* src/local.mk (src_fold_LDADD): Remove $(MBRTOWC_LIB) since it is
already added to LDADD. Remove $(LIBC32CONV) and $(LIBUNISTRING) which
were for an uncommitted patch which used Gnulib's mbfile module.
The meaning of non-file permission umask bits is implementation defined.
On Bionic libc, attempting to set them triggers a FORTIFY runtime check.
$ nohup true
FORTIFY: umask: called with invalid mask -601
Aborted nohup true
* src/nohup.c: (main) Avoid setting non-permission bits in umask.
Just clear the umask to ensure we create nohup.out with u+rw,
as we restore the original umask before the exec().
* tests/misc/nohup.sh: Add a test case.
* NEWS: Mention the bug fix.
* src/basenc.c (has_padding): A more robust helper to
identify padding in the presence of trailing newlines.
(do_decode): Use has_padding() rather than just looking
at the last character.
* tests/basenc/base64.pl: Fully test commit v9.4-53-g378dc38f4
by ensuring partially padded data is diagnosed.
baddecode9 is the case fixed in this commit.
* NEWS: Mention the bug fix.
This implicitly tests the previous commit to
adjust how date(1) handles multiple named format options.
Currrently it tests the following are supported:
chown --quiet --silent
date --rfc-email --rfc-822 --rfc-2822
date --uct --utc --universal
dircolors --bourne-shell --sh
dircolors --csh --c-shell
head --quiet --silent
* tests/misc/option-aliases.sh: A new test to ensure all
option aliases supported by a command are supported.
* Reference the new test.
This avoids a test failure on FreeBSD 14, MacOS 15, and musl.
Fix suggested by Pádraig Brady in:
<https://bugs.gnu.org/79301#32>.
* tests/fold/fold-spaces.sh (isblank): New function. Only run the tests
if the character is treated as blank.
Fixes https://bugs.gnu.org/79301
To make the interface more concise and consistent,
while being backwards compatible.
* src/digest.c (main): Continue to support -a "sha###" but
also support -a "sha2" and treat it like "sha3", except in...
(output_file): ... maintain the legacy tags for better compatability.
* doc/coreutils.texi (cksum invocation): Document the -a sha2 option.
* tests/cksum/cksum-base64.pl: Adjust as per modified --help.
* tests/cksum/cksum-c.sh: Add new supported SHA2-### tagged variant.
* NEWS: Mention the new feature.
* src/digest.c: Include sha3.h.
(BLAKE2B_MAX_LEN): Rename to
DIGEST_MAX_LEN since it is also used for SHA-3.
(sha3_sum_stream): New function.
(enum Algorithm, algorithm_args, algorithm_args, algorithm_types)
algorithm_tags, algorithm_bits, cksumfns, cksum_output_fns): Add entries
for SHA-3.
(usage): Mention that SHA-3 is supported. Mention requirements for
--length with SHA-3.
(split_3): Use DIGEST_MAX_LEN instead of BLAKE2B_MAX_LEN. Determine the
length of the digest for SHA-3. Make sure it is 224, 256, 384, or 512.
(digest_file): Set the digest length in bytes. Use DIGEST_MAX_LEN
instead of BLAKE2B_MAX_LEN. Always append the digest length to SHA3 in
the output.
(main): Allow the use of --length with 'cksum -a sha3'. Use
DIGEST_MAX_LEN instead of BLAKE2B_MAX_LEN. Make sure it is 224, 256,
384, or 512.
* tests/cksum/cksum-base64.pl (@pairs): Add expected sha3 output.
(fmt): Modify the output to use SHA3-512 since that is the default.
(@Tests): Modify arguments for sha3 to use --length=512.
* tests/cksum/cksum-sha3.sh: New test, based on tests/cksum/b2sum.sh.
* tests/local.mk (all_tests): Add the test.
* bootstrap.conf: Add crypto/sha3.
* gnulib: Update to latest commit.
* NEWS: Mention the change.
* doc/coreutils.texi (cksum general options): Mention sha3 as a
supported argument to the -a option. Mention that 'cksum -a sha3'
supports the --length option. Mention that SHA-3 is considered secure.
NetBSD 10 and Solaris 11.4 treat non-breaking spaces as blank
characters unlike glibc.
* src/system.h: Include uchar.h.
(c32isnbspace): New function based on iswnbspace from src/wc.c.
* src/fold.c (fold_file): Use it.
* src/wc.c (iswnbspace): Remove function.
(maybe_c32isnbspace): New function.
(wc, main): Use it.
Fixes https://bugs.gnu.org/79300
* bootstrap.conf (gnulib_modules): Add issymlink and issymlinkat.
* src/copy.c: Include issymlink.h.
(copy_reg): Use issymlink instead of readlinkat.
* src/rmdir.c: Include issymlink.h.
(main): Use issymlink instead of readlink.
* src/tail.c: Include issymlink.h.
(recheck, any_symlinks): Use issymlink instead of readlink.
* src/test.c: Include issymlink.h.
(unary_operator): Use issymlink instead of readlink.
* src/seq.c (main): Avoid possibly innacurate conversion
to long double, for all digit start values.
* tests/seq/seq-long-double.sh: Add a test case.
* NEWS: Mention the improvement.
Fixes https://bugs.gnu.org/79369
Fix allocated size alignment in locales with multi-byte grouping chars.
Tested with: LC_ALL=sv_SE.utf8 ls --size --block-size=\'k
* src/ls.c (print_file_name_and_frills): Don't rely on
printf("%*s", width, string) to pad multi-byte strings appropriately.
Instead work out the padding required and use:
printf("%*s%s", padding, "", string) to pad multi-byte appropriately.
* tests/ls/block-size.sh: Add a test case.
* NEWS: Mention the bug fix.
Fixes https://bugs.gnu.org/79347
* src/digest.c (main): Don't saturate -l to BLAKE2B_MAX_LEN,
so that the subsequent bounds check is performed.
* tests/cksum/b2sum.sh: Add a test case.
* NEWS: Mention the fix introduced in commit v9.5-71-gf2c84fe63
* src/fold.c (fold_file): Continue the loop when we have buffered bytes
but nothing left to read from the file.
(adjust_column): Don't assume that the character is printable.
* tests/fold/fold-characters.sh: Add a new test case.
(bad_unicode): New function.
* src/fold.c (fold_file): Replace invalid characters with the original
byte read. Copy multibyte sequences that may not yet be read to the
start of the buffer before reading more bytes.
* tests/fold/fold-characters.sh: Add a test case.
Enforcing this interface behavior is worthwhile
irrespective of our current implementation,
to ensure future or other implementations conform.
* tests/fold/fold-characters.sh: Ensure the fold implementation
uses bounded memory.
* src/fold.c: Include ioblksize.h.
(fold_file): Use two IO_BUFSIZE-sized buffers. Use fread instead of
getline. Check for if we reached the end of file.
* tests/nproc/nproc-quota.sh: Also simulate sched_getscheduler()
as this will not be called on older or non linux, or
may return ENOSYS on Alpine.
Fixes https://bugs.gnu.org/79299
Related to commits v9.1-109-g879d2180d and v9.7-248-g306de6c26
* tests/cp/sparse-perf.sh: This edge case was missed a couple of times,
so add a test to ensure we attempt copy offload.
* src/fold.c: Include mcel.h.
(count_bytes): Remove variable.
(counting_mode, last_character_width): New variables.
(shortopts, long_options): Add the option.
(adjust_column): If --characters is in used account for number of
characters instead of their width.
(fold_file): Use getline and iterate over the result with mcel functions
to handle multibyte characters.
(main): Check for the option.
* src/local.mk (src_fold_LDADD): Add $(LIBC32CONV), $(LIBUNISTRING), and
$(MBRTOWC_LIB).
* tests/fold/fold-characters.sh: New file.
* tests/fold/fold-spaces.sh: New file.
* tests/fold/fold-nbsp.sh: New file.
* tests/local.mk (all_tests): Add the tests.
* NEWS: Mention the new option.
* doc/coreutils.texi (fold invocation): Likewise.
Better fix for problem reported by Jeremy Allison
<https://bugs.gnu.org/79267>.
* src/copy.c (struct scan_inference): New type, replacing
union scan_inference. All uses changed. This is so
infer_scantype can report the first hole's offset when known.
(lseek_copy): 5th arg is now struct scan_inference const *,
not just off_t. All uses changed.
(infer_scantype): If SEEK_SET+SEEK_HOLE do not find a hole,
fall back on ZERO_SCANTYPE.
This reverts part of the previous change.
* src/copy.c (lseek_copy): When calling sparse_copy, do not
ask it to scan for zeros unless --sparse=always, so that it
can use copy_file_range which can be far more efficient.
Problem reported by Jeremy Allison <https://bugs.gnu.org/79267>.
* src/copy.c (create_hole, sparse_copy): Omit arg PUNCH_HOLES,
as we always punch holes now. All uses changed.
(lseek_copy): When calling sparse_copy, scan for holes when
sparse_mode == SPARSE_AUTO, as that means we are making holes.
(copy_reg): Always punch any hole made at end.
* doc/sort-version.texi (Other version/natural sort implementations):
Use https in documentation link.
* tests/chmod/symlinks.sh: Use https in license text.
* NEWS: Mention the new feature.
* doc/coreutils.texi (nproc invocation): Mention that
cgroup CPU quotas can limit the reported number.
* gnulib: Update to new nproc gnulib implementation:
https://github.com/coreutils/gnulib/commit/9b07115f4a
* src/tsort.c: Don't include long-options.h since the previous commit
removed the call to parse_gnu_standard_options_only. This avoids a
sc_prohibit_long_options_without_use syntax-check failure.
This is for conformance to POSIX.1-2024
* src/tsort.c: Include getopt.h.
(main): Accept and ignore -w. Do not bother altering
the usage message, as the option is useless.
* tests/misc/tsort.pl (cycle-3): New test.
* src/realpath.c (longopts): Add the option.
(main): Likewise.
(usage): Add the option to the --help message.
* tests/misc/realpath.sh: Add a simple test.
* doc/coreutils.texi (realpath invocation): Mention the new option.
* NEWS: Likewise.
* doc/coreutils.texi (basenc invocation): Add an example
using --base58 to generate a unique ID. This also demonstrates
compound usage of the basenc command, to convert to/from binary.
A 58 character encoding that:
- avoids visually ambiguous 0OIl characters
- uses only alphanumeric characters
Described at:
- https://tools.ietf.org/html/draft-msporny-base58-03
This implementation uses GMP (or gnulib's gmp fallback).
Performance is good in comparison to other implementations.
For example when using libgmp on an i7-5600U system,
encoding is 530 times faster, and decoding 830 times faster
than the implementation using arbitrary precision ints in cpython 3.13.
Memory use is proportional to the size of input.
Encoding benchmarks:
$ time yes | head -c65535 | src/basenc --base58 -w0 >file.enc
real 0m0.018s
./configure --quiet --without-libgmp && make -j $(nproc)
$ time yes | head -c65535 | src/basenc --base58 -w0 >file.enc
real 0m3.431s
# dnf install python3-base58
$ time yes | head -c65535 | base58 >file.enc # cpython 3.13
real 0m9.700s
Decoding benchmarks:
$ time src/basenc --base58 -d <file.enc >/dev/null
real 0m0.010s
$ ./configure --without-libgmp && make # gnulib gmp
$ time src/basenc --base58 -d <file.enc >/dev/null
real 0m0.145s
$ time base58 -d <file.enc >/dev/null # cpython 3.13
real 0m8.302s
* src/basenc.c (base_decode_ctx_finalize, base_encode_ctx_init,
base_encode_ctx, base_encode_ctx_finalize): New functions to
provide more general processing functionality.
(base58_{de,en}code_ctx{_init,,_finalize}): New functions to
accumulate all input before calling ...
(base58_{de,en}code): ... the GMP based encoding/decoding routines.
(do_encode, do_decode): Call the ctx variants if enabled.
* doc/coreutils.texi (basenc invocation): Describe the new option,
and indicate the main use case being interactive user use.
* src/local.mk: Link basenc with GMP.
* tests/basenc/basenc.pl: Add test cases.
* NEWS: Mention the new feature.
* src/basenc.c (do_decode): With -i ensure we strip '=' chars
if there is no padding for the chosen encoding.
* tests/basenc/basenc.pl: Add a test case.
* NEWS: Mention the bug fix.
* src/basenc.c (base16_encode, z85_encoding, do_decode): Use
ATTRIBUTE_NONSTRING instead of ATTRIBUTE_NONSTRING.
* src/basenc.c (sc_prohibit-_gl-attributes): New rule for
'make syntax-check'.
The sparse code sometimes issued multiple lseeks against the
same file without doing anything in betwee. Optimize them away
by keeping track of the last hole output, in a way that
crosses the sparse_copy function call boundary.
* src/copy.c (sparse_copy): New arg hole_size, replacing old args
scan_holes and last_write_made_hole. All callers changed.
(sparse_copy, lseek_copy): Do not create hole at
end; let the caller deal with it. All callers changed.
(lseek_copy): New args hole_size and total_n_read. Caller changed.
(copy_reg): Create hole at end for both lseek_copy and sparse_copy.
The sparse code assumed that st_blksize was the minimum hole size.
However, st_blksize is an optimum I/O buffer size, not the file
system fundamental block size. Use ST_NBLOCKSIZE instead;
although it may underestimate the true block size that just slows
‘cp’ down a bit, without introducing bugs.
* src/copy.c (sparse_copy): Arg scan_holes replaces
the old hole_size arg. All callers changed.
(lseek_copy): Remove hole_size arg; no longer needed.
Caller changed.
* cfg.mk (sc_standard_outputs): Add a grep command for source files.
* src/du.c (main): Use standard input instead of stdin, standard output
instead of stdout, and standard error instead of stderr in messages.
* src/nohup.c (main): Likewise.
* src/sort.c (main): Likewise.
* src/split.c (main): Likewise.
* src/stdbuf.c (main): Likewise.
* src/wc.c (main): Likewise.
* tests/du/files0-from.pl (@Tests): Adjust test case to new messages.
* tests/sort/sort-files0-from.pl: Likewise.
* tests/wc/wc-files0-from.pl: Likewise.
'make syntax-check' complains:
src/tail.c
maint.mk: the above files include safe-read.h but don't use it
make: *** [maint.mk:737: sc_prohibit_safe_read_without_use] Error 1
The removal was missed for tail.c in recent commit d3c7072a09.
* src/tail.c (safe-read.h): Remove include.
This does not change behavior on POSIX platforms; it’s mostly to
make it clearer when we’re looking for file identity.
* src/cat.c (main):
* src/copy.c (struct dir_list, is_ancestor, copy_internal):
* src/tail.c (struct File_spec, record_open_fd, recheck)
(tail_forever_inotify, tail_file):
* src/test.c (binary_operator):
Use psame_inode, PSAME_INODE, or SAME_INODE instead of comparing
device and inode numbers by hand.
* src/tail.c (recheck, tail_file): Do not mark a file as tailable
merely because --retry is not in effect. Simplify internal logic.
This should not change behavior; it’s just for clarity and to
make the code match the comments better.
* src/tail.c (get_file_status): Remove, since after the changes
described below it would be called in just one place and it’s a
bit clearer to inline by hand.
(tail_file): Don’t call fstat after reading the file, as that
misses changes arriving between read and fstat. Instead, reuse
the fstat done before reading the file.
This matters only in some obscure cases hard to test for.
* src/tail.c (file_lines, pipe_lines, pipe_bytes, start_bytes)
(start_lines, tail_bytes, tail_lines, tail): New return
convention, which reports errno. All callers changed.
(recheck): Don’t lose track of errno if a regular file is
replaced by a symlink.
(get_file_status): Set errno to 0 on success.
(tail_file): Be more careful about f->errnum.
It is now -1 only if the failure was not due to
a system call failing.
When not already calling lstat for some other reason,
prefer readlink to lstat+S_ISLNK,
as readlink does not suffer from EOVERFLOW issues.
* src/rmdir.c (main):
* src/tail.c (recheck, any_symlinks):
* src/test.c (unary_operator):
Also, fix commentary to talk about “nonexistent” rather than
“dead” processes, since the code looks for the former not the
latter and the difference matters for zombies.
* src/tail.c (some_writers_exist): Rename from writers_are_dead,
negate the sense, don’t have a special and counterintuitive case
for !nbpids, remove PIDs found not to exist, and avoid some
though not all unlikely races when kernels reuse PIDs.
(tail_forever): Optimize via blocking I/O even if --pid was used,
so long as all the writers no longer exist.
(tail_forever, tail_forever_inotify): Simplify the writers_dead
logic; there is no need to have a local var to track this, since
we can use pids and nbpids now.
(parse_options): Also free and clear pids if !forever.
* src/tail.c (valid_file_spec, recheck, writers_are_dead)
(tail_forever, check_fspec, tail_forever_inotify, tail_file)
(parse_options, main): Be a bit more systematic about checking
for sign, rather than for exact equality or inequality,
when the sign is enough. Makes the code a bit clearer
now that -2 sometimes means success.
* src/tail.c (struct File_spec): New member read_pos, replacing
size, since the value was really a read position not a size.
All uses changed.
(xlseek): Move defn up.
(record_open_fd): If the read_pos (formerly) size arg is unknown,
compute it here if it is a regular file.
(file_lines): Return the resulting read pos (or -1 on failure)
instead of storing it via a pointer. Caller changed.
Simplify by using SEEK_CUR instead of SEEK_SET when that is easy.
Avoid reading the same data twice when there are not enough
lines in the file.
(pipe_lines): Return -2 on success, -1 on failure, rather than
updating a read pos via a pointer (which was weird for pipes anyway).
Caller changed.
(pipe_bytes, tail_bytes, tail_lines, tail):
Return -1 on failure, a file offset if known, and < -1 otherwise,
instead of storing a file offset via a pointer. Caller changed.
(pipe_bytes): Take initial file offset as an arg, or -1 if unknown.
(start_bytes, start_lines): Return -1 (not 1) on error, -2 (not
-1) on EOF, and do not accept pointer to read pos as an arg since
neither we nor our caller know the read pos. Callers changed.
(recheck): Do not assume a newly-opened file is at offset zero,
as this is not always true on Solaris.
(tail_forever, check_fspec): Use dump_remainder result only
on regular files, to prevent (very unlikely) overflow.
(tail_file): Remove no-longer-needed TAIL_TEST_SLEEP code.
In the old days, safe_read acted more like what full_read does now.
When that went away, some code that invoked safe_read should
have gone back to plain 'read' but I guess we never got around to it.
Simplify this code by going back to plain 'read'.
Use safe_read only in csplit.c, which has a signal handler
and where 'read' can therefore fail with EINTR.
Although safe_read also checks for oversize buffers,
that is better done via io_blksize.
* src/cat.c (simple_cat, cat):
* src/head.c (copy_fd, elide_tail_lines_pipe)
* src/tac.c (tac_seekable, copy_to_temp):
(elide_tail_lines_seekable, head_lines):
* src/tail.c (dump_remainder, file_lines, pipe_lines)
(pipe_bytes, start_bytes, start_lines, tail_forever_inotify):
* src/tr.c (plain_read):
Use plain 'read', not safe_read, since there is no
need to worry about signals or oversize requests.
Also, there is no longer a need to include safe-read.h.
* src/ioblksize.h: Include sys-limits.h, for SYS_BUFSIZE_MAX.
(io_blksize): Max out at SYS_BUFSIZE_MAX.
* src/tail.c (struct File_spec, xwrite_stdout, file_lines)
(pipe_lines, pipe_bytes, start_bytes, any_live_files)
(tail_forever, any_remote_file, any_non_remote_file)
(any_symlinks, any_non_regular_fifo, tailable_stdin)
(tail_forever_inotify, ignore_fifo_and_pipe, main):
Prefer a signed type to size_t, if possible.
Ordinarily this is idx_t, but use int when the value
must fit in int anyway.
(file_lines): Similarly for blksize_t, which had no business
being here anyway.
(main): Check for overflow in the oddball case where ptrdiff_t is
narrower than int.
Signed types let us debug better, by using -fsanitize=undefined.
* doc/local.mk (doc/constants.texi):
Adjust change from macro to enum.
* src/tail.c (COPY_TO_EOF, COPY_A_BUFFER)
(DEFAULT_MAX_N_UNCHANGED_STATS_BETWEEN_OPENS):
Now enum constants, not macros.
(COPY_TO_EOF, COPY_A_BUFFER): Now negative, not positive.
(count_t): New typedef. Use it instead of uintmax_t.
(COUNT_MAX): New macro; use it instead of UINTMAX_MAX.
(struct File_spec, max_n_unchanged_stats_between_opens)
(dump_remainder, file_lines, pipe_lines, pipe_bytes)
(start_bytes, start_lines, tail_forever, check_fspec)
(tail_forever_inotify, tail_bytes, tail_lines, tail, tail_file)
(parse_obsolete_option, parse_options, main):
Prefer count_t to uintmax_t.
* src/tail.c (file_lines): Fix an unlikely bug where ‘tail -n N’
could output more than N lines if standard input is a largish
regular file with large initial offset that starts with (say) N-1
lines after the initial offset, but grows to N+1 lines between the
fstat and read calls. In this case ‘tail -n N’ now outputs N-1
lines, not N+1; that is, it pretends the file grew after ‘tail’
read it. That is better than outputting more than N lines.
* src/tail.c: Use ‘prettyname’ consistently as the identifier
for a prettified file name, as opposed to ‘pretty_filename’,
‘pretty_name’, and ‘name’. This makes the code easier to follow.
(struct File_spec): New member prettyname.
(pretty_name): Remove.
All uses of pretty_name (f) replaced by f->prettyname.
(close_fd, fremote): Accept struct File_spec, not name.
All callers changed.
(main): Initialize the new prettyname member.
This is simpler/smaller than calling pretty_name everywhere.
* src/tail.c (tail_lines): Refactor to simplify the confusing
code for using SEEK_END when counting lines. The old code
had a ‘end_pos != 0’ expression that was always true.
* src/tail.c (tail_bytes): New function.
(tail_bytes, tail_lines, tail): Accept struct stat pointer from
caller instead of calling fstat ourselves. All callers changed.
(tail_file): Skip a call to fstat if fstat already failed.
* tests/tail/follow-stdin.sh: Adjust to match new behavior
on failure, which omits a redundant diagnostic.
When the user specifies -c N where 2**63 <= N, don’t give
up and use the slow method (which will exhaust memory if
the file is large). Instead, treat it as N = 2**63 - 1,
since that has equivalent effect.
* src/tail.c (tail_bytes): With -c N and large N, adjust
the code so that lseeks can still be used without
affecting correctness. Formerly the code gave up and
did a sequential pass through the whole input, which
could easily exhaust memory.
This better matches the treatment of POSIX form,
e.g., ‘tail +Nc’ is now like ‘tail -c +N’ even when N is large.
* src/tail.c: Don’t include xstrtol.h.
(parse_obsolete_option): Treat numbers greater than UINTMAX_MAX as
if they are UINTMAX_MAX. Parse the number by hand with saturating
arithmetic; nowadays that’s simpler than using xstrtoumax. There
is no need for a diagnostic now, as the error cannot happen any more.
* tests/tail/tail.pl (obs-plus-c3): New test.
* src/tail.c (tail_bytes): Simplify the -c+N case by treating
regular files like other files; if the lseek fails for whatever
reason, fall back on the unoptimized version instead of reporting
a fatal error.
This affects behavior only on memory objects, which are not in Linux.
Formerly the code would use lseek on these objects,
but POSIX says the result of lseek is unspecified on them,
and in QNX lseek has undefined behavior on typed memory objects.
* src/head.c (elide_tail_bytes_file, elide_tail_lines_file):
Omit unnecessary uses of presume_input_pipe.
Improve some out-of-date comments.
(head): Do not assume a file is seekable merely because its
st_size is usable. Instead, seek only on regular files.
* src/od.c (skip): Do not seek on memory objects.
This is not true on Solaris when opening /dev/stdin; it could be
at a nonzero file offset. Arguably Linux should do likewise.
* src/wc.c (wc): Omit last arg, and act as if it is always -1.
All uses changed.
* src/head.c (diagnose_read_failure): New function.
Use it when possible, to simplify callers.
(head_bytes): Use copy_fd and diagnose_copy_fd_failure instead of
reinventing the wheel.
* NEWS: Improve wording to not imply POSIXLY_CORRECT=1 readlink -q will
be verbose.
* src/readlink.c (usage): Mention the affect of POSIXLY_CORRECT on
-s (--silent), -q (--quiet), and -v (--verbose) in the help message.
(main): Remove spurious newline added by previous commit.
* doc/coreutils.texi (readlink invocation): Mention that -s (--silent)
and -q (--quiet) are not the default when POSIXLY_CORRECT is set.
* src/readlink.c (main): Set verbose if the POSIXLY_CORRECT environment
variable is set.
* tests/readlink/readlink-posix.sh: New file.
* tests/local.mk (all_tests): Add it.
* NEWS: Mention the change.
* doc/coreutils.texi (readlink invocation): Document the behavior of
POSIXLY_CORRECT.
* src/system.h (x_timestyle_args): A new function refactored from ...
* src/ls.c (decode_switches): ... here.
* src/du.c: Use refactored x_timestyle_args() to output a custom error.
Addresses https://bugs.gnu.org/79113
This was seen to trigger the EXIT trap on cygwin
* cfg.mk (sc_prohibit_colon_redirection): Disallow all
cases of :> in tests/
* tests/cksum/md5sum-bsd.sh: Adjust to avoid more
stringent syntax check.
This should help avoid further audit confusion,
such as was just fixed by removing a FIXME.
* src/test.c (enum binop): New type.
(get_mtime): Return a struct timespec instead of returning
a bool and storing a struct timespec. All callers changed.
(binop): Return an int recording either success (an enum binop)
or failure (-1). All callers changed.
(binary_operator): Accept an enum binop, so that we need
not recompute the op type. All callers changed. Simplify.
the fallback error in binary_operator() was unreachable, since invalid
binary operators are rejected in three_arguments() via binop() which
returns false and prevents entry into binary_operator().
this dead code was unreachable and safe to remove
ref: line 636 where "binary operator expected" is generated.
Signed-off-by: Harry Fellowes <harryfellowes1@gmail.com>
Copyright-paperwork-exempt: Yes
* gnulib: Update to latest to get the fix to
propagate the appropriate error message.
* tests/install/basic-1.sh: Add a test case.
* NEWS: Mention the bug fix.
Addresses https://bugs.gnu.org/79072
* src/date.c (show_date_helper): Add a use_c_locale parameter.
(batch_convert): Add a format_in_c_locale parameter.
(main): Set format_in_c_locale to true if any of the options --rfc-3339,
--iso-8601, -R is seen.
Problem reported by Cosima Neidahl <https://bugs.gnu.org/78985#13>.
* tests/sort/sort-float.sh: At top level, use C locale at first.
(dbl_minima_order): Assume C locale.
Use string comparison for the fractional parts.
2025-07-10 Paul Eggert <eggert@cs.ucla.edu>
tests: fix integer overflow in sort-float
Problem reported by Cosima Neidahl <https://bugs.gnu.org/78985>.
* tests/sort/sort-float.sh (dbl_minima_order):
Use expr instead of test, to avoid problems with integers
too large for the shell.
Problem reported by Cosima Neidahl <https://bugs.gnu.org/78985>.
* tests/sort/sort-float.sh (dbl_minima_order):
Use expr instead of test to compare fractions,
to avoid problems with integers too large for the shell.
* src/factor.c (factor_using_pollard_rho)
(factor_using_pollard_rho2, mp_factor_using_pollard_rho):
Use int_fast64_t for internal counters rather than int, as int
could overflow on some somewhat-practical examples. Problem
discovered on a hypothetical platform where W_TYPE_SIZE is neither
32 nor 64, when factoring 0x7ffffffffffffeab7fffffffffff7369 ==
170141183460469225450570946617781744489, causing k to overflow in
mp_factor_using_pollard_rho. Presumably a similar problem exists
in the previous stable coreutils 9.7, too, on 32-bit platforms
with somewhat-larger test cases, though I haven’t take the
somewhat-extensive CPU time to discover it.
* src/factor.c (mp_factor_using_pollard_rho):
Fix recently-introduced aliasing bug by computing q
before g gets updated in place. Problem discovered
on a hypothetical platform where W_TYPE_SIZE
is neither 32 nor 64.
* src/factor.c: Do not include c-ctype.h.
(strtouuint): Don’t bother generating a number on
error; just return a strtol_error value other than LONGINT_OK.
Speed up overflow checking.
* src/factor.c (primes_ptab): New table of primes, replacing
primes_diff and primes_diff8. All uses changed. This is simpler
and should improve performance slightly. Although this limits the
table’s primes to 2**15 instead of to 668221, the limit can easily
grow to 2**32 by changing the type of ‘prime’, without hurting
performance significantly compared to the primes_diff and
primes_diff8 approach.
* src/make-prime-list.c (output_primes):
For each prime p, output p instead of two differences.
* src/factor.c (factor_using_pollard_rho)
(factor_using_pollard_rho2, mp_factor_using_pollard_rho):
Use int, not unsigned long int, for counters that won’t go above
2**31 on practical platforms. This yields a significant speedup
on GCC 15 x86-64, and using signed values allows for automatic
checks for overflow when using gcc -fsanitize=undefined.
* src/factor.c (mp_factor_insert_ui): Rename fom
mp_factor_insert_ui, and change arg type from unsigned long int to
mp_limb_t. All uses changed. This avoids creating and freeing a
small mpz_t.
(mp_factor_using_division): Add a static assert requiring that
that mp_limb_t be wide enough, which it should be (and is in
standard GMP).
* src/factor.c (factor_using_pollard_rho)
(factor_using_pollard_rho2): Use mp_limb_t, not unsigned long int,
for a parameter. This avoids some casts, and avoids a theoretical
bug where converting to mp_limb_t loses info.
lt2 a bit more natural, given the current implementation.
* src/factor.c (lt2): New function.
(ge2): Rewrite in terms of lt2.
(gt2): Remove. All callers changed to use lt2.
* src/factor.c (umul_ppmm): When !USE_LONGLONG_H so we need to
define this, speed things up if there is an unsigned type uuroom_t
wide enough to hold two words. Do not make a similar change for
udiv_qrnnd, as it is not performance critical and anyway on GCC 15
x86-64 that would mean a subroutine call.
* src/factor.c (gcd_odd, gcd2_odd): Speed up, given that the
second argument is always odd.
(gcd_odd): Avoid recomputing a temporary.
(gcd2_odd): Test for zero only if a multiple of B.
This saves an ‘assume’.
* src/factor.c (submod2): Use ckd_sub to subtract by hand rather
than using sub_ddmmss plus a compare. This speeds things up a
bit, on x86-64 with GCC 15 anyway.
* src/make-prime-list.c (output_primes): Output
SQUARE_OF_FIRST_OMITTED_PRIME, not FIRST_OMITTED_PRIME. All uses
changed. This way, the uses don’t need to worry about casts to
avoid overflow.
* src/factor.c (make_uuint2): New function.
(powm2, millerrabin2): Pass two-word args as uuints,
not as mp_limb_t const [2] pointers. All uses changed.
(prime2_p): Rework to use the new API, fixing a FIXME.
* src/factor.c (factor_using_division, mp_finish_up_in_single)
(mp_finish_in_single, factor_using_pollard_rho)
(factor_using_pollard_rho2, factor_up, factor):
Put FACTORS arg first, for consistency.
This is just a refactoring and should not affect speed.
* src/factor.c (print_factors): Use single-precision code only
when word size is 32 or 64, as it isn't tested for other sizes and
is known to not work when it the size is 128. This change does
not affect any known practical platforms.
Suggested for consideration by Torbjörn Granlund in:
https://lists.gnu.org/r/coreutils/2025-01/msg00000.html
* src/factor.c (PROVE_PRIMALITY): Now defaults to false.
(mp_prime_p): Help the compiler by telling it mpz_prob_prime_p
returns nonnegative.
* tests/factor/create-test.sh (bigprime): Test 2^400 - 593,
since that’s now practical.
* tests/local.mk (factor_tests): Add new test.
* src/factor.c (USE_BAILLIE_PSW): New constant.
(prime_p, prime2_p): Use it, i.e., always use Baillie-PSW.
Do so by using mp_prime_p. Do not tell GCC these functions
are pure; the pure mark was present only to pacify GCC
and is no longer needed now that thes functions call mp_prime_p.
Inspired by a proposal by Torbjörn Granlund in:
https://lists.gnu.org/r/coreutils/2025-01/msg00000.html
* src/factor.c (mp_prime_p): Use mpz_probab_prime_p rather than doing
it by hand, as mpz_probab_prime_p uses Baillie-PSW which is a win.
This removes the need for the ret2 label and goto.
* src/factor.c (prime_p, prime2_p, mp_prime_p): Do not skip the
flag_prove_primality test for the last prime in the table, i.e.,
when r == PRIMES_PTAB_ENTRIES - 1.
(mp_prime_p): There is no longer a need for the ret1 label or goto.
mini-gmp lacs mpn_tdiv_qr, so supply an emulation of it
when using mini-gmp.
* src/factor.c (copy_mpn_from_mpz, mpn_tdiv_qr) [!mpn_tdiv_qr]:
New functions.
These changes are taken from a proposal by Torbjörn Granlund in:
https://lists.gnu.org/r/coreutils/2025-01/msg00000.html
On my x86-64 platform, they improve speed by more than 8× when
factoring 340282366920938463463374607431768211457.
* src/factor.c (mp_modadd, mp_modsub, mp_modadd_1, mp_mulredc):
New functions.
(MP_FACTOR_USING_POLLARD_RHO_N_MAX): New macro.
(mp_factor_using_pollard_rho): Act on mpn not mpz, and on
mp_limb_t not unsigned long int. Reorder args. All uses changed.
This is mostly to make the code a bit easier to read.
It shrinks the size of factor.o by 0.7% on my x86-64 platform
though it doesn’t affect CPU performance significantly.
* src/factor.c (mp_no_factors): Rename from mp_factor_init.
(mp_no_factors, mp_factor_using_division, mp_factor):
Return struct mp_factors rather than modifying one passed by reference.
All uses changed.
* src/factor.c (mp_factor_using_division):
When continuing in single precision, don’t divide by primes that
were already cast out in multiple precision.
On my platform this gave a 2.5% speedup when factoring
2**128 + 172261 = 4999 * 68070087401668026297934508388031283,
as W_TYPE_SIZE == 64 and 4999 is the last prime in the primes table.
* src/factor.c (mp_factor_using_division): Use index into
primes_diff consistently with other uses. This is mostly just a
style change, and likely doesn’t change the generated machine code.
This does not affect performance much, but it should allow future
performance improvements.
* src/factor.c (factor_using_division): Two new args I and P,
which generalize this function. All uses changed.
(mp_finish_up_in_single, factor_up): New functions, like the
non-*_up* versions but with two new args PRIME_IDX and PRIME.
They mostly just have the old body of the old non-*_up_ versions.
(mp_finish_in_single, factor): Rewrite in terms of the new functions.
This significantly improves performance when a number exceeds
2**(W_TYPE_SIZE - 1) and is the product of a prime less than
FIRST_OMITTED_PRIME and another prime less than 2**(W_TYPE_SIZE - 1).
On my platform, for example, it doubled the speed of factoring
4999 * (2**128 - 159).
* src/factor.c (mp_size, mp_finish_in_single): New functions.
(mp_factor_using_division, mp_factor_using_pollard_rho):
Finish using single precision when possible.
* tests/factor/factor.pl (lt-5000-times-128-bit): New test.
* src/factor.c (factor_insert_refind):
Use idx_t for indexes into primes_diff,
for consistency with other indexes into primes_diff.
This has no practical effect unless the primes_diff
table becomes unreasonably large.
Support a multiplicity argument in the mp case, too.
This helps keeps the two cases in sync, for maintenance.
* src/factor.c (mp_factor_insert, mp_factor_insert_ui):
New arg M, for multiplicity. All callers changed.
Use something other than a macro when that is easy and won’t hurt
performance.
* src/factor.c (__ll_B, __ll_lowpart, _ll_highpart) [!USE_LONGLONG_H]:
(MAX_NFACTS, highbit_to_mask, factor_insert, PRIMES_PTAB_ENTRIES):
Make these enums, or constants, or static functions instead of macros.
(highbit_to_mask): Rename from HIGHBIT_TO_MASK. All uses changed.
* src/factor.c (factor_insert_multiplicity):
Adjust to keep in sync with mp_factor_insert changes below,
by adding 1 to the index and using memmove to move.
(mp_factor_insert): Omit redundant call to mpz_cmp.
Prefer idx_t (always nonnegative) to ptrdiff_t,
by adding 1 to the indexes.
Prefer mpz_init_set to mpz_init+mpz_set.
Use memmove to move, rather than doing it by hand.
* src/factor.c (MAX_NFACTS): Allow word size of 128 bits,
even if this is only theoretical now.
Check that struct factors’s unsigned char counts won’t overflow.
