please test: coreutils-5.93-1

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view

please test: coreutils-5.93-1

Eric Blake (cygwin)
Hash: SHA1

I've uploaded a test version of coreutils, 5.93-1.  This is a new, stable
upstream release, with a number of changes from 5.3.0.  However, this
build depends on features that have been added to cygwin since
cygwin-1.5.18-1 (such as getline and futimes), so I am leaving 5.3.0-9 as
current until cygwin-1.5.19 is released.  To use this release, you MUST
install a recent snapshot of cygwin (20051020 or later).  If you don't
know what this means, then stick with 5.3.0-9.  A list of changes from the
NEWS file appears below; see also /usr/share/doc/coreutils-5.93/.

This version also has a new cygwin-specific --append-exe option to ls(1)
(and dir, vdir) and stat(1); if a command-line argument does not have
.exe, but the file on the system does, then using this option will make
the listing show the .exe.  I found this addition to my ~/.bashrc useful
to use the new options (the spacing is chosen so that bash doesn't treat
the next word on the command line as an alias):

ls --append-exe -d . >/dev/null 2>&1 && append_exe=' --append-exe'
alias ls="ls {your favorite options here}$append_exe"
alias stat="stat$append_exe"
unset append_exe

- --
Life is short - so eat dessert first!

Eric Blake             [hidden email]
volunteer cygwin coreutils maintainer
Version: GnuPG v1.4.1 (Cygwin)
Comment: Public key at
Comment: Using GnuPG with Thunderbird -


* Major changes in release 5.93 (2005-11-06) [stable]

** Bug fixes

  dircolors no longer segfaults upon an attempt to use the new

  du no longer overflows a counter when processing a file larger than
  2^31-1 on some 32-bit systems (at least some AIX 5.1 configurations).

  md5sum once again defaults to using the ` ' non-binary marker
  (rather than the `*' binary marker) by default on Unix-like systems.

  mkdir -p and install -d no longer exit nonzero when asked to create
  a directory like `nonexistent/.'

  rm emits a better diagnostic when (without -r) it fails to remove
  a directory on e.g., Solaris 9/10 systems.

  tac now works when stdin is a tty, even on non-Linux systems

  "tail -c 2 FILE" and "touch 0101000000" now operate as POSIX
  1003.1-2001 requires, even when coreutils is conforming to older
  POSIX standards, as the newly-required behavior is upward-compatible
  with the old.

** Build-related bug fixes

  installing .mo files would fail

* Major changes in release 5.92 (2005-10-22) [stable]

** Bug fixes

  chmod now diagnoses an invalid mode string starting with an octal digit

  dircolors now properly quotes single-quote characters

* Major changes in release 5.91 (2005-10-17) [stable candidate]

** Bug fixes

  "mkdir -p /a/b/c" no longer fails merely because a leading prefix
  directory (e.g., /a or /a/b) exists on a read-only file system.

** Removed options

  tail's --allow-missing option has been removed.  Use --retry instead.

  stat's --link and -l options have been removed.
  Use --dereference (-L) instead.

** Deprecated options

  Using ls, du, or df with the --kilobytes option now evokes a warning
  that the long-named option is deprecated.  Use `-k' instead.

  du's long-named --megabytes option now evokes a warning.
  Use -m instead.

* Major changes in release 5.90 (2005-09-29) [unstable]

** Bring back support for `head -NUM', `tail -NUM', etc. even when
  conforming to POSIX 1003.1-2001.  The following changes apply only
  when conforming to POSIX 1003.1-2001; there is no effect when
  conforming to older POSIX versions.

  The following usages now behave just as when conforming to older POSIX:

    date -I
    expand -TAB1[,TAB2,...]
    fold -WIDTH
    head -NUM
    join -j FIELD
    join -j1 FIELD
    join -j2 FIELD
    join -o FIELD_NAME1 FIELD_NAME2...
    nice -NUM
    od -w
    pr -S
    split -NUM
    tail -[NUM][bcl][f] [FILE]

  The following usages no longer work, due to the above changes:

    date -I TIMESPEC  (use `date -ITIMESPEC' instead)
    od -w WIDTH       (use `od -wWIDTH' instead)
    pr -S STRING      (use `pr -SSTRING' instead)

  A few usages still have behavior that depends on which POSIX standard is
  being conformed to, and portable applications should beware these
  problematic usages.  These include:

    Problematic       Standard-conforming replacement, depending on
       usage            whether you prefer the behavior of:
                      POSIX 1003.2-1992    POSIX 1003.1-2001
    sort +4           sort -k 5            sort ./+4
    tail +4           tail -n +4           tail ./+4
    tail - f          tail f               [see (*) below]
    tail -c 4         tail -c 10 ./4       tail -c4
    touch 12312359 f  touch -t 12312359 f  touch ./12312359 f
    uniq +4           uniq -s 4            uniq ./+4

    (*) "tail - f" does not conform to POSIX 1003.1-2001; to read
    standard input and then "f", use the command "tail -- - f".

  These changes are in response to decisions taken in the January 2005
  Austin Group standardization meeting.  For more details, please see
  "Utility Syntax Guidelines" in the Minutes of the January 2005
  Meeting <>.

** Binary input and output are now implemented more consistently.
  These changes affect only platforms like MS-DOS that distinguish
  between binary and text files.

  The following programs now always use text input/output:

    expand unexpand

  The following programs now always use binary input/output to copy data:

    cp install mv shred

  The following programs now always use binary input/output to copy
  data, except for stdin and stdout when it is a terminal.

    head tac tail tee tr
    (cat behaves similarly, unless one of the options -bensAE is used.)

  cat's --binary or -B option has been removed.  It existed only on
  MS-DOS-like platforms, and didn't work as documented there.

  md5sum and sha1sum now obey the -b or --binary option, even if
  standard input is a terminal, and they no longer report files to be
  binary if they actually read them in text mode.

