* NEWS: "cp" no longer considers a destination symlink to be the
same as the referenced file when copying links or making backups.
* src/copy.c (copy_reg): When following a symlink, use the
followed name in later chown etc. requests, so that the created
file is affected, rather than the symlink. Use O_NOFOLLOW on
source when not dereferencing symlinks; this avoids a race.
Preserve errno correctly when doing multiple open attempts on the
destination.
(copy_internal): Follow destination symlinks only when copying a
regular file and only when we don't intend to remove or rename the
destination first, regardless of whether following source
symlinks; this is because since POSIX and tradition (e.g.,
FreeBSD) say we should ordinarily follow destination symlinks if
the system calls would ordinarily do so.
* src/copy.h (struct cp_options): Add comment that 'dereference'
is only for source files.
* src/cp.c (usage): Note that --derereference etc. are only for
source files.
(make_dir_parents_private): Follow symlinks, regardless of whether
--dereference is specified, because these are destination symlinks.
* tests/cp/same-file: Adjust tests to match revised behavior.
Filter out perror output since it might vary from host to host.
Use sed alone instead of also using echo.
* doc/coreutils.texi (cp invocation): Document the behavior better when
the destination is a symlink. Clarify source versus destination
symlinks. Describe the new behavior for destination symlinks.
2007-06-15 Jim Meyering <jim@meyering.net>
* src/copy.c: Include "canonicalize.h".
(copy_reg): Use canonicalize_filename_mode to follow the symlink,
so that we can always open with O_EXCL and avoid a race.
* NEWS: Mention the bug fix.
* src/copy.c (copy_reg): When open fails with EEXIST, the destination
is lstat'able, and a symlink, call open again, but now without O_EXCL.
* tests/cp/thru-dangling: New file, to test for the above fix.
* tests/cp/Makefile.am (TESTS): Add thru-dangling.
* THANKS: Add Michael McLagan.
Bug report from Michael McLagan in <http://bugzilla.redhat.com/243588>.
give a better diagnostic for a field-number/offset of 0.
* NEWS: Mention the fix.
* src/cut.c (ADD_RANGE_PAIR): Add an explicit check for 0.
Based on a patch from James Youngman.
* tests/misc/cut: Add tests for the above.
Also, diagnose the '-' in "cut -f -" as an invalid range, rather
than interpreting it as the unlimited range, "1-".
* NEWS: Mention these changes.
* src/cut.c (set_fields): Don't interpret an accumulator "value"
of 0 as an unspecified range endpoint.
Give better diagnostics.
Adjust a comment so that it is true also for 64-bit size_t.
* tests/cut/Test.pm: Add tests for the above.
* NEWS: Mention this.
* src/stty.c (strtoul_tcflag_t, strtoul_cc_t): New functions.
(recover_mode): Use those functions (not sscanf), to parse the
string robustly.
* tests/stty/invalid: New file. Test for the above.
* tests/stty/Makefile.am (TESTS): Add invalid.
* .x-sc_prohibit_atoi_atof: Don't exempt stty.c from this check.
Add tests/stty/invalid so we don't have to obfuscate the comment
about sscanf therein.
* Makefile.maint (sc_prohibit_atoi_atof): Mention sscanf in the
diagnostic, too.
* tests/uniq/Test.pm: When possible, create a "-z"-testing variant
of each existing test.
(2z, 3z, 4z, 5z, 20z, 122, 123): New tests from James Youngman.
* src/tr.c (main): Remove invalid assertion triggered by e.g.,
tr -c a '[b*256]'. There's nothing wrong with having Set2 larger
than Set1. Reported by Guntram Blohm.
* tests/tr/Test.pm (no-abort-1): Test for the above.
* NEWS: Mention this bug fix.
* THANKS: Add Guntram Blohm.
* src/split.c (line_bytes_split): Don't create an empty line
afterwards if the last buffer happens to be exactly full.
* tests/misc/split-fail: Add a test case for this.