* src/factor.c (USE_LONGLONG_H):
Default to false on unusual (but standard-conforming)
platforms that lack int64_t etc.
(UWtype, UHWtype): Now typedefs, not macros.
(UQItype): Remove.
(SItype, USItype, DItype, UDItype): Use standard C types.
This simplifies things slightly by using uuint for
some two-word integers.
* src/factor.c (strtouuint): Accept uuint *, not two mp_limb_t *.
All callers changed.
(print_factors_single): Accept uuint, not two limbs.
All callers changed.
(print_factors): Use simpler test for high bit,
one that need not worry about promoting to int.
Simplify by using GMP’s word type instead of pretending to roll our own.
* src/factor.c (wide_uuint): Remove. All uses replaced by mp_limb_t.
(umul_ppmm) [!umul_ppmm]: Don’t assume unsigned long is at least half
as wide as mp_limb_t. This simpler anyway.
(strtouuint): Rename from strto2wide_uint. All uses changed.
Remove experimental code for 128-bit words as it does not work and
we lack time to figure out why. Instead, ensure that words are
the same size as with GMP.
* src/factor.c (USE_INT128): Remove. All uses removed.
(wide_uint, W_TYPE_SIZE): Define to be the same as GMP.
(MP_LIMB_MAX): New macro. Check that it matches W_TYPE_SIZE.
(USE_LONGLONG_H): Default to true.
(UHWtype) [USE_LONGLONG_H]: Define to unsigned int, same as GMP.
(prime_p): Go back to not worrying about 128-bit words,
since GMP doesn’t worry and doesn’t use them.
(lbuf_putbitcnt): New function, since we cannot assume
that bitcnt_t fits into mp_limb_t.
(print_factors): Use it.
* src/make-prime-list.c (output_primes):
Don’t assume that wide_uint’s maximum is UINTMAX_MAX.
* src/factor.c (struct mp_factors): e (multiplicity) member
is now of type mp_bitcnt_t, not unsigned long int, since
its value is at most a bit count. All uses changed.
* src/factor.c (BIG_POWER_OF_10, LOG_BIG_POWER_OF_10):
Place fewer restrictions on BIG_POWER_OF_10.
This is only for currently-theoretical hosts;
it shouldn’t affect machine code on practical platforms.
* src/factor.c (wide_int): Remove, since it gets in the
way of using mp_limb_t for words. All uses removed.
(submod2, HIGHBIT_TO_MASK, divexact_21):
Rewrite without using wide_int.
This shouldn't change the machine code these days,
as compilers are pretty smart about isolating the
top bit of an unsigned int.
In practice there’s no bug but we might as well avoid the
undefined behavior.
* src/factor.c (hi_is_set): New static function.
(factor_insert_large, prime2_p, print_factors_single): Use it.
* src/local.mk: Similarly to commit v8.22-156-g09937e9d0
track speedlist.h with nodist_src_stty_SOURCES and DISTCLEANFILES
to ensure the make distcheck manifest comparison passes.
Addresses https://bug.gnu.org/78960
* src/local.mk: Use the coarser BUILT_SOURCES mechanism
to generate speedlist.h, rather than a specific dependency
(which did seem to work for parallel builds).
Fixes https://bugs.gnu.org/78960
* src/od.c (print_function_type): New type. Use it for convenience.
(width_bytes): Omit duplicate entries, such as ‘double’ vs ‘long
double’ on macOS. Problem reported by Bruno Haible
<https://bugs.gnu.org/78933>.
(decode_one_format): Cast null pointer to print_function_type
to pacify Apple clang-1400.0.29.202.
* src/local.mk: Adjust the dependency so that speedlist.h
is built irrespective of the object file name.
Note we could use BUILT_SOURCES for this,
but it's better to have this more accurate dependency.
Reinstate check removed in commit 56aa549a0 so that we
disallow -f2 when configured with utils_cv_ieee_16_bit_supported=no.
Otherwise the output routines will consume floats,
i.e. 4 bytes at a time. Without this extra check
the tests/od/od-endian.sh will fail with this configuration.
* src/od.c (decode_one_format): Reinstate the explicit check
for this configuration edge case.
Problem reported by Pádraig Brady <https://bugs.gnu.org/78880#43>.
This patch doesn’t fix any bugs; it merely pacifies GCC.
* src/od.c (ispec_to_format): New function, replacing
the old ISPEC_TO_FORMAT macro. All uses changed.
This part of the change is just refactoring.
(decode_one_format): Pacify à la ispec_to_format.
* src/od.c (width_bytes, decode_one_format): Don’t assume a signed
type has the same size as the corresponding unsigned type.
This has no effect on practical platforms; it’s just for
consistency there.
* src/od.c (address_base, address_pad_len, format_address):
Initialize statically rather than dynamically.
(limit_bytes_to_format): Remove. All uses replaced by
checking sign of end_offset.
(max_bytes_to_format): Remove static var. Now local to ‘main’.
(end_offset): -1 now means no limit. All uses changed.
On x86-64 (for example) print_long, print_long_long, and
print_intmax all behave identically, so give GCC enough info so
that it generates code for just one of these functions.
* src/od.c (enum size_spec): Arrange for enum values to
be the same if they represent types that behave the same.
(width_bytes, ISPEC_TO_FORMAT, decode_one_format):
Match the enum size_spec changes.
* src/od.c (FMT_BYTES_ALLOCATED): Use a simpler formula.
Although slightly too generous, the storage wasted is very small
and it pacifies gcc -Wformat-overflow=2.
(bytes_to_oct_digits, bytes_to_signed_dec_digits)
(bytes_to_unsigned_dec_digits, bytes_to_hex_digoits): Remove.
All uses replaced by algorithmic calculations, which are good
enough: they are valid for integers up to 2620 bits (!) and might
be slightly conservative for wider integers. Remove related
static_asserts, which are no longer needed.
This has practical effect only on hypothetical platforms where
uintmax_t is wider than unsigned long long int.
* src/od.c (enum size_spec): New constant INTMAX.
(MAX_INTEGRAL_TYPE_WIDTH): Now equals UINTMAX_WIDTH.
(FMT_BYTES_ALLOCATED): Allow for the extra "l" in "%lld".
Also, fix off-by-two error in size calculation.
(width_bytes, integral_type_size): Add entries for uintmax_t.
(print_intmax): New function.
(decode_one_function): Use it.
(ISPEC_TO_FORMAT): New arg Max_fmt. All uses changed.
* src/od.c (NO_SIZE): Make it explicitly 0, as the
initializers now rely on this.
(MAX_INTEGRAL_TYPE_SIZE): Remove. All uses replaced by
ARRAY_CARDINALITY (integral_type_size) - 1.
Move static assertion down to where this can be used.
(integral_type_size, fp_type_size): Make them const,
and initialize them statically.
(main): Omit no-longer-needed initialization code.
* src/od.c (MAX_ADDRESS_LENGTH, pseudo_offset, n_bytes_to_skip)
(max_bytes_to_format, end_offset, skip, format_address_none)
(format_address_std, format_address_paren, format_address_label)
(write_block, parse_old_offset, dump, dump_strings, main):
Prefer intmax_t to uintmax_t. This makes no practical difference,
and lets -fsanitize=undefined check for signed integer overflow.
(skip, dump): Remove no-longer-needed casts.
(xstr2nonneg): New static function. All callers of xstrtoumax
now call this function instead.
(main): Use ckd_add to detect signed integer overflow, since
the unsigned trick no longer works reliably.
Let xstrtol_fatal report the overflow, instead of doing
it by hand ourselves.
* src/od.c (parse_old_offset): First arg is now char *,
not char const *. If a decimal number, temporarily
modify the string so that xstrtoumax does not complain
about the '.'.
* tests/od/od.pl: Test for the bug.
* src/od.c (print_n_spaces, pad_at, pad_at_overflow):
New static functions.
(struct tspec, PRINT_FIELDS, print_named_ascii, print_ascii)
(decode_one_format, write_block, main):
Use idx_t, not int, for counts that depend on the number
of bytes in an object.
(decode_one_format): Use print_n_spaces to output spaces.
(PRINT_FIELDS, print_named_ascii, print_ascii):
Use pad_at to avoid integer overflow.
(write_block): Do not use %*s to pad, as the total pad might
exceed INT_MAX. Instead, pad by hand with putchar (' ').
(main): Use pad_at_overflow to report integer overflow due to
oversize -w. Use better way to tell whether -w is used,
without needing IF_LINT.
* tests/od/big-w.sh: New test.
* tests/local.mk (all_tests): Add it.
Also, fix minor related typos.
* src/od.c (MAX_INTEGRAL_TYPE_SIZE, MAX_ADDRESS_LENGTH):
Now a constant, not a macro.
(MAX_INTEGRAL_TYPE_WIDTH): New constant. Use it instead of
CHAR_BIT, so as not to assume that uintmax_t and unsigned long
long int are hole-free. This doesn’t matter on practical porting
targets, though there is still a mainframe or two that have holes.
(FMT_BYTES_ALLOCATED): Fix typo by changing "jd" to "jo".
Fix off-by-one typo in static assertion.
It’s long been safe to assume C99+ support for long long int.
* .gitignore: Remove m4/longlong.m4.
* bootstrap.conf (buildreq): Boost git prereq from 1.4.4 to 1.5.5,
syncing with Gnulib.
(bootstrap_post_import_hook): Remove m4/longlong.m4.
* m4/jm-macros.m4 (gl_CHECK_ALL_TYPES):
No need to require AC_TYPE_UNSIGNED_LONG_LONG_INT.
* src/factor.c (DItype, UDItype):
* src/od.c (main):
Assume HAVE_LONG_LONG_INT.
* src/od.c: (unsigned_long_long_int):
Remove. All uses replaced with unsigned long long int.
* src/od.c (dump, dump_strings): Use idx_t allocators
rather than size_t allocators, to avoid unchecked integer
overflow on theoretical platforms where SIZE_MAX < IDX_MAX.
* doc/coreutils.texi (cksum common options): Reorder and tweak the info
to make it clearer that --check does not support the legacy crc output
from the cksum command.
Reported at https://bugs.debian.org/1108363
* src/od.c (main): Don't pass LONGINT_OK to xstrtol_fatal(),
as otherwise it will abort().
* tests/od/od.pl: Add test cases.
* NEWS: Mention the bug fix.
Addresses https://bugs.gnu.org/78879
* src/od.c (dump_strings): There are three related issues here
due to not accounting for the terminating NUL char appropriately.
1. Ensure BUF always has enough space for the terminating NUL.
This avoids CWE-122: Heap-based Buffer Overflow,
where we wrote a single NUL byte directly after the allocated buffer.
I.e., there should be no buffer overflow with:
printf '%100s' | od -N100 -S1
2. Ensure we support -S == -N (END_OFFSET - STRING_MIN == ADDRESS):
I.e., there should be output with:
printf '%100s' | od -N10 -S10
3. Ensure we always output a valid address by ensuring
the ADDRESS and I variables are kept in sync.
I.e., this should output address 0000000 not 1777777777777777777777:
printf '%100s' | od -N10 -S1
As well as fixing these we simplify by using a single loop
to read the data, rather than two.
* doc/coreutils.texi (od invocation): Clarify that -N
implicitly NUL terminates strings.
* tests/od/od-N.sh: Add test cases.
* NEWS: Mention the bug fixes.
Fixes https://bugs.gnu.org/78880
Add a comment making it clear that it is perfectly safe to add
additional speeds to the canned list in the speedgen script, and under
what conditions it is appropriate to do so.
Signed-off-by: "H. Peter Anvin" (Intel) <hpa@zytor.com>
For generating src/speedtest.h, try a sequence of command-line options
until (or unless) one of them works:
-dM: gcc, clang and derived compilers, icc classic
-xdumpmacros: Sun Studio (writes to stderr!)
-qshowmacros: IBM XL classic
-PD: MSVC (usable with a wrapper such as cccl from the SWIG project)
Because Sun Studio -xdumpmacros unconditionally writes to stderr,
capture stderr output instead of sending it to /dev/null. This is
perfectly safe, even in the presence of stray stderr output, because:
1. speedgen ignores input that is not of the form #define B<number>
2. even if a line of that format would somehow spuriously appear,
the only outcome is that the generated C code will probe for a few
more macros.
Signed-off-by: H. Peter Anvin (Intel) <hpa@zytor.com>
* doc/coreutils.texi (stty invocation): Remove now imprecise
list of speeds given we may now support higher or arbitrary speeds.
Mention that we may support higher or arbitrary speeds.
* src/tty.c (TTY_USAGE): Rename from TTY_FAILURE, since this
is used only for usage failures. All uses changed.
(TTY_TTYNAME_FAILURE): New constant.
(main): Remove no-longer-needed assignment of ENOENT to errno.
Make status-setting clearer too.
Report an error if ttyname fails even though stdin is a terminal,
instead of silently pretending that stdin is not a terminal.
* tests/tty/tty.sh: Test for this issue. This should fix Bug#78244.
* tests/stty/stty-invalid.sh: Adjust to what is now invalid.
* tests/stty/stty.sh: Add checks for valid speed variants.
* tests/stty/bad-speed.sh: New test to ensure unsupported speeds
are diagnosed.
Support the case where speed_t is simply a number, and in that case
assume that arbitrary values can be passed. This is assumed to be the
case when all known speed_t macros equal their own value.
Try to probe for a variety of speed_t constants by trying to coax
$(CC) into emitting macro definitions (-E -dM). If this is not
supported, use a fairly extensive list of constants as a
fallback. This both improves the test for arbitrary speed support, as
well as allowing proper operation in the case where the constants are
not plain numbers and allows for handing enumerated speed constants
that were not known a priori when the source code was written.
A simple shell script (mostly using sed) is used to turn the list of
constants (probed and predefined) into a pair of conversion functions,
baud_to_value() and value_to_baud(); string_to_baud() is then
reimplemented as a wrapper around the latter.
* src/.gitignore: Add generated speedlist.h.
* src/local.mk: Generate speedlist.h.
* src/speedgen: Portable shell script to generate speedlist.h.
* src/stty.c: Adjust string_to_baud to
convert from arbitrary numeric values.
* src/termios.c: A helper used when generating speedlist.h
* NEWS: Mention the improvement.
Problem reported by Collin Funk in:
https://lists.gnu.org/r/bug-gnulib/2025-06/msg00094.html
* tests/date/date-debug.sh: Also allow NetBSD 10 mktime behavior.
Although NetBSD contradicts POSIX, POSIX is likely wrong here and
I vaguely recall that there’s a POSIX correction in progress
that will allow the NetBSD behavior.
* doc/coreutils.texi: Use @dots{} after [OPTION],
to be consistent with man pages.
* src/printenv.c (Usage): Remove unneeded "...".
* src/timeout.c (Usage): Add needed "...",
and also remove redundant [OPTION] only form.
* src/chroot.c (Usage): Likewise.
Fixes https://bugs.gnu.org/78628
This compiler lacks support for the x86-64 VPCLMULQDQ instruction;
compiles fail with "error: invalid cpu feature string for builtin".
Problem reported by Peter Dyballa, and fix suggested by
Pádraig Brady <https://bugs.gnu.org/78562#8>.
* configure.ac (USE_AVX512_CRC32):
Also check that __builtin_cpu_supports ("vpclmulqdq") compiles.
Recent changes in Gnulib guarantee O_DIRECTORY != 0.
* gl/lib/targetdir.c (target_directory_operand):
* src/dd.c (usage):
* src/ln.c (main):
Simplify now that O_DIRECTORY must be nonzero.
Do not link to fchownat, fchmodat, xreallocarray.
While we’re at it, depend as little as possible on problematic
functions like fileno, strtoumax. Use only functions that
were around in C89.
Problem reported by Peter Dyballa <https://bugs.gnu.org/78509#59>.
* src/libstdbuf.c: Include stddef.h, stdlib.h, gettext.h.
Do not include stdint.h or system.h.
(_): New macro, since we no longer include system.h.
Undef strtoul, not strtoumax.
(fileno_to_name): Remove. All uses removed.
(apply_mode): New arg STREAM_NAME, so that we no longer need
to depend on fileno which might run into a Gnulib workaround.
Don’t allocate more than ULONG_MAX - 1 bytes, as that’s
not helpful in the real world and this lessens dependency
on newer features like strtoumax.
These tests were seen to hang on PPC Mac OS X 10.4.11
* tests/cp/existing-perm-race.sh: Protect fifo interactions
with a 1 minute timeout.
* tests/cp/file-perm-race.sh: Likewise.
Reported in https://bugs.gnu.org/78509
* src/sort.c (begfield): Check pointer adjustment
to avoid Out-of-range pointer offset (CWE-823).
(limfield): Likewise.
* tests/sort/sort-field-limit.sh: Add a new test,
which triggers with ASAN or Valgrind.
* tests/local.mk: Reference the new test.
* NEWS: Mention bug fix introduced in v7.2 (2009).
Fixes https://bugs.gnu.org/78507
On so-far-only-theoretical platforms with 128-bit uintmax_t,
'factor' would misbehave by not factoring enough.
Work around the bug (at a performance cost)
and document the issue. I hope someone with
more time and expertise can fix the performance cost
that this introduces.
To reproduce the correctness bug, build with
'gcc -DUSE_INT128 -DEXHIBIT_INT128_BUG';
'make check' should fail due to the new test case.
* src/factor.c (USE_INT128): New macro.
(wide_uint, wide_int, W_TYPE_SIZE, WIDE_UINT_MAX):
Define to proper values if USE_INT128.
(prime_p) [!EXHIBIT_INT128_BUG]: Work around bug
with 128-bit wide_uint, at some performance cost.
* tests/factor/factor.pl (bug-with-128-bit-uintmax_t):
New test.
This prepares for using unsigned __int128 instead of uintmax_t.
It doesn’t change behavior.
* src/factor.c (wide_uint, wide_int):
New typedefs, replacing all uses of uintmax_t, intmax_t.
(WIDE_UINT_MAX): New macro, replacing all uses of UINTMAX_MAX.
(USE_LONGLONG_H): Don’t use LONG_MAX and INTMAX_MAX (which
should have been ULONG_MAX and UINTMAX_MAX anyway).
Instead, use W_TYPE_SIZE and ULONG_WIDTH, as that will be
easier to work with if we use unsigned __int128.
(binv): Also unroll for the 128-bit case.
* src/make-prime-list.c (print_wide_uint):
Print "wide_uint" not "uintmax_t".
* src/local.mk: Compile the make-prime-list in a temporary directory
using $(BUILD_CC) so it can be run even when $(CC) is a cross-compiler.
Add some comments.
(noinst_PROGRAMS): Remove src/make-prime-list.
(src_make_prime_list_LDADD): Remove variable.
* src/make-prime-list.c: Don't include config.h and attributes.h.
Remove unnecessary #undefs
(ATTRIBUTE_CONST, ATTRIBUTE_MALLOC): Define for the host compiler.
Fixes https://bugs.gnu.org/78377
* .gitignore: Add the soon to be generated lib/fts_.h
* cfg.mk (sc_gitignore_missing): Special case the mapping
of fts.in.h to fts_.h.
Fixes https://bugs.gnu.org/78398
* src/factor.c (mulredc2): Remove two ‘affirm’ calls that didn’t
match the accompanying comment, and one of which has a false
positive if UINTMAX_WIDTH == 128 and we factor 2**128 + 1.
* configure.ac (GNULIB_TEST_WARN_CFLAGS):
Do not use -Wsuggest-attribute=cold, -Wsuggest-attribute=const,
-Wsuggest-attribute=format as they produce false positives with
GCC 15 x86-64. Use -Wmissing-variable-declarations, as it no
longer seems to hurt.
* configure.ac: Simplify the setup of GCC’s -W... options,
by removing options no longer needed for GCC 15 (when
configuring with --enable-gcc-warnings) or GCC 12 (without).
I hope this is good enough nowadays.
Add GCC 15’s -Wtrailing-whitespace, since that works for us.
* init.cfg (require_membership_in_two_groups_): Skip
overflow gids used in user namespaces, as one can't
chgrp() to these.
Fixes https://bugs.gnu.org/78225
* tests/ls/acl.sh: In constrained environments, like with
setxattr() disabled for example, require_acl_ does not
check for the required functionality, so use the more
stringent require_setfacl_ instead.
Reported at https://github.com/NixOS/nixpkgs/issues/394953
* tests/misc/csplit-heap.sh: More memory is required to avoid
a false failure on CheriBSD with its heap accounting overhead.
This is confirmed to still trigger with the original memory leak
being tested.
Avoid warnings like this from GCC 15:
src/basenc.c:1139:20: error: initializer-string for array of 'char'
truncates NUL terminator but destination lacks 'nonstring' attribute
(9 chars into 8 available) [-Werror=unterminated-string-initialization]
* src/basenc.c (z85_encoding, do_decode): Mark two more variables as
non-terminated.
Also, change sleep and tail to not sleep less than requested.
* bootstrap.conf (gnulib_modules): Add dtimespec-bound.
* gl/lib/dtimespec-bound.c, gl/lib/dtimespec-bound.h:
* gl/modules/dtimespec-bound: New files.
* src/sleep.c, src/tail.c, src/timeout.c: Include dtimespec-bound.h.
* src/sleep.c, src/tail.c: Don’t include xstrtod.h.
* src/sleep.c (apply_suffix, main):
* src/tail.c (parse_options):
* src/timeout.c (apply_time_suffix):
Don’t sleep less than the true number of seconds.
* src/timeout.c: Don’t include ctype.h.
(is_negative): Remove; no longer needed.
(parse_duration): Use a slightly looser bound on the timeout, one
that doesn’t need -lm on GNU/Linux. Clear errno before calling
cl_strtod.
This was seen to add about 100,000 ns to the startup time,
on a 2.6 GHz i7-5600U with glibc 2.40.
* .gitignore: Remove /lib/fenv.h.
* bootstrap.conf: Remove fenv-rounding and signbit deps.
* src/local.mk: Remove fenv lib dependency.
* src/timeout.c (is_negative): A new helper function to
be equivalent of signbit in the underflow case.
(parse_duration): Remove the rounding up logic,
as a nanosecond here or there has no significance.
* doc/coreutils.texi (sleep invocation): Mention that suffixes are
best avoided with hex arguments.
(timeout invocation): Likewise.
* tests/misc/sleep.sh: Ensure 'd' is not interpreted as "day".
* .gitignore: Add /lib/fenv.h to ignore list.
* tests/timeout/timeout-parameters.sh: Use a sleep length of 10s
to be consistent with the pattern where we use this larger time
when it does not slow down a test, but also provides protection
against a hung test, and better avoidance of false failures due
to races on very loaded systems. Also fix the setting of FAIL.
* tests/timeout/timeout-large-parameters.sh: Remove duplicated test.
This handles timeouts like 16777216.000000001 correctly;
formerly the subsecond part of that timeout was ignored.
* bootstrap.conf (gnulib_modules): Add fenv-rounding, signbit.
* src/local.mk (src_timeout_LDADD): Append $(FENV_ROUNDING_LIBM).
* src/timeout.c: Include fenv.h, math.h.
Don’t include xstrtod.h, as xstrtod’s checking now gets in the way.
(parse_duration): Round up when calling cl_strtod.
Check for -1e-1000. Don’t double-round 1e-9.
* tests/timeout/timeout-parameters.sh: Test for -0.1,
-1e-1000, 1e-1000.
* cfg.mk (codespell_ignore_words_list): Ignore false-positives.
(exclude_file_name_regexp--sc_codespell): Skip some file names.
* doc/coreutils.texi (mktemp invocation): Use "alphanumeric" which is
consistent with the rest of the documentation.
* src/expand-common.c: Fix typo.
* src/ls.c: Likewise.
* tests/split/l-chunk-root.sh: Likewise.
* tests/sort/sort-h-thousands-sep.sh: sv_SE defaults to UTF-8
on macOS 18, so avoid the test for multi-byte separators.
Fixes https://bugs.gnu.org/77509
* src/ls.c (gobble_file): Indicating unknown ACL info with '?'
suffices for the edge case of a file being removed while reading,
or older cygwin when reading through dangling symlinks.
Reported by Corinna Vinschen.
This was seen on termux on Android with ./configure --disable-xattr
where listxattr() and getxattr() returned ENOTSUP.
Then the valid security context obtained by file_has_aclinfo()
was discounted, and problematically then freed multiple times.
Reported at https://github.com/termux/termux-packages/issues/23752
* src/ls.c (file_has_aclinfo_cache): Only discount the returned
acl info when all components are defaulted due to being unsupported.
This isn't strictly historically accurate
but most practical these days, especially since
systemd uses this as its default TERM type.
See https://github.com/coreutils/coreutils/issues/96
Tested with:
$ LS_COLORS= COLORTERM= TERM=vt220 src/ls --color
$ COLORTERM= TERM=vt220 src/dircolors
* src/dircolors.hin: Add vt220.
* configure.ac (LIBCRYPTO_SONAME): Store library name in cache so we
do not end up with an empty value for it when a cache file is used.
The configure variable name is changed from utils_cv_dlopen_libcrypto
to utils_cv_libcrypto_soname.
* who.c (scan_entries): Account for guessed tty names (e.g.
'sshd pts/1') from the readutmp module when using the systemd backend.
* bootstrap.conf (gnulib_modules): Add str_endswith.
* News: Mention the bug fix.
Addresses https://bugzilla.redhat.com/2343998
Current Gnulib arranges for fts debugging if GNULIB_FTS_DEBUG
is defined, so key off that rather than off DU_DEBUG.
* src/du.c (fts_debug): Remove decl, as Gnulib does this now.
(FTS_CROSS_CHECK): Remove; all uses removed.
(FTS_DEBUG) [!GNULIB_FTS_DEBUG]: Remove.
(long_options) [GNULIB_FTS_DEBUG]: Add a ---debug option.
(du_files): Call fts_cross_check only if fts_debug and GNULIB_FTS_DEBUG.
(main): Set fts_debug if GNULIB_FTS_DEBUG, not DU_DEBUG.
* src/chown-core.c, src/copy.c, src/cp-hash.c, src/csplit.c:
* src/expand-common.c, src/find-mount-point.c, src/force-link.c:
* src/group-list.c, src/iopoll.c, src/operand2sig.c:
* src/show-date.c, src/wc_avx2.c:
Omit unnecessary ‘extern ’ at the start of function defns.
This is less wordy, makes it a bit easier to grep for issues such
as the missing consistency checking in cksum.
* src/cksum_avx2.c, src/cksum_avx512.c, src/cksum_pclmul.c:
* src/cksum_vmull.c:
Include cksum.h instead of copying its decls/includes by hand.
This is a better way to ensure consistency among defns and uses.
* src/cksum.c [CRCTAB]: Include only config.h and stdio.h,
to simplify the crctab-generating code.
[!CRCTAB]: Do not include stdint.h or stdio.h, as cksum.h does it now.
(BIT, r, crc_remainder, main) [CRCTAB]: Use unsigned int, not
uint_fast32_t or uint32_t, as this is good enough for GNU where
unsigned int is guaranteed to be at least 32 bits, and this way we
needn’t worry about mismatches between %08x formats and uint_fast32_t.
(main) [CRCTAB]: Prefer more-local decls. Do not output
unnecessary directives to include stdint.h or stdio.h.
No need for ‘return EXIT_SUCCESS;’ nowadays.
* src/crctab.c: Regenerate.
* src/cat.c (main): Do not fail with plain ‘cat’ where input and
output are both /dev/tty, if the output happens to have O_APPEND set.
Problem reported by lilydjwg <https://bugs.gnu.org/76255>.
Also, don’t report an error if the seek position is at or after EOF,
even if O_APPEND is set.
* local.mk: Add the --loose-indent option, which results in help2man
avoiding extraneous new lines in expr.1, id.1, numfmt.1, shred.1,
tail.1, and timeout.1.
Fixes https://bugs.gnu.org/74107
* man/help2man: sync changes to commit 8fe02612
The main change here is to Use \f(CR for monospace text
when using groff in troff mode.
Previously \f(CW was used, but that's not portable.
I thought of a way to pacify -Wswitch-enum without much trouble.
Either add all the enums, or if that’s too verbose use ‘switch (+E)’
to indicate to the reader that there need not be a case for
every enum value. Since this approach improves static checking,
make the change everywhere and check it with -Wswitch-enum.
* configure.ac: Compile with -Wswitch-enum if it works and
--enable-gcc-warnings. No need to remove -Wswitch-default
since Gnulib no longer adds it.
* src/chmod.c (describe_change):
* src/chown-core.c (describe_change):
* src/copy.c (copy_debug_string, copy_debug_sparse_string):
* src/df.c (decode_output_arg, get_dev):
* src/du.c (main):
* src/factor.c (print_factors):
* src/head.c (diagnose_copy_fd_failure):
* src/ls.c (time_type_to_statx, calc_req_mask)
(decode_line_length, get_funky_string, parse_ls_color)
(gobble_file, print_long_format):
* src/split.c (main):
* src/sync.c (sync_arg):
* src/tr.c (is_char_class_member):
* src/wc.c (main):
Add switch cases to pacify -Wswitch-enum.
* src/copy.c (copy_debug_string, copy_debug_sparse_string):
Add unreachable () for unreachable cases.
* src/digest.c (main):
* src/od.c (decode_one_format):
* src/tr.c (get_next, get_spec_stats):
switch (E) → switch (+E).
* src/digest.c (main):
* src/tr.c (get_next):
Omit unnecessary ‘default: break;’ that merely pacified GCC,
as the new pacification style is better.
* src/ls.c (decode_line_length):
Add default unreachable case to prevent warning that function
might not return a value.
(gobble_file): Distinguish DEREF_NEVER from unreachable cases.
* src/ls.c (usage): Use parentheses to be less ambiguous as
to what are WORDs and equivalent short options. This is also
consistent with the description of --sort and --indicator-style.
Fixes https://bugs.gnu.org/75916
* src/system.h (emit_ancillary_info): Output
PACKAGE_PACKAGER_BUG_REPORTS if the build is configured
--with-packager-bug-reports.
Reported by Bruno Haible.
* src/sort.c (key_warnings): Remove "note " from the start
of a usually informational message, as this simplifies translation.
* tests/sort/sort-debug-warn.sh": Adjust accordingly.
Fixes https://bugs.gnu.org/75763
* tests/env/env-S.pl (cf): Remove uses of 'my' after the variable has
been declared.
* tests/factor/factor.pl (t): Likewise.
* tests/misc/fold.pl (prog): Remove duplicate assignment.
* cfg.mk (sc_prohibit_bare_set): A new syntax check to
ensure we protect use of set with '--', so that args
beginning with '-' are not interpreted as options,
and if no args are present, all existing args are cleared.
* tests/cp/symlink-slash.sh: Add -- to unprotected use of set.
* tests/ls/ls-time.sh: Likewise.
* tests/ls/symlink-slash.sh: Likewise.
* tests/mkdir/perm.sh: Likewise.
* tests/mkdir/selinux.sh: Likewise.
* tests/mkdir/smack-no-root.sh: Likewise.
* tests/mkdir/smack-root.sh: Likewise.
* tests/mv/part-hardlink.sh: Likewise.
* tests/nice/nice.sh: Likewise.
* tests/stty/stty-row-col.sh: Likewise.
* src/ls.c (main): Flag that we need to stat()
if we're going to get security context (call file_has_aclinfo_cache).
(file_has_aclinfo_cache): Be defensive and only lookup the device
for the file if the stat has been performed.
(has_capability_cache): Likewise.
* tests/ls/selinux-segfault.sh: Add a test case.
* NEWS: Mention the bug fix.
Reported by Bruno Haible.
* tests/head/head-c.sh: Pass a more similar operation
to get_min_ulimit_v_, so we get a more appropriate limit.
This was seen to be significant with CheriBSD.
* src/yes.c (main): Don't reuse the argv array as CHERI's
capability bounds do not allow for that, failing like:
$ yes $(seq 156) | head -n1
In-address space security exception (core dumped)
This was flagged on CheriBSD on ARM Morello with the error:
"In-address space security exception (core dumped)"
triggered with: tac -s '' /dev/null
* src/tac.c (main): Ensure we don't read beyond the
end of the supplied optarg.
* tests/df/skip-duplicates.sh: Just skip this test if we fail
to build the shared lib. This fails on Solaris 11 at least
due to no HAVE_MNTENT_H. Note HAVE_SYS_MNTENT_H does not
suffice for this wrapper code.
* tests/df/no-mtab-status.sh: Likewise.
* init.cfg (ulimit_supported_): skip_ if the ulimit -v
takes too long, which was seen with bash 5.2 on Solaris 11,
where fork() returned EAGAIN under memory constraints,
and bash retried for about 16 seconds.
(get_min_ulimit_v_): Break early if skipped.
* tests/misc/write-errors.sh: Be more conservative and
skip on failure to determine min ulimit.
* tests/tail/tail-c.sh: On Solaris 11, tail -c 4096 /dev/urandom,
will induce an lseek(,-4096,SEEK_END) which returns -4096 without
setting errno, and a subsequent read() then gives EINVAL.
Since tailing the end of a psuedo device is an edge case,
we just verify that we don't spin reading the device forever.
* src/cksum.c (cksum_fp_t): New typedef.
(pclmul_supported, avx2_supported, avx512_supported)
(vmull_supported): Return this new type instead of bool.
All callers changed. That way, callers do not need to
refer to functions like cksum_avx512 that might not
exist on this platform. Although GCC optimizes such
references away, the C standard does not require this
optimization.
* src/ls.c (print_dir): Fix bug: file_failure can set errno to
something other than EOVERFLOW but the code assumed it didn’t.
Also, omit ENOENT bug workaround with glibc 2.3 and later,
for consistency with Gnulib.
* src/tail.c (tail_file): Fix precedence issue introduced
in commit v9.5-231-g177fcec66 so that we pass correct flags to open().
Effectively this meant we would have dropped the O_BINARY flag
on windows, since O_RDONLY is generally 0.
Issue spotted by coverity.
* gnulib: Pick up gnulib commit f11caad4fd which ensures
we diagnose the actual utility name, and not just "coreutils"
when in single binary mode. This adjustment is required
since gnulib commit 959152ba37 which enforced use of gnulib's
error() once verror is used, and gnulib's error() always
outputs the base name of the command, which the new gnulib
commit now keeps up to date.
* tests/misc/write-errors.sh: Reset SIGPIPE to the default (terminate)
disposition, so that the test doesn't erroneously fail due to an
ignored SIGPIPE in the test environment.
* gnulib: Update to latest to pick up gnulib commit 05c63bc908
which ensures accurate determination of the presence of NFSv4 ACLs.
* NEWS: Adjust accordingly.
Related to https://bugs.gnu.org/74692
* src/csplit.c (get_first_line_in_buffer): Don't exit here
upon empty input, rather indicate no input in the return
to let callers handle in a more consistent fashion.
* NEWS: Mention the bug fix.
* tests/csplit/csplit.sh: Add a test case.
Reported by Daniel Hofstetter.
NFS (on Linux 6.12 at least) was seen to return EACCES
from listxattr() for files without read access.
We started using listxattr() in coreutils 9.4.
* src/ls.c (gobble_file): Map EACCES from file_has_aclinfo()
to '?', rather than displaying the error.
* doc/coreutils.texi (ls invocation): Document the '?' flag.
* NEWS: Mention the bug fix.
Addresses https://bugs.gnu.org/74692
* src/tail.c (tail_file): Open files with O_NONBLOCK
if we might need async processing.
(pipe_bytes): Ignore EAGAIN read() errors.
(pipe_lines): Likewise.
* tests/tail/pid-pipe.sh: Add a new test.
* tests/local.mk: Reference the new test.
* NEWS: Mention the bug fix.
Reported by Berhard Voelker.
* src/copy.h: Change update member from bool to enum.
* src/copy.c: s/interactive == I_ALWAYS_NO/update == UPDATE_NONE_FAIL/;
s/interactive == I_ALWAYS_SKIP/update == UPDATE_NONE/;
s/update/update == UPDATE_OLDER/;
* src/install.c: Init with UPDATE_ALL, rather than false.
* src/cp.c: Likewise. Simply parse -f,-i,-n to x.interactive,
and parse --update to x.update.
* src/mv.c: Likewise.
* tests/cp/cp-i.sh: Add a test case where -n --update -i
honors the --update option, which would previously have been
ignored due to the preceding -n.
Since coreutils 9.3 we had --update={all,older} override -i.
In coreutils 9.5 this was expanded to -u
(to make it consistent with --update=older).
This patch reinstates things so that -i combines with -u instead.
I.e. have -i be protective, rather than selective (like -u).
The -f option of mv is similarly adjusted in this patch,
so now --update does not override any of -f,-i,-n.
* NEWS: Mention the bug fix.
* src/cp.c (main): Don't have -u disable prompting.
* src/mv.c (main): Likewise.
* tests/cp/cp-i.sh: Add a test case for -i.
* tests/mv/update.sh: Likewise.
* tests/mv/i-3.sh. Add a test case for -f.