** Changes for better conformance to POSIX

  cp, ln, mv, rm changes:

    Leading white space is now significant in responses to yes-or-no questions.
    For example, if "rm" asks "remove regular file `foo'?" and you respond
    with " y" (i.e., space before "y"), it counts as "no".

  dd changes:

    On a QUIT or PIPE signal, dd now exits without printing statistics.

    On hosts lacking the INFO signal, dd no longer treats the USR1
    signal as if it were INFO when POSIXLY_CORRECT is set.

    If the file F is non-seekable and contains fewer than N blocks,
    then before copying "dd seek=N of=F" now extends F with zeroed
    blocks until F contains N blocks.

  fold changes:

    When POSIXLY_CORRECT is set, "fold file -3" is now equivalent to
    "fold file ./-3", not the obviously-erroneous "fold file ./-w3".

  ls changes:

    -p now marks only directories; it is equivalent to the new option
    --indicator-style=slash.  Use --file-type or
    --indicator-style=file-type to get -p's old behavior.

  nice changes:

    Documentation and diagnostics now refer to "nicenesses" (commonly
    in the range -20...19) rather than "nice values" (commonly 0...39).

  nohup changes:

    nohup now ignores the umask when creating nohup.out.

    nohup now closes stderr if it is a terminal and stdout is closed.

    nohup now exits with status 127 (not 1) when given an invalid option.

  pathchk changes:

    It now rejects the empty name in the normal case.  That is,
    "pathchk -p ''" now fails, and "pathchk ''" fails unless the
    current host (contra POSIX) allows empty file names.

    The new -P option checks whether a file name component has leading "-",
    as suggested in interpretation "Austin-039:XCU:pathchk:pathchk -p"
    It also rejects the empty name even if the current host accepts it; see

    The --portability option is now equivalent to -p -P.

** Bug fixes

  chmod, mkdir, mkfifo, and mknod formerly mishandled rarely-used symbolic
  permissions like =xX and =u, and did not properly diagnose some invalid
  strings like g+gr, ug,+x, and +1.  These bugs have been fixed.

  csplit could produce corrupt output, given input lines longer than 8KB

  dd now computes statistics using a realtime clock (if available)
  rather than the time-of-day clock, to avoid glitches if the
  time-of-day is changed while dd is running.  Also, it avoids
  using unsafe code in signal handlers; this fixes some core dumps.

  expr and test now correctly compare integers of unlimited magnitude.

  expr now detects integer overflow when converting strings to integers,
  rather than silently wrapping around.

  ls now refuses to generate time stamps containing more than 1000 bytes, to
  foil potential denial-of-service attacks on hosts with very large stacks.

  "mkdir -m =+x dir" no longer ignores the umask when evaluating "+x",
  and similarly for mkfifo and mknod.

  "mkdir -p /tmp/a/b dir" no longer attempts to create the `.'-relative
  directory, dir (in /tmp/a), when, after creating /tmp/a/b, it is unable
  to return to its initial working directory.  Similarly for "install -D
  file /tmp/a/b/file".

  "pr -D FORMAT" now accepts the same formats that "date +FORMAT" does.

  stat now exits nonzero if a file operand does not exist

** Improved robustness

  Date no longer needs to allocate virtual memory to do its job,
  so it can no longer fail due to an out-of-memory condition,
  no matter how large the result.

** Improved portability

  hostid now prints exactly 8 hexadecimal digits, possibly with leading zeros,
  and without any spurious leading "fff..." on 64-bit hosts.

  nice now works on Darwin 7.7.0 in spite of its invalid definition of NZERO.

  `rm -r' can remove all entries in a directory even when it is on a
  file system for which readdir is buggy and that was not checked by
  coreutils' old configure-time run-test.

  sleep no longer fails when resumed after being suspended on linux-,
  in spite of that kernel's buggy nanosleep implementation.

** New features

  chmod -w now complains if its behavior differs from what chmod a-w
  would do, and similarly for chmod -r, chmod -x, etc.

  cp and mv: the --reply=X option is deprecated

  date accepts the new option --rfc-3339=TIMESPEC.  The old --iso-8602 (-I)
  option is deprecated; it still works, but new applications should avoid it.
  date, du, ls, and pr's time formats now support new %:z, %::z, %:::z
  specifiers for numeric time zone offsets like -07:00, -07:00:00, and -07.

  dd has new iflag= and oflag= flags "binary" and "text", which have an
  effect only on nonstandard platforms that distinguish text from binary I/O.

  dircolors now supports SETUID, SETGID, STICKY_OTHER_WRITABLE,
  OTHER_WRITABLE, and STICKY, with ls providing default colors for these
  categories if not specified by dircolors.

  du accepts new options: --time[=TYPE] and --time-style=STYLE

  join now supports a NUL field separator, e.g., "join -t '\0'".
  join now detects and reports incompatible options, e.g., "join -t x -t y",

  ls no longer outputs an extra space between the mode and the link count
  when none of the listed files has an ACL.

  md5sum --check now accepts multiple input files, and similarly for sha1sum.

  If stdin is a terminal, nohup now redirects it from /dev/null to
  prevent the command from tying up an OpenSSH session after you logout.

  "rm -FOO" now suggests "rm ./-FOO" if the file "-FOO" exists and
  "-FOO" is not a valid option.

  stat -f -c %S outputs the fundamental block size (used for block counts).
  stat -f's default output format has been changed to output this size as well.
  stat -f recognizes file systems of type XFS and JFS

  "touch -" now touches standard output, not a file named "-".

  uname -a no longer generates the -p and -i outputs if they are unknown.

Unsubscribe info:
Problem reports: