The error diagnostic
"rm: cannot remove directory: '.'"
does not give the user a hint for the reason.
Issue a clearer error message.
* src/remove.c (rm_fts): Enhance the error diagnostic in the above
case to emphasize that skipping is done deliberately.
In the corresponding comment, mention that POSIX mandates this
behavior. Likewise in the subsequent comment for skipping "/".
* doc/coreutils.texi (rm invocation): In the paragraph describing
the above behavior, mention that POSIX mandates it.
This adds support for using a constant or "stick" parity bit.
* src/stty.c (usage): Mention the new flag.
* tests/misc/stty.sh: Adjust for the new flag.
* NEWS: Mention the improvement.
* docs/coreutils.texi (stty invocation): Mention the new flag.
Since the I/O overhead is significant to the relatively
simple processing done by this utility, use fputs() rather
than fputc() to output '\n'.
Time to process a 100MiB file was measured to
decrease from 0.417s to 0.383s, i.e. an 8% improvement.
Related to these changes, is a processing improvement in
gnulib, which increases throughput by 60% when processing
full buffers, which improves processing of a 100MiB file
with standard wrapped output, down to 0.256s.
http://git.sv.gnu.org/gitweb/?p=gnulib.git;a=commit;h=43fd1e7b
Also increase the encoding buffer size from 3 to 30KiB.
This was seen to give a further 8% improvement, taking
processing time down to 0.235s in the wrapped output case.
The decoding size buffer is not adjusted,
due to the noted caveat with --ignore-garbage.
* src/base64.c (BLOCKSIZE): Split into ENC_ and DEC_ variants,
with the former increased from 3KiB to 30KiB.
(wrap_write): Use the simpler fputc() rather than fputs()
to output the '\n' character. Also check against EOF
rather than < 0 for errors.
(do_encode): Likewise.
* NEWS: Mention the large increase in performance, which
with the I/O improvements in coreutils and the processing
improvement in gnulib, amount to about a 60% throughput increase.
* src/shred.c (dopass): Exit early to avoid redundant heap
allocation, and more importantly avoiding a file sync
when we're writting no data, as this can have side effects.
Also with --verbose, this avoids printing of "pass status"
which could be confusing as to whether data was actually written.
* tests/misc/shred-passes.sh: Ensure the status for data
passes are not written when not doing any data writes.
* src/shred.c (dopass): In the periodic pattern case increase the
I/O block size from 12KiB to 60KiB (also a multiple of 3 and 4096).
* NEWS: Adjust accordingly.
Since direct I/O is now enabled with commit v8.21-139-gebaf961
we must handle the case where we write an odd size at the
end of a file (with --exact), or we specify an odd --size that
is larger than 64KiB, or in the very unlikely case of a device
with an odd size. This issue was present since direct I/O
support was first added in v5.3.0, but latent since v6.0.
Theoretically this could have also been an issue after that on
systems which didn't have alignment constraints, but did have
size constraints for direct I/O.
* src/shred.c (dopass): On the first pass for a file, always
retry a write that fails with EINVAL, so we handle direct I/O
failure at either the start or end of the file. Adjust the comment
as the original case is out of date and implicitly handled
by this more general fix.
* tests/misc/shred-exact.sh: Add a test case.
* NEWS: Add a "bug fix" entry for shred since there are
two related issues now fixed.
* src/copy.c (create_hard_link): Add a bool 'dereference' parameter,
and pass AT_SYMLINK_FOLLOW as 'flags' to linkat() when dereference
is true.
(should_dereference): Add new 'bool' function to determine if a
file should be dereferenced or not.
(copy_internal): Use the above new should_dereference() and remember
its return value in a new local bool 'dereference' variable. Use that
in all three calls to create_hard_link().
* src/cp.c (main): after parsing the options, if x.dereference is
still DEFEF_UNDEFINED and the x.recursive is true, then only set
x.dereference to DEREF_NEVER iff --link was not specified.
* doc/coreutils.texi (cp invocation): Mention that cp(1) does not
follow symbolic links in the source when --link is specified.
Likewise in the description of the -R option when used together with
that option.
* tests/cp/same-file.sh: Adapt the expected results for the -fl,
the -bl and the -bfl tests.
* tests/cp/link-deref.sh: Add a new test.
* tests/local.mk (all_tests): Reference the above new test.
* NEWS (Changes in behavior): Mention the change.
This fixes http://bugs.gnu.org/15173
Co-authored-by: Bernhard Voelker <mail@bernhard-voelker.de>
Commit v5.92-1057-g43d487b introduced a regression
in coreutils 6.0 where it removed the page alignment
of the buffer to write, thus disabling direct I/O.
