Merging upstream version 0.23.

Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>

3 files changed, 750 insertions, 519 deletions

diff --git a/doc/tarlz.1 b/doc/tarlz.1
index d7cc093..d23b164 100644
--- a/doc/tarlz.1
+++ b/doc/tarlz.1
@@ -1,10 +1,10 @@
 .\" DO NOT MODIFY THIS FILE!  It was generated by help2man 1.47.16.
-.TH TARLZ "1" "January 2022" "tarlz 0.22" "User Commands"
+.TH TARLZ "1" "September 2022" "tarlz 0.23" "User Commands"
 .SH NAME
 tarlz \- creates tar archives with multimember lzip compression
 .SH SYNOPSIS
 .B tarlz
-[\fI\,options\/\fR] [\fI\,files\/\fR]
+\fI\,operation \/\fR[\fI\,options\/\fR] [\fI\,files\/\fR]
 .SH DESCRIPTION
 Tarlz is a massively parallel (multi\-threaded) combined implementation of
 the tar archiver and the lzip compressor. Tarlz uses the compression library
@@ -28,7 +28,7 @@ archive, skipping over the damaged members, just like the standard
 (uncompressed) tar. Moreover, the option '\-\-keep\-damaged' can be used to
 recover as much data as possible from each damaged member, and lziprecover
 can be used to recover some of the damaged members.
-.SH OPTIONS
+.SS "Operations:"
 .TP
 \fB\-\-help\fR
 display this help and exit
@@ -39,26 +39,36 @@ output version information and exit
 \fB\-A\fR, \fB\-\-concatenate\fR
 append archives to the end of an archive
 .TP
-\fB\-B\fR, \fB\-\-data\-size=\fR<bytes>
-set target size of input data blocks [2x8=16 MiB]
-.TP
 \fB\-c\fR, \fB\-\-create\fR
 create a new archive
 .TP
-\fB\-C\fR, \fB\-\-directory=\fR<dir>
-change to directory <dir>
-.TP
 \fB\-d\fR, \fB\-\-diff\fR
 find differences between archive and file system
 .TP
-\fB\-\-ignore\-ids\fR
-ignore differences in owner and group IDs
-.TP
 \fB\-\-delete\fR
 delete files/directories from an archive
 .TP
-\fB\-\-exclude=\fR<pattern>
-exclude files matching a shell pattern
+\fB\-r\fR, \fB\-\-append\fR
+append files to the end of an archive
+.TP
+\fB\-t\fR, \fB\-\-list\fR
+list the contents of an archive
+.TP
+\fB\-x\fR, \fB\-\-extract\fR
+extract files/directories from an archive
+.TP
+\fB\-z\fR, \fB\-\-compress\fR
+compress existing POSIX tar archives
+.TP
+\fB\-\-check\-lib\fR
+check version of lzlib and exit
+.SH OPTIONS
+.TP
+\fB\-B\fR, \fB\-\-data\-size=\fR<bytes>
+set target size of input data blocks [2x8=16 MiB]
+.TP
+\fB\-C\fR, \fB\-\-directory=\fR<dir>
+change to directory <dir>
 .TP
 \fB\-f\fR, \fB\-\-file=\fR<archive>
 use archive file <archive>
@@ -66,9 +76,6 @@ use archive file <archive>
 \fB\-h\fR, \fB\-\-dereference\fR
 follow symlinks; archive the files they point to
 .TP
-\fB\-\-mtime=\fR<date>
-use <date> as mtime for files added to archive
-.TP
 \fB\-n\fR, \fB\-\-threads=\fR<n>
 set number of (de)compression threads [2]
 .TP
@@ -81,21 +88,9 @@ don't subtract the umask on extraction
 \fB\-q\fR, \fB\-\-quiet\fR
 suppress all messages
 .TP
-\fB\-r\fR, \fB\-\-append\fR
-append files to the end of an archive
-.TP
-\fB\-t\fR, \fB\-\-list\fR
-list the contents of an archive
-.TP
 \fB\-v\fR, \fB\-\-verbose\fR
 verbosely list files processed
 .TP
-\fB\-x\fR, \fB\-\-extract\fR
-extract files/directories from an archive
-.TP
-\fB\-z\fR, \fB\-\-compress\fR
-compress existing POSIX tar archives
-.TP
 \fB\-0\fR .. \fB\-9\fR
 set compression level [default 6]
 .TP
@@ -126,32 +121,44 @@ use <owner> name/ID for files added to archive
 \fB\-\-group=\fR<group>
 use <group> name/ID for files added to archive
 .TP
+\fB\-\-exclude=\fR<pattern>
+exclude files matching a shell pattern
+.TP
+\fB\-\-ignore\-ids\fR
+ignore differences in owner and group IDs
+.TP
+\fB\-\-ignore\-overflow\fR
+ignore mtime overflow differences on 32\-bit
+.TP
 \fB\-\-keep\-damaged\fR
 don't delete partially extracted files
 .TP
 \fB\-\-missing\-crc\fR
 exit with error status if missing extended CRC
 .TP
+\fB\-\-mtime=\fR<date>
+use <date> as mtime for files added to archive
+.TP
 \fB\-\-out\-slots=\fR<n>
 number of 1 MiB output packets buffered [64]
 .TP
-\fB\-\-check\-lib\fR
-compare version of lzlib.h with liblz.{a,so}
-.TP
 \fB\-\-warn\-newer\fR
 warn if any file is newer than the archive
 .PP
-Exit status: 0 for a normal exit, 1 for environmental problems (file not
-found, files differ, invalid flags, I/O errors, etc), 2 to indicate a
-corrupt or invalid input file, 3 for an internal consistency error (e.g. bug)
-which caused tarlz to panic.
+If no archive is specified, tarlz tries to read it from standard input or
+write it to standard output.
+.PP
+Exit status: 0 for a normal exit, 1 for environmental problems
+(file not found, files differ, invalid command line options, I/O errors,
+etc), 2 to indicate a corrupt or invalid input file, 3 for an internal
+consistency error (e.g., bug) which caused tarlz to panic.
 .SH "REPORTING BUGS"
 Report bugs to lzip\-bug@nongnu.org
 .br
 Tarlz home page: http://www.nongnu.org/lzip/tarlz.html
 .SH COPYRIGHT
 Copyright \(co 2022 Antonio Diaz Diaz.
-Using lzlib 1.13\-rc1
+Using lzlib 1.13
 License GPLv2+: GNU GPL version 2 or later <http://gnu.org/licenses/gpl.html>
 .br
 This is free software: you are free to change and redistribute it.
diff --git a/doc/tarlz.info b/doc/tarlz.info
index 79661cc..d71c0a3 100644
--- a/doc/tarlz.info
+++ b/doc/tarlz.info
@@ -1,6 +1,6 @@
 This is tarlz.info, produced by makeinfo version 4.13+ from tarlz.texi.
 
-INFO-DIR-SECTION Data Compression
+INFO-DIR-SECTION Archiving
 START-INFO-DIR-ENTRY
 * Tarlz: (tarlz).               Archiver with multimember lzip compression
 END-INFO-DIR-ENTRY
@@ -11,7 +11,7 @@ File: tarlz.info,  Node: Top,  Next: Introduction,  Up: (dir)
 Tarlz Manual
 ************
 
-This manual is for Tarlz (version 0.22, 5 January 2022).
+This manual is for Tarlz (version 0.23, 23 September 2022).
 
 * Menu:
 
@@ -69,9 +69,9 @@ archive, but it has the following advantages:
    * The resulting multimember tar.lz archive can be decompressed in
      parallel, multiplying the decompression speed.
 
-   * New members can be appended to the archive (by removing the EOF
-     member), and unwanted members can be deleted from the archive. Just
-     like an uncompressed tar archive.
+   * New members can be appended to the archive (by removing the
+     end-of-archive member), and unwanted members can be deleted from the
+     archive. Just like an uncompressed tar archive.
 
    * It is a safe POSIX-style backup format. In case of corruption, tarlz
      can extract all the undamaged members from the tar.lz archive,
@@ -99,19 +99,24 @@ File: tarlz.info,  Node: Invoking tarlz,  Next: Portable character set,  Prev: I
 
 The format for running tarlz is:
 
-     tarlz [OPTIONS] [FILES]
+     tarlz OPERATION [OPTIONS] [FILES]
 
 All operations except '--concatenate' and '--compress' operate on whole
 trees if any FILE is a directory. All operations except '--compress'
-overwrite output files without warning.
-
-   On archive creation or appending tarlz archives the files specified, but
-removes from member names any leading and trailing slashes and any file name
-prefixes containing a '..' component. On extraction, leading and trailing
-slashes are also removed from member names, and archive members containing
-a '..' component in the file name are skipped. Tarlz detects when the
-archive being created or enlarged is among the files to be dumped, appended
-or concatenated, and skips it.
+overwrite output files without warning. If no archive is specified, tarlz
+tries to read it from standard input or write it to standard output. Tarlz
+refuses to read archive data from a terminal or write archive data to a
+terminal. Tarlz detects when the archive being created or enlarged is among
+the files to be archived, appended, or concatenated, and skips it.
+
+   Tarlz does not use absolute file names nor file names above the current
+working directory (perhaps changed by option '-C'). On archive creation or
+appending tarlz archives the files specified, but removes from member names
+any leading and trailing slashes and any file name prefixes containing a
+'..' component. On extraction, leading and trailing slashes are also
+removed from member names, and archive members containing a '..' component
+in the file name are skipped. Tarlz does not follow symbolic links during
+extraction; not even symbolic links replacing intermediate directories.
 
    On extraction and listing, tarlz removes leading './' strings from
 member names in the archive or given in the command line, so that
@@ -122,8 +127,7 @@ member names in the archive or given in the command line, so that
 setting is used. For example '-9 --solid --uncompressed -1' is equivalent
 to '-1 --solid'.
 
-   tarlz supports the following options: *Note Argument syntax:
-(arg_parser)Argument syntax.
+   tarlz supports the following operations:
 
 '--help'
      Print an informative help message describing the options and exit.
@@ -140,39 +144,22 @@ to '-1 --solid'.
      standard output. All the archives involved must be regular (seekable)
      files, and must be either all compressed or all uncompressed.
      Compressed and uncompressed archives can't be mixed. Compressed
-     archives must be multimember lzip files with the two end-of-file
+     archives must be multimember lzip files with the two end-of-archive
      blocks plus any zero padding contained in the last lzip member of each
-     archive. The intermediate end-of-file blocks are removed as each new
-     archive is concatenated. If the archive is uncompressed, tarlz parses
-     and skips tar headers until it finds the end-of-file blocks. Exit with
+     archive. The intermediate end-of-archive blocks are removed as each
+     new archive is concatenated. If the archive is uncompressed, tarlz
+     parses tar headers until it finds the end-of-archive blocks. Exit with
      status 0 without modifying the archive if no FILES have been specified.
 
