+@end group
+@end smallexample
+
+@noindent
+or
+
+@smallexample
+@group
+$ @kbd{tar --directory sourcedir --create --file=- . \
+       | tar --directory targetdir --extract --file=-}
+@end group
+@end smallexample
+
+@node intermediate directories
+@appendixsec Restoring Intermediate Directories
+
+A common concern is how to extract permissions and ownerships of
+intermediate directories when extracting only selected members from
+the archive.  To illustrate this, consider the following archive:
+
+@example
+@group
+# tar tvf A.tar
+drwxr-xr-x root/root         0 2017-11-16 14:39 foo/
+dr-xr-x--- gray/user         0 2017-11-16 14:39 foo/bar/
+-rw-r--r-- gray/user        10 2017-11-16 14:40 foo/bar/file
+@end group
+@end example
+
+Suppose you extract only the file @file{foo/bar/file}, while being
+@samp{root}:
+
+@example
+# @kbd{tar xvf A.tar foo/bar/file}
+foo/bar/file
+@end example
+
+Now, let's inspect the content of the created directories:
+
+@example
+@group
+# find foo -ls
+427257    0 drwxr-xr-x   3 root     root    16 Nov 17 16:10 foo
+427258    0 drwxr-xr-x   2 root     root    17 Nov 17 16:10 foo/bar
+427259    0 -rw-r--r--   1 gray     user    10 Nov  6 14:40 foo/bar/file
+@end group
+@end example
+
+The requested file is restored, including its ownership and
+permissions. The intermediate directories, however, are created with
+the default permissions, current timestamp and owned by the current
+user. This is because by the time @command{tar} has reached the requested file,
+it had already skipped the entries for its parent directories, so it
+has no iformation about their ownership and modes.
+
+To restore meta information about the intermediate directories,
+you'll need to specify them explicitly in the command line and use the
+@option{--no-recursive} option (@pxref{recurse}) to avoid extracting
+their content.
+
+To automate this process, @cite{Neal P. Murphy} proposed the following
+shell script@footnote{The original version of the script can be
+seen at @uref{http://lists.gnu.org/archive/html/bug-tar/2016-11/msg00024.html}}:
+
+@example
+@group
+#! /bin/sh
+(while read path
+ do
+   path=`dirname $path`
+   while [ -n "$path" -a "$path" != "." ]
+   do
+     echo $path
+     path=`dirname $path`
+   done
+ done < $2 | sort | uniq) |
+ tar -x --no-recursion -v -f $1 -T - -T $2
+@end group
+@end example
+
+The script takes two arguments: the name of the archive file, and the
+name of the file list file.
+
+To complete our example, the file list will contain single line:
+
+@example
+foo/bar/file
+@end example
+
+Supposing its name is @file{file.list} and the script is named
+@file{restore.sh}, you can invoke it as follows: 
+
+@example
+# @kbd{sh restore.sh A.tar file.list}
+@end example

doc/rendition.texi

@@ -1,6 +1,5 @@
 @c This is part of GNU tar manual.
-@c Copyright 1992, 1994-1997, 1999-2004, 2006, 2013-2014 Free Software
-@c Foundation, Inc.
+@c Copyright 1992--2021 Free Software Foundation, Inc.
 @c See file tar.texi for copying conditions.
 
 @c This file contains support for 'renditions' by Fran@,{c}ois Pinard

doc/snapshot.texi

@@ -1,5 +1,5 @@
 @c This is part of the paxutils manual.
-@c Copyright (C) 2005, 2007, 2014 Free Software Foundation, Inc.
+@c Copyright (C) 2005--2021 Free Software Foundation, Inc.
 @c Written by Sergey Poznyakoff
 @c This file is distributed under GFDL 1.1 or any later version
 @c published by the Free Software Foundation.
@@ -116,7 +116,7 @@ epoch.  These are followed by arbitrary number of directory records.
 particular directory.  Parts of a directory record are delimited with
 @acronym{ASCII} 0 characters.  The following table describes each
 part.  The @dfn{Number} type in this table stands for a decimal integer
-in @acronym{ASCII} notation.  (Negative values are preceeded with a "-"
+in @acronym{ASCII} notation.  (Negative values are preceded with a "-"
 character, while positive values have no leading punctuation.)
 
 @multitable @columnfractions 0.25 0.15 0.6
@@ -146,7 +146,7 @@ is compiled.  To see the specific ranges allowed for a particular
 @option{--show-snapshot-field-ranges} option:
 
 @smallexample
-$ @kbd{tar --show-shapshot-field-ranges}
+$ @kbd{tar --show-snapshot-field-ranges}
 This tar's snapshot file field ranges are
    (field name      => [ min, max ]):
 

doc/sparse.texi

@@ -1,5 +1,5 @@
 @c This is part of the paxutils manual.
-@c Copyright (C) 2006, 2014 Free Software Foundation, Inc.
+@c Copyright (C) 2006--2021 Free Software Foundation, Inc.
 @c This file is distributed under GFDL 1.1 or any later version
 @c published by the Free Software Foundation.
 
@@ -143,7 +143,7 @@ format, it will also extract a file containing extension header
 attributes.  This file can be used to expand the file to its original
 state.  However, posix-aware @command{tar}s will usually ignore the
 unknown variables, which makes restoring the file more
-difficult.  @xref{extracting sparse v.0.x, Extraction of sparse
+difficult.  @xref{extracting sparse v0x, Extraction of sparse
 members in v.0.0 format}, for the detailed description of how to
 restore such members using non-GNU @command{tar}s.
 @end enumerate
@@ -175,7 +175,7 @@ The real name of the sparse file is stored in the variable
 @code{GNU.sparse.name}.  Thus, those @command{tar} implementations
 that are not aware of GNU extensions will at least extract the files
 into separate directories, giving the user a possibility to expand it
-afterwards.  @xref{extracting sparse v.0.x, Extraction of sparse
+afterwards.  @xref{extracting sparse v0x, Extraction of sparse
 members in v.0.1 format}, for the detailed description of how to
 restore such members using non-GNU @command{tar}s.
 
@@ -218,12 +218,12 @@ The real name of the sparse file is stored in the variable
 variable @code{GNU.sparse.realsize}.
 
 The sparse map itself is stored in the file data block, preceding the actual
-file data.  It consists of a series of octal numbers of arbitrary length, delimited
+file data.  It consists of a series of decimal numbers delimited
 by newlines. The map is padded with nulls to the nearest block boundary.
 
-The first number gives the number of entries in the map. Following are map entries,
-each one consisting of two numbers giving the offset and size of the
-data block it describes.
+The first number gives the number of entries in the map. Following are
+map entries, each one consisting of two numbers giving the offset and
+size of the data block it describes.
 
 The format is designed in such a way that non-posix aware @command{tar}s and @command{tar}s not
 supporting @code{GNU.sparse.*} keywords will extract each sparse file

doc/tar-snapshot-edit.texi

@@ -1,5 +1,5 @@
 @c This is part of the paxutils manual.
-@c Copyright (C) 2007, 2014 Free Software Foundation, Inc.
+@c Copyright (C) 2007--2021 Free Software Foundation, Inc.
 @c This file is distributed under GFDL 1.1 or any later version
 @c published by the Free Software Foundation.
 

doc/tar.1

@@ -1,5 +1,5 @@
 .\" This file is part of GNU tar. -*- nroff -*-
-.\" Copyright 2013-2014 Free Software Foundation, Inc.
+.\" Copyright 2013-2021 Free Software Foundation, Inc.
 .\"
 .\" GNU tar is free software; you can redistribute it and/or modify
 .\" it under the terms of the GNU General Public License as published by
@@ -13,7 +13,7 @@
 .\"
 .\" You should have received a copy of the GNU General Public License
 .\" along with this program.  If not, see <http://www.gnu.org/licenses/>.
-.TH TAR 1 "February 22, 2014" "TAR" "GNU TAR Manual"
+.TH TAR 1 "July 13, 2020" "TAR" "GNU TAR Manual"
 .SH NAME
 tar \- an archiving utility
 .SH SYNOPSIS
@@ -37,7 +37,7 @@ tar \- an archiving utility
 \fBtar\fR \fB\-x\fR [\fB\-f\fR \fIARCHIVE\fR] [\fIOPTIONS\fR] [\fIMEMBER\fR...]
 .SS GNU-style usage
 .sp
-\fBtar\fR {\fB\-\-catenate\fR|\fB\-\-concatenate} [\fIOPTIONS\fR] \fIARCHIVE\fR \fIARCHIVE\fR
+\fBtar\fR {\fB\-\-catenate\fR|\fB\-\-concatenate\fR} [\fIOPTIONS\fR] \fIARCHIVE\fR \fIARCHIVE\fR
 .sp
 \fBtar\fR \fB\-\-create\fR [\fB\-\-file\fR \fIARCHIVE\fR] [\fIOPTIONS\fR] [\fIFILE\fR...]
 .sp
@@ -111,7 +111,7 @@ into the archive file
 verbosely listing the files being archived:
 .PP
 .EX
-.B tar cfv a.tar /etc
+tar cfv etc.tar /etc
 .EE
 .PP
 In
@@ -133,9 +133,11 @@ The example command above written in the
 could look like:
 .PP
 .EX
-.B tar -cvf a.tar /etc
+tar -cvf etc.tar /etc
+.EE
 or
-.B tar -c -v -f a.tar /etc
+.EX
+tar -c -v -f etc.tar /etc
 .EE
 .PP
 In
@@ -152,11 +154,11 @@ method.
 Here are several ways of writing the example command in this style:
 .PP
 .EX
-.B tar --create --file a.tar --verbose /etc
+tar --create --file etc.tar --verbose /etc
 .EE
 or (abbreviating some options):
 .EX
-.B tar --cre --file=a.tar --verb /etc
+tar --cre --file=etc.tar --verb /etc
 .EE
 .PP
 The options in all three styles can be intermixed, although doing so
@@ -175,7 +177,7 @@ resulting archive might be unusable with non-GNU implementations of
 \fBtar\fR.  Notice also that when more than one archive is given, the
 members from archives other than the first one will be accessible in
 the resulting archive only if using the \fB\-i\fR
-(\fB\-\-ignore\-zeros) option.
+(\fB\-\-ignore\-zeros\fR) option.
 
 Compressed archives cannot be concatenated.
 .TP
@@ -198,7 +200,7 @@ short option equivalent.
 .TP
 \fB\-r\fR, \fB\-\-append\fR
 Append files to the end of an archive.  Arguments have the same
-meaning as for \fB\-c\fR (\fB\-\-create).
+meaning as for \fB\-c\fR (\fB\-\-create\fR).
 .TP
 \fB\-t\fR, \fB\-\-list\fR
 List the contents of an archive.  Arguments are optional.  When given,
@@ -219,7 +221,10 @@ There is no short option equivalent for this option.
 \fB\-u\fR, \fB\-\-update\fR
 Append files which are newer than the corresponding copy in the
 archive.  Arguments have the same meaning as with \fB\-c\fR and
-\fB\-r\fR options.
+\fB\-r\fR options.  Notice, that newer files don't replace their
+old archive copies, but instead are appended to the end of archive.
+The resulting archive can thus contain several members of the
+same name, corresponding to various versions of the same file.
 .TP
 \fB\-x\fR, \fB\-\-extract\fR, \fB\-\-get\fR
 Extract files from an archive.  Arguments are optional.  When given,
@@ -253,12 +258,18 @@ exist when creating an archive, it will be created and all files will
 be added to the resulting archive (the \fBlevel 0\fR dump).  To create
 incremental archives of non-zero level \fBN\fR, create a copy of the
 snapshot file created during the level \fBN-1\fR, and use it as
-\fIFILE\fR. 
+\fIFILE\fR.
 
 When listing or extracting, the actual contents of \fIFILE\fR is not
 inspected, it is needed only due to syntactical requirements.  It is
 therefore common practice to use \fB/dev/null\fR in its place.
 .TP
+\fB\-\-hole\-detection\fR=\fIMETHOD\fR
+Use \fIMETHOD\fR to detect holes in sparse files.  This option implies
+\fB\-\-sparse\fR.  Valid values for \fIMETHOD\fR are \fBseek\fR and
+\fBraw\fR.  Default is \fBseek\fR with fallback to \fBraw\fR when not
+applicable.
+.TP
 \fB\-G\fR, \fB\-\-incremental\fR
 Handle old GNU-format incremental backups.
 .TP
@@ -278,7 +289,7 @@ effect only if the archive is open for reading (e.g. with
 .B \-\-list
 or
 .B \-\-extract
-options).		    
+options).
 .TP
 \fB\-\-no\-check\-device\fR
 Do not check device numbers when creating incremental archives.
@@ -291,7 +302,7 @@ Process only the \fIN\fRth occurrence of each file in the
 archive.  This option is valid only when used with one of the
 following subcommands: \fB\-\-delete\fR, \fB\-\-diff\fR,
 \fB\-\-extract\fR or \fB\-\-list\fR and when a list of files is given
-either on the command line or via the \fB\-\fRT option.  The default
+either on the command line or via the \fB\-T\fR option.  The default
 \fIN\fR is \fB1\fR.
 .TP
 \fB\-\-restrict\fR
@@ -327,6 +338,9 @@ Don't replace existing files when extracting.
 \fB\-\-keep\-newer\-files\fR
 Don't replace existing files that are newer than their archive copies.
 .TP
+\fB\-\-keep\-directory\-symlink\fR
+Don't replace existing symlinks to directories when extracting.
+.TP
 \fB\-\-no\-overwrite\-dir\fR
 Preserve metadata of existing directories.
 .TP
@@ -347,6 +361,9 @@ Recursively remove all files in the directory prior to extracting it.
 \fB\-\-remove\-files\fR
 Remove files from disk after adding them to the archive.
 .TP
+\fB\-\-skip\-old\-files
+Don't replace existing files when extracting, silently skip over them.
+.TP
 \fB\-U\fR, \fB\-\-unlink\-first\fR
 Remove each file prior to extracting over it.
 .TP
@@ -368,7 +385,7 @@ Extract files to standard output.
 Pipe extracted files to \fICOMMAND\fR.  The argument is the pathname
 of an external program, optionally with command line arguments.  The
 program will be invoked and the contents of the file being extracted
-supplied to it on its standard output.  Additional data will be
+supplied to it on its standard input.  Additional data will be
 supplied via the following environment variables:
 .RS
 .TP
@@ -448,9 +465,10 @@ Format of the archive being processed.  One of:
 .BR posix ,
 .BR ustar ,
 .BR v7 .
+.TP
 .B TAR_SUBCOMMAND
 A short option (with a leading dash) describing the operation \fBtar\fR is
-executing. 
+executing.
 .RE
 .SS Handling of file attributes
 .TP
@@ -464,15 +482,40 @@ Delay setting modification times and permissions of extracted
 directories until the end of extraction.  Use this option when
 extracting from an archive which has unusual member ordering.
 .TP
-\fB\-\-group\fR=\fINAME\fR
-Force \fINAME\fR as group for added files.
+\fB\-\-group\fR=\fINAME\fR[:\fIGID\fR]
+Force \fINAME\fR as group for added files.  If \fIGID\fR is not
+supplied, \fINAME\fR can be either a user name or numeric GID.  In
+this case the missing part (GID or name) will be inferred from the
+current host's group database.
+
+When used with \fB\-\-group\-map\fR=\fIFILE\fR, affects only those
+files whose owner group is not listed in \fIFILE\fR.
+.TP
+\fB\-\-group\-map\fR=\fIFILE\fR
+Read group translation map from \fIFILE\fR.  Empty lines are ignored.
+Comments are introduced with \fB#\fR sign and extend to the end of line.
+Each non-empty line in \fIFILE\fR defines translation for a single
+group.  It must consist of two fields, delimited by any amount of whitespace:
+
+.EX
+\fIOLDGRP\fR \fINEWGRP\fR[\fB:\fINEWGID\fR]
+.EE
+
+\fIOLDGRP\fR is either a valid group name or a GID prefixed with
+\fB+\fR.  Unless \fINEWGID\fR is supplied, \fINEWGRP\fR must also be
+either a valid group name or a \fB+\fIGID\fR.  Otherwise, both
+\fINEWGRP\fR and \fINEWGID\fR need not be listed in the system group
+database.
+
+As a result, each input file with owner group \fIOLDGRP\fR will be
+stored in archive with owner group \fINEWGRP\fR and GID \fINEWGID\fR.
 .TP
 \fB\-\-mode\fR=\fICHANGES\fR
 Force symbolic mode \fICHANGES\fR for added files.
 .TP
 \fB\-\-mtime\fR=\fIDATE-OR-FILE\fR
 Set mtime for added files.  \fIDATE-OR-FILE\fR is either a date/time
-in almost arbitrary formate, or the name of an existing file.  In the
+in almost arbitrary format, or the name of an existing file.  In the
 latter case the mtime of that file will be used.
 .TP
 \fB\-m\fR, \fB\-\-touch\fR
@@ -491,15 +534,37 @@ Apply the user's umask when extracting permissions from the archive
 \fB\-\-numeric\-owner\fR
 Always use numbers for user/group names.
 .TP
-\fB\-\-owner\fR=\fINAME\fR
-Force \fINAME\fR as owner for added files.
+\fB\-\-owner\fR=\fINAME\fR[:\fIUID\fR]
+Force \fINAME\fR as owner for added files.  If \fIUID\fR is not
+supplied, \fINAME\fR can be either a user name or numeric UID.  In
+this case the missing part (UID or name) will be inferred from the
+current host's user database.
+
+When used with \fB\-\-owner\-map\fR=\fIFILE\fR, affects only those
+files whose owner is not listed in \fIFILE\fR.
+.TP
+\fB\-\-owner\-map\fR=\fIFILE\fR
+Read owner translation map from \fIFILE\fR.  Empty lines are ignored.
+Comments are introduced with \fB#\fR sign and extend to the end of line.
+Each non-empty line in \fIFILE\fR defines translation for a single
+UID.  It must consist of two fields, delimited by any amount of whitespace:
+
+.EX
+\fIOLDUSR\fR \fINEWUSR\fR[\fB:\fINEWUID\fR]
+.EE
+
+\fIOLDUSR\fR is either a valid user name or a UID prefixed with
+\fB+\fR.  Unless \fINEWUID\fR is supplied, \fINEWUSR\fR must also be
+either a valid user name or a \fB+\fIUID\fR.  Otherwise, both
+\fINEWUSR\fR and \fINEWUID\fR need not be listed in the system user
+database.
+
+As a result, each input file owned by \fIOLDUSR\fR will be
+stored in archive with owner name \fINEWUSR\fR and UID \fINEWUID\fR.
 .TP
 \fB\-p\fR, \fB\-\-preserve\-permissions\fR, \fB\-\-same\-permissions\fR
 extract information about file permissions (default for superuser)
 .TP
-\fB\-\-preserve\fR
-Same as both \fB\-p\fR and \fB\-s\fR.
-.TP
 \fB\-\-same\-owner\fR
 Try extracting files with the same ownership as exists in the archive
 (default for superuser).
@@ -524,6 +589,34 @@ Using \fB\-\-sort=inode\fR reduces the number of disk seeks made when
 creating the archive and thus can considerably speed up archivation.
 This sorting order is supported only if the underlying system provides
 the necessary information.
+.SS Extended file attributes
+.TP
+.B \-\-acls
+Enable POSIX ACLs support.
+.TP
+.B \-\-no\-acls
+Disable POSIX ACLs support.
+.TP
+.B \-\-selinux
+Enable SELinux context support.
+.TP
+.B \-\-no-selinux
+Disable SELinux context support.
+.TP
+.B \-\-xattrs
+Enable extended attributes support.
+.TP
+.B \-\-no\-xattrs
+Disable extended attributes support.
+.TP
+.BI \-\-xattrs\-exclude= PATTERN
+Specify the exclude pattern for xattr keys.  \fIPATTERN\fR is a POSIX
+regular expression, e.g. \fB\-\-xattrs\-exclude='^user\.'\fR, to exclude
+attributes from the user namespace.
+.TP
+.BI \-\-xattrs\-include= PATTERN
+Specify the include pattern for xattr keys.  \fIPATTERN\fR is a POSIX
+regular expression.
 .SS Device selection and switching
 .TP
 \fB\-f\fR, \fB\-\-file\fR=\fIARCHIVE\fR
@@ -557,7 +650,7 @@ instead.  You can do so by giving the following command line option:
 --rsh-command=/usr/bin/ssh
 .EE
 
-The remote mashine should have the
+The remote machine should have the
 .BR rmt (8)
 command installed.  If its pathname does not match \fBtar\fR's
 default, you can inform \fBtar\fR about the correct pathname using the
@@ -567,7 +660,7 @@ option.
 \fB\-\-force\-local\fR
 Archive file is local even if it has a colon.
 .TP
-\fB\-\fRF, \fB\-\-info\-script\fR=\fICOMMAND\fR, \fB\-\-new\-volume\-script\fR=\fICOMMAND\fR
+\fB\-F\fR, \fB\-\-info\-script\fR=\fICOMMAND\fR, \fB\-\-new\-volume\-script\fR=\fICOMMAND\fR
 Run \fICOMMAND\fR at the end of each tape (implies \fB\-M\fR).  The
 command can include arguments.  When started, it will inherit \fBtar\fR's
 environment plus the following variables:
@@ -601,7 +694,7 @@ executing.
 .B TAR_FD
 File descriptor which can be used to communicate the new volume name
 to
-.BR tar .  
+.BR tar .
 .RE
 .RS
 
@@ -609,7 +702,7 @@ If the info script fails, \fBtar\fR exits; otherwise, it begins writing
 the next volume.
 .RE
 .TP
-\fB\-L\fR, \fB\-\-tape\-length\fR=\fN\fR
+\fB\-L\fR, \fB\-\-tape\-length\fR=\fIN\fR
 Change tape after writing \fIN\fRx1024 bytes.  If \fIN\fR is followed
 by a size suffix (see the subsection
 .B Size suffixes
@@ -639,7 +732,7 @@ When this option is used in conjunction with
 .BR \-\-multi\-volume ,
 .B tar
 will keep track of which volume of a multi-volume archive it is
-working in \fIFILE\fR. 
+working in \fIFILE\fR.
 .SS Device blocking
 .TP
 \fB\-b\fR, \fB\-\-blocking\-factor\fR=\fIBLOCKS\fR
@@ -647,7 +740,7 @@ Set record size to \fIBLOCKS\fRx\fB512\fR bytes.
 .TP
 \fB\-B\fR, \fB\-\-read\-full\-records\fR
 When listing or extracting, accept incomplete input records after
-end-of-file marker.  
+end-of-file marker.
 .TP
 \fB\-i\fR, \fB\-\-ignore\-zeros\fR
 Ignore zeroed blocks in archive.  Normally two consecutive 512-blocks
@@ -690,7 +783,8 @@ Same as \fB\-\-format=v7\fR.
 \fB\-\-pax\-option\fR=\fIkeyword\fR[[:]=\fIvalue\fR][,\fIkeyword\fR[[:]=\fIvalue\fR]]...
 Control pax keywords when creating \fBPAX\fR archives (\fB\-H
 pax\fR).  This option is equivalent to the \fB\-o\fR option of the
-.BR pax (1) utility.
+.BR pax (1)
+utility.
 .TP
 \fB\-\-posix\fR
 Same as \fB\-\-format=posix\fR.
@@ -703,7 +797,7 @@ use \fITEXT\fR as a globbing pattern for volume name.
 \fB\-a\fR, \fB\-\-auto\-compress\fR
 Use archive suffix to determine the compression program.
 .TP
-\fB\-\fRI, \fB\-\-use\-compress\-program\fI=\fICOMMAND\fR
+\fB\-I\fR, \fB\-\-use\-compress\-program\fI=\fICOMMAND\fR
 Filter data through \fICOMMAND\fR.  It must accept the \fB\-d\fR
 option, for decompression.  The argument can contain command line
 options.
@@ -735,9 +829,13 @@ Do not use archive suffix to determine the compression program.
 Filter the archive through
 .BR gzip (1).
 .TP
-\fB\-\fRZ, \fB\-\-compress\fR, \fB\-\-uncompress\fR
+\fB\-Z\fR, \fB\-\-compress\fR, \fB\-\-uncompress\fR
 Filter the archive through
 .BR compress (1).
+.TP
+\fB\-\-zstd\fR
+Filter the archive through
+.BR zstd (1).
 .SS Local file selection
 .TP
 \fB\-\-add\-file\fR=\fIFILE\fR
@@ -768,7 +866,8 @@ environment variable.  If it is not set, \fBexisting\fR is assumed.
 .RE
 .TP
 \fB\-C\fR, \fB\-\-directory\fR=\fIDIR\fR
-Change to directory DIR.
+Change to \fIDIR\fR before performing any operations.  This option is
+order-sensitive, i.e. it affects all options that follow.
 .TP
 \fB\-\-exclude\fR=\fIPATTERN\fR
 Exclude files matching \fIPATTERN\fR, a
@@ -841,13 +940,31 @@ Avoid descending automatically in directories.
 \fB\-\-no\-unquote\fR
 Do not unquote input file or member names.
 .TP
+\fB\-\-no\-verbatim\-files\-from\fR
+Treat each line read from a file list as if it were supplied in the
+command line.  I.e., leading and trailing whitespace is removed and,
+if the resulting string begins with a dash, it is treated as \fBtar\fR
+command line option.
+
+This is the default behavior.  The \fB\-\-no\-verbatim\-files\-from\fR
+option is provided as a way to restore it after
+\fB\-\-verbatim\-files\-from\fR option.
+
+This option is positional: it affects all \fB\-\-files\-from\fR
+options that occur after it in, until \fB\-\-verbatim\-files\-from\fR
+option or end of line, whichever occurs first.
+
+It is implied by the \fB\-\-no\-null\fR option.
+.TP
 \fB\-\-null\fR
-Instruct subsequent \fB\-T\fR options to read null-terminated names,
-disable handling of the \fB\-C\fR option read from the file.
+Instruct subsequent \fB\-T\fR options to read null-terminated names
+verbatim (disables special handling of names that start with a dash).
+
+See also \fB\-\-verbatim\-files\-from\fR.
 .TP
 \fB\-N\fR, \fB\-\-newer\fR=\fIDATE\fR, \fB\-\-after\-date\fR=\fIDATE\fR
 Only store files newer than DATE.  If \fIDATE\fR starts with \fB/\fR
-or \fB.\fR it is taken to be a file name; the ctime of that file is
+or \fB.\fR it is taken to be a file name; the mtime of that file is
 used as the date.
 .TP
 \fB\-\-one\-file\-system\fR
@@ -865,10 +982,43 @@ unless overridden by environment variable \fBSIMPLE_BACKUP_SUFFIX\fR.
 .TP
 \fB\-T\fR, \fB\-\-files\-from\fR=\fIFILE\fR
 Get names to extract or create from \fIFILE\fR.
+
+Unless specified otherwise, the \fIFILE\fR must contain a list of
+names separated by ASCII \fBLF\fR (i.e. one name per line).  The
+names read are handled the same way as command line arguments.  They
+undergo quote removal and word splitting, and any string that starts
+with a \fB\-\fR is handled as \fBtar\fR command line option.
+
+If this behavior is undesirable, it can be turned off using the
+\fB\-\-verbatim\-files\-from\fR option.
+
+The \fB\-\-null\fR option instructs \fBtar\fR that the names in
+\fIFILE\fR are separated by ASCII \fBNUL\fR character, instead of
+\fBLF\fR.  It is useful if the list is generated by
+.BR find (1)
+.B \-print0
+predicate.
 .TP
 \fB\-\-unquote\fR
 Unquote file or member names (default).
 .TP
+\fB\-\-verbatim\-files\-from\fR
+Treat each line obtained from a file list as a file name, even if it
+starts with a dash.  File lists are supplied with the
+\fB\-\-files\-from\fR (\fB\-T\fR) option.  The default behavior is to
+handle names supplied in file lists as if they were typed in the
+command line, i.e. any names starting with a dash are treated as
+\fBtar\fR options.  The \fB\-\-verbatim\-files\-from\fR option
+disables this behavior.
+
+This option affects all \fB\-\-files\-from\fR options that occur after
+it in the command line.  Its effect is reverted by the
+\fB\-\-no\-verbatim\-files\-from} option.
+
+This option is implied by the \fB\-\-null\fR option.
+
+See also \fB\-\-add\-file\fR.
+.TP
 \fB\-X\fR, \fB\-\-exclude\-from\fR=\fIFILE\fR
 Exclude files matching patterns listed in FILE.
 .SS  File name transformations
@@ -876,7 +1026,7 @@ Exclude files matching patterns listed in FILE.
 \fB\-\-strip\-components\fR=\fINUMBER\fR
 Strip \fINUMBER\fR leading components from file names on extraction.
 .TP
-\fB\-\-transform\fR=\fIEXPRESSION\dR, \fB\-\-xform\fR=\fIEXPRESSION\fR
+\fB\-\-transform\fR=\fIEXPRESSION\fR, \fB\-\-xform\fR=\fIEXPRESSION\fR
 Use sed replace \fIEXPRESSION\fR to transform file names.
 .SS File name matching options
 These options affect both exclude and include patterns.
@@ -912,6 +1062,9 @@ Display progress messages every \fIN\fRth record (default 10).
 \fB\-\-checkpoint\-action\fR=\fIACTION\fR
 Run \fIACTION\fR on each checkpoint.
 .TP
+\fB\-\-clamp\-mtime\fR
+Only set time when the file is more recent than what was given with \-\-mtime.
+.TP
 \fB\-\-full\-time\fR
 Print file time to its full resolution.
 .TP
@@ -954,7 +1107,7 @@ Show file or archive names after transformation by \fB\-\-strip\fR and
 Print total bytes after processing the archive.  If \fISIGNAL\fR is
 given, print total bytes when this signal is delivered.  Allowed
 signals are:
-.BR SIGHUP , 
+.BR SIGHUP ,
 .BR SIGQUIT ,
 .BR SIGINT ,
 .BR SIGUSR1 ", and"
@@ -965,7 +1118,11 @@ The \fBSIG\fR prefix can be omitted.
 Print file modification times in UTC.
 .TP
 \fB\-v\fR, \fB\-\-verbose\fR
-Verbosely list files processed.
+Verbosely list files processed.  Each instance of this option on the
+command line increases the verbosity level by one.  The maximum
+verbosity level is 3.  For a detailed discussion of how various
+verbosity levels affect tar's output, please refer to \fBGNU Tar
+Manual\fR, subsection 2.5.1 "\fBThe \-\-verbose Option\fR".
 .TP
 \fB\-\-warning\fR=\fIKEYWORD\fR
 Enable or disable warning messages identified by \fIKEYWORD\fR.  The
@@ -1018,9 +1175,18 @@ Keywords applicable for \fBtar --create\fR:
 .TP
 .B file-changed
 "%s: file changed as we read it"
+.TP
+.B failed-read
+Suppresses warnings about unreadable files or directories. This
+keyword applies only if used together with the
+.B \-\-ignore\-failed\-read
+option.
 .HP
 Keywords applicable for \fBtar --extract\fR:
 .TP
+.B existing\-file
+"%s: skipping existing file"
+.TP
 .B timestamp
 "%s: implausibly old time stamp %s"
 .br
@@ -1048,7 +1214,7 @@ default (unless \fB\-\-verbose\fR is used).  A common example of what
 you can get when using this warning is:
 
 .EX
-$ \fBtar --warning=decompress-program -x -f archive.Z
+$ tar --warning=decompress-program -x -f archive.Z
 tar (child): cannot run compress: No such file or directory
 tar (child): trying gzip
 .EE
@@ -1139,8 +1305,8 @@ failure during backup to a remote device.
 .BR lzop (1),
 .BR rmt (8),
 .BR symlink (7),
-.BR tar (5),
-.BR xz (1).
+.BR xz (1),
+.BR zstd (1).
 .PP
 Complete \fBtar\fR manual: run
 .B info tar
@@ -1156,7 +1322,7 @@ found at:
 .SH "BUG REPORTS"
 Report bugs to <bug\-tar@gnu.org>.
 .SH COPYRIGHT
-Copyright \(co 2013 Free Software Foundation, Inc.
+Copyright \(co 2013-2019 Free Software Foundation, Inc.
 .br
 .na
 License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
@@ -1171,4 +1337,3 @@ There is NO WARRANTY, to the extent permitted by law.
 .\" time-stamp-end: "\""
 .\" time-stamp-line-limit: 20
 .\" end:
-

doc/texify.sed

@@ -1,4 +1,4 @@
-# Copyright 2006-2007, 2013-2014 Free Software Foundation, Inc.
+# Copyright 2006-2021 Free Software Foundation, Inc.
 
