Log Message:
coreutils: Update to 9.0

Changelog:
* Noteworthy changes in release 9.0 (2021-09-24) [stable]

** Bug fixes

  chmod -v no longer misreports modes of dangling symlinks.
  [bug introduced in coreutils-5.3.0]

  cp -a --attributes-only now never removes destination files,
  even if the destination files are hardlinked, or the source
  is a non regular file.
  [bug introduced in coreutils-8.6]

  csplit --suppress-matched now elides the last matched line
  when a specific number of pattern matches are performed.
  [bug introduced with the --suppress-matched feature in coreutils-8.22]

  df no longer outputs duplicate remote mounts in the presence of bind mounts.
  [bug introduced in coreutils-8.26]

  df no longer mishandles command-line args that it pre-mounts
  [bug introduced in coreutils-8.29]

  du no longer crashes on XFS file systems when the directory hierarchy is
  heavily changed during the run.
  [bug introduced in coreutils-8.25]

  env -S no longer crashes when given unusual whitespace characters
  [bug introduced in coreutils-8.30]

  expr no longer mishandles unmatched \(...\) in regular expressions.
  [bug introduced in coreutils-6.0]

  ls no longer crashes when printing the SELinux context for unstatable files.
  [bug introduced in coreutils-6.9.91]

  mkdir -m no longer mishandles modes more generous than the umask.
  [bug introduced in coreutils-8.22]

  nl now handles single character --section-delimiter arguments,
  by assuming a second ':' character has been specified, as specified by POSIX.
  [This bug was present in "the beginning".]

  pr again adjusts tabs in input, to maintain alignment in multi column output.
  [bug introduced in coreutils-6.9]

  rm no longer skips an extra file when the removal of an empty directory fails.
  [bug introduced by the rewrite to use fts in coreutils-8.0]

  split --number=K/N will again correctly split chunk K of N to stdout.
  Previously a chunk starting after 128KiB, output the wrong part of the file.
  [bug introduced in coreutils-8.26]

  tail -f no longer overruns a stack buffer when given too many files
  to follow and ulimit -n exceeds 1024.
  [bug introduced in coreutils-7.5]

  tr no longer crashes when using --complement with certain
  invalid combinations of case character classes.
  [bug introduced in coreutils-8.6]

  basenc --base64 --decode no longer silently discards decoded characters
  on (1024*5) buffer boundaries
  [bug introduced in coreutils-8.31]

** Changes in behavior

  cp and install now default to copy-on-write (COW) if available.

  cp, install and mv now use the copy_file_range syscall if available.
  Also, they use lseek+SEEK_HOLE rather than ioctl+FS_IOC_FIEMAP on sparse
  files, as lseek is simpler and more portable.

  On GNU/Linux systems, ls no longer issues an error message on a
  directory merely because it was removed.  This reverts a change
  that was made in release 8.32.

  ptx -T no longer attempts to substitute old-fashioned TeX escapes
  for 8-bit non-ASCII alphabetic characters.  TeX indexes should
  instead use '\usepackage[latin1]{inputenc}' or equivalent.

  stat will use decomposed (major,minor) device numbers in its default format.
  This is less ambiguous, and more consistent with ls.

  sum [-r] will output a file name, even if only a single name is passed.
  This is consistent with sum -s, cksum, and other sum(1) implementations.

** New Features

  cksum now supports the -a (--algorithm) option to select any
  of the existing sum, md5sum, b2sum, sha*sum implementations etc.
  cksum now subsumes all of these programs, and coreutils
  will introduce no future standalone checksum utility.

  cksum -a now supports the 'sm3' argument, to use the SM3 digest algorithm.

  cksum --check now supports auto detecting the digest type to use,
  when verifying tagged format checksums.

  expr and factor now support bignums on all platforms.

  ls --classify now supports the "always", "auto", or \ 
"never" flags,
  to support only outputting classifier characters if connected to a tty.

  ls now accepts the --sort=width option, to sort by file name width.
  This is useful to more compactly organize the default vertical column output.

  ls now accepts the --zero option, to terminate each output line with
  NUL instead of newline.

  nl --line-increment can now take a negative number to decrement the count.

  stat supports more formats for representing decomposed device numbers.
  %Hd,%Ld and %Hr,%Lr will output major,minor device numbers and device types
  respectively.  %d corresponds to st_dev and %r to std_rdev.

** Improvements

  cat --show-ends will now show \r\n as ^M$.  Previously the \r was taken
  literally, thus overwriting the first character in the line with '$'.

  cksum [-a crc] is now up to 4 times faster by using a slice by 8 algorithm,
  and at least 8 times faster where pclmul instructions are supported.
  A new --debug option will indicate if pclmul is being used.

  md5sum --check now supports checksum files with CRLF line endings.
  This also applies to cksum, sha*sum, and b2sum.

  df now recognizes these file systems as remote:
  acfs, coda, fhgfs, gpfs, ibrix, ocfs2, and vxfs.

  rmdir now clarifies the error if a symlink_to_dir/ has not been traversed.
  This is the case on GNU/Linux systems, where the trailing slash is ignored.

  stat and tail now know about the "devmem", "exfat", \ 
"secretmem", "vboxsf",
  and "zonefs" file system types.  stat -f -c%T now reports the file system
  type, and tail -f uses polling for "vboxsf" and inotify for the others.

  timeout now supports sub-second timeouts on macOS.

  wc is up to 5 times faster when counting only new line characters,
  where avx2 instructions are supported.
  A new --debug option will indicate if avx2 is being used.