* NEWS: mention this.
ls -x DIR would sometimes output the wrong string in place of the
first entry.
* NEWS: Mention the bug fix.
* tests/ls/x-option: New file.
* tests/ls/Makefile.am (TESTS): Add x-option.
Ulrich Drepper in <http://bugzilla.redhat.com/232652>.
* tests/misc/runcon-no-reorder: New file. Test for the above.
* tests/misc/Makefile.am (TESTS): Add runcon-no-reorder.
* tests/Makefile.am (t9): New target, to run tests/cp/cp-a-selinux.
(all_t): Add t9.
Use a directory on a loopback device mounted with -o context=...
* tests/cp/cp-a-selinux: Since this test now runs mount and umount,
it is a root-only one.
* tests/cp/Makefile.am (TESTS): Add cp-a-selinux.
* tests/selinux: New file.
* tests/Makefile.am (EXTRA_DIST): Add selinux.
* tests/misc/selinux: Source the new script, rather than open coding it.
Change how "cp -a" and "cp --preserve=context" work with SELinux.
Now, cp -a attempts to preserve context, but failure to do so does
not change cp's exit status. However "cp --preserve=context" is
similar, but failure *does* cause cp to exit with nonzero status.
* src/copy.h (struct cp_options) [require_preserve_context]: New member.
* src/copy.c (copy_reg, copy_internal): Implement the above.
* src/mv.c (cp_option_init): Initialize the new member.
* src/install.c (cp_option_init): Likewise.
* src/cp.c (cp_option_init): Likewise.
(decode_preserve_arg): Set it or reset it.
FIXME: add an on-writable-NFS-only test
I started with the patches from Red Hat.
The entries below tell how the code evolved.
* src/ls.c (print_long_format, print_file_name_and_frills): When
there is no security context (due to getfilecon/lgetfilecon failing
with e.g. ENOTSUP), print it as "?", not "".
* src/ls.c (print_file_name_and_frills): Make -Z work without -l.
(length_of_file_name_and_frills): Likewise.
* src/ls.c: Remove the --lcontext and --scontext options.
Change the way -Z, --context work so that it no longer implies -l.
Thus, -Z -l will work like -lcontext and -Z without -l will work
like --scontext.
Adjust tests to reflect new 'ls -l' syntax -- affects only
systems with SELinux when operating on a file with no ACL.
These tests assumed that everything before the first space on
each line is the 10-byte mode string. But there may also be a "+"
in the 11th column, just before the space. However, note that this
is not new. The same thing would have happened even without the
change below, when listing a file with an ACL.
* tests/chmod/equals, tests/cp/cp-parents, tests/cp/fail-perm:
* tests/cp/link-preserve, tests/install/basic-1, tests/misc/mknod:
* tests/mkdir/parents, tests/mkdir/special-1, tests/mv/partition-perm:
Don't make compilation depend on USE_ACL. An SELinux security
context counts as an "alternate access control method", so ls
must output a "+" for each file with a security context.
* src/ls.c [struct fileinfo] (have_acl): Declare unconditionally.
(FILE_HAS_ACL): Remove macro definition. Use f->have_acl directly.
(gobble_file): Record whether a file has a security context, and
update the condition used to determine whether to print the "+".
(gobble_file): Call getfilecon/lgetfilecon also when
format == long_format, so that we get the "+".
* src/ls.c (gobble_file): Add a comment explaining why (with a
security context option) ls doesn't exit nonzero due to e.g.,
getfilecon failing with errno == ENOTSUP.
* src/ls.c (gobble_file): Ignore failure of getfilecon if it's due
to ENOTSUP.
* src/ls.c (gobble_file): Factor out three small blocks using
getfilecon and lgetfilecon.
Don't ignore return value from getfilecon and lgetfilecon.
* src/ls.c (print_long_format): Don't use ?: (empty 2nd arg with C
ternary operator).
(print_scontext_format): Likewise.