 # This file is part of GNU tar.
 

doc/value.texi

@@ -1,6 +1,5 @@
 @c This is part of GNU tar manual.
-@c Copyright 1992, 1994-1997, 1999-2006, 2013-2014 Free Software
-@c Foundation, Inc.
+@c Copyright 1992--2021 Free Software Foundation, Inc.
 @c See file tar.texi for copying conditions.
 
 @macro GNUTAR

gnu/Makefile.am

@@ -1,3 +0,0 @@
-# Make the subset of Gnulib that GNU tar needs.
-include gnulib.mk
-AM_CFLAGS += $(GNULIB_WARN_CFLAGS) $(WERROR_CFLAGS)

gnulib.modules

@@ -1,7 +1,7 @@
 # List of gnulib modules needed for GNU tar.
 # A module name per line. Empty lines and comments are ignored.
 
-# Copyright 2005-2010, 2012-2014 Free Software Foundation, Inc.
+# Copyright 2005-2021 Free Software Foundation, Inc.
 
 # This file is part of GNU tar.
 
@@ -18,8 +18,8 @@
 # You should have received a copy of the GNU General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 
-acl
 alloca
+areadlinkat-with-size
 argmatch
 argp
 argp-version-etc
@@ -37,6 +37,7 @@ fchownat
 fcntl-h
 fdopendir
 fdutimensat
+file-has-acl
 fileblocks
 fnmatch-gnu
 fprintftime
@@ -65,8 +66,6 @@ modechange
 obstack
 openat
 parse-datetime
-posix_openpt
-ptsname
 priv-set
 progname
 quote
@@ -84,18 +83,20 @@ stat-time
 stdbool
 stdint
 stpcpy
+stdopen
 strdup-posix
 strerror
+strnlen
 strtoimax
 strtol
 strtoul
 strtoumax
 symlinkat
 timespec
+timespec-sub
 unlinkat
 unlinkdir
 unlocked-io
-unlockpt
 utimensat
 version-etc-fsf
 xalloc

lib/Makefile.am

@@ -1,7 +1,6 @@
 # Makefile for GNU tar library.			-*- Makefile -*-
 
-# Copyright 1994-1997, 1999-2001, 2003-2007, 2009-2010, 2013-2014 Free
-# Software Foundation, Inc.
+# Copyright 1994-2021 Free Software Foundation, Inc.
 
 # This file is part of GNU tar.
 
@@ -33,7 +32,6 @@ AM_CFLAGS = $(GNULIB_WARN_CFLAGS) $(WERROR_CFLAGS)
 noinst_HEADERS = \
  paxlib.h\
  rmt.h\
- stdopen.h\
  system.h\
  system-ioctl.h\
  wordsplit.h\
@@ -41,10 +39,8 @@ noinst_HEADERS = \
 
 libtar_a_SOURCES = \
   paxerror.c paxexit-status.c paxlib.h paxnames.c \
-  prepargs.c prepargs.h \
   rtapelib.c \
   rmt.h \
-  stdopen.c stdopen.h \
   system.h system-ioctl.h \
   wordsplit.c\
   xattr-at.c

lib/attr-xattr.in.h

@@ -1,5 +1,5 @@
 /* Replacement <attr/xattr.h> for platforms that lack it.
-   Copyright 2012-2014 Free Software Foundation, Inc.
+   Copyright 2012-2021 Free Software Foundation, Inc.
 
    This program is free software: you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by

lib/prepargs.c

@@ -1,83 +0,0 @@
-/* Parse arguments from a string and prepend them to an argv.
-   Copyright 1999-2001, 2007, 2013-2014 Free Software Foundation, Inc.
-
-   This program is free software; you can redistribute it and/or modify
-   it under the terms of the GNU General Public License as published by
-   the Free Software Foundation; either version 3, or (at your option)
-   any later version.
-
-   This program is distributed in the hope that it will be useful,
-   but WITHOUT ANY WARRANTY; without even the implied warranty of
-   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-   GNU General Public License for more details.
-
-   You should have received a copy of the GNU General Public License
-   along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
-
-/* Written by Paul Eggert <eggert@twinsun.com>.  */
-
-#ifdef HAVE_CONFIG_H
-# include <config.h>
-#endif
-#include "prepargs.h"
-#include <sys/types.h>
-#include <xalloc.h>
-
-#if HAVE_STRING_H
-# include <string.h>
-#endif
-
-#include <ctype.h>
-
-/* Find the white-space-separated options specified by OPTIONS, and
-   using BUF to store copies of these options, set ARGV[0], ARGV[1],
-   etc. to the option copies.  Return the number N of options found.
-   Do not set ARGV[N].  If ARGV is null, do not store ARGV[0]
-   etc.  Backslash can be used to escape whitespace (and backslashes).  */
-static int
-prepend_args (char const *options, char *buf, char **argv)
-{
-  char const *o = options;
-  char *b = buf;
-  int n = 0;
-
-  for (;;)
-    {
-      while (isspace ((unsigned char) *o))
-	o++;
-      if (!*o)
-	return n;
-      if (argv)
-	argv[n] = b;
-      n++;
-
-      do
-	if ((*b++ = *o++) == '\\' && *o)
-	  b[-1] = *o++;
-      while (*o && ! isspace ((unsigned char) *o));
-
-      *b++ = '\0';
-    }
-}
-
-/* Prepend the whitespace-separated options in OPTIONS to the argument
-   vector of a main program with argument count *PARGC and argument
-   vector *PARGV.  */
-void
-prepend_default_options (char const *options, int *pargc, char ***pargv)
-{
-  if (options)
-    {
-      char *buf = xmalloc (strlen (options) + 1);
-      int prepended = prepend_args (options, buf, (char **) 0);
-      int argc = *pargc;
-      char * const *argv = *pargv;
-      char **pp = (char **) xmalloc ((prepended + argc + 1) * sizeof *pp);
-      *pargc = prepended + argc;
-      *pargv = pp;
-      *pp++ = *argv++;
-      pp += prepend_args (options, buf, pp);
-      while ((*pp++ = *argv++))
-	continue;
-    }
-}

lib/prepargs.h

@@ -1,3 +0,0 @@
-/* Parse arguments from a string and prepend them to an argv.  */
-
-void prepend_default_options (char const *, int *, char ***);

lib/stdopen.c

@@ -1,76 +0,0 @@
-/* stdopen.c - ensure that the three standard file descriptors are in use
-
-   Copyright 2005, 2007, 2013-2014 Free Software Foundation, Inc.
-
-   This program is free software; you can redistribute it and/or modify
-   it under the terms of the GNU General Public License as published by
-   the Free Software Foundation; either version 3, or (at your option)
-   any later version.
-
-   This program is distributed in the hope that it will be useful,
-   but WITHOUT ANY WARRANTY; without even the implied warranty of
-   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-   GNU General Public License for more details.
-
-   You should have received a copy of the GNU General Public License
-   along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
-
-/* Written by Paul Eggert and Jim Meyering.  */
-
-#ifdef HAVE_CONFIG_H
-# include <config.h>
-#endif
-
-#include "stdopen.h"
-
-#include <sys/types.h>
-#include <sys/stat.h>
-#include <fcntl.h>
-#include <unistd.h>
-#include <errno.h>
-
-/* Try to ensure that all of the standard file numbers (0, 1, 2)
-   are in use.  Without this, each application would have to guard
-   every call to open, dup, fopen, etc. with tests to ensure they
-   don't use one of the special file numbers when opening a file.
-   Return false if at least one of the file descriptors is initially
-   closed and an attempt to reopen it fails.  Otherwise, return true.  */
-bool
-stdopen (void)
-{
-  int fd;
-  bool ok = true;
-
-  for (fd = 0; fd <= 2; fd++)
-    {
-      if (fcntl (fd, F_GETFD) < 0)
-        {
-          if (errno != EBADF)
-            ok = false;
-          else
-            {
-              static const int contrary_mode[]
-                = { O_WRONLY, O_RDONLY, O_RDONLY };
-              int mode = contrary_mode[fd];
-              int new_fd;
-              /* Open /dev/null with the contrary mode so that the typical
-                 read (stdin) or write (stdout, stderr) operation will fail.
-                 With descriptor 0, we can do even better on systems that
-                 have /dev/full, by opening that write-only instead of
-                 /dev/null.  The only drawback is that a write-provoked
-                 failure comes with a misleading errno value, ENOSPC.  */
-              if (mode == O_RDONLY
-                  || (new_fd = open ("/dev/full", mode) != fd))
-                new_fd = open ("/dev/null", mode);
-              if (new_fd != fd)
-                {
-                  if (0 <= new_fd)
-                    close (new_fd);
-                  ok = false;
-                }
-            }
-        }
-    }
-
-  return ok;
-}

lib/stdopen.h

@@ -1,16 +0,0 @@
-#ifndef STDOPEN_H
-# define STDOPEN_H 1
-
-# include <stdbool.h>
-
-# ifdef __cplusplus
-extern "C" {
-# endif
-
-bool stdopen (void);
-
-# ifdef __cplusplus
-}
-# endif
-
-#endif

lib/wordsplit.h

@@ -1,5 +1,5 @@
 /* wordsplit - a word splitter
-   Copyright (C) 2009-2014 Free Software Foundation, Inc.
+   Copyright (C) 2009-2018 Sergey Poznyakoff
 
    This program is free software; you can redistribute it and/or modify it
    under the terms of the GNU General Public License as published by the
@@ -12,52 +12,119 @@
    GNU General Public License for more details.
 
    You should have received a copy of the GNU General Public License along
-   with this program. If not, see <http://www.gnu.org/licenses/>.
-
-   Written by Sergey Poznyakoff
-*/
+   with this program. If not, see <http://www.gnu.org/licenses/>. */
 
 #ifndef __WORDSPLIT_H
 #define __WORDSPLIT_H
 
 #include <stddef.h>
 
-struct wordsplit
+#if 2 < __GNUC__ + (7 <= __GNUC_MINOR__)
+# define __WORDSPLIT_ATTRIBUTE_FORMAT(spec) __attribute__ ((__format__ spec))
+#else
+# define __WORDSPLIT_ATTRIBUTE_FORMAT(spec) /* empty */
+#endif
+
+typedef struct wordsplit wordsplit_t;
+
+/* Structure used to direct the splitting.  Members marked with [Input]
+   can be defined before calling wordsplit(), those marked with [Output]
+   provide return values when the function returns.  If neither mark is
+   used, the member is internal and must not be used by the caller.
+
+   In the comments below, the identifiers in parentheses indicate bits that
+   must be set (or unset, if starting with !) in ws_flags (if starting with
+   WRDSF_) or ws_options (if starting with WRDSO_) to initialize or use the
+   given member.
+   
+   If not redefined explicitly, most of them are set to some reasonable
+   default value upon entry to wordsplit(). */
+struct wordsplit            
 {
-  size_t ws_wordc;
-  char **ws_wordv;
-  size_t ws_offs;
-  size_t ws_wordn;
-  int ws_flags;
-  const char *ws_delim;
-  const char *ws_comment;
-  const char *ws_escape;
-  void (*ws_alloc_die) (struct wordsplit * wsp);
+  size_t ws_wordc;          /* [Output] Number of words in ws_wordv. */
+  char **ws_wordv;          /* [Output] Array of parsed out words. */
+  size_t ws_offs;           /* [Input] (WRDSF_DOOFFS) Number of initial
+			       elements in ws_wordv to fill with NULLs. */
+  size_t ws_wordn;          /* Number of elements ws_wordv can accomodate. */ 
+  unsigned ws_flags;        /* [Input] Flags passed to wordsplit. */
+  unsigned ws_options;      /* [Input] (WRDSF_OPTIONS)
+			       Additional options. */
+  size_t ws_maxwords;       /* [Input] (WRDSO_MAXWORDS) Return at most that
+			       many words */
+  size_t ws_wordi;          /* [Output] (WRDSF_INCREMENTAL) Total number of
+			       words returned so far */
+
+  const char *ws_delim;     /* [Input] (WRDSF_DELIM) Word delimiters. */
+  const char *ws_comment;   /* [Input] (WRDSF_COMMENT) Comment characters. */
+  const char *ws_escape[2]; /* [Input] (WRDSF_ESCAPE) Characters to be escaped
+			       with backslash. */
+  void (*ws_alloc_die) (wordsplit_t *wsp);
+                            /* [Input] (WRDSF_ALLOC_DIE) Function called when
+			       out of memory.  Must not return. */
   void (*ws_error) (const char *, ...)
                    __attribute__ ((__format__ (__printf__, 1, 2)));
+                            /* [Input] (WRDSF_ERROR) Function used for error
+			       reporting */
   void (*ws_debug) (const char *, ...)
                    __attribute__ ((__format__ (__printf__, 1, 2)));
+                            /* [Input] (WRDSF_DEBUG) Function used for debug
+			       output. */
+  const char **ws_env;      /* [Input] (WRDSF_ENV, !WRDSF_NOVAR) Array of
+			       environment variables. */
 
-  const char **ws_env;
-  const char *(*ws_getvar) (const char *, size_t, void *);
-  void *ws_closure;
+  char **ws_envbuf;
+  size_t ws_envidx;
+  size_t ws_envsiz;
+  
+  int (*ws_getvar) (char **ret, const char *var, size_t len, void *clos);
+                            /* [Input] (WRDSF_GETVAR, !WRDSF_NOVAR) Looks up
+			       the name VAR (LEN bytes long) in the table of
+			       variables and if found returns in memory
+			       location pointed to by RET the value of that
+			       variable.  Returns WRDSE_OK (0) on success,
+			       and an error code (see WRDSE_* defines below)
+			       on error.  User-specific errors can be returned
+			       by storing the error diagnostic string in RET
+			       and returning WRDSE_USERERR.
+                               Whatever is stored in RET, it must be allocated
+			       using malloc(3). */
+  void *ws_closure;         /* [Input] (WRDSF_CLOSURE) Passed as the CLOS
+			       argument to ws_getvar and ws_command. */
+  int (*ws_command) (char **ret, const char *cmd, size_t len, char **argv,
+                     void *clos);
+	                    /* [Input] (!WRDSF_NOCMD) Returns in the memory
+			       location pointed to by RET the expansion of
+			       the command CMD (LEN bytes long).  If WRDSO_ARGV
+			       option is set, ARGV contains CMD split out to
+			       words.  Otherwise ARGV is NULL.
 
-  const char *ws_input;
-  size_t ws_len;
-  size_t ws_endp;
-  int ws_errno;
+			       See ws_getvar for a discussion of possible
+			       return values. */
+	
+  const char *ws_input;     /* Input string (the S argument to wordsplit. */  
+  size_t ws_len;            /* Length of ws_input. */
+  size_t ws_endp;           /* Points past the last processed byte in
+			       ws_input. */
+  int ws_errno;             /* [Output] Error code, if an error occurred. */
+  char *ws_usererr;         /* Points to textual description of
+			       the error, if ws_errno is WRDSE_USERERR.  Must
+			       be allocated with malloc(3). */
   struct wordsplit_node *ws_head, *ws_tail;
+                            /* Doubly-linked list of parsed out nodes. */
+  int ws_lvl;               /* Invocation nesting level. */
 };
 
-/* Wordsplit flags.  Only 2 bits of a 32-bit word remain unused.
-   It is getting crowded... */
+/* Initial size for ws_env, if allocated automatically */
+#define WORDSPLIT_ENV_INIT 16
+
+/* Wordsplit flags. */
 /* Append the words found to the array resulting from a previous
    call. */
 #define WRDSF_APPEND            0x00000001
-/* Insert we_offs initial NULLs in the array ws_wordv.
+/* Insert ws_offs initial NULLs in the array ws_wordv.
    (These are not counted in the returned ws_wordc.) */
 #define WRDSF_DOOFFS            0x00000002
-/* Don't do command substitution. Reserved for future use. */
+/* Don't do command substitution. */
 #define WRDSF_NOCMD             0x00000004
 /* The parameter p resulted from a previous call to
    wordsplit(), and wordsplit_free() was not called. Reuse the
@@ -65,10 +132,8 @@ struct wordsplit
 #define WRDSF_REUSE             0x00000008
 /* Print errors */
 #define WRDSF_SHOWERR           0x00000010
-/* Consider it an error if an undefined shell variable
-   is expanded. */
+/* Consider it an error if an undefined variable is expanded. */
 #define WRDSF_UNDEF             0x00000020
-
 /* Don't do variable expansion. */
 #define WRDSF_NOVAR             0x00000040
 /* Abort on ENOMEM error */
@@ -79,7 +144,7 @@ struct wordsplit
 #define WRDSF_SQUOTE            0x00000200
 /* Handle double quotes */
 #define WRDSF_DQUOTE            0x00000400
-/* Handle quotes and escape directives */
+/* Handle single and double quotes */
 #define WRDSF_QUOTE             (WRDSF_SQUOTE|WRDSF_DQUOTE)
 /* Replace each input sequence of repeated delimiters with a single
    delimiter */
@@ -107,56 +172,106 @@ struct wordsplit
 /* Don't split input into words.  Useful for side effects. */
 #define WRDSF_NOSPLIT           0x00400000
 /* Keep undefined variables in place, instead of expanding them to
-   empty string */
+   empty strings. */
 #define WRDSF_KEEPUNDEF         0x00800000
 /* Warn about undefined variables */
 #define WRDSF_WARNUNDEF         0x01000000
 /* Handle C escapes */
 #define WRDSF_CESCAPES          0x02000000
-
 /* ws_closure is set */
 #define WRDSF_CLOSURE           0x04000000
 /* ws_env is a Key/Value environment, i.e. the value of a variable is
    stored in the element that follows its name. */
 #define WRDSF_ENV_KV            0x08000000
-
 /* ws_escape is set */
 #define WRDSF_ESCAPE            0x10000000
-
 /* Incremental mode */
 #define WRDSF_INCREMENTAL       0x20000000
+/* Perform pathname and tilde expansion */
+#define WRDSF_PATHEXPAND        0x40000000
+/* ws_options is initialized */
+#define WRDSF_OPTIONS           0x80000000
 
 #define WRDSF_DEFFLAGS	       \
   (WRDSF_NOVAR | WRDSF_NOCMD | \
    WRDSF_QUOTE | WRDSF_SQUEEZE_DELIMS | WRDSF_CESCAPES)
 
-#define WRDSE_EOF        0
+/* Remove the word that produces empty string after path expansion */
+#define WRDSO_NULLGLOB        0x00000001
+/* Print error message if path expansion produces empty string */
+#define WRDSO_FAILGLOB        0x00000002
+/* Allow a leading period to be matched by metacharacters. */
+#define WRDSO_DOTGLOB         0x00000004
+/* ws_command needs argv parameter */
+#define WRDSO_ARGV            0x00000008
+/* Keep backslash in unrecognized escape sequences in words */
+#define WRDSO_BSKEEP_WORD     0x00000010
+/* Handle octal escapes in words */
+#define WRDSO_OESC_WORD       0x00000020
+/* Handle hex escapes in words */
+#define WRDSO_XESC_WORD       0x00000040
+
+/* ws_maxwords field is initialized */
+#define WRDSO_MAXWORDS        0x00000080
+
+/* Keep backslash in unrecognized escape sequences in quoted strings */
+#define WRDSO_BSKEEP_QUOTE    0x00000100
+/* Handle octal escapes in quoted strings */
+#define WRDSO_OESC_QUOTE      0x00000200
+/* Handle hex escapes in quoted strings */
+#define WRDSO_XESC_QUOTE      0x00000400
+
+#define WRDSO_BSKEEP          WRDSO_BSKEEP_WORD     
+#define WRDSO_OESC            WRDSO_OESC_WORD       
+#define WRDSO_XESC            WRDSO_XESC_WORD       
+
+/* Indices into ws_escape */
+#define WRDSX_WORD  0
+#define WRDSX_QUOTE 1
+
+/* Set escape option F in WS for words (Q==0) or quoted strings (Q==1) */
+#define WRDSO_ESC_SET(ws,q,f) ((ws)->ws_options |= ((f) << 4*(q)))
+/* Test WS for escape option F for words (Q==0) or quoted strings (Q==1) */
+#define WRDSO_ESC_TEST(ws,q,f) ((ws)->ws_options & ((f) << 4*(q)))
+
+#define WRDSE_OK         0
+#define WRDSE_EOF        WRDSE_OK
 #define WRDSE_QUOTE      1
 #define WRDSE_NOSPACE    2
-#define WRDSE_NOSUPP     3
-#define WRDSE_USAGE      4
-#define WRDSE_CBRACE     5
-#define WRDSE_UNDEF      6
-#define WRDSE_NOINPUT    7
+#define WRDSE_USAGE      3
+#define WRDSE_CBRACE     4
+#define WRDSE_UNDEF      5
+#define WRDSE_NOINPUT    6
+#define WRDSE_PAREN      7
+#define WRDSE_GLOBERR    8
+#define WRDSE_USERERR    9
 
-int wordsplit (const char *s, struct wordsplit *p, int flags);
-int wordsplit_len (const char *s, size_t len,
-		      struct wordsplit *p, int flags);
-void wordsplit_free (struct wordsplit *p);
-void wordsplit_free_words (struct wordsplit *ws);
+int wordsplit (const char *s, wordsplit_t *ws, unsigned flags);
+int wordsplit_len (const char *s, size_t len, wordsplit_t *ws, unsigned flags);
+void wordsplit_free (wordsplit_t *ws);
+void wordsplit_free_words (wordsplit_t *ws);
+void wordsplit_free_envbuf (wordsplit_t *ws);
+int wordsplit_get_words (wordsplit_t *ws, size_t *wordc, char ***wordv);
+
+static inline void wordsplit_getwords (wordsplit_t *ws, size_t *wordc, char ***wordv)
+  __attribute__ ((deprecated));
+
+static inline void
+wordsplit_getwords (wordsplit_t *ws, size_t *wordc, char ***wordv)
+{
+  wordsplit_get_words (ws, wordc, wordv);
+}
+
+int wordsplit_append (wordsplit_t *wsp, int argc, char **argv);
 
 int wordsplit_c_unquote_char (int c);
 int wordsplit_c_quote_char (int c);
-size_t wordsplit_c_quoted_length (const char *str, int quote_hex,
-				  int *quote);
-void wordsplit_general_unquote_copy (char *dst, const char *src, size_t n,
-				     const char *escapable);
-void wordsplit_sh_unquote_copy (char *dst, const char *src, size_t n);
-void wordsplit_c_unquote_copy (char *dst, const char *src, size_t n);
+size_t wordsplit_c_quoted_length (const char *str, int quote_hex, int *quote);
 void wordsplit_c_quote_copy (char *dst, const char *src, int quote_hex);
 
-void wordsplit_perror (struct wordsplit *ws);
-const char *wordsplit_strerror (struct wordsplit *ws);
+void wordsplit_perror (wordsplit_t *ws);
+const char *wordsplit_strerror (wordsplit_t *ws);
 
+void wordsplit_clearerr (wordsplit_t *ws);
 
 #endif

lib/xattr-at.c

@@ -1,7 +1,7 @@
 /* openat-style fd-relative functions for operating with extended file
    attributes.
 
-   Copyright 2012-2014 Free Software Foundation, Inc.
+   Copyright 2012-2021 Free Software Foundation, Inc.
 
    This program is free software: you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
@@ -31,6 +31,8 @@
 
 #include "openat-priv.h"
 
+#ifdef HAVE_XATTRS
+
 /* setxattrat */
 #define AT_FUNC_NAME setxattrat
 #define AT_FUNC_F1 setxattr
@@ -108,3 +110,5 @@
 #undef AT_FUNC_RESULT
 #undef AT_FUNC_POST_FILE_PARAM_DECLS
 #undef AT_FUNC_POST_FILE_ARGS
+
+#endif /* HAVE_XATTRS */

lib/xattr-at.h

@@ -1,7 +1,7 @@
 /* Prototypes for openat-style fd-relative functions for operating with
    extended file attributes.
 
-   Copyright 2012-2014 Free Software Foundation, Inc.
+   Copyright 2012-2021 Free Software Foundation, Inc.
 
    This program is free software: you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by

po/.gitignore

@@ -1,3 +1,6 @@
+/Makefile.in.in
+/Makevars.template~
+/Makefile.in.in~
 *.gmo
 *.mo
 *.po

po/POTFILES.in

@@ -1,7 +1,6 @@
 # List of files which contain translatable strings.
 
-# Copyright 1996, 1999-2000, 2003-2005, 2007, 2013-2014 Free Software
-# Foundation, Inc.
+# Copyright 1996-2021 Free Software Foundation, Inc.
 
 # This file is part of GNU tar.
 

scripts/Makefile.am

@@ -1,6 +1,6 @@
 # Make GNU tar scripts.
 
-# Copyright 2004, 2006-2007, 2013-2014 Free Software Foundation, Inc.
+# Copyright 2004-2021 Free Software Foundation, Inc.
 
 # This file is part of GNU tar.
 

scripts/backup-specs

@@ -82,7 +82,7 @@ SLEEP_MESSAGE="`awk '
    }' /dev/null`"
 
 
-# Copyright 2004, 2007, 2013-2014 Free Software Foundation, Inc.
+# Copyright 2004-2021 Free Software Foundation, Inc.
 
 # This file is part of GNU tar.
 

scripts/backup.in

@@ -1,7 +1,7 @@
 #! /bin/sh
 # Make backups.
 
-# Copyright 2004-2006, 2013 Free Software Foundation
+# Copyright 2004-2006, 2013, 2019 Free Software Foundation
 
 # This file is part of GNU tar.
 
@@ -149,13 +149,16 @@ message 20 "Variables:"
 message 20 "BACKUP_DIRS=$BACKUP_DIRS"
 message 20 "BACKUP_FILES=$BACKUP_FILES"
 
-# The buch of commands below is run in a subshell for which all output is
+# The bunch of commands below is run in a subshell for which all output is
 # piped through 'tee' to the logfile.  Doing this, instead of having
 # multiple pipelines all over the place, is cleaner and allows access to
 # the exit value from various commands more easily.
 (
  message 1 "preparing tapes"
- $MT_BEGIN "${TAPE_FILE}"
+ if ! $MT_BEGIN "${TAPE_FILE}"; then
+    echo >&2 "$0: tape initialization failed"
+    exit 1
+ fi 
  rm -f "${VOLNO_FILE}"
 
  message 1 "processing backup directories"
@@ -191,7 +194,7 @@ message 20 "BACKUP_FILES=$BACKUP_FILES"
      # 'rsh' doesn't exit with the exit status of the remote command.  What
      # stupid lossage.  TODO: think of a reliable workaround.
      if [ $? -ne 0 ] ; then
-	 echo "Backup of ${1} failed." 1>&2
+	 echo "$0: backup of ${1} failed." 1>&2
          # I'm assuming that the tar will have written an empty
          # file to the tape, otherwise I should do a cat here.
      else
@@ -238,17 +241,17 @@ message 20 "BACKUP_FILES=$BACKUP_FILES"
  else
      echo "No miscellaneous files specified"
  fi
-
  message 1 "final cleanup"
 
  $MT_REWIND "${TAPE_FILE}"
  $MT_OFFLINE "${TAPE_FILE}"
  echo "."
 ) 2>&1 | tee -a "${LOGFILE}"
+RC=$?
 
 if test "${ADMINISTRATOR}" != NONE; then
      echo "Sending the dump log to ${ADMINISTRATOR}"
      mail -s "Results of backup started ${startdate}" ${ADMINISTRATOR} < "${LOGFILE}"
 fi
-
+exit $RC
 # EOF

scripts/backup.sh.in

@@ -1,7 +1,7 @@
 #! /bin/sh
 # Make backups.
 
-# Copyright 2004-2006, 2013-2014 Free Software Foundation, Inc.
+# Copyright 2004-2021 Free Software Foundation, Inc.
 
 # This file is part of GNU tar.
 
@@ -311,7 +311,9 @@ backup_host() {
         CMD="exec ${TAR_PART1} -f \"${TAPE_FILE}\" $@"
         message 10 "CMD: $CMD"
         sh -c "$CMD"
-        message 10 "RC: $?"
+        RC=$?
+        message 10 "RC: $RC"
+        return $RC
     fi
 }
 

scripts/dump-remind.in

@@ -8,7 +8,7 @@
 # interested parties that a tape for the next volume of the backup needs to
 # be put in the tape drive.
 
-# Copyright 2004-2005, 2010, 2012-2014 Free Software Foundation, Inc.
+# Copyright 2004-2021 Free Software Foundation, Inc.
 
 # This file is part of GNU tar.
 

scripts/restore.in

@@ -1,7 +1,7 @@
 #! /bin/sh
 # Restore backups.
 
-# Copyright 2004, 2006, 2013-2014 Free Software Foundation, Inc.
+# Copyright 2004-2021 Free Software Foundation, Inc.
 
 # This file is part of GNU tar.
 

scripts/tar-snapshot-edit

@@ -1,6 +1,6 @@
 #! /usr/bin/perl -w
 # Display and edit the 'dev' field in tar's snapshots
-# Copyright 2007, 2011, 2013-2014 Free Software Foundation, Inc.
+# Copyright 2007-2021 Free Software Foundation, Inc.
 
 # This file is part of GNU tar.
 
@@ -408,7 +408,7 @@ sub check_field_values ($) {
     }
     if ( $msg ne "" ) {
 	$error_found = 1;
-	print "\n    shapshot file header:\n";
+	print "\n    snapshot file header:\n";
 	print $msg;
     }
 

scripts/tarcat

@@ -3,7 +3,7 @@
 # concatenates a GNU tar multi-volume archive into a single tar archive.
 # Author: Bruno Haible <bruno@clisp.org>, Sergey Poznyakoff <gray@gnu.org.ua>
 
-# Copyright 2004-2005, 2010, 2013-2014 Free Software Foundation, Inc.
+# Copyright 2004-2021 Free Software Foundation, Inc.
 
 # This file is part of GNU tar.
 

scripts/xsparse.c

@@ -1,7 +1,7 @@
 /* xsparse - expands compressed sparse file images extracted from GNU tar
    archives.
 
-   Copyright 2006-2007, 2010, 2013-2014 Free Software Foundation, Inc.
+   Copyright 2006-2021 Free Software Foundation, Inc.
 