We want to use direct I/O when possible to avoid
impacting the page cache at least, as we know we don't
want to cache the data we're writing.
* src/shred.c (dopass): Allocate the buffer on the heap,
while using a more general calculation to allow to have
the output size independent from the fillpattern() size
constraint of a multiple of 3. Also we dispense with the
union as it's no longer needed given we're aligning on
a page boundary and thus don't need to explicitly handle
uint32_t alignment.
* src/md5sum.c (usage): s/three/four/ in the message pertaining
to the --check related options. Also clarify that --strict
is just significant for the formatting of the checksum lines.
Also since we're changing both strings, move the --strict description
in with the description of the other options and order alphabetically.
* THANKS.in: Added reporter: Daniel Mach
* src/md5sum.c (main): Add a comment as to why we continue
to escape names that do not have '\n' but do have '\\' chars.
(print_filename): Use the predetermined boolean to decide
whether to escape or not, so that in the common case we
can output the file name directly, rather than inspecting each char.
* tests/misc/md5sum.pl: Add case to show '\\' chars cause escaping.
* tests/misc/sha1sum.pl: Likewise.
* src/mktemp.c (main): Use an exit() strategy consistent with the
previous clauses dealing with optional error messages to ensure
we exit with the correct status in all cases.
Prompted by the continuous integration build failure at:
http://hydra.nixos.org/build/6412979
The reason for having a --quiet option is to
suppress only some subset of possible errors.
The most useful separation here is with usage/internal errors,
and errors due to file creation etc. (i.e. I/O errors).
* src/mktemp.c (main): Match the --help and info docs and
only suppress the file/dir creation error messages.
* tests/misc/mktemp.pl: Adjust accordingly.
* src/mktemp.c (usage): Synchronize the -p option description with
the logic and info docs. I.E. that -p is just an alias of --tmpdir.
Also for consistency treat --tmpdir='' the same with or without -t.
I.E. always ignore the --tmpdir option if the param is empty.
Fixes http://bugs.gnu.org/15425
This regression was introduced in commit v6.7-71-g0928c24
* src/rm.c (main): Make the -I option behave like --interactive=once.
* tests/rm/interactive-once.sh: Add cases for single and multiple files.
* NEWS: Mention the bug fix.
Fixes http://bugs.gnu.org/9308
* src/chown.c (main): Since "name" parameters to parse_user_spec()
are now optional, just pass NULL for those unused parameters.
* src/chroot.c (main): Likewise.
* src/id.c (usage): Remove 'name' from the synopsis,
implying that one can also specify by user ID.
(main): Like chown(1), call parse_user_spec() to implement
user name or ID lookup with appropriate precedence.
* doc/coreutils.texi (id invocation): Mention that
a user ID is supported and how '+' affects lookup order.
* tests/misc/id-groups.sh: Remove test now subsumed into...
* tests/misc/id-uid.sh: New test covering new interface.
* tests/local.mk: Rename the test.
* NEWS: Mention the new feature.
Addresses http://bugs.gnu.org/15421
* src/group-list.h (print_group_list): Add a parameter for the
delimiter of type char.
* src/group-list.c (print_group_list): Likewise, and use it instead
of a white space character to delimit the group entries.
* src/groups.c (main): Pass white space character to print_group_list().
* src/id.c (longopts): Add array element for the new long option.
(usage): Document the new option. While at it, fix the alignment
of the descriptions to match that of HELP_OPTION_DESCRIPTION.
(main): Define the bool flag opt_zero indicating the use of the
new option. In the getopt_long loop, handle it.
Output an error diagnostic in the case the --zero option has been
specified together with the default format.
In the case of -gG, pass either a NUL or a white space character to
print_group_list() - depending on the above new flag.
Likewise change the printing of the final newline character: output
a NUL instead if the --zero option has been specified.
* doc/coreutils.texi (id invocation): Document the new option.
While at it, move the @exitstatus macro down after the macro
@primaryAndSupplementaryGroups in order to be consistent with
other texinfo documents.
(groups invocation): Move @exitstatus down after the macro
@primaryAndSupplementaryGroups here, too.
* tests/misc/id-zero.sh: Add new test exercising the new option.
* tests/local.mk (all_tests): Reference it.
* NEWS (New features): Mention the new option.
Fixes http://bugs.gnu.org/9987
If there is an error reading a directory that was referenced
through recursion, rather than directly on the command line,
then exit with the "less serious" exit code, rather than the
"serious" exit code reserved for command line arguments.
This issue was introduced in commit v5.2.1-1908-gb58dea5
* src/ls.c (print_dir): Ensure that the command_line_arg param
is false for directories being recursed into.