-'-B BYTES'
-'--data-size=BYTES'
-     Set target size of input data blocks for the option '--bsolid'. *Note
-     --bsolid::. Valid values range from 8 KiB to 1 GiB. Default value is
-     two times the dictionary size, except for option '-0' where it
-     defaults to 1 MiB. *Note Minimum archive sizes::.
+     Concatenating archives containing files in common results in two or
+     more tar members with the same name in the resulting archive, which
+     may produce nondeterministic behavior during multi-threaded extraction.
+     *Note mt-extraction::.
 
 '-c'
 '--create'
      Create a new archive from FILES.
 
-'-C DIR'
-'--directory=DIR'
-     Change to directory DIR. When creating or appending, the position of
-     each '-C' option in the command line is significant; it will change the
-     current working directory for the following FILES until a new '-C'
-     option appears in the command line. When extracting or comparing, all
-     the '-C' options are executed in sequence before reading the archive.
-     Listing ignores any '-C' options specified. DIR is relative to the
-     then current working directory, perhaps changed by a previous '-C'
-     option.
-
-     Note that a process can only have one current working directory (CWD).
-     Therefore multi-threading can't be used to create an archive if a '-C'
-     option appears after a relative file name in the command line.
-
 '-d'
 '--diff'
      Compare and report differences between archive and file system. For
@@ -188,10 +175,6 @@ to '-1 --solid'.
      on archive creation: 'tarlz -C / -d'. Alternatively, tarlz may be run
      from the root directory to perform the comparison.
 
-'--ignore-ids'
-     Make '--diff' ignore differences in owner and group IDs. This option is
-     useful when comparing an '--anonymous' archive.
-
 '--delete'
      Delete files and directories from an archive in place. It currently can
      delete only from uncompressed archives and from archives with files
@@ -210,12 +193,102 @@ to '-1 --solid'.
      be dangerous. A corrupt archive, a power cut, or an I/O error may cause
      data loss.
 
-'--exclude=PATTERN'
-     Exclude files matching a shell pattern like '*.o'. A file is considered
-     to match if any component of the file name matches. For example, '*.o'
-     matches 'foo.o', 'foo.o/bar' and 'foo/bar.o'. If PATTERN contains a
-     '/', it matches a corresponding '/' in the file name. For example,
-     'foo/*.o' matches 'foo/bar.o'.
+'-r'
+'--append'
+     Append files to the end of an archive. The archive must be a regular
+     (seekable) file either compressed or uncompressed. Compressed members
+     can't be appended to an uncompressed archive, nor vice versa. If the
+     archive is compressed, it must be a multimember lzip file with the two
+     end-of-archive blocks plus any zero padding contained in the last lzip
+     member of the archive. It is possible to append files to an archive
+     with a different compression granularity. Appending works as follows;
+     first the end-of-archive blocks are removed, then the new members are
+     appended, and finally two new end-of-archive blocks are appended to
+     the archive. If the archive is uncompressed, tarlz parses and skips
+     tar headers until it finds the end-of-archive blocks. Exit with status
+     0 without modifying the archive if no FILES have been specified.
+
+     Appending files already present in the archive results in two or more
+     tar members with the same name, which may produce nondeterministic
+     behavior during multi-threaded extraction. *Note mt-extraction::.
+
+'-t'
+'--list'
+     List the contents of an archive. If FILES are given, list only the
+     FILES given.
+
+'-x'
+'--extract'
+     Extract files from an archive. If FILES are given, extract only the
+     FILES given. Else extract all the files in the archive. To extract a
+     directory without extracting the files under it, use
+     'tarlz -xf foo --exclude='dir/*' dir'. Tarlz removes files and empty
+     directories unconditionally before extracting over them. Other than
+     that, it will not make any special effort to extract a file over an
+     incompatible type of file. For example, extracting a file over a
+     non-empty directory will usually fail.
+
+'-z'
+'--compress'
+     Compress existing POSIX tar archives aligning the lzip members to the
+     tar members with choice of granularity (--bsolid by default, --dsolid
+     works like --asolid). The input archives are kept unchanged. Existing
+     compressed archives are not overwritten. A hyphen '-' used as the name
+     of an input archive reads from standard input and writes to standard
+     output (unless the option '--output' is used). Tarlz can be used as
+     compressor for GNU tar using a command like
+     'tar -c -Hustar foo | tarlz -z -o foo.tar.lz'. Note that tarlz only
+     works reliably on archives without global headers, or with global
+     headers whose content can be ignored.
+
+     The compression is reversible, including any garbage present after the
+     end-of-archive blocks. Tarlz stops parsing after the first
+     end-of-archive block is found, and then compresses the rest of the
+     archive. Unless solid compression is requested, the end-of-archive
+     blocks are compressed in a lzip member separated from the preceding
+     members and from any non-zero garbage following the end-of-archive
+     blocks. '--compress' implies plzip argument style, not tar style. Each
+     input archive is compressed to a file with the extension '.lz' added
+     unless the option '--output' is used. When '--output' is used, only
+     one input archive can be specified. '-f' can't be used with
+     '--compress'.
+
+'--check-lib'
+     Compare the version of lzlib used to compile tarlz with the version
+     actually being used at run time and exit. Report any differences
+     found. Exit with error status 1 if differences are found. A mismatch
+     may indicate that lzlib is not correctly installed or that a different
+     version of lzlib has been installed after compiling tarlz. Exit with
+     error status 2 if LZ_API_VERSION and LZ_version_string don't match.
+     'tarlz -v --check-lib' shows the version of lzlib being used and the
+     value of LZ_API_VERSION (if defined). *Note Library version:
+     (lzlib)Library version.
+
+
+   tarlz supports the following options: *Note Argument syntax:
+(arg_parser)Argument syntax.
+
+'-B BYTES'
+'--data-size=BYTES'
+     Set target size of input data blocks for the option '--bsolid'. *Note
+     --bsolid::. Valid values range from 8 KiB to 1 GiB. Default value is
+     two times the dictionary size, except for option '-0' where it
+     defaults to 1 MiB. *Note Minimum archive sizes::.
+
+'-C DIR'
+'--directory=DIR'
+     Change to directory DIR. When creating or appending, the position of
+     each '-C' option in the command line is significant; it will change the
+     current working directory for the following FILES until a new '-C'
+     option appears in the command line. When extracting or comparing, all
+     the '-C' options are executed in sequence before reading the archive.
+     Listing ignores any '-C' options specified. DIR is relative to the
+     then current working directory, perhaps changed by a previous '-C'
+     option.
+
+     Note that a process can only have one current working directory (CWD).
+     Therefore multi-threading can't be used to create an archive if a '-C'
+     option appears after a relative file name in the command line.
 
 '-f ARCHIVE'
 '--file=ARCHIVE'
@@ -228,14 +301,6 @@ to '-1 --solid'.
      Archive or compare the files they point to instead of the links
      themselves.
 
-'--mtime=DATE'
-     When creating or appending, use DATE as the modification time for
-     files added to the archive instead of their actual modification times.
-     The value of DATE may be either '@' followed by the number of seconds
-     since the epoch, or a date in format 'YYYY-MM-DD HH:MM:SS', or the
-     name of an existing file starting with '.' or '/'. In the latter case,
-     the modification time of that file is used.
-
 '-n N'
 '--threads=N'
      Set the number of (de)compression threads, overriding the system's
@@ -268,65 +333,11 @@ to '-1 --solid'.
 '--quiet'
      Quiet operation. Suppress all messages.
 
-'-r'
-'--append'
-     Append files to the end of an archive. The archive must be a regular
-     (seekable) file either compressed or uncompressed. Compressed members
-     can't be appended to an uncompressed archive, nor vice versa. If the
-     archive is compressed, it must be a multimember lzip file with the two
-     end-of-file blocks plus any zero padding contained in the last lzip
-     member of the archive. It is possible to append files to an archive
-     with a different compression granularity. Appending works as follows;
-     first the end-of-file blocks are removed, then the new members are
-     appended, and finally two new end-of-file blocks are appended to the
-     archive. If the archive is uncompressed, tarlz parses and skips tar
-     headers until it finds the end-of-file blocks. Exit with status 0
-     without modifying the archive if no FILES have been specified.
-
-'-t'
-'--list'
-     List the contents of an archive. If FILES are given, list only the
-     FILES given.
-
 '-v'
 '--verbose'
      Verbosely list files processed. Further -v's (up to 4) increase the
      verbosity level.
 
-'-x'
-'--extract'
-     Extract files from an archive. If FILES are given, extract only the
-     FILES given. Else extract all the files in the archive. To extract a
-     directory without extracting the files under it, use
-     'tarlz -xf foo --exclude='dir/*' dir'. Tarlz will not make any special
-     effort to extract a file over an incompatible type of file. For
-     example, extracting a link over a directory will usually fail.
-     (Principle of least surprise).
-
-'-z'
-'--compress'
-     Compress existing POSIX tar archives aligning the lzip members to the
-     tar members with choice of granularity (--bsolid by default, --dsolid
-     works like --asolid). The input archives are kept unchanged. Existing
-     compressed archives are not overwritten. A hyphen '-' used as the name
-     of an input archive reads from standard input and writes to standard
-     output (unless the option '--output' is used). Tarlz can be used as
-     compressor for GNU tar using a command like
-     'tar -c -Hustar foo | tarlz -z -o foo.tar.lz'. Note that tarlz only
-     works reliably on archives without global headers, or with global
-     headers whose content can be ignored.
-
-     The compression is reversible, including any garbage present after the
-     EOF blocks. Tarlz stops parsing after the first EOF block is found,
-     and then compresses the rest of the archive. Unless solid compression
-     is requested, the EOF blocks are compressed in a lzip member separated
-     from the preceding members and from any non-zero garbage following the
-     EOF blocks. '--compress' implies plzip argument style, not tar style.
-     Each input archive is compressed to a file with the extension '.lz'
-     added unless the option '--output' is used. When '--output' is used,
-     only one input archive can be specified. '-f' can't be used with
-     '--compress'.
-
 '-0 .. -9'
      Set the compression level for '--create', '--append', and
      '--compress'. The default compression level is '-6'. Like lzip, tarlz
@@ -354,8 +365,8 @@ to '-1 --solid'.
 '--asolid'
      When creating or appending to a compressed archive, use appendable
      solid compression. All the files being added to the archive are
-     compressed into a single lzip member, but the end-of-file blocks are
-     compressed into a separate lzip member. This creates a solidly
+     compressed into a single lzip member, but the end-of-archive blocks
+     are compressed into a separate lzip member. This creates a solidly
      compressed appendable archive. Solid archives can't be created nor
      decoded in parallel.
 
@@ -375,20 +386,20 @@ to '-1 --solid'.
      When creating or appending to a compressed archive, compress each file
      specified in the command line separately in its own lzip member, and
      use solid compression for each directory specified in the command
-     line. The end-of-file blocks are compressed into a separate lzip
+     line. The end-of-archive blocks are compressed into a separate lzip
      member. This creates a compressed appendable archive with a separate
      lzip member for each file or top-level directory specified.
 