    This file is part of GNU tar.
 

src/Makefile.am

@@ -1,7 +1,6 @@
 # Makefile for GNU tar sources.
 
-# Copyright 1994-1997, 1999-2001, 2003, 2006-2007, 2009, 2013-2014 Free
-# Software Foundation, Inc.
+# Copyright 1994-2021 Free Software Foundation, Inc.
 
 # This file is part of GNU tar.
 
@@ -33,6 +32,7 @@ tar_SOURCES = \
  xheader.c\
  incremen.c\
  list.c\
+ map.c\
  misc.c\
  names.c\
  sparse.c\
@@ -52,7 +52,3 @@ AM_CFLAGS = $(WARN_CFLAGS) $(WERROR_CFLAGS)
 LDADD = ../lib/libtar.a ../gnu/libgnu.a $(LIBINTL) $(LIBICONV)
 
 tar_LDADD = $(LIBS) $(LDADD) $(LIB_CLOCK_GETTIME) $(LIB_EACCESS) $(LIB_SELINUX)
-
-if TAR_LIB_ATTR
-tar_LDADD += -lattr
-endif

src/arith.h

@@ -1,5 +1,5 @@
 /* Long integers, for GNU tar.
-   Copyright 1999, 2007, 2013-2014 Free Software Foundation, Inc.
+   Copyright 1999-2021 Free Software Foundation, Inc.
 
    This file is part of GNU tar.
 

src/buffer.c

@@ -1,7 +1,6 @@
 /* Buffer management for tar.
 
-   Copyright 1988, 1992-1994, 1996-1997, 1999-2010, 2013-2014 Free
-   Software Foundation, Inc.
+   Copyright 1988-2021 Free Software Foundation, Inc.
 
    This file is part of GNU tar.
 
@@ -105,13 +104,20 @@ bool write_archive_to_stdout;
 
 /* When creating a multi-volume archive, each 'bufmap' represents
    a member stored (perhaps partly) in the current record buffer.
+   Bufmaps are form a single-linked list in chronological order.
+
    After flushing the record to the output media, all bufmaps that
-   represent fully written members are removed from the list, then
-   the sizeleft and start numbers in the remaining bufmaps are updated.
+   represent fully written members are removed from the list, the
+   nblocks and sizeleft values in the bufmap_head and start values
+   in all remaining bufmaps are updated.  The information stored
+   in bufmap_head is used to form the volume header.
 
    When reading from a multi-volume archive, the list degrades to a
    single element, which keeps information about the member currently
-   being read.
+   being read.  In that case the sizeleft member is updated explicitly
+   from the extractor code by calling the mv_size_left function.  The
+   information from bufmap_head is compared with the volume header data
+   to ensure that subsequent volumes are fed in the right order.
 */
 
 struct bufmap
@@ -121,6 +127,7 @@ struct bufmap
   char *file_name;              /* Name of the stored file */
   off_t sizetotal;              /* Size of the stored file */
   off_t sizeleft;               /* Size left to read/write */
+  size_t nblocks;               /* Number of blocks written since reset */
 };
 static struct bufmap *bufmap_head, *bufmap_tail;
 
@@ -145,6 +152,7 @@ mv_begin_write (const char *file_name, off_t totsize, off_t sizeleft)
       bp->file_name = xstrdup (file_name);
       bp->sizetotal = totsize;
       bp->sizeleft = sizeleft;
+      bp->nblocks = 0;
     }
 }
 
@@ -155,8 +163,7 @@ bufmap_locate (size_t off)
 
   for (map = bufmap_head; map; map = map->next)
     {
-      if (!map->next
-	  || off < map->next->start * BLOCKSIZE)
+      if (!map->next || off < map->next->start * BLOCKSIZE)
 	break;
     }
   return map;
@@ -185,7 +192,10 @@ bufmap_reset (struct bufmap *map, ssize_t fixup)
   if (map)
     {
       for (; map; map = map->next)
-	map->start += fixup;
+	{
+	  map->start += fixup;
+	  map->nblocks = 0;
+	}
     }
 }
 
@@ -270,7 +280,8 @@ enum compress_type {
   ct_lzip,
   ct_lzma,
   ct_lzop,
-  ct_xz
+  ct_xz,
+  ct_zstd
 };
 
 static enum compress_type archive_compression_type = ct_none;
@@ -299,6 +310,7 @@ static struct zip_magic const magic[] = {
   { ct_lzma,     6, "\xFFLZMA" },
   { ct_lzop,     4, "\211LZO" },
   { ct_xz,       6, "\xFD" "7zXZ" },
+  { ct_zstd,     4, "\x28\xB5\x2F\xFD" },
 };
 
 #define NMAGIC (sizeof(magic)/sizeof(magic[0]))
@@ -314,6 +326,7 @@ static struct zip_program zip_program[] = {
   { ct_lzma,     XZ_PROGRAM,       "-J" },
   { ct_lzop,     LZOP_PROGRAM,     "--lzop" },
   { ct_xz,       XZ_PROGRAM,       "-J" },
+  { ct_zstd,     ZSTD_PROGRAM,     "--zstd" },
   { ct_none }
 };
 
@@ -391,7 +404,12 @@ check_compressed_archive (bool *pshort)
   /* Restore global values */
   read_full_records = sfr;
 
-  if (tar_checksum (record_start, true) == HEADER_SUCCESS)
+  if (record_start != record_end /* no files smaller than BLOCKSIZE */
+      && (strcmp (record_start->header.magic, TMAGIC) == 0
+          || strcmp (record_start->buffer + offsetof (struct posix_header,
+                                                      magic),
+                     OLDGNU_MAGIC) == 0)
+      && tar_checksum (record_start, true) == HEADER_SUCCESS)
     /* Probably a valid header */
     return ct_tar;
 
@@ -495,7 +513,7 @@ print_stats (FILE *fp, const char *text, tarlong numbytes)
   char abbr[LONGEST_HUMAN_READABLE + 1];
   char rate[LONGEST_HUMAN_READABLE + 1];
   int n = 0;
-  
+
   int human_opts = human_autoscale | human_base_1024 | human_SI | human_B;
 
   if (text && text[0])
@@ -512,12 +530,12 @@ print_stats (FILE *fp, const char *text, tarlong numbytes)
    before each data item (bytes read, written, deleted, in that order).
    EOR is a delimiter to output after each item (used only if deleting
    from the archive), EOL is a delimiter to add at the end of the output
-   line. */ 
+   line. */
 int
-format_total_stats (FILE *fp, const char **formats, int eor, int eol)
+format_total_stats (FILE *fp, char const *const *formats, int eor, int eol)
 {
   int n;
-  
+
   switch (subcommand_option)
     {
     case CREATE_SUBCOMMAND:
@@ -536,7 +554,7 @@ format_total_stats (FILE *fp, const char **formats, int eor, int eol)
 
 	fputc (eor, fp);
 	n++;
-	
+
         n += print_stats (fp, formats[TF_WRITE],
 			  prev_written + bytes_written);
 
@@ -570,7 +588,7 @@ format_total_stats (FILE *fp, const char **formats, int eor, int eol)
   return n;
 }
 
-const char *default_total_format[] = {
+static char const *const default_total_format[] = {
   N_("Total bytes read"),
   /* Amanda 2.4.1p1 looks for "Total bytes written: [0-9][0-9]*".  */
   N_("Total bytes written"),
@@ -865,12 +883,19 @@ _flush_write (void)
       if (map)
 	{
 	  size_t delta = status - map->start * BLOCKSIZE;
+	  ssize_t diff;
+	  map->nblocks += delta / BLOCKSIZE;
 	  if (delta > map->sizeleft)
 	    delta = map->sizeleft;
 	  map->sizeleft -= delta;
 	  if (map->sizeleft == 0)
-	    map = map->next;
-	  bufmap_reset (map, map ? (- map->start) : 0);
+	    {
+	      diff = map->start + map->nblocks;
+	      map = map->next;
+	    }
+	  else
+	    diff = map->start;
+	  bufmap_reset (map, - diff);
 	}
     }
   return status;
@@ -980,18 +1005,28 @@ short_read (size_t status)
 void
 flush_archive (void)
 {
-  size_t buffer_level = current_block->buffer - record_start->buffer;
-  record_start_block += record_end - record_start;
-  current_block = record_start;
-  record_end = record_start + blocking_factor;
+  size_t buffer_level;
 
   if (access_mode == ACCESS_READ && time_to_start_writing)
     {
       access_mode = ACCESS_WRITE;
       time_to_start_writing = false;
       backspace_output ();
+      if (record_end - record_start < blocking_factor)
+	{
+	  memset (record_end, 0,
+		  (blocking_factor - (record_end - record_start))
+		  * BLOCKSIZE);
+	  record_end = record_start + blocking_factor;
+	  return;
+	}
     }
 
+  buffer_level = current_block->buffer - record_start->buffer;
+  record_start_block += record_end - record_start;
+  current_block = record_start;
+  record_end = record_start + blocking_factor;
+
   switch (access_mode)
     {
     case ACCESS_READ:
@@ -1031,7 +1066,7 @@ backspace_output (void)
 
     /* Seek back to the beginning of this record and start writing there.  */
 
-    position -= record_size;
+    position -= record_end->buffer - record_start->buffer;
     if (position < 0)
       position = 0;
     if (rmtlseek (archive, position, SEEK_SET) != position)
@@ -1092,9 +1127,9 @@ close_archive (void)
 {
   if (time_to_start_writing || access_mode == ACCESS_WRITE)
     {
-      flush_archive ();
-      if (current_block > record_start)
-        flush_archive ();
+      do
+	flush_archive ();
+      while (current_block > record_start);
     }
 
   compute_duration ();
@@ -1112,6 +1147,16 @@ close_archive (void)
   bufmap_free (NULL);
 }
 
+void
+write_fatal_details (char const *name, ssize_t status, size_t size)
+{
+  write_error_details (name, status, size);
+  if (rmtclose (archive) != 0)
+    close_error (*archive_name_cursor);
+  sys_wait_for_child (child_pid, false);
+  fatal_exit ();
+}
+
 /* Called to initialize the global volume number.  */
 void
 init_volume_number (void)
@@ -1253,8 +1298,7 @@ change_tape_menu (FILE *read_file)
               sys_spawn_shell ();
               break;
             }
-          /* FALL THROUGH */
-
+	  FALLTHROUGH;
         default:
           fprintf (stderr, _("Invalid input. Type ? for help.\n"));
         }
@@ -1406,7 +1450,10 @@ try_new_volume (void)
 
   header = find_next_block ();
   if (!header)
-    return false;
+    {
+      WARN ((0, 0, _("This does not look like a tar archive")));
+      return false;
+    }
 
   switch (header->header.typeflag)
     {
@@ -1416,7 +1463,7 @@ try_new_volume (void)
 	if (read_header (&header, &dummy, read_header_x_global)
 	    != HEADER_SUCCESS_EXTENDED)
 	  {
-	    ERROR ((0, 0, _("This does not look like a tar archive")));
+	    WARN ((0, 0, _("This does not look like a tar archive")));
 	    return false;
 	  }
 
@@ -1445,7 +1492,7 @@ try_new_volume (void)
 	    break;
 
 	  default:
-	    ERROR ((0, 0, _("This does not look like a tar archive")));
+	    WARN ((0, 0, _("This does not look like a tar archive")));
 	    return false;
 	  }
         break;
@@ -1455,18 +1502,17 @@ try_new_volume (void)
       if (!read_header0 (&dummy))
         return false;
       tar_stat_destroy (&dummy);
-      assign_string (&volume_label, current_header->header.name);
+      ASSIGN_STRING_N (&volume_label, current_header->header.name);
       set_next_block_after (header);
       header = find_next_block ();
       if (header->header.typeflag != GNUTYPE_MULTIVOL)
         break;
-      /* FALL THROUGH */
-
+      FALLTHROUGH;
     case GNUTYPE_MULTIVOL:
       if (!read_header0 (&dummy))
         return false;
       tar_stat_destroy (&dummy);
-      assign_string (&continued_file_name, current_header->header.name);
+      ASSIGN_STRING_N (&continued_file_name, current_header->header.name);
       continued_file_size =
         UINTMAX_FROM_HEADER (current_header->header.size);
       continued_file_offset =
@@ -1480,8 +1526,14 @@ try_new_volume (void)
   if (bufmap_head)
     {
       uintmax_t s;
-      if (!continued_file_name
-          || strcmp (continued_file_name, bufmap_head->file_name))
+      if (!continued_file_name)
+	{
+	  WARN ((0, 0, _("%s is not continued on this volume"),
+		 quote (bufmap_head->file_name)));
+	  return false;
+	}
+
+      if (strcmp (continued_file_name, bufmap_head->file_name))
         {
           if ((archive_format == GNU_FORMAT || archive_format == OLDGNU_FORMAT)
               && strlen (bufmap_head->file_name) >= NAME_FIELD_SIZE
@@ -1603,15 +1655,7 @@ match_volume_label (void)
 		      quote (volume_label_option)));
       if (label->header.typeflag == GNUTYPE_VOLHDR)
 	{
-	  if (memchr (label->header.name, '\0', sizeof label->header.name))
-	    assign_string (&volume_label, label->header.name);
-	  else
-	    {
-	      volume_label = xmalloc (sizeof (label->header.name) + 1);
-	      memcpy (volume_label, label->header.name,
-		      sizeof (label->header.name));
-	      volume_label[sizeof (label->header.name)] = 0;
-	    }
+	  ASSIGN_STRING_N (&volume_label, label->header.name);
 	}
       else if (label->header.typeflag == XGLTYPE)
 	{
@@ -1647,8 +1691,7 @@ _write_volume_label (const char *str)
       memset (label, 0, BLOCKSIZE);
 
       strcpy (label->header.name, str);
-      assign_string (&current_stat_info.file_name,
-                     label->header.name);
+      assign_string (&current_stat_info.file_name, label->header.name);
       current_stat_info.had_trailing_slash =
         strip_trailing_slashes (current_stat_info.file_name);
 
@@ -1679,7 +1722,6 @@ add_chunk_header (struct bufmap *map)
 {
   if (archive_format == POSIX_FORMAT)
     {
-      off_t block_ordinal;
       union block *blk;
       struct tar_stat_info st;
 
@@ -1689,16 +1731,15 @@ add_chunk_header (struct bufmap *map)
       st.stat.st_uid = getuid ();
       st.stat.st_gid = getgid ();
       st.orig_file_name = xheader_format_name (&st,
-                                               "%d/GNUFileParts.%p/%f.%n",
+                                               "%d/GNUFileParts/%f.%n",
                                                volno);
       st.file_name = st.orig_file_name;
       st.archive_file_size = st.stat.st_size = map->sizeleft;
 
-      block_ordinal = current_block_ordinal ();
       blk = start_header (&st);
       if (!blk)
         abort (); /* FIXME */
-      finish_header (&st, blk, block_ordinal);
+      simple_finish_header (write_extended (false, &st, blk));
       free (st.orig_file_name);
     }
 }
@@ -1720,15 +1761,19 @@ gnu_add_multi_volume_header (struct bufmap *map)
 {
   int tmp;
   union block *block = find_next_block ();
+  size_t len = strlen (map->file_name);
 
-  if (strlen (map->file_name) > NAME_FIELD_SIZE)
-    WARN ((0, 0,
-           _("%s: file name too long to be stored in a GNU multivolume header, truncated"),
-           quotearg_colon (map->file_name)));
+  if (len > NAME_FIELD_SIZE)
+    {
+      WARN ((0, 0,
+	     _("%s: file name too long to be stored in a GNU multivolume header, truncated"),
+	     quotearg_colon (map->file_name)));
+      len = NAME_FIELD_SIZE;
+    }
 
   memset (block, 0, BLOCKSIZE);
 
-  strncpy (block->header.name, map->file_name, NAME_FIELD_SIZE);
+  memcpy (block->header.name, map->file_name, len);
   block->header.typeflag = GNUTYPE_MULTIVOL;
 
   OFF_TO_CHARS (map->sizeleft, block->header.size);

src/checkpoint.c

@@ -1,6 +1,6 @@
 /* Checkpoint management for tar.
 
-   Copyright 2007, 2013-2014 Free Software Foundation, Inc.
+   Copyright 2007-2021 Free Software Foundation, Inc.
 
    This file is part of GNU tar.
 
@@ -23,6 +23,7 @@
 #include <sys/ioctl.h>
 #include <termios.h>
 #include "fprintftime.h"
+#include <signal.h>
 
 enum checkpoint_opcode
   {
@@ -32,7 +33,8 @@ enum checkpoint_opcode
     cop_ttyout,
     cop_sleep,
     cop_exec,
-    cop_totals
+    cop_totals,
+    cop_wait
   };
 
 struct checkpoint_action
@@ -43,6 +45,7 @@ struct checkpoint_action
   {
     time_t time;
     char *command;
+    int signal;
   } v;
 };
 
@@ -52,6 +55,16 @@ static unsigned checkpoint;
 /* List of checkpoint actions */
 static struct checkpoint_action *checkpoint_action, *checkpoint_action_tail;
 
+/* State of the checkpoint system */
+enum {
+  CHKP_INIT,       /* Needs initialization */
+  CHKP_COMPILE,    /* Actions are being compiled */
+  CHKP_RUN         /* Actions are being run */
+};
+static int checkpoint_state;
+/* Blocked signals */
+static sigset_t sigs;
+
 static struct checkpoint_action *
 alloc_action (enum checkpoint_opcode opcode)
 {
@@ -85,6 +98,12 @@ checkpoint_compile_action (const char *str)
 {
   struct checkpoint_action *act;
 
+  if (checkpoint_state == CHKP_INIT)
+    {
+      sigemptyset (&sigs);
+      checkpoint_state = CHKP_COMPILE;
+    }
+  
   if (strcmp (str, ".") == 0 || strcmp (str, "dot") == 0)
     alloc_action (cop_dot);
   else if (strcmp (str, "bell") == 0)
@@ -117,6 +136,12 @@ checkpoint_compile_action (const char *str)
     }
   else if (strcmp (str, "totals") == 0)
     alloc_action (cop_totals);
+  else if (strncmp (str, "wait=", 5) == 0)
+    {
+      act = alloc_action (cop_wait);
+      act->v.signal = decode_signal (str + 5);
+      sigaddset (&sigs, act->v.signal);
+    }
   else
     FATAL_ERROR ((0, 0, _("%s: unknown checkpoint action"), str));
 }
@@ -124,15 +149,24 @@ checkpoint_compile_action (const char *str)
 void
 checkpoint_finish_compile (void)
 {
-  if (checkpoint_option)
+  if (checkpoint_state == CHKP_INIT
+      && checkpoint_option
+      && !checkpoint_action)
     {
-      if (!checkpoint_action)
-	/* Provide a historical default */
-	checkpoint_compile_action ("echo");
+      /* Provide a historical default */
+      checkpoint_compile_action ("echo");
+    }
+
+  if (checkpoint_state == CHKP_COMPILE)
+    {
+      sigprocmask (SIG_BLOCK, &sigs, NULL);
+
+      if (!checkpoint_option)
+	/* set default checkpoint rate */
+	checkpoint_option = DEFAULT_CHECKPOINT;
+
+      checkpoint_state = CHKP_RUN;
     }
-  else if (checkpoint_action)
-    /* Otherwise, set default checkpoint rate */
-    checkpoint_option = DEFAULT_CHECKPOINT;
 }
 
 static const char *checkpoint_total_format[] = {
@@ -390,6 +424,13 @@ run_checkpoint_actions (bool do_write)
 	case cop_totals:
 	  compute_duration ();
 	  print_total_stats ();
+	  break;
+
+	case cop_wait:
+	  {
+	    int n;
+	    sigwait (&sigs, &n);
+	  }
 	}
     }
 }

src/common.h

@@ -1,7 +1,6 @@
 /* Common declarations for the tar program.
 
-   Copyright 1988, 1992-1994, 1996-1997, 1999-2010, 2012-2014 Free
-   Software Foundation, Inc.
+   Copyright 1988-2021 Free Software Foundation, Inc.
 
    This file is part of GNU tar.
 
@@ -44,6 +43,12 @@
 # define GLOBAL extern
 #endif
 
+#if 7 <= __GNUC__
+# define FALLTHROUGH __attribute__ ((__fallthrough__))
+#else
+# define FALLTHROUGH ((void) 0)
+#endif
+
 #define TAREXIT_SUCCESS PAXEXIT_SUCCESS
 #define TAREXIT_DIFFERS PAXEXIT_DIFFERS
 #define TAREXIT_FAILURE PAXEXIT_FAILURE
@@ -211,13 +216,20 @@ GLOBAL bool multi_volume_option;
    do not get archived (also see after_date_option above).  */
 GLOBAL struct timespec newer_mtime_option;
 
-/* If true, override actual mtime (see below) */
-GLOBAL bool set_mtime_option;
-/* Value to be put in mtime header field instead of the actual mtime */
+enum set_mtime_option_mode
+{
+  USE_FILE_MTIME,
+  FORCE_MTIME,
+  CLAMP_MTIME,
+};
+
+/* Override actual mtime if set to FORCE_MTIME or CLAMP_MTIME */
+GLOBAL enum set_mtime_option_mode set_mtime_option;
+/* Value to use when forcing or clamping the mtime header field. */
 GLOBAL struct timespec mtime_option;
 
-/* Return true if newer_mtime_option is initialized.  */
-#define NEWER_OPTION_INITIALIZED(opt) (0 <= (opt).tv_nsec)
+/* Return true if mtime_option or newer_mtime_option is initialized.  */
+#define TIME_OPTION_INITIALIZED(opt) (0 <= (opt).tv_nsec)
 
 /* Return true if the struct stat ST's M time is less than
    newer_mtime_option.  */
@@ -280,6 +292,19 @@ GLOBAL bool sparse_option;
 GLOBAL unsigned tar_sparse_major;
 GLOBAL unsigned tar_sparse_minor;
 
+enum hole_detection_method
+  {
+    HOLE_DETECTION_DEFAULT,
+    HOLE_DETECTION_RAW,
+    HOLE_DETECTION_SEEK
+  };
+
+GLOBAL enum hole_detection_method hole_detection;
+
+/* The first entry in names.c:namelist specifies the member name to
+   start extracting from. Set by add_starting_file() upon seeing the
+   -K option.
+*/
 GLOBAL bool starting_file_option;
 
 /* Specified maximum byte length of each tape volume (multiple of 1024).  */
@@ -315,6 +340,9 @@ GLOBAL const char *volume_label_option;
 
 /* Other global variables.  */
 
+/* Force POSIX-compliance */
+GLOBAL bool posixly_correct;
+
 /* File descriptor for archive file.  */
 GLOBAL int archive;
 
@@ -398,10 +426,6 @@ GLOBAL bool show_transformed_names_option;
    timestamps from archives with an unusual member order. It is automatically
    set for incremental archives. */
 GLOBAL bool delay_directory_restore_option;
-
-/* Warn about implicit use of the wildcards in command line arguments.
-   (Default for tar prior to 1.15.91, but changed afterwards */
-GLOBAL bool warn_regex_usage;
 
 /* Declarations for each module.  */
 
@@ -453,7 +477,7 @@ void set_start_time (void);
 #define TF_READ    0
 #define TF_WRITE   1
 #define TF_DELETED 2
-int format_total_stats (FILE *fp, const char **formats, int eor, int eol);
+int format_total_stats (FILE *fp, char const *const *formats, int eor, int eol);
 void print_total_stats (void);
 
 void mv_begin_write (const char *file_name, off_t totsize, off_t sizeleft);
@@ -523,6 +547,8 @@ void extract_archive (void);
 void extract_finish (void);
 bool rename_directory (char *src, char *dst);
 
+void remove_delayed_set_stat (const char *fname);
+
 /* Module delete.c.  */
 
 void delete_archive_members (void);
@@ -605,7 +631,11 @@ void skip_member (void);
 
 #define min(a, b) ((a) < (b) ? (a) : (b))
 #define max(a, b) ((a) < (b) ? (b) : (a))
+
+char const *quote_n_colon (int n, char const *arg);
 void assign_string (char **dest, const char *src);
+void assign_string_n (char **string, const char *value, size_t n);
+#define ASSIGN_STRING_N(s,v) assign_string_n (s, v, sizeof (v))
 int unquote_string (char *str);
 char *zap_slashes (char *name);
 char *normalize_filename (int cdidx, const char *name);
@@ -618,8 +648,6 @@ typedef struct namebuf *namebuf_t;
 namebuf_t namebuf_create (const char *dir);
 void namebuf_free (namebuf_t buf);
 char *namebuf_name (namebuf_t buf, const char *name);
-void namebuf_add_dir (namebuf_t buf, const char *name);
-char *namebuf_finish (namebuf_t buf);
 
 const char *tar_dirname (void);
 
@@ -714,7 +742,21 @@ int set_file_atime (int fd, int parentfd, char const *file,
 
 /* Module names.c.  */
 
-extern size_t name_count;
+enum files_count
+  {
+    FILES_NONE,
+    FILES_ONE,
+    FILES_MANY
+  };
+extern enum files_count filename_args;
+
+/* Return true if there are file names in the list */
+static inline bool
+name_more_files (void)
+{
+  return filename_args != FILES_NONE;
+}
+
 extern struct name *gnu_list_name;
 
 void gid_to_gname (gid_t gid, char **gname);
@@ -723,14 +765,14 @@ void uid_to_uname (uid_t uid, char **uname);
 int uname_to_uid (char const *uname, uid_t *puid);
 
 void name_init (void);
-void name_add_name (const char *name, int matching_flags);
-void name_add_dir (const char *name);
-void name_add_file (const char *name, int term);
+bool name_more_files (void);
+void name_add_name (const char *name);
 void name_term (void);
 const char *name_next (int change_dirs);
 void name_gather (void);
 struct name *addname (char const *string, int change_dir,
 		      bool cmdline, struct name *parent);
+void add_starting_file (char const *file_name);
 void remname (struct name *name);
 bool name_match (const char *name);
 void names_notfound (void);
@@ -739,7 +781,7 @@ void collect_and_sort_names (void);
 struct name *name_scan (const char *name);
 struct name const *name_from_list (void);
 void blank_name_list (void);
-char *new_name (const char *dir_name, const char *name);
+char *make_file_name (const char *dir_name, const char *name);
 size_t stripped_prefix_len (char const *file_name, size_t num);
 bool all_names_found (struct tar_stat_info *st);
 
@@ -748,10 +790,12 @@ bool is_avoided_name (char const *name);
 
 bool contains_dot_dot (char const *name);
 
-#define ISFOUND(c) ((occurrence_option == 0) ? (c)->found_count : \
-                    (c)->found_count == occurrence_option)
-#define WASFOUND(c) ((occurrence_option == 0) ? (c)->found_count : \
-                     (c)->found_count >= occurrence_option)
+#define ISFOUND(c) (occurrence_option == 0			\
+		    ? (c)->found_count != 0			\
+		    : (c)->found_count == occurrence_option)
+#define WASFOUND(c) (occurrence_option == 0			\
+		     ? (c)->found_count != 0			\
+		     : (c)->found_count >= occurrence_option)
 
 /* Module tar.c.  */
 
@@ -769,7 +813,46 @@ const char *subcommand_string (enum subcommand c);
 void set_exit_status (int val);
 
 void request_stdin (const char *option);
-void more_options (int argc, char **argv);
+
+int decode_signal (const char *);
+
+/* Where an option comes from: */
+enum option_source
+  {
+    OPTS_ENVIRON,        /* Environment variable TAR_OPTIONS */
+    OPTS_COMMAND_LINE,   /* Command line */
+    OPTS_FILE            /* File supplied by --files-from */
+  };
+
+/* Option location */
+struct option_locus
+{
+  enum option_source source;  /* Option origin */
+  char const *name;           /* File or variable name */
+  size_t line;                /* Number of input line if source is OPTS_FILE */
+  struct option_locus *prev;  /* Previous occurrence of the option of same
+				 class */
+};
+
+struct tar_args        /* Variables used during option parsing */
+{
+  struct option_locus *loc;
+
+  struct textual_date *textual_date; /* Keeps the arguments to --newer-mtime
+					and/or --date option if they are
+					textual dates */
+  bool o_option;                   /* True if -o option was given */
+  bool pax_option;                 /* True if --pax-option was given */
+  bool compress_autodetect;        /* True if compression autodetection should
+				      be attempted when creating archives */
+  char const *backup_suffix_string;   /* --suffix option argument */
+  char const *version_control_string; /* --backup option argument */
+};
+
+#define TAR_ARGS_INITIALIZER(loc)              \
+  { loc, NULL, false, false, false, NULL, NULL }
+
+void more_options (int argc, char **argv, struct option_locus *loc);
 
 /* Module update.c.  */
 
@@ -789,6 +872,7 @@ void xheader_store (char const *keyword, struct tar_stat_info *st,
 void xheader_read (struct xheader *xhdr, union block *header, off_t size);
 void xheader_write (char type, char *name, time_t t, struct xheader *xhdr);
 void xheader_write_global (struct xheader *xhdr);
+void xheader_forbid_global (void);
 void xheader_finish (struct xheader *hdr);
 void xheader_destroy (struct xheader *hdr);
 char *xheader_xhdr_name (struct tar_stat_info *st);
@@ -893,6 +977,7 @@ void checkpoint_flush_actions (void);
 #define WARN_EXISTING_FILE       0x00100000
 #define WARN_XATTR_WRITE         0x00200000
 #define WARN_RECORD_SIZE         0x00400000
+#define WARN_FAILED_READ         0x00800000
 
 /* These warnings are enabled by default in verbose mode: */
 #define WARN_VERBOSE_WARNINGS    (WARN_RENAME_DIRECTORY|WARN_NEW_DIRECTORY|\
@@ -904,10 +989,12 @@ void set_warning_option (const char *arg);
 
 extern int warning_option;
 
+#define WARNING_ENABLED(opt) (warning_option & (opt))
+
 #define WARNOPT(opt,args)			\
   do						\
     {						\
-      if (warning_option & opt) WARN (args);	\
+      if (WARNING_ENABLED(opt)) WARN (args);	\
     }						\
   while (0)
 
@@ -926,9 +1013,15 @@ extern void (*fatal_exit_hook) (void);
 
 void excfile_add (const char *name, int flags);
 void info_attach_exclist (struct tar_stat_info *dir);
-void info_cleanup_exclist (struct tar_stat_info *dir);
 void info_free_exclist (struct tar_stat_info *dir);
 bool excluded_name (char const *name, struct tar_stat_info *st);
 void exclude_vcs_ignores (void);
 
+/* Module map.c */
+void owner_map_read (char const *name);
+int owner_map_translate (uid_t uid, uid_t *new_uid, char const **new_name);
+void group_map_read (char const *file);
+int group_map_translate (gid_t gid, gid_t *new_gid, char const **new_name);
+
+
 _GL_INLINE_HEADER_END

src/compare.c

@@ -1,7 +1,6 @@
 /* Diff files from a tar archive.
 
-   Copyright 1988, 1992-1994, 1996-1997, 1999-2001, 2003-2007,
-   2009-2010, 2012-2014 Free Software Foundation, Inc.
+   Copyright 1988-2021 Free Software Foundation, Inc.
 
    This file is part of GNU tar.
 