* NEWS: Mention the bug fix.
Fixes http://bugs.gnu.org/15249
* src/copy.c (copy_internal): Change mention of the removed --reply=no
option, to the similar in this context --no-clobber.
* src/sort.c: SI and IEC suffixes can now be mixed when --human-numeric.
* doc/coreutils.texi (seq invocation): Add a sentence clarifying
that seq terminates when LAST becomes smaller than the current number
plus INCREMENT.
* src/seq.c (usage): Likewise.
Fixes http://bugs.gnu.org/15068
Also slightly rephrase some descriptions for extra clarity, and
add more consistent indentation.
* src/df.c (usage): Semicolon, no final period.
* src/du.c (usage): Likewise, plus indentation and clarifying words.
* src/ls.c (usage): Semicolon, rephrasings, added parentheses for
clarity, indentation.
* src/rm.c (usage): Semicolons.
* src/tail.c (usage): Adjust -f description to prefer explanatory
language instead of option syntax.
Run "make update-copyright".
* src/numfmt.c: Update copyright year number range. This file has
obviously been added to coreutils after and without the annual update.
* tests/misc/numfmt.pl: Likewise.
Also do not end option descriptions with a period, properly indent
continuation lines, and make some tiny clarifications.
* src/du.c (usage): Lowercase after semicolon.
* src/ls.c (usage): Semicolons instead of periods, small rephrasing
and two hyphens for clarity, proper indentation.
* src/mktemp.c (usage): Semicolons and lowercase.
* src/od.c (usage): Semicolons.
* src/ptx.c (usage): Use the standard phrase, clarify default option.
* src/setuidgid.c (usage): Properly indent continuation line.
* src/split.c (usage): Semicolons, lowercase, no final period.
* src/stat.c (usage): Semicolons, lowercase.
* src/tail.c (usage): Proper indentation, one shorter rephrasing,
semicolons, no final periods.
* src/timeout.c (usage): Properly indent, semicolons, no final periods.
Fixes http://bugs.gnu.org/14976
* src/csplit.c (find_lines): Assert that load_buffer() updates the
global buffers, thus "b" will be non NULL, thus suppressing subsequent
NULL pointer derefence warnings.
(process_regexp): Avoid a redundant assignment of the "line" pointer.
(process_line_count): Likewise. Also reduce the "line" pointer scope.
This new option can be used to find directories with a huge
amount of files. The GNU find utility has the printf format
"%h" which prints the number of entries in a directory, but
this is non-cumulative and doesn't handle hard links.
* src/du.c (struct duinfo): Add new member for counting inodes.
(duinfo_init): Initialize inodes member with Zero.
(duinfo_set): Set inodes counter to 1.
(duinfo_add): Sum up the 2 given inodes counters.
(opt_inodes): Add new boolean flag to remember if the --inodes
option has been specified.
(INODES_OPTION): Add new enum value to be used ...
(long_options): ... here.
(usage): Add description of the new option.
(print_size): Pass inodes counter or size to print_only_size,
depending on the inodes mode.
(process_file): Adapt threshold handling: with --inodes, print or
elide the entries according to the struct member inodes.
(main): Add a case for accepting the new INODES_OPTION.
Print a warning diagnostic when --inodes is used together with the
option --apparent-size or -b.
Reset the output_block_size to 1 ... and thus ignoring the
options -m and -k.
* tests/du/inodes.sh: Add a new test.
* tests/local.mk (all_tests): Mention it.
* doc/coreutils.texi (du invocation): Document the new option.
* NEWS: Mention the new option.
src/copy.c (copy_internal): Use rmdir() rather than unlink()
when the source is a directory, so that empty directories
are replaced in the destination as per POSIX.
* tests/mv/part-rename.sh: Augment with various combinations.
* NEWS: Mention the bug fix.
Fixes http://bugs.gnu.org/14763
Include the number of arguments which rm received in the "Remove all
arguments?" prompt. This is useful in the, presumably, common case
where the arguments were not provided by hand, but instead were the
result of various shell expansions. A simple, if somewhat contrived,
example (assuming rm is aliased to rm -I) could be:
rm * .o
where the prompt "Remove 120 arguments?" is more likely to make
the user catch the problem.
* src/rm.c (main): Include correctly pluralized n_files
in the output message. Also remove the now redudant "all".
* tests/rm/interactive-always.sh: Adjust to the new prompt.
* tests/rm/interactive-once.sh: Likewise.
* src/dd.c (STATUS_NONE): Simplify the enum so that
it's more general than just suppressing transfer counts.
Then test this in all locations where non fatal diagnostics
are output.
* tests/dd/misc.sh: Ensure the diagnostic about
being unable to skip past the end of input is suppressed.
* NEWS: Mention the change in behavior.
Fixes http://bugs.gnu.org/14897
* src/head.c (elide_tail_lines_file): For seekable empty files,
or seekable files where the current offset is after the
end of the file, return immediately. Previously the short
circuit code could not be reached due to logic error.
Spotted by coverity.
Similarly to commit v8.21-84-g8d2da3f in src/uptime.c
avoid a "definitely lost" error from valgrind. Note this
only happens with pinky when compiled without optimization,
in which case certain paths aren't eliminated casuing
valgrind to trigger the message. Note also that coverity
flags this "resource leak" too.
* src/pinky.c (short_pinky): free utmp_buf for developer builds.
Avoid Valgrind reports of "definitely lost" items
and while at it, free all discarded mount entries
to minimize the amount of memory used.
* src/df.c (filter_mount_list): Use the newly exported
free_mount_entry() from gnulib to free all mount entries
as they're discarded.
Prompted by the continuous integration build failure at:
http://hydra.nixos.org/build/5508873
* src/shuf.c (write_random_numbers): Convert to an int type
that matches the prinft format spec.
main(): Process new option. Replace input_numbers_option_used()
with a local variable. Re-organize argument processing.
usage(): Describe the new option.
(write_random_numbers): A new function to generate a
permutation of the specified input range with repetition.
(write_random_lines): Likewise for stdin and --echo.
(write_permuted_numbers): New function refactored from
write_permuted_output().
(write_permuted_lines): Likewise.
* tests/misc/shuf.sh: Add tests for --repetitions option.
* doc/coreutils.texi: Mention --repetitions, add examples.
* TODO: Mention an optimization to avoid needing to
read all of the input into memory with --repetitions.
* NEWS: Mention new shuf option.
* src/df.c (filter_mount_list): Initialize devlist->dev_num correctly
when unable to stat() a mount point. This will avoid possible invalid
deduplication done on the list due to use of uninitialized memory.
* tests/df/skip-duplicates.sh: Ensure this code path is exercised.
Also refactor the test to be table driven.
* NEWS: Mention the bug fix.
* src/mkdir.c (main): Move the variable ret and issuing the error
message into the body of the scontext-related if-block.
* src/mkfifo.c (main): Likewise.
* src/mknod.c (main): Likewise.
Consolidate all smack routines and checks in a module.
We replace and wrap the most commonly used smack routines,
which allows removing ifdefs throughout the code.
* gl/lib/smack.h: A new header containing the implementation
of the wrapped and replacement routines. Note the is_smack_enabled()
routine should be optimized out at compile time when compiled
on a system without libsmack.
* gl/modules/smack: Describe the new module and move the
configure time code here from ...
* m4/jm-macros.m4: ... here.
* bootstrap.conf: Reference the new module.
* src/id.c: Use the routines without ifdefs where possible.
* src/ls.c: Likewise.
* src/mkdir.c: Likewise.
* src/mkfifo.c: Likewise.
* src/mknod.c: Likewise.
Enable creation of SMACK security context with -Z command-line switch
if SMACK is enabled.
* mkdir.c (main): Set process security context to given SMACK label.
* mkfifo.c (main): Likewise.
* mknod.c (main): Likewise.
* src/local.mk: link mk{dir, fifo, nod} with libsmack.
* NEWS: Mention the new feature.
Enable showing of file SMACK security with '-Z' command-line switch
if SMACK is enabled. Showing SMACK context of a file does not strictly
require SMACK to be enabled but this required to make choice whether to
show SELinux or SMACK security context.
* src/ls.c (getfilecon_cache): Retrieve SMACK context if available.
(gobble_file): Handle SMACK context similarly to SELinux context.
* src/local.mk: Link lsl with libsmack.
* NEWS: Mention the new feature.
* .mailmap: Merge the Author's 2 email addresses.
This is consistent with the documented interface and
avoids any ambiguity in a user thinking that stdbuf without options
might reset to a "standard" buffering setup.
* src/stdbuf.c (set_libstdbuf_options): Indicate with the return value
whether any env variables were actually set.
(main): Fail unless some env variables were set.
* tests/misc/stdbuf.sh: Ensure this constraint is enforced.
* NEWS: Mention the small change in behavior.
* src/truncate.c (usage): Mention that --size is in bytes which
is by far the most common usage.
* doc/coreutils.texi (truncate invocation): Likewise. Also cross
reference the --io-blocks option.
Reported in http://bugs.gnu.org/14686
