On most unix- and linux-based kernels, ls -i DIR_CONTAINING_MOUNT_POINT
would print the wrong inode number for any entry that is a mount point.
It would do that by relying on readdir's dirent.d_ino values, while
most readdir implementations return the inode number of the underlying,
inaccessible directory. Thus, it is not consistent with what you'd
get when applying stat to the same entry. This bug led to surprising
results like "ls -i" and "ls -i --color" printing different numbers (ls
must usually "stat" a file to colorize its name). This change makes it
so that on offending systems, ls must stat non-command-line-arguments
for which otherwise it would be able to use "for free" dirent.d_ino
values. Regardless of this change, ls is already required to stat every
command-line argument. Note: versions of GNU ls prior to coreutils-6.0
did not perform the invalid optimization, and hence always printed
correct inode numbers. Thus, for the sake of correctness, ls -i is
forgoing the readdir optimization, for any kernel (including linux!)
with POSIX-nonconforming readdir. Note that currently, only Cygwin has
been agile enough to conform.
* src/ls.c (RELIABLE_D_INO): Define.
(print_dir): Use it.
For plenty of discussion, see this long thread:
http://thread.gmane.org/gmane.comp.gnu.coreutils.bugs/14020
This bug was introduced by the 2006-02-26 commit, 33eb3efe:
"In ls, avoid calling stat for --inode (-i), when possible."
* tests/ls/readdir-mountpoint-inode: New test.
* tests/Makefile.am (TESTS): Add it.
* tests/ls/stat-vs-dirent: Don't suppress failure of this test,
now that ls -i is fixed. Though note that it doesn't test well,
since it compares only the always-stat'd command-line arguments.
* NEWS (Bug fixes): Mention it.
* src/copy.c (copy_reg): When cloning only skip the data copying
* tests/cp/reflink-perm: New test to check times and modes copied
* tests/Makefile.am: Reference the new test
* NEWS: Mention the fix
* doc/coreutils.texi (cp invocation): Document the new
"auto" and "always" options to --reflink.
* src/copy.c (copy_reg): Fall back to a standard copy
when reflink() is not supported and --reflink=auto specified.
* src/copy.h [struct cp_options] (reflink): Change type s/bool/enum/.
* src/cp.c (usage): Describe the --reflink={always,auto} options
and expand a little on what --reflink does.
(main): parse the new parameters to --reflink and allow all
--sparse options with --reflink=auto.
* src/install.c (cp_option_init): Init the enum instead of bool.
* src/mv.c (cp_option_init): Likewise.
* tests/cp/reflink-auto: A new test for falling back to normal copy.
* tests/Makefile.am: Reference the new test.
* NEWS: Mention the new feature.
* Makefile.am (.version, dist-hook, gen-ChangeLog): Use $(AM_V_GEN)
and $(AM_V_at), so that automake's silent-rules option (make V=1/V=0)
now controls whether the commands are printed at build time.
(THANKS-to-translators, check-ls-dircolors): Likewise.
* tests/test-lib.sh (require_selinux_enforcing_): New function.
* tests/mkdir/selinux: Use it.
Otherwise, this test would fail on Rawhide with SELinux disabled.
* tests/tail-2/assert: Avoid spurious failure due to race condition.
Rather than sleeping for 1 second and crossing fingers,
wait explicitly for backgrounded tail process to start.
Otherwise, this test would fail under heavy load.
Now that we prohibit indentation via TABs, there's no need for
Emacs indent-tabs-mode setting lines, so prohibit those, too.
* cfg.mk (sc_prohibit_emacs__indent_tabs_mode__setting): New rule.
...when run on a kernel older than what was implied by headers and
libraries tested at configure time.
* src/copy.c (utimens_symlink): Ignore failure when errno == ENOSYS.
* NEWS (Bug fixes): Mention it.
Reported by Todd Zullinger and Kamil Dudka.
Details in this thread:
http://thread.gmane.org/gmane.linux.redhat.fedora.devel/119834
* tests/cp/cp-mv-enotsup-xattr: Upon a set-up failiure, rather than
failing the test with a "framework failure" diagnostic, just skip it.
Russell Whitaker reported that this test failed on slackware.
* src/install.c (extra_mode): Be careful to return only a 0 or 1
value, since this is a "bool" function, and there are still some
compilers for which this is necessary. Without this change,
Bernhard Voelker reported that the Forte Developer 7 C 5.4 2002/03/09
compiler would produce a malfunctioning "install" binary. Details in
http://thread.gmane.org/gmane.comp.gnu.coreutils.bugs/17710/focus=17760
Now that "make check" no longer builds programs like arch, ...
* src/Makefile.am (dist-hook): Depend on $(all_programs), so that the
subsequent build in man/ doesn't fail due to lack of a program like
arch that is not going to be installed.
* NEWS: Mention the change.
* doc/coreutils.texi: Document the new --line-increment option.
* src/nl.c (struct option): Add --line-increment,
(usage): Describe it,
(main): Use it.
* src/Makefile.am (check): Don't depend on check-AUTHORS.
That check fails on systems for which a program like stdbuf is
not built. Instead, move this check to "make syntax-check".
(sc_check-AUTHORS): Rename rule from check-AUTHORS.
* cfg.mk (sc_check-AUTHORS): New rule.
Reported by Berhnard Voelker.
* src/sort.c (find_unit_order): Spell out 256-element static
initializer, rather than relying on C99 syntax.
Required for Forte Developer 7 C 5.4 2002/03/09 on Solaris 10.
Reported by Bernhard Voelker.
* src/copy.c (copy_internal): On systems without utimensat don't
use utimens on a symlink, as that would dereference the symlink.
* tests/cp/abuse: To work around possible attribute preservation
failures breaking the test, use cp -dR rather than cp -a.
* NEWS: Remove the description associated with the removed
experimental code which unconditionally tried to reflink() on copy.
Also clarify where --reflink works exactly.
* src/tail.c (tail_forever_inotify): Remove the redundant and
incorrect error check of the return from inotify_add_watch().
Also initialize the wd member of each File_spec to an invalid value.
Reported by C de-Avillez.