Fixes https://bugs.gnu.org/70887
* doc/coreutils.texi (mv invocation): Be less ambiguous,
in that -f is significant for any replacement operation
on the destination, not just unlinking.
Update to latest gnulib with new copyright year.
Run "make update-copyright" and then...
* gnulib: Update included in this commit as copyright years
are the only change from the previous gnulib commit.
* tests/init.sh: Sync with gnulib to pick up copyright year.
* bootstrap: Likewise.
* tests/sample-test: Adjust to use the single most recent year.
* src/numfmt.c (simple_strtod_human): Only look for 'i'
after detecting a suffix.
* tests/misc/numfmt.pl: Add a test case.
* NEWS: Mention the bug fix.
Reported at https://bugs.debian.org/1091758
Problem reported by Tim Connors <https://bugs.gnu.org/75208>.
* doc/coreutils.texi (Date conversion specifiers):
* src/date.c (usage):
Warn about ambiguous formats like %D.
* gl/lib/strnumcmp-in.h (ISDIGIT):
* src/system.h (ISDIGIT): Remove. All uses replaced by c_isdigit,
with appropriate inclusions of c-ctype.h. This is more regular,
and is more portable to existing (but unlikely) platforms where
INT_MAX == UINT_MAX.
The 0 (EXIT) signal is valid as input
(and useful to determine existence of a pid),
so list it along with other signals.
* doc/coreutils.texi (signal specifications): Document 0, "EXIT".
* src/kill.c (list_signals): Start loops at 0, not 1.
* tests/misc/kill.sh: Add a test case.
* NEWS: Mention the change in behavior.
* src/tail.c (tail_forever): Without --retry, exit with failure
status like we do for the inotify case (since v8.11-15-g61de57cd2).
This is also consistent with the failure exit if no file was
accessible at tail startup.
* tests/tail/follow-stdin.sh: Tweak due to earlier exit.
* tests/tail/follow-name.sh: Test with and without inotify.
* NEWS: Mention the bug fix.
* bootstrap.conf: Depend on crc-x86_64 rather than crc.
* gnulib: Update to latest.
* src/cksum.c (crc32b_sum_stream): Add --debug info.
* NEWS: Mention the performance improvement.
Require --retry to continue to track files upon rename.
We already unfollowed a file if it was renamed
to another file system (unlinked), so this makes the behavior
consistent if renaming to a file in the same file system.
I.e. --follow=name without --retry, means unfollow if the
name is unlinked or moved, so this change ensures that
behavior for all rename cases.
Related commits: v8.0-121-g3b997a9bc, v8.23-161-gd313a0b24
* src/tail.c (tail_forever_notify): Remove watch for a renamed file
if --retry is not specified.
* tests/tail/F-vs-rename.sh: Related test cleanup.
* tests/tail/follow-name.sh: Add a test case.
* NEWS: Mention the bug fix.
Fixes https://bugs.gnu.org/74653
* tests/ls/dired.sh: macOS normalizes unicode characters to decomposed
(NFD) form when storing names in the file system, which breaks the
round-trip comparison employed by the test. So instead use a character
which does not decompose; verified with:
echo æ | uconv -f utf8 -t utf8 -x nfd | od -Ax -tx1z
* tests/dd/skip-seek-past-file.sh: Do not assume that
seek to exactly OFF_T_MAX should fail; it works on macOS 12.6
and POSIX allows this. Come to think of it, it should work
on Solaris too, if someone ever comes across a Solaris host
with a file system that allows such files.
Oracle Developer Studio 12.6 for sparc mishandles
‘sizeof ((char []) {'x', 'y'})’: it says
“warning: null dimension: sizeof()” and then generates
the wrong length in data. Work around the compiler bug
by counting sizes by hand, which may be a bit clearer anyway,
if a bit more error-prone.
* src/ls.c (BIN_STR): Remove.
(color_indicator): Spell out instead of using BIN_STR.
* tests/printf/printf-cov.pl: Since gnulib commit v1.0-1103-ge5b82978e2
we avoid iconv() on ASCII range 0x32 - 0x7F inclusive, so adjust
this test to fall within that range.
Addresses https://bugs.gnu.org/74428
* src/df.c (replace_control_chars):
* src/dircolors.c (parse_line):
* src/printf.c (print_esc):
* src/ptx.c (unescape_string):
* src/stat.c (print_it):
* src/tr.c (star_digits_closebracket):
Omit to_uchar calls that aren’t needed, because the parent
expression works with ‘char’ as well as with ‘unsigned char’.
Also, port better to macOS.
* src/printf.c (verify_numeric): Don’t assume that when s == end
then errno is zero; it is EINVAL on macOS, and POSIX allows this.
(print_direc): Treat missing arg as zero for numeric conversions,
and as an empty string for the others.
(print_formatted): Use null pointer, not an empty string,
to represent missing arg.
* tests/printf/printf.sh: Test empty and space widths and precisions.
* src/printf.c (struct arg_cursor): New struct.
(get_curr_arg): New function.
(print_formatted): Use it instead of ...
(GET_CURR_ARG, SET_CURR_ARG): ... these removed macros.
This makes the code a bit easier to follow, and any efficiency
cost should be minimal.
* tests/ls/selinux-segfault.sh: Move recent addition to ...
* tests/ls/selinux.sh: ... this new test that uses require_selinux_
to skip appropriately when we've built without selinux support.
Also add a non root test that checks we output '.' along with the
mode for files, to indicate a security context is present.
* tests/local.mk (Reference the new test).
This also fixes a problem with ls -Z when configured with
--disable-acl, reported by Pádraig Brady
<https://bugs.gnu.org/73418#52>.
* src/ls.c (gobble_file): Pass ACL_GET_SCONTEXT to
file_has_aclinfo, if -Z is used.
Problem reported by Pádraig Brady <https://bugs.gnu.org/73418#35>.
This bug was fixed by the recent gnulib update.
* tests/ls/selinux-segfault.sh:
Also test for ls -Z on broken symlinks.
* src/shuf.c (main): In dev mode call randint_all_free()
to avoid false failure with valgrind 3.16.1 at least.
Note this partially reinstates commit v9.0-109-g0106b5a4b.
This was noticed on a debian 11 system running CI tests.
* src/seq.c: Include full-write.h.
(seq_fast): Since we’re doing all the buffering work anyway,
we might as well use syscalls instead of stdio to write.
Use full_write instead of fwrite.
* src/seq.c (seq_fast): Simplify by using an output buffer of
known size (BUFSIZ) on the stack, rather than a heap buffer that
might grow. For the number buffer, don’t bother appending NUL
since nobody uses the NUL, and xpalloc from nullptr not p0 since
we need to move the buffer data by hand anyway.
* src/seq.c (incr): Change API to make the code easier to follow,
and also to avoid undefined behavior on hypothetical platforms
where '9' == INT_MAX (!). Caller changed.
* src/shuf.c (RESERVOIR_LINES_INCREMENT): Remove.
All uses removed.
(read_input_reservoir_sampling, main):
Prefer idx_t to size_t for sizes related to xpalloc.
(read_input_reservoir_sampling): Prefer xpalloc to xnrealloc.
* src/pwd.c (struct file_name, file_name_prepend):
Prefer idx_t to size_t for sizes related to xpalloc,
(file_name_init): Don’t overflow if PATH_MAX == INT_MAX.
(file_name_prepend): Prefer xpalloc to by-hand resizing.
Simplify by using memcpy return value.
* src/du.c (prev_level, process_file):
Prefer idx_t to size_t for sizes related to xpalloc,
and to nesting levels (since that’s what fts_level does anyway).
(process_file): Prefer xpalloc to xnrealloc.
* src/df.c (ncolumns, nrows, print_table, get_header, get_dev):
Prefer idx_t to size_t for sizes related to xpalloc.
(ncolumns_alloc, nrows_alloc): New static vars.
(alloc_table_row, alloc_field): Prefer xpalloc to xnrealloc.
* src/pr.c (buff_allocated, main):
Prefer idx_t to size_t for sizes.
(main, store_char): Use xpalloc, not x2realloc.
(init_store_cols): Check for multiplication overflow ourselves
and use ximalloc, not xnmalloc. This is a bit simpler.
* src/system.h (X2REALLOC): Remove; no longer used.
* src/set-fields.c (n_frp, n_frp_allocated, complement_rp, set_fields):
Prefer idx_t to ptrdiff_t/size_t for nonnegative sizes.
(add_range_pair): Use xpalloc, not x2nrealloc.
* src/ptx.c (line_width, gap_size, WORD, WORD_TABLE)
(maximum_word_length, reference_max_width, occurs_alloc)
(number_of_occurs, half_line_width, truncation_string_length)
(compare_words, search_table, digest_word_file)
(find_occurs_in_text, fix_output_parameters)
(generate_all_output, main, find_occurs_in_text)
(fix_output_parameters, generate_all_output):
Prefer idx_t to ptrdiff_t/size_t for nonnegative sizes.
(first, second): Remove macros, replacing them with locals.
(search_table): Use hi (for highest + 1) to simplify.
Avoid unlikely overflow by not computing lo + hi.
(digest_word_file, find_occurs_in_text): Use xpalloc, not x2nrealloc.
* src/od.c (n_specs, n_specs_allocated, write_block, get_lcm, main):
Use idx_t instead of size_t for some sizes, so that we can
use xpalloc.
(decode_format_string): Prefer xpalloc to X2NREALLOC.
* src/expand-common.c (get_next_tab_column): Check for tab
stop overflow here. All callers changed to not check.
* src/expand.c (expand): Use colno for column number.
Found when testing on a new platform with a new file system.
* src/ls.c (file_has_aclinfo_cache): For failures, also cache
return value, scontext, and scontext_err, and when using cached
values make sure buf and size have reasonable values for
aclinfo_free etc.
* bootstrap.conf (gnulib_modules): Add acl-permissions, which
supplies acl_errno_valid, and which we are already using
indirectly via file-has-acl.
* src/ls.c (errno_unsupported): Remove. All calls replaced
by !acl_errno_valid.
* src/expand-common.h (colno): New typedef.
All uses of uintmax_t for column numbers replaced by colno.
* src/expand-common.c (add_tab_stop): Use xpalloc
instead of X2NREALLOC, and use ckd_add to check for overflow.
* src/expand-common.c (max_column_width, n_tabs_allocated)
(first_free_tab, add_tab_stop, parse_tab_stops, validate_tab_stops)
(get_next_tab_column):
Use idx_t for sizes. All uses changed.
(add_tab_stop): Use xpalloc instead of X2NREALLOC.
Use ckd_add to check for overflow, instead of doing it by hand.
* gl/lib/heap.c (struct heap, heap_alloc, heap_insert)
(heapify_down, heapify_up): Prefer idx_t to size_t for sizes.
(heap_insert): Use xpalloc instead of x2nrealloc.
(heapify_down): Return void since no caller cares about value.
* gl/modules/heap: Depend on idx.
$ echo -n '123456789' | cksum --raw -a crc32b | basenc --base16
CBF43926
* bootstrap.conf: Explicitly depend on the crc module.
* doc/coreutils.texi (cksum): Add "crc32b" as an argument to -a.
* src/cksum.c (crc32b_sum_stream): A new function similar to
crc_sum_stream, but which does not include the length in
the CRC calculation.
* src/cksum.h: Add crc32b_sum_stream prototype.
* src/digest.c: Add "crc32b" as an argument to -a.
* tests/cksum/cksum.sh: Refactor to test both crc and crc32b.
* tests/cksum/cksum-a.sh: Add "crc32b" case.
* tests/cksum/cksum-base64.pl: Likewise.
* tests/misc/read-errors.sh: Likewise.
* NEWS: Mention the new feature.
* src/timeout.c (usage): Fix typo of period with comma.
* tests/timeout/timeout.sh: Only test a single option variant,
as tests/misc/usage_vs_getopt.sh suffices for basic option validation.
* src/timeout.c: Support -f and -p short options, corresponding to
--foreground and --preserve-status respectively. This adds
compatability with POSIX 2024 and OpenBSD.
(usage): Separate translations, and reorder the option descriptions.
* doc/coreutils.texi (timeout invocation): Adjust accordingly,
and also reorder the option descriptions alphabetically.
* tests/timeout/timeout.sh: Also test short option variants.
* cfg.mk (sc_gldist_missing): Add a new target to ensure we don't
forget to distribute any new gl/ files.
* gl/local.mk: Remove generation comment since it's
now encapsulated in the syntax-check, which outputs a consumable
diff to make any future adjustments.
Also adjust ordering to that of the C locale used in the syntax check.
'make distcheck' would fail since commit 75b34c77e4, because the
comparison by check-ls-dircolors fails.
* Makefile.am (check-ls-dircolors): Adjust sed(1) expression to the
changed data initialization.
* tests/ls/no-cap.sh: Move to being a root only test, since
commit v9.5-132-g2a6bed933 we now need to call setcap
to make the test effective. Otherwise we would have always
just skipped the test.
Update gnulib submodule to latest. This changes the file_has_aclinfo
API, so at the same time do the following changes to ls.c, which
adjusts to these changes among other things.
* src/ls.c (filetype_d_type, d_type_filetype): New static constants.
(format_needs_capability): New static var.
(main): Set and use it. Don’t set format_needs_stat merely
because print_scontext, as we needn’t call stat to get the
scontext. Instead, set format_needs_type if print_scontext but
not format_needs_stat.
(print_dir): Use new static tables to determine filetype
more efficiently.
(file_has_aclinfo_cache): Adjust to Gnulib file_has_aclinfo API change.
(gobble_file): Check stat if format_needs_type but the type is
unknown. Be conservative, and when deciding whether to check stat
but the type is unknown, assume it might be directory. Similarly
for normal files when classifying; if the type is unknown assume
it might be normal. Use new static constants and IFTODT to
compute filetype more straightforwardly. Get ACLs and check for
capability less often.
(get_color_indicator): Omit unnecessary call to is_colored (C_CAP),
since f->has_capability can be true only if is_colored (C_CAP).
* src/ls.c (gobble_file): Minor refactoring. Last arg is now null
pointer, not "", for no directory. All callers changed.
Avoid need for cast from char const * to char *.
* src/ls.c (BIN_STR): New macro, replacing LEN_STR_PAIR.
All uses changed. This avoids the need to store the
trailing \0 in each string. This change is more for clarity,
to make it clear the \0 is not needed.
The recent commit v9.5-119-g4ce432ad8 restricted capability checking
to only files with XATTR_NAME_CAPS set. If this is done then we need
to adjust tests/ls/no-cap.sh so that it doesn't always skip. More
problematically XATTR_NAME_CAPS was only determined in long listing
mode, thus breaking capability coloring in short listing mode
as evidenced by the failing tests/ls/capability.sh test.
Note capability checking does have a large overhead, but we've
disabled capability checking by default anyway through the default
color configuration since v9.0-187-g6b5134770
So for these reasons revert to checking capabilities as before.
* src/ls.c (gobble_file): Check for capabilities in all modes
if enabled in color config.
The description of -k regressed in coreutils 9.0
* doc/coreutils.texi (ls invocation): Fix incomplete paragraph
describing -k introduced by a mistake in commit v8.32-180-g1625916a1.
Where rpl_fopen() is used rather than fopen(),
wrapping fopen() is ineffective.
Note rpl_fopen() is used as of glibc-2.39 at least
(due to fflush and fclose being replaced).
* tests/df/no-mtab-status.sh: Wrap open() rather than fopen().
* tests/df/skip-duplicates.sh: Likewise.
Prompted by the following 'make syntax-check' failure:
cppi: src/factor.c: line 175: not properly indented
cppi: src/factor.c: line 176: not properly indented
maint.mk: incorrect preprocessor indentation
make: *** [cfg.mk:750: sc_preprocessor_indentation] Error 1
* src/factor.c: Filter through 'cppi -a'.
* src/ls.c: Do not include <selinux/selinux.h> or "smack.h".
Include <linux/attr.h> if HAVE_LINUX_ATTR_H, for XATTR_NAME_CAPS.
(free_ent): Use aclinfo_scontext_free to free f->scontext.
(getfilecon_cache): Remove; no longer needed.
(file_has_aclinfo_cache): Rename from file_has_acl_cache,
and use file_has_aclinfo instead of file_has_acl. All uses changed.
(gobble_file): Use file_has_aclinfo instead of file_has_acl, so
that we get more info about the file before deciding whether to
issue further syscalls for it. Let file_has_aclinfo worry about
smack and SELinux. Call has_capability only if the xattr list
mentions XATTR_NAME_CAPS.
* src/factor.c (lbuf_putint_append): New function, with
most of the old lbuf_putint body. Do the umaxtostr stuff
by hand so that we needn’t worry about the trailing NUL.a
Do the string copy by hand since the string is so short.
(lbuf_putint): Reimplement in terms of lbuf_putint_append.
Omit last arg, which is no longer needed. All callers changed.
(print_uuint): Rewrite to avoid recursion, using
lbuf_putint_append for the usual case.
* src/factor.c (W_TYPE_SIZE): Simplify by always defining
to UINTMAX_WIDTH.
(W): Remove. All uses replaced by W_TYPE_SIZE.
We no longer need one of its static_asserts.
Previously, the code used stdio buffers for gmp numbers,
and did its own buffering for smaller numbers. This meant
for more flushing than was needed. The code now uses its
own buffering for all standard output, which makes for
less flushing and fewer writes.
* src/factor.c (lbuf_half_flush): New function, taken from the
body of lbuf_putnl.
(lbuf_putnl): Use it.
(lbuf_putmpz): New function, to output an mpz without using stdio.
(print_factors): Output via functions instead of via stdio.
* src/factor.c (struct lbuf_, lbuf, lbuf_alloc): Remove.
All uses removed.
(FACTOR_PIPE_BUF): Now a constant instead of a macro.
Increase to PIPE_BUF if available.
(lbuf_buf, lbuffered): New static vars, replacing lbuf.
All uses changed.
(lbuf_flush): Avoid unlikely recursion on write failure.
(lbuf_putc): Now simply adds a byte to the buffer.
(lbuf_putnl): Do the work of the old lbuf_putc ('\n').
Use changed. Use memrchr to find the newline.
(lbuf_putint): Widths are now int, not size_t.
* src/factor.c (uuset): New function.
(mod2): Return uuint rather than having half the returned value
stored via a pointer. This makes the code a bit easier to read
and can help the compiler avoid aliasing issues. All callers changed.
This refactors to add a new type, a uintmax_t pair, which
can simplify some code without slowing it down.
* src/factor.c (uuint): New type.
(lo, hi, hiset, make_uuint): New functions.
(struct factors.plarge): Use the new type. All uses changed.
* bootstrap.conf (gnulib_modules): Add stdc_trailing_zeros.
* cfg.mk (_gl_TS_unmarked_extern_vars): Remove factor_clz_tab,
as it’s no longer present.
* src/factor.c: Include stdbit.h.
(__clz_tab, factor_clz_tab): Remove.
(ASSERT, UHWtype, __GMP_DECLSPEC): Use simpler way to pacify
-Wunused-macros.
(count_leading_zeros, count_trailing_zeros):
Remove. All uses replaced by stdc_leading_zeros, stdc_trailing_zeros.
(factor_using_division, prime2_p): Add a couple of ‘assume’s
so that GCC knows the stdc_* calls are nonzero and can
optimize accordingly.
* src/factor.c (mod2): Work even if cntd <= cnta. The old version
of the code assumed that shifts by N had unspecified behavior
unless 0 <= N < wordsize. Although this assumption is portable to
all known practical platforms, the C standard says these shifts
have undefined behavior and some pedantic platforms check this.
* tests/factor/create-test.sh:
* tests/local.mk (factor_tests): New test t37.
* src/sort.c (usage): Don't mention the ambiguous "manual",
rather "full documentation", echoing the language at the
bottom of each coreutils man page.
Fixes https://bugs.gnu.org/72914
* man/env.x: Avoid confusion in the [OPTIONS] section
by renaming to [SCRIPT OPTION HANDLING], and removing info
regarding default signal handling, which is best
restricted to the full info manual.
Addresses https://bugs.gnu.org/72914
* src/echo.c (usage): Use printf(1) rather than 'printf',
which is marked up more appropriately, and can be
referenced by some man page readers.
Fixes https://bugs.gnu.org/72914
* src/printf.c (print_formatted): Add support for %i$ indexed args.
* tests/printf/printf-indexed.sh: Add a new file of test cases.
* tests/local.mk: Reference the new test file.
* doc/coreutils.texi (printf invocation): Mention how mixed
processing of indexed and sequential references are supported,
unlike the printf(2) library function.
* NEWS: Mention the new (POSIX:2024) feature.
Addresses https://bugs.gnu.org/73068
This issue was noticed with -flto on GCC 14.2.1
* gl/lib/xdectoint.c (__xnumtoint): Only inspect the
returned value if LONGINT_INVALID is not set,
as the returned value is uninitialized in that case.
Fixes https://bugs.gnu.org/72842
* NEWS: Mention the change in behavior.
* src/install.c (need_copy): s/lstat/stat/ for the source.
* tests/install/install-C.sh: Add test cases
(and improve existing test case which wan't valid
due to the existing non standard modes on test files).
Addresses https://bugs.gnu.org/72707
* boostrap.conf (gnulib.modules): Add xvasprintf, which
had been omitted by mistake.
* src/copy.c, src/dd.c, src/test.c: Don't include verror.h,
as Gnulib removed it.
Support overriding previous sorting options
with an explicit --sort=name option.
* doc/coreutils.texi (ls invocation): Document the new option.
* src/ls.c (usage): Likewise.
(sort_args): Add the "name" entry, and sort to be consistent
with the ordering presented in --help.
* tests/ls/ls-time.sh: Add test cases.
* NEWS: Mention the new feature.
Suggested by: Tzvetelin Katchov
* src/printf.c: Remove redundant comment.
State explicitly that the leading 0 is the exception
from normal escape processing. Remove a full stop for consistency.
* doc/coreutils.texi (printf invocation): Add a reference
to C99 string escapes since these are not mentioned
in the referenced glibc printf info. Also explicitly state
the leading 0 exception. Also use NNN rather than OOO
to be consistent with the --help documentation.
Also remove and extraneous '\' and fix grammar in the info
regarding the ninth bit.
Addresses https://bugs.gnu.org/72657
Although these patches don’t affect user-visible behavior,
they do clean up the source code a bit, and the
machine code should be a tiny bit more efficient.
* src/cat.c (simple_cat, cat):
* src/csplit.c (read_input):
* src/head.c (copy_fd, elide_tail_bytes_pipe)
(elide_tail_lines_pipe, elide_tail_lines_seekable, head_bytes)
(head_lines):
* src/install.c (have_same_content):
* src/tac-pipe.c (buf_init_from_stdin):
* src/tac.c (tac_seekable, copy_to_temp):
* src/tail.c (dump_remainder, file_lines, pipe_lines)
(pipe_bytes, start_bytes, start_lines, tail_forever_inotify):
* src/tr.c (plain_read):
Adjust to recent Gnulib changes by using new types
for safe_read, safe_write, full_read, full_write.
Not clear that the overflows could be exploited,
but they made the code confusing.
* src/head.c (elide_tail_bytes_pipe): Don’t convert uintmax_t
to size_t first thing; wait until it’s known the value will fit,
and then use idx_t rather than size_t to prefer signed types.
Prefer idx_t in nearby code, too.
Rename locals n_elide_0 to n_elide (for consistency elsewhere)
and n_elide to in_elide.
Remove bogus (SIZE_MAX < n_elide + READ_BUFSIZE) test;
in the typical case where n_elide’s type was the same as
that of SIZE_MAX, the test never succeeded, and in the
less-common case where n_elide was wider than size_t,
the addition could silently overflow, causing the test
to fail when it should succeed. The test is not needed anyway now.
Add static asserts to document code assumptions.
Redo the ! (n_elide <= HEAD_TAIL_PIPE_BYTECOUNT_THRESHOLD) case
so that it works with enormous values of n_elide even on
32-bit platforms; for example, n_bufs is now uintmax_t not size_t.
Simplify by using xpalloc instead of by-hand code.
Remove bogus ‘if (rem)’ test, as rem is always nonzero.
* src/tail.c (tail_lines): If skipping all input, use lseek if
possible.
(parse_options): Allow counts to exceed 2**64.
(main): Don’t subtract 1 from UINTMAX_MAX, since it stands
for infinity in this context.
(main): Also don’t read anything when given infinite elisions.
* tests/tail/tail.pl: Adjust to match new behavior. Rename err-5
test to big-c and expect the invocation to succeed, since ‘tail
-c99999999999999999999’ now succeeds instead of (unnecessarily)
failing.
* src/head.c (head): Optimize for -n-HUGE, where HUGE exceeds
2**64 - 2.
(string_to_integer): Return UINTMAX_MAX for too-large numbers,
instead of failing.
(main): Omit no-lnger-necessary test for byte count overflow.
Also, prepare for allowing some arguments to overflow
without that being an error.
* gl/lib/xdectoint.c: Do not include stddef.h,
since we no longer use ‘unreachable’.
(xnumtoimax, xnumtoumax, __xnumtoint):
New arg FLAGS. All callers changed.
Stop using __xdectoint_signed. All definers removed.
* gl/lib/xdectoint.h (XTOINT_MIN_QUIET, XTOINT_MAX_QUIET)
(XTOINT_MIN_RANGE, XTOINT_MAX_RANGE): New flag constants.
* src/fmt.c (main):
* src/fold.c (main):
* src/nl.c (main):
* src/pr.c (getoptnum):
* src/split.c (main):
Use XTOINT_MIN_RANGE and XTOINT_MAX_RANGE if appropriate.
* src/pr.c (getoptnum): Return int rather than returning void
and storing through int *.
* src/stty.c (apply_settings):
Use ckd_add to check for overflow instead of doing it by hand.
(integer_arg): Accept and return uintmax_t, not unsigned long.
The test writes to the disk and means the space used changes. If this
crosses a number boundary, the heading spacing can change:
-Filesystem 1024-blocks Used Available Capacity Mounted on
+Filesystem 1024-blocks Used Available Capacity Mounted on
* tests/df/df-P.sh: Squash spaces with tr to avoid alignment variations.
* gl/lib/randperm.c: Include <stdckdint.h>.
(randperm_bound): Return SIZE_MAX if the multiplication overflows.
Do not overflow when converting bit count to byte count.
Problem reported by Daniel Carpenter <https://bugs.gnu.org/72445>.
* gl/lib/randread.c (randread_new): Fill the ISAAC buffer
instead of storing at most BYTES_BOUND bytes into it.
We already support reproducible builds since commit v8.24-99-gc1b3d6587,
and this adjusts that change to also support reproducible
tarball contents with subsequent runs of `make dist`.
* Makefile.am: Don't create a varying .timestamp file, instead ...
* man/local.mk: Rely on the timestamp of the .tarball-version file.
Fixes https://bugs.gnu.org/72232
On these file systems the atime is always zero.
Problem found with ZFS on Ubuntu 24.04 LTS.
* tests/stat/stat-birthtime.sh (check_timestamps_updated):
* tests/stat/stat-nanoseconds.sh:
Work even if atimes are always zero.
* tests/stat/stat-nanoseconds.sh:
Fix typo: print_ver_ called before init.sh sourced.
* src/ls.c: Track if --time=mtime is explicitly specified,
so that we can apply the GNU extension of sorting by the
specified time, when not displaying (-l not specified),
and not explicitly sorting (-t not specified).
* tests/ls/ls-time.sh: Add / Update test cases.
Fixes https://bugs.gnu.org/71803
* doc/coreutils.texi (ls invocation): Document the "c-maybe"
--quoting-style, which was added as an option in 2008.
Reported at https://bugs.debian.org/1074334
The ERE used lacks the grouping of the extensions and therefore would
also match files where the first two patterns are not at the end of
the line:
grep -E '\.sh|\.pl|\.xpl$'
* cfg.mk (sc_tests_list_consistency): Add grouping (...) around the
sub-patterns. While at it, also remove the redundant escaping, i.e.,
\$$ -> $$ to be consistent with the rest of this file.
* src/stat.c (human_fstype): Rename "FUSEBLK" to "FUSE" to sync with
kernel adjustments. Add "bcachefs", and "pidfs". Both are local,
with the latter being similar to "proc" which is also local.
* NEWS: Mention the change in behavior, and the improvement.
Problem reported by Dan Jaobson (Bug#71171).
* doc/coreutils.texi: Clarify that directory entries are sorted,
not command-line arguments.
* src/ls.c (usage): Be less chatty about -U and
about --group-directories-first.
* src/local.mk: Avoid overriding automake generated DEPENDENCIES,
so that it applies its adjustments to LDADD to avoid propagating
flags (like -Wl,-rpath) into make targets. This was seen on FreeBSD
where LIBINTL is set to:
/usr/local/lib/libintl.so -Wl,-rpath -Wl,/usr/local/lib
Instead let automake generate a sanitized src_coreutils_DEPENDENCIES
(based on LDADD), which we then augment with the EXTRA_... variable.
<stdbit.h> is in C23 and should be more portable in the long run,
now that Gnulib supports it.
* bootstrap.conf (gnulib_modules): Remove count-leading-zeros.
Add stdc_leading_zeros.
* gl/lib/randperm.c, src/ioblksize.h:
Include stdbit.h instead of count-leading-zeros.h.
* gl/lib/randperm.c (floor_lg): Remove; no longer needed.
(randperm_bound): Use stdc_bit_width instead of floor_lg;
* gl/modules/randperm (Depends-on): Remove count-leading-zeros.
Add stdc_bit_width.
* src/ioblksize.h (io_blksize): Use stdc_leading_zeros_ull
instead of count_leading_zeros_ll.
* configure.ac: Disable GCC 14’s -Wmissing-variable-declarations
in the test directory, as it’s not worth the aggravation there.
Likewise for GCC's -Wsuggest-attribute=cold.
* src/sort.c: Ignore -Wmissing-variable-declarations only
with GCC 14 and newer, since it didn’t exist earlier.
Ignore the warning only when including md5.h, where it
needs to be ignored, as the warning might be useful elsewhere.
* src/cksum.c (main) [CRCTAB]: Generate updated crctab.c (see below).
* src/crctab.c: Include cksum.h, to check consistency
between decl and defn. Include stdio.h since cksum.h needs it.
This patch is part of work done for a project from Google Summer of
Code, see the project details at
<https://summerofcode.withgoogle.com/programs/2024/projects/E9Jp7RUx>.
* src/sleep.c (usage): Directly mention the floating-point option,
which is typical for sleeping milliseconds.
Also reorganize the text to be 3 lines rather than 4.
Similarly to commit v9.4-143-gfcfba90d0,
and enabled for AVX by commit v9.5-25-g0e4450103.
This was seen to improve AVX performance by about 10%
on an AMD 7800X3D (Ryzen 7 (2023)) CPU,
while having neutral AVX performance,
on an Intel i7-5600U (Broadwell-U (2015)) CPU.
With avx not enabled, this gives about a 3% performance boost,
on an Intel i7-5600U.
* src/wc.c: Use the centrally configured optimum buffer size.
* src/wc_avx2.c: Likewise.
* NEWS: Mention the change in performance.
* src/wc_avx2.c (wc_lines_avx2): Change from
_mm256_sub_epi8() + _mm256_sad_epu8() to
_mm256_movemask_epi8() + __builtin_popcount().
This will allow adjusting the I/O size above 16KiB.
* configure.ac: Align check with routines used in wc_avx2.c.
* src/show-date.{h,c}: Declaration and definition of show_date.
* src/du.c: Wse the common show_date instead of the previous local
function.
* src/date.c: Wse the common show_date via a wrapper show_date_helper.
* src/local.mk: Corresponding adjustments.
* src/cp.c: Add the entries for the --update=none-fail option.
* tests/mv/update.sh: Add a test case.
* NEWS: Mention the bug fix.
Fixes https://bugs.gnu.org/70727
* gnulib: Update to support bootstrapping with python by default.
* bootstrap: Sync with gnulib.
* cfg.mk: Don't force python implementation with `make world`,
rather rely on the auto selection of python if appropriate.
* doc/coreutils.texi (sort invocation): Give a DSU example
for sorting names which may have a variable number of fields.
Addresses https://bugs.gnu.org/70532
* cfg.mk: Add a new "world" default target so that one
can bootstrap (using the python implementation), configure,
and make, by using `make -f cfg.mk`.
* gnulib: Update to latest primarily to test the
bootstrap python implementation which is now in beta test.
* README-hacking: Document the `make -f cfg.mk` shortcut.
Problem reported by Ionut Nicula in:
https://bugs.gnu.org/70477
* src/tail.c (tail_bytes): Do not loop forever on commands
like 'tail -c 4096 /dev/zero'.
* tests/tail/tail-c.sh: Test this fix.
* src/join.c (main): s/field/file/ in the error message
for -a and -v, introduced in TEXTUTILS-1_13-24-g6f63d53e1.
Reported at https://bugs.debian.org/1068864
* src/cat.c (main):
Improve test for when copying will exhaust the output device.
Do not rely on st_size, which is unreliable in /proc.
Use lseek instead; this is good enough here.
* tests/cat/cat-self.sh: Test the relaxation of the heuristic
for self-copying.
od was seen to abort() on glibc on ia64 and m68k with the error:
Fatal glibc error: printf_fp.c:501 (__printf_fp_buffer_1):
assertion failed:
cy == 1 || (p.frac[p.fracsize - 2] == 0 && p.frac[0] == 0)
* tests/od/od-multiple-t.sh: Avoid outputting long double floats
to avoid undefined behavior. 'float' and 'double' are standardized
by IEEE 754 (except on Linux/m68k) and don't have undefined values.
* tests/misc/numfmt.pl: Verify that printf field width specs
count characters and not bytes before enabling locale tests.
This was seen on FreeBSD 14.0 and Solaris 11 OpenIndiana.
Reported by Bruno Haible
* src/ls.c (print_dir): readdir() on FreeBSD 14 was
seen to pass ENOENT through. ENOENT in this context
means "Directory unlinked but still open".
Reported by Bruno Haible with tests/ls/removed-directory.sh
There is a mismatch between isblank() used by tr and c32isblank() now
used by uniq on Solaris 11 OpenIndiana. isblank() was seen to return
true for non breaking space, while c32isblank() returned false.
Interestingly on Solaris, non breaking space is considered a blank
character, and isblank() and c32isblank() honor this in all locales.
* tests/uniq/uniq.pl: Adjust the blank check to use join(1) rather than
tr(1), as join uses the same blank determination routines as uniq(1).
This issue was introduced in commit v8.19-145-g24ebca6
* src/copy.c (copy_internal): On systems that don't support ACLs,
the fallback default chmod done on directories should maintain
the set-group-ID, as that's generally auto-set by the system.
* NEWS: Mention the fix.
Reported by Bruno Haible on Alpine (with tests/cp/preserve-mode.sh)
* tests/misc/numfmt.pl: Some systems with the fr_FR.UTF-8
locale installed, do not have a thousands grouping character defined.
In this case we skip the locale tests which depend on a non empty
grouping character.
* tests/chmod/symlinks.sh: The count of adjusted modes was
one more on systems where symlink modes can be adjusted.
Therefore only include the non symlinks in the count.
* src/chown-core.h (emit_from_option_description): The conditional
string composition here caused issues for translators.
Instead move to a more general description ...
(src/chown.c (usage): ... here.
Fixes https://bugs.gnu.org/69985
* tests/chmod/symlinks.sh: Ensure this new test is immune
to setgid directories by resetting modes with =777 rather than 777.
Also output more debugging in all failure cases.
* tests/mv/mv-exchange.sh: Canonicalize different
"operation not supported" messages, so we can ignore correctly.
Reported by Bruno Haible on AIX, NetBSD, and OpenBSD.
Also, improve quality of diagnostics.
Problems/suggestions by Bernhard Voelker in
<https://bugs.gnu.org/69532#82>.
* src/copy.c (emit_verbose): New arg FORMAT. All uses changed,
to improve quality of diagnostics when --exchange is used.
(copy_internal): Don’t try to optimize --exchange so much; this
simplifies the code and keeps it closer to the non --exchange case.
* src/copy.c (copy_internal): Simplify logic for copying
from directory to non-directory or vice versa, and always
diagnose with both source and destination file names.
Using the shell's exec -a feature can be awkward
so add support for setting overriding argv[0].
This gives env full control over the arguments it passes.
* src/env.c: Accept -a,--argv0 and set argv[0] appropriately.
* tests/env/env.sh: Add test cases.
* doc/coreutils.texi (env invocation): Describe -a,--argv0.
* NEWS: Mention the new feature.
* doc/coreutils.texi (pr invocation): Explicitly state that
multicolumn output will convert spaces to TABs, and show that
this can be undone with the `pr -t -e` or `expand` commands.
Suggested by Douglas McIlroy in https://bugs.gnu.org/69807
* src/copy.h (struct cp_options): New member 'exchange'.
* src/copy.c (copy_internal): Support the new member.
* src/mv.c (EXCHANGE_OPTION): New constant.
(long_options): Add --exchange.
(usage): Document --exchange.
(main): Support --exchange.
* tests/mv/mv-exchange.sh: New test case.