 '--no-solid'
      When creating or appending to a compressed archive, compress each file
-     separately in its own lzip member. The end-of-file blocks are
+     separately in its own lzip member. The end-of-archive blocks are
      compressed into a separate lzip member. This creates a compressed
      appendable archive with a lzip member for each file.
 
 '--solid'
      When creating or appending to a compressed archive, use solid
      compression. The files being added to the archive, along with the
-     end-of-file blocks, are compressed into a single lzip member. The
+     end-of-archive blocks, are compressed into a single lzip member. The
      resulting archive is not appendable. No more files can be later
      appended to the archive. Solid archives can't be created nor decoded
      in parallel.
@@ -406,22 +417,50 @@ to '-1 --solid'.
      If GROUP is not a valid group name, it is decoded as a decimal numeric
      group ID.
 
+'--exclude=PATTERN'
+     Exclude files matching a shell pattern like '*.o'. A file is considered
+     to match if any component of the file name matches. For example, '*.o'
+     matches 'foo.o', 'foo.o/bar' and 'foo/bar.o'. If PATTERN contains a
+     '/', it matches a corresponding '/' in the file name. For example,
+     'foo/*.o' matches 'foo/bar.o'. Multiple '--exclude' options can be
+     specified.
+
+'--ignore-ids'
+     Make '--diff' ignore differences in owner and group IDs. This option is
+     useful when comparing an '--anonymous' archive.
+
+'--ignore-overflow'
+     Make '--diff' ignore differences in mtime caused by overflow on 32-bit
+     systems with a 32-bit time_t.
+
 '--keep-damaged'
      Don't delete partially extracted files. If a decompression error
      happens while extracting a file, keep the partial data extracted. Use
      this option to recover as much data as possible from each damaged
      member. It is recommended to run tarlz in single-threaded mode
-     (-threads=0) when using this option.
+     (--threads=0) when using this option.
 
 '--missing-crc'
-     Exit with error status 2 if the CRC of the extended records is missing.
-     When this option is used, tarlz detects any corruption in the extended
-     records (only limited by CRC collisions). But note that a corrupt
-     'GNU.crc32' keyword, for example 'GNU.crc33', is reported as a missing
-     CRC instead of as a corrupt record. This misleading 'Missing CRC'
-     message is the consequence of a flaw in the POSIX pax format; i.e.,
-     the lack of a mandatory check sequence in the extended records. *Note
-     crc32::.
+     Exit with error status 2 if the CRC of the extended records is
+     missing. When this option is used, tarlz detects any corruption in the
+     extended records (only limited by CRC collisions). But note that a
+     corrupt 'GNU.crc32' keyword, for example 'GNU.crc33', is reported as a
+     missing CRC instead of as a corrupt record. This misleading
+     'Missing CRC' message is the consequence of a flaw in the POSIX pax
+     format; i.e., the lack of a mandatory check sequence of the extended
+     records. *Note crc32::.
+
+'--mtime=DATE'
+     When creating or appending, use DATE as the modification time for
+     files added to the archive instead of their actual modification times.
+     The value of DATE may be either '@' followed by the number of seconds
+     since (or before) the epoch, or a date in format
+     '[-]YYYY-MM-DD HH:MM:SS' or '[-]YYYY-MM-DDTHH:MM:SS', or the name of
+     an existing reference file starting with '.' or '/' whose modification
+     time is used. The time of day 'HH:MM:SS' in the date format is
+     optional and defaults to '00:00:00'. The epoch is
+     '1970-01-01 00:00:00 UTC'. Negative seconds or years define a
+     modification time before the epoch.
 
 '--out-slots=N'
      Number of 1 MiB output packets buffered per worker thread during
@@ -431,17 +470,6 @@ to '-1 --solid'.
      more memory. Valid values range from 1 to 1024. The default value is
      64.
 
-'--check-lib'
-     Compare the version of lzlib used to compile tarlz with the version
-     actually being used at run time and exit. Report any differences
-     found. Exit with error status 1 if differences are found. A mismatch
-     may indicate that lzlib is not correctly installed or that a different
-     version of lzlib has been installed after compiling tarlz. Exit with
-     error status 2 if LZ_API_VERSION and LZ_version_string don't match.
-     'tarlz -v --check-lib' shows the version of lzlib being used and the
-     value of LZ_API_VERSION (if defined). *Note Library version:
-     (lzlib)Library version.
-
 '--warn-newer'
      During archive creation, warn if any file being archived has a
      modification time newer than the archive creation time. This option
@@ -453,9 +481,9 @@ to '-1 --solid'.
 
 
    Exit status: 0 for a normal exit, 1 for environmental problems (file not
-found, files differ, invalid flags, I/O errors, etc), 2 to indicate a
-corrupt or invalid input file, 3 for an internal consistency error (e.g.
-bug) which caused tarlz to panic.
+found, files differ, invalid command line options, I/O errors, etc), 2 to
+indicate a corrupt or invalid input file, 3 for an internal consistency
+error (e.g., bug) which caused tarlz to panic.
 
 
 File: tarlz.info,  Node: Portable character set,  Next: File format,  Prev: Invoking tarlz,  Up: Top
@@ -473,9 +501,7 @@ The set of characters from which portable file names are constructed.
 characters, respectively.
 
    File names are identifiers. Therefore, archiving works better when file
-names use only the portable character set without spaces added. Unicode is
-for human consumption. It should be avoided in computing environments,
-specially in file names. *Note why not Unicode: (moe)why not Unicode.
+names use only the portable character set without spaces added.
 
 
 File: tarlz.info,  Node: File format,  Next: Amendments to pax format,  Prev: Portable character set,  Up: Top
@@ -512,10 +538,11 @@ limitations of the ustar format.
    Each tar member contains one file archived, and is represented by the
 following sequence:
 
-   * An optional extended header block with extended header records. This
-     header block is of the form described in pax header block, with a
-     typeflag value of 'x'. The extended header records are included as the
-     data for this header block.
+   * An optional extended header block followed by one or more blocks that
+     contain the extended header records as if they were the contents of a
+     file; i.e., the extended header records are included as the data for
+     this header block. This header block is of the form described in pax
+     header block, with a typeflag value of 'x'.
 
    * A header block in ustar format that describes the file. Any fields
      defined in the preceding optional extended header records override the
@@ -529,9 +556,11 @@ split over two or more lzip members, the archive must be decoded
 sequentially. *Note Multi-threaded decoding::.
 
    At the end of the archive file there are two 512-byte blocks filled with
-binary zeros, interpreted as an end-of-archive indicator. These EOF blocks
-are either compressed in a separate lzip member or compressed along with
-the tar members contained in the last lzip member.
+binary zeros, interpreted as an end-of-archive indicator. These EOA blocks
+are either compressed in a separate lzip member or compressed along with the
+tar members contained in the last lzip member. For a compressed archive to
+be recognized by tarlz as appendable, the last lzip member must contain
+between 512 and 32256 zeros alone.
 
    The diagram below shows the correspondence between each tar member
 (formed by one or two headers plus optional data) in the tar archive and
@@ -540,7 +569,7 @@ compression is used: *Note File format: (lzip)File format.
 
 tar
 +========+======+=================+===============+========+======+========+
-| header | data | extended header | extended data | header | data |   EOF  |
+| header | data | extended header | extended data | header | data |   EOA  |
 +========+======+=================+===============+========+======+========+
 
 tar.lz
@@ -572,25 +601,57 @@ space, equal-sign, and newline.
 
    These are the <keyword> values currently supported by tarlz:
 
+'atime'
+     The signed decimal representation of the access time of the following
+     file in seconds since (or before) the epoch, obtained from the function
+     'stat'. The atime record is created only for files with a modification
+     time outside of the ustar range. *Note ustar-mtime::.
+
+'gid'
+     The unsigned decimal representation of the group ID of the group that
+     owns the following file. The gid record is created only for files with
+     a group ID greater than 2_097_151 (octal 7777777). *Note
+     ustar-uid-gid::.
+
 'linkpath'
-     The pathname of a link being created to another file, of any type,
+     The file name of a link being created to another file, of any type,
      previously archived. This record overrides the field 'linkname' in the
      following ustar header block. The following ustar header block
      determines the type of link created. If typeflag of the following
      header block is 1, it will be a hard link. If typeflag is 2, it will
      be a symbolic link and the linkpath value will be used as the contents
-     of the symbolic link.
+     of the symbolic link. The linkpath record is created only for links
+     with a link name that does not fit in the space provided by the ustar
+     header.
+
+'mtime'
+     The signed decimal representation of the modification time of the
+     following file in seconds since (or before) the epoch, obtained from
+     the function 'stat'. This record overrides the field 'mtime' in the
+     following ustar header block. The mtime record is created only for
+     files with a modification time outside of the ustar range. *Note
+     ustar-mtime::.
 
 'path'
-     The pathname of the following file. This record overrides the fields
-     'name' and 'prefix' in the following ustar header block.
+     The file name of the following file. This record overrides the fields
+     'name' and 'prefix' in the following ustar header block. The path
+     record is created for files with a name that does not fit in the space
+     provided by the ustar header, but is also created for files that
+     require any other extended record so that the fields 'name' and
+     'prefix' in the following ustar header block can be zeroed.
 
 'size'
      The size of the file in bytes, expressed as a decimal number using
      digits from the ISO/IEC 646:1991 (ASCII) standard. This record
-     overrides the size field in the following ustar header block. The size
-     record is used only for files with a size value greater than
-     8_589_934_591 (octal 77777777777). This is 2^33 bytes or larger.
+     overrides the field 'size' in the following ustar header block. The
+     size record is created only for files with a size value greater than
+     8_589_934_591 (octal 77777777777); that is, 8 GiB (2^33 bytes) or
+     larger.
+
+'uid'
+     The unsigned decimal representation of the user ID of the file owner
+     of the following file. The uid record is created only for files with a
+     user ID greater than 2_097_151 (octal 7777777). *Note ustar-uid-gid::.
 
 'GNU.crc32'
      CRC32-C (Castagnoli) of the extended header data excluding the 8 bytes
@@ -643,18 +704,18 @@ and groups, tarlz will use the byte values in these names unmodified.
 character strings except when all characters in the array contain non-null
 characters including the last character.
 
-   The fields 'prefix' and 'name' produce the pathname of the file. A new
-pathname is formed, if prefix is not an empty string (its first character
-is not null), by concatenating prefix (up to the first null character), a
-slash character, and name; otherwise, name is used alone. In either case,
-name is terminated at the first null character. If prefix begins with a
-null character, it is ignored. In this manner, pathnames of at most 256
-characters can be supported. If a pathname does not fit in the space
-provided, an extended record is used to store the pathname.
+   The fields 'name' and 'prefix' produce the file name. A new file name is
+formed, if prefix is not an empty string (its first character is not null),
+by concatenating prefix (up to the first null character), a slash
+character, and name; otherwise, name is used alone. In either case, name is
+terminated at the first null character. If prefix begins with a null
+character, it is ignored. In this manner, file names of at most 256
+characters can be supported. If a file name does not fit in the space
+provided, an extended record is used to store the file name.
 
