directory before opening ".", to avoid trouble with file system
implementations that cache readdir results at opendir-time.
* src/remove.c (AD_pop_and_chdir): Add a file descriptor parameter.
Don't update **DIRP. Don't call fdopendir here.
(remove_dir): Call fdopendir here instead.
Report and patch from Mikulas Patocka:
<http://lists.gnu.org/archive/html/bug-coreutils/2006-12/msg00170.html>
as POSIX requires.
* src/dd.c (scanargs): Implement it.
* tests/dd/misc (outbytes): Test it.
* doc/coreutils.texi (dd invocation): Specify that bs=N
overrides later ibs and obs, undoing part of the
previous change. (The behavior was wrong.)
* src/remove.c (remove_entry): Handle EACCES for a non-directory, too.
Karl Berry reported that a cross-partition "mv /etc/issue ~"
failed with the um,... suboptimal diagnostic,
"mv: cannot remove `/etc/issue': Not a directory".
* tests/rm/Makefile.am (TESTS): Add fail-eacces.
* tests/rm/fail-eacces: New file.
* NEWS: Mention that both mv and rm are affected.
* src/cut.c (cut_fields): Set file-scoped global to NULL after
freeing it. This avoids a double-free (and core dump on some systems)
for this usage: "echo 1>a; echo 2>b; cut -f2- a b". Reported by
James Hunt in <http://bugzilla.redhat.com/220312>.
* NEWS: List this bug fix.
* THANKS: Mention him.
* tests/misc/cut: New file.
* tests/misc/Makefile.am (TESTS): Add cut.
* src/chgrp.c (main): Don't prohibit -RLh, aka -RL with --no-dereference.
* src/chown.c (main): Likewise.
* src/chown-core.c (change_file_owner): Add to a comment.
* tests/chown/preserve-root: Add tests.
* doc/coreutils.texi (Treating / specially): With --preserve-root,
chgrp and chown will not modify "/", even through a symlink.
* src/chmod.c (process_file): Do honor the --preserve-root option.
* src/chown-core.c (change_file_owner): Likewise, but here, also
handle the case in which a traversal would go "through" a symlink
to root. Reported by Matthew M. Boedicker
* tests/chown/preserve-root: Test for the above.
* tests/chown/Makefile.am (TESTS): Add preserve-root.
so that commands like "sort -k 18446744073709551616" no longer fail merely
because 18446744073709551616 doesn't fit in uintmax_t. The trick is that
these fields can all be treated as effectively infinity; their exact
values don't matter, since no internal buffer can be that long.
* src/join.c (string_to_join_field): Verify that SIZE_MAX <= ULONG_MAX
if the code assumes this. Silently truncate too-large values to SIZE_MAX,
as the remaining code will do the right thing in this case.
* src/sort.c (parse_field_count): Likewise.
* src/uniq.c (size_opt, main): Likewise.
* tests/join/Test.pm (bigfield): New test.
* tests/sort/Test.pm (bigfield): New test.
* tests/uniq/Test.pm (121): New test.
Signed-off-by: Jim Meyering <jim@meyering.net>
* NEWS: Document the cp -p fix for special bits.
* src/copy.c (set_owner): Now returns a three-way result, so
that the caller can clear the special bits. All callers changed.
(copy_reg): Don't set the special bits if chown failed.
(copy_internal): Likewise.
* tests/cp/special-bits: Test this fix.
Signed-off-by: Jim Meyering <jim@meyering.net>
* m4/jm-macros.m4 (coreutils_MACROS): Check for fchmod.
* src/copy.c (fchmod_or_lchmod): New function.
(copy_reg): New arg OMITTED_PERMISSIONS. All uses changed.
Omit confusing and unused ", dst_mode" arg to 'open' without O_CREAT.
When creating a file, use O_EXCL, so we're more likely to detect
funny business by other processes. At the end, if permissions
were omitted, chmod them back in.
(copy_internal): If the ownership might change, omit some permissions
at first, then restore them after chowning the file.