* tests/local.mk (all_tests): Add it.
This is an issue with -[H]R mode, where an attacker
may replace a traversed file with a symlink
between where we stat() the file and chmod() the file.
* src/chmod.c (process_file): Remove the first !S_ISLNK guard
as that's now just an optimization, and instead consistently
apply fchmodat() to files/symlinks. Ensure AT_SYMLINK_NOFOLLOW
is set when traversing in default (-H) mode.
* NEWS: Mention the bug fix.
Fixes https://bugs.gnu.org/11108
There have been various requests to add -h to avoid following symlinks
for security reasons. This wasn't provided previously as chmod(1)
already ignored symlinks unless specified on the command line.
Note chmod defaults to -H mode rather than the chown default of -P,
as usually chown can work directly on symlinks and so defaults
to not traversing those specified on the command line.
Note FreeBSD chmod does default to -P mode, but we retain the -H mode
default also for compatibility with existing chmod behavior.
Adding -HLP will allow chmod to disable traversing CLI symlinks to dirs.
Adding -h will allow to disable following CLI symlinks to files/dirs,
also operating on all symlinks on systems that support that.
Adding --dereference will be significant with -H (the default). I.e.
symlinks to dirs not recursed, but symlinks are dereferenced.
Adding these options will also be consistent with chown(1), chgrp(1),
and chmod(1) on other systems.
Note since chmod(1) currently ignores symlinks by default,
and -h is primarily a mechanism to avoid following symlinks, rather than
for operating on the symlink itself, we make -h try to chmod a symlink,
but ignore ENOTSUP. In that way we're consistent with chown(1)
where it also ignores ENOTSUP for symlinks, and we don't fail when
trying to be extra secure with command line params.
* doc/coreutils.texi (chmod invocation): Reference the -H,-L,-P
descriptions, and adjust the corresponding macros to say
the default is -H or -P as appropriate.
Add --dereference and -h,--no-dereference descriptions.
* man/chmod.x: Adjust discussion of symlink handling.
* src/chmod.c (main): Accept new options and set
fts flags appropriately.
(process_file): Process / dereference symlinks as necessary.
* src/system.h (emit_symlink_recurse_options): A new function
refactored from chown.c and chmod.c usage().
* tests/chmod/symlinks.sh: New test for the new options.
* tests/local.mk: Reference the new test.
* NEWS: Mention the new feature.
* src/basenc.c (base16_encode, base2msbf_encode, base2lsbf_encode):
Ensure we don't overflow the output buffer, whose length is
passed in the OUTLEN parameter. This issue was flagged by clang
with -Wunused-but-set-parameter.
Behave like who(1) in requiring --lookup to enable this
often slow feature. pinky(1) is supposed to be lightweight after all.
* doc/coreutils.texi (who invocation): Adjust the description to no
longer reference dialup, and be more general about the still significant
delays.
(pinky invocation): Reference the same --lookup description.
* src/pinky.c (main): Accept --lookup to enable DNS lookups.
* NEWS: Mention the change in behavior.
Fixes https://bugs.debian.org/628815
Following v5.2.1-679-g7e29ef8b8 symlinks specified on the command line
no longer induce an error if lchown() is not supported on the system.
* doc/coreutils.texi (chown invocation, chgrp invocation): Adjust
accordingly, and also use a macro to avoid duplication.
* src/chown-core.c: Use our more standard is_ENOTSUP() wrapper
in the code related to this.
Some signals with values less that the max signal number for the system
do not have defined names. For example, currently on amd64 Linux,
signals 32 and 33 do not have defined names, and Android has a wider
gap of undefined names where it reserves some realtime signals.
Previously the signal listing in env ended up reusing the name
of the last printed valid signal (the repeated HUP below):
$ env --list-signal-handling true
HUP ( 1): IGNORE
HUP (32): BLOCK
HUP (38): IGNORE
..and the corresponding signal numbers were rejected as operands for the
env, kill, and timeout commands.
This patch removes the requirement that sig2str returns 0 for a signal
number associated with an operand. This allows unnamed signals to be in
the sets `env' attempts to manipulate when a --*-signal option is used
with no argument, and kill(1) and timeout(1) to send such unnamed
signals.
* src/operand2sig.c (operand2sig): Drop signame argument, accept all
signal numbers <= SIGNUM_BOUND. All callers updated.
* src/env.c (parse_signal_action_params, reset_signal_handlers)
(parse_block_signal_params, set_signal_proc_mask)
(list_signal_handling): Accept all signal numbers <= SIGNUM_BOUND,
use SIG%d for printing if necessary.
* src/kill.c (list_signals, main): Likewise.
(send_signals): Check errno from kill(3) for bad signo.
* src/timeout.c (main): Update operand2sig call.
* tests/misc/kill.sh: Test listing all signal numbers.
* NEWS: Mention the improvement.
* configure.ac: Wrap the following with AC_CACHE_VAL,
so that they can be cached / overridden. We use
the "utils_cv_" prefix as they're coreutils specific overrides.
utils_cv_avx2_intrinsic_exists,
utils_cv_brain_16_bit_supported,
utils_cv_ieee_16_bit_supported,
utils_cv_pclmul_intrinsic_exists,
utils_cv_stdbuf_supported.
Recent clang provides __bf16 on aarch64 but it is broken.
If built with -O0, the conversion is wrong:
$ printf '\x3F\x80' | od --end=big -An -tfB | tr -d ' '
1.875
If built with -O1 or higher, compilation fails:
fatal error: error in backend:
Cannot select: 0xb400007a58d29780: f32 = fp_extend 0xb40000...
0xb40000...: bf16,ch = CopyFromReg 0xb40000..., Register:bf16 %13
0xb40000...: bf16 = Register %13
In function: print_bfloat
The latter issue does not cause the existing configure test to fail
because the promotion is optimized out.
* configure.ac: Ensure 16 bit float promotion code does not get
optimized out, and produces an expected result.
* src/timeout.c (main): Block cleanup signals earlier so that cleanup()
is not runnable until monitored_pid is in a deterministic state.
This ensures we always send a termination signal to the child
once it's forked.
* NEWS: Mention the bug fix.
Reported at https://github.com/coreutils/coreutils/issues/82
* src/timeout.c (cleanup): Handle the case where monitored_pid
might be -1, which could happen if a signal was received
immediately after a failed fork() call. In that case it would
send the termination signal to all processes that the timeout
process has permission to send signals too.
* NEWS: Mention the bug fix.
* src/digest.c (main): If --binary was enabled with a previous --tag,
then reset the binary mode to auto select if --untagged then specified.
* tests/cksum/cksum-a.sh: Add a test case.
Since this functionality is recently available
in the exch(1) utility from util-linux,
it was thought best not to complicate mv with it.
This reverts commit 6cd2d5e533
* src/digest.c (main): Only validate the last used --length
for being a multiple of 8.
* tests/cksum/b2sum.sh: Add a test case.
Fixes https://bugs.gnu.org/69546
renameat2() syscall allows atomically swapping 2 paths on one
file system. Expose this ability to the user with --swap.
* doc/coreutils.texi: Describe mv --swap option.
* src/mv.c (main): Support --swap.
* tests/mv/mv-swap.sh: Add test for mv -x.
* tests/local.mk: Reference new test.
* NEWS: Mention the new option.
* src/ioblksize.h: Add updated test results and
increase value from 128KiB to 256KiB, which was last
updated 10 years ago.
* NEWS: Mention the improvement.
* src/mktemp.c (main): When --suffix is specified, TEMPLATE
points to the meraged buffer DEST_NAME. As X's in the suffix are
not significant to the generated random characters, the diagnostic
for too few X's should only refer to the template portion.
* tests/misc/mktemp.pl: Adjust accordingly.
* NEWS: Mention the bug fix.
* tests/df/problematic-chars.sh: Rely on gnulib setting
this to "none" where not usable.
* tests/misc/sleep.sh: Likewise.
* tests/printf/printf-mb.sh: Likewise.
* tests/printf/printf-quote.sh: Likewise.
* tests/sort/sort-debug-keys.sh: Likewise.
* configure.ac: Test where to find the dlopen function. Set LIB_DL.
Use it in the DLOPEN_LIBCRYPTO test.
* src/local.mk (src_sort_LDADD): Add $(LIB_DL).
* src/cp.c (main): Add support for --update=none-fail to provide the
functionality of diagnosing files in the destination,
and exiting with failure status.
(usage): Mark -n as deprecated.
* src/mv.c: Likewise.
* src/copy.h: Add UPDATE_NONE_FAIL definition.
* src/system.h (emit_update_parameters_note): Add --update=none-fail
description.
* doc/coreutils.texi (cp invocation): Likewise.
Also mention why -n is deprecated.
* tests/mv/update.sh: Add a test case, including precedence
with -n and other --update options.
* tests/cp/cp-i.sh: Verify that --backup and --update=none{,-fail}
are mutually exclusive.
* tests/mv/mv-n.sh: Likewise.
* NEWS: Mention the new feature.
Addresses https://bugs.gnu.org/62572
One should link the versioned lib at runtime,
and the unversioned lib at build time,
as the unversioned lib may not be installed,
and better couples the binary with the required version.
* configure.ac: Define LIBCRYPTO_SONAME, determined from
the test binary linked with -lcrypto. Also document
why we use SHA512() in the check, rather than MD5().
* src/sort.c (link_libcrypto): Use the versioned lib in dlopen().
* cfg.mk: Exclude the ptr_MD5_* symbols added in
commit v9.4-130-g7f57ac2d2, as there is no way
to declare these static given they way they're defined.
This saves time in the usual case, which does not need -lcrypto.
* configure.ac (DLOPEN_LIBCRYPTO): New macro.
* src/sort.c [DLOPEN_LIBCRYPTO && HAVE_OPENSSL_MD5]: New macros
MD5_Init, MD5_Update, MD5_Final. Include "md5.h" after defining
them. Include <dlfcn.h>, and define new functions link_failure
and symbol_address.
(link_libcrypto): New function.
(random_md5_state_init): Call it before using crypto functions.
When recursively copying files into OS trees, it often happens that
some subdirectory of the source directory is a symlink in the target
directory. Currently, cp will fail in that scenario with the error:
"cannot overwrite non-directory %s with directory %s"
However, we'd like cp in this scenario to follow the destination
directory symlink and copy the files into the symlinked directory
instead. Let's support this by adding a new option
--keep-directory-symlink that makes cp follow destination directory
symlinks.
We name the option --keep-directory-symlink to keep consistent with
tar which has the same option with the same effect.
* doc/coreutils.texi (cp invocation): Describe the new option.
* src/copy.h: Add the new setting.
* src/copy.h: Adjust to follow symlinks if setting enabled.
* src/cp.c (usage): Describe the new option.
(main): Accept the new option.
* tests/cp/keep-directory-symlink.sh: A new test.
* tests/local.mk: Reference the new test.
* NEWS: Mention the new feature.
* src/ls.c (decode_switches): Remove pragmas. They are no longer
needed to pacify GCC 13.2.1 with --enable-gcc-checking, and there’s
little point keeping them around for older GCC versions.
* gl/lib/fadvise.c: Remove pragma that works around GCC bug 83559
<https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83559>.
This bug was fixed in GCC 9, and we needn’t worry about
--enable-gcc-warnings for compilers that old.
* src/test.c: Likewise.
* doc/coreutils.texi (shred invocation): Fix the example
to correctly close file descriptor 3.
* THANKS.in: Remove old email since now recorded in repo history.
Reported at https://bugs.debian.org/1063837
* src/env.c (parse_signal_action_params, parse_signal_block_params):
Rename OPTARG to ARG so that it does not conflict with OPTARG used by
getopt.
Copyright-paperwork-exempt: Yes
* configure.ac: Ensure the compiler can promote 16 bit floating point
types to float, before enabling that code in od. This was an issue
with clang 16 at least.
* src/od.c: Adjust for the new defines.
* tests/od/od-float.sh: Likewise. Also port to the dash shell,
whose inbuilt printf doesn't support hex escapes.
This was introduced in coreutils 9.2 through commit v9.1-184-g40bf1591b,
and was fixed in coreutils 9.5 through commit v9.4-111-gc4c5ed8f4.
This issue has been assigned CVE-2024-0684.
* NEWS: Mention the bug fix.
* tests/split/line-bytes.sh: Add a test case.
Reported by Valentin Metz.
ulimit -v is generally not supported with ASAN, giving errors like:
"ReserveShadowMemoryRange failed while trying to map 0x... bytes.
Perhaps you're using ulimit -v"
* tests/cp/link-heap.sh: Mention ASAN as a possible reason for skipping.
* tests/csplit/csplit-heap.sh: Likewise.
* tests/cut/cut-huge-range.sh: Likewise.
* tests/dd/no-allocate.sh: Likewise.
* tests/printf/printf-surprise.sh: Likewise.
* tests/rm/many-dir-entries-vs-OOM.sh: Likewise.
* tests/head/head-c.sh: Only skip the part of the test needing ulimit.
* tests/split/line-bytes.sh: Likewise.
* src/split.c (line_bytes_split): Do not shrink hold buffer.
If it’s large for this batch it’s likely to be large for the next
batch, and for ‘split’ it’s not worth the complexity/CPU hassle to
shrink it. Do not assume hold_size can be bufsize.
* src/date.c (res_width): This function computes its result solely
from the value of its parameter and qualifies for the const attribute.
* src/tee.c (get_next_out): This function has no side effect and
qualifies for the pure attribute.
* THANKS.in: Remove duplicate now that author has a commit in the repo.
Those two functions were flagged by GCC 12.3.0,
though not by GCC 13.2.1.
Update to latest gnulib with new copyright year.
Run "make update-copyright" and then...
* gnulib: Update included in this commit as copyright years
are the only change from the previous gnulib commit.
* tests/init.sh: Sync with gnulib to pick up copyright year.
* bootstrap: Manually update copyright year,
until we fully sync with gnulib at a later stage.
* tests/sample-test: Adjust to use the single most recent year.
* configure.ac: Clang now seems to have -Wformat-extra-args,
-Wimplicit-const-int-float-conversion, and
-Wtautological-constant-out-of-range-compare on by default,
so disable them even if --enable-gcc-warnings is not used.
Rely on Gnulib’s check for clang rather than rolling our own.
* gl/lib/strnumcmp-in.h (numcompare): After commit v9.0-8-g6cafb122f,
we need to treat characters as signed to avoid invalid comparisons
between negative integers and unsigned characters.
* NEWS: Mention the bug fix.
* tests/misc/numfmt.pl: Following on from v9.4-86-g615167cc4,
adjust this test accordingly. This test was being skipped
on some systems, and so only noticed now.
Reported by Jim Meyering.
* tests/misc/numfmt.pl: Determine the thousands grouping character
in use, rather than skipping locale tests when it's not a space.
For example fr_FR.UTF-8 uses "NARROW NO-BREAK SPACE" as the grouping
char on modern glibc systems at least.
* tests/sort/sort-h-thousands-sep.sh: Likewise.
chown is a close superset of chgrp functionality,
so merge sources to avoid unwanted divergence in future.
This removes about 300 lines in chgrp.c
* build-aux/gen-single-binary.sh: Generate new rules for chgrp.
* cfg.mk: Exclude new wrappers.
* po/POTFILES.in: Remove chgrp.c
* src/chgrp.c: Remove.
* src/chown-chgrp.c: New wrapper.
* src/chown-chown.c: Likewise.
* src/chown.c (main): Prepend ':' for chgrp(1).
* src/chown.h: Define both operating modes.
(usage): Adjust depending on utility being called.
* src/coreutils-chgrp.c: Likewise.
* src/local.mk: Reference new wrappers.
Do not perform SELinux context translation for operations not involving
user input or output. Context translation converts MCS/MLS labels into
human readable form, which is useful for user facing applications like
ls(1) or the --context=CTX argument of cp(1).
* src/copy.c (set_process_security_ctx): Use raw selinux variants.
* src/install.c (need_copy): Likewise.
(setdefaultfilecon): Likewise.
* src/selinux.c (computecon): Likewise.
(defaultcon): Likewise.
* tests/cp/no-ctx.sh: Add raw variants to preload lib.
* NEWS: Mention the improvement.
* doc/coreutils.texi (cp invocation): Reference the related --update
option, like we had already done in mv invocation.
* src/cp.c (usage): State clearly what --no-clobber does,
indicating it's protection focused, rather than being update focused.
* doc/coreutils.texi (chown invocation): Convert --from option
description to a macro and call from ...
(chgrp description): ... here.
* src/chown-core.h (emit_from_option_description): A new function
refactored from ...
* src/chown.c (usage): ... here, and called from ...
* src/chgrp.c (usage): ... here.
(main): Accept the --from option as chown(1) does.
* po/POTFILES.in: Add chown-core.h as now translated.
* tests/chown/basic.sh: Decouple the root user from id 0.
* tests/chgrp/from.sh: A new test largely based on chown/basic.sh.
* tests/local.mk: Reference the new test.
* NEWS: Mention the new feature.
Suggested by Ed Neville.
* configure.ac: Remove obsolete macro call.
Recent autoconf warns that it is obsolete.
AC_PROG_CPP sets up the -traditional-cpp option if required.
GCC ignores -traditional since commit f458d1d5 (2002).
Fixes https://bugs.gnu.org/67756
The description of -f regressed in coreutils 9.0
* doc/coreutils.texi (ls invocation): Detail which options
are enabled/disabled with -f.
* src/ls.c (usage): Likewise.
(decode_switches): Update comments.
Fixes https://bugs.gnu.org/67765
* src/touch.c (usage): Reorganise the description to be similar to
the format used for the ls --time description, which formats better
when converted to a man page. Also separate the description
to allow for more granular translations.
Fixes https://bugs.gnu.org/67656
* src/tail.c (file_lines): Ensure we use a buffer size >= PAGE_SIZE when
searching backwards to avoid seeking within a file,
which on sysfs files is accepted but also returns no data.
* tests/tail/tail-sysfs.sh: Add a new test.
* tests/local.mk: Reference the new test.
* NEWS: Mention the bug fix.
Fixes https://bugs.gnu.org/67490
For consistency with the "SI" standard, and with other coreutils
which output a lowercase 'k' in "SI" mode.
* src/numfmt.c (suffix_power): Treat 'k' like 'K' on input.
(double_to_human): Output lowercase 'k' in SI mode.
(usage): Adjust accordingly.
* doc/coreutils.texi: Mention 'k' accepted, and printed in SI mode.
* tests/misc/numfmt.pl: Adjust accordingly.
* NEWS: Mention the change in behavior.
Fixes https://bugs.gnu.org/47103
-w counted bytes not characters, which is wrong in multibyte locales.
This bug exists even in Fedora, which is why the recently-added
test cases from Fedora didn’t catch it.
* src/uniq.c (find_field): New arg PLEN. All callers changed.
Compute length of field correctly in multi-byte locales.
(different): Don’t worry about check_chars; find_field now does that.
* tests/uniq/uniq.pl: Test for this bug.
* src/uniq.c (skip_fields, skip_chars, check_chars, count_occurrences)
(output_unique, output_first_repeated, output_later_repeated)
(delimit_groups): Initialize statically, rather than in ‘main’.
This shrinks the executable a bit.
* src/uniq.c (enum countmode): Remove this type.
(count_occurrences): New static var, replacing the old countmode,
and of type boolean instead of a two-value enum type that was
confusing (and which caused a hard-to-test bug when the count
exceeded INTMAX_MAX - 1). All uses changed.
* src/uniq.c (skip_fields, skip_chars, check_chars, size_opt)
(find_field, different, writeline, check_file, main):
Prefer signed to unsigned integer types, since this allows
for better runtime checking with -fsanitize=undefined.
* src/system.h: Include <stdckdint.h>, since the new
DECIMAL_DIGIT_ACCUMULATE uses it.
Do not include stdckdint.h from files that also include system.h.
(DECIMAL_DIGIT_ACCUMULATE): Omit last arg, which is no longer needed.
Reimplement by using C23-style stdckdint.h’s ckd_mul and ckd_add,
as that’s more standard and is more likely to generate better code.
* src/pinky.c (count_ampersands): Simplify and return idx_t.
(create_fullname): Compute proper destination string size,
basically, by adding (ulen - 1) * ampersands rather than ulen *
(ampersands - 1). Problem found on CHERI-64.
* src/ls.c (print_long_format): Use correct column width,
introduced due to a copy/paste error in commit v9.4-2-gcbb6dfec5
* tests/ls/size-align.sh: Add a test.
* tests/local.mk: Reference the new test.
Fixes https://bugs.gnu.org/66919
* src/join.c (xfields): Simplify and fix bug with fields
that start with a NUL byte when -t is not used.
* tests/misc/join-utf8.sh: Also test when -t is not used,
and when a field starts with NUL.
* NEWS: Mention this.
* bootstrap.conf (gnulib_modules): Remove cu-ctype, as this module
is now more trouble than it’s worth. All uses removed.
Add skipchars.
* gl/lib/cu-ctype.c, gl/lib/cu-ctype.h, gl/modules/cu-ctype:
Remove.
* gl/lib/skipchars.c, gl/lib/skipchars.h, gl/modules/skipchars:
* tests/misc/join-utf8.sh:
New files.
* src/join.c: Include skipchars.h and mcel.h instead of cu-ctype.h.
(tab): Now mcel_t, not int. All uses changed.
(output_separator, output_seplen): New static vars.
(eq_tab, newline_or_blank, comma_or_blank): New functions.
(xfields, prfields, prjoin, add_field_list, main):
Support multi-byte characters.
* src/numfmt.c: Include ctype.h, skipchars.h.
Do not include cu-ctype.h.
(newline_or_blank): New function.
(next_field): Support multi-byte characters.
* src/sort.c: Include ctype.h instead of cu-ctype.h.
(inittables): Open-code field_sep since it no longer exists.
‘sort’ is not multi-byte safe yet, but when it is this code
will need revamping anyway.
* src/uniq.c: Include mcel.h and skipchars.h instead of cu-ctype.h.
(newline_or_blank): New function.
(find_field): Support multi-byte characters.
* tests/local.mk (all_tests): Add tests/misc/join-utf8.sh
* src/dircolors.c: Include c-ctype.h, not ctype.h.
(parse_line): Use c_isspace, not isspace, as the .dircolors
file format (which does not seem to be documented!) appears
to be ASCII.
Include ctype.h only in files that need it. Many of its uses
are incorrect, as they assume single-byte locales. The idea is
to remove the incorrect uses later, when there is time.
* src/chroot.c, src/csplit.c, src/dd.c, src/digest.c, src/dircolors.c:
* src/expand-common.c, src/expand.c, src/fmt.c, src/fold.c, src/ls.c:
* src/od.c, src/pinky.c, src/pr.c, src/ptx.c, src/seq.c:
* src/set-fields.c, src/split.c, src/stdbuf.c, src/test.c:
* src/tr.c, src/truncate.c, src/unexpand.c, src/wc.c:
Include ctype.h.
* src/system.h: Do not include ctype.h.
include ctype.h.o
This is so that we don’t need to have every source file
include ctype.h.
* bootstrap.conf (gnulib_modules): Add cu-ctype.
* gl/lib/cu-ctype.c, gl/lib/cu-ctype.h, gl/modules/cu-ctype:
New files.
* src/join.c, src/numfmt.c, src/sort.c, src/uniq.c:
Include cu-ctype.h, for field_sep.
* src/system.h (field_sep): Remove; now supplied by cu-ctype.
* src/digest.c (valid_digits, split_3):
* src/echo.c (main):
* src/printf.c (print_esc):
* src/ptx.c (unescape_string):
* src/stat.c (print_it):
When the code is supposed to support only POSIX-locale hex digits,
use c_isxdigit rather than isxdigit. Include c-ctype.h as needed.
This defends against oddball locales where isxdigit != c_isxdigit.
This will make decoding more resilient to corruption
whether due to transmission errors or nefarious adjustment.
See https://eprint.iacr.org/2022/361.pdf
* gnulib: Update to commit 3f463202bd enforcing canonical encoding.
* tests/basenc/base64.pl: Add test cases, and adjust existing cases.
* NEWS: Mention the change in behavior.
This sped up ‘basenc -d --base16’ by 60% on my old platform,
AMD Phenom II X4 910e, Fedora 38.
* src/basenc.c (struct base16_decode_context): Simplify by
omitting have_nibble. ‘nibble’ is now negative if it’s missing.
All uses changed.
(B16): New macro, inspired by ../lib/base64.c.
(base16_to_int): New static var, likewise.
(isubase16): Reimplement using base16_to_int, since isxdigit is
not guaranteed to succeed on the chars we want when the locale is
oddball.
(base16_decode_ctx): Tune by using base16_to_int and by
This tends to generate better code, at least on x86-64,
because callers are just as fast and callees can avoid a conversion.
* src/basenc.c: The following renamings also change the arg type
from char to unsigned char. All uses changed.
(isubase): Rename from isbase.
(isubase64url): Rename from isbase64url.
(isubase32hex): Rename from isbase32hex.
(isubase16): Rename from isbase16.
(isuz85): Rename from isz85.
(isubase2): Rename from isbase2.
2023-10-24 Paul Eggert <eggert@cs.ucla.edu>
* src/basenc.c (struct base16_decode_context):
Simplify by storing -1 for missing nibbles. All uses changed.
* src/basenc.c (base16_decode_ctx): Convert to uppercase
before converting from hex.
* tests/basenc/basenc.pl: Add a test case.
* NEWS: Mention the change in behavior.
Addresses https://bugs.gnu.org/66698
Padding of encoded data is useful in cases where
base64 encoded data is concatenated / streamed.
I.e. where there are padding chars _within_ the stream.
In other cases padding is optional and can be inferred.
Note we continue to treat partial padding as invalid,
as that would be indicative of truncation.
* src/basenc.c (do_decode): Auto pad the end of the input.
* NEWS: Mention the change in behavior.
* tests/misc/base64.pl: Adjust to not fail for missing padding.
Addresses https://bugs.gnu.org/66265
* src/wc.c: Do not include assure.h. Replace the only
use of ‘assure’ with ‘unreachable’ which is good enough.
(wc, main): Remove labels and gotos. This doesn’t affect
performance in any way I can measure, and makes the code
a bit easier to follow.
Prefer signed to unsigned integers, to make it easier to catch
integer overflow errors.
* src/wc.c: Do not include safe-read.
(total_lines_overflow, total_words_overflow, total_chars_overflow)
(total_bytes_overflow): Now bool, not uintmax_t. All uses changed.
(max_line_length): Now intmax_t, not uintmax_t. All uses changed.
The total_... vars are still uintmax_t because overflow into them
is checked.
(page_size): Now idx_t, not size_t.
(wc_lines, wc, get_input_fstatus, compute_number_width, main):
Prefer signed to unsigned ints where either should do.
(wc_lines, wc): Use read rather than safe_read, since we don’t
need safe_read’s checks for huge buffers.
(wc): Redo call to mbrtoc32 to lessen the number of comparisons
against its returned value. Do this partly by keeping a pointer
to the end of the buffer rather than a count. Simplify
overflow-checking code.
(compute_number_width): Check for integer overflow.
Don’t assume size_t fits into unsigned long.
* src/wc.h (struct wc_lines): Prefer signed integers.
* src/wc_avx2.c: Do not include safe-read.h.
(wc_lines_avx2): Prefer signed integers. Use read, not safe_read.
* src/wc.c: Use "#include <...>" for files not in the current dir.
Include "wc.h" instead of declaring wc_lines_avx2 by hand.
(wc_lines): New API, with no file name (no longer needed) and
with a return struct rather than arg pointers. All uses changed.
Use avx2_supported directly instead of using a function pointer.
Exploit C99-style declarations after statements.
Multiply by 15 rather than dividing; it’s faster and more accurate
and cannot overflow here.
(wc): Simplify based on wc_lines API change.
* src/wc.h: New file.
* src/wc_avx2.c: Include it, to check API better.
(wc_lines_avx2): Use new API. All uses changed. Exploit C99.
Make locals more local.
* src/factor.c (struct mp_factors): New member nalloc.
(mp_factor_init): Initialize it.
* src/factor.c (mp_factor_insert):
* src/tail.c (parse_options): Use xpalloc to avoid quadratic
worst-case behavior on reallocation.
* src/tail.c (pids_alloc): New static var.
* src/wc.c (SUPPORT_OLD_MBRTOWC): Remove. All uses removed.
(wc): Simplify by assuming C99-or-later behavior for mbrtoc32,
which after all is a C11 API. Fix the !SUPPORT_OLD_MBRTOWC
code, which evidently was never tested seriously.
The 3× speedup was measured by invoking 'wc $(find * -type f)'
on the coreutils sources etc. on an Ubuntu 23.04 x86-64.
These changes also speed up wc 20% in UTF-8 locales.
* src/wc.c (wc_isprint, wc_isspace): New static vars.
(wc): Use them for speed.
(main): Initialize them if needed.
(isnbspace): Remove; no longer used.
* bootstrap.conf (gnulib_modules): Remove c32isprint.
* src/wc.c (wc): Consider all non-white-space characters
to be word constituents, even if they are not printable.
POSIX requires this, and it is what BSD does.
Partly do this by simplifying the check for a word,
by counting word starts rather than word ends.
* tests/wc/wc.pl: Test for the bug.
* m4/jm-macros.m4: Do not check for ftruncate, iswspace,
mkfifo, mbrlen, sysctl. Coreutils no longer uses the
corresponding HAVE_* macros, typically because Gnulib
handles them now.
* src/wc.c (iswspace): Remove; unused.
This should work better on non-glibc platforms that don’t
use Unicode for wchar_t. However, POSIX appears to prohibit
this for printf.c so leave that alone.
* bootstrap.conf (gnulib_modules): Add btoc32, c32iscntrl,
c32isprint, c32isspace, c32width, mbrtoc32. Remove btoc, wcwidth.
* src/df.c, src/ls.c, src/wc.c:
Include uchar.h instead of wchar.h and wctype.h.
* src/df.c (replace_invalid_chars):
* src/ls.c (quote_name_buf):
* src/wc.c (isnbspace, wc):
Use char32_t instead of wchar_t.
* src/wc.c: Don’t have special #ifs for platforms where
MB_LEN_MAX is 1. On these platforms, MB_CUR_MAX is 1 as well,
so the compiler should optimize away all multi-byte code.
This causes Gnulib code to also use mcel, which is more consistent.
* bootstrap.conf (avoided_gnulib_modules): Avoid mbuiter
and mbuiterf, since we can now just use mcel. This avoids
the need to ship and compile mbchar and these modules.
(gnulib_modules): Change mcel to mcel-prefer.
* src/wc.c: Do not include mbchar.h.
(wc): Check for ASCII characters instead of using is_basic.
Other parts of Gnulib and coreutils already assume the encoding
is upward compatible with ASCII, and the old code wouldn’t
have worked anyway with shift-JIS.
The mcel API is simpler and corresponds more closely to how
Emacs etc. behave when the input has encoding errors,
since it treats each encoding-error byte separately.
* bootstrap.conf (gnulib_modules): Add mcel.
* src/expr.c: Include mcel.h instead of mbuiter.h.
(mbs_logical_cspn, mbs_logical_substr, mbs_offset_to_chars):
Use mcel API.
(mbs_logical_substr): Use ximemdup0 so as not to waste memory in
the result, fixing a FIXME.
On gcc 10 the following build failure occurs:
"error: a label can only be part of a statement
and a declaration is not a statement"
This is because the current code is non standards conforming,
but GCC >= 11 will compile it (even with the -Wpedantic option).
This issue is tracked for GCC at:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111526
* src/tail.c (parse_options): Avoid a declaration after label,
by using a surrounding block.
tail can watch multiple files, but currently only a single writer. It
can be useful to watch files from multiple writers, or even processes
not directly related to the files (e.g. watch log files written by a
server process, for the duration of a test driven by a separate
client).
* src/tail.c (writers_are_dead): New function.
(tail_forever): Use it to wait for writers.
(tail_forever_inotify): As above.
(parse_options): Manage --pid options in an array.
* doc/coreutils.texi: Update documentation.
* tests/tail/pid.sh: Add a variant with two PIDs.
* News: Mention the new feature.
Currently --dired is silently ignored
with conflicting output formats
* src/ls.c (decode_switches): Set default format and hyperlink mode
when the --dired option is specified.
* tests/ls/dired.sh: Check that formats are implied / overridden.
* NEWS: Mention the change in behavior.
* doc/coreutils.texi (ls invocation): Adjust --dired description.
* src/copy.c (copy_internal): When cp -f's fstatat fails on the
destination with ELOOP, report an error immediately when fstatat
used AT_SYMLINK_NOFOLLOW, as the later unlinkat would fail too.
This bug occurs only when temporarily setting the mode to the
intersection of old and new modes when changing ownership.
* src/copy.c (owner_failure_ok): Treat EACCES like EPERM.
* src/chown-core.c (restricted_chown): Don’t assume fchown exists.
The Gnulib doc says that nowadays this is needed only for
ports to mingw and MSVC 14, but it’s an easy port so let’s do it.
Remove Gnulib modules that coreutils code no longer use directly.
Some of these are used indirectly, but gnulib-tool should do that.
* bootstrap.conf (gnulib_modules): Remove calloc-gnu, cloexec,
getgroups, getpass-gnu, getugroups, getusershell, gnu-mae,
group-member, lchown, mgetgroups, netinet_in, readlink,
realloc-gnu, rename, rpmatch, stpncpy, tzset, wchar-single,
wcswidth.
* configure.ac (GNULIB_WCHAR_SINGLE_LOCALE): Define.
This can improve performance, while dropping support for
rare encodings on non-GNU platforms. Nowadays these encodings
are typically not worth the hassle.
* src/local.mk (src_timeout_LDADD, src_dd_LDADD)
(src_shred_LDADD, src_sync_LDADD): Use TIMER_TIME_LIB
and FDATASYNC_LIB instead of LIB_TIMER_TIME and
LIB_FDATASYNC.
Omit checks no longer needed now that we use strsignal.
* configure.ac: Do not check for strsignal-related decls.
* src/kill.c (sys_siglist, strsignal): Remove.
This simplifies memory allocation a bit, and removes an arbitrary
limitation from numfmt, which formerly limited cell output to 127
bytes.
* bootstrap.conf (gnulib_modules): Remove mbsalign, strncat.
Add strnlen (the code already used strnlen directly, and we were
saved only because Gnulib used the module indirectly)
* gl/lib/mbsalign.c, gl/lib/mbsalign.h, gl/modules/mbsalign:
* gl/modules/mbsalign-tests, gl/tests/test-mbsalign.c: Remove.
* src/df.c, src/ls.c: Do not include mbsalign.h.
(MBSWIDTH_FLAGS): New constant, now used for all
mbswidth calls. All callers changed to check for -1 return.
* src/df.c (struct field_data_t): ‘width’ is now int not size_t,
since mbswidth can’t do widths greater than INT_MAX anyway.
Replace ‘align’ with ‘align_right’. All uses changed.
(print_table): Redo to avoid the need for ambsalign.
(get_header, get_dev): mbswidth returns int, not size_t.
* src/ls.c (MAX_MON_WIDTH): Remove; no longer used.
(abmon_init): Use strnlen to cheaply discard too-long month names.
Align by hand instead of using mbsalign.
* src/numfmt.c: Include stdckdint.h, mbswidth.h.
Do not include mbsalign.h.
(padding_buffer_size): Now idx_t. All uses changed.
(padding_width): Now intmax_t, since it’s no longer an object
size. Its sign now records alignment. All uses changed.
(zero_padding_width): Now int, since it’s given to sprintf.
All uses changed.
(padding_alignment): Remove; it’s now taken from padding_width’s sign.
(double_to_human): Return string length. BUF_SIZE arg is now idx_t.
Include suffix in output. All callers changed. Simplify by not
calling strncat or stpcpy. Calculate fmt size bound more carefully.
(setup_padding_buffer): Remove. All uses removed.
(parse_format_string): Use intmax_t, not long, for pad.
On overflow, set widths to large values that cause later code
to do the right thing, rather than separately checking for
overflow here.
(prepare_padded_number): Return bool, not int 0/1. New arg
PADDING. All uses changed. Do not limit padded output to 127
bytes; instead, use xpalloc to expand the output buffer.
(print_padded_number): New arg PADDING. All uses changed.
(process_suffixed_number): Simplify.
(main): Take extremum if xstrtoimax overflows, as this does
the right thing.
* tests/misc/numfmt.pl: New test suf-20 to test for truncation bug.
Remove tests pad-3.2, fmt-err-7, as they’re no longer invalid but
are quite expensive.
Most of this just affects commentary and documentations. The only
significant behavior change is translating author names via
proper_name_lite rather than proper_name_utf8, or not translating
them at all. proper_name_lite is good enough for coreutils and
avoids the bloat that had coreutils not using Gnulib proper_name.
* bootstrap.conf (gnulib_modules): Use propername-lite instead
of propername.
(XGETTEXT_OPTIONS): Look for proper_name_lite instead of for
proper_name_utf8.