-   The field 'linkname' does not use the prefix to produce a pathname. If
-the linkname does not fit in the 100 characters provided, an extended record
-is used to store the linkname.
+   The field 'linkname' does not use the prefix to produce a file name. If
+the link name does not fit in the 100 characters provided, an extended
+record is used to store the link name.
 
    The field 'mode' provides 12 access permission bits. The following table
 shows the symbolic name of each bit and its octal value:
@@ -667,7 +728,9 @@ S_IRGRP    00040   S_IWGRP    00020   S_IXGRP    00010
 S_IROTH    00004   S_IWOTH    00002   S_IXOTH    00001
 
    The fields 'uid' and 'gid' are the user and group IDs of the owner and
-group of the file, respectively.
+group of the file, respectively. If the file uid or gid are greater than
+2_097_151 (octal 7777777), an extended record is used to store the uid or
+gid.
 
    The field 'size' contains the octal representation of the size of the
 file in bytes. If the field 'typeflag' specifies a file of type '0'
@@ -680,7 +743,10 @@ header. If the file size is larger than 8_589_934_591 bytes
 
    The field 'mtime' contains the octal representation of the modification
 time of the file at the time it was archived, obtained from the function
-'stat'.
+'stat'. If the modification time is negative or larger than 8_589_934_591
+(octal 77777777777) seconds since the epoch, an extended record is used to
+store the modification time. The ustar range of mtime goes from
+'1970-01-01 00:00:00 UTC' to '2242-03-16 12:56:31 UTC'.
 
    The field 'chksum' contains the octal representation of the value of the
 simple sum of all bytes in the header logical record. Each byte in the
@@ -694,7 +760,8 @@ file archived:
      Regular file.
 
 ''1''
-     Hard link to another file, of any type, previously archived.
+     Hard link to another file, of any type, previously archived. Hard
+     links must not contain file data.
 
 ''2''
      Symbolic link.
@@ -712,8 +779,8 @@ file archived:
 
 ''7''
      Reserved to represent a file to which an implementation has associated
-     some high-performance attribute. Tarlz treats this type of file as a
-     regular file (type 0).
+     some high-performance attribute (contiguous file). Tarlz treats this
+     type of file as a regular file (type 0).
 
 
    The field 'magic' contains the ASCII null-terminated string "ustar". The
@@ -735,9 +802,9 @@ Tarlz creates safe archives that allow the reliable detection of invalid or
 corrupt metadata during decoding even when the integrity checking of lzip
 can't be used because the lzip members are only decompressed partially, as
 it happens in parallel '--diff', '--list', and '--extract'. In order to
-achieve this goal, tarlz makes some changes to the variant of the pax
-format that it uses. This chapter describes these changes and the concrete
-reasons to implement them.
+achieve this goal and avoid some other flaws in the pax format, tarlz makes
+some changes to the variant of the pax format that it uses. This chapter
+describes these changes and the concrete reasons to implement them.
 
 
 5.1 Add a CRC of the extended records
@@ -775,45 +842,73 @@ In order to allow the extraction of pax archives by a tar utility conforming
 to the POSIX-2:1993 standard, POSIX.1-2008 recommends selecting extended
 header field values that allow such tar to create a regular file containing
 the extended header records as data. This approach is broken because if the
-extended header is needed because of a long file name, the fields 'prefix'
-and 'name' will be unable to contain the full pathname of the file.
-Therefore the files corresponding to both the extended header and the
-overridden ustar header will be extracted using truncated file names,
-perhaps overwriting existing files or directories. It may be a security risk
-to extract a file with a truncated file name.
+extended header is needed because of a long file name, the fields 'name'
+and 'prefix' will be unable to contain the full file name. (Some tar
+implementations store the truncated name in the field 'name' alone,
+truncating the name to only 100 bytes instead of 256). Therefore the files
+corresponding to both the extended header and the overridden ustar header
+will be extracted using truncated file names, perhaps overwriting existing
+files or directories. It may be a security risk to extract a file with a
+truncated file name.
 
    To avoid this problem, tarlz writes extended headers with all fields
-zeroed except size, chksum, typeflag, magic and version. This prevents old
-tar programs from extracting the extended records as a file in the wrong
-place. Tarlz also sets to zero those fields of the ustar header overridden
-by extended records. Finally, tarlz skips members without name when decoding
-except when listing. This is needed to detect certain format violations
-during parallel extraction.
-
-   If an extended header is required for any reason (for example a file size
-larger than 8 GiB or a link name longer than 100 bytes), tarlz moves the
-file name also to the extended header to prevent an ustar tool from trying
-to extract the file or link. This also makes easier during parallel decoding
-the detection of a tar member split between two lzip members at the boundary
-between the extended header and the ustar header.
+zeroed except 'size' (which contains the size of the extended records),
+'chksum', 'typeflag', 'magic', and 'version'. In particular, tarlz sets the
+fields 'name' and 'prefix' to zero. This prevents old tar programs from
+extracting the extended records as a file in the wrong place. Tarlz also
+sets to zero those fields of the ustar header overridden by extended
+records. Finally, tarlz skips members with zeroed 'name' and 'prefix' when
+decoding, except when listing. This is needed to detect certain format
+violations during parallel extraction.
+
+   If an extended header is required for any reason (for example a file
+size of 8 GiB or larger, or a link name longer than 100 bytes), tarlz also
+moves the file name to the extended records to prevent an ustar tool from
+trying to extract the file or link. This also makes easier during parallel
+decoding the detection of a tar member split between two lzip members at
+the boundary between the extended header and the ustar header.
 
 
 5.3 As simple as possible (but not simpler)
 ===========================================
 
 The tarlz format is mainly ustar. Extended pax headers are used only when
-needed because the length of a file name or link name, or the size of a file
-exceed the limits of the ustar format. Adding 1 KiB of extended headers to
-each member just to record subsecond timestamps seems wasteful for a backup
-format. Moreover, minimizing the overhead may help recovering the archive
-with lziprecover in case of corruption.
+needed because the length of a file name or link name, or the size or other
+attribute of a file exceed the limits of the ustar format. Adding 1 KiB of
+extended header and records to each member just to save subsecond
+timestamps seems wasteful for a backup format. Moreover, minimizing the
+overhead may help recovering the archive with lziprecover in case of
+corruption.
 
    Global pax headers are tolerated, but not supported; they are parsed and
 ignored. Some operations may not behave as expected if the archive contains
 global headers.
 
 
-5.4 Avoid misconversions to/from UTF-8
+5.4 Improve reproducibility
+===========================
+
+Pax includes by default the process ID of the pax process in the ustar name
+of the extended headers, making the archive not reproducible. Tarlz stores
+the true name of the file just once, either in the ustar header or in the
+extended records, making it easier to produce reproducible archives.
+
+   Pax allows an extended record to have length x-1 or x if x is a power of
+ten; '99<97_bytes>' or '100<97_bytes>'. Tarlz minimizes the length of the
+record and always produces a length of x-1 in these cases.
+
+
+5.5 No data in hard links
+=========================
+
+Tarlz does not allow data in hard link members. The data (if any) must be in
+the member determining the type of the file (which can't be a link). If all
+the names of a file are stored as hard links, the type of the file is lost.
+Not allowing data in hard links also prevents invalid actions like
+extracting file data for a hard link to a symbolic link or to a directory.
+
+
+5.6 Avoid misconversions to/from UTF-8
 ======================================
 
 There is no portable way to tell what charset a text string is coded into.
@@ -968,12 +1063,12 @@ headers must provide their own integrity checking.
 Multi-threaded extraction may produce different output than single-threaded
 extraction in some cases:
 
-   During multi-threaded extraction, several independent processes are
+   During multi-threaded extraction, several independent threads are
 simultaneously reading the archive and creating files in the file system.
 The archive is not read sequentially. As a consequence, any error or
-weirdness in the archive (like a corrupt member or an EOF block in the
-middle of the archive) won't be usually detected until part of the archive
-beyond that point has been processed.
+weirdness in the archive (like a corrupt member or an end-of-archive block
+in the middle of the archive) won't be usually detected until part of the
+archive beyond that point has been processed.
 
    If the archive contains two or more tar members with the same name,
 single-threaded extraction extracts the members in the order they appear in
@@ -986,6 +1081,9 @@ unspecified which of the tar members is extracted.
 names resolve to the same file in the file system), the result is undefined.
 (Probably the resulting file will be mangled).
 
+   Extraction of a hard link may fail if it is extracted before the file it
+links to.
+
 
 File: tarlz.info,  Node: Minimum archive sizes,  Next: Examples,  Prev: Multi-threaded decoding,  Up: Top
 
@@ -1054,7 +1152,7 @@ Example 4: Create a compressed appendable archive containing directories
 'dir1', 'dir2' and 'dir3' with a separate lzip member per directory. Then
 append files 'a', 'b', 'c', 'd' and 'e' to the archive, all of them
 contained in a single lzip member. The resulting archive 'archive.tar.lz'
-contains 5 lzip members (including the EOF member).
+contains 5 lzip members (including the end-of-archive member).
 
      tarlz --dsolid -cf archive.tar.lz dir1 dir2 dir3
      tarlz --asolid -rf archive.tar.lz a b c d e
@@ -1081,7 +1179,7 @@ Example 7: Extract files 'a' and 'c', and the whole tree under directory
 Example 8: Copy the contents of directory 'sourcedir' to the directory
 'destdir'.
 