@@ -51,6 +50,8 @@ diff_init (void)
     read_directory_file ();
 }
 
+enum { QUOTE_ARG, QUOTE_NAME };
+
 /* Sigh about something that differs by writing a MESSAGE to stdlis,
    given MESSAGE is nonzero.  Also set the exit status if not already.  */
 void
@@ -60,7 +61,7 @@ report_difference (struct tar_stat_info *st, const char *fmt, ...)
     {
       va_list ap;
 
-      fprintf (stdlis, "%s: ", quotearg_colon (st->file_name));
+      fprintf (stdlis, "%s: ", quote_n_colon (QUOTE_NAME, st->file_name));
       va_start (ap, fmt);
       vfprintf (stdlis, fmt, ap);
       va_end (ap);
@@ -263,18 +264,20 @@ diff_link (void)
       && !sys_compare_links (&file_data, &link_data))
     report_difference (&current_stat_info,
 		       _("Not linked to %s"),
-		       quote (current_stat_info.link_name));
+		       quote_n_colon (QUOTE_ARG,
+				      current_stat_info.link_name));
 }
 
 #ifdef HAVE_READLINK
 static void
 diff_symlink (void)
 {
+  char buf[1024];
   size_t len = strlen (current_stat_info.link_name);
-  char *linkbuf = alloca (len + 1);
+  char *linkbuf = len < sizeof buf ? buf : xmalloc (len + 1);
 
-  int status = readlinkat (chdir_fd, current_stat_info.file_name,
-			   linkbuf, len + 1);
+  ssize_t status = readlinkat (chdir_fd, current_stat_info.file_name,
+			       linkbuf, len + 1);
 
   if (status < 0)
     {
@@ -285,8 +288,11 @@ diff_symlink (void)
       report_difference (&current_stat_info, NULL);
     }
   else if (status != len
-	   || strncmp (current_stat_info.link_name, linkbuf, len) != 0)
+	   || memcmp (current_stat_info.link_name, linkbuf, len) != 0)
     report_difference (&current_stat_info, _("Symlink differs"));
+
+  if (linkbuf != buf)
+    free (linkbuf);
 }
 #endif
 
@@ -473,8 +479,7 @@ diff_archive (void)
       ERROR ((0, 0, _("%s: Unknown file type '%c', diffed as normal file"),
 	      quotearg_colon (current_stat_info.file_name),
 	      current_header->header.typeflag));
-      /* Fall through.  */
-
+      FALLTHROUGH;
     case AREGTYPE:
     case REGTYPE:
     case GNUTYPE_SPARSE:

src/create.c

@@ -1,7 +1,6 @@
 /* Create a tar archive.
 
-   Copyright 1985, 1992-1994, 1996-1997, 1999-2001, 2003-2007,
-   2009-2010, 2012-2014 Free Software Foundation, Inc.
+   Copyright 1985-2021 Free Software Foundation, Inc.
 
    This file is part of GNU tar.
 
@@ -22,6 +21,7 @@
 
 #include <system.h>
 
+#include <areadlink.h>
 #include <quotearg.h>
 
 #include "common.h"
@@ -201,8 +201,8 @@ to_base256 (int negative, uintmax_t value, char *where, size_t size)
 #define MODE_TO_CHARS(val, where) mode_to_chars (val, where, sizeof (where))
 #define UID_TO_CHARS(val, where) uid_to_chars (val, where, sizeof (where))
 
-#define UNAME_TO_CHARS(name,buf) string_to_chars (name, buf, sizeof(buf))
-#define GNAME_TO_CHARS(name,buf) string_to_chars (name, buf, sizeof(buf))
+#define UNAME_TO_CHARS(name, buf) string_to_chars (name, buf, sizeof (buf))
+#define GNAME_TO_CHARS(name, buf) string_to_chars (name, buf, sizeof (buf))
 
 static bool
 to_chars (int negative, uintmax_t value, size_t valsize,
@@ -517,8 +517,8 @@ start_private_header (const char *name, size_t size, time_t t)
   MODE_TO_CHARS (S_IFREG|S_IRUSR|S_IWUSR|S_IRGRP|S_IROTH, header->header.mode);
   UID_TO_CHARS (0, header->header.uid);
   GID_TO_CHARS (0, header->header.gid);
-  strncpy (header->header.magic, TMAGIC, TMAGLEN);
-  strncpy (header->header.version, TVERSION, TVERSLEN);
+  memcpy (header->header.magic, TMAGIC, TMAGLEN);
+  memcpy (header->header.version, TVERSION, TVERSLEN);
   return header;
 }
 
@@ -541,15 +541,19 @@ write_gnu_long_link (struct tar_stat_info *st, const char *p, char type)
   size_t size = strlen (p) + 1;
   size_t bufsize;
   union block *header;
-  char *tmpname;
 
   header = start_private_header ("././@LongLink", size, 0);
-  uid_to_uname (0, &tmpname);
-  UNAME_TO_CHARS (tmpname, header->header.uname);
-  free (tmpname);
-  gid_to_gname (0, &tmpname);
-  GNAME_TO_CHARS (tmpname, header->header.gname);
-  free (tmpname);
+  if (! numeric_owner_option)
+    {
+      static char *uname, *gname;
+      if (!uname)
+	{
+	  uid_to_uname (0, &uname);
+	  gid_to_gname (0, &gname);
+	}
+      UNAME_TO_CHARS (uname, header->header.uname);
+      GNAME_TO_CHARS (gname, header->header.gname);
+    }
 
   strcpy (header->buffer + offsetof (struct posix_header, magic),
 	  OLDGNU_MAGIC);
@@ -706,7 +710,7 @@ write_extended (bool global, struct tar_stat_info *st, union block *old_header)
     {
       type = XHDTYPE;
       p = xheader_xhdr_name (st);
-      t = st->stat.st_mtime;
+      t = set_mtime_option ? mtime_option.tv_sec : st->stat.st_mtime;
     }
   xheader_write (type, p, t, &st->xhdr);
   free (p);
@@ -740,17 +744,17 @@ union block *
 start_header (struct tar_stat_info *st)
 {
   union block *header;
+  char const *uname = NULL;
+  char const *gname = NULL;
 
   header = write_header_name (st);
   if (!header)
     return NULL;
 
   /* Override some stat fields, if requested to do so.  */
+  owner_map_translate (st->stat.st_uid, &st->stat.st_uid, &uname);
+  group_map_translate (st->stat.st_gid, &st->stat.st_gid, &gname);
 
-  if (owner_option != (uid_t) -1)
-    st->stat.st_uid = owner_option;
-  if (group_option != (gid_t) -1)
-    st->stat.st_gid = group_option;
   if (mode_option)
     st->stat.st_mode =
       ((st->stat.st_mode & ~MODE_ALL)
@@ -822,7 +826,24 @@ start_header (struct tar_stat_info *st)
   }
 
   {
-    struct timespec mtime = set_mtime_option ? mtime_option : st->mtime;
+    struct timespec mtime;
+
+    switch (set_mtime_option)
+      {
+      case USE_FILE_MTIME:
+	mtime = st->mtime;
+	break;
+
+      case FORCE_MTIME:
+	mtime = mtime_option;
+	break;
+
+      case CLAMP_MTIME:
+	mtime = timespec_cmp (st->mtime, mtime_option) > 0
+	           ? mtime_option : st->mtime;
+	break;
+      }
+
     if (archive_format == POSIX_FORMAT)
       {
 	if (MAX_OCTAL_VAL (header->header.mtime) < mtime.tv_sec
@@ -895,8 +916,8 @@ start_header (struct tar_stat_info *st)
 
     case POSIX_FORMAT:
     case USTAR_FORMAT:
-      strncpy (header->header.magic, TMAGIC, TMAGLEN);
-      strncpy (header->header.version, TVERSION, TVERSLEN);
+      memcpy (header->header.magic, TMAGIC, TMAGLEN);
+      memcpy (header->header.version, TVERSION, TVERSLEN);
       break;
 
     default:
@@ -909,13 +930,13 @@ start_header (struct tar_stat_info *st)
     }
   else
     {
-      if (owner_name_option)
-	st->uname = xstrdup (owner_name_option);
+      if (uname)
+	st->uname = xstrdup (uname);
       else
 	uid_to_uname (st->stat.st_uid, &st->uname);
 
-      if (group_name_option)
-	st->gname = xstrdup (group_name_option);
+      if (gname)
+	st->gname = xstrdup (gname);
       else
 	gid_to_gname (st->stat.st_gid, &st->gname);
 
@@ -1114,7 +1135,7 @@ dump_dir0 (struct tar_stat_info *st, char const *directory)
     return;
 
   info_attach_exclist (st);
-  
+
   if (incremental_option && archive_format != POSIX_FORMAT)
     blk->header.typeflag = GNUTYPE_DUMPDIR;
   else /* if (standard_option) */
@@ -1198,7 +1219,7 @@ dump_dir0 (struct tar_stat_info *st, char const *directory)
 	    char const *entry;
 	    size_t entry_len;
 	    size_t name_len;
-	    
+
 	    name_buf = xstrdup (st->orig_file_name);
 	    name_size = name_len = strlen (name_buf);
 
@@ -1327,15 +1348,15 @@ create_archive (void)
 {
   struct name const *p;
 
-  trivial_link_count = name_count <= 1 && ! dereference_option;
+  trivial_link_count = filename_args != FILES_MANY && ! dereference_option;
 
   open_archive (ACCESS_WRITE);
   buffer_write_global_xheader ();
 
   if (incremental_option)
     {
-      size_t buffer_size = 1000;
-      char *buffer = xmalloc (buffer_size);
+      size_t buffer_size = 0;
+      char *buffer = NULL;
       const char *q;
 
       collect_and_sort_names ();
@@ -1350,12 +1371,8 @@ create_archive (void)
 	  {
 	    struct tar_stat_info st;
 	    size_t plen = strlen (p->name);
-	    if (buffer_size <= plen)
-	      {
-		while ((buffer_size *= 2) <= plen)
-		  continue;
-		buffer = xrealloc (buffer, buffer_size);
-	      }
+	    while (buffer_size <= plen)
+	      buffer = x2realloc (buffer, &buffer_size);
 	    memcpy (buffer, p->name, plen);
 	    if (! ISSLASH (buffer[plen - 1]))
 	      buffer[plen++] = DIRECTORY_SEPARATOR;
@@ -1373,23 +1390,21 @@ create_archive (void)
 					   open_searchdir_flags);
 			  if (fd < 0)
 			    {
-			      open_diag (p->name);
+			      file_removed_diag (p->name, !p->parent,
+						 open_diag);
 			      break;
 			    }
 			  st.fd = fd;
 			  if (fstat (fd, &st.stat) != 0)
 			    {
-			      stat_diag (p->name);
+			      file_removed_diag (p->name, !p->parent,
+						 stat_diag);
 			      break;
 			    }
 			  st.orig_file_name = xstrdup (p->name);
 			}
-		      if (buffer_size < plen + qlen)
-			{
-			  while ((buffer_size *=2 ) < plen + qlen)
-			    continue;
-			  buffer = xrealloc (buffer, buffer_size);
- 			}
+		      while (buffer_size < plen + qlen)
+			buffer = x2realloc (buffer, &buffer_size);
 		      strcpy (buffer + plen, q + 1);
 		      dump_file (&st, q + 1, buffer);
 		    }
@@ -1472,8 +1487,8 @@ dump_hard_link (struct tar_stat_info *st)
 	  /* We found a link.  */
 	  char const *link_name = safer_name_suffix (duplicate->name, true,
 	                                             absolute_names_option);
-
-	  duplicate->nlink--;
+	  if (duplicate->nlink)
+	    duplicate->nlink--;
 
 	  block_ordinal = current_block_ordinal ();
 	  assign_string (&st->link_name, link_name);
@@ -1510,7 +1525,8 @@ file_count_links (struct tar_stat_info *st)
       char *linkname = NULL;
       struct link *lp;
 
-      assign_string (&linkname, st->orig_file_name);
+      assign_string (&linkname, safer_name_suffix (st->orig_file_name, true,
+						   absolute_names_option));
       transform_name (&linkname, XFORM_LINK);
 
       lp = xmalloc (offsetof (struct link, name)
@@ -1837,22 +1853,17 @@ dump_file0 (struct tar_stat_info *st, char const *name, char const *p)
 #ifdef HAVE_READLINK
   else if (S_ISLNK (st->stat.st_mode))
     {
-      char *buffer;
-      int size;
-      size_t linklen = st->stat.st_size;
-      if (linklen != st->stat.st_size || linklen + 1 == 0)
-	xalloc_die ();
-      buffer = (char *) alloca (linklen + 1);
-      size = readlinkat (parentfd, name, buffer, linklen + 1);
-      if (size < 0)
+      st->link_name = areadlinkat_with_size (parentfd, name, st->stat.st_size);
+      if (!st->link_name)
 	{
+	  if (errno == ENOMEM)
+	    xalloc_die ();
 	  file_removed_diag (p, top_level, readlink_diag);
 	  return;
 	}
-      buffer[size] = '\0';
-      assign_string (&st->link_name, buffer);
       transform_name (&st->link_name, XFORM_SYMLINK);
-      if (NAME_FIELD_SIZE - (archive_format == OLDGNU_FORMAT) < size)
+      if (NAME_FIELD_SIZE - (archive_format == OLDGNU_FORMAT)
+	  < strlen (st->link_name))
 	write_long_link (st);
 
       xattrs_selinux_get (parentfd, name, st, 0);

src/delete.c

@@ -1,7 +1,6 @@
 /* Delete entries from a tar archive.
 
-   Copyright 1988, 1992, 1994, 1996-1997, 2000-2001, 2003-2006, 2010,
-   2013-2014 Free Software Foundation, Inc.
+   Copyright 1988-2021 Free Software Foundation, Inc.
 
    This file is part of GNU tar.
 
@@ -148,6 +147,23 @@ write_recent_bytes (char *data, size_t bytes)
     write_record (1);
 }
 
+static inline void
+flush_file (void)
+{
+  off_t blocks_to_skip;
+
+  set_next_block_after (current_header);
+  blocks_to_skip = (current_stat_info.stat.st_size
+			  + BLOCKSIZE - 1) / BLOCKSIZE;
+
+  while (record_end - current_block <= blocks_to_skip)
+    {
+      blocks_to_skip -= (record_end - current_block);
+      flush_archive ();
+    }
+  current_block += blocks_to_skip;
+}
+
 void
 delete_archive_members (void)
 {
@@ -156,7 +172,6 @@ delete_archive_members (void)
 
   /* FIXME: Should clean the routine before cleaning these variables :-( */
   struct name *name;
-  off_t blocks_to_skip = 0;
   off_t blocks_to_keep = 0;
   int kept_blocks_in_record;
 
@@ -164,11 +179,12 @@ delete_archive_members (void)
   open_archive (ACCESS_UPDATE);
   acting_as_filter = strcmp (archive_name_array[0], "-") == 0;
 
+  /* Skip to the first member that matches the name list. */
   do
     {
       enum read_header status = read_header (&current_header,
-                                             &current_stat_info,
-                                             read_header_x_raw);
+					     &current_stat_info,
+					     read_header_x_raw);
 
       switch (status)
 	{
@@ -182,13 +198,12 @@ delete_archive_members (void)
 	      break;
 	    }
 	  name->found_count++;
-	  if (!ISFOUND(name))
+	  if (!ISFOUND (name))
 	    {
 	      skip_member ();
 	      break;
 	    }
-
-	  /* Fall through.  */
+	  FALLTHROUGH;
 	case HEADER_SUCCESS_EXTENDED:
 	  logical_status = status;
 	  break;
@@ -199,7 +214,7 @@ delete_archive_members (void)
 	      set_next_block_after (current_header);
 	      break;
 	    }
-	  /* Fall through.  */
+	  FALLTHROUGH;
 	case HEADER_END_OF_FILE:
 	  logical_status = HEADER_END_OF_FILE;
 	  break;
@@ -210,14 +225,12 @@ delete_archive_members (void)
 	    {
 	    case HEADER_STILL_UNREAD:
 	      WARN ((0, 0, _("This does not look like a tar archive")));
-	      /* Fall through.  */
-
+	      FALLTHROUGH;
 	    case HEADER_SUCCESS:
 	    case HEADER_SUCCESS_EXTENDED:
 	    case HEADER_ZERO_BLOCK:
 	      ERROR ((0, 0, _("Skipping to next header")));
-	      /* Fall through.  */
-
+	      FALLTHROUGH;
 	    case HEADER_FAILURE:
 	      break;
 
@@ -247,15 +260,12 @@ delete_archive_members (void)
 
       if (logical_status == HEADER_SUCCESS)
 	{
-	  /* FIXME: Pheew!  This is crufty code!  */
 	  logical_status = HEADER_STILL_UNREAD;
-	  goto flush_file;
+	  flush_file ();
 	}
 
-      /* FIXME: Solaris 2.4 Sun cc (the ANSI one, not the old K&R) says:
-	 "delete.c", line 223: warning: loop not entered at top
-	 Reported by Bruno Haible.  */
-      while (1)
+      /* Skip matching members and move the rest up the archive. */
+      while (logical_status != HEADER_END_OF_FILE)
 	{
 	  enum read_header status;
 
@@ -263,105 +273,108 @@ delete_archive_members (void)
 
 	  if (current_block == record_end)
 	    flush_archive ();
+
 	  status = read_header (&current_header, &current_stat_info,
-	                        read_header_auto);
+				read_header_auto);
 
-	  xheader_decode (&current_stat_info);
-
-	  if (status == HEADER_ZERO_BLOCK && ignore_zeros_option)
+	  switch (status)
 	    {
-	      set_next_block_after (current_header);
-	      continue;
-	    }
-	  if (status == HEADER_END_OF_FILE || status == HEADER_ZERO_BLOCK)
-	    {
-	      logical_status = HEADER_END_OF_FILE;
-	      break;
-	    }
+	    case HEADER_STILL_UNREAD:
+	    case HEADER_SUCCESS_EXTENDED:
+	      abort ();
 
-	  if (status == HEADER_FAILURE)
-	    {
-	      ERROR ((0, 0, _("Deleting non-header from archive")));
-	      set_next_block_after (current_header);
-	      continue;
-	    }
+	    case HEADER_SUCCESS:
+	      /* Found another header.  */
+	      xheader_decode (&current_stat_info);
 
-	  /* Found another header.  */
-
-	  if ((name = name_scan (current_stat_info.file_name)) != NULL)
-	    {
-	      name->found_count++;
-	      if (ISFOUND(name))
+	      if ((name = name_scan (current_stat_info.file_name)) != NULL)
 		{
-		flush_file:
-		  set_next_block_after (current_header);
-		  blocks_to_skip = (current_stat_info.stat.st_size
-				    + BLOCKSIZE - 1) / BLOCKSIZE;
-
-		  while (record_end - current_block <= blocks_to_skip)
+		  name->found_count++;
+		  if (ISFOUND (name))
 		    {
-		      blocks_to_skip -= (record_end - current_block);
-		      flush_archive ();
+		      flush_file ();
+		      break;
 		    }
-		  current_block += blocks_to_skip;
-		  blocks_to_skip = 0;
-		  continue;
 		}
-	    }
-	  /* Copy header.  */
+	      /* Copy header.  */
 
-	  if (current_stat_info.xhdr.size)
-	    {
-	      write_recent_bytes (current_stat_info.xhdr.buffer,
-				  current_stat_info.xhdr.size);
-	    }
-	  else
-	    {
-	      write_recent_blocks (recent_long_name, recent_long_name_blocks);
-	      write_recent_blocks (recent_long_link, recent_long_link_blocks);
-	    }
-	  new_record[new_blocks] = *current_header;
-	  new_blocks++;
-	  blocks_to_keep
-	    = (current_stat_info.stat.st_size + BLOCKSIZE - 1) / BLOCKSIZE;
-	  set_next_block_after (current_header);
-	  if (new_blocks == blocking_factor)
-	    write_record (1);
-
-	  /* Copy data.  */
-
-	  kept_blocks_in_record = record_end - current_block;
-	  if (kept_blocks_in_record > blocks_to_keep)
-	    kept_blocks_in_record = blocks_to_keep;
-
-	  while (blocks_to_keep)
-	    {
-	      int count;
-
-	      if (current_block == record_end)
+	      if (current_stat_info.xhdr.size)
 		{
-		  flush_read ();
-		  current_block = record_start;
-		  kept_blocks_in_record = blocking_factor;
-		  if (kept_blocks_in_record > blocks_to_keep)
-		    kept_blocks_in_record = blocks_to_keep;
+		  write_recent_bytes (current_stat_info.xhdr.buffer,
+				      current_stat_info.xhdr.size);
 		}
-	      count = kept_blocks_in_record;
-	      if (blocking_factor - new_blocks < count)
-		count = blocking_factor - new_blocks;
-
-	      if (! count)
-		abort ();
-
-	      memcpy (new_record + new_blocks, current_block, count * BLOCKSIZE);
-	      new_blocks += count;
-	      current_block += count;
-	      blocks_to_keep -= count;
-	      kept_blocks_in_record -= count;
-
+	      else
+		{
+		  write_recent_blocks (recent_long_name,
+				       recent_long_name_blocks);
+		  write_recent_blocks (recent_long_link,
+				       recent_long_link_blocks);
+		}
+	      new_record[new_blocks] = *current_header;
+	      new_blocks++;
+	      blocks_to_keep
+		= (current_stat_info.stat.st_size + BLOCKSIZE - 1) / BLOCKSIZE;
+	      set_next_block_after (current_header);
 	      if (new_blocks == blocking_factor)
 		write_record (1);
+
+	      /* Copy data.  */
+
+	      kept_blocks_in_record = record_end - current_block;
+	      if (kept_blocks_in_record > blocks_to_keep)
+		kept_blocks_in_record = blocks_to_keep;
+
+	      while (blocks_to_keep)
+		{
+		  int count;
+
+		  if (current_block == record_end)
+		    {
+		      flush_read ();
+		      current_block = record_start;
+		      kept_blocks_in_record = blocking_factor;
+		      if (kept_blocks_in_record > blocks_to_keep)
+			kept_blocks_in_record = blocks_to_keep;
+		    }
+		  count = kept_blocks_in_record;
+		  if (blocking_factor - new_blocks < count)
+		    count = blocking_factor - new_blocks;
+
+		  if (! count)
+		    abort ();
+
+		  memcpy (new_record + new_blocks, current_block,
+			  count * BLOCKSIZE);
+		  new_blocks += count;
+		  current_block += count;
+		  blocks_to_keep -= count;
+		  kept_blocks_in_record -= count;
+
+		  if (new_blocks == blocking_factor)
+		    write_record (1);
+		}
+	      break;
+
+	    case HEADER_ZERO_BLOCK:
+	      if (ignore_zeros_option)
+		set_next_block_after (current_header);
+	      else
+		logical_status = HEADER_END_OF_FILE;
+	      break;
+
+	    case HEADER_END_OF_FILE:
+	      logical_status = HEADER_END_OF_FILE;
+	      break;
+
+	    case HEADER_FAILURE:
+	      ERROR ((0, 0, _("Deleting non-header from archive")));
+	      set_next_block_after (current_header);
+	      break;
+
+	    default:
+	      abort ();
 	    }
+	  tar_stat_destroy (&current_stat_info);
 	}
 
       if (logical_status == HEADER_END_OF_FILE)

src/exclist.c

@@ -1,6 +1,6 @@
 /* Per-directory exclusion files for tar.
 
-   Copyright 2014 Free Software Foundation, Inc.
+   Copyright 2014-2021 Free Software Foundation, Inc.
 
    This file is part of GNU tar.
 
@@ -43,7 +43,7 @@ struct excfile
   char name[1];
 };
 
-struct excfile *excfile_head, *excfile_tail;
+static struct excfile *excfile_head, *excfile_tail;
 
 void
 excfile_add (const char *name, int flags)
@@ -72,7 +72,7 @@ info_attach_exclist (struct tar_stat_info *dir)
   struct excfile *file;
   struct exclist *head = NULL, *tail = NULL, *ent;
   struct vcs_ignore_file *vcsfile;
-    
+
   if (dir->exclude_list)
     return;
   for (file = excfile_head; file; file = file->next)
@@ -102,16 +102,17 @@ info_attach_exclist (struct tar_stat_info *dir)
 
 	  if (vcsfile->initfn)
 	    vcsfile->data = vcsfile->initfn (vcsfile->data);
-	  
+
 	  if (add_exclude_fp (vcsfile->addfn, ex, fp,
-			      EXCLUDE_WILDCARDS|EXCLUDE_ANCHORED, '\n',
+			      FNM_FILE_NAME|EXCLUDE_WILDCARDS|EXCLUDE_ANCHORED,
+			      '\n',
 			      vcsfile->data))
 	    {
 	      int e = errno;
 	      FATAL_ERROR ((0, e, "%s", quotearg_colon (file->name)));
 	    }
 	  fclose (fp);
-	  
+
 	  ent = xmalloc (sizeof (*ent));
 	  ent->excluded = ex;
 	  ent->flags = file->flags == EXCL_DEFAULT
@@ -129,34 +130,6 @@ info_attach_exclist (struct tar_stat_info *dir)
   dir->exclude_list = head;
 }
 
-void
-info_cleanup_exclist (struct tar_stat_info *dir)
-{
-  struct exclist *ep = dir->exclude_list;
-
-  while (ep)
-    {
-      struct exclist *next = ep->next;
-      
-      if (ep->flags & EXCL_NON_RECURSIVE)
-	{
-	  
-	  /* Remove the entry */
-	  if (ep->prev)
-	    ep->prev->next = ep->next;
-	  else
-	    dir->exclude_list = ep->next;
-
-	  if (ep->next)
-	    ep->next->prev = ep->prev;
-
-	  free_exclude (ep->excluded);
-	  free (ep);
-	}
-      ep = next;
-    }
-}
-
 void
 info_free_exclist (struct tar_stat_info *dir)
 {
@@ -172,7 +145,7 @@ info_free_exclist (struct tar_stat_info *dir)
 
   dir->exclude_list = NULL;
 }
-  
+
 
 /* Return nonzero if file NAME is excluded.  */
 bool
@@ -183,7 +156,7 @@ excluded_name (char const *name, struct tar_stat_info *st)
   char *bname = NULL;
   bool result;
   int nr = 0;
-  
+
   name += FILE_SYSTEM_PREFIX_LEN (name);
 
   /* Try global exclusion list first */
@@ -192,7 +165,7 @@ excluded_name (char const *name, struct tar_stat_info *st)
 
   if (!st)
     return false;
-  
+
   for (result = false; st && !result; st = st->parent, nr = EXCL_NON_RECURSIVE)
     {
       for (ep = st->exclude_list; ep; ep = ep->next)
@@ -201,7 +174,7 @@ excluded_name (char const *name, struct tar_stat_info *st)
 	    continue;
 	  if ((result = excluded_file_name (ep->excluded, name)))
 	    break;
-	  
+
 	  if (!rname)
 	    {
 	      rname = name;
@@ -229,8 +202,8 @@ cvs_addfn (struct exclude *ex, char const *pattern, int options, void *data)
 {
   struct wordsplit ws;
   size_t i;
-    
-  if (wordsplit (pattern, &ws, 
+
+  if (wordsplit (pattern, &ws,
 		 WRDSF_NOVAR | WRDSF_NOCMD | WRDSF_SQUEEZE_DELIMS))
     return;
   for (i = 0; i < ws.ws_wordc; i++)
@@ -278,22 +251,18 @@ bzr_addfn (struct exclude *ex, char const *pattern, int options, void *data)
 static void *
 hg_initfn (void *data)
 {
-  int *hgopt;
   static int hg_options;
-  
-  if (!data)
-    hgopt = &hg_options;
-
+  int *hgopt = data ? data : &hg_options;
   *hgopt = EXCLUDE_REGEX;
   return hgopt;
 }
-  
+
 static void
 hg_addfn (struct exclude *ex, char const *pattern, int options, void *data)
 {
   int *hgopt = data;
   size_t len;
-  
+
   while (isspace (*pattern))
     ++pattern;
   if (*pattern == 0 || *pattern == '#')
@@ -318,27 +287,27 @@ hg_addfn (struct exclude *ex, char const *pattern, int options, void *data)
 
       --len;
       p = xmalloc (len+1);
-      memcpy (p, pattern, len); 
+      memcpy (p, pattern, len);
       p[len] = 0;
       pattern = p;
       exclude_add_pattern_buffer (ex, p);
       options |= FNM_LEADING_DIR|EXCLUDE_ALLOC;
     }
-  
+
   add_exclude (ex, pattern,
 	       ((*hgopt == EXCLUDE_REGEX)
 		? (options & ~EXCLUDE_WILDCARDS)
 		: (options & ~EXCLUDE_REGEX)) | *hgopt);
 }
 
-struct vcs_ignore_file vcs_ignore_files[] = {
+static struct vcs_ignore_file vcs_ignore_files[] = {
   { ".cvsignore", EXCL_NON_RECURSIVE, cvs_addfn, NULL, NULL },
   { ".gitignore", 0, git_addfn, NULL, NULL },
   { ".bzrignore", 0, bzr_addfn, NULL, NULL },
-  { ".hgignore",  0, hg_addfn, hg_initfn , NULL },
+  { ".hgignore",  0, hg_addfn, hg_initfn, NULL },
   { NULL, 0, git_addfn, NULL, NULL }
 };
-  
+
 static struct vcs_ignore_file *
 get_vcs_ignore_file (const char *name)
 {

src/exit.c

@@ -1,6 +1,6 @@
 /* Exit from GNU tar.
 
-   Copyright 2009, 2013-2014 Free Software Foundation, Inc.
+   Copyright 2009-2021 Free Software Foundation, Inc.
 
    This file is part of GNU tar.
 

src/extract.c

@@ -1,7 +1,6 @@
 /* Extract files from a tar archive.
 
-   Copyright 1988, 1992-1994, 1996-2001, 2003-2007, 2010, 2012-2014 Free
-   Software Foundation, Inc.
+   Copyright 1988-2021 Free Software Foundation, Inc.
 
    This file is part of GNU tar.
 