(print_scontext): Declare to be "bool", not int. Adjust uses.
* src/Makefile.am (dir_LDADD, ls_LDADD, vdir_LDADD): Add $(LIB_SELINUX).
* tests/misc/chcon: New file.
* tests/misc/chcon-fail: New file.
* tests/Makefile.am (check-root): Run new, root-only misc/chcon test.
* tests/misc/Makefile.am (TESTS): Add chcon and chcon-fail.
* tests/misc/Makefile.am (TESTS): Add selinux.
* tests/misc/selinux: New file.
* tests/help-version: Skip chcon.
* tests/du/slink: When using df --local and df --type=TYPE,
test only the exit code. Don't bother with stdout.
Prompted by a report by Thomas Schwinge of an inaccurate diagnostic.
* tests/mkdir/p-3: When setting up an unreadable "." in an
inaccessible parent, make the parent inaccessible *after* making "."
unreadable. Otherwise, running "chmod a-r ." in an already-
inaccessible parent would fail on NFS with "Stale NFS file handle".
Signed-off-by: Jim Meyering <jim@meyering.net>
* NEWS: Describe how the fix affects pr.
* src/pr.c (init_parameters): The --merge (-m) option does
not imply --expand-tabs (-e), so don't set "untabify_input".
Reported by Wis Macomson.
* tests/misc/pr: New file. Test for the above fix.
* tests/misc/Makefile.am (TESTS): Add pr.
* THANKS: Update.
(overwrite_prompt): Say "try to overwrite", not "overwrite", to
make it clearer that the attempt may fail. Problem reported by
Dan Jacobson in:
http://lists.gnu.org/archive/html/bug-coreutils/2007-03/msg00130.html
Output symbolic mode as well as numeric.
* tests/mv/i-2 (fail): Adjust to new prompt format.
inaccessible parent, make the parent inaccessible *after* making "."
unreadable. Otherwise, running "chmod a-r ." in an already-
inaccessible parent would fail on NFS with "Stale NFS file handle".
Reported by Bob Proulx.
* tests/chgrp/basic: Don't let failure by chgrp to set the
group of a symlink make this test fail. Do give a diagnostic.
In the chgrp-no-change-ctime test, add darwin7.9.0 as another
known-failing system.
When failing on some other system, print $host_triplet, too.
Reported by Peter Fales.
* tests/misc/pwd-unreadable-parent: Rather than trying to decide
whether this test has a chance of succeeding, run it only when
$(host_os) is linux-gnu. It was failing on powerpc-apple-darwin8.8.0
* tests/misc/Makefile.am (TESTS_ENVIRONMENT): Add host_os=$(host_os).
* tests/misc/pwd-unreadable-parent: Test $REPLACE_GETCWD = 0,
rather than for __GETCWD_PREFIX in config.h (the latter is no
longer defined, ever, due to gnulib changes).
* tests/misc/Makefile.am (TESTS_ENVIRONMENT): Define REPLACE_GETCWD.
* bootstrap.conf: Remove findprog.
* doc/coreutils.texi (sort invocation): The default is to not
compress. Don't treat "" specially.
* src/sort.c: Don't include findprog.h.
(create_temp): Compress only if the user specified --compress-program.
* tests/misc/sort-compress: Adjusts tests to match new behavior.
* src/shuf.c (main): Move declaration of input_lines to
function scope, and initialize to NULL, so we can free it.
[lint]: Free all malloc'd memory.
* tests/misc/shuf: Also test shuf's -e and -i options.
* doc/coreutils.texi (dd invocation): Warn that noatime might not be
reliable.
* src/dd.c (flags, usage): Look at O_NOATIME, not
HAVE_WORKING_O_NOATIME, to decide whether to support the noatime
flag, so that dd attempts O_NOATIME even if the build file system
does not support it. Problem reported by Jim Meyering today in
bug-coreutils.
* tests/dd/misc: Generate a warning, not a failure, if noatime
exists but fails.