-     tarlz -C sourcedir -c . | tarlz -C destdir -x
+     tarlz -C sourcedir --uncompressed -cf - . | tarlz -C destdir -xf -
 
 
 Example 9: Compress the existing POSIX archive 'archive.tar' and write the
@@ -1091,6 +1189,18 @@ other members can still be extracted).
 
      tarlz -z --no-solid archive.tar
 
+
+Example 10: Compress the archive 'archive.tar' and write the output to
+'foo.tar.lz'.
+
+     tarlz -z -o foo.tar.lz archive.tar
+
+
+Example 11: Concatenate and compress two archives 'archive1.tar' and
+'archive2.tar', and write the output to 'foo.tar.lz'.
+
+     tarlz -A archive1.tar archive2.tar | tarlz -z -o foo.tar.lz
+
 
 File: tarlz.info,  Node: Problems,  Next: Concept index,  Prev: Examples,  Up: Top
 
@@ -1133,23 +1243,26 @@ Concept index
 
 
 Tag Table:
-Node: Top223
-Node: Introduction1214
-Node: Invoking tarlz4022
-Ref: --data-size6436
-Ref: --bsolid16388
-Node: Portable character set21224
-Node: File format22019
-Ref: key_crc3226944
-Node: Amendments to pax format32572
-Ref: crc3233236
-Ref: flawed-compat34547
-Node: Program design37348
-Node: Multi-threaded decoding41273
-Node: Minimum archive sizes45764
-Node: Examples47902
-Node: Problems49918
-Node: Concept index50473
+Node: Top216
+Node: Introduction1210
+Node: Invoking tarlz4029
+Ref: --data-size12880
+Ref: --bsolid17192
+Node: Portable character set22788
+Node: File format23431
+Ref: key_crc3230188
+Ref: ustar-uid-gid33452
+Ref: ustar-mtime34254
+Node: Amendments to pax format36254
+Ref: crc3236963
+Ref: flawed-compat38274
+Node: Program design42364
+Node: Multi-threaded decoding46289
+Ref: mt-extraction49570
+Node: Minimum archive sizes50876
+Node: Examples53014
+Node: Problems55381
+Node: Concept index55936
 
 End Tag Table
 
diff --git a/doc/tarlz.texi b/doc/tarlz.texi
index bfa6d9d..5bdd2af 100644
--- a/doc/tarlz.texi
+++ b/doc/tarlz.texi
@@ -6,10 +6,10 @@
 @finalout
 @c %**end of header
 
-@set UPDATED 5 January 2022
-@set VERSION 0.22
+@set UPDATED 23 September 2022
+@set VERSION 0.23
 
-@dircategory Data Compression
+@dircategory Archiving
 @direntry
 * Tarlz: (tarlz).               Archiver with multimember lzip compression
 @end direntry
@@ -96,9 +96,9 @@ The resulting multimember tar.lz archive can be decompressed in
 parallel, multiplying the decompression speed.
 
 @item
-New members can be appended to the archive (by removing the EOF
-member), and unwanted members can be deleted from the archive. Just
-like an uncompressed tar archive.
+New members can be appended to the archive (by removing the
+end-of-archive member), and unwanted members can be deleted from the
+archive. Just like an uncompressed tar archive.
 
 @item
 It is a safe POSIX-style backup format. In case of corruption, tarlz
@@ -111,7 +111,7 @@ lziprecover can be used to recover some of the damaged members.
 @item
 A multimember tar.lz archive is usually smaller than the corresponding
 solidly compressed tar.gz archive, except when individually
-compressing files smaller than about 32 KiB.
+compressing files smaller than about @w{32 KiB}.
 @end itemize
 
 Tarlz protects the extended records with a Cyclic Redundancy Check (CRC) in
@@ -133,21 +133,28 @@ the format of the archive is compatible with tarlz.
 The format for running tarlz is:
 
 @example
-tarlz [@var{options}] [@var{files}]
+tarlz @var{operation} [@var{options}] [@var{files}]
 @end example
 
 @noindent
 All operations except @samp{--concatenate} and @samp{--compress} operate on
 whole trees if any @var{file} is a directory. All operations except
-@samp{--compress} overwrite output files without warning.
-
-On archive creation or appending tarlz archives the files specified, but
-removes from member names any leading and trailing slashes and any file name
-prefixes containing a @samp{..} component. On extraction, leading and
-trailing slashes are also removed from member names, and archive members
-containing a @samp{..} component in the file name are skipped. Tarlz detects
-when the archive being created or enlarged is among the files to be dumped,
-appended or concatenated, and skips it.
+@samp{--compress} overwrite output files without warning. If no archive is
+specified, tarlz tries to read it from standard input or write it to
+standard output. Tarlz refuses to read archive data from a terminal or write
+archive data to a terminal. Tarlz detects when the archive being created or
+enlarged is among the files to be archived, appended, or concatenated, and
+skips it.
+
+Tarlz does not use absolute file names nor file names above the current
+working directory (perhaps changed by option @samp{-C}). On archive creation
+or appending tarlz archives the files specified, but removes from member
+names any leading and trailing slashes and any file name prefixes containing
+a @samp{..} component. On extraction, leading and trailing slashes are also
+removed from member names, and archive members containing a @samp{..}
+component in the file name are skipped. Tarlz does not follow symbolic links
+during extraction; not even symbolic links replacing intermediate
+directories.
 
 On extraction and listing, tarlz removes leading @samp{./} strings from
 member names in the archive or given in the command line, so that
@@ -158,11 +165,7 @@ If several compression levels or @samp{--*solid} options are given, the last
 setting is used. For example @w{@samp{-9 --solid --uncompressed -1}} is
 equivalent to @w{@samp{-1 --solid}}.
 
-tarlz supports the following
-@uref{http://www.nongnu.org/arg-parser/manual/arg_parser_manual.html#Argument-syntax,,options}:
-@ifnothtml
-@xref{Argument syntax,,,arg_parser}.
-@end ifnothtml
+tarlz supports the following operations:
 
 @table @code
 @item --help
@@ -180,40 +183,22 @@ specified with the option @samp{-f}, the input archives are concatenated to
 standard output. All the archives involved must be regular (seekable) files,
 and must be either all compressed or all uncompressed. Compressed and
 uncompressed archives can't be mixed. Compressed archives must be
-multimember lzip files with the two end-of-file blocks plus any zero padding
-contained in the last lzip member of each archive. The intermediate
-end-of-file blocks are removed as each new archive is concatenated. If the
-archive is uncompressed, tarlz parses and skips tar headers until it finds
-the end-of-file blocks. Exit with status 0 without modifying the archive if
+multimember lzip files with the two end-of-archive blocks plus any zero
+padding contained in the last lzip member of each archive. The intermediate
+end-of-archive blocks are removed as each new archive is concatenated. If
+the archive is uncompressed, tarlz parses tar headers until it finds the
+end-of-archive blocks. Exit with status 0 without modifying the archive if
 no @var{files} have been specified.
 
-@anchor{--data-size}
-@item -B @var{bytes}
-@itemx --data-size=@var{bytes}
-Set target size of input data blocks for the option @samp{--bsolid}.
-@xref{--bsolid}. Valid values range from @w{8 KiB} to @w{1 GiB}. Default
-value is two times the dictionary size, except for option @samp{-0} where it
-defaults to @w{1 MiB}. @xref{Minimum archive sizes}.
+Concatenating archives containing files in common results in two or more tar
+members with the same name in the resulting archive, which may produce
+nondeterministic behavior during multi-threaded extraction.
+@xref{mt-extraction}.
 
 @item -c
 @itemx --create
 Create a new archive from @var{files}.
 
-@item -C @var{dir}
-@itemx --directory=@var{dir}
-Change to directory @var{dir}. When creating or appending, the position of
-each @samp{-C} option in the command line is significant; it will change the
-current working directory for the following @var{files} until a new
-@samp{-C} option appears in the command line. When extracting or comparing,
-all the @samp{-C} options are executed in sequence before reading the
-archive. Listing ignores any @samp{-C} options specified. @var{dir} is
-relative to the then current working directory, perhaps changed by a
-previous @samp{-C} option.
-
-Note that a process can only have one current working directory (CWD).
-Therefore multi-threading can't be used to create an archive if a @samp{-C}
-option appears after a relative file name in the command line.
-
 @item -d
 @itemx --diff
 Compare and report differences between archive and file system. For each tar
@@ -228,10 +213,6 @@ be used in combination with @samp{--diff} when absolute file names were used
 on archive creation: @w{@samp{tarlz -C / -d}}. Alternatively, tarlz may be
 run from the root directory to perform the comparison.
 
-@item --ignore-ids
-Make @samp{--diff} ignore differences in owner and group IDs. This option is
-useful when comparing an @samp{--anonymous} archive.
-
 @item --delete
 Delete files and directories from an archive in place. It currently can
 delete only from uncompressed archives and from archives with files
@@ -249,12 +230,109 @@ To delete a directory without deleting the files under it, use
 may be dangerous. A corrupt archive, a power cut, or an I/O error may cause
 data loss.
 
-@item --exclude=@var{pattern}
-Exclude files matching a shell pattern like @samp{*.o}. A file is considered
-to match if any component of the file name matches. For example, @samp{*.o}
-matches @samp{foo.o}, @samp{foo.o/bar} and @samp{foo/bar.o}. If
-@var{pattern} contains a @samp{/}, it matches a corresponding @samp{/} in
-the file name. For example, @samp{foo/*.o} matches @samp{foo/bar.o}.
+@item -r
+@itemx --append
+Append files to the end of an archive. The archive must be a regular
+(seekable) file either compressed or uncompressed. Compressed members can't
+be appended to an uncompressed archive, nor vice versa. If the archive is
+compressed, it must be a multimember lzip file with the two end-of-archive
+blocks plus any zero padding contained in the last lzip member of the
+archive. It is possible to append files to an archive with a different
+compression granularity. Appending works as follows; first the
+end-of-archive blocks are removed, then the new members are appended, and
+finally two new end-of-archive blocks are appended to the archive. If the
+archive is uncompressed, tarlz parses and skips tar headers until it finds
+the end-of-archive blocks. Exit with status 0 without modifying the archive
+if no @var{files} have been specified.
+
+Appending files already present in the archive results in two or more tar
+members with the same name, which may produce nondeterministic behavior
+during multi-threaded extraction. @xref{mt-extraction}.
+
+@item -t
+@itemx --list
+List the contents of an archive. If @var{files} are given, list only the
+@var{files} given.
+
+@item -x
+@itemx --extract
+Extract files from an archive. If @var{files} are given, extract only the
+@var{files} given. Else extract all the files in the archive. To extract a
+directory without extracting the files under it, use
+@w{@samp{tarlz -xf foo --exclude='dir/*' dir}}. Tarlz removes files and
+empty directories unconditionally before extracting over them. Other than
+that, it will not make any special effort to extract a file over an
+incompatible type of file. For example, extracting a file over a non-empty
+directory will usually fail.
+
+@item -z
+@itemx --compress
+Compress existing POSIX tar archives aligning the lzip members to the tar
+members with choice of granularity (---bsolid by default, ---dsolid works
+like ---asolid). The input archives are kept unchanged. Existing compressed
+archives are not overwritten. A hyphen @samp{-} used as the name of an input
+archive reads from standard input and writes to standard output (unless the
+option @samp{--output} is used). Tarlz can be used as compressor for GNU tar
+using a command like @w{@samp{tar -c -Hustar foo | tarlz -z -o foo.tar.lz}}.
+Note that tarlz only works reliably on archives without global headers, or
+with global headers whose content can be ignored.
+
+The compression is reversible, including any garbage present after the
+end-of-archive blocks. Tarlz stops parsing after the first end-of-archive
+block is found, and then compresses the rest of the archive. Unless solid
+compression is requested, the end-of-archive blocks are compressed in a lzip
+member separated from the preceding members and from any non-zero garbage
+following the end-of-archive blocks. @samp{--compress} implies plzip
+argument style, not tar style. Each input archive is compressed to a file
+with the extension @samp{.lz} added unless the option @samp{--output} is
+used. When @samp{--output} is used, only one input archive can be specified.
+@samp{-f} can't be used with @samp{--compress}.
+
+@item --check-lib
+Compare the
+@uref{http://www.nongnu.org/lzip/manual/lzlib_manual.html#Library-version,,version of lzlib}
+used to compile tarlz with the version actually being used at run time and
+exit. Report any differences found. Exit with error status 1 if differences
+are found. A mismatch may indicate that lzlib is not correctly installed or
+that a different version of lzlib has been installed after compiling tarlz.
+Exit with error status 2 if LZ_API_VERSION and LZ_version_string don't
+match. @w{@samp{tarlz -v --check-lib}} shows the version of lzlib being used
+and the value of LZ_API_VERSION (if defined).
+@ifnothtml
+@xref{Library version,,,lzlib}.
+@end ifnothtml
+
+@end table
+
+tarlz supports the following
+@uref{http://www.nongnu.org/arg-parser/manual/arg_parser_manual.html#Argument-syntax,,options}:
+@ifnothtml
+@xref{Argument syntax,,,arg_parser}.
+@end ifnothtml
+
+@table @code
+@anchor{--data-size}
+@item -B @var{bytes}
+@itemx --data-size=@var{bytes}
+Set target size of input data blocks for the option @samp{--bsolid}.
+@xref{--bsolid}. Valid values range from @w{8 KiB} to @w{1 GiB}. Default
+value is two times the dictionary size, except for option @samp{-0} where it
+defaults to @w{1 MiB}. @xref{Minimum archive sizes}.
+
+@item -C @var{dir}
+@itemx --directory=@var{dir}
+Change to directory @var{dir}. When creating or appending, the position of
+each @samp{-C} option in the command line is significant; it will change the
+current working directory for the following @var{files} until a new
+@samp{-C} option appears in the command line. When extracting or comparing,
+all the @samp{-C} options are executed in sequence before reading the
+archive. Listing ignores any @samp{-C} options specified. @var{dir} is
+relative to the then current working directory, perhaps changed by a
+previous @samp{-C} option.
+
+Note that a process can only have one current working directory (CWD).
+Therefore multi-threading can't be used to create an archive if a @samp{-C}
+option appears after a relative file name in the command line.
 