@@ -109,7 +108,7 @@ struct delayed_set_stat
     struct xattr_array *xattr_map;
     /* Length and contents of name.  */
     size_t file_name_len;
-    char file_name[1];
+    char *file_name;
   };
 
 static struct delayed_set_stat *delayed_set_stat_head;
@@ -195,7 +194,7 @@ extr_init (void)
 
 /* Use fchmod if possible, fchmodat otherwise.  */
 static int
-fd_chmod (int fd, char const *file, mode_t mode, int atflag)
+fd_i_chmod (int fd, char const *file, mode_t mode, int atflag)
 {
   if (0 <= fd)
     {
@@ -206,6 +205,42 @@ fd_chmod (int fd, char const *file, mode_t mode, int atflag)
   return fchmodat (chdir_fd, file, mode, atflag);
 }
 
+/* A version of fd_i_chmod which gracefully handles several common error
+   conditions.  Additional argument TYPEFLAG is the type of file in tar
+   notation.
+ */
+static int
+fd_chmod(int fd, char const *file_name, int mode, int atflag, int typeflag)
+{
+  int chmod_errno = fd_i_chmod (fd, file_name, mode, atflag) == 0 ? 0 : errno;
+
+  /* On Solaris, chmod may fail if we don't have PRIV_ALL, because
+     setuid-root files would otherwise be a backdoor.  See
+     http://opensolaris.org/jive/thread.jspa?threadID=95826
+     (2009-09-03).  */
+  if (chmod_errno == EPERM && (mode & S_ISUID)
+      && priv_set_restore_linkdir () == 0)
+    {
+      chmod_errno = fd_i_chmod (fd, file_name, mode, atflag) == 0 ? 0 : errno;
+      priv_set_remove_linkdir ();
+    }
+
+  /* Linux fchmodat does not support AT_SYMLINK_NOFOLLOW, and
+     returns ENOTSUP even when operating on non-symlinks, try
+     again with the flag disabled if it does not appear to be
+     supported and if the file is not a symlink.  This
+     introduces a race, alas.  */
+  if (atflag && typeflag != SYMTYPE && ! implemented (chmod_errno))
+    chmod_errno = fd_i_chmod (fd, file_name, mode, 0) == 0 ? 0 : errno;
+
+  if (chmod_errno && (typeflag != SYMTYPE || implemented (chmod_errno)))
+    {
+      errno = chmod_errno;
+      return -1;
+    }
+  return 0;
+}
+
 /* Use fchown if possible, fchownat otherwise.  */
 static int
 fd_chown (int fd, char const *file, uid_t uid, gid_t gid, int atflag)
@@ -260,35 +295,8 @@ set_mode (char const *file_name,
 
       if (current_mode != mode)
 	{
-	  int chmod_errno =
-	    fd_chmod (fd, file_name, mode, atflag) == 0 ? 0 : errno;
-
-	  /* On Solaris, chmod may fail if we don't have PRIV_ALL, because
-	     setuid-root files would otherwise be a backdoor.  See
-	     http://opensolaris.org/jive/thread.jspa?threadID=95826
-	     (2009-09-03).  */
-	  if (chmod_errno == EPERM && (mode & S_ISUID)
-	      && priv_set_restore_linkdir () == 0)
-	    {
-	      chmod_errno =
-		fd_chmod (fd, file_name, mode, atflag) == 0 ? 0 : errno;
-	      priv_set_remove_linkdir ();
-	    }
-
-	  /* Linux fchmodat does not support AT_SYMLINK_NOFOLLOW, and
-	     returns ENOTSUP even when operating on non-symlinks, try
-	     again with the flag disabled if it does not appear to be
-	     supported and if the file is not a symlink.  This
-	     introduces a race, alas.  */
-	  if (atflag && typeflag != SYMTYPE && ! implemented (chmod_errno))
-	    chmod_errno = fd_chmod (fd, file_name, mode, 0) == 0 ? 0 : errno;
-
-	  if (chmod_errno
-	      && (typeflag != SYMTYPE || implemented (chmod_errno)))
-	    {
-	      errno = chmod_errno;
-	      chmod_error_details (file_name, mode);
-	    }
+	  if (fd_chmod (fd, file_name, mode, atflag, typeflag))
+	    chmod_error_details (file_name, mode);
 	}
     }
 }
@@ -393,6 +401,24 @@ set_stat (char const *file_name,
   xattrs_selinux_set (st, file_name, typeflag);
 }
 
+/* Find the direct ancestor of FILE_NAME in the delayed_set_stat list.
+ */
+static struct delayed_set_stat *
+find_direct_ancestor (char const *file_name)
+{
+  struct delayed_set_stat *h = delayed_set_stat_head;
+  while (h)
+    {
+      if (! h->after_links
+	  && strncmp (file_name, h->file_name, h->file_name_len) == 0
+	  && ISSLASH (file_name[h->file_name_len])
+	  && (last_component (file_name) == file_name + h->file_name_len + 1))
+	break;
+      h = h->next;
+    }
+  return h;
+}
+
 /* For each entry H in the leading prefix of entries in HEAD that do
    not have after_links marked, mark H and fill in its dev and ino
    members.  Assume HEAD && ! HEAD->after_links.  */
@@ -441,26 +467,56 @@ delay_set_stat (char const *file_name, struct tar_stat_info const *st,
 		mode_t mode, int atflag)
 {
   size_t file_name_len = strlen (file_name);
-  struct delayed_set_stat *data =
-    xmalloc (offsetof (struct delayed_set_stat, file_name)
-	     + file_name_len + 1);
-  data->next = delayed_set_stat_head;
+  struct delayed_set_stat *data;
+
+  for (data = delayed_set_stat_head; data; data = data->next)
+    if (strcmp (data->file_name, file_name) == 0)
+      break;
+
+  if (data)
+    {
+      if (data->interdir)
+	{
+	  struct stat real_st;
+	  if (fstatat (chdir_fd, data->file_name,
+		       &real_st, data->atflag) != 0)
+	    {
+	      stat_error (data->file_name);
+	    }
+	  else
+	    {
+	      data->dev = real_st.st_dev;
+	      data->ino = real_st.st_ino;
+	    }
+	}
+    }
+  else
+    {
+      data = xmalloc (sizeof (*data));
+      data->next = delayed_set_stat_head;
+      delayed_set_stat_head = data;
+      data->file_name_len = file_name_len;
+      data->file_name = xstrdup (file_name);
+      data->after_links = false;
+      if (st)
+	{
+	  data->dev = st->stat.st_dev;
+	  data->ino = st->stat.st_ino;
+	}
+    }
+
   data->mode = mode;
   if (st)
     {
-      data->dev = st->stat.st_dev;
-      data->ino = st->stat.st_ino;
       data->uid = st->stat.st_uid;
       data->gid = st->stat.st_gid;
       data->atime = st->atime;
       data->mtime = st->mtime;
     }
-  data->file_name_len = file_name_len;
   data->current_mode = current_mode;
   data->current_mode_mask = current_mode_mask;
   data->interdir = ! st;
   data->atflag = atflag;
-  data->after_links = 0;
   data->change_dir = chdir_current;
   data->cntx_name = NULL;
   if (st)
@@ -492,8 +548,6 @@ delay_set_stat (char const *file_name, struct tar_stat_info const *st,
       data->xattr_map = NULL;
       data->xattr_map_size = 0;
     }
-  strcpy (data->file_name, file_name);
-  delayed_set_stat_head = data;
   if (must_be_dot_or_slash (file_name))
     mark_after_links (data);
 }
@@ -507,7 +561,7 @@ repair_delayed_set_stat (char const *dir,
 			 struct stat const *dir_stat_info)
 {
   struct delayed_set_stat *data;
-  for (data = delayed_set_stat_head;  data;  data = data->next)
+  for (data = delayed_set_stat_head; data; data = data->next)
     {
       struct stat st;
       if (fstatat (chdir_fd, data->file_name, &st, data->atflag) != 0)
@@ -537,6 +591,56 @@ repair_delayed_set_stat (char const *dir,
 	  quotearg_colon (dir)));
 }
 
+static void
+free_delayed_set_stat (struct delayed_set_stat *data)
+{
+  free (data->file_name);
+  xheader_xattr_free (data->xattr_map, data->xattr_map_size);
+  free (data->cntx_name);
+  free (data->acls_a_ptr);
+  free (data->acls_d_ptr);
+  free (data);
+}
+
+void
+remove_delayed_set_stat (const char *fname)
+{
+  struct delayed_set_stat *data, *next, *prev = NULL;
+  for (data = delayed_set_stat_head; data; data = next)
+    {
+      next = data->next;
+      if (chdir_current == data->change_dir
+	  && strcmp (data->file_name, fname) == 0)
+	{
+	  free_delayed_set_stat (data);
+	  if (prev)
+	    prev->next = next;
+	  else
+	    delayed_set_stat_head = next;
+	  return;
+	}
+      else
+	prev = data;
+    }
+}
+
+static void
+fixup_delayed_set_stat (char const *src, char const *dst)
+{
+  struct delayed_set_stat *data;
+  for (data = delayed_set_stat_head; data; data = data->next)
+    {
+      if (chdir_current == data->change_dir
+	  && strcmp (data->file_name, src) == 0)
+	{
+	  free (data->file_name);
+	  data->file_name = xstrdup (dst);
+	  data->file_name_len = strlen (dst);
+	  return;
+	}
+    }
+}
+
 /* After a file/link/directory creation has failed, see if
    it's because some required directory was not present, and if so,
    create all required directories.  Return zero if all the required
@@ -691,10 +795,9 @@ maybe_recoverable (char *file_name, bool regular, bool *interdir_made)
 	    break;
 	  stp = &st;
 	}
-
       /* The caller tried to open a symbolic link with O_NOFOLLOW.
 	 Fall through, treating it as an already-existing file.  */
-
+      FALLTHROUGH;
     case EEXIST:
       /* Remove an old file, if the options allow this.  */
 
@@ -711,8 +814,7 @@ maybe_recoverable (char *file_name, bool regular, bool *interdir_made)
 	case KEEP_NEWER_FILES:
 	  if (file_newer_p (file_name, stp, &current_stat_info))
 	    break;
-	  /* FALL THROUGH */
-
+	  FALLTHROUGH;
 	case DEFAULT_OLD_FILES:
 	case NO_OVERWRITE_DIR_OLD_FILES:
 	case OVERWRITE_OLD_FILES:
@@ -723,6 +825,7 @@ maybe_recoverable (char *file_name, bool regular, bool *interdir_made)
 	case UNLINK_FIRST_OLD_FILES:
 	  break;
 	}
+      FALLTHROUGH;
 
     case ENOENT:
       /* Attempt creating missing intermediate directories.  */
@@ -746,13 +849,13 @@ maybe_recoverable (char *file_name, bool regular, bool *interdir_made)
    in advance dramatically improves the following  performance of reading and
    writing a file).  If not restoring permissions, invert the INVERT_PERMISSIONS
    bits from the file's current permissions.  TYPEFLAG specifies the type of the
-   file.  FILE_CREATED indicates set_xattr has created the file */
+   file.  Returns non-zero when error occurs (while un-available xattrs is not
+   an error, rather no-op).  Non-zero FILE_CREATED indicates set_xattr has
+   created the file. */
 static int
 set_xattr (char const *file_name, struct tar_stat_info const *st,
            mode_t invert_permissions, char typeflag, int *file_created)
 {
-  int status = 0;
-
 #ifdef HAVE_XATTRS
   bool interdir_made = false;
 
@@ -760,17 +863,32 @@ set_xattr (char const *file_name, struct tar_stat_info const *st,
     {
       mode_t mode = current_stat_info.stat.st_mode & MODE_RWX & ~ current_umask;
 
-      do
-        status = mknodat (chdir_fd, file_name, mode ^ invert_permissions, 0);
-      while (status && maybe_recoverable ((char *)file_name, false,
-                                          &interdir_made));
+      for (;;)
+        {
+          if (!mknodat (chdir_fd, file_name, mode ^ invert_permissions, 0))
+            {
+              /* Successfully created file */
+              xattrs_xattrs_set (st, file_name, typeflag, 0);
+              *file_created = 1;
+              return 0;
+            }
 
-      xattrs_xattrs_set (st, file_name, typeflag, 0);
-      *file_created = 1;
+          switch (maybe_recoverable ((char *)file_name, false, &interdir_made))
+            {
+              case RECOVER_OK:
+                continue;
+              case RECOVER_NO:
+                skip_member ();
+                open_error (file_name);
+                return 1;
+              case RECOVER_SKIP:
+                return 0;
+            }
+        }
     }
 #endif
 
-  return(status);
+  return 0;
 }
 
 /* Fix the statuses of all directories whose statuses need fixing, and
@@ -846,11 +964,7 @@ apply_nonancestor_delayed_set_stat (char const *file_name, bool after_links)
 	}
 
       delayed_set_stat_head = data->next;
-      xheader_xattr_free (data->xattr_map, data->xattr_map_size);
-      free (data->cntx_name);
-      free (data->acls_a_ptr);
-      free (data->acls_d_ptr);
-      free (data);
+      free_delayed_set_stat (data);
     }
 }
 
@@ -861,7 +975,7 @@ is_directory_link (const char *file_name)
   struct stat st;
   int e = errno;
   int res;
-  
+
   res = (fstatat (chdir_fd, file_name, &st, AT_SYMLINK_NOFOLLOW) == 0 &&
 	 S_ISLNK (st.st_mode) &&
 	 fstatat (chdir_fd, file_name, &st, 0) == 0 &&
@@ -870,6 +984,26 @@ is_directory_link (const char *file_name)
   return res;
 }
 
+/* Given struct stat of a directory (or directory member) whose ownership
+   or permissions of will be restored later, return the temporary permissions
+   for that directory, sufficiently restrictive so that in the meantime
+   processes owned by other users do not inadvertently create files under this
+   directory that inherit the wrong owner, group, or permissions from the
+   directory.
+
+   If not root, though, make the directory writeable and searchable at first,
+   so that files can be created under it.
+*/
+static inline int
+safe_dir_mode (struct stat const *st)
+{
+  return ((st->st_mode
+	   & (0 < same_owner_option || 0 < same_permissions_option
+	      ? S_IRWXU
+	      : MODE_RWX))
+	  | (we_are_root ? 0 : MODE_WXUSR));
+}
+
 /* Extractor functions for various member types */
 
 static int
@@ -899,18 +1033,7 @@ extract_dir (char *file_name, int typeflag)
   else if (typeflag == GNUTYPE_DUMPDIR)
     skip_member ();
 
-  /* If ownership or permissions will be restored later, create the
-     directory with restrictive permissions at first, so that in the
-     meantime processes owned by other users do not inadvertently
-     create files under this directory that inherit the wrong owner,
-     group, or permissions from the directory.  If not root, though,
-     make the directory writeable and searchable at first, so that
-     files can be created under it.  */
-  mode = ((current_stat_info.stat.st_mode
-	   & (0 < same_owner_option || 0 < same_permissions_option
-	      ? S_IRWXU
-	      : MODE_RWX))
-	  | (we_are_root ? 0 : MODE_WXUSR));
+  mode = safe_dir_mode (&current_stat_info.stat);
 
   for (;;)
     {
@@ -926,6 +1049,7 @@ extract_dir (char *file_name, int typeflag)
       if (errno == EEXIST
 	  && (interdir_made
 	      || keep_directory_symlink_option
+	      || old_files_option == NO_OVERWRITE_DIR_OLD_FILES
 	      || old_files_option == DEFAULT_OLD_FILES
 	      || old_files_option == OVERWRITE_OLD_FILES))
 	{
@@ -933,7 +1057,7 @@ extract_dir (char *file_name, int typeflag)
 
 	  if (keep_directory_symlink_option && is_directory_link (file_name))
 	    return 0;
-	  
+
 	  if (deref_stat (file_name, &st) == 0)
 	    {
 	      current_mode = st.st_mode;
@@ -946,6 +1070,31 @@ extract_dir (char *file_name, int typeflag)
 		      repair_delayed_set_stat (file_name, &st);
 		      return 0;
 		    }
+		  else if (old_files_option == NO_OVERWRITE_DIR_OLD_FILES)
+		    {
+		      /* Temporarily change the directory mode to a safe
+			 value, to be able to create files in it, should
+			 the need be.
+		      */
+		      mode = safe_dir_mode (&st);
+		      status = fd_chmod(-1, file_name, mode,
+					AT_SYMLINK_NOFOLLOW, DIRTYPE);
+		      if (status == 0)
+			{
+			  /* Store the actual directory mode, to be restored
+			     later.
+			  */
+			  current_stat_info.stat = st;
+			  current_mode = mode & ~ current_umask;
+			  current_mode_mask = MODE_RWX;
+			  atflag = AT_SYMLINK_NOFOLLOW;
+			  break;
+			}
+		      else
+			{
+			  chmod_error_details (file_name, mode);
+			}
+		    }
 		  break;
 		}
 	    }
@@ -1091,11 +1240,7 @@ extract_file (char *file_name, int typeflag)
       int file_created = 0;
       if (set_xattr (file_name, &current_stat_info, invert_permissions,
                      typeflag, &file_created))
-        {
-          skip_member ();
-          open_error (file_name);
-          return 1;
-        }
+        return 1;
 
       while ((fd = open_output_file (file_name, typeflag, mode,
                                      file_created, &current_mode,
@@ -1178,20 +1323,69 @@ extract_file (char *file_name, int typeflag)
   return status;
 }
 
+/* Find a delayed_link structure corresponding to the source NAME.
+   Such a structure exists in the delayed link list only if the link
+   placeholder file has been created. Therefore, try to stat the NAME
+   first. If it doesn't exist, there is no matching entry in the list.
+   Otherwise, look for the entry in list which has the matching dev
+   and ino numbers.
+   
+   This approach avoids scanning the singly-linked list in obvious cases
+   and does not rely on comparing file names, which may differ for
+   various reasons (e.g. relative vs. absolute file names).
+ */
+static struct delayed_link *
+find_delayed_link_source (char const *name)
+{
+  struct delayed_link *dl;
+  struct stat st;
+
+  if (!delayed_link_head)
+    return NULL;
+  
+  if (fstatat (chdir_fd, name, &st, AT_SYMLINK_NOFOLLOW))
+    {
+      if (errno != ENOENT)
+	stat_error (name);
+      return NULL;
+    }
+  
+  for (dl = delayed_link_head; dl; dl = dl->next)
+    {
+      if (dl->dev == st.st_dev && dl->ino == st.st_ino)
+	break;
+    }
+  return dl;
+}
+  
 /* Create a placeholder file with name FILE_NAME, which will be
    replaced after other extraction is done by a symbolic link if
    IS_SYMLINK is true, and by a hard link otherwise.  Set
    *INTERDIR_MADE if an intermediate directory is made in the
-   process.  */
+   process.
+   Install the created struct delayed_link after PREV, unless the
+   latter is NULL, in which case insert it at the head of the delayed
+   link list.
+*/
 
 static int
-create_placeholder_file (char *file_name, bool is_symlink, bool *interdir_made)
+create_placeholder_file (char *file_name, bool is_symlink, bool *interdir_made,
+			 struct delayed_link *prev)
 {
   int fd;
   struct stat st;
 
   while ((fd = openat (chdir_fd, file_name, O_WRONLY | O_CREAT | O_EXCL, 0)) < 0)
     {
+      if (errno == EEXIST && find_delayed_link_source (file_name))
+	{
+	  /* The placeholder file has already been created.  This means
+	     that the link being extracted is a duplicate of an already
+	     processed one.  Skip it.
+	   */
+	  return 0;
+	}
+      
       switch (maybe_recoverable (file_name, false, interdir_made))
 	{
 	case RECOVER_OK:
@@ -1220,8 +1414,16 @@ create_placeholder_file (char *file_name, bool is_symlink, bool *interdir_made)
 	xmalloc (offsetof (struct delayed_link, target)
 		 + strlen (current_stat_info.link_name)
 		 + 1);
-      p->next = delayed_link_head;
-      delayed_link_head = p;
+      if (prev)
+	{
+	  p->next = prev->next;
+	  prev->next = p;
+	}
+      else
+	{
+	  p->next = delayed_link_head;
+	  delayed_link_head = p;
+	}
       p->dev = st.st_dev;
       p->ino = st.st_ino;
       p->birthtime = get_stat_birthtime (&st);
@@ -1236,7 +1438,7 @@ create_placeholder_file (char *file_name, bool is_symlink, bool *interdir_made)
 	}
       p->change_dir = chdir_current;
       p->sources = xmalloc (offsetof (struct string_list, string)
-			    + strlen (file_name) + 1);
+			     + strlen (file_name) + 1);
       p->sources->next = 0;
       strcpy (p->sources->string, file_name);
       p->cntx_name = NULL;
@@ -1245,14 +1447,11 @@ create_placeholder_file (char *file_name, bool is_symlink, bool *interdir_made)
       p->acls_a_len = 0;
       p->acls_d_ptr = NULL;
       p->acls_d_len = 0;
-      xheader_xattr_copy (&current_stat_info, &p->xattr_map, &p->xattr_map_size);
+      xheader_xattr_copy (&current_stat_info, &p->xattr_map,
+			  &p->xattr_map_size);
       strcpy (p->target, current_stat_info.link_name);
 
-      h = delayed_set_stat_head;
-      if (h && ! h->after_links
-	  && strncmp (file_name, h->file_name, h->file_name_len) == 0
-	  && ISSLASH (file_name[h->file_name_len])
-	  && (last_component (file_name) == file_name + h->file_name_len + 1))
+      if ((h = find_direct_ancestor (file_name)) != NULL)
 	mark_after_links (h);
 
       return 0;
@@ -1267,12 +1466,16 @@ extract_link (char *file_name, int typeflag)
   bool interdir_made = false;
   char const *link_name;
   int rc;
-
+  struct delayed_link *dl;
+  
   link_name = current_stat_info.link_name;
 
   if (! absolute_names_option && contains_dot_dot (link_name))
-    return create_placeholder_file (file_name, false, &interdir_made);
-
+    return create_placeholder_file (file_name, false, &interdir_made, NULL);
+  dl = find_delayed_link_source (link_name);
+  if (dl)
+    return create_placeholder_file (file_name, false, &interdir_made, dl);
+  
   do
     {
       struct stat st1, st2;
@@ -1334,7 +1537,7 @@ extract_symlink (char *file_name, int typeflag)
   if (! absolute_names_option
       && (IS_ABSOLUTE_FILE_NAME (current_stat_info.link_name)
 	  || contains_dot_dot (current_stat_info.link_name)))
-    return create_placeholder_file (file_name, true, &interdir_made);
+    return create_placeholder_file (file_name, true, &interdir_made, NULL);
 
   while (symlinkat (current_stat_info.link_name, chdir_fd, file_name) != 0)
     switch (maybe_recoverable (file_name, false, &interdir_made))
@@ -1427,47 +1630,23 @@ extract_fifo (char *file_name, int typeflag)
 }
 #endif
 
-static int
-extract_volhdr (char *file_name, int typeflag)
-{
-  skip_member ();
-  return 0;
-}
-
-static int
-extract_failure (char *file_name, int typeflag)
-{
-  return 1;
-}
-
-static int
-extract_skip (char *file_name, int typeflag)
-{
-  skip_member ();
-  return 0;
-}
-
 typedef int (*tar_extractor_t) (char *file_name, int typeflag);
 
 
-
 /* Prepare to extract a file. Find extractor function.
-   Return zero if extraction should not proceed.  */
+   Return true to proceed with the extraction, false to skip the current
+   member.  */
 
-static int
+static bool
 prepare_to_extract (char const *file_name, int typeflag, tar_extractor_t *fun)
 {
-  int rc = 1;
-
-  if (EXTRACT_OVER_PIPE)
-    rc = 0;
+  tar_extractor_t extractor = NULL;
 
   /* Select the extractor */
   switch (typeflag)
     {
     case GNUTYPE_SPARSE:
-      *fun = extract_file;
-      rc = 1;
+      extractor = extract_file;
       break;
 
     case AREGTYPE:
@@ -1476,106 +1655,106 @@ prepare_to_extract (char const *file_name, int typeflag, tar_extractor_t *fun)
       /* Appears to be a file.  But BSD tar uses the convention that a slash
 	 suffix means a directory.  */
       if (current_stat_info.had_trailing_slash)
-	*fun = extract_dir;
+	extractor = extract_dir;
       else
-	{
-	  *fun = extract_file;
-	  rc = 1;
-	}
+	extractor = extract_file;
       break;
 
     case SYMTYPE:
-      *fun = extract_symlink;
+      extractor = extract_symlink;
       break;
 
     case LNKTYPE:
-      *fun = extract_link;
+      extractor = extract_link;
       break;
 
 #if S_IFCHR
     case CHRTYPE:
       current_stat_info.stat.st_mode |= S_IFCHR;
-      *fun = extract_node;
+      extractor = extract_node;
       break;
 #endif
 
 #if S_IFBLK
     case BLKTYPE:
       current_stat_info.stat.st_mode |= S_IFBLK;
-      *fun = extract_node;
+      extractor = extract_node;
       break;
 #endif
 
 #if HAVE_MKFIFO || defined mkfifo
     case FIFOTYPE:
-      *fun = extract_fifo;
+      extractor = extract_fifo;
       break;
 #endif
 
     case DIRTYPE:
     case GNUTYPE_DUMPDIR:
-      *fun = extract_dir;
+      extractor = extract_dir;
       if (current_stat_info.is_dumpdir)
 	delay_directory_restore_option = true;
       break;
 
     case GNUTYPE_VOLHDR:
-      *fun = extract_volhdr;
-      break;
-
+      return false;
+      
     case GNUTYPE_MULTIVOL:
       ERROR ((0, 0,
 	      _("%s: Cannot extract -- file is continued from another volume"),
 	      quotearg_colon (current_stat_info.file_name)));
-      *fun = extract_skip;
-      break;
+      return false;
 
     case GNUTYPE_LONGNAME:
     case GNUTYPE_LONGLINK:
       ERROR ((0, 0, _("Unexpected long name header")));
-      *fun = extract_failure;
-      break;
+      return false;
 
     default:
       WARNOPT (WARN_UNKNOWN_CAST,
 	       (0, 0,
 		_("%s: Unknown file type '%c', extracted as normal file"),
 		quotearg_colon (file_name), typeflag));
-      *fun = extract_file;
+      extractor = extract_file;
     }
 
-  /* Determine whether the extraction should proceed */
-  if (rc == 0)
-    return 0;
-
-  switch (old_files_option)
+  if (EXTRACT_OVER_PIPE)
     {
-    case UNLINK_FIRST_OLD_FILES:
-      if (!remove_any_file (file_name,
-                            recursive_unlink_option ? RECURSIVE_REMOVE_OPTION
-                                                      : ORDINARY_REMOVE_OPTION)
-	  && errno && errno != ENOENT)
-	{
-	  unlink_error (file_name);
-	  return 0;
-	}
-      break;
-
-    case KEEP_NEWER_FILES:
-      if (file_newer_p (file_name, 0, &current_stat_info))
-	{
-	  WARNOPT (WARN_IGNORE_NEWER,
-		   (0, 0, _("Current %s is newer or same age"),
-		    quote (file_name)));
-	  return 0;
-	}
-      break;
-
-    default:
-      break;
+      if (extractor != extract_file)
+	return false;
     }
+  else
+    {
+      switch (old_files_option)
+	{
+	case UNLINK_FIRST_OLD_FILES:
+	  if (!remove_any_file (file_name,
+				recursive_unlink_option
+				  ? RECURSIVE_REMOVE_OPTION
+				  : ORDINARY_REMOVE_OPTION)
+	      && errno && errno != ENOENT)
+	    {
+	      unlink_error (file_name);
+	      return false;
+	    }
+	  break;
 
-  return 1;
+	case KEEP_NEWER_FILES:
+	  if (file_newer_p (file_name, 0, &current_stat_info))
+	    {
+	      WARNOPT (WARN_IGNORE_NEWER,
+		       (0, 0, _("Current %s is newer or same age"),
+			quote (file_name)));
+	      return false;
+	    }
+	  break;
+
+	default:
+	  break;
+	}
+    }
+  *fun = extractor;
+  
+  return true;
 }
 
 /* Extract a file from the archive.  */
@@ -1584,12 +1763,20 @@ extract_archive (void)
 {
   char typeflag;
   tar_extractor_t fun;
+  bool skip_dotdot_name;
 
   fatal_exit_hook = extract_finish;
 
   set_next_block_after (current_header);
 
+  skip_dotdot_name = (!absolute_names_option
+		      && contains_dot_dot (current_stat_info.orig_file_name));
+  if (skip_dotdot_name)
+    ERROR ((0, 0, _("%s: Member name contains '..'"),
+	    quotearg_colon (current_stat_info.orig_file_name)));
+
   if (!current_stat_info.file_name[0]
+      || skip_dotdot_name
       || (interactive_option
 	  && !confirm ("extract", current_stat_info.file_name)))
     {
@@ -1630,13 +1817,14 @@ extract_archive (void)
 
   if (prepare_to_extract (current_stat_info.file_name, typeflag, &fun))
     {
-      if (fun && (*fun) (current_stat_info.file_name, typeflag)
-	  && backup_option)
-	undo_last_backup ();
+      if (fun (current_stat_info.file_name, typeflag) == 0)
+	return;
     }
   else
     skip_member ();
 
+  if (backup_option)
+    undo_last_backup ();
 }
 
 /* Extract the links whose final extraction were delayed.  */
@@ -1709,8 +1897,8 @@ apply_delayed_links (void)
 	  sources = next;
 	}
 
-   xheader_xattr_free (ds->xattr_map, ds->xattr_map_size);
-   free (ds->cntx_name);
+      xheader_xattr_free (ds->xattr_map, ds->xattr_map_size);
+      free (ds->cntx_name);
 
       {
 	struct delayed_link *next = ds->next;
@@ -1741,7 +1929,9 @@ extract_finish (void)
 bool
 rename_directory (char *src, char *dst)
 {
-  if (renameat (chdir_fd, src, chdir_fd, dst) != 0)
+  if (renameat (chdir_fd, src, chdir_fd, dst) == 0)
+    fixup_delayed_set_stat (src, dst);
+  else
     {
       int e = errno;
       bool interdir_made;

src/incremen.c

@@ -1,7 +1,6 @@
 /* GNU dump extensions to tar.
 
-   Copyright 1988, 1992-1994, 1996-1997, 1999-2001, 2003-2009, 2013-2014
-   Free Software Foundation, Inc.
+   Copyright 1988-2021 Free Software Foundation, Inc.
 
    This file is part of GNU tar.
 