* cfg.mk (local-checks-to-skip): Remove
sc_proper_name_utf8_requires_ICONV, since we no longer use
proper_name_utf8.
(old_NEWS_hash): Update.
(sc_check-I18N-AUTHORS): Remove; no longer needed.
* tests/sort/sort-continue.sh: Use ulimit -n 7 not -n 6. On
Solaris 10 'sort' uses Gnulib mkostemp, which calls Gnulib
getrandom, which opens /dev/urandom to calculate the temp file's
name, which means 'sort' needs one more file descriptor to work.
* tests/cksum/md5sum-bsd.sh: Avoid part of test dealing with backslashes
in file names, on systems where backslash is a directory separator.
Issue reported by Bruno Haible on cygwin.
Following commit v9.3-80-g5e1e0993b which makes cksum
match the output of the standalone utilities...
* doc/coreutils.texi (cksum output modes): Remove the mention
that cksum never outputs a binary indicator, as that's no longer the
case.
* tests/cksum/b2sum.sh: Avoid outputting a binary indicator.
* tests/cksum/sm3sum.pl: Likewise.
* src/system.h (write_error): Also call fpurge(), which was seen to
be needed on FreeBSD 13.1 to avoid duplicated write errors.
* src/head.c (xwrite_stdout): Likewise.
* bootstrap.conf: Depend on fpurge.
Reported by Bruno Haible.
* doc/coreutils.texi: Reorg so that 'cksum invocation' is the
main node listing all options and output formats, which is then
referenced by the descriptions of the standalone utilities.
Use macros in the description of the standalone utilities
rather than referencing 'md5sum invocation' to be more direct.
* src/cp.c (main): Set default reflink mode appropriately
with --sparse=never.
* src/copy.c (infer_scantype): Add a comment to related code.
* tests/cp/sparse-2.sh: Add a test case.
* NEWS: Mention the bug.
* tests/sort/sort-debug-keys.sh: Decimal point was seen to be '.'
on fr_FR.UTF-8 on Alpine Linux 3.18, so add an extra guard
to ensure we've a ',' as the decimal point on this locale.
Fixes https://bugs.gnu.org/65310
* src/uptime.c (print_uptime): Check for overflow
when computing uptime. Use C99-style decl after statements.
Do not let an idx_t value go negative.
(print_uptime, uptime): Be more generous about read_utmp failures,
or when read_utmp does not report the boot time. Instead of
failing, warn but keep going, printing the information that we did
get, and then exit with nonzero status.
(print_uptime): Return the desired exit status. Caller changed.
* src/uptime.c: Don't include c-strtod.h.
(print_uptime): Don't read /proc/uptime, because the value it provides
does not change when a date adjustment occurs.
* bootstrap.conf (gnulib_modules): Remove 'uptime'.
Older systems that had issues with these like HP-UX and Solaris 8
are now obsolete, and can easily apply patches to provide support.
Also we've used %td since coreutils 9.1, with no reported issues.
* cfg.mk (sc_prohibit-c99-printf-format): Remove to allow use of
%[jtz] size specifiers, which allows for cleaner code
by avoiding the need to cast to PRI?MAX etc.
* NEWS: Separate out the description of the _existing_ issues
with outputting timestamps on 32 bit systems, from _future_ issues
outputting timestamps on all systems. Also move this to the
"improvement" section, since it's not really a coreutils
specific issue, and also is a build time configurable option.
When we are only interested in entries of type USER_PROCESS, tell
read_utmp that it does not need to determine the boot time.
* src/pinky.c (short_pinky): Pass option READ_UTMP_USER_PROCESS.
* src/users.c (users): Likewise.
* src/who.c (who): Likewise, if calling list_entries_who.
(This patch is coauthored with Bruno Haible,
with original version at <https://bugs.gnu.org/64937#>.)
This updates the gnulib submodule to latest.
For year-2038 safety on Linux/{x86,arm},
this adds an --enable-systemd option to ‘configure’.
The idea is that this sort of thing will become the default
after it has been tested more.
* configure.ac: Don't test whether struct utmp and struct utmpx
have the ut_host field; this is now done in gnulib's readutmp module.
* src/local.mk: Link the programs 'pinky', 'uptime', 'users',
'who' with $(READUTMP_LIB).
* src/pinky.c, src/who.c:
Test HAVE_STRUCT_XTMP_UT_HOST instead of HAVE_UT_HOST.
* src/pinky.c (print_entry):
* src/who.c (print_user, print_deadprocs, print_login)
(print_initspawn, scan_entries):
Support the situation where ut_line is a 'char *' rather than a
'char[]' of fixed size. Likewise for ut_user and ut_host.
(make_id_equals_comment): Likewise for ut_id.
* src/pinky.c (print_entry):
* src/who.c (print_user):
Open /dev to simplify looking up its entries.
Don’t use printf if the output might in theory be longer than INT_MAX.
* src/pinky.c (scan_entries, short_pinky):
* src/uptime.c (print_uptime, uptime):
* src/users.c (list_entries_users, users):
* src/who.c (who):
Use idx_t where new read_utmp needs it.
* src/system.h (STREQ_LEN): Add comment that last arg can be -1.
* src/uptime.c (print_uptime): Prefer signed types.
Fix unlikely bug on platforms with 32-bit long and 64-bit time_t
if the idle time exceeds 2**31 days (about 6 million years...).
* src/cut.c: Complete the error-handling improvements started in
commit e0a4a60af5, by adding a couple of remaining checks for putchar().
While there, sprinkle a few rather useful comments, and perform a few
small code cleanups, to make the code and the comments more uniform
and more conformant to the official coding style. Also make the help
message slightly more uniform.
* src/ioblksize.h: Avoid syntax check with redundant idx.h inclusion.
* src/od.c (FMT_BYTES_ALLOCATED): Increase by two to avoid:
error: '%s' directive writing between 1 and 2 bytes into a region
of size between 1 and 4 [-Werror=format-overflow=]
(maint): Use %td to print idx_t rather than invalid %jt format.
* src/pinky.c (idle_string): Prefer intmax_t to unsigned long int;
this avoids an overflow on platforms where unsigned long is 32
bits and time_t is 64 bits (the bug could occur on such a system
that was idle for more than 6 million years, so it’s a bit
hard to supply a test case...).
* src/numfmt.c (suffix_power_char, powerld, expld)
(simple_strtod_int, double_to_human, prepare_padded_number)
(process_suffixed_number): Prefer signed types.
(process_suffixed_number): Fix an unlikely bug if an
arg has exactly 2**32 spaces at the start.
* src/kill.c (list_signals):
Prefer signed types. This avoids undefined behavior on
theoretical platforms where unsigned and signed int have
different representations.
* src/od.c: Include stdckdint.h.
(bytes_to_oct_digits, bytes_to_signed_dec_digits)
(bytes_to_unsigned_dec_digits, bytes_to_hex_digits):
Use ‘char’ for these small constants.
(simple_strtoi): Rename from simple_strtoul. Convert to int
instead of unsigned long; that’s good enough. All uses changed.
Simplify by using ckd_mul and ckd_add to check for overflow.
(main): Prefer signed types to unsigned.
* src/cksum.c (main):
* src/df.c (decode_output_arg):
* src/digest.c (valid_digits):
Prefer idx_t to unsigned types when the value is an index
into an array.
When it’s easy, prefer signed types to unsigned, as
they are less confusing and allow overflow checking.
* src/factor.c (struct mp_factors, udiv_qrnnd)
(count_leading_zeros, count_trailing_zeros)
(factor_insert_multiplicity, mp_factor_clear, mp_factor_insert)
(factor_insert_refind, factor_using_division)
(mp_factor_using_division, powm2, millerrabin, millerrabin2)
(mp_millerrabin, prime_p, prime2_p, mp_prime_p, isqrt, isqrt2)
(invtab, q_freq, factor_using_squfof, strto2uintmax)
(print_factors_single, main):
Prefer signed integers to unsigned.
Change stzncpy’s implementation to match its comment, in the case
where SRC + LEN would overflow. This case never happens in coreutils.
* src/system.h (stzncpy): Work even if SRC + LEN would overflow.
This can be reproduced by getting the read() above 2G,
which induces a short read, thus triggering the erroneous failure.
$ truncate -s 5G 5G
$ cat 5G | TMPDIR=$PWD tac | wc -c
tac: /tmp/tacFt7txA: read error: Illegal seek
0
With the fix in place we now get:
$ cat 5G | TMPDIR=$PWD src/tac | wc -c
5368709120
* src/tac.c (tac_seekable): Use full_read() to handle short reads.
* NEWS: Mention the bug fix.
Reported at https://bugs.debian.org/1042546
Problem reported by Nir Oren <https://bugs.gnu.org/64785>.
* src/copy.c (copy_internal): Use a more-specific diagnostic when
a rename fails due to a problem that must be due to the
destination, avoiding user confusion in cases like 'mv dir x'
where x is a nonempty directory.
* tests/mv/dir2dir.sh: Adjust to match.
tail -n/-c +NUM, is different from tail -n/-c NUM,
and head -n/-c NUM, and head -n/c -NUM, in that it
specifies a 1 based index rather than a count to skip/include.
So clarify this in tail --help and tail info manual.
Note we also mention this gotcha at:
https://www.pixelbeat.org/docs/coreutils-gotchas.html#tail
* doc/coreutils.texi (tail invocation): Give examples for -c/-n +NUM,
to make it clear one has to specify a number 1 larger than
might be expected.
* src/tail.c (usage): State the skip at start edge case more clearly
in the -n description. -c is not often used with tail so we leave
full explanation of that to the info manual. Also split the string
to simplify translation.
* tests/split/l-chunk.sh: Move the "expensive" portion to ...
* tests/split/l-chunk-root.sh: .. A new test split from l-chunk.sh
which uses an isolated TMPDIR, rather than exhausting /tmp,
as that gives false positive failures with some other coreutils tests
like tac-2-nonseekable.sh and shuf-reservoir.sh at least.
* tests/local.mk: Reference the new test.
* bootstrap.conf: Depend on tmpdir rather than tmpfile,
as the standard tmpfile() doesn't honor $TMPDIR.
* src/split.c (copy_to_tmpfile): Adjust to call temp_stream() rather
than tmpfile();
* NEWS: Mention the improvement.
This also refactors temp_stream() to its own module,
in preparation for use by split.
* src/tac.c: Refactor temp_stream() out to ...
* src/temp-stream.c: ... A new module mostly refactored from tac,
but uses tmpdir to more robustly support $TMPDIR,
while falling back to /tmp if not available.
* src/temp-stream.h: The new module interface.
* src/local.mk: Reference the new module from tac.
* tests/tac/tac.pl: Adjust to non failing missing $TMPDIR.
* po/POTFILES.in: Reference the new module with translatable strings.
* NEWS: Mention the user visible improvements to tac TMPDIR handling.
One needs to include stdlib--.h if using mkstemp()
lest one hits esoteric bugs with closed stdin etc.
* cfg.mk (sc_require_stdlib_safer): Add a new syntax check.
(sc_require_stdio_safer): Fix this; broken since commit fa7ed969c3.
* src/join.c (prjoin): Check for write errors after each line.
* tests/misc/write-errors.sh: enable the test for join.
* NEWS: Mention the improvement.
* src/comm.c (writeline): Simplify by removing the unneeded STREAM
parameter. Call write_error() upon ferror().
(compare_files): Adjust to simplified writeline().
* tests/misc/write-errors.sh: Enable comm test.
* NEWS: Mention the improvement.
* src/cut.c (cut_bytes): Diagnose errors from fwrite() and putchar().
(cut_fields): Likewise.
* tests/misc/write-errors.sh: Enable the test for cut,
and augment to cover both cut_bytes() and cut_fields().
* NEWS: Mention the improvement.
* src/uniq.c (write_line): Check the output from fwrite() immediately.
(check_file): Likewise.
* tests/misc/write-errors.sh: Enable the test case.
* NEWS: Mention the improvement.
* src/od.c (dump): Check for write errors after each block written,
to exit early even with large / unbounded inputs.
* tests/misc/write-errors.sh: enable od check.
* NEWS: Mention the improvement.
Fixes https://bugs.gnu.org/64540
* cfg.mk (sc_some_programs_must_avoid_exit_failure): Adjust to
avoid false positive.
(sc_prohibit_exit_write_error): A new syntax check to prohibit
open coding error(..., "write error"); instead directing to use...
* src/system.h (write_error): ... a new function to clear stdout errors
before we explicitly diagnose a write error and exit.
* src/basenc.c: Use write_error() to ensure no repeated diagnostics.
* src/cat.c: Likewise.
* src/expand.c: Likewise.
* src/factor.c: Likewise.
* src/paste.c: Likewise.
* src/seq.c: Likewise.
* src/shuf.c: Likewise.
* src/split.c: Likewise.
* src/tail.c: Likewise.
* src/tr.c: Likewise.
* src/unexpand.c: Likewise.
* tests/misc/write-errors.sh: Remove TODOs for the fixed utilities:
expand, factor, paste, shuf, tr, unexpand.
* src/digest.c (problematic_chars): This recently introduced
function does not modify state so is pure, even though GCC 13.1 at least
did not warn about that attribute being appropriate.
* src/digest.c (digest_check): Also escape in the case that the
file name contains '\'.
* tests/cksum/md5sum-bsd.sh: Add a test case.
* doc/coreutils.texi (md5um invocation): Clarify escaping operation.
* NEWS: Mention the bug fix.
Fixes https://bugs.gnu.org/64392
Support -b, --binary, and -t, --text
to allow full emulation of older utilities with:
exec cksum -a $algo --untagged "$@"
Note this would diverge from OpenBSD's support of cksum -b.
* src/digest.c: Change -b to mean --binary, not --base64 in all cases.
Accept -b and -t in all cases. Keep --binary and --text undocumented
for cksum.
* tests/cksum/cksum-base64.pl: s/-b/--base64/.
* tests/cksum/cksum-a.sh: Ensure cksum supports -b and -t appropriately.
* NEWS: Mention the change in behavior.
* bootstrap.conf (gnulib_modules): Add stdckdint.
Also, in C source code, prefer C23 macros like ckd_add
to their Gnulib near-equivalents like INT_ADD_WRAPV.
Include <stdckdint.h> as needed.
* src/who.c (idle_string): Avoid signed integer overflow
if the superuser messes with the clock in bizarre ways.
Remove an ‘assume’ that wasn’t correct under this scenario.
* src/cut.c (cut_file):
* src/nl.c (nl_file): Pacify GCC Bug#109613 in a better way, by
narrowing the coverage of the ‘assume’ so that bugs in the
no-longer-covered part are not masked.
* src/fmt.c (get_paragraph):
* src/stty.c (display_changed, display_all): Omit calls to
‘assume’ that are present only to pacify false positives by Parfait
<https://labs.oracle.com/pls/apex/f?p=94065:12:17236785746387:13>,
which went in-house in 2012 and never came back.
Now that Gnulib’s ‘error’ module does proper static checking
for not returning, we need no longer use the ‘die’ macro.
This makes code easier to read for people that are used to ‘error’.
* cfg.mk (error_fns, exclude_file_name_regexp): Remove ‘die’.
(sc_die_EXIT_FAILURE): Remove.
* src/die.h: Remove. All includes removed. All calls to ‘die’
changed back to calls to ‘error’.
* src/install.c (get_ids): Use quoteaf (problem found with
make syntax-check).
* src/system.h: Include error.h, since some of our macros call ‘error’.
Stop including error.h elsewhere.
This modernizes the source code somewhat, to take advantage
of advances in GCC over the years, and Gnulib’s ‘assure’ module.
Include assure.h in files that now need it.
Do not include assert.h directly; it’s no longer needed.
* bootstrap.conf (gnulib_modules): Add ‘assure’.
* gl/lib/randread.c (randread_error):
* src/chmod.c (describe_change):
* src/chown-core.c (describe_change):
* src/cp.c (decode_preserve_arg):
* src/head.c (diagnose_copy_fd_failure):
* src/ls.c (parse_ls_color):
* src/od.c (decode_one_format):
* src/split.c (main):
* src/test.c (binary_operator, posixtest):
Prefer affirm to abort, since it has better diagnostics in the
normal case and better performance with -DNDEBUG.
* gl/lib/xdectoint.c, src/die.h: Include stddef.h, for unreachable.
* gl/lib/xdectoint.c: Do not include verify.h; no longer needed.
* gl/lib/xdectoint.c (__xnumtoint):
* src/die.h (die):
Prefer C23 unreachable () to assume (false).
* gl/lib/xfts.c (xfts_open):
* src/basenc.c (base32hex_encode):
* src/copy.c (abandon_move, copy_internal, valid_options):
* src/cut.c (cut_fields):
* src/df.c (alloc_field, decode_output_arg, get_dev):
* src/du.c (process_file, main):
* src/echo.c (usage):
* src/factor.c (udiv_qrnnd, mod2, gcd2_odd, factor_insert_large)
(mulredc2, factor_using_pollard_rho, isqrt2, div_smallq)
(factor_using_squfof):
* src/iopoll.c (iopoll_internal, fwrite_wait):
* src/join.c (add_field):
* src/ls.c (dev_ino_pop, main, gobble_file, sort_files):
* src/mv.c (do_move):
* src/od.c (decode_format_string, read_block, dump, main):
* src/remove.c (rm):
* src/rm.c (main):
* src/sort.c (stream_open):
* src/split.c (next_file_name, lines_chunk_split):
* src/stdbuf.c (main):
* src/stty.c (set_speed):
* src/tac-pipe.c (line_ptr_decrement, line_ptr_increment):
* src/touch.c (touch):
* src/tr.c (find_bracketed_repeat, get_next)
(validate_case_classes, get_spec_stats, string2_extend, main):
* src/tsort.c (search_item, tsort):
* src/wc.c (main):
Prefer affirm to assert, as it allows for better static
checking when compiling with -DNDEBUG.
* src/chown-core.c (change_file_owner):
* src/df.c (get_field_list):
* src/expr.c (printv, null, tostring, toarith, eval2):
* src/ls.c (time_type_to_statx, calc_req_mask, get_funky_string)
(print_long_format):
* src/numfmt.c (simple_strtod_fatal):
* src/od.c (decode_one_format):
* src/stty.c (mode_type_flag):
* src/tail.c (xlseek):
* src/tr.c (is_char_class_member, get_next, get_spec_stats)
(string2_extend):
Prefer unreachable () to abort () or assert (false) when merely
pacifying the compiler, e.g., in a switch statement on an enum
where all cases are covered.
* src/copy.c (valid_options): Now returns void; the bool was useless.
Caller no longer needs to assert.
* src/csplit.c (find_line):
* src/expand-common.c (next_file):
* src/shred.c (incname):
* src/sort.c (main):
* src/tr.c (append_normal_char, append_range, append_char_class)
(append_repeated_char, append_equiv_class):
* src/tsort.c (search_item):
Omit assert, since the hardware will check for us.
* src/df.c (header_mode): Now the enum type it should have been.
* src/du.c (process_file):
* src/ls.c (assert_matching_dev_ino):
* src/tail.c (valid_file_spec):
* src/tr.c (validate_case_classes):
Mark defns with MAYBE_UNUSED if they’re not used when -DNDEBUG.
* src/factor.c (prime_p, prime2_p, mp_prime_p): Now ATTRIBUTE_PURE.
Prefer affirm to error+abort. No need to translate this diagnostic.
* src/fmt.c (get_paragraph):
* src/stty.c (display_changed, display_all, sane_mode):
* src/who.c (idle_string):
Prefer assume to assert, since the goal is merely pacification
and assert doesn’t pacify anyway if -DNDEBUG is used.
* src/join.c (decode_field_spec):
Omit unreachable abort.
* src/ls.c (assert_matching_dev_ino, main):
* src/tr.c (get_next):
Prefer assure to assert, since the check is relatively expensive
and won’t help static analysis.
* src/ls.c (main):
Prefer static_assert to assert of a constant expression.
(format_inode): Redo to make it clear that buflen doesn’t matter,
and that buf must have a certain number of bytes. All callers changed.
This pacifies -Wformat-overflow.
* src/od.c (decode_one_format):
Omit an assert that tested for obviously undefined behavior,
as the compiler could optimize it away anyway.
* src/od.c (decode_one_format, decode_format_string):
Prefer ATTRIBUTE_NONNULL to runtime checking.
* src/stat.c: Do not include <stddef.h> since system.h does that now.
* src/sync.c (sync_arg):
Prefer unreachable () to assert (true), which was a typo.
* src/system.h: Include stddef.h, for unreachable.
* src/tail.c (xlseek): Simplify by relying on ‘error’ to exit.
* src/digest.c (split_3): Reinstate the check for whitespace after the
digest portion of the line, so that we exit early before inspecting
the file name which would be outside the passed buffer in the case
where the input does not contain a newline.
* tests/cksum/b2sum.sh: Add a test case.
* NEWS: Mention the bug fix.
* THANKS.in: Add Frank Busse who has reported multiple bugs using KLEE.
Fixes https://bugs.gnu.org/64229
* NEWS: Mention the error message to aid those searching
for solutions to the issue, and mention cksum also
as that was confirmed to fix the error with the adjusted
cpu feature detection, as discussed at https://bugs.debian.org/1037264
* src/cksum.c: Cleanup syntax-check failure from previous commit.
* src/cksum.c (cksum_pclmul) [!CRCTAB && !USE_PCLMUL_CRC32]:
Remove macro.
(cksum_fp): No longer file-scope.
(pclmul_supported): Define only if USE_PCLMUL_CRC32.
This omits the debug output "using generic hardware support"
for simplicity and consistency with wc’s output.
(crc_sum_stream) [!USE_PCLMUL_32]: No need for static function pointer.
* src/wc.c (wc_lines_p) [USE_AVX2_WC_LINECOUNT]: No longer file-scope.
(wc) [USE_AVX2_WC_LINECOUNT]: Check for avx2 support at most once,
which was surely the code’s original intent.
(wc) [!USE_AVX2_WC_LINECOUNT]: No need for static function pointer.
This fixes a typo in the previous patch.
Problem reported by Pádraig Brady <https://bugs.gnu.org/64058#11>.
* src/cksum.c (pclmul_supported): Also require AVX support
to use cksum_pclmul.
Problem reported by Dave Hansen <https://bugs.gnu.org/64058>.
Apply similar change to cksum and pclmul, too.
* NEWS: Mention wc fix.
* configure.ac (cpuid_exists, get_cpuid_count_exists):
Remove. All uses removed, since we no longer use __get_cpuid or
__get_cpuid_count.
(pclmul_intrinsic_exists, avx2_intrinsic_exists): Set to no if
__builtin_cpu_supports calls cannot be compiled.
(HAVE_PCLMUL_INTRINSIC, HAVE_AVX2_INTRINSIC): Remove; unused.
Simplify surrounding code because of this.
* src/cksum.c (pclmul_supported):
* src/wc.c (avx2_supported):
Use __builtin_cpu_supports instead of doing it by hand.
Simplify surrounding code because of this.
* src/dircolors.hin: Sort backup section by extension.
Treat .dpkg-new and .dpkg-tmp as backup files.
Treat .crdownload (Chromium based browsers' partial download)
as a backup file.
* src/dd.c (parse_integer): Use recursion to support more than two
multipliers. Also protect suffix[-1] access to ensure we don't
inspect before the passed string.
* tests/dd/bytes.sh: Add test cases.
* doc/coreutils.texi (dd invocation): Note the support for specifying
many multipliers in a number.
* NEWS: Mention the bug fix.
Fixes https://bugs.debian.org/1037275
* doc/coreutils.texi (od invocation): Remove mention of ASCII,
as all printable characters in unibyte locales are output.
* src/od.c (usage): Clarify that only NUL terminated strings
are displayed, and that it's printable chars, not only graphic chars
that are output. I.e., spaces are output also if part of the string.
Reported at https://bugs.ddebian.org/1037217
* src/ls.c (gobble_file): Ensure we lstat() a symlink
specified on the command line, if we receive ELOOP from stat().
* tests/ls/symlink-loop.sh: Add a new test.
* tests/local.mk: Reference the new test.
* NEWS: Mention the bug fix.
Fixes https://bugs.gnu.org/63931
* src/ls.c (gobble_file): stat() symlinks directly,
rather than their targets. This will be more consistent
with how symlinks are generally accessed.
(make_link_name): Remove no longer used function.
Addresses https://bugs.gnu.org/63931
* NEWS: Mention the improvement in reinstating runtime avoidance
of copy_file_range(), that came with the last gnulib update,
picking up gnulib commit fb034b35.
Note mktemp --suffix has the same inconsistency,
but mktemp -d does support creating dirs
so probably best to leave that as is.
* src/split.c (main): Check for trailing /.
* tests/split/additional-suffix.sh: Augment the test.
Reported in https://bugs.debian.org/1036827
* src/dd.c: Don't include no longer used error.h.
Use quoteaf() rather than quote() to quote appropriate for the shell
and to avoid the syntax-check failure,
* src/stty.c: Use quoteaf() rather than quotef()
to have more consistent quoting of the invalid arg.
src/cfg.mk (sc_error_quotes, sc_error_shell_quotes,
sc_error_shell_always_quotes): Include "die" and "diagnose"
in the class of error functions to check arguments for.
* src/dd.c (_GL_NO_INLINE_ERROR): Remove; no longer needed.
(diagnose): Rename from nl_error and omit first arg since it is
always zero. All uses changed.
(error): Remove macro.
Following on from commit v9.0-15-gaa31b919c
which updated README-prereq...
* bootstrap.conf: Add an explicit requirement on m4.
Add an explicit requirement on texi2pdf which is often
packaged separately to makeinfo and induces a failure
far down the distribution phase if not present.
Replace the rsync dependency with wget,
which gnulib changed to in 2018.
This reverts commit 800c86d5, as that was deemed too invasive.
We do keep the change to tee.c to allow using -O3 without warnings.
For other optimization options like -O0, -Og, -O1, -Os,
one can use WERROR_CFLAGS= to stop warnings inducing a build failure.
Allow easily building a debug build for example with:
make CFLAGS='-O0 -ggdb'
False -Wmaybe-uninitialized warnings hit in different
places depending on the compiler passes used.
These changes were tested with gcc 10.2.1, 12.2.1, and 13.1.1 like:
for o in g s z fast 0 1 2 3; do
make clean && make -j$(nproc) CFLAGS="-O$o" || break
done
* src/digest.c: Disable -Wmaybe-uninitialized that gives
false positive here at -O0.
* src/ln.c: Avoid -Wmaybe-uninitialized that gives
false positive here at -O1.
* src/pr.c: Likewise.
* src/sort.c: Likewise.
* src/tee.c: Avoid -Wmaybe-uninitialized that gives
false positive here at -O3 on gcc 13.1.1 at least.
* src/cp.c: Avoid -Wmaybe-uninitialized that gives
false positive here at -Os on gcc 13.1.1 at least.
* src/copy.c: Avoid -Wmaybe-uninitialized that gives
false positive here at -Og on gcc 13.1.1 at least.
* src/head.c: Likewise.
* src/paste.c: Likewise.
Tested on gcc 13.1.1 with: make CFLAGS='-O0 -ggdb'
* configure.ac: Disable -Wstringop-overflow for gnulib.
This warning is far too problematic in my experience:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88443
and triggers with gcc -O0 with versions 12,13 at least.
As split is often dealing with large files,
ensure we indicate to the kernel our sequential access pattern.
This was seen to operate 5% faster when reading from SSD,
as tested with:
dd bs=1M count=2K if=/dev/urandom of=big.in
for split in split.orig split; do
# Ensure big file is not cached
dd of=big.in oflag=nocache conv=notrunc,fdatasync count=0 status=none
# Test read efficiency
CWD=$PWD; (cd /dev/shm && time $CWD/src/$split -n2 $CWD/big.in)
done
real 0m9.039s
user 0m0.055s
sys 0m3.510s
real 0m8.568s
user 0m0.056s
sys 0m3.752s
* src/split.c (main): Use fdadvise to help the kernel
choose a more appropriate readahead buffer.
* NEWS: Mention the improvement.
This fixes failures in "very-expensive" tests on FTS with many
directory entries:
FAIL: tests/rm/ext3-perf
FAIL: tests/rm/many-dir-entries-vs-OOM
The following shows the problem in the former of the above tests:
$ mkdir d && seq 400000 | env -C d xargs touch )
$ rm -rf d
rm: traversal failed: d: Operation not supported
Gnulib commit 3f0950f65abb (2023-04-26) introduced this regression
which was fixed again with gnulib commit d4d8abb39eb0.
See discussion in
<https://lists.gnu.org/r/bug-gnulib/2023-05/msg00040.html>
* bootstrap.conf (gnulib_modules): Change "year2038-required" to
"year2038-recommended"; the module has been replaced.
* gnulib: Update to latest.
* tests/init.sh: Likewise.
Before:
$ pr --expand-tabs=
pr: '-e' extra characters or invalid number in the argument:
‘SHELL=/bin/bash’: Value too large for defined data type
After:
$ pr --expand-tabs=
pr: '-e': Invalid argument: ‘’
* src/pr.c (getoptarg): Ensure we don't parse beyond the
end of an empty argument, thus outputting arbitrary stack
info in subsequent error messages.
Addresses https://bugs.debian.org/1035596
This doesn’t change behavior; it just clarifies the code a bit.
* src/cp.c (re_protect): New arg DST_SRC_NAME, for clarity, and so
that we need to skip '/'s only once. Caller changed.
Rename a couple of local variables to try to make things clearer.
* src/cp.c (re_protect): Ensure copy_acl() is passed an absolute path.
* tests/cp/cp-parents.sh: Add a test case.
* NEWS: Mention the bug.
Fixes https://bugs.gnu.org/63245
* README: State that DEBUG=yes is particularly useful with perl tests.
* tests/split/l-chunk.sh: Use the more standard $DEBUG variable
rather than an internal $DEBUGGING variable.
* configure.ac (WERROR_CFLAGS): Omit mention of
-Wno-analyzer-double-free, -Wno-analyzer-null-dereference, and
-Wno-analyzer-use-after-free as manywarnings no longer uses them.
* src/csplit.c, src/fmt.c, src/make-prime-list.c, src/nohup.c:
Add pragmas to pacify GCC 13 when coreutils is configured
with --enable-gcc-warnings='expensive'.
* src/chmod.c (main): Use xpalloc instead of X2REALLOC,
and make the corresponding variables signed instead of unsigned.
When reallocating the buffer, this grows it by a factor of 1.5, not 2.
This also pacifies gcc -Wanalyzer-null-dereference.
* src/csplit.c (load_buffer): Refactor for clarity.
This also xpacifies gcc -Wanalyzer-use-of-uninitialized-value.
When reallocating the buffer, grow it by a factor of 1.5, not 2.
* tests/misc/read-errors.sh: Exercise more modes of
various utilities for better read error coverage.
* tests/split/fail.sh: Remove part refactored into the above test.
Avoid the following error with -mno-ssse3:
inlining failed in call to 'always_inline' '_mm_shuffle_epi8':
target specific option mismatch
* configure.ac: Ensure we use ssse3 specific code when
checking whether to enable the pclmul cksum implementation.
* src/pr.c (init_parameters): Ensure we avoid a 0 lines_per_body
which was possible when adjusting for double spacing.
That caused print_page() to always return true,
causing an infinite loop.
* tests/pr/pr-tests.pl: Add a test case.
* NEWS: Mention the fix.
Fixes https://bugs.debian.org/1034808
Since skipping of files is central to the operation of -i and -u,
and with -u one may be updating few files out of many,
reinstate the verbosity of this functionality as it was before 9.3.
* src/copy.c (copy_internal): Only output "skipped" message
with --debug. Also adjust so message never changes with --debug.
* tests/cp/cp-i.sh: Adjust accordingly.
* tests/mv/mv-n.sh: Likewise.
* tests/cp/debug.sh: Add explicit test case for message.
* NEWS: Mention the change in behavior.
When run under QEmu emulation emulated /proc files have
unstable inode numbers.
* tests/cp/proc-short-read.sh: Skip if unstable inode numbers detected.
* src/install.c (strip): Prepend "./" to file names with a leading "-".
* tests/install/strip-program.sh: Add a test case.
* NEWS: Mention the bug fix.
Reported in https://bugs.debian.org/1034429
* tests/misc/tty-eof.pl: Ensure we don't erroneously
skip commands with parameters.
Comment as to why cut(1) is treated differently.
Adjust expect calls to not wait needlessly for cut output.
* tests/cp/sparse-2.sh: Don't depend on the copy taking
<= allocation of the source. Instead leverage --debug
to check that zero detection is being enabled.
Fix a build failure seen on gcc 3.4 on Solaris 10 at least.
* src/crctab.c: Ensure we include config.h for all compilation units.
This is now required for new _Noreturn usage in gnulib for stdint.h.
* src/cksum.c: Update generation code to ensure config.h included.
* cfg.mk: Remove crctab.c exclusion from the config.h check.
* src/wc.c (wc): Update the offset when not reading,
and do read if we can't update the offset.
* tests/misc/wc-proc.sh: Add a test case.
* NEWS: Mention the bug fix.
Fixes https://bugs.gnu.org/61300
* NEWS: Mention the change in behavior to issue a "not replaced"
error diagnostic with -n, and the "skipped" message with -v.
* src/copy.c (copy_internal): Adjust to output the "skipped" messages
depending on -i, -n, -u.
* tests/cp/cp-i.sh: Adjust accordingly.
* tests/mv/mv-n.sh: Likewise.
Add --update=none which is equivalent to the --no-clobber behavior
from before coreutils 9.2. I.e. existing files are unconditionally
skipped, and them not being replaced does not affect the exit status.
* src/copy.h [enum Update_type]: A new type to support parameters
to the --update command line option.
[enum Interactive]: Add I_ALWAYS_SKIP.
* src/copy.c: Treat I_ALWAYS_SKIP like I_ALWAYS_NO (-n),
except that we don't fail when skipping.
* src/system.h (emit_update_parameters_note): A new function
to output the description of the new --update parameters.
* src/cp.c (main): Parse --update arguments, ensuring that
-n takes precedence if specified.
(usage): Describe the new option. Also allude that
-u is related in the -n description.
* src/mv.c: Accept the new --update parameters and
update usage() accordingly.
* doc/coreutils.texi (cp invocation): Describe the new --update
parameters. Also reference --update from the --no-clobber description.
(mv invocation): Likewise.
* tests/mv/update.sh: Test the new parameters.
* NEWS: Mention the new feature.
Addresses https://bugs.gnu.org/62572
* gnulib: Reference the latest gnulib including the
fix to the backupfile module in commit 94496522.
* tests/cp/backup-dir.sh: Add a test to ensure
we rename appropriately when backing up through subdirs.
* NEWS: Mention the bug fix.
Fixes https://bugs.gnu.org/62607
* tests/misc/tee.sh: Call cleanup_ in all cases to ensure
there are no overlapping interactions on the fifo that
might impact later parts of the test. This was seen to
cause issue with dash on musl libc.
Addresses https://bugs.gnu.org/62542
* tests/misc/csplit-heap.sh: More memory is required to avoid
a false failure on some systems. Noticed with musl libc
with bash as the shell. This is confirmed to still easily
trigger with the original memory leak being tested.
Addresses https://bugs.gnu.org/62542
* src/wc.c (wc): Use INT_ADD_WRAPV() to detect overflow.
(main): Upon overflow, saturate the total, print a diagnostic,
and set exit status.
* tests/misc/wc-total.sh: Add a test case, which operates
on BTRFS and 64 bit systems at least.
Reported at https://bugs.debian.org/1027100
* NEWS: Mention the fix.
* src/dircolors.c: Fail upon read error from getline().
* tests/misc/dircolors.sh: Add a new test.
* tests/local.mk: Reference the new test.
On restricted systems like android or some containers,
FICLONE could return EPERM, EACCES, or ENOTTY,
which would have induced the command to fail to copy
rather than falling back to a more standard copy.
* src/copy.c (is_terminal_failure): A new function refactored
from handle_clone_fail().
(is_CLONENOTSUP): Merge in the handling of EACCES, ENOTTY, EPERM
as they also pertain to determination of whether cloning is supported
if we ever use this function in that context.
(handle_clone_fail): Use is_terminal_failure() in all cases,
so that we assume a terminal failure in less errno cases.
* NEWS: Mention the bug fix.
Addresses https://bugs.gnu.org/62404
This applies to all checksumming utilities,
where we incorrectly report all subsequent files as checking 'OK'
once any file has passed a digest check.
The exit status was not impacted, only the printed status.
* src/digest.c (digest_check): Use the correct state variable
to determine if the _current_ file has passed or not.
* tests/misc/md5sum.pl: Add a test case.
Fixes https://bugs.gnu.org/62403
Similarly to the fix to tests/rmdir/ignore.sh in c0e5f8c59,
tee should not be expected to fail when run with read-only outputs
when run as root.
* tests/misc/tee.sh: Add uid_is_privileged_ guard around test for
read-only outputs.
* tests/ls/stat-free-symlinks.sh: Filter out syscalls that
return ENOSYS, as that was seen with statx() on Debian 10.13
on mips64, and resulted in overcounting of stat calls.