 @item -f @var{archive}
 @itemx --file=@var{archive}
@@ -266,14 +344,6 @@ argument reads from standard input or writes to standard output.
 Follow symbolic links during archive creation, appending or comparison.
 Archive or compare the files they point to instead of the links themselves.
 
-@item --mtime=@var{date}
-When creating or appending, use @var{date} as the modification time for
-files added to the archive instead of their actual modification times. The
-value of @var{date} may be either @samp{@@} followed by the number of
-seconds since the epoch, or a date in format @w{@samp{YYYY-MM-DD HH:MM:SS}},
-or the name of an existing file starting with @samp{.} or @samp{/}. In the
-latter case, the modification time of that file is used.
-
 @item -n @var{n}
 @itemx --threads=@var{n}
 Set the number of (de)compression threads, overriding the system's default.
@@ -305,64 +375,11 @@ permissions specified in the archive.
 @itemx --quiet
 Quiet operation. Suppress all messages.
 
-@item -r
-@itemx --append
-Append files to the end of an archive. The archive must be a regular
-(seekable) file either compressed or uncompressed. Compressed members can't
-be appended to an uncompressed archive, nor vice versa. If the archive is
-compressed, it must be a multimember lzip file with the two end-of-file
-blocks plus any zero padding contained in the last lzip member of the
-archive. It is possible to append files to an archive with a different
-compression granularity. Appending works as follows; first the end-of-file
-blocks are removed, then the new members are appended, and finally two new
-end-of-file blocks are appended to the archive. If the archive is
-uncompressed, tarlz parses and skips tar headers until it finds the
-end-of-file blocks. Exit with status 0 without modifying the archive if no
-@var{files} have been specified.
-
-@item -t
-@itemx --list
-List the contents of an archive. If @var{files} are given, list only the
-@var{files} given.
-
 @item -v
 @itemx --verbose
 Verbosely list files processed. Further -v's (up to 4) increase the
 verbosity level.
 
-@item -x
-@itemx --extract
-Extract files from an archive. If @var{files} are given, extract only the
-@var{files} given. Else extract all the files in the archive. To extract a
-directory without extracting the files under it, use
-@w{@samp{tarlz -xf foo --exclude='dir/*' dir}}. Tarlz will not make any
-special effort to extract a file over an incompatible type of file. For
-example, extracting a link over a directory will usually fail. (Principle of
-least surprise).
-
-@item -z
-@itemx --compress
-Compress existing POSIX tar archives aligning the lzip members to the tar
-members with choice of granularity (---bsolid by default, ---dsolid works
-like ---asolid). The input archives are kept unchanged. Existing compressed
-archives are not overwritten. A hyphen @samp{-} used as the name of an input
-archive reads from standard input and writes to standard output (unless the
-option @samp{--output} is used). Tarlz can be used as compressor for GNU tar
-using a command like @w{@samp{tar -c -Hustar foo | tarlz -z -o foo.tar.lz}}.
-Note that tarlz only works reliably on archives without global headers, or
-with global headers whose content can be ignored.
-
-The compression is reversible, including any garbage present after the EOF
-blocks. Tarlz stops parsing after the first EOF block is found, and then
-compresses the rest of the archive. Unless solid compression is requested,
-the EOF blocks are compressed in a lzip member separated from the preceding
-members and from any non-zero garbage following the EOF blocks.
-@samp{--compress} implies plzip argument style, not tar style. Each input
-archive is compressed to a file with the extension @samp{.lz} added unless
-the option @samp{--output} is used. When @samp{--output} is used, only one
-input archive can be specified. @samp{-f} can't be used with
-@samp{--compress}.
-
 @item -0 .. -9
 Set the compression level for @samp{--create}, @samp{--append}, and
 @samp{--compress}. The default compression level is @samp{-6}. Like lzip,
@@ -392,7 +409,7 @@ appended to an uncompressed archive, nor vice versa.
 @item --asolid
 When creating or appending to a compressed archive, use appendable solid
 compression. All the files being added to the archive are compressed into a
-single lzip member, but the end-of-file blocks are compressed into a
+single lzip member, but the end-of-archive blocks are compressed into a
 separate lzip member. This creates a solidly compressed appendable archive.
 Solid archives can't be created nor decoded in parallel.
 
@@ -405,58 +422,85 @@ compressed data block must contain an integer number of tar members. Block
 compression is the default because it improves compression ratio for
 archives with many files smaller than the block size. This option allows
 tarlz revert to default behavior if, for example, it is invoked through an
-alias like @samp{tar='tarlz --solid'}. @xref{--data-size}, to set the target
-block size.
+alias like @w{@samp{tar='tarlz --solid'}}. @xref{--data-size}, to set the
+target block size.
 
 @item --dsolid
 When creating or appending to a compressed archive, compress each file
 specified in the command line separately in its own lzip member, and use
 solid compression for each directory specified in the command line. The
-end-of-file blocks are compressed into a separate lzip member. This creates
-a compressed appendable archive with a separate lzip member for each file or
-top-level directory specified.
+end-of-archive blocks are compressed into a separate lzip member. This
+creates a compressed appendable archive with a separate lzip member for each
+file or top-level directory specified.
 
 @item --no-solid
 When creating or appending to a compressed archive, compress each file
-separately in its own lzip member. The end-of-file blocks are compressed
+separately in its own lzip member. The end-of-archive blocks are compressed
 into a separate lzip member. This creates a compressed appendable archive
 with a lzip member for each file.
 
 @item --solid
 When creating or appending to a compressed archive, use solid compression.
-The files being added to the archive, along with the end-of-file blocks, are
-compressed into a single lzip member. The resulting archive is not
+The files being added to the archive, along with the end-of-archive blocks,
+are compressed into a single lzip member. The resulting archive is not
 appendable. No more files can be later appended to the archive. Solid
 archives can't be created nor decoded in parallel.
 
 @item --anonymous
-Equivalent to @samp{--owner=root --group=root}.
+Equivalent to @w{@samp{--owner=root --group=root}}.
 
 @item --owner=@var{owner}
-When creating or appending, use @var{owner} for files added to the
-archive. If @var{owner} is not a valid user name, it is decoded as a
-decimal numeric user ID.
+When creating or appending, use @var{owner} for files added to the archive.
+If @var{owner} is not a valid user name, it is decoded as a decimal numeric
+user ID.
 
 @item --group=@var{group}
-When creating or appending, use @var{group} for files added to the
-archive. If @var{group} is not a valid group name, it is decoded as a
-decimal numeric group ID.
+When creating or appending, use @var{group} for files added to the archive.
+If @var{group} is not a valid group name, it is decoded as a decimal numeric
+group ID.
+
+@item --exclude=@var{pattern}
+Exclude files matching a shell pattern like @samp{*.o}. A file is considered
+to match if any component of the file name matches. For example, @samp{*.o}
+matches @samp{foo.o}, @samp{foo.o/bar} and @samp{foo/bar.o}. If
+@var{pattern} contains a @samp{/}, it matches a corresponding @samp{/} in
+the file name. For example, @samp{foo/*.o} matches @samp{foo/bar.o}.
+Multiple @samp{--exclude} options can be specified.
+
+@item --ignore-ids
+Make @samp{--diff} ignore differences in owner and group IDs. This option is
+useful when comparing an @samp{--anonymous} archive.
+
+@item --ignore-overflow
+Make @samp{--diff} ignore differences in mtime caused by overflow on 32-bit
+systems with a 32-bit time_t.
 
 @item --keep-damaged
 Don't delete partially extracted files. If a decompression error happens
 while extracting a file, keep the partial data extracted. Use this option to
 recover as much data as possible from each damaged member. It is recommended
-to run tarlz in single-threaded mode (--threads=0) when using this option.
+to run tarlz in single-threaded mode (---threads=0) when using this option.
 
 @item --missing-crc
-Exit with error status 2 if the CRC of the extended records is missing.
-When this option is used, tarlz detects any corruption in the extended
-records (only limited by CRC collisions). But note that a corrupt
-@samp{GNU.crc32} keyword, for example @samp{GNU.crc33}, is reported as a
-missing CRC instead of as a corrupt record. This misleading
-@samp{Missing CRC} message is the consequence of a flaw in the POSIX pax
-format; i.e., the lack of a mandatory check sequence in the extended
-records. @xref{crc32}.
+Exit with error status 2 if the CRC of the extended records is missing. When
+this option is used, tarlz detects any corruption in the extended records
+(only limited by CRC collisions). But note that a corrupt @samp{GNU.crc32}
+keyword, for example @samp{GNU.crc33}, is reported as a missing CRC instead
+of as a corrupt record. This misleading @w{@samp{Missing CRC}} message is
+the consequence of a flaw in the POSIX pax format; i.e., the lack of a
+mandatory check sequence of the extended records. @xref{crc32}.
+
+@item --mtime=@var{date}
+When creating or appending, use @var{date} as the modification time for
+files added to the archive instead of their actual modification times. The
+value of @var{date} may be either @samp{@@} followed by the number of
+seconds since (or before) the epoch, or a date in format
+@w{@samp{[-]YYYY-MM-DD HH:MM:SS}} or @samp{[-]YYYY-MM-DDTHH:MM:SS}, or the
+name of an existing reference file starting with @samp{.} or @samp{/} whose
+modification time is used. The time of day @samp{HH:MM:SS} in the date
+format is optional and defaults to @samp{00:00:00}. The epoch is
+@w{@samp{1970-01-01 00:00:00 UTC}}. Negative seconds or years define a
+modification time before the epoch.
 