@@ -39,7 +38,15 @@ enum children
 #define DIRF_FOUND    0x0004    /* directory is found on fs */
 #define DIRF_NEW      0x0008    /* directory is new (not found
 				   in the previous dump) */
-#define DIRF_RENAMED  0x0010    /* directory is renamed */
+#define DIRF_RENAMED  0x0010    /* Last target in a chain of renames */
+/* A directory which is renamed from another one is recognized by its
+   orig member, which is not-NULL.  This directory may eventually be
+   the source for another rename, in which case it will be pointed to by
+   the orig member of another directory structure.  The last directory
+   in such a chain of renames (the one which is not pointed to by any
+   other orig) is marked with the DIRF_RENAMED flag.  This marks a starting
+   point from which append_incremental_renames starts encoding renames for
+   this chain. */
 
 #define DIR_IS_INITED(d) ((d)->flags & DIRF_INIT)
 #define DIR_IS_NFS(d) ((d)->flags & DIRF_NFS)
@@ -496,6 +503,7 @@ procdir (const char *name_buffer, struct tar_stat_info *st,
 			    quote_n (1, d->name)));
 		  directory->orig = d;
 		  DIR_SET_FLAG (directory, DIRF_RENAMED);
+		  DIR_CLEAR_FLAG (d, DIRF_RENAMED);
 		  dirlist_replace_prefix (d->name, name_buffer);
 		}
 	      directory->children = CHANGED_CHILDREN;
@@ -538,6 +546,7 @@ procdir (const char *name_buffer, struct tar_stat_info *st,
 			quote_n (1, d->name)));
 	      directory->orig = d;
 	      DIR_SET_FLAG (directory, DIRF_RENAMED);
+	      DIR_CLEAR_FLAG (d, DIRF_RENAMED);
 	      dirlist_replace_prefix (d->name, name_buffer);
 	    }
 	  directory->children = CHANGED_CHILDREN;
@@ -650,7 +659,7 @@ makedumpdir (struct directory *directory, const char *dir)
 
   if (directory->children == ALL_CHILDREN)
     dump = NULL;
-  else if (DIR_IS_RENAMED (directory))
+  else if (directory->orig)
     dump = directory->orig->idump ?
            directory->orig->idump : directory->orig->dump;
   else
@@ -735,7 +744,7 @@ scan_directory (struct tar_stat_info *st)
     savedir_error (dir);
 
   info_attach_exclist (st);
-  
+
   tmp = xstrdup (dir);
   zap_slashes (tmp);
 
@@ -879,43 +888,36 @@ obstack_code_rename (struct obstack *stk, char const *from, char const *to)
 static void
 store_rename (struct directory *dir, struct obstack *stk)
 {
-  if (DIR_IS_RENAMED (dir))
+  struct directory *prev, *p;
+
+  /* Detect eventual cycles. If the chain forms a cycle, prev points to
+     the entry DIR is renamed from.*/
+  for (prev = dir; prev && prev->orig != dir; prev = prev->orig)
+    ;
+
+  if (prev == NULL)
     {
-      struct directory *prev, *p;
-
-      /* Detect eventual cycles and clear DIRF_RENAMED flag, so these entries
-	 are ignored when hit by this function next time.
-	 If the chain forms a cycle, prev points to the entry DIR is renamed
-	 from. In this case it still retains DIRF_RENAMED flag, which will be
-	 cleared in the 'else' branch below */
-      for (prev = dir; prev && prev->orig != dir; prev = prev->orig)
-	DIR_CLEAR_FLAG (prev, DIRF_RENAMED);
-
-      if (prev == NULL)
-	{
-	  for (p = dir; p && p->orig; p = p->orig)
-	    obstack_code_rename (stk, p->orig->name, p->name);
-	}
-      else
-	{
-	  char *temp_name;
-
-	  DIR_CLEAR_FLAG (prev, DIRF_RENAMED);
-
-	  /* Break the cycle by using a temporary name for one of its
-	     elements.
-	     First, create a temp name stub entry. */
-	  temp_name = dir_name (dir->name);
-	  obstack_1grow (stk, 'X');
-	  obstack_grow (stk, temp_name, strlen (temp_name) + 1);
-
-	  obstack_code_rename (stk, dir->name, "");
-
-	  for (p = dir; p != prev; p = p->orig)
-	    obstack_code_rename (stk, p->orig->name, p->name);
-
-	  obstack_code_rename (stk, "", prev->name);
-	}
+      for (p = dir; p && p->orig; p = p->orig)
+	obstack_code_rename (stk, p->orig->name, p->name);
+    }
+  else
+    {
+      char *temp_name;
+      
+      /* Break the cycle by using a temporary name for one of its
+	 elements.
+	 First, create a temp name stub entry. */
+      temp_name = dir_name (dir->name);
+      obstack_1grow (stk, 'X');
+      obstack_grow (stk, temp_name, strlen (temp_name) + 1);
+      
+      obstack_code_rename (stk, dir->name, "");
+      
+      for (p = dir; p != prev; p = p->orig)
+	obstack_code_rename (stk, p->orig->name, p->name);
+      
+      obstack_code_rename (stk, "", prev->name);
+      free (temp_name);
     }
 }
 
@@ -941,7 +943,8 @@ append_incremental_renames (struct directory *dir)
     size = 0;
 
   for (dp = dirhead; dp; dp = dp->next)
-    store_rename (dp, &stk);
+    if (DIR_IS_RENAMED (dp))
+      store_rename (dp, &stk);
 
   /* FIXME: Is this the right thing to do when DIR is null?  */
   if (dir && obstack_object_size (&stk) != size)
@@ -997,10 +1000,10 @@ read_incr_db_01 (int version, const char *initbuf)
   newer_mtime_option = decode_timespec (buf, &ebuf, false);
 
   if (! valid_timespec (newer_mtime_option))
-    ERROR ((0, errno, "%s:%ld: %s",
-	    quotearg_colon (listed_incremental_option),
-	    lineno,
-	    _("Invalid time stamp")));
+    FATAL_ERROR ((0, errno, "%s:%ld: %s",
+		  quotearg_colon (listed_incremental_option),
+		  lineno,
+		  _("Invalid time stamp")));
   else
     {
       if (version == 1 && *ebuf)
@@ -1042,9 +1045,9 @@ read_incr_db_01 (int version, const char *initbuf)
 	  mtime = decode_timespec (strp, &ebuf, false);
 	  strp = ebuf;
 	  if (!valid_timespec (mtime) || *strp != ' ')
-	    ERROR ((0, errno, "%s:%ld: %s",
-		    quotearg_colon (listed_incremental_option), lineno,
-		    _("Invalid modification time")));
+	    FATAL_ERROR ((0, errno, "%s:%ld: %s",
+			  quotearg_colon (listed_incremental_option), lineno,
+			  _("Invalid modification time")));
 
 	  errno = 0;
 	  u = strtoumax (strp, &ebuf, 10);
@@ -1052,9 +1055,9 @@ read_incr_db_01 (int version, const char *initbuf)
 	    errno = ERANGE;
 	  if (errno || strp == ebuf || *ebuf != ' ')
 	    {
-	      ERROR ((0, errno, "%s:%ld: %s",
-		      quotearg_colon (listed_incremental_option), lineno,
-		      _("Invalid modification time (nanoseconds)")));
+	      FATAL_ERROR ((0, errno, "%s:%ld: %s",
+			    quotearg_colon (listed_incremental_option), lineno,
+			    _("Invalid modification time (nanoseconds)")));
 	      mtime.tv_nsec = -1;
 	    }
 	  else
@@ -1068,17 +1071,17 @@ read_incr_db_01 (int version, const char *initbuf)
 			 TYPE_MINIMUM (dev_t), TYPE_MAXIMUM (dev_t));
       strp = ebuf;
       if (errno || *strp != ' ')
-	ERROR ((0, errno, "%s:%ld: %s",
+	FATAL_ERROR ((0, errno, "%s:%ld: %s",
 		quotearg_colon (listed_incremental_option), lineno,
-		_("Invalid device number")));
+		      _("Invalid device number")));
 
       ino = strtosysint (strp, &ebuf,
 			 TYPE_MINIMUM (ino_t), TYPE_MAXIMUM (ino_t));
       strp = ebuf;
       if (errno || *strp != ' ')
-	ERROR ((0, errno, "%s:%ld: %s",
-		quotearg_colon (listed_incremental_option), lineno,
-		_("Invalid inode number")));
+	FATAL_ERROR ((0, errno, "%s:%ld: %s",
+		      quotearg_colon (listed_incremental_option), lineno,
+		      _("Invalid inode number")));
 
       strp++;
       unquote_string (strp);
@@ -1155,11 +1158,14 @@ read_num (FILE *fp, char const *fieldname,
     }
 
   if (c)
-    FATAL_ERROR ((0, 0,
-		  _("%s: byte %s: %s %s followed by invalid byte 0x%02x"),
-		  quotearg_colon (listed_incremental_option),
-		  offtostr (ftello (fp), offbuf),
-		  fieldname, buf, c));
+    {
+      unsigned uc = c;
+      FATAL_ERROR ((0, 0,
+		    _("%s: byte %s: %s %s followed by invalid byte 0x%02x"),
+		    quotearg_colon (listed_incremental_option),
+		    offtostr (ftello (fp), offbuf),
+		    fieldname, buf, uc));
+    }
 
   *pval = strtosysint (buf, NULL, min_val, max_val);
   conversion_errno = errno;
@@ -1541,9 +1547,10 @@ dumpdir_ok (char *dumpdir)
     {
       if (expect && *p != expect)
 	{
+	  unsigned char uc = *p;
 	  ERROR ((0, 0,
 		  _("Malformed dumpdir: expected '%c' but found %#3o"),
-		  expect, *p));
+		  expect, uc));
 	  return false;
 	}
       switch (*p)
@@ -1578,7 +1585,7 @@ dumpdir_ok (char *dumpdir)
 	  if (expect != 'T')
 	    {
 	      ERROR ((0, 0,
-		      _("Malformed dumpdir: 'T' not preceeded by 'R'")));
+		      _("Malformed dumpdir: 'T' not preceded by 'R'")));
 	      return false;
 	    }
 	  if (p[1] == 0 && !has_tempdir)
@@ -1708,7 +1715,7 @@ try_purge_directory (char const *directory_name)
       const char *entry;
       struct stat st;
       free (p);
-      p = new_name (directory_name, cur);
+      p = make_file_name (directory_name, cur);
 
       if (deref_stat (p, &st) != 0)
 	{

src/list.c

@@ -1,7 +1,6 @@
 /* List a tar archive, with support routines for reading a tar archive.
 
-   Copyright 1988, 1992-1994, 1996-2001, 2003-2007, 2010, 2012-2014 Free
-   Software Foundation, Inc.
+   Copyright 1988-2021 Free Software Foundation, Inc.
 
    This file is part of GNU tar.
 
@@ -23,7 +22,7 @@
 #include <system.h>
 #include <inttostr.h>
 #include <quotearg.h>
-
+#include <time.h>
 #include "common.h"
 
 union block *current_header;	/* points to current archive header */
@@ -120,22 +119,24 @@ enforce_one_top_level (char **pfile_name)
 {
   char *file_name = *pfile_name;
   char *p;
-  
+
   for (p = file_name; *p && (ISSLASH (*p) || *p == '.'); p++)
     ;
 
-  if (!*p)
-    return;
-
-  if (strncmp (p, one_top_level_dir, strlen (one_top_level_dir)) == 0)
+  if (*p)
     {
       int pos = strlen (one_top_level_dir);
-      if (ISSLASH (p[pos]) || p[pos] == 0)
-	return;
-    }
+      if (strncmp (p, one_top_level_dir, pos) == 0)
+	{
+	  if (ISSLASH (p[pos]) || p[pos] == 0)
+	    return;
+	}
 
-  *pfile_name = new_name (one_top_level_dir, file_name);
-  normalize_filename_x (*pfile_name);
+      *pfile_name = make_file_name (one_top_level_dir, file_name);
+      normalize_filename_x (*pfile_name);
+    }
+  else
+    *pfile_name = xstrdup (one_top_level_dir);
   free (file_name);
 }
 
@@ -193,7 +194,7 @@ read_and (void (*do_something) (void))
 	  decode_header (current_header, &current_stat_info,
 			 &current_format, 1);
 	  if (! name_match (current_stat_info.file_name)
-	      || (NEWER_OPTION_INITIALIZED (newer_mtime_option)
+	      || (TIME_OPTION_INITIALIZED (newer_mtime_option)
 		  /* FIXME: We get mtime now, and again later; this causes
 		     duplicate diagnostics if header.mtime is bogus.  */
 		  && ((mtime.tv_sec
@@ -216,7 +217,7 @@ read_and (void (*do_something) (void))
 		  if (show_omitted_dirs_option)
 		    WARN ((0, 0, _("%s: Omitting"),
 			   quotearg_colon (current_stat_info.file_name)));
-		  /* Fall through.  */
+		  FALLTHROUGH;
 		default:
 		  skip_member ();
 		  continue;
@@ -271,8 +272,7 @@ read_and (void (*do_something) (void))
 	    {
 	    case HEADER_STILL_UNREAD:
 	      ERROR ((0, 0, _("This does not look like a tar archive")));
-	      /* Fall through.  */
-
+	      FALLTHROUGH;
 	    case HEADER_ZERO_BLOCK:
 	    case HEADER_SUCCESS:
 	      if (block_number_option)
@@ -408,26 +408,27 @@ read_header (union block **return_block, struct tar_stat_info *info,
 	     enum read_header_mode mode)
 {
   union block *header;
-  union block *header_copy;
   char *bp;
   union block *data_block;
   size_t size, written;
-  union block *next_long_name = 0;
-  union block *next_long_link = 0;
+  union block *next_long_name = NULL;
+  union block *next_long_link = NULL;
   size_t next_long_name_blocks = 0;
   size_t next_long_link_blocks = 0;
-
+  enum read_header status = HEADER_SUCCESS;
+  
   while (1)
     {
-      enum read_header status;
-
       header = find_next_block ();
       *return_block = header;
       if (!header)
-	return HEADER_END_OF_FILE;
+	{
+	  status = HEADER_END_OF_FILE;
+	  break;
+	}
 
       if ((status = tar_checksum (header, false)) != HEADER_SUCCESS)
-	return status;
+	break;
 
       /* Good block.  Decode file size and return.  */
 
@@ -437,7 +438,10 @@ read_header (union block **return_block, struct tar_stat_info *info,
 	{
 	  info->stat.st_size = OFF_FROM_HEADER (header->header.size);
 	  if (info->stat.st_size < 0)
-	    return HEADER_FAILURE;
+	    {
+	      status = HEADER_FAILURE;
+	      break;
+	    }
 	}
 
       if (header->header.typeflag == GNUTYPE_LONGNAME
@@ -447,10 +451,14 @@ read_header (union block **return_block, struct tar_stat_info *info,
 	  || header->header.typeflag == SOLARIS_XHDTYPE)
 	{
 	  if (mode == read_header_x_raw)
-	    return HEADER_SUCCESS_EXTENDED;
+	    {
+	      status = HEADER_SUCCESS_EXTENDED;
+	      break;
+	    }
 	  else if (header->header.typeflag == GNUTYPE_LONGNAME
 		   || header->header.typeflag == GNUTYPE_LONGLINK)
 	    {
+	      union block *header_copy;
 	      size_t name_size = info->stat.st_size;
 	      size_t n = name_size % BLOCKSIZE;
 	      size = name_size + BLOCKSIZE;
@@ -517,7 +525,10 @@ read_header (union block **return_block, struct tar_stat_info *info,
 	      xheader_decode_global (&xhdr);
 	      xheader_destroy (&xhdr);
 	      if (mode == read_header_x_global)
-		return HEADER_SUCCESS_EXTENDED;
+		{
+		  status = HEADER_SUCCESS_EXTENDED;
+		  break;
+		}
 	    }
 
 	  /* Loop!  */
@@ -536,6 +547,7 @@ read_header (union block **return_block, struct tar_stat_info *info,
 	      name = next_long_name->buffer + BLOCKSIZE;
 	      recent_long_name = next_long_name;
 	      recent_long_name_blocks = next_long_name_blocks;
+	      next_long_name = NULL;
 	    }
 	  else
 	    {
@@ -567,6 +579,7 @@ read_header (union block **return_block, struct tar_stat_info *info,
 	      name = next_long_link->buffer + BLOCKSIZE;
 	      recent_long_link = next_long_link;
 	      recent_long_link_blocks = next_long_link_blocks;
+	      next_long_link = NULL;
 	    }
 	  else
 	    {
@@ -578,9 +591,12 @@ read_header (union block **return_block, struct tar_stat_info *info,
 	    }
 	  assign_string (&info->link_name, name);
 
-	  return HEADER_SUCCESS;
+	  break;
 	}
     }
+  free (next_long_name);
+  free (next_long_link);
+  return status;
 }
 
 #define ISOCTAL(c) ((c)>='0'&&(c)<='7')
@@ -630,10 +646,12 @@ decode_header (union block *header, struct tar_stat_info *stat_info,
   stat_info->stat.st_mode = mode;
   stat_info->mtime.tv_sec = TIME_FROM_HEADER (header->header.mtime);
   stat_info->mtime.tv_nsec = 0;
-  assign_string (&stat_info->uname,
-		 header->header.uname[0] ? header->header.uname : NULL);
-  assign_string (&stat_info->gname,
-		 header->header.gname[0] ? header->header.gname : NULL);
+  assign_string_n (&stat_info->uname,
+		   header->header.uname[0] ? header->header.uname : NULL,
+		   sizeof (header->header.uname));
+  assign_string_n (&stat_info->gname,
+		   header->header.gname[0] ? header->header.gname : NULL,
+		   sizeof (header->header.gname));
 
   xheader_xattr_init (stat_info);
 
@@ -689,7 +707,6 @@ decode_header (union block *header, struct tar_stat_info *stat_info,
 	}
     }
 
-  stat_info->archive_file_size = stat_info->stat.st_size;
   xheader_decode (stat_info);
 
   if (sparse_member_p (stat_info))
@@ -1049,15 +1066,11 @@ tartime (struct timespec t, bool full_time)
     {
       if (full_time)
 	{
-	  sprintf (buffer, "%04ld-%02d-%02d %02d:%02d:%02d",
-		   tm->tm_year + 1900L, tm->tm_mon + 1, tm->tm_mday,
-		   tm->tm_hour, tm->tm_min, tm->tm_sec);
+	  strftime (buffer, sizeof buffer, "%Y-%m-%d %H:%M:%S", tm);
 	  code_ns_fraction (ns, buffer + strlen (buffer));
 	}
       else
-	sprintf (buffer, "%04ld-%02d-%02d %02d:%02d",
-		 tm->tm_year + 1900L, tm->tm_mon + 1, tm->tm_mday,
-		 tm->tm_hour, tm->tm_min);
+	strftime (buffer, sizeof buffer, "%Y-%m-%d %H:%M", tm);
       return buffer;
     }
 
@@ -1218,18 +1231,7 @@ simple_print_header (struct tar_stat_info *st, union block *blk,
 	  && !numeric_owner_option)
 	user = st->uname;
       else
-	{
-	  /* Try parsing it as an unsigned integer first, and as a
-	     uid_t if that fails.  This method can list positive user
-	     ids that are too large to fit in a uid_t.  */
-	  uintmax_t u = from_header (blk->header.uid,
-				     sizeof blk->header.uid, 0,
-				     0, UINTMAX_MAX,
-				     false, false);
-	  user = (u != -1
-		  ? STRINGIFY_BIGINT (u, uform)
-		  : imaxtostr (UID_FROM_HEADER (blk->header.uid), uform));
-	}
+	user = STRINGIFY_BIGINT (st->stat.st_uid, uform);
 
       if (st->gname
 	  && st->gname[0]
@@ -1237,18 +1239,7 @@ simple_print_header (struct tar_stat_info *st, union block *blk,
 	  && !numeric_owner_option)
 	group = st->gname;
       else
-	{
-	  /* Try parsing it as an unsigned integer first, and as a
-	     gid_t if that fails.  This method can list positive group
-	     ids that are too large to fit in a gid_t.  */
-	  uintmax_t g = from_header (blk->header.gid,
-				     sizeof blk->header.gid, 0,
-				     0, UINTMAX_MAX,
-				     false, false);
-	  group = (g != -1
-		   ? STRINGIFY_BIGINT (g, gform)
-		   : imaxtostr (GID_FROM_HEADER (blk->header.gid), gform));
-	}
+	group = STRINGIFY_BIGINT (st->stat.st_gid, gform);
 
       /* Format the file size or major/minor device numbers.  */
 
@@ -1465,7 +1456,7 @@ test_archive_label (void)
       decode_header (current_header,
 		     &current_stat_info, &current_format, 0);
       if (current_header->header.typeflag == GNUTYPE_VOLHDR)
-	assign_string (&volume_label, current_header->header.name);
+	ASSIGN_STRING_N (&volume_label, current_header->header.name);
 
       if (volume_label)
 	{

src/map.c

@@ -0,0 +1,283 @@
+/* Owner/group mapping for tar
+
+   Copyright 2015-2021 Free Software Foundation, Inc.
+
+   This file is part of GNU tar.
+
+   GNU tar is free software; you can redistribute it and/or modify
+   it under the terms of the GNU General Public License as published by
+   the Free Software Foundation; either version 3 of the License, or
+   (at your option) any later version.
+
+   GNU tar is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+   GNU General Public License for more details.
+
+   You should have received a copy of the GNU General Public License
+   along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
+
+#include <system.h>
+#include "common.h"
+#include "wordsplit.h"
+#include <hash.h>
+#include <pwd.h>
+
+struct mapentry
+{
+  uintmax_t orig_id;
+  uintmax_t new_id;
+  char *new_name;
+};
+
+static size_t
+map_hash (void const *entry, size_t nbuckets)
+{
+  struct mapentry const *map = entry;
+  return map->orig_id % nbuckets;
+}
+
+static bool
+map_compare (void const *entry1, void const *entry2)
+{
+  struct mapentry const *map1 = entry1;
+  struct mapentry const *map2 = entry2;
+  return map1->orig_id == map2->orig_id;
+}
+
+static int
+parse_id (uintmax_t *retval,
+	  char const *arg, char const *what, uintmax_t maxval,
+	  char const *file, unsigned line)
+{
+  uintmax_t v;
+  char *p;
+  
+  errno = 0;
+  v = strtoumax (arg, &p, 10);
+  if (*p || errno)
+    {
+      error (0, 0, _("%s:%u: invalid %s: %s"),  file, line, what, arg);
+      return -1;
+    }
+  if (v > maxval)
+    {
+      error (0, 0, _("%s:%u: %s out of range: %s"), file, line, what, arg);
+      return -1;
+    }
+  *retval = v;
+  return 0;
+}
+
+static void
+map_read (Hash_table **ptab, char const *file,
+	  uintmax_t (*name_to_id) (char const *), char const *what,
+	  uintmax_t maxval)
+{
+  FILE *fp;
+  char *buf = NULL;
+  size_t bufsize = 0;
+  ssize_t n;
+  struct wordsplit ws;
+  int wsopt;
+  unsigned line;
+  int err = 0;
+  
+  fp = fopen (file, "r");
+  if (!fp)
+    open_fatal (file);
+
+  ws.ws_comment = "#";
+  wsopt = WRDSF_COMMENT | WRDSF_NOVAR | WRDSF_NOCMD | WRDSF_SQUEEZE_DELIMS
+          | WRDSF_QUOTE;
+  line = 0;
+  while ((n = getline (&buf, &bufsize, fp)) > 0)
+    {
+      struct mapentry *ent;
+      uintmax_t orig_id, new_id;
+      char *name = NULL;
+      char *colon;
+      
+      ++line;
+      if (wordsplit (buf, &ws, wsopt))
+	FATAL_ERROR ((0, 0, _("%s:%u: cannot split line: %s"),
+		      file, line, wordsplit_strerror (&ws)));
+      wsopt |= WRDSF_REUSE;
+      if (ws.ws_wordc == 0)
+	continue;
+      if (ws.ws_wordc != 2)
+	{
+	  error (0, 0, _("%s:%u: malformed line"), file, line);
+	  err = 1;
+	  continue;
+	}
+
+      if (ws.ws_wordv[0][0] == '+')
+	{
+	  if (parse_id (&orig_id, ws.ws_wordv[0]+1, what, maxval, file, line)) 
+	    {
+	      err = 1;
+	      continue;
+	    }
+	}
+      else if (name_to_id)
+	{
+	  orig_id = name_to_id (ws.ws_wordv[0]);
+	  if (orig_id == UINTMAX_MAX)
+	    {
+	      error (0, 0, _("%s:%u: can't obtain %s of %s"),
+		     file, line, what, ws.ws_wordv[0]);
+	      err = 1;
+	      continue;
+	    }
+	}
+
+      colon = strchr (ws.ws_wordv[1], ':');
+      if (colon)
+	{
+	  if (colon > ws.ws_wordv[1])
+	    name = ws.ws_wordv[1];
+	  *colon++ = 0;
+	  if (parse_id (&new_id, colon, what, maxval, file, line)) 
+	    {
+	      err = 1;
+	      continue;
+	    }
+	}
+      else if (ws.ws_wordv[1][0] == '+')
+	{
+	  if (parse_id (&new_id, ws.ws_wordv[1], what, maxval, file, line)) 
+	    {
+	      err = 1;
+	      continue;
+	    }
+	}
+      else
+	{
+	  name = ws.ws_wordv[1];
+	  new_id = name_to_id (ws.ws_wordv[1]);
+	  if (new_id == UINTMAX_MAX)
+	    {
+	      error (0, 0, _("%s:%u: can't obtain %s of %s"),
+		     file, line, what, ws.ws_wordv[1]);
+	      err = 1;
+	      continue;
+	    }
+	}
+
+      ent = xmalloc (sizeof (*ent));
+      ent->orig_id = orig_id;
+      ent->new_id = new_id;
+      ent->new_name = name ? xstrdup (name) : NULL;
+      
+      if (!((*ptab
+	     || (*ptab = hash_initialize (0, 0, map_hash, map_compare, 0)))
+	    && hash_insert (*ptab, ent)))
+	xalloc_die ();
+    }
+  if (wsopt & WRDSF_REUSE)
+    wordsplit_free (&ws);
+  fclose (fp);
+  if (err)
+    FATAL_ERROR ((0, 0, _("errors reading map file")));
+}
+
+/* UID translation */
+
+static Hash_table *owner_map;
+
+static uintmax_t
+name_to_uid (char const *name)
+{
+  struct passwd *pw = getpwnam (name);
+  return pw ? pw->pw_uid : UINTMAX_MAX;
+}
+
+void
+owner_map_read (char const *file)
+{
+  map_read (&owner_map, file, name_to_uid, "UID", TYPE_MAXIMUM (uid_t));
+}
+
+int
+owner_map_translate (uid_t uid, uid_t *new_uid, char const **new_name)
+{
+  int rc = 1;
+  
+  if (owner_map)
+    {
+      struct mapentry ent, *res;
+  
+      ent.orig_id = uid;
+      res = hash_lookup (owner_map, &ent);
+      if (res)
+	{
+	  *new_uid = res->new_id;
+	  *new_name = res->new_name;
+	  return 0;
+	}
+    }
+
+  if (owner_option != (uid_t) -1)
+    {
+      *new_uid = owner_option;
+      rc = 0;
+    }
+  if (owner_name_option)
+    {
+      *new_name = owner_name_option;
+      rc = 0;
+    }
+
+  return rc;
+}
+
+/* GID translation */
+
+static Hash_table *group_map;
+
+static uintmax_t
+name_to_gid (char const *name)
+{
+  struct group *gr = getgrnam (name);
+  return gr ? gr->gr_gid : UINTMAX_MAX;
+}
+
+void
+group_map_read (char const *file)
+{
+  map_read (&group_map, file, name_to_gid, "GID", TYPE_MAXIMUM (gid_t));
+}
+
+int
+group_map_translate (gid_t gid, gid_t *new_gid, char const **new_name)
+{
+  int rc = 1;
+  
+  if (group_map)
+    {
+      struct mapentry ent, *res;
+  
+      ent.orig_id = gid;
+      res = hash_lookup (group_map, &ent);
+      if (res)
+	{
+	  *new_gid = res->new_id;
+	  *new_name = res->new_name;
+	  return 0;
+	}
+    }
+
+  if (group_option != (uid_t) -1)
+    {
+      *new_gid = group_option;
+      rc = 0;
+    }
+  if (group_name_option)
+    {
+      *new_name = group_name_option;
+      rc = 0;
+    }
+  
+  return rc;
+}

src/misc.c

@@ -1,7 +1,6 @@
 /* Miscellaneous functions, not really specific to GNU tar.
 
-   Copyright 1988, 1992, 1994-1997, 1999-2001, 2003-2007, 2009-2010,
-   2012-2014 Free Software Foundation, Inc.
+   Copyright 1988-2021 Free Software Foundation, Inc.
 
    This program is free software; you can redistribute it and/or modify it
    under the terms of the GNU General Public License as published by the
@@ -29,8 +28,15 @@
 # define DOUBLE_SLASH_IS_DISTINCT_ROOT 0
 #endif
 
+static void namebuf_add_dir (namebuf_t, char const *);
+static char *namebuf_finish (namebuf_t);
 static const char *tar_getcdpath (int);
 
+char const *
+quote_n_colon (int n, char const *arg)
+{
+  return quotearg_n_style_colon (n, get_quoting_style (NULL), arg);
+}
 
 /* Handling strings.  */
 
@@ -43,6 +49,20 @@ assign_string (char **string, const char *value)
   *string = value ? xstrdup (value) : 0;
 }
 
+void
+assign_string_n (char **string, const char *value, size_t n)
+{
+  free (*string);
+  if (value)
+    {
+      size_t l = strnlen (value, n);
+      char *p = xmalloc (l + 1);
+      memcpy (p, value, l);
+      p[l] = 0;
+      *string = p;
+    }
+}
+
 #if 0
 /* This function is currently unused; perhaps it should be removed?  */
 
@@ -294,8 +314,6 @@ normalize_filename (int cdidx, const char *name)
       size_t copylen;
       bool need_separator;
 
-      if (!cdpath)
-	call_arg_fatal ("getcwd", ".");
       copylen = strlen (cdpath);
       need_separator = ! (DOUBLE_SLASH_IS_DISTINCT_ROOT
 			  && copylen == 2 && ISSLASH (cdpath[1]));
@@ -586,7 +604,12 @@ safer_rmdir (const char *file_name)
       return -1;
     }
 
-  return unlinkat (chdir_fd, file_name, AT_REMOVEDIR);
+  if (unlinkat (chdir_fd, file_name, AT_REMOVEDIR) == 0)
+    {
+      remove_delayed_set_stat (file_name);
+      return 0;
+    }
+  return -1;
 }
 
 /* Remove FILE_NAME, returning 1 on success.  If FILE_NAME is a directory,
@@ -650,7 +673,7 @@ remove_any_file (const char *file_name, enum remove_option option)
 		 (entrylen = strlen (entry)) != 0;
 		 entry += entrylen + 1)
 	      {
-		char *file_name_buffer = new_name (file_name, entry);
+		char *file_name_buffer = make_file_name (file_name, entry);
 		int r = remove_any_file (file_name_buffer,
                                          RECURSIVE_REMOVE_OPTION);
 		int e = errno;
@@ -715,7 +738,7 @@ maybe_backup_file (const char *file_name, bool this_is_the_archive)
       && (S_ISBLK (file_stat.st_mode) || S_ISCHR (file_stat.st_mode)))
     return true;
 
-  after_backup_name = find_backup_file_name (file_name, backup_type);
+  after_backup_name = find_backup_file_name (chdir_fd, file_name, backup_type);
   if (! after_backup_name)
     xalloc_die ();
 
@@ -885,8 +908,6 @@ chdir_count (void)
 int
 chdir_arg (char const *dir)
 {
-  char *absdir;
-
   if (wd_count == wd_alloc)
     {
       if (wd_alloc == 0)
@@ -896,7 +917,7 @@ chdir_arg (char const *dir)
       if (! wd_count)
 	{
 	  wd[wd_count].name = ".";
-	  wd[wd_count].abspath = xgetcwd ();
+	  wd[wd_count].abspath = NULL;
 	  wd[wd_count].fd = AT_FDCWD;
 	  wd_count++;
 	}
@@ -913,22 +934,8 @@ chdir_arg (char const *dir)
 	return wd_count - 1;
     }
 
-
-  /* If the given name is absolute, use it to represent this directory;
-     otherwise, construct a name based on the previous -C option.  */
-  if (IS_ABSOLUTE_FILE_NAME (dir))
-    absdir = xstrdup (dir);
-  else if (wd[wd_count - 1].abspath)
-    {
-      namebuf_t nbuf = namebuf_create (wd[wd_count - 1].abspath);
-      namebuf_add_dir (nbuf, dir);
-      absdir = namebuf_finish (nbuf);
-    }
-  else
-    absdir = 0;
-
   wd[wd_count].name = dir;
-  wd[wd_count].abspath = absdir;
+  wd[wd_count].abspath = NULL;
   wd[wd_count].fd = 0;
   return wd_count++;
 }
@@ -1022,9 +1029,47 @@ tar_getcdpath (int idx)
     {
       static char *cwd;
       if (!cwd)
-	cwd = xgetcwd ();
+	{
+	  cwd = xgetcwd ();
+	  if (!cwd)
+	    call_arg_fatal ("getcwd", ".");
+	}
       return cwd;
     }
+
+  if (!wd[idx].abspath)
+    {
+      int i;
+      int save_cwdi = chdir_current;
+      
+      for (i = idx; i >= 0; i--)
+	if (wd[i].abspath)
+	  break;
+      
+      while (++i <= idx)
+	{
+	  chdir_do (i);
+	  if (i == 0)
+	    {
+	      if ((wd[i].abspath = xgetcwd ()) == NULL)
+		call_arg_fatal ("getcwd", ".");
+	    }
+	  else if (IS_ABSOLUTE_FILE_NAME (wd[i].name))
+	    /* If the given name is absolute, use it to represent this
+	       directory; otherwise, construct a name based on the
+	       previous -C option.  */
+	    wd[i].abspath = xstrdup (wd[i].name);
+	  else
+	    {
+	      namebuf_t nbuf = namebuf_create (wd[i - 1].abspath);
+	      namebuf_add_dir (nbuf, wd[i].name);
+	      wd[i].abspath = namebuf_finish (nbuf);
+	    }
+	}
+
+      chdir_do (save_cwdi);
+    }
+	   
   return wd[idx].abspath;
 }
 
@@ -1032,7 +1077,10 @@ void
 close_diag (char const *name)
 {
   if (ignore_failed_read_option)
-    close_warn (name);
+    {
+      if (WARNING_ENABLED(WARN_FAILED_READ))
+	close_warn (name);
+    }
   else
     close_error (name);
 }
@@ -1041,7 +1089,10 @@ void
 open_diag (char const *name)
 {
   if (ignore_failed_read_option)
-    open_warn (name);
+    {
+      if (WARNING_ENABLED(WARN_FAILED_READ))
+	open_warn (name);
+    }
   else
     open_error (name);
 }
@@ -1050,7 +1101,10 @@ void
 read_diag_details (char const *name, off_t offset, size_t size)
 {
   if (ignore_failed_read_option)
-    read_warn_details (name, offset, size);
+    {
+      if (WARNING_ENABLED(WARN_FAILED_READ))
+	read_warn_details (name, offset, size);
+    }
   else
     read_error_details (name, offset, size);
 }
@@ -1059,7 +1113,10 @@ void
 readlink_diag (char const *name)
 {
   if (ignore_failed_read_option)
-    readlink_warn (name);
+    {
+      if (WARNING_ENABLED(WARN_FAILED_READ))
+	readlink_warn (name);
+    }
   else
     readlink_error (name);
 }
@@ -1068,7 +1125,10 @@ void
 savedir_diag (char const *name)
 {
   if (ignore_failed_read_option)
-    savedir_warn (name);
+    {
+      if (WARNING_ENABLED(WARN_FAILED_READ))
+	savedir_warn (name);
+    }
   else
     savedir_error (name);
 }
@@ -1077,7 +1137,10 @@ void
 seek_diag_details (char const *name, off_t offset)
 {
   if (ignore_failed_read_option)
-    seek_warn_details (name, offset);
+    {
+      if (WARNING_ENABLED(WARN_FAILED_READ))
+	seek_warn_details (name, offset);
+    }
   else
     seek_error_details (name, offset);
 }
@@ -1086,7 +1149,10 @@ void
 stat_diag (char const *name)
 {
   if (ignore_failed_read_option)
-    stat_warn (name);
+    {
+      if (WARNING_ENABLED(WARN_FAILED_READ))
+	stat_warn (name);
+    }
   else
     stat_error (name);
 }
@@ -1106,13 +1172,6 @@ file_removed_diag (const char *name, bool top_level,
     diagfn (name);
 }
 
-void
-write_fatal_details (char const *name, ssize_t status, size_t size)
-{
-  write_error_details (name, status, size);
-  fatal_exit ();
-}
-
 /* Fork, aborting if unsuccessful.  */
 pid_t
 xfork (void)
@@ -1197,7 +1256,7 @@ namebuf_name (namebuf_t buf, const char *name)
   return buf->buffer;
 }
 
-void
+static void
 namebuf_add_dir (namebuf_t buf, const char *name)
 {
   static char dirsep[] = { DIRECTORY_SEPARATOR, 0 };
@@ -1210,7 +1269,7 @@ namebuf_add_dir (namebuf_t buf, const char *name)
   buf->dir_length += strlen (name);
 }
 
-char *
+static char *
 namebuf_finish (namebuf_t buf)
 {
   char *res = buf->buffer;

src/sparse.c

@@ -1,6 +1,6 @@
 /* Functions for dealing with sparse files
 
-   Copyright 2003-2007, 2010, 2013-2014 Free Software Foundation, Inc.
+   Copyright 2003-2021 Free Software Foundation, Inc.
 