* src/stty.c (main): Use static structures to ensure
they're initialized (to zero), so that random data is
not displayed, or compared resulting in a inaccurate
failure reported to users. This was seen on musl libc
where some parts of the termios c_cc array were
not initialized by tcgetattr().
Reported by Bruno Haible.
* tests/misc/tee.sh: AIX doesn't support detecting
closed outputs either with poll() or select() so avoid
testing that functionality.
* tests/tail-2/pipe-f.sh: Likewise.
Since SELinux version 3.5, the return value of context_str(3) is
declared as const; see:
https://github.com/SELinuxProject/selinux/commit/dd98fa322766
Therefore, GCC complains (here with -Werror):
src/selinux.c: In function 'defaultcon':
src/selinux.c:152:16: error: assignment discards 'const' qualifier \
from pointer target type [-Werror=discarded-qualifiers]
152 | if (!(constr = context_str (tcontext)))
| ^
src/selinux.c: In function 'restorecon_private':
src/selinux.c:252:16: error: assignment discards 'const' qualifier \
from pointer target type [-Werror=discarded-qualifiers]
252 | if (!(constr = context_str (tcontext)))
| ^
* src/selinux.c (defaultcon): Define CONSTR as const.
(restorecon_private): Likewise.
* src/sum.c (output_bsd): On sparc64 for example,
a crc of 0 was output due to casting an int variable
to uint16_t and thus operating on the wrong end of the variable.
Instead use explicit assignment to the narrower type
to ensure we get the appropriate data.
(output_sysv): Likewise.
Reported by Bruno Haible.
* iopoll.c (fclose_wait): Rename from confusing fclose_nonblock name.
Also adjust to do no operations on the stream after fclose()
as this is undefined. Instead use fflush() to determine EAGAIN status.
(fwrite_wait): Renamed from confusing fwrite_nonblock name.
* src/dircolors.hin: Make the separate sections of the self
documenting dircolors database more apparent,
by adding heading comments, and appropriate separation.
Following on from commit v8.29-45-g24053fbd8 which unconditionally
used case insensitive extension matching, support selective
case sensitive matching when there are separate extension cases
defined with different display sequences.
* src/dircolors.hin: Document how file name suffixes are matched.
Note this is displayed with `dircolors --print-database` which
the texi info recommends to use for details.
* src/ls.c (parse_ls_color): Postprocess the list to
mark entries for case sensitive matching,
and also adjust so that unmatchable entries are more quickly ignored.
(get_color_indicator): Use exact matching rather than
case insensitive matching if so marked.
* tests/ls/color-ext.sh: Add test cases.
* NEWS: Mention the change in behavior.
Addresses https://bugs.gnu.org/33123
* tests/du/threshold.sh: Directories are assumed to be
of size 0 with --apparent since commit v9.1-187-g110bcd283
so remove --apparent cases from this test.
Non blocking outputs can be seen for example
when piping telnet through tee to a terminal.
In that case telnet sets its input to nonblocking mode,
which results in tee's output being nonblocking,
in which case in may receive an EAGAIN error upon write().
The same issue was seen with mpirun.
The following can be used to reproduce this
locally at a terminal (in most invocations):
$ { dd iflag=nonblock count=0 status=none;
dd bs=10K count=10 if=/dev/zero status=none; } |
tee || echo fail >/dev/tty
* src/iopoll.c (iopoll_internal): A new function refactored from
iopoll(), to also support a mode where we check the output
descriptor is writeable.
(iopoll): Now refactored to just call iopoll_internal().
(fwait_for_nonblocking_write): A new internal function which
uses iopoll_internal() to wait for writeable output
if an EAGAIN or EWOULDBLOCK was received.
(fwrite_nonblock): An fwrite() wrapper which uses
fwait_for_nonblocking_write() to handle EAGAIN.
(fclose_nonblock): Likewise.
src/iopoll.h: Add fclose_nonblock, fwrite_nonblock.
src/tee.c: Call fclose_nonblock() and fwrite_nonblock wrappers,
instead of the standard functions.
* tests/misc/tee.sh: Add a test case.
* NEWS: Mention the improvement.
The idea was suggested by Kamil Dudka in
https://bugzilla.redhat.com/1615467
* bootstrap.conf (gnulib_modules): Add free-posix, tmpfile.
* src/split.c (copy_to_tmpfile): New function.
(input_file_size): Use it to split larger files when sizes cannot
easily be determined via fstat or lseek. See Bug#61386#235.
* tests/split/l-chunk.sh: Mark tests of /dev/zero as
very expensive since they exhaust /tmp.
This was introduced recently with commit v9.1-166-g6b12e62d9
* src/tee.c (tee_files): Check the return from fopen()
before passing to fileno() etc.
* tests/misc/tee.sh: Add a test case.
Problem reported by Pádraig Brady (Bug#61386#226).
* src/split.c (parse_chunk): Use die instead of error.
(main): Quote a string.
* tests/local.mk (all_root_tests): Move du/apparent.sh from here ...
(all_tests): ... to here.
Problem reported by Christoph Anton Mitterer (Bug#61884).
* src/du.c (process_file): When counting apparent sizes, count
only usable st_size members.
* tests/du/apparent.sh: New file.
* tests/local.mk (all_root_tests): Add it.
* src/split.c (create): Avoid fstat + ftruncate in the usual case
where the output file does not already exist, by trying
to create it with O_EXCL first. This costs a failed open
in the unusual case where the output file already exists,
but that’s OK.
Prefer signed types to uintmax_t, as this allows for better
runtime checking with gcc -fsanitize=undefined.
Also, when an integer overflows just use the maximal value
when the code will do the right thing anyway.
* src/split.c (set_suffix_length, bytes_split, lines_split)
(line_bytes_split, lines_chunk_split, bytes_chunk_extract)
(lines_rr, parse_chunk, main):
Prefer a signed type (typically intmax_t) to uintmax_t.
(strtoint_die): New function.
(OVERFLOW_OK): New macro. Use it elsewhere, where we now allow
LONGINT_OVERFLOW because the code then does the right thing on all
practical platforms (they have int wide enough so that it cannot
be practically exhausted). We can do this now that we can safely
assume intmax_t has at least 64 bits.
(parse_n_units): New function.
(parse_chunk, main): Use it.
(main): Do not worry about integer overflow when the code
will do the right thing anyway with the extreme value.
Just use the extreme value.
* tests/split/fail.sh: Adjust to match new behavior.
* src/split.c (bytes_split, lines_chunk_split)
(bytes_chunk_extract, main): Prefer ssize_t to size_t when
representing the return value of ‘read’. Use a negative value
instead of SIZE_MAX to indicate a missing value.
* src/split.c: Include sys-limits.h, not safe-read.h.
(input_file_size, bytes_split, lines_split, line_bytes_split)
(lines_chunk_split, bytes_chunk_extract, lines_rr): Call read, not
safe_read, since safe_read no longer buys us anything.
(main): Reject outlandish buffer sizes right away,
rather than allocating huge buffers and never using them.
* src/split.c (closeout): There should be no need for a special
case for ECHILD, since we never wait for the same child twice.
Simplify with this in mind.
Ignore and default SIGPIPE, rather than blocking and unblocking it.
* src/split.c (default_SIGPIPE):
New static var, replacing oldblocked and newblocked.
(create): Use it.
(main): Set it.
* src/split.c (input_file_size): Do not bother with lseek if the
initial read probe reaches EOF, since the file size is known then.
This works better on macOS, which doesn’t allow lseek on /dev/null.
Do not special-case size-zero files, as the issue can occur
with any size file (though /proc files are the most common).
If the current position is past end of file, treat this as
size zero regardless of whether the file has a usable st_size.
Pass through lseek -1 return values rather than using ‘return -1’;
this makes the code a bit easier to analyze (and a bit faster).
Avoid undefined behavior if the size calculation overflows.
(lines_chunk_split): Do not bother with lseek if it would have
no effect if successful. This works better on macOS, which
doesn’t allow lseek on /dev/null.
* tests/split/l-chunk.sh: Adjust to match fixed behavior.
* src/split.c (bytes_split): New arg REM_BYTES.
Use this to split more evenly. All callers changed.
(lines_chunk_split, bytes_chunk_extract):
Be consistent with new byte_split.
* tests/split/b-chunk.sh, tests/split/l-chunk.sh: Test new behavior.
* src/split.c (lines_chunk_split): Simplify by having chunk_end
point to the first byte after the chunk, rather than to the last
byte of the chunk. This will reduce confusion once we allow
chunks to be empty.
* src/tee.c (pipe_check): Make this a local var instead
of a static var. This suppresses a -Wmaybe-uninitialized
diagnostic with gcc 12.2.1 20221121 (Red Hat 12.2.1-4).
(main): Don’t set pipe_check unnecessarily if a later
-p option overrides an earlier one that wants pipe_check.
Problem discovered when I investigated the GCC warning.
* src/tail.c (check_output_alive): Reuse iopoll()
rather than directly calling poll() or select().
* src/iopoll.c (iopoll): Refactor to support non blocking operation,
or ignoring descriptors by passing a negative value.
* src/iopoll.h (iopoll): Adjust to support a BLOCK parameter.
* src/tee.c (tee_files): Adjust iopoll() call to explicitly block.
* src/local.mk: Have tail depend on iopoll.c.
* src/tee.c (usage): Change from describing one (non pipe) aspect
to the more general point of being the option to use if working with
pipes, and referencing the more detailed info below.
* doc/coreutils.texi (tee invocation): s/standard/appropriate/ since
the standard operation with pipes is to exit immediately upon write
error. s/early/immediately/ as it's ambiguous as to what "early"
is in relation to.
If input is intermittent (a tty, pipe, or socket), and all remaining
outputs are pipes (eg, >(cmd) process substitutions), exit early when
they have all become broken pipes (and thus future writes will fail),
without waiting for more input to become available, as future write
attempts to these outputs will fail (SIGPIPE/EPIPE).
Only provide this enhancement when pipe errors are ignored (-p mode).
Note that only one output needs to be monitored at a time with iopoll(),
as we only want to exit early if _all_ outputs have been removed.
* src/tee.c (pipe_check): New global for iopoll mode.
(main): enable pipe_check for -p, as long as output_error ignores EPIPE,
and input is suitable for iopoll().
(get_next_out): Helper function for finding next valid output.
(fail_output, tee_files): Break out write failure/output removal logic
to helper function.
(tee_files): Add out_pollable array to track which outputs are suitable
for iopoll() (ie, that are pipes); track first output index that is
still valid; add iopoll() broken pipe detection before calling read(),
removing an output that becomes a broken pipe.
* src/local.mk (src_tee_SOURCES): include src/iopoll.c.
* NEWS: Mention tee -p enhancement in Improvements.
* doc/coreutils.texi: Mention the new early exit behavior in the nopipe
modes for the tee -p option.
Suggested-by: Arsen Arsenović <arsen@aarsen.me>
When a program's output becomes a broken pipe, future attempts to write
to that ouput will fail (SIGPIPE/EPIPE). Once it is known that all
future write attepts will fail (due to broken pipes), in many cases it
becomes pointless to wait for further input for slow devices like ttys.
Ideally, a program could use this information to exit early once it is
known that future writes will fail.
Introduce iopoll() to wait on a pair of fds (input & output) for input
to become ready or output to become a broken pipe.
This is relevant when input is intermittent (a tty, pipe, or socket);
but if input is always ready (a regular file or block device), then
a read() will not block, and write failures for a broken pipe will
happen normally.
Introduce iopoll_input_ok() to check whether an input fd is relevant
for iopoll().
Experimentally, broken pipes are only detectable immediately for pipes,
but not sockets. Errors for other file types will be detected in the
usual way, on write failure.
Introduce iopoll_output_ok() to check whether an output fd is suitable
for iopoll() -- namely, whether it is a pipe.
iopoll() is best implemented with a native poll(2) where possible, but
fall back to a select(2)-based implementation platforms where there are
portability issues. See also discussion in tail.c.
In general, adding a call to iopoll() before a read() in filter programs
also allows broken pipes to "propagate" backwards in a shell pipeline.
* src/iopoll.c, src/iopoll.h (iopoll): New function implementing broken
pipe detection on output while waiting for input.
(IOPOLL_BROKEN_OUTPUT, IOPOLL_ERROR): Return codes for iopoll().
(IOPOLL_USES_POLL): Macro for poll() vs select() implementation.
(iopoll_input_ok): New function to check whether an input fd is relevant
for iopoll().
(iopoll_output_ok): New function to check whether an input fd is
suitable for iopoll().
* src/local.mk (noinst_HEADERS): add src/iopoll.h.
* NEWS: Mention the fts fix to avoid the following assert
in rm on mem pressure:
Program terminated with signal SIGSEGV, Segmentation fault.
at ../lib/cycle-check.c:60
assure (state->magic == CC_MAGIC);
* gnulib: Update to the latest to pick up fts commit f17d3977.
* tests/rm/empty-inacc.sh: Ensure we're not reading from stdin
when we're relying on no prompt to proceed. Also change the
file being tested so that a failure in one test doesn't impact
following tests causing a framework failure.
gdb was seen to hang intermittently on macOS 12.
Also gdb requires signing on newer macOS systems:
https://sourceware.org/gdb/wiki/PermissionsDarwin
So restrict its use on macOS systems for now.
* tests/rm/r-root.sh: Skip on darwin systems.
* tests/tail-2/inotify-race.sh: Restrict the test to
inotify capable systems to avoid the hang with some gdbs.
* tests/tail-2/inotify-race.sh: Likewise.
Upcomming gnulib changes may disable SEEK_HOLE
even if the system supports it, so dynamically
check if we've SEEK_HOLE enabled.
* init.cfg (seek_data_capable_): SEEK_DATA may be disabled in the build
if the system support is deemed insufficient, so also use `cp --debug`
to determine if it's enabled.
* tests/cp/sparse-2.sh: Adjust to a more general diagnostic.
* tests/cp/sparse-extents-2.sh: Likewise.
* tests/cp/sparse-extents.sh: Likewise.
* tests/cp/sparse-perf.sh: Likewise.
How a file is copied is dependent on the sparseness of the file,
what file system it is on, what file system the destination is on,
the attributes of the file, and whether they're being copied or not.
Also the --reflink and --sparse options directly impact the operation.
Given it's hard to reason about the combination of all of the above,
the --debug option is useful for users to directly identify if
copy offloading, reflinking, or sparse detection are being used.
It will also be useful for tests to directly query if
these operations are supported.
The new output looks as follows:
$ src/cp --debug src/cp file.sparse
'src/cp' -> 'file.sparse'
copy offload: yes, reflink: unsupported, sparse detection: no
$ truncate -s+1M file.sparse
$ src/cp --debug file.sparse file.sparse.cp
'file.sparse' -> 'file.sparse.cp'
copy offload: yes, reflink: unsupported, sparse detection: SEEK_HOLE
$ src/cp --reflink=never --debug file.sparse file.sparse.cp
'file.sparse' -> 'file.sparse.cp'
copy offload: avoided, reflink: no, sparse detection: SEEK_HOLE
* doc/coreutils.texi (cp invocation): Describe the --debug option.
(mv invocation): Likewise.
(install invocation): Likewise.
* src/copy.h: Add a new DEBUG member to cp_options, to control
whether to output debug info or not.
* src/copy.c (copy_debug): A new global structure to
unconditionally store debug into from the last copy_reg operations.
(copy_debug_string, emit_debug): New functions to print debug info.
* src/cp.c: if ("--debug") x->debug=true;
* src/install.c: Likewise.
* src/mv.c: Likewise.
* tests/cp/debug.sh: Add a new test.
* tests/local.mk: Reference the new test.
* NEWS: Mention the new feature.
* src/remove.c (prompt, rm_fts): In the dir-handling code of both of
these functions, relax a "get_dir_status (...) == DS_EMPTY" condition
to instead test only "get_dir_status (...) != 0", enabling flow control
to reach the prompt function also for unreadable directories. However,
that function itself also needed special handling for this case:
(prompt): Handle empty, inaccessible directories properly,
deleting them with -d (--dir), and prompting about whether to delete
with -i (--interactive).
* tests/rm/empty-inacc.sh: Add tests for the new code.
Reported by наб <nabijaczleweli@nabijaczleweli.xyz> in
bugs.debian.org/1015273
* NEWS (Bug fixes): Mention this.
* tests/chmod/setgid.sh: Try all the groups you’re a member of,
in case id -g returns 4294967295 (nogroup) which is special
and does not let you chgrp a file to it.
* init.cfg (groups): Port better to macOS 12, where
group 4294967295 (nogroup) is special: you can be a member
without being able to chgrp files to the group.
* src/copy.c: Some changes if HAVE_FCLONEFILEAT && !USE_XATTR.
(fd_has_acl): New function.
(CLONE_ACL): Default to 0.
(copy_reg): Use CLONE_NOFOLLOW to avoid races like CVE-2021-30995
<https://www.trendmicro.com/en_us/research/22/a/
analyzing-an-old-bug-and-discovering-cve-2021-30995-.html>.
Use CLONE_ACL if available and working, falling back to cloning
without it if it fails due to EINVAL.
If the only problem with fclonefileat is that it would create the
file with the wrong timestamp, or with too few permissions,
do that but fix the timestamp and permissions afterwards,
rather than falling back on a traditional copy.
* src/copy.c (infer_scantype): Do not set *SCAN_INFERENCE
when returning a value other than LSEEK_SCANTYPE.
This is just minor refactoring; it simplifies the code a bit.
Callers are uneffected.
doc: document --preserve=mode better
* src/tail (tail_forever): Attempt to read() from non blocking
single non regular file, which shouldn't block, but also
read data even when the mtime doesn't change.
* NEWS: Mention the improvement.
* THANKS.in: Thanks for detailed testing.
This was seen to be an issue when following a
symlink that was being updated to point to
different underlying devices.
* src/tail.c (recheck): Guard the lseek() call to only
be performed for regular files.
* NEWS: Mention the bug fix.
--raw output is the most composable format, and also is a
robust way to discard the file name without parsing (escaped) output.
Examples:
$ cksum --raw -a crc "$afile" | basenc --base16
4ACFC4F0
$ cksum --raw -a crc "$afile" | basenc --base2msbf
01001010110011111100010011110000
$ cksum --raw -a sha256 "$bfile" | basenc --base32
AAAAAAAADHLGRHAILLQWLAY6SNH7OY5OI2RKNQLSWPY3MCUM4JXQ====
* doc/coreutils.texi (cksum invocation): Describe the new feature.
* src/digest.c (output_file): Inspect the new RAW_DIGEST global,
and output the bytes directly if set.
* src/cksum.c (output_crc): Likewise.
* src/sum.c (output_bsd, output_sysv): Likewise.
* tests/misc/cksum-raw.sh: A new test.
* tests/local.mk: Reference the new test.
* NEWS: Mention the new feature.
* src/uptime.c (print_uptime): Following gnulib commit 9041103
HAVE_UTMP_H will always be defined. Therefore key on whether
the utmp.ut_type member is present.
* boottime.m4 (GNULIB_BOOT_TIME): Assume utmp.h is present.
* src/digest.c [HASH_ALGO_CKSUM]: Include "base64.h"
[HASH_ALGO_CKSUM] (base64_digest): New global.
[HASH_ALGO_CKSUM] (enum BASE64_DIGEST_OPTION): New enum.
[HASH_ALGO_CKSUM] (long_options): Add "base64".
(valid_digits): Rename from hex_digits, now taking an input length argument.
Adjust callers.
(bsd_split_3): Rename arg from hex_digits to digest.
Add new *d_len parameter for length of extracted digest.
Move "i" declaration down to first use.
(split_3): Rename arg from hex_digits to digest.
Add new *d_len parameter for length of extracted digest.
Instead of relying on "known" length of digest to find the following
must-be-whitespace byte, search for the first whitespace byte.
[HASH_ALGO_CKSUM] (output_file): Handle base64_digest.
[HASH_ALGO_CKSUM] (main): Set base64_digest.
[HASH_ALGO_CKSUM] (b64_equal): New function.
(hex_equal): New function, factored out of digest_check.
(digest_check) Factored part into b64_equal and hex_equal.
Rename local hex_digest to digest.
* tests/misc/cksum-base64.pl: Add tests.
* tests/local.mk (all_tests): Add to the list.
* cfg.mk (_cksum): Define.
(exclude_file_name_regexp--sc_prohibit_test_backticks): Exempt new test.
(exclude_file_name_regexp--sc_long_lines): Likewise.
* doc/coreutils.texi (cksum invocation): Document it.
(md5sum invocation) [--check]: Mention digest encoding auto-detect.
* NEWS (New Features): Mention this.
This reverts the previous change, so that when a file
is skipped due to -u, this is not considered a failure.
* doc/coreutils.texi: Document this.
* src/copy.c (copy_internal): If --update says to skip,
treat this as success instead of failure.
* tests/mv/update.sh, tests/cp/slink-2-slink.sh:
Revert previous change, to match reverted behavior.
* NEWS, doc/coreutils.texi: Document this.
* src/copy.c (copy_internal):
* src/ln.c (do_link): Return false when skipping action due to
--interactive or --no-clobber.
* tests/cp/cp-i.sh, tests/cp/preserve-link.sh:
* tests/cp/slink-2-slink.sh, tests/mv/i-1.pl, tests/mv/i-5.sh:
* tests/mv/mv-n.sh, tests/mv/update.sh:
Adjust expectations of exit status to match revised behavior.
* src/digest.c (digest_check): Locals n_misformatted_lines and
n_improperly_formatted_lines were declared and set/incremented
identically. Remove declaration of the latter. Use the other instead.
Wishlist item from Mike Frysinger (Bug#61050).
* src/copy.c (copy_internal):
Do not fall back on copying if x->no_copy.
* src/copy.h (struct cp_options): New member no_copy.
* src/mv.c (NO_COPY_OPTION): New constant.
(long_options, usage, main): Support --no-copy.
* tests/mv/no-copy.sh: New test.
* tests/local.mk (all_tests): Add it.
* src/csplit.c (usage): Use "suppress" rather than "remove"
when describing -z so it's more apparent that the effect
is a particular numbered file is not created, rather than
being removed later. I.e., don't suggest -z may induce
gaps in file numbering.
Reported at https://bugs.debian.org/1029103
* src/copy.c (sparse_copy): Fallback to standard copy upon ENOENT,
which was seen intermittently across CIFS file systems.
* NEWS: Mention the bug fix, though qualify it as an "issue"
rather than a bug, as coreutils is likely only highlighting
a CIFS bug in this case.
Fixes https://bugs.gnu.org/60455
* src/copy.c (handle_clone_fail): A new function refactored
from copy_reg() to handle failures from FICLONE or fclonefileat().
Fail with all errors from FICLONE, unless they're from the set
indicating the file system or file do not support the clone operation.
Also fail with errors from fclonefileat() (dest_dest < 0)
if they're from the set indicating a transient failure for the file.
(copy_ref): Call handle_clone_fail() after fclonefileat() and FICLONE.
(sparse_copy): Call the refactored is_CLONENOTSUP()
which is now also used by the new handle_clone_fail() function.
* NEWS: Mention the bug fix. Also mention explicitly
the older --reflink=auto default change to aid searching.
* cfg.mk: Adjust old_NEWS_hash with `make update-NEWS-hash`.
Fixes https://bugs.gnu.org/60489
* src/dd.c (parse_integer): Support Q,R suffixes.
* src/od.c (main): Likewise.
* src/split.c (main): Likewise.
* src/stdbuf.c (parse_size): Likewise.
* src/truncate.c (main): Likewise.
* src/sort.c (specify_size_size): Likewise.
Also line length syntax check fix.
* tests/misc/numfmt.pl: Adust top end large number checks
to the new largest values.
* doc/coreutils.texi (numfmt invocation): Add a numfmt example.
* NEWS: Tweak to aid searchability.
* src/dd, src/head.c, src/od.c, src/sort.c, src/stdbuf.c, src/tail.c:
(usage):
* src/system.h (emit_size_note):
Mention new SI prefixes.
* src/du.c (main):
* src/head.c (head_file):
* src/numfmt.c (suffix_power, suffix_power_char, prepare_padded_number):
* src/shred.c (main):
* src/sort.c (unit_order):
* src/tail.c (parse_options):
Support new SI prefixes.
* src/numfmt.c (MAX_ACCEPTABLE_DIGITS): Increase to 33.
(zero_and_valid_suffixes, valid_suffixes): New constants,
with new SI prefixes.
(valid_suffix, unit_to_umax): Use them.
(prepare_padded_number): Diagnose "999Q" instead of "999Y".
* tests/misc/numfmt.pl, tests/misc/sort.pl:
Adjust tests to match new max.
Newer grep(1) complains:
$ make sc_prohibit_test_minus_ao
/usr/bin/grep: warning: * at start of expression
prohibit_test_minus_ao
* cfg.mk (exclude_file_name_regexp--sc_prohibit_test_minus_ao): Fix
expression inroduced in v8.24-120-g3205bb178, and narrow down the file
pattern to the 'doc/' directory.
* bootstrap.conf (gnulib_modules): Add count-leading-zeros,
which was already an indirect dependency, since ioblksize.h
now uses it directly.
* src/ioblksize.h: Include count-leading-zeros.h.
(io_blksize): Treat impossible blocksizes as IO_BUFSIZE.
When growing a blocksize to IO_BUFSIZE, keep it a multiple of the
stated blocksize. Work around the ZFS performance bug.
* NEWS: Mention the bug fix.
Problem reported by Korn Andras at https://bugs.gnu.org/59382
Update to latest gnulib with new copyright year.
Run "make update-copyright" and then...
* tests/init.sh: Sync with gnulib to pick up copyright year.
* bootstrap: Manually update copyright year,
until we fully sync with gnulib at a later stage.
* tests/sample-test: Adjust to use the single most recent year.
* src/stty.c (wrapf): Adjust the comparison by 1,
to account for the space we're adding.
* tests/misc/stty.sh: Add a test case.
* NEWS: Mention the fix.
Reported in https://bugs.debian.org/1027442
This was seen to vastly improve performance
on NFS 4.2 systems by allowing server side copies,
with partially sparse files (avidemux generated mp4 files).
* src/copy.c (lseek_copy): Also set hole_size to 0,
i.e. enable copy_file_range(), with --sparse=auto (the default),
to enable copy offload in this case, as we've strong signal
from SEEK_DATA that we're operating on actual data and not holes here.
* NEWS: Mention the improvement.
Fixes https://bugs.gnu.org/60416
* src/wc.c (wc): Use off_t rather than size_t
when calculating where to seek to, so that
we don't seek to a too low offset on systems
where size_t < off_t, which would result in
many read() calls to determine the file size.
* tests/misc/wc-proc.sh: Add a test case
sufficient for 32 bit systems at least.
* NEWS: Mention the bug fix.
Reported at https://bugs.debian.org/1027101
* tests/cp/proc-short-read.sh: Kernel on ARMv7 Processor rev 3 (v7l)
spells it "BogoMIPS", so allow any capitalization. Patch from
Zach van Rijn in <https://bugs.gnu.org/60339>.
* doc/coreutils.texi, doc/sort-version.texi: Prefer on "x -- y" to
"x---y" in prose, as the result is more readable in Emacs.
Fix some instances of unescaped ‘-’ that should be minus, not
hyphen. Fix some other instances that should be en dash. No
spaces around en dash when it’s a range.
* cfg.mk (sc_texi_long_option_escaped): A new check to
avoid future instances of this.
* doc/coreutils.texi (Common options): Rearrange this menu
to be less repetitive in each description, and avoid long lines.
Addresses https://bugs.gnu.org/59262
Problem reported by Antonio Diaz Diaz (bug#59262).
* doc/coreutils.texi: Use markup in menus to prevent
‘--’ from turning into an em dash, and to be more
consistent.
Note using iconv(1) rather than recode(1) is not appropriate
for this example, as the required functionality is only
available on libiconv's iconv implementation, which is
not installed on most systems.
* doc/coreutils.texi (printf invocation): Use env rather than
/usr/local/bin for the printf command. Escape '%' so more robust.
Also use a locale that exists on modern systems.
Previously this was restricted to the C99 universal character subset,
which restricted most values <= 0x9F, as that simplifies the C lexer.
However printf(1) doesn't need this restriction.
Note also the bash builtin printf already supports all values <= 0x9F.
* src/printf.c (main): Relax the restriction on points <= 0x9F.
* doc/coreutils.texi (printf invocation): Adjust description.
* tests/misc/printf-cov.pl: Adjust accordingly. Add new cases.
* NEWS: Mention the change in behavior.
Reported at https://bugs.debian.org/1022857
* doc/coreutils.texi (sort invocation): Mention in the
multi invocation sort example that the -V GNU extension
could be used to sort IPv4 addresses, and thus simplify
to a single invocation.
* src/system.h (emit_exec_status): A new function to
output standard "Exit status:" info for commands that exec others.
* doc/coreutils.texi (Exit status): Add "ls" and "runcon"
to the list of commands with non standard exit status.
* src/numfmt.c (main): Call initialize_exit_failure() explicitly
to better indicate this utility may exit with something other than
EXIT_FAILURE.
* src/timeout.c (usage): Use more consistent capitalization.
* src/chroot.c: Call emit_exec_status().
* src/env.c: Likewise.
* src/nice.c: Likewise.
* src/nohup.c: Likewise.
* src/runcon.c: Likewise.
* src/stdbuf.c: Likewise.
* src/runcon.c (main): Call initialize_exit_failure(),
so we use an appropriate exit status upon failure to close stdout.
This should have been part of recent commit ea3ee6df.
* tests/misc/help-version.sh: Adjust test case accordingly.
* src/getlimits.c: Don't call initialize_exit_failure()
as it's not needed for standard EXIT_FAILURE returns.
Also use the function variant that diagnoses invalid options.
without this option, control of when the total is output
is quite awkward. Consider trying to suppress the total line,
which could be achieved with something like:
wc-no-total() { wc "$@" /dev/null | head -n-2; }
As well as being non obvious, it's also non general.
It would give a non failure, but zero count if passed a file on stdin.
Also it doesn't work in conjunction with the --files0-from option,
which would need to be handled differently with something like:
{ find files -print0; printf '%s\0' /dev/null; } |
wc --files0-from=- |
head -n2
Also getting just the total can be awkward as file names
are only suppressed when processing stdin, and
also a total line is only printed if processing more than one file.
For completness this might be achieved currently with:
wc-only-total() {
wc "$@" |
tail -n1 |
sed 's/^ *//; s/ [^ 0-9]*$//'
}
* src/wc.c: Add new --total option.
* tests/misc/wc-total.sh: New test suite for the new option.
* tests/local.mk: Reference the new test.
* doc/coreutils.texi (wc invocation): Document the new option.
* THANKS.in: Add suggestor.
* NEWS: Mention the new feature.
* .gitignore: Add new headers from gnulib.
* src/basenc.c: Adjust line length due to replacement
of 'verify' with 'static_assert'.
* src/od.c: Likewise.
I ran into this problem when attempting to recursively
remove a directory in a filesystem on flaky hardware.
Although the underlying readdir syscall failed with errno == EIO,
rm issued no diagnostic about the I/O error.
Without this patch I see this behavior:
$ rm -fr baddir
rm: cannot remove 'baddir': Directory not empty
$ rm -ir baddir
rm: descend into directory 'baddir'? y
rm: remove directory 'baddir'? y
rm: cannot remove 'baddir': Directory not empty
With this patch I see the following behavior, which
lets the user know about the I/O error when rm tries
to read baddir's directory entries:
$ rm -fr baddir
rm: cannot remove 'baddir': Input/output error
$ rm -ir baddir
rm: cannot remove 'baddir': Input/output error
* src/remove.c (Ternary): Remove. All uses removed.
(get_dir_status): New static function.
(prompt): Last arg is now directory status, not ternary.
Return RM_USER_ACCEPTED if user explicitly accepted.
All uses changed.
Report any significant error in directory status right away.
(prompt, rm_fts): Use get_dir_status to get directory status lazily.
(excise): Treat any FTS_DNR errno as being more descriptive, not
just EPERM and EACCESS. For example, EIO is more descriptive.
(rm_fts): Distinguish more clearly between explicit and implied
user OK.
* src/remove.h (RM_USER_ACCEPTED): New constant.
(VALID_STATUS): Treat it as valid.
* src/system.h (is_empty_dir): Remove, replacing with ...
(directory_status): ... this more-general function.
All uses changed. Avoid undefined behavior of looking at
a non-null readdir pointer after corresponding closedir.
* tests/rm/rm-readdir-fail.sh: Adjust test of internals
to match current behavior.
* bootstrap.conf (gnulib_modules): Remove alignof, which isn’t
needed since coreutils source modules don’t include alignof.h.
Add stdalign, since they depend on alignof working without
stdalign.h.
* gl/lib/fadvise.h, gl/lib/smack.h, src/blake2/blake2-impl.h:
Do not include config.h from a .h file. config.h is supposed
to be included once, at the start of compilation and before
any other file.
* src/stty.c (check_speed): If difference input and output speeds
are specified, then validate the system supports that, before
interacting with the device.
* src/stty.c (eq_mode): A new function to compare
equivalence of two modes.
(main): Use eq_mode() rather than memcmp() to compare
two modes. Also use stack variables rather than implicitly
initialized static variables. Also remove all uses of
the SPEED_WAS_SET hack since we now more robustly compare modes.
* NEWS: Update the [io]speed fix entry.
Reported at https://bugs.debian.org/1019468
* src/stty.c (main): Move internal TESTING code that showed
the new and old mode, upon failure to apply the new mode,
to being runtime controlled with the ---debug option.
Also augment the display to show which items were not
set as expected.
* doc/coreutils.texi (stty invocation): Say that "drain"
is treated as an option, rather than a line setting,
and so option processing rules apply to it.
Reported in https://bugs.debian.org/1018803
* src/stty.c (apply_settings): Validate [io]speed arguments
against the internal accepted set.
(set_speed): Check the cfset[io]speed() return value so
that we validate against the system supported set.
* tests/misc/stty-invalid.sh: Add a test case.
* NEWS: Mention the bug fix.
Reported in https://bugs.debian.org/1018790
* src/tail.c (check_output_alive): Add a guard that would
trigger on most platforms, to detect if we're using the
gnulib poll module. That's currently problematic in the
way it emulates poll() using select() and would cause
issues on macOS and AIX at least as poll() is replaced there.
* src/comm.c (compare_files): Handle the single character
--output-delimeter case separately so that NUL is appropriately
handled.
* doc/coreutils.texi (comm invocation): Fix the description
of --output-delimiter to say an empty delimeter is treated
as a NUL separator, rather than being disallowed.
* tests/misc/comm.pl: Add a test case.
Reported at https://bugs.debian.org/1014008
* src/runcon.c: Use EXIT_CANCELED (125) instead of EXIT_FAILURE (1),
so that errors specific to runcon can be distinguished,
from those of the invoked program.
* doc/coreutils.texi (runcon invocation): Fix the Exit status
description to say we return 125 (not 127) for internal errors.
* tests/misc/runcon-no-reorder.sh: Add a test case.
The README was becoming too long and contained
quite a bit of info only pertaining to rarely used systems, so...
* README: Split out install specific info to README-install.
Also remove a few stale lines, and reorder a few items.
* README-install: A new file split from README.
* Makefile.am [EXTRA_DIST]: Explicitly reference new README-install
file for distribution, since automake only auto adds README.
* TODO: Reference the HPUX info now in README-install.
* src/ls.c (usage): Don't mention "modification" in the
description of ctime (-c), as it's confusing with mtime.
Mention "metadata" when discussing "change" time to
disambiguate from data change time.
* doc/coreutils.texi (ls invocation): State that --time=creation
falls back to using mtime where not available.
This behaviour is correctly documented when doing `cp --help`.
There is no `--reflink=when` option.
* doc/coreutils.texi (cp invocation): Fix document stating
that `--reflink` is equivalent to `--reflink=always`.
It's useful to treat empty and missing arguments differently.
Missing means all signals, while empty means no signals and
so is a no-op. It's useful to treat empty arguments like
this, so that dynamically specified arguments like the following
are supported
env --ignore-signals "$SIGS_TO_IGNORE"
Note `env --ignore-signals=` is treated as an empty argument.
* doc/coreutils.texi (env invocation): Empty args are treated
differently to missing arguments, so call that out explicitly.
* src/env.c (usage): Likewise.
Addresses https://bugs.debian.org/1016049
* src/date.c (usage): Specify that --date, --file, --reference,
and --resolution are mutually exclusive. This is also useful
documentation to group similar options.
* doc/coreutils.texi (Options for date): Likewise.
Addresses https://bugs.gnu.org/55401
* src/date.c: (main): Track and diagnose whether any
-d or -s options are dropped, as users may think
multiple options are supported, given they can be relative.
* tests/misc/date-debug.sh: Add a test case.
* NEWS: Mention the improvement.
* src/runcon.c (main): With -c avoid searching the path
to ensure the file specified to --compute is executed.
* tests/misc/runcon-compute.sh: Add a new test.