 @item --out-slots=@var{n}
 Number of @w{1 MiB} output packets buffered per worker thread during
@@ -465,20 +509,6 @@ number of packets may increase compression speed if the files being archived
 are larger than @w{64 MiB} compressed, but requires more memory. Valid
 values range from 1 to 1024. The default value is 64.
 
-@item --check-lib
-Compare the
-@uref{http://www.nongnu.org/lzip/manual/lzlib_manual.html#Library-version,,version of lzlib}
-used to compile tarlz with the version actually being used at run time and
-exit. Report any differences found. Exit with error status 1 if differences
-are found. A mismatch may indicate that lzlib is not correctly installed or
-that a different version of lzlib has been installed after compiling tarlz.
-Exit with error status 2 if LZ_API_VERSION and LZ_version_string don't
-match. @w{@samp{tarlz -v --check-lib}} shows the version of lzlib being used
-and the value of LZ_API_VERSION (if defined).
-@ifnothtml
-@xref{Library version,,,lzlib}.
-@end ifnothtml
-
 @item --warn-newer
 During archive creation, warn if any file being archived has a modification
 time newer than the archive creation time. This option may slow archive
@@ -496,10 +526,10 @@ keyword appearing in the same block of extended records.
 
 @end table
 
-Exit status: 0 for a normal exit, 1 for environmental problems (file not
-found, files differ, invalid flags, I/O errors, etc), 2 to indicate a
-corrupt or invalid input file, 3 for an internal consistency error (e.g. bug)
-which caused tarlz to panic.
+Exit status: 0 for a normal exit, 1 for environmental problems
+(file not found, files differ, invalid command line options, I/O errors,
+etc), 2 to indicate a corrupt or invalid input file, 3 for an internal
+consistency error (e.g., bug) which caused tarlz to panic.
 
 
 @node Portable character set
@@ -518,13 +548,7 @@ The last three characters are the period, underscore, and hyphen-minus
 characters, respectively.
 
 File names are identifiers. Therefore, archiving works better when file
-names use only the portable character set without spaces added. Unicode is
-for human consumption. It should be
-@uref{http://www.gnu.org/software/moe/manual/moe_manual.html#why-not-Unicode,,avoided}
-in computing environments, specially in file names.
-@ifnothtml
-@xref{why not Unicode,,,moe}.
-@end ifnothtml
+names use only the portable character set without spaces added.
 
 
 @node File format
@@ -552,29 +576,30 @@ bytes (for example 512).
 
 @sp 1
 A tar.lz file consists of a series of lzip members (compressed data sets).
-The members simply appear one after another in the file, with no
-additional information before, between, or after them.
+The members simply appear one after another in the file, with no additional
+information before, between, or after them.
 
-Each lzip member contains one or more tar members in a simplified POSIX
-pax interchange format. The only pax typeflag value supported by tarlz
-(in addition to the typeflag values defined by the ustar format) is
-@samp{x}. The pax format is an extension on top of the ustar format that
-removes the size limitations of the ustar format.
+Each lzip member contains one or more tar members in a simplified POSIX pax
+interchange format. The only pax typeflag value supported by tarlz (in
+addition to the typeflag values defined by the ustar format) is @samp{x}.
+The pax format is an extension on top of the ustar format that removes the
+size limitations of the ustar format.
 
 Each tar member contains one file archived, and is represented by the
 following sequence:
 
 @itemize @bullet
 @item
-An optional extended header block with extended header records. This
-header block is of the form described in pax header block, with a
-typeflag value of @samp{x}. The extended header records are included as
-the data for this header block.
+An optional extended header block followed by one or more blocks that
+contain the extended header records as if they were the contents of a file;
+i.e., the extended header records are included as the data for this header
+block. This header block is of the form described in pax header block, with
+a typeflag value of @samp{x}.
 
 @item
-A header block in ustar format that describes the file. Any fields
-defined in the preceding optional extended header records override the
-associated fields in this header block for this file.
+A header block in ustar format that describes the file. Any fields defined
+in the preceding optional extended header records override the associated
+fields in this header block for this file.
 
 @item
 Zero or more blocks that contain the contents of the file.
@@ -586,9 +611,11 @@ is split over two or more lzip members, the archive must be decoded
 sequentially. @xref{Multi-threaded decoding}.
 
 At the end of the archive file there are two 512-byte blocks filled with
-binary zeros, interpreted as an end-of-archive indicator. These EOF
-blocks are either compressed in a separate lzip member or compressed
-along with the tar members contained in the last lzip member.
+binary zeros, interpreted as an end-of-archive indicator. These EOA blocks
+are either compressed in a separate lzip member or compressed along with the
+tar members contained in the last lzip member. For a compressed archive to
+be recognized by tarlz as appendable, the last lzip member must contain
+between 512 and 32256 zeros alone.
 
 The diagram below shows the correspondence between each tar member (formed
 by one or two headers plus optional data) in the tar archive and each
@@ -602,7 +629,7 @@ used:
 @verbatim
 tar
 +========+======+=================+===============+========+======+========+
-| header | data | extended header | extended data | header | data |   EOF  |
+| header | data | extended header | extended data | header | data |   EOA  |
 +========+======+=================+===============+========+======+========+
 
 tar.lz
@@ -620,7 +647,7 @@ Similarly, if several records with the same keyword appear in the same
 block of extended records, only the last record for the repeated keyword
 takes effect. The other records for the repeated keyword are ignored.@*
 A global header inserted between an extended header and an ustar header.@*
-An extended header just before the EOF blocks.
+An extended header just before the end-of-archive blocks.
 @end ignore
 
 @sp 1
@@ -635,7 +662,7 @@ archive extraction. @xref{flawed-compat}.
 
 The pax extended header data consists of one or more records, each of
 them constructed as follows:@*
-@samp{"%d %s=%s\n", <length>, <keyword>, <value>}
+@w{@samp{"%d %s=%s\n", <length>, <keyword>, <value>}}
 
 The fields <length> and <keyword> in the record must be limited to the
 portable character set (@pxref{Portable character set}). The field <length>
@@ -647,25 +674,53 @@ space, equal-sign, and newline.
 These are the <keyword> values currently supported by tarlz:
 
 @table @code
+@item atime
+The signed decimal representation of the access time of the following file
+in seconds since (or before) the epoch, obtained from the function
+@samp{stat}. The atime record is created only for files with a modification
+time outside of the ustar range. @xref{ustar-mtime}.
+
+@item gid
+The unsigned decimal representation of the group ID of the group that owns
+the following file. The gid record is created only for files with a group ID
+greater than 2_097_151 (octal 7777777). @xref{ustar-uid-gid}.
+
 @item linkpath
-The pathname of a link being created to another file, of any type,
+The file name of a link being created to another file, of any type,
 previously archived. This record overrides the field @samp{linkname} in the
-following ustar header block. The following ustar header block
-determines the type of link created. If typeflag of the following header
-block is 1, it will be a hard link. If typeflag is 2, it will be a
-symbolic link and the linkpath value will be used as the contents of the
-symbolic link.
+following ustar header block. The following ustar header block determines
+the type of link created. If typeflag of the following header block is 1, it
+will be a hard link. If typeflag is 2, it will be a symbolic link and the
+linkpath value will be used as the contents of the symbolic link. The
+linkpath record is created only for links with a link name that does not fit
+in the space provided by the ustar header.
+
+@item mtime
+The signed decimal representation of the modification time of the following
+file in seconds since (or before) the epoch, obtained from the function
+@samp{stat}. This record overrides the field @samp{mtime} in the following
+ustar header block. The mtime record is created only for files with a
+modification time outside of the ustar range. @xref{ustar-mtime}.
 
 @item path
-The pathname of the following file. This record overrides the fields
-@samp{name} and @samp{prefix} in the following ustar header block.
+The file name of the following file. This record overrides the fields
+@samp{name} and @samp{prefix} in the following ustar header block. The path
+record is created for files with a name that does not fit in the space
+provided by the ustar header, but is also created for files that require any
+other extended record so that the fields @samp{name} and @samp{prefix} in
+the following ustar header block can be zeroed.
 
 @item size
-The size of the file in bytes, expressed as a decimal number using
-digits from the ISO/IEC 646:1991 (ASCII) standard. This record overrides
-the size field in the following ustar header block. The size record is
-used only for files with a size value greater than 8_589_934_591
-@w{(octal 77777777777)}. This is 2^33 bytes or larger.
+The size of the file in bytes, expressed as a decimal number using digits
+from the ISO/IEC 646:1991 (ASCII) standard. This record overrides the field
+@samp{size} in the following ustar header block. The size record is created
+only for files with a size value greater than 8_589_934_591
+@w{(octal 77777777777)}; that is, @w{8 GiB} (2^33 bytes) or larger.
+
+@item uid
+The unsigned decimal representation of the user ID of the file owner of the
+following file. The uid record is created only for files with a user ID
+greater than 2_097_151 (octal 7777777). @xref{ustar-uid-gid}.
 
 @anchor{key_crc32}
 @item GNU.crc32
@@ -722,18 +777,18 @@ The fields @samp{name}, @samp{linkname}, and @samp{prefix} are
 null-terminated character strings except when all characters in the array
 contain non-null characters including the last character.
 
-The fields @samp{prefix} and @samp{name} produce the pathname of the file. A
-new pathname is formed, if prefix is not an empty string (its first
-character is not null), by concatenating prefix (up to the first null
-character), a slash character, and name; otherwise, name is used alone. In
-either case, name is terminated at the first null character. If prefix
-begins with a null character, it is ignored. In this manner, pathnames of at
-most 256 characters can be supported. If a pathname does not fit in the
-space provided, an extended record is used to store the pathname.
+The fields @samp{name} and @samp{prefix} produce the file name. A new file
+name is formed, if prefix is not an empty string (its first character is not
+null), by concatenating prefix (up to the first null character), a slash
+character, and name; otherwise, name is used alone. In either case, name is
+terminated at the first null character. If prefix begins with a null
+character, it is ignored. In this manner, file names of at most 256
+characters can be supported. If a file name does not fit in the space
+provided, an extended record is used to store the file name.
 
-The field @samp{linkname} does not use the prefix to produce a pathname. If
-the linkname does not fit in the 100 characters provided, an extended record
-is used to store the linkname.
+The field @samp{linkname} does not use the prefix to produce a file name. If
+the link name does not fit in the 100 characters provided, an extended
+record is used to store the link name.
 