    This program is free software; you can redistribute it and/or modify it
    under the terms of the GNU General Public License as published by the
@@ -208,9 +208,9 @@ sparse_add_map (struct tar_stat_info *st, struct sp_array const *sp)
   st->sparse_map_avail = avail + 1;
 }
 
-/* Scan the sparse file and create its map */
+/* Scan the sparse file byte-by-byte and create its map. */
 static bool
-sparse_scan_file (struct tar_sparse_file *file)
+sparse_scan_file_raw (struct tar_sparse_file *file)
 {
   struct tar_stat_info *st = file->stat_info;
   int fd = file->fd;
@@ -221,41 +221,38 @@ sparse_scan_file (struct tar_sparse_file *file)
 
   st->archive_file_size = 0;
 
-  if (ST_NBLOCKS (st->stat) == 0)
-    offset = st->stat.st_size;
-  else
+  if (!tar_sparse_scan (file, scan_begin, NULL))
+    return false;
+
+  while ((count = blocking_read (fd, buffer, sizeof buffer)) != 0
+         && count != SAFE_READ_ERROR)
     {
-      if (!tar_sparse_scan (file, scan_begin, NULL))
-	return false;
+      /* Analyze the block.  */
+      if (zero_block_p (buffer, count))
+        {
+          if (sp.numbytes)
+            {
+              sparse_add_map (st, &sp);
+              sp.numbytes = 0;
+              if (!tar_sparse_scan (file, scan_block, NULL))
+                return false;
+            }
+        }
+      else
+        {
+          if (sp.numbytes == 0)
+            sp.offset = offset;
+          sp.numbytes += count;
+          st->archive_file_size += count;
+          if (!tar_sparse_scan (file, scan_block, buffer))
+            return false;
+        }
 
-      while ((count = blocking_read (fd, buffer, sizeof buffer)) != 0
-	     && count != SAFE_READ_ERROR)
-	{
-	  /* Analyze the block.  */
-	  if (zero_block_p (buffer, count))
-	    {
-	      if (sp.numbytes)
-		{
-		  sparse_add_map (st, &sp);
-		  sp.numbytes = 0;
-		  if (!tar_sparse_scan (file, scan_block, NULL))
-		    return false;
-		}
-	    }
-	  else
-	    {
-	      if (sp.numbytes == 0)
-		sp.offset = offset;
-	      sp.numbytes += count;
-	      st->archive_file_size += count;
-	      if (!tar_sparse_scan (file, scan_block, buffer))
-		return false;
-	    }
-
-	  offset += count;
-	}
+      offset += count;
     }
 
+  /* save one more sparse segment of length 0 to indicate that
+     the file ends with a hole */
   if (sp.numbytes == 0)
     sp.offset = offset;
 
@@ -264,6 +261,115 @@ sparse_scan_file (struct tar_sparse_file *file)
   return tar_sparse_scan (file, scan_end, NULL);
 }
 
+static bool
+sparse_scan_file_wholesparse (struct tar_sparse_file *file)
+{
+  struct tar_stat_info *st = file->stat_info;
+  struct sp_array sp = {0, 0};
+
+  /* Note that this function is called only for truly sparse files of size >= 1
+     block size (checked via ST_IS_SPARSE before).  See the thread
+     http://www.mail-archive.com/bug-tar@gnu.org/msg04209.html for more info */
+  if (ST_NBLOCKS (st->stat) == 0)
+    {
+      st->archive_file_size = 0;
+      sp.offset = st->stat.st_size;
+      sparse_add_map (st, &sp);
+      return true;
+    }
+
+  return false;
+}
+
+#ifdef SEEK_HOLE
+/* Try to engage SEEK_HOLE/SEEK_DATA feature. */
+static bool
+sparse_scan_file_seek (struct tar_sparse_file *file)
+{
+  struct tar_stat_info *st = file->stat_info;
+  int fd = file->fd;
+  struct sp_array sp = {0, 0};
+  off_t offset = 0;
+  off_t data_offset;
+  off_t hole_offset;
+
+  st->archive_file_size = 0;
+
+  for (;;)
+    {
+      /* locate first chunk of data */
+      data_offset = lseek (fd, offset, SEEK_DATA);
+
+      if (data_offset == (off_t)-1)
+        /* ENXIO == EOF; error otherwise */
+        {
+          if (errno == ENXIO)
+            {
+              /* file ends with hole, add one more empty chunk of data */
+              sp.numbytes = 0;
+              sp.offset = st->stat.st_size;
+              sparse_add_map (st, &sp);
+              return true;
+            }
+          return false;
+        }
+
+      hole_offset = lseek (fd, data_offset, SEEK_HOLE);
+
+      /* according to specs, if FS does not fully support
+	 SEEK_DATA/SEEK_HOLE it may just implement kind of "wrapper" around
+	 classic lseek() call.  We must detect it here and try to use other
+	 hole-detection methods. */
+      if (offset == 0 /* first loop */
+          && data_offset == 0
+          && hole_offset == st->stat.st_size)
+        {
+          lseek (fd, 0, SEEK_SET);
+          return false;
+        }
+
+      sp.offset = data_offset;
+      sp.numbytes = hole_offset - data_offset;
+      sparse_add_map (st, &sp);
+
+      st->archive_file_size += sp.numbytes;
+      offset = hole_offset;
+    }
+
+  return true;
+}
+#endif
+
+static bool
+sparse_scan_file (struct tar_sparse_file *file)
+{
+  /* always check for completely sparse files */
+  if (sparse_scan_file_wholesparse (file))
+    return true;
+
+  switch (hole_detection)
+    {
+    case HOLE_DETECTION_DEFAULT:
+    case HOLE_DETECTION_SEEK:
+#ifdef SEEK_HOLE
+      if (sparse_scan_file_seek (file))
+        return true;
+#else
+      if (hole_detection == HOLE_DETECTION_SEEK)
+	WARN((0, 0,
+	      _("\"seek\" hole detection is not supported, using \"raw\".")));
+      /* fall back to "raw" for this and all other files */
+      hole_detection = HOLE_DETECTION_RAW;
+#endif
+      FALLTHROUGH;
+    case HOLE_DETECTION_RAW:
+      if (sparse_scan_file_raw (file))
+	return true;
+    }
+
+  return false;
+}
+
 static struct tar_sparse_optab const oldgnu_optab;
 static struct tar_sparse_optab const star_optab;
 static struct tar_sparse_optab const pax_optab;
@@ -321,6 +427,30 @@ sparse_dump_region (struct tar_sparse_file *file, size_t i)
 			     bufsize);
 	  return false;
 	}
+      else if (bytes_read == 0)
+	{
+	  char buf[UINTMAX_STRSIZE_BOUND];
+	  struct stat st;
+	  size_t n;
+	  if (fstat (file->fd, &st) == 0)
+	    n = file->stat_info->stat.st_size - st.st_size;
+	  else
+	    n = file->stat_info->stat.st_size
+	      - (file->stat_info->sparse_map[i].offset
+		 + file->stat_info->sparse_map[i].numbytes
+		 - bytes_left);
+	  
+	  WARNOPT (WARN_FILE_SHRANK,
+		   (0, 0,
+		    ngettext ("%s: File shrank by %s byte; padding with zeros",
+			      "%s: File shrank by %s bytes; padding with zeros",
+			      n),
+		    quotearg_colon (file->stat_info->orig_file_name),
+		    STRINGIFY_BIGINT (n, buf)));
+	  if (! ignore_failed_read_option)
+	    set_exit_status (TAREXIT_DIFFERS);
+	  return false;
+	}
 
       memset (blk->buffer + bytes_read, 0, BLOCKSIZE - bytes_read);
       bytes_left -= bytes_read;
@@ -358,9 +488,9 @@ sparse_extract_region (struct tar_sparse_file *file, size_t i)
 	  return false;
 	}
       set_next_block_after (blk);
+      file->dumped_size += BLOCKSIZE;
       count = blocking_write (file->fd, blk->buffer, wrbytes);
       write_size -= count;
-      file->dumped_size += count;
       mv_size_left (file->stat_info->archive_file_size - file->dumped_size);
       file->offset += count;
       if (count != wrbytes)
@@ -492,6 +622,12 @@ check_sparse_region (struct tar_sparse_file *file, off_t beg, off_t end)
 			     rdsize);
 	  return false;
 	}
+      else if (bytes_read == 0)
+	{
+	  report_difference (file->stat_info, _("Size differs"));
+	  return false;
+	}
+      
       if (!zero_block_p (diff_buffer, bytes_read))
 	{
 	  char begbuf[INT_BUFSIZE_BOUND (off_t)];
@@ -503,6 +639,7 @@ check_sparse_region (struct tar_sparse_file *file, off_t beg, off_t end)
 
       beg += bytes_read;
     }
+
   return true;
 }
 
@@ -529,6 +666,7 @@ check_data_region (struct tar_sparse_file *file, size_t i)
 	  return false;
 	}
       set_next_block_after (blk);
+      file->dumped_size += BLOCKSIZE;      
       bytes_read = safe_read (file->fd, diff_buffer, rdsize);
       if (bytes_read == SAFE_READ_ERROR)
 	{
@@ -539,10 +677,14 @@ check_data_region (struct tar_sparse_file *file, size_t i)
 			     rdsize);
 	  return false;
 	}
-      file->dumped_size += bytes_read;
+      else if (bytes_read == 0)
+	{
+	  report_difference (&current_stat_info, _("Size differs"));
+	  return false;
+	}
       size_left -= bytes_read;
       mv_size_left (file->stat_info->archive_file_size - file->dumped_size);
-      if (memcmp (blk->buffer, diff_buffer, rdsize))
+      if (memcmp (blk->buffer, diff_buffer, bytes_read))
 	{
 	  report_difference (file->stat_info, _("Contents differ"));
 	  return false;
@@ -560,7 +702,7 @@ sparse_diff_file (int fd, struct tar_stat_info *st)
   off_t offset = 0;
 
   if (!tar_sparse_init (&file))
-    return dump_status_not_implemented;
+    return false;
 
   file.stat_info = st;
   file.fd = fd;
@@ -1107,7 +1249,8 @@ pax_decode_header (struct tar_sparse_file *file)
       union block *blk;
       char *p;
       size_t i;
-
+      off_t start;
+      
 #define COPY_BUF(b,buf,src) do                                     \
  {                                                                 \
    char *endp = b->buffer + BLOCKSIZE;                             \
@@ -1123,8 +1266,9 @@ pax_decode_header (struct tar_sparse_file *file)
        if (src == endp)                                            \
 	 {                                                         \
 	   set_next_block_after (b);                               \
-           file->dumped_size += BLOCKSIZE;                         \
            b = find_next_block ();                                 \
+           if (!b)                                                 \
+             FATAL_ERROR ((0, 0, _("Unexpected EOF in archive"))); \
            src = b->buffer;                                        \
 	   endp = b->buffer + BLOCKSIZE;                           \
 	 }                                                         \
@@ -1134,9 +1278,11 @@ pax_decode_header (struct tar_sparse_file *file)
    dst[-1] = 0;                                                    \
  } while (0)
 
+      start = current_block_ordinal ();
       set_next_block_after (current_header);
-      file->dumped_size += BLOCKSIZE;
       blk = find_next_block ();
+      if (!blk)
+        FATAL_ERROR ((0, 0, _("Unexpected EOF in archive")));
       p = blk->buffer;
       COPY_BUF (blk,nbuf,p);
       if (!decode_num (&u, nbuf, TYPE_MAXIMUM (size_t)))
@@ -1172,6 +1318,8 @@ pax_decode_header (struct tar_sparse_file *file)
 	  sparse_add_map (file->stat_info, &sp);
 	}
       set_next_block_after (blk);
+
+      file->dumped_size += BLOCKSIZE * (current_block_ordinal () - start);
     }
 
   return true;

src/suffix.c

@@ -1,5 +1,5 @@
 /* This file is part of GNU tar.
-   Copyright 2007, 2009, 2013-2014 Free Software Foundation, Inc.
+   Copyright 2007-2021 Free Software Foundation, Inc.
 
    Written by Sergey Poznyakoff.
 
@@ -45,6 +45,8 @@ static struct compression_suffix compression_suffixes[] = {
   { S(lzo,  LZOP) },
   { S(xz,   XZ) },
   { S(txz,  XZ) }, /* Slackware */
+  { S(zst,  ZSTD) },
+  { S(tzst, ZSTD) },
   { NULL }
 #undef S
 #undef __CAT2__
@@ -59,7 +61,7 @@ find_compression_suffix (const char *name, size_t *ret_len)
     {
       size_t len;
       struct compression_suffix *p;
-      
+
       suf++;
       len = strlen (suf);
 
@@ -98,10 +100,14 @@ strip_compression_suffix (const char *name)
 {
   char *s = NULL;
   size_t len;
+  struct compression_suffix const *p = find_compression_suffix (name, &len);
 
-  if (find_compression_suffix (name, &len))
+  if (p)
     {
-      if (strncmp (name + len - 4, ".tar", 4) == 0)
+      /* Strip an additional ".tar" suffix, but only if the just-stripped
+	 "outer" suffix did not begin with "t".  */
+      if (len > 4 && strncmp (name + len - 4, ".tar", 4) == 0
+	  && p->suffix[0] != 't')
 	len -= 4;
       if (len == 0)
 	return NULL;
@@ -111,4 +117,3 @@ strip_compression_suffix (const char *name)
     }
   return s;
 }
-  

src/system.c

@@ -1,6 +1,6 @@
 /* System-dependent calls for tar.
 
-   Copyright 2003-2008, 2010, 2013-2014 Free Software Foundation, Inc.
+   Copyright 2003-2021 Free Software Foundation, Inc.
 
    This program is free software; you can redistribute it and/or modify it
    under the terms of the GNU General Public License as published by the
@@ -26,13 +26,14 @@
 static _Noreturn void
 xexec (const char *cmd)
 {
-  struct wordsplit ws;
+  char *argv[4];
 
-  ws.ws_env = (const char **) environ;
-  if (wordsplit (cmd, &ws, (WRDSF_DEFFLAGS | WRDSF_ENV) & ~WRDSF_NOVAR))
-    FATAL_ERROR ((0, 0, _("cannot split string '%s': %s"),
-		  cmd, wordsplit_strerror (&ws)));
-  execvp (ws.ws_wordv[0], ws.ws_wordv);
+  argv[0] = (char *) "/bin/sh";
+  argv[1] = (char *) "-c";
+  argv[2] = (char *) cmd;
+  argv[3] = NULL;
+
+  execv ("/bin/sh", argv);
   exec_fatal (cmd);
 }
 
@@ -330,6 +331,7 @@ sys_child_open_for_compress (void)
   pid_t grandchild_pid;
   pid_t child_pid;
 
+  signal (SIGPIPE, SIG_IGN);
   xpipe (parent_pipe);
   child_pid = xfork ();
 

src/tar.h

@@ -1,7 +1,6 @@
 /* GNU tar Archive Format description.
 
-   Copyright 1988-1989, 1991-1997, 2000-2001, 2003-2007, 2012-2014 Free
-   Software Foundation, Inc.
+   Copyright 1988-2021 Free Software Foundation, Inc.
 
    This file is part of GNU tar.
 
@@ -327,6 +326,14 @@ struct tar_stat_info
   size_t sparse_map_size;   /* Size of the sparse map */
   struct sp_array *sparse_map;
 
+  off_t real_size;          /* The real size of sparse file */
+  bool  real_size_set;      /* True when GNU.sparse.realsize is set in
+			       archived file */
+
+  bool  sparse_name_done;   /* Set to true if 'GNU.sparse.name' header was
+                               processed pax header parsing.  Following 'path'
+                               header (lower priority) will be ignored. */
+
   size_t xattr_map_size;   /* Size of the xattr map */
   struct xattr_array *xattr_map;
 

src/transform.c

@@ -1,5 +1,5 @@
 /* This file is part of GNU tar.
-   Copyright 2006-2008, 2013-2014 Free Software Foundation, Inc.
+   Copyright 2006-2021 Free Software Foundation, Inc.
 
    This program is free software; you can redistribute it and/or modify it
    under the terms of the GNU General Public License as published by the
@@ -100,7 +100,7 @@ add_segment (struct transform *tf)
 }
 
 static void
-add_literal_segment (struct transform *tf, char *str, char *end)
+add_literal_segment (struct transform *tf, const char *str, const char *end)
 {
   size_t len = end - str;
   if (len)
@@ -208,6 +208,8 @@ parse_transform_expr (const char *expr)
     }
 
   delim = expr[1];
+  if (!delim)
+    USAGE_ERROR ((0, 0, _("Invalid transform expression")));
 
   /* Scan regular expression */
   for (i = 2; expr[i] && expr[i] != delim; i++)
@@ -272,7 +274,7 @@ parse_transform_expr (const char *expr)
       USAGE_ERROR ((0, 0, _("Invalid transform expression: %s"), errbuf));
     }
 
-  if (str[0] == '^' || str[strlen (str) - 1] == '$')
+  if (str[0] == '^' || (i > 2 && str[i - 3] == '$'))
     tf->transform_type = transform_first;
 
   free (str);
@@ -378,13 +380,15 @@ parse_transform_expr (const char *expr)
 	      break;
 
 	    default:
-	      /* Try to be nice */
-	      {
-		char buf[2];
-		buf[0] = '\\';
-		buf[1] = *cur;
-		add_literal_segment (tf, buf, buf + 2);
-	      }
+	      if (*cur == delim)
+		add_char_segment (tf, delim);
+	      else
+		{
+		  char buf[2];
+		  buf[0] = '\\';
+		  buf[1] = *cur;
+		  add_literal_segment (tf, buf, buf + 2);
+		}
 	      cur++;
 	      break;
 	    }
@@ -400,6 +404,7 @@ parse_transform_expr (const char *expr)
 	cur++;
     }
   add_literal_segment (tf, beg, cur);
+  free(str);
 
   return p;
 }
@@ -547,7 +552,7 @@ _single_transform_name_to_obstack (struct transform *tf, char *input)
 			default:
 			  break;
 			}
-		      /*FALL THROUGH*/
+		      FALLTHROUGH;
 
 		    case ctl_upcase:
 		    case ctl_locase:

src/unlink.c

@@ -1,6 +1,6 @@
 /* Unlink files.
 
-   Copyright 2009, 2013-2014 Free Software Foundation, Inc.
+   Copyright 2009-2021 Free Software Foundation, Inc.
 
    This file is part of GNU tar.
 
@@ -32,6 +32,10 @@ struct deferred_unlink
 				       entry got added to the queue */
   };
 
+#define IS_CWD(p) \
+  ((p)->is_dir \
+   && ((p)->file_name[0] == 0 || strcmp ((p)->file_name, ".") == 0))
+
 /* The unlink queue */
 static struct deferred_unlink *dunlink_head, *dunlink_tail;
 
@@ -60,6 +64,24 @@ dunlink_alloc (void)
   return p;
 }
 
+static void
+dunlink_insert (struct deferred_unlink *anchor, struct deferred_unlink *p)
+{
+  if (anchor)
+    {
+      p->next = anchor->next;
+      anchor->next = p;
+    }
+  else 
+    {
+      p->next = dunlink_head;
+      dunlink_head = p;
+    }
+  if (!p->next)
+    dunlink_tail = p;
+  dunlink_count++;
+}
+
 static void
 dunlink_reclaim (struct deferred_unlink *p)
 {
@@ -73,7 +95,7 @@ flush_deferred_unlinks (bool force)
 {
   struct deferred_unlink *p, *prev = NULL;
   int saved_chdir = chdir_current;
-
+  
   for (p = dunlink_head; p; )
     {
       struct deferred_unlink *next = p->next;
@@ -86,12 +108,11 @@ flush_deferred_unlinks (bool force)
 	    {
 	      const char *fname;
 
-	      if (p->dir_idx
-		  && (p->file_name[0] == 0
-		      || strcmp (p->file_name, ".") == 0))
+	      if (p->dir_idx && IS_CWD (p))
 		{
-		  fname = tar_dirname ();
-		  chdir_do (p->dir_idx - 1);
+		  prev = p;
+		  p = next;
+		  continue;
 		}
 	      else
 		fname = p->file_name;
@@ -103,16 +124,18 @@ flush_deferred_unlinks (bool force)
 		    case ENOENT:
 		      /* nothing to worry about */
 		      break;
+		    case EEXIST:
+		      /* OpenSolaris >=10 sets EEXIST instead of ENOTEMPTY
+			 if trying to remove a non-empty directory */
+#if defined ENOTEMPTY && ENOTEMPTY != EEXIST
 		    case ENOTEMPTY:
-		      if (!force)
-			{
-			  /* Keep the record in list, in the hope we'll
-			     be able to remove it later */
-			  prev = p;
-			  p = next;
-			  continue;
-			}
-		      /* fall through */
+#endif
+		      /* Keep the record in list, in the hope we'll
+			 be able to remove it later */
+		      prev = p;
+		      p = next;
+		      continue;
+
 		    default:
 		      rmdir_error (fname);
 		    }
@@ -139,6 +162,34 @@ flush_deferred_unlinks (bool force)
     }
   if (!dunlink_head)
     dunlink_tail = NULL;
+  else if (force)
+    {
+      for (p = dunlink_head; p; )
+	{
+	  struct deferred_unlink *next = p->next;
+	  const char *fname;
+
+	  chdir_do (p->dir_idx);
+	  if (p->dir_idx && IS_CWD (p))
+	    {
+	      fname = tar_dirname ();
+	      chdir_do (p->dir_idx - 1);
+	    }
+	  else
+	    fname = p->file_name;
+
+	  if (unlinkat (chdir_fd, fname, AT_REMOVEDIR) != 0)
+	    {
+	      if (errno != ENOENT)
+		rmdir_error (fname);
+	    }
+	  dunlink_reclaim (p);
+	  dunlink_count--;
+	  p = next;
+	}
+      dunlink_head = dunlink_tail = NULL;
+    }	  
+	    
   chdir_do (saved_chdir);
 }
 
@@ -146,6 +197,7 @@ void
 finish_deferred_unlinks (void)
 {
   flush_deferred_unlinks (true);
+  
   while (dunlink_avail)
     {
       struct deferred_unlink *next = dunlink_avail->next;
@@ -171,10 +223,17 @@ queue_deferred_unlink (const char *name, bool is_dir)
   p->is_dir = is_dir;
   p->records_written = records_written;
 
-  if (dunlink_tail)
-    dunlink_tail->next = p;
+  if (IS_CWD (p))
+    {
+      struct deferred_unlink *q, *prev;
+      for (q = dunlink_head, prev = NULL; q; prev = q, q = q->next)
+	if (IS_CWD (q) && q->dir_idx < p->dir_idx)
+	  break;
+      if (q)
+	dunlink_insert (prev, p);
+      else
+	dunlink_insert (dunlink_tail, p);
+    }
   else
-    dunlink_head = p;
-  dunlink_tail = p;
-  dunlink_count++;
+    dunlink_insert (dunlink_tail, p);
 }

src/update.c

@@ -1,7 +1,6 @@
 /* Update a tar archive.
 
-   Copyright 1988, 1992, 1994, 1996-1997, 1999-2001, 2003-2005, 2007,
-   2010, 2013-2014 Free Software Foundation, Inc.
+   Copyright 1988-2021 Free Software Foundation, Inc.
 
    This file is part of GNU tar.
 
@@ -111,7 +110,7 @@ update_archive (void)
 
   name_gather ();
   open_archive (ACCESS_UPDATE);
-  buffer_write_global_xheader ();
+  xheader_forbid_global ();
 
   while (!found_end)
     {
@@ -152,7 +151,7 @@ update_archive (void)
 
 			    for (p = dirp; *p; p += strlen (p) + 1)
 			      addname (namebuf_name (nbuf, p),
-				       0, false, NULL);
+				       name->change_dir, false, NULL);
 
 			    namebuf_free (nbuf);
 			    free (dirp);
@@ -186,13 +185,11 @@ update_archive (void)
 	    {
 	    case HEADER_STILL_UNREAD:
 	      WARN ((0, 0, _("This does not look like a tar archive")));
-	      /* Fall through.  */
-
+	      FALLTHROUGH;
 	    case HEADER_SUCCESS:
 	    case HEADER_ZERO_BLOCK:
 	      ERROR ((0, 0, _("Skipping to next header")));
-	      /* Fall through.  */
-
+	      FALLTHROUGH;
 	    case HEADER_FAILURE:
 	      break;
 

src/utf8.c

@@ -1,6 +1,6 @@
 /* Charset handling for GNU tar.
 
-   Copyright 2004, 2006-2007, 2013-2014 Free Software Foundation, Inc.
+   Copyright 2004-2021 Free Software Foundation, Inc.
 
    This file is part of GNU tar.
 
@@ -35,11 +35,14 @@
 # define iconv_open(tocode, fromcode) ((iconv_t) -1)
 
 # undef iconv
-# define iconv(cd, inbuf, inbytesleft, outbuf, outbytesleft) ((size_t) 0)
+# define iconv(cd, inbuf, inbytesleft, outbuf, outbytesleft) (errno = ENOSYS, (size_t) -1)
 
 # undef iconv_close
 # define iconv_close(cd) 0
 
+# undef iconv_t
+# define iconv_t int
+
 #endif
 
 
@@ -64,10 +67,9 @@ bool
 utf8_convert (bool to_utf, char const *input, char **output)
 {
   char ICONV_CONST *ib;
-  char *ob;
+  char *ob, *ret;
   size_t inlen;
   size_t outlen;
-  size_t rc;
   iconv_t cd = utf8_init (to_utf);
 
   if (cd == 0)
@@ -80,11 +82,27 @@ utf8_convert (bool to_utf, char const *input, char **output)
 
   inlen = strlen (input) + 1;
   outlen = inlen * MB_LEN_MAX + 1;
-  ob = *output = xmalloc (outlen);
+  ob = ret = xmalloc (outlen);
   ib = (char ICONV_CONST *) input;
-  rc = iconv (cd, &ib, &inlen, &ob, &outlen);
+  /* According to POSIX, "if iconv() encounters a character in the input
+     buffer that is valid, but for which an identical character does not
+     exist in the target codeset, iconv() shall perform an
+     implementation-defined conversion on this character." It will "update
+     the variables pointed to by the arguments to reflect the extent of the
+     conversion and return the number of non-identical conversions performed".
+     On error, it returns -1. 
+     In other words, non-zero return always indicates failure, either because
+     the input was not fully converted, or because it was converted in a
+     non-reversible way.
+   */
+  if (iconv (cd, &ib, &inlen, &ob, &outlen) != 0)
+    {
+      free (ret);
+      return false;
+    }
   *ob = 0;
-  return rc != -1;
+  *output = ret;
+  return true;
 }
 
 

src/warning.c

@@ -1,6 +1,6 @@
 /* Warnings for GNU tar.
 
-   Copyright 2009, 2012-2014 Free Software Foundation, Inc.
+   Copyright 2009-2021 Free Software Foundation, Inc.
 
    This file is part of GNU tar.
 