* tests/local.mk: Reference the new test.
* NEWS: Mention the bug fix.
Reported in https://bugs.debian.org/1013924
* src/remove.c: Include stat-time.h.
(cache_fstatat, cache_stat_init): Use negative st->st_atim.tv_sec to
determine whether the stat is cached, not negative st->st_size.
On non-POSIX platforms that lack st_atim.tv_sec, don’t bother to cache.
This follows up on comments by Pádraig Brady (bug#56391).
* src/copy.c (copy_reg): When --reflink=always removes a file
due to an FICLONE failure, do not remove a nonempty file.
* src/shuf.c: Do not include xdectoint.h.
(main): Improve diagnostic for ‘shuf -i -10-10’. Without this
patch, the diagnostic was “shuf: invalid input range: ‘’” which is
not helpful. Now it is “shuf: invalid input range: ‘-10-10’”.
* cfg.mk (begword, endword): New macros.
(sc_prohibit_stat_macro_address, sc_prohibit_fail_0)
(sc_prohibit_short_facl_mode_spec, sc_require_stdio_safer)
(sc_prohibit_sleep, sc_prohibit_framework_failure)
(sc_marked_devdiagnostics):
* build-aux/gen-single-binary.sh:
Prefer POSIX-compatible EREs to GNU extensions like \w and \<.
Problem reported by pkoraou@gmail.com (Bug#55910).
* src/copy.c (copy_internal): Treat a relative destination name ""
as if it were "." for the purpose of directory-relative syscalls
like fstatat that might might refer to the destination directory.
* src/sort.c (keycompare, compare): Don’t overflow if -r is
specified and a comparison function returns INT_MIN, as this
causes the comparison to have undefined behavior (typically the
reverse of correct). glibc memcmp on s390x reportedly returns
INT_MIN in some cases, so this is not a purely academic issue.
* src/comm.c (compare_files):
* src/join.c (keycmp):
* src/ls.c (off_cmp):
* src/ptx.c (compare_words, compare_occurs):
* src/set-fields.c (compare_ranges):
Prefer ((a > b) - (a < b)) to variants like (a < b ? -1 : a > b)
as it’s typically faster these days.
* src/sort.c (keycompare): Rework to avoid gotos.
This also shrinks the machine code a bit (112 bytes)
with GCC 12 x86-64 -O2. Nowadays compilers are smart
enough to coalesce jumps so we need not do it by hand.
* src/sort.c (keycompare): Rework to pacify a GCC 12
-Wmaybe-uninitialized false positive, by coalescing some minor
duplicate code and eliminating a branch. This should execute an
insn or two less in the usual case.
When factoring numbers that have a large 2^n factor, it can be hard to
eyeball just how many 2's there are. Add an option to print each prime
power factor in the p^e format (omitting the exponent when it is 1).
* src/factor.c: Add -h, --exponents option for printing in p^e format.
* doc/coreutils.texi (factor invocation): Document the new option.
* tests/misc/factor.pl: Add test case.
* THANKS.in: Add previous suggester
(https://lists.gnu.org/r/coreutils/2017-11/msg00015.html).
Suggested-by: Emanuel Landeholm <emanuel.landeholm@gmail.com>
Signed-off-by: Rasmus Villemoes <rasmus.villemoes@prevas.dk>
Problem reported by Giulio Genovese (Bug#55212).
* src/sort.c (nan_compare): To compare NaNs, simply printf+strcmp.
This avoids the problem of padding bits and unspecified behavior.
Args are now long double instead of char *; caller changed.
Found with -flto and --enable-gcc-warnings.
* src/pr.c (getoptarg): Fix misuse of xstrtol, which does not
necessarily set tmp_long on errror, and does not set errno in any
reliable way. The previous code might access uninitialized
storage; on typical platforms this merely causes it to possibly
print the wrong diagnostic.
* doc/coreutils.texi: Set txicodequoteundirected and
txicodequotebacktick so that ' and ` in code examples appear
as-is, rather than being transliterated to ’ and ‘. E.g., prefer
“... this is equivalent to ‘tr '\303\266' '\305\201'’ and ...” to
“... this is equivalent to ‘tr ’\303\266’ ’\305\201’’ and ...”
in PDF output.
* doc/coreutils.texi (Character arrays): Avoid using shell
notation like $'\u7530' since this isn’t in POSIX yet. Instead,
use ö and Ł which should work in all texinfo output formats.
This option has changed from ignoring only ENOTEMPTY|EEXIST
(i.e. ignore errors _solely_ due to dir not empty),
to ignoring some other errors from more protected dirs
that are not empty. That adjustment was made to better
support use with --parents, to essentially remove as much of
a hierarchy as possible, without erroring as we hit more
protected non empty parent dirs.
That functionality adjustment was originally discussed at:
https://lists.gnu.org/r/bug-coreutils/2008-01/msg00283.html
* src/rmdir.c (usage): Adjust to be more accurate to current behavior.
Also adjust --parents option to be easier to read.
* doc/coreutils.texi (rmdir invocation): Likewise.
Reported at https://github.com/coreutils/coreutils/issues/40
Avoid "Unicode character U+#1 not supported, sorry" error
when converting from texi to dvi or pdf.
* doc/coreutils.texi (tr invocation): Avoid the @U{XXXX}
texi representation, as even though info and html can represent
these characters directly, there are conversion errors
for pdf and dvi. Instead use the more abstract shell
$'\uXXXX' representation.
* src/tail.c (check_output_alive): Use poll() on Solaris.
Also handle POLLHUP, which Solaris returns in this case.
* tests/tail-2/pipe-f.sh: Use `head -n2` rather than `sed 2q`
as Solaris sed does not exit in this case.
* NEWS: Mention the improvement.
Reported by Bruno Haible.
This improves on the fix for --target-directory diagnostics bugs on
Solaris 11. Problem reported by Bruno Haible and Pádraig Brady; see:
https://lists.gnu.org/r/coreutils/2022-04/msg00044.html
Also, omit some unnecessary stat calls.
* gl/lib/targetdir.c (target_directory_operand): If !O_DIRECTORY,
do not bother calling open if stat failed with errno != EOVERFLOW.
Rename is_a_dir to try_to_open since that’s closer to what it means.
If the open failed with EACCES and we used O_SEARCH, look at stat
results to see whether errno should be ENOTDIR for better diagnostics.
Treat EOVERFLOW as an “I don’t know whether it’s a directory and
there’s no easy way to find out” rather than as an error.
* gl/lib/targetdir.c (target_directory_operand): New arg ST.
All callers changed.
* src/cp.c (do_copy):
* src/mv.c (main):
Avoid unnecessary stat call if target_directory_operand already
got the status.
Move target directory code out of system.h to a new targetdir module.
This doesn’t change functionality.
* bootstrap.conf (gnulib_modules): Add targetdir.
* src/cp.c, src/install.c, src/mv.c: Include targetdir.h.
* src/system.h (must_be_working_directory, target_directory_operand)
(targetdir_dirfd_valid): Move from here ...
* gl/lib/targetdir.c, gl/lib/targetdir.h, gl/modules/targetdir:
... to these new files.
* src/system.h (target_directory_operand): Also check with stat()
on systems with O_SEARCH, to avoid open("file", O_SEARCH|O_DIRECTORY)
returning EACCES rather than ENOTDIR, which was seen on Solaris 11.4
when operating on non dirs without execute bit set.
* NEWS: Remove related bug entry, as that issue was only introduced
after coreutils v9.0 was released.
Reported by Bruno Haible.
* tests/misc/env.sh: Verify with another command that
execvp() doesn not return ENOENT, before testing the
exit code from the command in question.
* tests/misc/nice-fail.sh: Likewise.
* tests/misc/stdbuf.sh: Likewise.
* tests/misc/timeout-parameters.sh: Likewise.
* tests/misc/md5sum-newline.pl: Avoid binary '*' tags when
comparing checksums.
* tests/misc/md5sum-bsd.sh: Avoid binary '*' tags so that we correctly
trigger the ambiguity test.
Reported by Bruno Haible
* tests/Coreutils.pm: Ensure an unset $TERM env var,
which is required on perl 5.22.2 on Solaris 11 OpenIndiana at least,
where TERM was being reset to 'dumb'.
Reported By Bruno Haible.
commit v9.0-66-ge2daa8f79 introduced an issue, for example
where cp could hang when overwriting a destination fifo,
when it would try to open() the fifo on systems
like Solaris 10 that didn't support the O_DIRECTORY flag.
This is still racy on such systems, but only in the
case where a directory is replaced by a fifo in
the small window between stat() and open().
* src/system.h (target_directory_operand): On systems without
O_DIRECTORY, ensure the file is a directory before attempting to open().
* tests/cp/special-f.sh: Protect cp with timeout(1),
as cp was seen to hang when trying to overwrite an existing fifo.
* NEWS: Mention the bug fix.
* doc/coreutils.texi (install invocation): For the --compare option,
clarify that the ownership or permissions of the source files don't
matter. Also don't imply --owner or --group need to be specified
for --compare to be effective.
* src/install.c (usage): Add more detail on what's being compared.
Fixes https://bugs.gnu.org/50889
Lookup of file-based capabilities adds 30% overhead to the common
case of ls --color usage. Since the use of file capabilities is
very rare, it doesn't make sense to pay this cost in the common
case. It's better to use getcap to inspect capabilities, and the
following run shows only 8 files using capabilities on my fedora
35 distro (14 years after the feature was introduced to the linux
kernel).
$ getcap -r /
/usr/bin/arping = cap_net_raw+p
/usr/bin/clockdiff = cap_net_raw+p
/usr/bin/gnome-keyring-daemon = cap_ipc_lock+ep
/usr/bin/gnome-shell = cap_sys_nice+ep
/usr/bin/newgidmap = cap_setgid+ep
/usr/bin/newuidmap = cap_setuid+ep
/usr/sbin/mtr-packet = cap_net_raw+ep
/usr/sbin/suexec = cap_setgid,cap_setuid+ep
* src/dircolors.hin: Set "CAPABILITY" to "00", to indicate unused.
* src/ls.c: Set the default C_CAP color to not colored.
* NEWS: Mention the change in behavior.
* src/dircolors.hin: Add patterns for suffixes for "backup files".
The color used is so they stand out less than non-backup files,
and bright black works well on both light and dark backgrounds.
* THANKS.in: Remove duplicate.
Fixes https://bugs.gnu.org/54521
* src/join.c (usage): Clarify that -e is not sufficient
to enable output of missing fields from one of the inputs.
Rather the -12jo options are required to explicitly
enable output of those fields.
Fixes https://bugs.gnu.org/54625
* tests/misc/printf-mb.sh: Given we shortcut the single char
(invalid multi-byte) case, add a case to ensure we're correctly
checking the return from mbrtowc().
* src/printf.c (STRTOX): Update to support multi-byte chars.
* tests/misc/printf-mb.sh: Add a new test.
* tests/local.mk: Reference the new test.
* NEWS: Mention the improvement.
Fixes https://bugs.gnu.org/54388
* src/test.c (usage): State that -rwx is determined by
user access, rather than permission bits.
* doc/coreutils.texi (Access permission tests): Likewise.
* man/test.x [SEE ALSO]: access(2).
Fixes https://bugs.gnu.org/54338
* cfg.mk (sc_die_EXIT_FAILURE): Generalize to match any EXIT_ define,
and also relax to ignore error() usage with ternary operator.
* src/chroot.c (main): Use () to avoid the sc_error_quotes check.
Revert to the default behavior before the introduction of statx().
* src/stat.c (do_stat): Set AT_NO_AUTOMOUNT without --cached=never.
* doc/coreutils.texi (stat invocation): Mention the automount
behavior with --cached=never.
* NEWS: Mention the change in behavior.
Fixes https://bugs.gnu.org/54287
statx() has different defaults wrt automounting
compared to stat() or lstat(), so explicitly
set the AT_NO_AUTOMOUNT flag to suppress that behavior,
and avoid unintended operations or potential errors.
* src/ls.c (do_statx): Pass AT_NO_AUTOMOUNT to avoid this behavior.
* NEWS: Mention the change in behavior.
Fixes https://bugs.gnu.org/54286
Signed-off-by: Rohan Sable <rsable@redhat.com>
On macOS, isspace(0x85) returns true,
which results in splitting within multi-byte characters.
* src/fmt.c (get_line): s/isspace/c_isspace/.
* tests/fmt/non-space.sh: Add a new test.
* tests/local.mk: Reference new test.
* NEWS: Mention the fix.
Addresses https://bugs.gnu.org/54124
* tests/misc/wc-nbsp.sh: Only the en_US.iso8859-1 form
is accepted on macOS 10.15.7 at least. GNU/Linux also
accepts ISO-8859-1 (and canonicalizes the charmap to this).
This implements my suggestion in Bug#54112.
* src/dd.c (usage): Document the change.
(parse_integer, scanargs): Implement the change.
Omit some now-obsolete checks for invalid flags.
* tests/dd/bytes.sh: Test the new behavior, while retaining
checks for the now-obsolete usage.
* tests/dd/nocache_eof.sh: Avoid now-obsolete usage.
Alias iseek=N to skip=N, oseek=N to seek=N (Bug#45648).
* src/dd.c (scanargs): Parse iseek= and oseek=.
* tests/dd/skip-seek.pl (sk-seek5): New test case.
Do not allocate I/O buffer if copy_file_range suffices.
* src/copy.c (sparse_copy, lseek_copy): Buffer arg is now char **
instead of char *, and buffer is now allocated only if needed.
All uses changed.
It's more common to use bold style than not,
for references to other man pages.
Ideally each man page renderer would highlight references,
but currently some rely on styles in the page itself.
* man/help2man: Implement a --bold-refs option that
will mark up references like "name(1)" with bold
style around the "name" component.
* man/local.mk: Pass --bold-refs to our help2man unless disabled.
* configure.ac: Add a --disable-bold-man-page-references option.
Addresses https://bugs.gnu.org/53977
COLORTERM is an environment used usually to expose truecolor support in
terminal emulators. Therefore support matches on that in addition
to TERM. Also set the default COLORTERM match pattern so that
we apply colors if COLORTERM is any value.
This implicitly supports a terminal like "foot"
without a need for an explicit TERM entry.
* NEWS: Mention the new feature.
* src/dircolors.c (main): Match COLORTERM like we do for TERM.
* src/dircolors.hin: Add default config to match any COLORTERM.
* tests/misc/dircolors.pl: Add test cases.
Problem reported by Dan Jacobson (Bug#48248).
* doc/coreutils.texi (tr invocation): Improve documentation for
tr's failure to support multibyte characters POSIX-style.
* doc/coreutils.texi (tr invocation), src/tr.c (usage):
Use terminology closer to POSIX's.
* NEWS: Mention the new feature.
* doc/coreutils.texi (dircolors invocation): Describe the new
--print-ls-colors option.
* src/dircolors.c (print_ls_colors): A new global to select
between shell or terminal output.
(append_entry): A new function refactored from dc_parse_stream()
to append the entry in the appropriate format.
(dc_parse_stream): Adjust to call append_entry().
* tests/misc/dircolors.pl: Add test cases.
since gnulib commit ff208d546a,
related to coreutils commit v9.0-143-gabde15969
we no longer maintain numeric IDs through chopt->{user,group}_name.
Therefore we need to adjust to ensure tests/chown/basic.sh passes.
* src/chown-core.c (uid_to_str, gid_to_str): New helper functions
to convert numeric id to string.
(change_file_owner): Use the above new functions to pass
numeric ids to describe_change().
This also affects ls -v in some corner cases.
Problems reported by Michael Debertol <https://bugs.gnu.org/49239>.
While looking into this, I spotted some more areas where the
code and documentation did not agree, or where the documentation
was unclear. In some cases I changed the code; in others
the documentation. I hope things are nailed down better now.
* doc/sort-version.texi: Distinguish more carefully between
characters and bytes. Say that non-identical strings can
compare equal, since they now can. Improve readability in
various ways. Make it clearer that a suffix can be the
entire string.
* src/ls.c (cmp_version): Fall back on strcmp if filevercmp
reports equality, since filevercmp is no longer a total order.
* src/sort.c (keycompare): Use filenvercmp, to treat NULs correctly.
* tests/misc/ls-misc.pl (v_files):
Adjust test to match new behavior.
* tests/misc/sort-version.sh: Add tests for stability,
and for sorting with NUL bytes.
* src/system.h (emit_ancillary_info): While supported if entered
manually, the "[" character is not highlighted as part of a
URL by default in terminals, so avoid using it.
Addresses https://bugs.gnu.org/53946
* src/system.h: Adjust the alignment of the --help
and --version option descriptions, to start at column 21.
This better aligns with the descriptions of most commands,
and also aligns with the minimum column a description must
start at to ensure a blank line is not output when a description
follows an option on a line by itself.
* src/rmdir.c (usage): Move description to column 21,
so that a --long-option on its own line without a
trailing description, doesn't have an erroneous blank
line inserted between the option and description.
Also group descriptions with blank lines rather than indents,
so that man pages don't have erroneous blank lines
added within the description.
Addresses https://bugs.gnu.org/53946
* src/ls.c (usage): Use blank lines to group multi-line
option descriptions, rather than indenting.
This results in more consistent alignment of descriptions,
and also avoids erroneous new lines in generated in man pages.
Addresses https://bugs.gnu.org/53946
* doc/coreutils.texi, doc/sort-version.texi:
Capitalize “Coreutils”.
* doc/sort-version.texi: Don’t emphasize natural sort so much,
since Coreutils has just version sort.
Use the term “lexicographic” instead of “alphabetic” or “standard”.
Suggest combining ‘V’ with ‘b’, and show why ‘b’ is needed.
Use shorter titles for sections, as GNU Emacs displays info poorly
when titles are too long to fit in a line.
Use @samp instead of @code for samples of data.
Do not use @samp{@code{...}}; @samp{...} should suffice and
double-nesting looks bad with Emacs.
Omit blank lines in examples that would not be present
in actual shell sessions.
Quote with `` and '', not with " or with '.
Mention dpkg --compare-versions more prominently.
Don’t rely on "\n" being equivalent to "\\n" in shell args.
Prefer Unicode name for hyphen-minus.
* doc/coreutils.texi (date invocation): Remove @var{...} usage,
as that capitalizes in the representation and thus somewhat
ambiguates the format wrt Month and Minute. This also avoids
a syntax check failure about redundant capitalization in @var{}.
Problem reported by Dan Jacobson (Bug#51288).
* doc/coreutils.texi (date invocation, Setting the time)
(Options for date):
* src/date.c (usage): Improve doc.
Problem reported by Vladimir D. Seleznev (Bug#53631).
* src/id.c (main): Do not canonicalize user name before
deciding what groups the user belongs to.
* configure.ac: Move the single-binary code before the
gcc-warnings code, so that the latter can depend on the former.
Suppress -Wsuggest-attribute=noreturn with single binaries,
to avoid diagnostics like the following:
src/expr.c: In function 'single_binary_main_expr':
error: function might be candidate for attribute 'noreturn'
Problem reported by Pádraig Brady in:
https://lists.gnu.org/r/coreutils/2022-01/msg00061.html
* src/basenc.c (finish_and_exit, do_encode, do_decode):
* src/comm.c (compare_files):
* src/tsort.c (tsort):
* src/uptime.c (uptime):
Mark with _Noreturn. Otherwise, unoptimized compilations may warn
that the calling renamed-main function doesn't return a value,
when !lint and when single-binary.
* src/dd.c (parse_integer): Avoid undefined behavior
that accesses an uninitialized ‘n’ when e == LONGINT_INVALID.
Return more-accurate error code when INTMAX_MAX < n.
* src/hostname.c (sethostname): Provide a substitute on all
platforms, to simplify the mainline code.
(main): Simplify. Remove an IF_LINT.
Use main_exit rather than return.
* src/factor.c (factor_using_squfof) [USE_SQUFOF]:
Use plain assert (...), not IF_LINT (assert (...)).
This code is currently never compiled or executed,
so this is merely a symbolic cleanup.
* src/cut.c (enum operating_mode, operating_mode)
(output_delimiter_specified, cut_stream):
Remove; no longer needed.
(output_delimiter_default): New static var. Code can now
use ‘output_delimiter_string != output_delimiter_default’
instead of ‘output_delimiter_specified’.
(cut_file): New arg CUT_STREAM. Caller changed.
(main): Simplify. Coalesce duplicate code. Redo to avoid need
for IF_LINT, or for the static var. No need to xstrdup optarg.
* src/basenc.c (finish_and_exit): New function.
(do_encode, do_decode): Use it. Accept new INFILE arg. Remove
no-longer-needed IF_LINT code. Exit when done. Caller changed.
Also, close a no-longer-needed file descriptor when falling
back from inotify.
* src/tail.c (tail_forever_inotify): Return void, not bool. Exit
on fatal error, or on successful completion. Accept an extra
argument pointing to a hash table that the caller should free on
non-fatal error; this simplifies cleanup. Don’t bother setting
errno when returning. Caller changed.
(main): Omit no-longer-needed IF_LINT code. Close inotify
descriptor if inotify fails; this fixes a file descriptor leak and
means we needn’t call inotify_rm_watch. Use main_exit, not return.
* src/copy.c (dest_info_free, src_info_free) [lint]:
Remove. All uses removed.
(copy_internal): Pacify only Clang and Coverity; GCC doesn’t need it.
* src/cp-hash.c (forget_all) [lint]: Remove. All uses removed.
* src/cp.c, src/install.c, src/ln.c, src/mv.c (main):
Use main_exit, not return.
* src/sort.c (pipe_fork, keycompare, sort, main):
Remove lint code that no longer seems to be needed.
(sort): Unconditionally compile ifdef lint code that is needed
to free storage even when not linting.
(main): Use main_exit, not return.
* src/ptx.c (unescape_string): Rename from copy_unescaped_string,
and unescape the string in place. Callers changed. This way,
we needn’t allocate storage and thus needn’t worry about
-fsanitizer=leak.
* src/tsort.c (struct item.balance): Now signed char to save space.
(struct item.printed): New member.
(new_item): Initialize k->printed to false. Simplify via xzalloc.
(scan_zeros): Use k->printed rather than nulling out string.
(tsort): Move exiting code here ...
(main): ... from here.
(tsort) [lint]: Omit no-longer-needed code. Instead, set head->printed.
This introduces a new macro main_exit, which is useful
for pacifying gcc -fsanitizer=lint and in some cases
means we can remove some ‘IF_LINT’ and ‘ifdef lint’ code.
* src/expr.c (main): Use main_exit, not return.
(docolon): Omit an IF_LINT that GCC no longer needs.
* src/system.h (main_exit): New macro.
Use more constrained argument matching
to improve forward compatibility and robustness.
For example it's better that `cksum -a sha3` is _not_
equivalent to `cksum -a sha386`, so that a user
specifying `-a sha3` on an older cksum would not be surprised.
Also argmatch() is used when parsing tags from lines like:
SHA3 (filename) = abcedf....
so it's more robust that older cksum instances to fail
earlier in the parsing process, when parsing output from
possible future cksum implementations that might support SHA3.
* src/digest.c (algorithm_from_tag): Use argmatch_exact()
to ensure we don't match abbreviated algorithms.
(main): Likewise.
* tests/misc/cksum-a.sh: Add a test case.
When the destination for mv is a directory, use functions like openat
to access the destination files, when such functions are available.
This should be more efficient and should avoid some race conditions.
Likewise for 'install'.
* src/cp.c (must_be_working_directory, target_directory_operand)
(target_dirfd_valid): Move from here ...
* src/system.h: ... to here, so that install and mv can use them.
Make them inline so GCC doesn’t complain.
* src/install.c (lchown) [HAVE_LCHOWN]: Remove; no longer needed.
(need_copy, copy_file, change_attributes, change_timestamps)
(install_file_in_file, install_file_in_dir):
New args for directory-relative names. All uses changed.
Continue to pass full names as needed, for diagnostics and for
lower-level functions that do not support directory-relative names.
(install_file_in_dir): Update *TARGET_DIRFD as needed.
(main): Handle target-directory in the new, cp-like way.
* src/mv.c (remove_trailing_slashes): Remove static var; now local.
(do_move): New args for directory-relative names. All uses changed.
Continue to pass full names as needed, for diagnostics and for
lower-level functions that do not support directory-relative names.
(movefile): Remove; no longer needed.
(main): Handle target-directory in the new, cp-like way.
* tests/install/basic-1.sh:
* tests/mv/diag.sh: Adjust to match new diagnostic wording.
Problem reported by Sworddragon (Bug#51345).
* src/dd.c (cleanup): Synchronize output unless dd has been interrupted.
(synchronize_output): New function, split out from dd_copy.
Update conversions_mask so synchronization is done at most once.
(main): Do not die with the output file open, since we want to be
able to synchronize it before exiting. Synchronize output before
exiting.
Problem reported by Sworddragon (Bug#51482).
* src/dd.c (reported_w_bytes): New var.
(print_xfer_stats): Set it.
(dd_copy): Print a final progress report if useful before
synchronizing output data.
* src/csplit.c: Prefer signed integers to unsigned for sizes
when either will do. Check for some unlikely overflows.
(INCR_SIZE): Remove; no longer used.
(free_buffer): Also free the arg, simplifying callers.
(get_new_buffer): Use xpalloc instead of computing new
size by hand. Add ATTRIBUTE_DEALLOC.
(delete_all_files, close_output_file):
If unlink fails with ENOENT, treat it as success.
(close_output_file): If unlink fails, decrement count anyway.
(parse_repeat_count, parse_patterns): Check for int overflow.
(check_format_conv_type): Use signed format.
Use the new Gnulib modules alignalloc and xalignalloc
to simplify some memory allocation.
Also, fix some unlikely integer overflow problems.
* bootstrap.conf (gnulib_modules): Add alignalloc, xalignalloc.
* src/cat.c, src/copy.c, src/dd.c, src/shred.c, src/split.c:
Include alignalloc.h.
* src/cat.c (main):
* src/copy.c (copy_reg):
* src/dd.c (alloc_ibuf, alloc_obuf):
* src/shred.c (dopass):
* src/split.c (main):
Use alignalloc/xalignalloc/alignfree instead of doing page
alignment by hand.
* src/cat.c (main):
Check for integer overflow in page size calculations.
* src/dd.c (INPUT_BLOCK_SLOP, OUTPUT_BLOCK_SLOP, MAX_BLOCKSIZE):
(real_ibuf, real_obuf) [lint]:
Remove; no longer needed.
(cleanup) [lint]:
(scanargs): Simplify.
* src/ioblksize.h (io_blksize): Do not allow blocksizes largest
than the largest power of two that fits in idx_t and size_t.
* src/shred.c (PAGE_ALIGN_SLOP, PATTERNBUF_SIZE): Remove.
* src/copy.c (copy_reg): Remove a ‘free’ call that does nothing
because its argument is always a null pointer, starting with
2007-11-1608:31:15Z!jim@meyering.net.
Simplify byte-swapping, so that the code no longer needs to
allocate a page before the input buffer.
* src/dd.c (SWAB_ALIGN_OFFSET, char_is_saved, saved_char): Remove.
All uses removed.
(INPUT_BLOCK_SLOP): Simplify to just page_size.
(alloc_ibuf, dd_copy): Adjust to new swab_buffer API.
(swab_buffer): New arg SAVED_BYTE, taking the place of the old
global variables. Do not access BUF[-1].
* src/dd.c: Prefer signed to unsigned types where either will do,
as this helps improve checking with gcc -fsanitize=undefined.
Limit the signed types to their intended ranges.
(MAX_BLOCKSIZE): Don’t exceed IDX_MAX - slop either.
(input_offset_overflow): Remove; overflow now denoted by negative.
(parse_integer): Return INTMAX_MAX on overflow, instead of unspecified.
Do not falsely report overflow for ‘00x99999999999999999999999999999’.
* tests/dd/misc.sh: New test for 00xBIG.
* tests/dd/skip-seek-past-file.sh: Adjust to new diagnostic wording.
New test for BIGxBIG.
* gl/lib/randint.h (randint_all_new):
Do not declare with _GL_ATTRIBUTE_NONNULL (), as
the arg can be a null pointer. This fixes a typo added in
2021-11-01T05:30:28Z!eggert@cs.ucla.edu.
Gnulib now replaces copy_file_range on buggy hosts
so there is no need for Coreutils to worry about the bug.
* src/copy.c: Do not include sys/utsname.h, xstrtol.h.
(functional_copy_file_range): Remove. All uses now
simply call copy_file_range.
Somehow ‘make check’ didn’t catch these the first few times.
* src/copy.c (copy_dir): Don’t pass null pointer to
copy_internal where it now expects non-null if move mode.
* src/cp.c (make_dir_parents_private): Initialize *attr_list
before recentely-added quick return.
'cp A B' attempts to open B as a directory, to see whether to
write to B/A instead of to B. In the common case where the
open fails with ENOENT, do not bother to stat B afterwards
since the stat should also fail with ENOENT.
* src/copy.c (copy_internal, copy): Change bool arg about
nonexistent destination to a 3-way int argument. All callers changed.
(copy_internal): Do not bother to stat a destination already known
to not exist when following symlinks.
When copying to a directory, use functions like openat to access
the destination files, when such functions are available. This
should be more efficient and should avoid some race conditions.
* bootstrap.conf (gnulib_modules): Add areadlinkat-with-size,
fchmodat, fchownat, mkdirat, mkfifoat, utimensat.
* src/copy.c (lchown) [!HAVE_LCHOWN]:
* src/copy.c, src/system.h (rpl_mkfifo, mkfifo) [!HAVE_MKFIFO]:
Remove. All uses removed.
(utimens_symlink): Remove; we shouldn’t have to worry about
those obsolete systems any more. All uses replaced by utimensat.
* src/copy.c (copy_dir, set_owner, fchmod_or_lchmod, copy_reg)
(same_file_ok, writable_destination, overwrite_ok, abandon_move)
(create_hard_link, src_is_dst_backup, copy_internal, copy):
* src/cp.c (make_dir_parents_private, re_protect):
New args for directory-relative names. All uses changed.
Continue to pass full names as needed, for diagnostics and for
lower-level functions like qset_acl that do not support
directory-relative names.
* src/copy.c (copy_reg): Prefer readlinkat to lstatat for merely
checking whether a file is a symlink, to avoid EOVERFLOW issues.
(subst_suffix): New function.
(create_hard_link): Accept a null SRC_NAME as meaning that if it
is needed it needs to be constructed from SRC_RELNAME, DST_NAME,
and DST_RELNAME.
(source_is_dst_backup): Use subst_suffix instead of doing it by hand.
(copy_internal): Remember and use directory-relative names instead
of full names.
* src/cp.c (lchown) [!HAVE_LCHOWN]: Remove. All uses removed.
(must_be_working_directory): New function.
(target_directory_operand): Simply take file name as arg,
and return a file descriptor or negative number on failure;
open with O_DIRECTORY to obtain any file descriptor.
All uses changed.
(target_dirfd_valid): New function.
(do_copy): Use these new functions to obtain a file descriptor
for any target directory, and use directory-relative names
for that directory.
(main): Omit no-longer-needed stat when --target-directory,
as do_copy now does this.
* src/ln.c (O_PATHSEARCH): Move from here ...
* src/system.h: ... to here.
* tests/cp/fail-perm.sh: Adjust to change in diagnostic wording,
and add a test for --no-target-directory.
Commit 2f438fa9f5 (basenc: A new program
complementary to base64/base32) introduced a typo in the list of allowed
commit message prefixes, accidentally changing "basename" to
"nbasename". Revert it back to the correct "basename".
Run "make update-copyright" and then...
* gnulib: Update to latest with copyright year adjusted.
* tests/init.sh: Sync with gnulib to pick up copyright year.
* bootstrap: Likewise.
* tests/sample-test: Adjust to use the single most recent year.
mainly to get updated copyright year
* doc/fdl.texi: Sync from gnulib.
* .gitignore: Add lib/unictype, as bitmap.h therein is depended on
since gnulib commit f698ea71
* NEWS, doc/coreutils.texi (Options for date): Mention this.
* src/date.c (RESOLUTION_OPTION): New constant.
(DEBUG_DATE_PARSING_OPTION): Rename from DEBUG_DATE_PARSING.
All uses changed.
(long_options, usage, main): Support --resolution.
* NEWS, doc/coreutils.texi: Mention this.
* bootstrap.conf (gnulib_modules): Add gettime-res.
* src/date.c (res_width, adjust_resolution): New functions.
(main): Adjust %-N to be %9N, or whatever, before using it.
This fixes a porting bug introduced in
2019-08-12T03:29:00Z!bruno@clisp.org.
Problem discovered on AIX 7.1.
* src/local.mk (LDADD): Add $(LIB_MBRTOWC), since pretty much
every command uses quotearg or mbrtowc or whatever.
(src_sort_LDADD): Add $(LIBPMULTITHREAD) and
$(LIB_PTHREAD_SIGMASK) instead of $(LIBTHREAD).
When configured with --enable-single-binary tools issue incorrect prctl:
prctl(PR_SET_KEEPCAPS, ...) = -1 EINVAL (Invalid argument)
PR_SET_MM_ARG_START is not a prctl 'option' parameter, it's 'arg2'
parameter for the option PR_SET_MM. It also has to have 'arg4' and
'arg5' set to 0 explicitly, otherwise the kernel also returns -EINVAL.
* src/coreutils.c (launch_program): Fix prctl arguments.
* NEWS: Mention the improvement.
Fixes https://bugs.gnu.org/52800
Living so close to Hollywood I know that "colorize"
means adding color to something that was already monochrome,
whereas "color" means to give color to something.
Coreutils apps color text instead of colorizing it.
* scripts/git-hooks/commit-msg: Count UTF-8 characters rather
than bytes to avoid erroneously rejecting as "longer than 72" a
log message line like the UTF-8 one for id.c just prior. It has
77 bytes but only 67 characters.
(check_msg): Read in "utf8" mode. Also include actual length
in the diagnostic.
(main): Don't loop when stdout is redirected, as it is when
invoked via vc-dwim.
Paul Eggert reported privately both the error of counting bytes
rather than chars and the re_edit loop when failing via vc-dwim.
* bootstrap.conf (gnulib_modules): Remove
non-recursive-gnulib-prefix-hack.
(gnulib_tool_option_extras): Add --automake-subdir.
(bootstrap_post_import_hook): No need to massage lib/gnulib.mk.
Problem reported by Jakub Sokołowski (bug #52330).
* src/uname.c [__APPLE__]: Don’t include sys/syctl.h,
mach/machine.h, mach-o/arch.h.
(print_element_env): New function. With __APPLE__, it defers to the
env var UNAME_MACHINE (if given) for uname -m, and similarly for -nrsv.
(main): Use it. For -p with __APPLE__, rely on predefined macros
and omit any 64-bit indication, for compatibility with macOS uname.
* configure.ac: Check for fclonefileat.
* src/copy.c [HAVE_FCLONEFILEAT && !USE_XATTR]:
Include <sys/clonefile.h>.
(copy_reg): If possible, use fclonefileat to clone.
This fixes a bug that I introduced in
2006-12-06T19:44:08Z!eggert@cs.ucla.edu.
* src/copy.c (USE_XATTR): New macro.
(copy_reg): Use it to help the compiler. Prefer open u+w to a
later chmod u=rw; u+r isn’t needed for xattr. For the later u-r,
do only one (or zero) chmod calls instead of two (or one).
In the last chmod, respect the umask instead of ignoring it.
* tests/cp/preserve-mode.sh: Test for the bug.
Prefer MAYBE_UNUSED to _GL_UNUSED, since the C2x syntax
will be [[maybe_unused]] at the start of the declaration,
and we want to look forward to that. All uses of _GL_UNUSED
either changed to MAYBE_UNUSED, or (when not needed) removed.
This fixes an issue introduced in the fix for Bug#11100.
* NEWS: Mention this.
* src/copy.c (copy_reg): Fix obscure bug where open-without-CREAT
failed with ENOENT and we forget to call set_process_security_ctx
before calling open-with-CREAT. Also, don’t bother to unlink
DST_NAME if open failed with ENOENT; and if unlink fails with
ENOENT, don’t consider that to be an error (someone else could
have removed the file for us, and that’s OK). Also, don’t worry
about move mode, since we use O_EXCL|O_CREAT and so won’t open
an existing file.
* tests/misc/env-signal-handler.sh: Use retry_delay_ to
avoid a false failure under load, where env hasn't setup
the SIGINT handling before timeout(1) sends the SIGINT.
Fixes https://bugs.gnu.org/51793
* cfg.mk (exclude_file_name_regexp--sc_system_h_headers):
Add chown-core.h to the regexp, to better decouple from system.h.
* src/env.c: Remove minmax.h include already included in system.h.
* src/libstdbuf.c: Likewise.
* src/prog-fprintf.h: Remove doubled semicolon.
Add _GL_ATTRIBUTE_NONNULL, _GL_ATTRIBUTE_MALLOC,
_GL_ATTRIBUTE_DEALLOC, _GL_ATTRIBUTE_DALLOC_FREE,
_GL_ATTRIBUTE_RETURNS_NONNULL to .h files when appropriate.