 The field @samp{mode} provides 12 access permission bits. The following
 table shows the symbolic name of each bit and its octal value:
@@ -746,8 +801,10 @@ table shows the symbolic name of each bit and its octal value:
 @item S_IROTH @tab 00004 @tab S_IWOTH @tab 00002 @tab S_IXOTH @tab 00001
 @end multitable
 
+@anchor{ustar-uid-gid}
 The fields @samp{uid} and @samp{gid} are the user and group IDs of the owner
-and group of the file, respectively.
+and group of the file, respectively. If the file uid or gid are greater than
+2_097_151 (octal 7777777), an extended record is used to store the uid or gid.
 
 The field @samp{size} contains the octal representation of the size of the
 file in bytes. If the field @samp{typeflag} specifies a file of type '0'
@@ -758,9 +815,13 @@ to 0 or ignores it, and does not store or expect any logical records
 following the header. If the file size is larger than 8_589_934_591 bytes
 @w{(octal 77777777777)}, an extended record is used to store the file size.
 
+@anchor{ustar-mtime}
 The field @samp{mtime} contains the octal representation of the modification
 time of the file at the time it was archived, obtained from the function
-@samp{stat}.
+@samp{stat}. If the modification time is negative or larger than
+8_589_934_591 @w{(octal 77777777777)} seconds since the epoch, an extended
+record is used to store the modification time. The ustar range of mtime goes
+from @w{@samp{1970-01-01 00:00:00 UTC}} to @w{@samp{2242-03-16 12:56:31 UTC}}.
 
 The field @samp{chksum} contains the octal representation of the value of
 the simple sum of all bytes in the header logical record. Each byte in the
@@ -775,7 +836,8 @@ file archived:
 Regular file.
 
 @item '1'
-Hard link to another file, of any type, previously archived.
+Hard link to another file, of any type, previously archived. Hard links must
+not contain file data.
 
 @item '2'
 Symbolic link.
@@ -792,9 +854,9 @@ Directory.
 FIFO special file.
 
 @item '7'
-Reserved to represent a file to which an implementation has associated
-some high-performance attribute. Tarlz treats this type of file as a
-regular file (type 0).
+Reserved to represent a file to which an implementation has associated some
+high-performance attribute (contiguous file). Tarlz treats this type of file
+as a regular file (type 0).
 
 @end table
 
@@ -817,9 +879,9 @@ Tarlz creates safe archives that allow the reliable detection of invalid or
 corrupt metadata during decoding even when the integrity checking of lzip
 can't be used because the lzip members are only decompressed partially, as
 it happens in parallel @samp{--diff}, @samp{--list}, and @samp{--extract}.
-In order to achieve this goal, tarlz makes some changes to the variant of
-the pax format that it uses. This chapter describes these changes and the
-concrete reasons to implement them.
+In order to achieve this goal and avoid some other flaws in the pax format,
+tarlz makes some changes to the variant of the pax format that it uses. This
+chapter describes these changes and the concrete reasons to implement them.
 
 @sp 1
 @anchor{crc32}
@@ -857,23 +919,28 @@ to the POSIX-2:1993 standard, POSIX.1-2008 recommends selecting extended
 header field values that allow such tar to create a regular file containing
 the extended header records as data. This approach is broken because if the
 extended header is needed because of a long file name, the fields
-@samp{prefix} and @samp{name} will be unable to contain the full pathname of
-the file. Therefore the files corresponding to both the extended header and
-the overridden ustar header will be extracted using truncated file names,
-perhaps overwriting existing files or directories. It may be a security risk
-to extract a file with a truncated file name.
+@samp{name} and @samp{prefix} will be unable to contain the full file name.
+(Some tar implementations store the truncated name in the field @samp{name}
+alone, truncating the name to only 100 bytes instead of 256). Therefore the
+files corresponding to both the extended header and the overridden ustar
+header will be extracted using truncated file names, perhaps overwriting
+existing files or directories. It may be a security risk to extract a file
+with a truncated file name.
 
 To avoid this problem, tarlz writes extended headers with all fields zeroed
-except size, chksum, typeflag, magic and version. This prevents old tar
-programs from extracting the extended records as a file in the wrong place.
-Tarlz also sets to zero those fields of the ustar header overridden by
-extended records. Finally, tarlz skips members without name when decoding
-except when listing. This is needed to detect certain format violations
-during parallel extraction.
-
-If an extended header is required for any reason (for example a file size
-larger than @w{8 GiB} or a link name longer than 100 bytes), tarlz moves the
-file name also to the extended header to prevent an ustar tool from trying
+except @samp{size} (which contains the size of the extended records),
+@samp{chksum}, @samp{typeflag}, @samp{magic}, and @samp{version}. In
+particular, tarlz sets the fields @samp{name} and @samp{prefix} to zero.
+This prevents old tar programs from extracting the extended records as a
+file in the wrong place. Tarlz also sets to zero those fields of the ustar
+header overridden by extended records. Finally, tarlz skips members with
+zeroed @samp{name} and @samp{prefix} when decoding, except when listing.
+This is needed to detect certain format violations during parallel
+extraction.
+
+If an extended header is required for any reason (for example a file size of
+@w{8 GiB} or larger, or a link name longer than 100 bytes), tarlz also moves
+the file name to the extended records to prevent an ustar tool from trying
 to extract the file or link. This also makes easier during parallel decoding
 the detection of a tar member split between two lzip members at the boundary
 between the extended header and the ustar header.
@@ -882,17 +949,39 @@ between the extended header and the ustar header.
 @section As simple as possible (but not simpler)
 
 The tarlz format is mainly ustar. Extended pax headers are used only when
-needed because the length of a file name or link name, or the size of a file
-exceed the limits of the ustar format. Adding @w{1 KiB} of extended headers
-to each member just to record subsecond timestamps seems wasteful for a
-backup format. Moreover, minimizing the overhead may help recovering the
-archive with lziprecover in case of corruption.
+needed because the length of a file name or link name, or the size or other
+attribute of a file exceed the limits of the ustar format. Adding @w{1 KiB}
+of extended header and records to each member just to save subsecond
+timestamps seems wasteful for a backup format. Moreover, minimizing the
+overhead may help recovering the archive with lziprecover in case of
+corruption.
 
 Global pax headers are tolerated, but not supported; they are parsed and
 ignored. Some operations may not behave as expected if the archive contains
 global headers.
 
 @sp 1
+@section Improve reproducibility
+
+Pax includes by default the process ID of the pax process in the ustar name
+of the extended headers, making the archive not reproducible. Tarlz stores
+the true name of the file just once, either in the ustar header or in the
+extended records, making it easier to produce reproducible archives.
+
+Pax allows an extended record to have length x-1 or x if x is a power of
+ten; @samp{99<97_bytes>} or @samp{100<97_bytes>}. Tarlz minimizes the length
+of the record and always produces a length of x-1 in these cases.
+
+@sp 1
+@section No data in hard links
+
+Tarlz does not allow data in hard link members. The data (if any) must be in
+the member determining the type of the file (which can't be a link). If all
+the names of a file are stored as hard links, the type of the file is lost.
+Not allowing data in hard links also prevents invalid actions like
+extracting file data for a hard link to a symbolic link or to a directory.
+
+@sp 1
 @section Avoid misconversions to/from UTF-8
 
 There is no portable way to tell what charset a text string is coded into.
@@ -1047,17 +1136,18 @@ corresponding to the tar member header. This is another reason why the tar
 headers must provide their own integrity checking.
 
 @sp 1
+@anchor{mt-extraction}
 @section Limitations of multi-threaded extraction
 
 Multi-threaded extraction may produce different output than single-threaded
 extraction in some cases:
 
-During multi-threaded extraction, several independent processes are
-simultaneously reading the archive and creating files in the file system. The
-archive is not read sequentially. As a consequence, any error or weirdness
-in the archive (like a corrupt member or an EOF block in the middle of the
-archive) won't be usually detected until part of the archive beyond that
-point has been processed.
+During multi-threaded extraction, several independent threads are
+simultaneously reading the archive and creating files in the file system.
+The archive is not read sequentially. As a consequence, any error or
+weirdness in the archive (like a corrupt member or an end-of-archive block
+in the middle of the archive) won't be usually detected until part of the
+archive beyond that point has been processed.
 
 If the archive contains two or more tar members with the same name,
 single-threaded extraction extracts the members in the order they appear in
@@ -1070,6 +1160,9 @@ If the same file is extracted through several paths (different member names
 resolve to the same file in the file system), the result is undefined.
 (Probably the resulting file will be mangled).
 
+Extraction of a hard link may fail if it is extracted before the file it
+links to.
+
 
 @node Minimum archive sizes
 @chapter Minimum archive sizes required for multi-threaded block compression
@@ -1123,8 +1216,8 @@ tarlz -cf archive.tar.lz a b c
 
 @sp 1
 @noindent
-Example 2: Append files @samp{d} and @samp{e} to the multimember
-compressed archive @samp{archive.tar.lz}.
+Example 2: Append files @samp{d} and @samp{e} to the multimember compressed
+archive @samp{archive.tar.lz}.
 
 @example
 tarlz -rf archive.tar.lz d e
@@ -1148,7 +1241,7 @@ Example 4: Create a compressed appendable archive containing directories
 directory. Then append files @samp{a}, @samp{b}, @samp{c}, @samp{d} and
 @samp{e} to the archive, all of them contained in a single lzip member.
 The resulting archive @samp{archive.tar.lz} contains 5 lzip members
-(including the EOF member).
+(including the end-of-archive member).
 
 @example
 tarlz --dsolid -cf archive.tar.lz dir1 dir2 dir3
@@ -1184,11 +1277,11 @@ tarlz -xf archive.tar.lz a c dir1
 
 @sp 1
 @noindent
-Example 8: Copy the contents of directory @samp{sourcedir} to the
-directory @samp{destdir}.
+Example 8: Copy the contents of directory @samp{sourcedir} to the directory
+@samp{destdir}.
 
 @example
-tarlz -C sourcedir -c . | tarlz -C destdir -x
+tarlz -C sourcedir --uncompressed -cf - . | tarlz -C destdir -xf -
 @end example
 
 @sp 1
@@ -1202,6 +1295,24 @@ the other members can still be extracted).
 tarlz -z --no-solid archive.tar
 @end example
 
+@sp 1
+@noindent
+Example 10: Compress the archive @samp{archive.tar} and write the output to
+@samp{foo.tar.lz}.
+
+@example
+tarlz -z -o foo.tar.lz archive.tar
+@end example
+
+@sp 1
+@noindent
+Example 11: Concatenate and compress two archives @samp{archive1.tar} and
+@samp{archive2.tar}, and write the output to @samp{foo.tar.lz}.
+
+@example
+tarlz -A archive1.tar archive2.tar | tarlz -z -o foo.tar.lz
+@end example
+
 
 @node Problems
 @chapter Reporting bugs