@@ -47,6 +47,7 @@ static char const *const warning_args[] = {
   "existing-file",
   "xattr-write",
   "record-size",
+  "failed-read",
   NULL
 };
 
@@ -74,7 +75,8 @@ static int warning_types[] = {
   WARN_DECOMPRESS_PROGRAM,
   WARN_EXISTING_FILE,
   WARN_XATTR_WRITE,
-  WARN_RECORD_SIZE
+  WARN_RECORD_SIZE,
+  WARN_FAILED_READ
 };
 
 ARGMATCH_VERIFY (warning_args, warning_types);

src/xattrs.c

@@ -1,6 +1,6 @@
 /* Support for extended attributes.
 
-   Copyright (C) 2006-2014 Free Software Foundation, Inc.
+   Copyright (C) 2006-2021 Free Software Foundation, Inc.
 
    This file is part of GNU tar.
 
@@ -246,13 +246,36 @@ xattrs__acls_set (struct tar_stat_info const *st,
   acl_free (acl);
 }
 
+/* Cleanup textual representation of the ACL in VAL by eliminating tab
+   characters and comments */
+static void
+xattrs_acls_cleanup (char *val, size_t *plen)
+{
+  char *p, *q;
+
+  p = q = val + strcspn (val, "#\t");
+  while (*q)
+    {
+      if (*q == '\t')
+	q++;
+      else if (*q == '#')
+	{
+	  while (*q != '\n')
+	    q++;
+	}
+      else
+	*p++ = *q++;
+    }
+  *plen = p - val;
+  *p++ = 0;
+}
+
 static void
 xattrs__acls_get_a (int parentfd, const char *file_name,
                     struct tar_stat_info *st,
                     char **ret_ptr, size_t * ret_len)
 {
   char *val = NULL;
-  ssize_t len;
   acl_t acl;
 
   if (!(acl = acl_get_file_at (parentfd, file_name, ACL_TYPE_ACCESS)))
@@ -262,7 +285,7 @@ xattrs__acls_get_a (int parentfd, const char *file_name,
       return;
     }
 
-  val = acl_to_text (acl, &len);
+  val = acl_to_text (acl, NULL);
   acl_free (acl);
 
   if (!val)
@@ -272,8 +295,7 @@ xattrs__acls_get_a (int parentfd, const char *file_name,
     }
 
   *ret_ptr = xstrdup (val);
-  *ret_len = len;
-
+  xattrs_acls_cleanup (*ret_ptr, ret_len);
   acl_free (val);
 }
 
@@ -284,7 +306,6 @@ xattrs__acls_get_d (int parentfd, char const *file_name,
                     char **ret_ptr, size_t * ret_len)
 {
   char *val = NULL;
-  ssize_t len;
   acl_t acl;
 
   if (!(acl = acl_get_file_at (parentfd, file_name, ACL_TYPE_DEFAULT)))
@@ -294,7 +315,7 @@ xattrs__acls_get_d (int parentfd, char const *file_name,
       return;
     }
 
-  val = acl_to_text (acl, &len);
+  val = acl_to_text (acl, NULL);
   acl_free (acl);
 
   if (!val)
@@ -304,8 +325,7 @@ xattrs__acls_get_d (int parentfd, char const *file_name,
     }
 
   *ret_ptr = xstrdup (val);
-  *ret_len = len;
-
+  xattrs_acls_cleanup (*ret_ptr, ret_len);
   acl_free (val);
 }
 #endif /* HAVE_POSIX_ACLS */
@@ -336,6 +356,7 @@ acls_one_line (const char *prefix, char delim,
       obstack_grow (&stk, prefix, pref_len);
       obstack_grow (&stk, aclstring, move);
 
+      pos += move + 1;
       aclstring += move + 1;
     }
 
@@ -434,8 +455,12 @@ xattrs_clear_setup (void)
   clear_mask_map (&xattrs_setup.excl);
 }
 
-/* get all xattrs from file given by FILE_NAME or FD (when non-zero).  This
-   includes all the user.*, security.*, system.*, etc. available domains */
+static bool xattrs_masked_out (const char *kw, bool archiving);
+
+/* get xattrs from file given by FILE_NAME or FD (when non-zero)
+   xattrs are checked against the user supplied include/exclude mask
+   if no mask is given this includes all the user.*, security.*, system.*,
+   etc. available domains */
 void
 xattrs_xattrs_get (int parentfd, char const *file_name,
                    struct tar_stat_info *st, int fd)
@@ -480,8 +505,6 @@ xattrs_xattrs_get (int parentfd, char const *file_name,
               size_t len = strlen (attr);
               ssize_t aret = 0;
 
-              /* Archive all xattrs during creation, decide at extraction time
-               * which ones are of interest/use for the target filesystem. */
               while (((fd == 0)
                       ? ((aret = lgetxattrat (parentfd, file_name, attr,
                                               val, asz)) == -1)
@@ -492,7 +515,10 @@ xattrs_xattrs_get (int parentfd, char const *file_name,
                 }
 
               if (aret != -1)
-                xheader_xattr_add (st, attr, val, aret);
+                {
+                  if (!xattrs_masked_out (attr, true))
+                    xheader_xattr_add (st, attr, val, aret);
+                }
               else if (errno != ENOATTR)
                 call_arg_warn ((fd == 0) ? "lgetxattrat"
                                : "fgetxattr", file_name);
@@ -735,6 +761,8 @@ xattrs_print (struct tar_stat_info const *st)
     {
       fprintf (stdlis, "  a: ");
       acls_one_line ("", ',', st->acls_a_ptr, st->acls_a_len);
+      if (st->acls_a_len && st->acls_d_len)
+	fprintf (stdlis, ",");
       acls_one_line ("default:", ',', st->acls_d_ptr, st->acls_d_len);
       fprintf (stdlis, "\n");
     }

src/xattrs.h

@@ -1,6 +1,6 @@
 /* Support for extended attributes.
 
-   Copyright (C) 2006-2014 Free Software Foundation, Inc.
+   Copyright (C) 2006-2021 Free Software Foundation, Inc.
 
    This file is part of GNU tar.
 

src/xheader.c

@@ -1,7 +1,6 @@
 /* POSIX extended headers for tar.
 
-   Copyright (C) 2003-2007, 2009-2010, 2012-2014 Free Software
-   Foundation, Inc.
+   Copyright (C) 2003-2021 Free Software Foundation, Inc.
 
    This file is part of GNU tar.
 
@@ -185,6 +184,9 @@ xheader_set_keyword_equal (char *kw, char *eq)
   bool global = true;
   char *p = eq;
 
+  if (eq == kw)
+    USAGE_ERROR ((0, 0, _("Malformed pax option: %s"), quote (kw)));
+    
   if (eq[-1] == ':')
     {
       p--;
@@ -255,7 +257,7 @@ char *
 xheader_format_name (struct tar_stat_info *st, const char *fmt, size_t n)
 {
   char *buf;
-  size_t len = strlen (fmt);
+  size_t len;
   char *q;
   const char *p;
   char *dirp = NULL;
@@ -266,43 +268,51 @@ xheader_format_name (struct tar_stat_info *st, const char *fmt, size_t n)
   char nbuf[UINTMAX_STRSIZE_BOUND];
   char const *nptr = NULL;
 
-  for (p = fmt; *p && (p = strchr (p, '%')); )
+  len = 0;
+  for (p = fmt; *p; p++)
     {
-      switch (p[1])
+      if (*p == '%' && p[1])
 	{
-	case '%':
-	  len--;
-	  break;
-
-	case 'd':
-	  if (st)
+	  switch (*++p)
 	    {
-	      if (!dirp)
-		dirp = dir_name (st->orig_file_name);
-	      dir = safer_name_suffix (dirp, false, absolute_names_option);
-	      len += strlen (dir) - 2;
+	    case '%':
+	      len++;
+	      break;
+
+	    case 'd':
+	      if (st)
+		{
+		  if (!dirp)
+		    dirp = dir_name (st->orig_file_name);
+		  dir = safer_name_suffix (dirp, false, absolute_names_option);
+		  len += strlen (dir);
+		}
+	      break;
+
+	    case 'f':
+	      if (st)
+		{
+		  base = last_component (st->orig_file_name);
+		  len += strlen (base);
+		}
+	      break;
+
+	    case 'p':
+	      pptr = umaxtostr (getpid (), pidbuf);
+	      len += pidbuf + sizeof pidbuf - 1 - pptr;
+	      break;
+
+	    case 'n':
+	      nptr = umaxtostr (n, nbuf);
+	      len += nbuf + sizeof nbuf - 1 - nptr;
+	      break;
+	      
+	    default:
+	      len += 2;
 	    }
-	  break;
-
-	case 'f':
-	  if (st)
-	    {
-	      base = last_component (st->orig_file_name);
-	      len += strlen (base) - 2;
-	    }
-	  break;
-
-	case 'p':
-	  pptr = umaxtostr (getpid (), pidbuf);
-	  len += pidbuf + sizeof pidbuf - 1 - pptr - 2;
-	  break;
-
-	case 'n':
-	  nptr = umaxtostr (n, nbuf);
-	  len += nbuf + sizeof nbuf - 1 - nptr - 2;
-	  break;
 	}
-      p++;
+      else
+	len++;
     }
 
   buf = xmalloc (len + 1);
@@ -359,29 +369,46 @@ xheader_format_name (struct tar_stat_info *st, const char *fmt, size_t n)
   return buf;
 }
 
+/* Table of templates for the names of POSIX extended headers.
+   Indexed by the the type of the header (per-file or global)
+   and POSIX compliance mode (0 or q depending on whether
+   POSIXLY_CORRECT environment variable is set. */
+static const char *header_template[][2] = {
+  /* Individual header templates: */
+  { "%d/PaxHeaders/%f", "%d/PaxHeaders.%p/%f" },
+  /* Global header templates: */
+  { "/GlobalHead.%n", "/GlobalHead.%p.%n" }
+};
+/* Indices to the above table */
+enum {
+  pax_file_header,
+  pax_global_header
+};
+/* Return the name for the POSIX extended header T */
+#define HEADER_TEMPLATE(t) header_template[t][posixly_correct]
+
 char *
 xheader_xhdr_name (struct tar_stat_info *st)
 {
   if (!exthdr_name)
-    assign_string (&exthdr_name, "%d/PaxHeaders.%p/%f");
+    assign_string (&exthdr_name, HEADER_TEMPLATE (pax_file_header));
   return xheader_format_name (st, exthdr_name, 0);
 }
 
-#define GLOBAL_HEADER_TEMPLATE "/GlobalHead.%p.%n"
-
 char *
 xheader_ghdr_name (void)
 {
   if (!globexthdr_name)
     {
       size_t len;
+      const char *global_header_template = HEADER_TEMPLATE (pax_global_header);
       const char *tmp = getenv ("TMPDIR");
       if (!tmp)
 	tmp = "/tmp";
-      len = strlen (tmp) + sizeof (GLOBAL_HEADER_TEMPLATE); /* Includes nul */
+      len = strlen (tmp) + strlen (global_header_template) + 1;
       globexthdr_name = xmalloc (len);
       strcpy(globexthdr_name, tmp);
-      strcat(globexthdr_name, GLOBAL_HEADER_TEMPLATE);
+      strcat(globexthdr_name, global_header_template);
     }
 
   return xheader_format_name (NULL, globexthdr_name, global_header_count + 1);
@@ -458,6 +485,14 @@ xheader_write_global (struct xheader *xhdr)
     }
 }
 
+/* Forbid modifications of the global extended header */
+void
+xheader_forbid_global (void)
+{
+  if (keyword_global_override_list)
+    USAGE_ERROR ((0, 0, _("can't update global extended header record")));
+}
+
 void
 xheader_xattr_init (struct tar_stat_info *st)
 {
@@ -755,6 +790,16 @@ xheader_decode (struct tar_stat_info *st)
 	continue;
     }
   run_override_list (keyword_override_list, st);
+
+  /* The archived (effective) file size is always set directly in tar header
+     field, possibly overridden by "size" extended header - in both cases,
+     result is now decoded in st->stat.st_size */
+  st->archive_file_size = st->stat.st_size;
+
+  /* The real file size (given by stat()) may be redefined for sparse
+     files in "GNU.sparse.realsize" extended header */
+  if (st->real_size_set)
+    st->stat.st_size = st->real_size;
 }
 
 static void
@@ -803,11 +848,11 @@ xheader_store (char const *keyword, struct tar_stat_info *st,
   t = locate_handler (keyword);
   if (!t || !t->coder)
     return;
-  if (xheader_keyword_deleted_p (keyword)
-      || xheader_keyword_override_p (keyword))
+  if (xheader_keyword_deleted_p (keyword))
     return;
   xheader_init (&st->xhdr);
-  t->coder (st, keyword, &st->xhdr, data);
+  if (!xheader_keyword_override_p (keyword))
+    t->coder (st, keyword, &st->xhdr, data);
 }
 
 void
@@ -1017,7 +1062,7 @@ xheader_string_end (struct xheader *xhdr, char const *keyword)
     }
   x_obstack_blank (xhdr, p);
   x_obstack_1grow (xhdr, '\n');
-  cp = obstack_next_free (xhdr->stk) - xhdr->string_length - p - 1;
+  cp = (char*) obstack_next_free (xhdr->stk) - xhdr->string_length - p - 1;
   memmove (cp + p, cp, xhdr->string_length);
   cp = stpcpy (cp, np);
   *cp++ = ' ';
@@ -1280,15 +1325,33 @@ path_coder (struct tar_stat_info const *st, char const *keyword,
   code_string (st->file_name, keyword, xhdr);
 }
 
+static void
+raw_path_decoder (struct tar_stat_info *st, char const *arg)
+{
+  decode_string (&st->orig_file_name, arg);
+  decode_string (&st->file_name, arg);
+  st->had_trailing_slash = strip_trailing_slashes (st->file_name);
+}
+
+
 static void
 path_decoder (struct tar_stat_info *st,
 	      char const *keyword __attribute__((unused)),
 	      char const *arg,
 	      size_t size __attribute__((unused)))
 {
-  decode_string (&st->orig_file_name, arg);
-  decode_string (&st->file_name, arg);
-  st->had_trailing_slash = strip_trailing_slashes (st->file_name);
+  if (! st->sparse_name_done)
+    raw_path_decoder (st, arg);
+}
+
+static void
+sparse_path_decoder (struct tar_stat_info *st,
+                     char const *keyword __attribute__((unused)),
+                     char const *arg,
+                     size_t size __attribute__((unused)))
+{
+  st->sparse_name_done = true;
+  raw_path_decoder (st, arg);
 }
 
 static void
@@ -1360,7 +1423,10 @@ sparse_size_decoder (struct tar_stat_info *st,
 {
   uintmax_t u;
   if (decode_num (&u, arg, TYPE_MAXIMUM (off_t), keyword))
-    st->stat.st_size = u;
+    {
+      st->real_size_set = true;
+      st->real_size = u;
+    }
 }
 
 static void
@@ -1443,13 +1509,13 @@ sparse_map_decoder (struct tar_stat_info *st,
 		    size_t size __attribute__((unused)))
 {
   int offset = 1;
+  struct sp_array e;
 
   st->sparse_map_avail = 0;
   while (1)
     {
       intmax_t u;
       char *delim;
-      struct sp_array e;
 
       if (!ISDIGIT (*arg))
 	{
@@ -1717,7 +1783,7 @@ struct xhdr_tab const xhdr_tab[] = {
   { "uname",    uname_coder,    uname_decoder,    0, false },
 
   /* Sparse file handling */
-  { "GNU.sparse.name",       path_coder, path_decoder,
+  { "GNU.sparse.name",       path_coder, sparse_path_decoder,
     XHDR_PROTECTED, false },
   { "GNU.sparse.major",      sparse_major_coder, sparse_major_decoder,
     XHDR_PROTECTED, false },

tests/.gitignore

@@ -8,3 +8,8 @@ argcv.c
 argcv.h
 genfile.c
 genfile
+download
+ttyemu
+checkseekhole
+ckmtime
+/compress-*.at

tests/Makefile.am

@@ -1,6 +1,6 @@
 # Makefile for GNU tar regression tests.
 
-# Copyright 1996-1997, 1999-2001, 2003-2007, 2009, 2012-2013 Free Software
+# Copyright 1996-2021 Free Software Foundation, Inc.
 
 # This file is part of GNU tar.
 
@@ -17,9 +17,13 @@
 # You should have received a copy of the GNU General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 
-EXTRA_DIST = $(TESTSUITE_AT) testsuite package.m4 star/README star/quicktest.sh
+EXTRA_DIST = $(TESTSUITE_AT) \
+  testsuite package.m4 star/README star/quicktest.sh \
+  compress.m4
+
 DISTCLEANFILES       = atconfig $(check_SCRIPTS)
 MAINTAINERCLEANFILES = Makefile.in $(TESTSUITE)
+CLEANFILES =
 
 ## ------------ ##
 ## package.m4.  ##
@@ -42,28 +46,42 @@ $(srcdir)/package.m4: $(top_srcdir)/configure.ac
 ## ------------ ##
 
 TESTSUITE_AT = \
+ testsuite.at\
+ compress.m4\
  T-cd.at\
+ T-dir00.at\
+ T-dir01.at\
  T-empty.at\
  T-null.at\
+ T-null2.at\
  T-rec.at\
+ T-recurse.at\
  T-zfile.at\
  T-nonl.at\
  T-mult.at\
  T-nest.at\
- testsuite.at\
+ add-file.at\
  append.at\
  append01.at\
  append02.at\
  append03.at\
  append04.at\
+ append05.at\
  backup01.at\
  chtype.at\
  comprec.at\
+ comperr.at\
+ checkpoint/defaults.at\
+ checkpoint/interval.at\
+ checkpoint/dot.at\
+ checkpoint/dot-compat.at\
+ checkpoint/dot-int.at\
  delete01.at\
  delete02.at\
  delete03.at\
  delete04.at\
  delete05.at\
+ difflink.at\
  exclude.at\
  exclude01.at\
  exclude02.at\
@@ -100,8 +118,15 @@ TESTSUITE_AT = \
  extrac17.at\
  extrac18.at\
  extrac19.at\
+ extrac20.at\
+ extrac21.at\
+ extrac22.at\
+ extrac23.at\
+ extrac24.at\
  filerem01.at\
  filerem02.at\
+ dirrem01.at\
+ dirrem02.at\
  gzip.at\
  grow.at\
  incremental.at\
@@ -114,9 +139,10 @@ TESTSUITE_AT = \
  incr07.at\
  incr08.at\
  incr09.at\
+ incr10.at\
+ incr11.at\
  indexfile.at\
  ignfail.at\
- iotty.at\
  label01.at\
  label02.at\
  label03.at\
@@ -136,6 +162,7 @@ TESTSUITE_AT = \
  lustar01.at\
  lustar02.at\
  lustar03.at\
+ map.at\
  multiv01.at\
  multiv02.at\
  multiv03.at\
@@ -144,19 +171,27 @@ TESTSUITE_AT = \
  multiv06.at\
  multiv07.at\
  multiv08.at\
+ multiv09.at\
+ multiv10.at\
+ numeric.at\
  old.at\
  onetop01.at\
  onetop02.at\
  onetop03.at\
  onetop04.at\
+ onetop05.at\
  opcomp01.at\
  opcomp02.at\
  opcomp03.at\
  opcomp04.at\
  opcomp05.at\
  opcomp06.at\
+ positional01.at\
+ positional02.at\
+ positional03.at\
  options.at\
  options02.at\
+ options03.at\
  owner.at\
  pipe.at\
  recurse.at\
@@ -166,6 +201,7 @@ TESTSUITE_AT = \
  rename03.at\
  rename04.at\
  rename05.at\
+ rename06.at\
  remfiles01.at\
  remfiles02.at\
  remfiles03.at\
@@ -198,16 +234,24 @@ TESTSUITE_AT = \
  sparse02.at\
  sparse03.at\
  sparse04.at\
+ sparse05.at\
+ sparse06.at\
+ sparse07.at\
  sparsemv.at\
  sparsemvp.at\
  spmvp00.at\
  spmvp01.at\
  spmvp10.at\
+ sptrcreat.at\
+ sptrdiff00.at\
+ sptrdiff01.at\
  time01.at\
+ time02.at\
  truncate.at\
  update.at\
  update01.at\
  update02.at\
+ update03.at\
  volsize.at\
  volume.at\
  verbose.at\
@@ -215,6 +259,8 @@ TESTSUITE_AT = \
  version.at\
  xform-h.at\
  xform01.at\
+ xform02.at\
+ xform03.at\
  star/gtarfail.at\
  star/gtarfail2.at\
  star/multi-fail.at\
@@ -226,6 +272,8 @@ TESTSUITE_AT = \
  xattr03.at\
  xattr04.at\
  xattr05.at\
+ xattr06.at\
+ xattr07.at\
  acls01.at\
  acls02.at\
  acls03.at\
@@ -233,6 +281,9 @@ TESTSUITE_AT = \
  selacl01.at\
  capabs_raw01.at
 
+distclean-local:
+	-rm -rf download
+
 TESTSUITE = $(srcdir)/testsuite
 
 AUTOTEST = $(AUTOM4TE) --language=autotest
@@ -255,7 +306,7 @@ check-full:
 #check_SCRIPTS = tar
 
 # Run the test suite on the *installed* tree.
-installcheck-local:
+installcheck-local: $(check_PROGRAMS)
 	$(SHELL) $(TESTSUITE) $(TESTSUITEFLAGS) AUTOTEST_PATH=$(exec_prefix)/bin
 
 
@@ -263,15 +314,10 @@ installcheck-local:
 ## genfile      ##
 ## ------------ ##
 
-check_PROGRAMS = genfile
-
-if TAR_COND_GRANTPT
-check_PROGRAMS += ttyemu
-endif
+check_PROGRAMS = genfile checkseekhole ckmtime
 
 genfile_SOURCES = genfile.c argcv.c argcv.h
-
-ttyemu_SOURCES = ttyemu.c
+checkseekhole_SOURCES = checkseekhole.c
 
 localedir = $(datadir)/locale
 AM_CPPFLAGS = \

tests/T-cd.at

@@ -1,7 +1,7 @@
 # Process this file with autom4te to create testsuite. -*- Autotest -*-
 #
 # Test suite for GNU tar.
-# Copyright 2013-2014 Free Software Foundation, Inc.
+# Copyright 2013-2021 Free Software Foundation, Inc.
 #
 # This file is part of GNU tar.
 #
@@ -19,7 +19,7 @@
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 
 AT_SETUP([-C in file lists])
-AT_KEYWORDS([files-from T-cd])
+AT_KEYWORDS([files-from T-cd chdir])
 
 AT_TAR_CHECK([
 AT_SORT_PREREQ

tests/T-dir00.at

@@ -0,0 +1,46 @@
+# Process this file with autom4te to create testsuite. -*- Autotest -*-
+#
+# Test suite for GNU tar.
+# Copyright 2014-2021 Free Software Foundation, Inc.
+
+# This file is part of GNU tar.
+
+# GNU tar is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 3 of the License, or
+# (at your option) any later version.
+
+# GNU tar is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+
+# You should have received a copy of the GNU General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+# Tar 1.27 and 1.28 did not extract files under directory members listed
+# in the file read by --file-from.
+#
+# Reported-by: Jean-Louis Martineau <martineau@zmanda.com>
+# References: <541AE02C.2050008@zmanda.com>,
+#             http://lists.gnu.org/archive/html/bug-tar/2014-09/msg00006.html
+
+AT_SETUP([recursive extraction from --files-from])
+AT_KEYWORDS([files-from extract T-dir T-dir00])
+AT_TAR_CHECK([
+AT_SORT_PREREQ
+mkdir dir
+genfile -f dir/file1
+genfile -f dir/file2
+tar cf archive dir
+rm -rf dir
+echo dir > list
+tar xfTv archive list | sort
+],
+[0],
+[dir/
+dir/file1
+dir/file2
+])
+AT_CLEANUP
+

tests/T-dir01.at

@@ -0,0 +1,46 @@
+# Process this file with autom4te to create testsuite. -*- Autotest -*-
+#
+# Test suite for GNU tar.
+# Copyright 2014-2021 Free Software Foundation, Inc.
+
+# This file is part of GNU tar.
+
+# GNU tar is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 3 of the License, or
+# (at your option) any later version.
+
+# GNU tar is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+
+# You should have received a copy of the GNU General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+# Tar 1.27 and 1.28 did not remove trailing slashes from file names
+# obtained with the --file-from option.
+#
+# Reported-by: Jean-Louis Martineau <martineau@zmanda.com>
+# References: <541AE02C.2050008@zmanda.com>,
+#             http://lists.gnu.org/archive/html/bug-tar/2014-09/msg00006.html
+
+AT_SETUP([trailing slash in --files-from])
+AT_KEYWORDS([files-from extract T-dir T-dir01])
+AT_TAR_CHECK([
+AT_SORT_PREREQ
+mkdir dir
+genfile -f dir/file1
+genfile -f dir/file2
+tar cf archive dir
+rm -rf dir
+echo dir/ > list
+tar xfTv archive list | sort
+],
+[0],
+[dir/
+dir/file1
+dir/file2
+])
+AT_CLEANUP
+

tests/T-empty.at

@@ -1,7 +1,7 @@
 # Process this file with autom4te to create testsuite. -*- Autotest -*-
 
 # Test suite for GNU tar.
-# Copyright 2006-2007, 2013-2014 Free Software Foundation, Inc.
+# Copyright 2006-2021 Free Software Foundation, Inc.
 
 # This file is part of GNU tar.
 

tests/T-mult.at

@@ -1,7 +1,7 @@
 # Process this file with autom4te to create testsuite. -*- Autotest -*-
 #
 # Test suite for GNU tar.
-# Copyright 2013-2014 Free Software Foundation, Inc.
+# Copyright 2013-2021 Free Software Foundation, Inc.
 #
 # This file is part of GNU tar.
 #

tests/T-nest.at

@@ -1,7 +1,7 @@
 # Process this file with autom4te to create testsuite. -*- Autotest -*-
 #
 # Test suite for GNU tar.
-# Copyright 2013-2014 Free Software Foundation, Inc.
+# Copyright 2013-2021 Free Software Foundation, Inc.
 #
 # This file is part of GNU tar.
 #

tests/T-nonl.at

@@ -1,7 +1,7 @@
 # Process this file with autom4te to create testsuite. -*- Autotest -*-
 #
 # Test suite for GNU tar.
-# Copyright 2013-2014 Free Software Foundation, Inc.
+# Copyright 2013-2021 Free Software Foundation, Inc.
 #
 # This file is part of GNU tar.
 #
@@ -29,9 +29,8 @@ AT_KEYWORDS([files-from nonewline nonl T-nonl])
 
 AT_TAR_CHECK([
 genfile --length=0 --file empty
-AS_ECHO_N(c) > 1.nonl
-echo d > 2.nonl
-AS_ECHO_N(e) >> 2.nonl
+printf c > 1.nonl
+printf 'd\ne' > 2.nonl
 touch a b c d e
 AT_DATA([filelist],[a
 b
@@ -59,4 +58,3 @@ c
 [],[],[],[ustar])
 
 AT_CLEANUP
-

tests/T-null.at

@@ -1,7 +1,7 @@
 # Process this file with autom4te to create testsuite. -*- Autotest -*-
 
 # Test suite for GNU tar.
-# Copyright 2006-2007, 2013-2014 Free Software Foundation, Inc.
+# Copyright 2006-2021 Free Software Foundation, Inc.
 
 # This file is part of GNU tar.
 

tests/T-null2.at

@@ -0,0 +1,46 @@
+# This file is part of test suite for GNU tar. -*- Autotest -*-
+# Copyright 2015-2021 Free Software Foundation, Inc.
+#
+# GNU tar is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 3 of the License, or
+# (at your option) any later version.
+#
+# GNU tar is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+AT_SETUP([--null enables verbatim reading])
+AT_KEYWORDS([files-from null T-null2 T-verbatim])
+
+# According to the docs, --null should read each line from the file
+# list verbatim.  This feature was broken by commit 26538c9b (tar version
+# 1.27).
+
+AT_TAR_CHECK([
+AT_DATA([file-list],[a
+-b
+--c d
+:\\.jpg
+])
+
+genfile -f a
+genfile -f -b
+genfile -f '--c d'
+genfile -f ':\\.jpg'
+
+cat file-list | tr '\n' '\0' | tar -c -f archive -v --null -T -
+],
+[0],
+[a
+-b
+--c d
+:\\\\.jpg
+],
+[],[],[],[ustar]) # Testing one format is enough
+
+AT_CLEANUP

tests/T-rec.at

@@ -1,7 +1,7 @@
 # Process this file with autom4te to create testsuite. -*- Autotest -*-
 #
 # Test suite for GNU tar.
-# Copyright 2013-2014 Free Software Foundation, Inc.
+# Copyright 2013-2021 Free Software Foundation, Inc.
 #
 # This file is part of GNU tar.
 #

tests/T-recurse.at

@@ -0,0 +1,90 @@
+# Process this file with autom4te to create testsuite. -*- Autotest -*-
+
+# Test suite for GNU tar.
+# Copyright 2015-2021 Free Software Foundation, Inc.
+
+# This file is part of GNU tar.
+
+# GNU tar is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 3 of the License, or
+# (at your option) any later version.
+
+# GNU tar is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+
+# You should have received a copy of the GNU General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+# Description: Test interaction of --recursion and --no-recursion options
+# together with --files-from option.  This is complementary to recurs02.at test
+# case. References:
+# <alpine.LSU.2.11.1502201029580.29773@nerf60.vanv.qr>
+# http://lists.gnu.org/archive/html/bug-tar/2015-06/msg00006.html
+
+AT_SETUP([files-from & recurse: toggle])
+AT_KEYWORDS([recurse T-recurse files-from])
+
+AT_TAR_CHECK([
+mkdir directory1 directory2
+touch directory1/file directory2/file
+
+AT_DATA([F1],[--no-recursion
+directory1/
+--recursion
+directory2/
+])
+
+AT_DATA([F2A],[directory1/
+])
+
+AT_DATA([F2B],[directory2/
+])
+
+a=archive
+tar cf "$a" --files-from F1
+tar tf "$a"
+
+a=archive2
+tar cf "$a" --no-recursion -T F2A --recursion -T F2B
+tar tf "$a"
+],
+[0],
+[directory1/
+directory2/
+directory2/file
+directory1/
+directory2/
+directory2/file
+])
+
+AT_CLEANUP
+
+
+AT_SETUP([toggle --recursion (not) from -T])
+AT_KEYWORDS([recurse T-recurse T-recurse2 files-from])
+
+AT_TAR_CHECK([
+mkdir directory1 directory2
+touch directory1/file directory2/file
+
+AT_DATA([F1],[--no-recursion
+directory1/
+])
+
+AT_DATA([F2],[directory2/
+])
+
+tar cf archive -T F1 --recursion -T F2
+tar tf archive
+
+],
+[0],
+[directory1/
+directory2/
+directory2/file
+])
+
+AT_CLEANUP

tests/T-zfile.at

@@ -1,7 +1,7 @@
 # Process this file with autom4te to create testsuite. -*- Autotest -*-
 #
 # Test suite for GNU tar.
-# Copyright 2013-2014 Free Software Foundation, Inc.
+# Copyright 2013-2021 Free Software Foundation, Inc.
 #
 # This file is part of GNU tar.
 #