* src/cp.c (make_dir_parents_private): Likewise.
* src/copy.c (cached_umask): New function.
* src/copy.h (cached_umask): New decl.
while trying to retain the readability of C99 as much as possible.
* src/remove.c (close_preserve_errno): Remove.
(fd_to_subdirp): Rewrite to avoid the need for decl after statement.
Signed-off-by: Paul Eggert <eggert@cs.ucla.edu>
ln: creating hard link `f' => `f': No such file or directory
* src/ln.c (do_link): Give a better diagnostic in this unusual case.
(do_link): Rename local: s/lstat_ok/dest_lstat_ok/.
* tests/ln/Makefile.am (TESTS): Add hard-backup.
* tests/ln/hard-backup: New test for the above.
* src/dd.c (dd_copy): Declare real_buf and real_obuf to be static,
so we need not free them at all. This is easier than freeing
both buffers at each of the early "return"s.
* src/sort.c (main): Don't allocate memory for each new key here.
(insertkey): Allocate memory for each key here, instead.
(key_init): Rename from new_key. Don't allocate.
* src/system.h (ftello) [!HAVE_FSEEKO && !defined ftello]:
Define inline replacement function.
This (along with a yesterday's fix for autoconf's
_AC_SYS_LARGEFILE_MACRO_VALUE macro) makes it so coreutils
now builds once more on Tru64 V4.0. Reported by Nelson Beebe.
Suggested by Steve McIntyre in <http://bugs.debian.org/392925>.
* src/remove.h (struct rm_options) [one_file_system]: New member.
* src/rm.c (rm_option_init): Initialize it.
(usage): Document the option.
* src/mv.c (rm_option_init): Likewise.
* src/remove.c (remove_dir): With --one-file-system and --recursive,
for each directory command line argument, do not affect a file system
different from that of the starting directory. And give a diagnostic.
* src/rm.c (ONE_FILE_SYSTEM): New enum.
(main): Handle new option.
* tests/rm/one-file-system: Test the above.
* tests/rm/Makefile.am (TESTS): Add one-file-system.
* tests/Makefile.am (check-root): Add the rm/one-file-system
test to the list.
(EXTRA_DIST): Add other-fs-tmpdir.
* tests/mv/setup: Removed. Renamed to...
* tests/other-fs-tmpdir: ...this new file.
* tests/mv/Makefile.am (EXTRA_DIST): Remove setup.
* tests/mv/acl: Reflect renaming: use ../other-fs-tmpdir.
* tests/mv/backup-is-src: Likewise.
* tests/mv/hard-link-1: Likewise.
* tests/mv/leak-fd: Likewise.
* tests/mv/mv-special-1: Likewise.
* tests/mv/part-fail: Likewise.
* tests/mv/part-hardlink: Likewise.
* tests/mv/part-rename: Likewise.
* tests/mv/part-symlink: Likewise.
* tests/mv/partition-perm: Likewise.
* tests/mv/to-symlink: Likewise.
* tests/mv/into-self-2: Likewise.
[doc/ChangeLog]
* coreutils.texi (rm invocation): Describe --one-file-system.
on some file system types) also with the --preserve-root option
of chown or chgrp.
* src/chown-core.c (change_file_owner): Compare fts_statp-based
dev/ino against root dev/ino only for directories.
(chown_files): Don't let the root_dev_ino setting influence whether
we use FTS_NOSTAT: fts always sets *fts_statp for a directory.
7777. This matches historical 'cp' behavior and avoids some
(though not all) implementation-defined behavior of mkdir.
* src/cp.c (make_dir_parents_private): Likewise.
* src/copy.c (copy_internal): Don't pass 'open' a mode greater
than 777. This is required by POSIX. It doesn't make any difference
in actual behavior on any host that I know of.
"fdatasync" option string from its description, so help2man formats
the derived man page properly. Reported by Samuel Thibault
in <http://bugs.debian.org/393649>.