* gl/lib/mbsalign.h, gl/lib/randperm.h, src/chown-core.h:
Include stdlib.h, for the benefit of _GL_ATTRIBUTE_DALLOC_FREE.
* gl/lib/randread.c (randread_free_body): New static function.
(randread_new, randread_free): Use it.
* src/copy.c (valid_options): Remove assert that is no longer
needed because it is now checked statically.
* configure.ac (WERROR_CFLAGS): Enable -Wsuggest-attribute=format
for lib/ and src/.
* src/copy.c (copy_attr_error, copy_attr_allerror):
Add ATTRIBUTE_FORMAT.
(copy_attr): Ignore -Wsuggest-attribute=format in the
small section of code that needs it ignored.
* src/test.c (test_syntax_error): Mark with ATTRIBUTE_FORMAT.
(binary_operator): Omit unnecessary NULL args, pacifying
-Wsuggest-attribute=format.
* src/system.h (__attribute__): Remove. Replace all uses that
rely on this by _GL_ATTRIBUTE_xxx or ATTRIBUTE_xxx.
(ATTRIBUTE_WARN_UNUSED_RESULT): Remove. Replace all uses by
NODISCARD.
This will help us make the transition to C2x, where some
attributes must come at the start of function decls.
Leave the attributes alone in .h files for now,
as the Gnulib tradition is to not expose attribute.h to users.
* bootstrap.conf (gnulib_modules): Add ‘attribute’.
* gl/lib/randperm.c, src/make-prime-list.c, src/system.h:
Include attribute.h.
* gl/lib/strnumcmp.c (strnumcmp): Remove _GL_ATTRIBUTE_PURE here,
as this belongs in the .h file.
* gl/lib/strnumcmp.h (strnumcmp): Add _GL_ATTRIBUTE_PURE here.
* src/sort.c (human_numcompare, numcompare): Now ATTRIBUTE_PURE;
discovered due to strnumcmp.h change.
* gl/lib/randperm.c, src/copy.c, src/dd.c, src/df.c, src/digest.c:
* src/env.c, src/expr.c, src/factor.c, src/ls.c:
* src/make-prime-list.c, src/numfmt.c, src/od.c, src/pathchk.c:
* src/pinky.c, src/pr.c, src/ptx.c, src/realpath.c, src/relpath.c:
* src/seq.c, src/sort.c, src/stat.c, src/stty.c, src/system.h:
* src/tr.c, src/uniq.c, src/wc.c:
In .c files, crefer ATTRIBUTE_CONST to _GL_ATTRIBUTE_CONST, and
similarly for ATTRIBUTE_FORMAT and ATTRIBUTE_PURE.
* src/system.h (FALLTHROUGH): Remove; attribute.h defines it.
New warnings are added related to the handling
of thousands grouping characters, decimal points, and sign characters.
Examples now diagnosed are:
$ printf '0,9\n1,a\n' | sort -nk1 --debug -t, -s
sort: key 1 is numeric and spans multiple fields
sort: field separator ‘,’ is treated as a group separator in numbers
1,a
_
0,9
___
$ printf '1,a\n0,9\n' | LC_ALL=fr_FR.utf8 sort -gk1 --debug -t, -s
sort: key 1 is numeric and spans multiple fields
sort: field separator ‘,’ is treated as a decimal point in numbers
0,9
___
1,a
__
$ printf '1.0\n0.9\n' | LC_ALL=fr_FR.utf8 sort -s -k1,1g --debug
sort: note numbers use ‘,’ as a decimal point in this locale
0.9
_
1.0
_
$ LC_ALL=fr_FR.utf8 sort -n --debug /dev/null
sort: text ordering performed using ‘fr_FR.utf8’ sorting rules
sort: note numbers use ‘,’ as a decimal point in this locale
sort: the multi-byte number group separator in this locale \
is not supported
$ sort --debug -t- -k1n /dev/null
sort: key 1 is numeric and spans multiple fields
sort: field separator ‘-’ is treated as a minus sign in numbers
sort: note numbers use ‘.’ as a decimal point in this locale
$ sort --debug -t+ -k1g /dev/null
sort: key 1 is numeric and spans multiple fields
sort: field separator ‘+’ is treated as a plus sign in numbers
sort: note numbers use ‘.’ as a decimal point in this locale
* src/sort.c (key_warnings): Add the warnings above.
* tests/misc/sort-debug-warn.sh: Add test cases.
Also check that all sort invocations succeed.
* NEWS: Mention the improvement.
Addresses https://bugs.gnu.org/51011
That was a false alarm due to a bug in FreeBSD 9.1 truss;
see Pádraig Brady’s report (Bug#51433#29).
* src/copy.c (lseek_copy, infer_scantype): Don’t bother checking
whether lseek returned -1. This doesn’t entirely revert the
previous change, as it keeps the code simplification of the
previous change while reverting the check for -1.
Problem reported by Pádraig Brady (Bug#51433#14).
* src/copy.c (lseek_copy, infer_scantype): Report an error if
lseek with SEEK_DATA or SEEK_HOLE returns less than -1,
as this is an lseek bug.
* src/echo.c (usage): Say printf(1) is preferred
due to being more standard and robust.
* man/echo.x [SEE ALSO]: Reference printf(1).
* doc/coreutils.texi (echo invocation): Mention in the
summary that echo is not robust when outputting
any string, and that printf is preferred.
Also expand on the examples showing how to
output a single '-n' string.
Addresses https://bugs.gnu.org/51311
* doc/coreutils.texi (timeout invocation): Clarify
that -k is ignored if either its duration or the
main timeout duration is 0.
Addresses https://bugs.gnu.org/51128
* src/timeout.c (main): Propagate the killed status from the child.
* doc/coreutils.texi (timeout invocation): Remove the
description of the --foreground specific handling of SIGKILL,
now that it's consistent with the default mode of operation.
* tests/misc/timeout.sh: Add a test case.
* NEWS: Mention the change in behavior.
Fixes https://bugs.gnu.org/51135
* doc/coreutils.texi (timeout invocation): Add detail on
how --foreground allows timeout(1) to use more standard
exit status as the uncatchable SIGKILL is not sent to itself.
Fixes https://bugs.gnu.org/51135
* gl/lib/strintcmp.c (strintcmp): Don’t assume that the input
cannot contain ((char) -1), as this equals '\377' when char is
signed (assuming 8-bit char).
* src/sort.c (decimal_point): Now char, to make it clear
that it’s always in char range now.
(NON_CHAR): New constant.
(traverse_raw_number): Return char not unsigned char;
this is simpler and could be faster. All callers changed.
(main): Do not convert decimal_point and thousands_sep to
unsigned char, as this can mishandle comparisons on
machines where char is signed and the input data contains
((char) -1). Use NON_CHAR, not -1, as an out-of-range value for
thousands_sep.
Use C11-style _Noreturn instead of the old ATTRIBUTE_NORETURN
macro. This pacifies clang on OpenBSD 6.9, which otherwise
complains "'noreturn' function does return" in some places.
* gl/lib/randread.c, src/system.h (ATTRIBUTE_NORETURN):
Remove. All uses either removed as GCC no longer needs them, or
changed to C11-style _Noreturn since Gnulib arranges for _Noreturn
globally nowadays.
* doc/coreutils.texi (ls invocation - general output formatting):
The option ordering was not changed when the option was renamed
from --null to --zero.
* init.cfg (seek_data_capable_): Add a timeout to ensure failure for
slow lseek(...SEEK_DATA) calls (even if that syscall isn't interrupted).
* tests/cp/sparse-perf.sh: Run the SEEK_DATA check on the
1TiB empty file to exclude both FreeBSD 9.1 which takes 35s,
and ZFS which requires a delay of about 5s between file creation
and use of SEEK_DATA to correctly determine it's empty (return ENXIO).
Also remove the stat size checks as they invalidate the test
due to cp never writing data due to it being always zeros,
and thus converted to holes in the output.
* src/chmod.c: Reorder enum so CH_NOT_APPLIED
can be treated as a non error.
* tests/chmod/ignore-symlink.sh: A new test.
* tests/local.mk: Reference the new test.
* NEWS: Mention the bug fix.
Fixes https://bugs.gnu.org/50784
* tests/cp/sparse-perf.sh: Avoid the case where
we saw SEEK_DATA take 35s to return a result
against a 1TB sparse file. This happened on
a FreeBSD 9.1 VM at least.
Reported by Nelson H. F. Beebe.
* src/cksum.c (crc_sum_stream): On sparc64 for example,
a crc of 0 was printed due to mismatch in size of
variable copied between generator and output functions.
uint_fast32_t is generally 64 bits on 64 bit systems,
so we copy through an int to ensure we don't use the wrong
end of a 64 bit variable.
Reported by Nelson H. F. Beebe
* bootstrap.conf: We only need poll on Linux and AIX
where poll is not replaced. Also resinstate dependence
on select so we can use it unconditionally.
* src/tail.c (check_output_alive): Reinstate use of select()
by default as poll was seen to be ineffective for this
application on macOS.
Fixes https://bugs.gnu.org/50714
* src/digest.c: Allow using the --untagged option with --check,
so that `cksum -a md5 --untagged` used to emulate md5sum for example,
may be augmented with the --check option. Also support the --tag
option with cksum, to allow overriding a previous --untagged setting.
* doc/coreutils.texi: Adjust accordingly.
* tests/misc/cksum-a.sh: Likewise.
* tests/tail-2/F-vs-rename.sh: Keep stdout and stderr separate,
so that interspersion doesn't impact regex checks. Also wait
for each file's data to be printed to avoid multiple writes
to a file to be printed in a single iteration, which would
impact the regex checks. Also we refactor the check function,
rather than repeatedly redefining variations.
This is wrong fix really, as only introducing delay I think.
* tests/tail-2/F-vs-rename.sh: Avoid a rare false failure
due to a race in the test. Now wait until tail has noticed
that b is replaced before writing to a, so that the subsequent
write of "y" to b will be displayed independently from
current contents of b ("x").
* tests/ls/removed-directory.sh: On FreeBSD 9.1 at least,
one gets ENOENT when trying to traverse the current removed dir
with ../, so instead reference the parent dir directly.
Here are the warnings:
src/chmod.c:175:3: error: 'ch.new_mode' may be used uninitialized in\
this function [-Werror=maybe-uninitialized]
175 | strmode (ch->new_mode, perms);
src/chmod.c:178:3: error: 'ch.old_mode' may be used uninitialized in\
this function [-Werror=maybe-uninitialized]
178 | strmode (ch->old_mode, old_perms);
* src/chmod.c (process_file): Initialize ch. Its new_mode and
old_mode fields could indeed be used uninitialized to form mode
strings, but those are used only when built from initialized members.
* src/digest.c (digest_check): Treat empty lines like comments,
as commented checksum files very often have empty lines.
* tests/misc/md5sum.pl: Adjust accordingly.
* src/digest.c: Always set the digest_length, so that
we check the correct number of hex digits when parsing
non tagged format checksums.
* tests/misc/cksum-a.sh: Add a test case. Also fix
up this test which was ineffective due to fail=1
being set in a subshell and ignored.
Support checksum files with CRLF line endings,
which is a common gotcha for using --check on windows,
or with checksum files generated on windows.
Note we escape \r here to support the original coreutils format
(with file name at EOL), and file names with literal
\r characters as the last character of their name.
* src/digest.c (filename_unescape): Convert \\r -> \r.
(print_filename): Escape \r -> \\r.
(output_file): Detect \r chars in file names.
(digest_check): Ignore literal \r char at EOL.
* tests/misc/md5sum.pl: Add a test case.
* tests/misc/sha1sum.pl: Likewise.
* NEWS: Mention the improvement.
This only practically matters on windows.
But given there are separate text handling options in cygwin,
keep the interface simple, and avoid exposing the
confusing binary/text difference here.
* doc/coreutils.texi (md5sum invocation): Mention that
--binary and --text are not supported by the cksum command.
* src/digest.c: Set flag to use binary mode by default.
(output_file): Don't distinguish text and binary modes with
' ' and '*', and just use ' ' always.
This format is a better default, since it results in simpler usage,
as you don't need to specify --tag on generation or -a on
checking invocations. Also it's a more general format supporting
mixed and length adjusted digests.
* doc/coreutils.texi (cksum invocation): Document a new --untagged
option, to use the older coreutils format.
(md5sum invocation): Mention that cksum doesn't support --tag.
* src/digest.c: Adjust cksum(1) to default to --tag,
and accept the new --untagged option.
* tests/misc/b2sum.sh: Adjust accordingly.
* tests/misc/cksum-a.sh: Likewise.
* tests/misc/cksum-c.sh: Likewise.
* src/cksum.h: Thread DELIM through the output functions.
* src/digest.c: Likewise.
* src/sum.c: Likewise.
* src/sum.h: Likewise.
* src/cksum.c: Likewise. Also adjust check to allow -z
with traditional output modes. Also ajust the global variable
name to avoid shadowing warnings.
* tests/misc/cksum-a.sh: Adjust accordingly.
This will be generally useful going forward, for sha3-256 etc.
* src/digest.c: Rename b2_length to digest_length, and
adjust/simplify the code to operate on this for both
b2sum and cksum -a blake2b.
Support `cksum --check FILE` without having to specify a digest
algorithm, allowing for more generic file check instructions.
This also supports mixed digest checksum files, supporting
more robust multi digest checks.
* src/digest.c (algorithm_from_tag): A new function to
identify the digest algorithm from a tagged format line.
(split3): Set the algorithm depending on tag, and update
the expected digest length accordingly.
* tests/misc/cksum-c.sh: Add a new test.
* tests/local.mk: Reference the new test.
* tests/misc/md5sum.pl: Adjust to more generic error.
* tests/misc/sha1sum.pl: Likewise.
* doc/coreutils.texi (md5sum invocation): Mention the new -c feature.
* NEWS: Mention the new feature.
Add message digest sm3, which uses the OSCCA SM3 secure
hash (OSCCA GM/T 0004-2012 SM3) generic hash transformation.
* bootstrap.conf: Add the sm3 module.
* doc/coreutils.texi: Mention the cksum -a option.
* src/digest.c: Provide support for --algorithm='sm3'.
* tests/misc/sm3sum.pl: Add a new test (from Tianjia Zhang)
* tests/local.mk: Reference the new test.
* NEWS: Mention the new feature.
Tested-by: Tianjia Zhang <tianjia.zhang@linux.alibaba.com>
* src/digest.c: Organize HASH_ALGO_CKSUM to be table driven,
and amalgamate all digest algorithms.
(main): Parse all options if HASH_ALGO_CKSUM, and disallow
--tag, --zero, and --check with the traditional bsd, sysv, and crc
checksums for now.
* src/local.mk: Reorganize to include all digest modules in cksum.
* tests/misc/cksum-a.sh: Add a new test.
* tests/misc/b2sum.sh: Update to default to checking with cksum,
as b2sum's implementation diverges a bit from the others.
* tests/local.mk: Reference the new test.
* doc/coreutils.texi (cksum invocation): Adjust the summary to
identify the new mode, and document the new --algorithm option.
* man/cksum.x: Adjust description to be more general.
* man/*sum.x: Add [See Also] section referencing cksum(1).
* NEWS: Mention the new feature.
* cfg.mk: Adjust cksum.c to not require config.h
and support a main (for crctab) without calling bindtextdomain().
* po/POTFILES.in: Remove cksum_pclmul.c since it no longer
concerns itself with diagnostics.
* src/cksum.c: Refactor to just providing stream digest,
and digest printing functionality.
* src/cksum.h: Adjust to the new interface.
* src/cksum_pclmul.c: Remove diagnostics, and determine errors
internally.
* src/crctab.c: Separate from cksum.h since that's now included
multiple times.
* src/digest.c: Provide cksum(1) functionality if -DHASH_ALGO_CKSUM
* src/local.mk: Adjust to new crctab.c and HASH_ALGO_CKSUM define.
This should have been part of commit v8.32-113-gb73b9fcb1
* doc/coreutils.texi (cksum invocation): Add the --debug description.
* src/cksum.c (usage): Likewise.
(main): Also give explicit indication when using generic hardware.
Since digest will be providing all digest functionality,
refactor sum.c into it.
* po/POTFILES.in: sum.c no longer has translatable strings so remove.
* src/digest.c: Call out to new stream interfaces in sum.c
* src/local.mk: Adjust sources for the sum binary.
* src/sum.c: Provide a stream interface for BSD and SYSV digests.
* src/sum.h: A new file to declare the exported functions in sum.c
md5sum.c will be the base for all digest functions,
so rename accordingly.
* src/md5sum.c: Rename to ...
* src/digest.c: ... renamed from md5sum.c
* scripts/git-hooks/commit-msg: Allow digest: commit prefix.
* po.POTFILES.in: Adjust to new name.
* src/local.mk: Likewise.
* doc/coreutils.texi (whoami invocation): Clarify it prints names,
not numeric IDs.
* man/whoami.x: Likewise.
* man/logname.x: Reference getlogin(3).
* src/logname.c: Clarify that it prints the login name,
rather than the name of the effective user ID.
Fixes https://bugs.gnu.org/48894
* tests/ls/stat-vs-dirent.sh: Skip the test if we can't stat(1),
as the file may have been removed, or have a malformed name
due to '\n' etc. in the file name.
* gnulib: Update to latest. This fixes a gnulib test failure in base64,
among other fixes.
* cfg.mk: Disable sc_indent as auto indent is too invasive for now.
Adjust to output the file name if any name parameter is passed.
This is consistent with sum -s, cksum, and sum implementations
on other platforms. This should not cause significant compat
issues, as multiple fields are already output, and so already
need to be parsed.
* src/sum.c (bsd_sum_file): Output the file name
if any name parameter is passed.
* tests/misc/sum.pl: Adjust accordingly.
* doc/coreutils.texi (sum invocation): Likewise.
* NEWS: Mention the change in behavior.
* tests/misc/help-version.sh: Test that /dev/full causes
shell printf to fail. This ports better to NetBSD 9.88.46,
where it doesn’t. Problem reported by Nelson H. F. Beebe.
* tests/misc/help-version.sh: Merge gzip-related changes
back from gzip/tests/help-version. This fixes problems
when TERM is not 'dumb', and should simplify maintenance.
Emil Lundberg <lundberg.emil@gmail.com> reports in
https://bugs.gnu.org/49741 about a 'basenc --base64 -d' decoding bug.
The input buffer length was not divisible by 3, resulting in
decoding errors.
* NEWS: Mention fix.
* src/basenc.c (DEC_BLOCKSIZE): Change from 1024*5 to 4200 (35*3*5*8)
which is divisible by 3,4,5,8 - satisfying both base32 and base64;
Use compile-time verify() macro to enforce the above.
* tests/misc/basenc.pl: Add test.
This patch modifies basenc to prefer signed integers to
unsigned, as signed are less error-prone.
This patch also updates Gnulib to to latest, which updates Gnulib’s
base32 and base64 modules to prefer signed to unsigned integers.
* src/basenc.c: Include idx.h.
(struct base2_decode_context): Use unsigned char, not unsigned
for an octet that must fit in an unsigned char.
(base_encode, struct base_decode_context)
(base64_decode_ctx_wrapper, prepare_inbuf, base64url_encode)
(base64url_decode_ctx_wrapper, base32_decode_ctx_wrapper)
(base32hex_encode, base32hex_decode_ctx_wrapper, base16_encode)
(base16_decode_ctx, z85_encode, Z85_HI_CTX_TO_32BIT_VAL)
(z85_decoding, z85_decode_ctx, base2msbf_encode)
(base2lsbf_encode, base2lsbf_decode_ctx, base2msbf_decode_ctx)
(wrap_write, do_encode, do_decode, main):
Prefer signed integers to unsigned.
(main): Treat extremely large wrap columns as if they were
infinite; that’s good enough. Since we’re now using xstrtoimax,
this allows ‘-w -0’ (same as ‘-w 0’).
* tests/misc/base64.pl (gen_tests): -w-0 is no longer an error.
find-mount-point.h rightly includes <stdlib.h> for its use
of _GL_ATTRIBUTE_DEALLOC_FREE, which uses free, yet that new
inclusion provoked a syntax-check failure. Exempt this header
file as we've done for others.
* cfg.mk (exclude_file_name_regexp--sc_system_h_headers):
Add find-mount-point.h to the regexp.
(sc_system_h_headers): Use grep -E, for a more readable regexp.
This better tests the SEEK_HOLE logic which
replaced the original fiemap hole identification logic.
Also it avoids a false failure in sparse-2.sh
on reflink supporting file systems, where we
try to correlate the file sizes produced by cp and dd.
* tests/cp/sparse-2.sh: s/cp/cp --reflink=never/
* tests/cp/sparse-extents-2.sh: Likewise.
* tests/cp/sparse-extents.sh: Likewise.
* tests/cp/sparse-perf.sh: Likewise.
* tests/cp/sparse.sh: Likewise.
Fixes https://github.com/coreutils/coreutils/issues/54
This is so that commands like ‘fmt - -’ read from stdin
both times, even when it is a tty. Fix some other minor
issues that are related.
* src/blake2/b2sum.c (main):
* src/cksum.c (cksum):
* src/cut.c (cut_file):
* src/expand-common.c (next_file):
* src/fmt.c (fmt):
* src/fold.c (fold_file):
* src/md5sum.c (digest_file, digest_check):
* src/nl.c (nl_file):
* src/od.c (check_and_close):
* src/paste.c (paste_parallel, paste_serial):
* src/pr.c (close_file):
* src/sum.c (bsd_sum_file):
Use clearerr on stdin so that stdin can be read multiple times
even if it is a tty. Do not assume that ferror preserves errno as
POSIX does not guarantee this. Coalesce duplicate diagnostic
calls.
* src/blake2/b2sum.c (main):
* src/fmt.c (main, fmt):
Report read error, even if it's merely fclose failure.
* src/fmt.c: Include die.h.
(fmt): New arg FILE. Close input (reporting error) if not stdin.
All callers changed.
* src/ptx.c (swallow_file_in_memory): Clear stdin's EOF flag.
* src/sort.c (xfclose): Remove unnecessary feof call.
Problem reported by Michael Debertol (Bug#50070).
* NEWS: Mention the fix.
* src/chmod.c (struct change_status): New struct, replacing the
old enum Change_status. All uses changed.
(describe_change): Distinguish between cases depending on
whether 'stat' or its equivalent succeeded. Report a line
of output even if 'stat' failed, as that matches the documentation.
Rework to avoid casts.
(process_file): Do not output nonsense modes computed from
uninitialized storage, removing a couple of IF_LINTs. Simplify by
defaulting to CH_NO_STAT.
If the command-line argument is automounted, df would use
stat info that became wrong after the following open.
* NEWS: Mention the fix (bug#50012).
* src/df.c (automount_stat_err): New function.
This fixes the hang on fifos in a better way, by using O_NONBLOCK.
(main): Use it.
We must delay handling when \r is the last character
of the buffer being processed, as the next character
may or may not be \n.
* src/cat.c (pending_cr): A new global to record whether
the last character processed (in -E mode) is '\r'.
(cat): Honor pending_cr when processing the start of the buffer.
(main): Honor pending_cr if no more files to process.
* tests/misc/cat-E.sh: Add test cases.
Fixes https://bugs.gnu.org/49925
Pacify GCC 11.1 -fanalyzer.
* src/uniq.c (check_file): Use simpler test to check whether this
is the first time through the loop. Although the old test was
correct, the new one is easier to understand and perhaps a tiny
bit more efficient.
Caught by GCC 11.1 -fanalyzer.
* src/numfmt.c (simple_strtod_int): Remove unnecessary test of
*endptr vs NULL. Presumably this was a typo and **endptr was
intended instead of *endptr, but an **endptr test is also
unnecessary since c_isdigit (0) returns false.
* doc/coreutils.texi (tr invocation): Provide a summary
list of the available options, which is useful to
provide a quick reminder for those already familiar
with the functionality of tr.
Fixes https://bugs.gnu.org/49764
In documentation and comments, don’t assume that secondary storage
devices are disk devices. Similarly, don’t assume that main memory
uses magnetic cores, which became obsolete in the 1970s.
* src/du.c (usage):
* src/ls.c (usage):
* src/shred.c (usage): Reword to avoid “disk” in usage messages.
* doc/coreutils.texi (ls invocation): Document implementation more
closely. Be more consistent about style. Omit some needless words.
* src/ls.c (usage): Don’t overdocument -f, as the details were wrong.
Omit -1 advice as it’s a bit obsolete now that we have --zero and
is a bit much for --usage output anyway.
* NEWS, doc/coreutils.texi (General output formatting):
* src/ls.c (usage):
Document this.
* src/ls.c (ZERO_OPTION): Rename from NULL_OPTION.
All uses changed.
(long_options): Rename --null to --zero.
(dired_dump_obstack, main, print_dir): Use '\n' instead of
eolbyte where eolbyte must equal '\n'.
(decode_switches): Decode --zero instead of --null.
--zero also implies -1, -N, --color=none, --show-control-chars.
Use easier-to-decipher code to set ‘format’ and ‘dired’.
Reject attempts to combine --dired and --zero.
* tests/local.mk: Adjust to test script renaming.
* tests/ls/zero-option.sh: Rename from tests/ls/null-option.sh,
and test --zero instead of --null.
* src/ls.c (enum time_type, enum sort_type, enum indicator_style)
(enum Dereference_symlink, ignore_mode):
Put ‘= 0’ after default values, since the code relies
on static storage defaulting to zero.
(enum sort_type): Reorder so that -1 can be used to represent unset.
(main): Test print_with_color after parse_ls_color may have reset it.
(decode_line_length): Return the line length instead of setting
static storage. All uses changed. Treat line lengths exceeding
PTRDIFF_MAX as infinite, to avoid pointer-subtraction glitches.
(stdout_isatty): New function, to avoid calling isatty twice.
(decode_switches): Calculate defaults more lazily, to avoid using
syscalls or getenv during startup unless the results are more
likely to be needed. Use -1 to indicate options that haven’t been
set on the command line yet. Move print_with_color test from
here to ‘main’. Suppress bogus GCC warning.
(getenv_quoting_style): Return the quoting style instead of
setting static storage.
(init_column_info): New arg MAX_COLS, to avoid recalculating it.
Caller changed.
* src/ls.c (dired_pos): Now off_t, not size_t, since it counts
output file offsets.
(dired_dump_obstack): This obstack's file offsets are now
off_t, not size_t.
(format_user_or_group, format_user_or_group_width):
ID arg is now uintmax_t, not unsigned long, since uid_t and
gid_t values might exceed ULONG_MAX.
(format_user_or_group_width): Use snprintf with NULL instead of
sprintf with a discarded buffer. This avoids a stack buffer,
and so should be safer.
Prefer functions or constants to macros where either will do.
That’s cleaner, and nowadays there’s no performance reason to
prefer macros. All uses changed.
* src/ls.c (INITIAL_TABLE_SIZE, MIN_COLUMN_WIDTH):
Now constants instead of macros.
(file_or_link_mode): New function, replacing the old macro
FILE_OR_LINK_MODE.
(dired_outbyte): New function, replacing the old macro DIRED_PUTCHAR.
(dired_outbuf): New function, replacing the old macro DIRED_FPUTS.
(dired_outstring): New function, replacing the old macro
DIRED_FPUTS_LITERAL.
(dired_indent): New function, replacing the old macro DIRED_INDENT.
(push_current_dired_pos): New function, replacing the old macro
PUSH_CURRENT_DIRED_POS.
(assert_matching_dev_ino): New function, replacing the old macro
ASSERT_MATCHING_DEV_INO.
(do_stat, do_lstat, stat_for_mode, stat_for_ino, fstat_for_ino)
(signal_init, signal_restore, cmp_ctime, cmp_mtime, cmp_atime)
(cmp_btime, cmp_size, cmp_name, cmp_extension)
(fileinfo_name_width, cmp_width, cmp_version):
No longer inline; compilers can deduce this well enough nowadays.
(main): Protect unused assert with ‘if (false)’ rather than
commenting it out, so that the compiler checks the code.
(print_dir): Output the space and newline in the same buffer
as the human-readable number they surround.
(dirfirst_check): New function, replacing the old macro
DIRFIRST_CHECK. Simplify by using subtraction.
(off_cmp): New function, replacing the old macro longdiff.
(print_long_format): No need to null-terminate the string now.
(format_user_or_group): Let printf count the bytes.
* src/ls.c (format_user_or_group_width, print_long_format):
Use return value from sprintf instead of calling strlen on
the resulting buffer, or inferring the length some other way.
As originally reported in <https://bugzilla.redhat.com/1962515>,
df invoked without -a printed duplicated entries for NFS mounts
of bind mounts. This is a regression from commit v8.25-54-g1c17f61ef99,
which introduced the use of a hash table.
The proposed patch makes sure that the devlist entry seen the last time
is used for comparison when eliminating duplicated mount entries. This
way it worked before introducing the hash table.
Patch co-authored by Roberto Bergantinos.
* src/ls.c (struct devlist): Introduce the seen_last pointer.
(devlist_for_dev): Return the devlist entry seen the last time if found.
(filter_mount_list): Remember the devlist entry seen the last time for
each hashed item.
* NEWS: Mention the bug fix.
Fixes https://bugs.gnu.org/49298
This fixes an unlikely stack out-of-bounds write reported by
Stepan Broz via Kamil Dudka (Bug#49209).
* bootstrap.conf (gnulib_modules): Replace select with poll.
* src/tail.c: Do not include <sys/select.h>.
[!_AIX]: Include poll.h.
(check_output_alive) [!_AIX]: Use poll instead of select.
(tail_forever_inotify): Likewise. Simplify logic, as there is no
need for a ‘while (len <= evbuf_off)’ loop.
* src/stat.c (default_format): Use decomposed decimal
representation (major,minor) in the default format.
This is least ambiguous for human interpretation,
and more consistent with ls for example.
Fixes https://bugs.gnu.org/48960
In preparation for changing the default device number
representation (to decomposed decimal), provide more
formatting options for device numbers.
These new (FreeBSD compat) formatting options are added:
%Hd major device number in decimal (st_dev)
%Ld minor device number in decimal (st_dev)
%Hr major device type in decimal (st_rdev)
%Lr minor device type in decimal (st_rdev)
%r (composed) device type in decimal (st_rdev)
%R (composed) device type in hex (st_rdev)
* doc/coreutils.texi (stat invocation): Document new formats.
* src/stat.c (print_it): Handle the new %H and %L modifiers.
(print_statfs): Adjust to passing the format as two chars
rather than an int. Using an int was introduced in commit db42ae78,
but using separate chars is cleaner and more extensible.
(print_stat): Likewise. Handle any modifiers and the new 'r' format.
(usage): Document the new formats.
* tests/misc/stat-fmt.sh: Add a test case for new modifiers.
Addresses https://bugs.gnu.org/48960
Coreutils mistakenly did not list xstrndup as a module
that it depends on directly. When the latest Gnulib removed
the dirname module's dependency on xstrndup, this mistake
caused coreutils to not build. Since all of Coreutils's
uses of xstrndup know the string length, xmemdup0 is a better
match for what's needed. Since the size args are typically
signed or derived from subtracting pointers, the new Gnulib
ximemdup0 function is a better match yet.
So, use ximemdup0 instead of xstrndup.
* src/cut.c, src/dircolors.c, src/expand-common.c, src/expand.c:
* src/numfmt.c, src/set-fields.c, src/unexpand.c:
Do not include xstrndup.h; no longer needed.
* src/dircolors.c (parse_line):
* src/expand-common.c (parse_tab_stops):
* src/numfmt.c (parse_format_string):
* src/set-fields.c (set_fields):
Use ximemdup0 instead of xstrndup.
This is now only used on 10 year old linux kernels,
and performs a sync before each copy.
* src/copy.c (extent_copy): Remove function and all callers.
* src/extent-scan.c: Remove.
* src/extent-scan.h: Remove.
* src/fiemap.h: Remove.
* src/local.mk: Adjust for removed files.
* NEWS: Adjust to say fiemap is removed.
copy_file_range() before Linux kernel release 5.3 had many issues,
as described at https://lwn.net/Articles/789527/, which was
referenced from https://lwn.net/Articles/846403/; a more general
article discussing the generality of copy_file_range().
Linux kernel 5.3 was released in September 2019, which is new enough
that we need to actively avoid older kernels.
* src/copy.c (functional_copy_file_range): A new function
that returns false for Linux kernels before version 5.3.
(sparse_copy): Call this new function to gate use of
copy_file_range().
* tests/cp/sparse-2.sh: Double check cp --sparse=always,
with dd conv=sparse, in the case where the former didn't
create a sparse file. Now that this test is being newly run
on macos, we're seeing a failure due to seek() not creating
holes on apfs unless the size is >= 16MiB.
fiemap is no longer the default copy implementation,
so check for SEEK_DATA support instead as that's preferred.
This will ensure better test coverage on systems without fiemap.
* init.cfg: Replace fiemap_capable_ with seek_data_capable_.
This is best supported with python 3 so prefer that.
* tests/seek-data-capable: A new test script checking for
SEEK_DATA support on the passed file name,
called from seek_data_capable_.
* tests/fiemap-capable: Remove no longer used probing script.
* tests/cp/fiemap-perf.sh: Renamed to tests/cp/sparse-perf.sh
* tests/cp/fiemap-2.sh: Renamed to tests/cp/sparse-2.sh
* tests/cp/fiemap-extents.sh: Renamed to tests/cp/sparse-extents.sh
* tests/cp/sparse-fiemap.sh: Renamed to tests/cp/sparse-extents-2.sh
* tests/cp/fiemap-FMR.sh: Renamed to tests/cp/copy-FMR.sh
* tests/local.mk: Reference the renamed tests.
* src/copy.c (sparse_copy): Upon EPERM from copy_file_range(),
fall back to a standard copy, which will give a more accurate
error as to whether the issue is with the source or destination.
Also this will avoid the issue where seccomp or apparmor are
not configured to handle copy_file_range(), in which case
the fall back standard copy would succeed without issue.
This specific issue with seccomp was noticed for example in:
https://github.com/golang/go/issues/40900
* src/copy.c (infer_scantype): Ensure we don't error out
if SEEK_DATA returns EOPNOTSUPP, on systems where this value
is distinct from ENOTSUP. Generally both of these should be checked.
* src/copy.c (sparse_copy): Ensure we fall back to
a standard copy if copy_file_range() returns ENOTSUP.
This generally is best checked when checking ENOSYS,
but it also seems to be a practical concern on Centos 7,
as a quick search gave https://bugzilla.redhat.com/1840284
Fixes a bits/long-double.h include build issue on some systems.
* bootstrap: Sync new --version option from gnulib.
* gnulib: Update to lastest.
Reported by Carl Edquist
The test program will compile successfully even if __get_cpuid_count
is not declared. The error for the missing symbol will only show up
at link time. Thus, use AC_LINK_IFELSE instead of AC_COMPILE_IFELSE.
* configure.ac (__get_cpuid_count check): Use C_LINK_IFELSE instead
of AC_COMPILE_IFELSE.
(__get_cpuid check): Likewise.
Ensure we call hash_free() to avoid valgrind and leak_sanitizer
"definitely lost" warnings. These were not real leaks as
we terminate immediately after, but we should avoid these
"definitely lost" warnings where possible.
* src/copy.c: Add dest_info_free() and src_info_free().
* src/copy.h: Declare the above.
* src/cp-hash.c: Don't define unless "lint" is defined.
* src/install.c: Call dest_info_free() in dev mode.
* src/mv.c: Likewise.
* src/cp.c: Likewise. Also call src_info_free().
* src/ln.c: Call hash_free() in dev mode.
* src/tail.c: Call hash_free() even if about to exit, in dev mode.
Fixes https://bugs.gnu.org/48189
* src/copy.c (sparse_copy): Don't use copy_file_range()
with --reflink=never as copy_file_range() may implicitly
use acceleration techniques like reflinking.
(extent_copy): Pass through whether we allow reflinking.
(lseek_copy): Likewise.
Fixes https://bugs.gnu.org/48164
* src/wc.c: (main): Handle the new --debug option.
Only call avx2_supported if needed.
(avx2_supported): Diagnose various failures and attempts.
* NEWS: Mention the new wc improvement and --debug option.
Use cpuid to detect CPU support for avx2 instructions.
Performance was seen to improve by 5x for a file with only newlines,
while the performance for a file with no such characters is unchanged.
* configure.ac [USE_AVX2_WC_LINECOUNT]: A new conditional,
set when __get_cpuid_count() and avx2 compiler intrinsics are supported.
* src/wc.c (avx2_supported): A new function using __get_cpuid_count()
to determine if avx2 instructions are supported.
(wc_lines): A new function refactored from wc(),
which implements the standard line counting logic,
and provides the fallback implementation for when avx2 is not supported.
* src/wc_avx2.c: A new module to implement using avx2 intrinsics.
* src/local.mk: Reference the new module. Note we build as a separate
lib so that it can be portably built with separate -mavx2 etc. flags.
Problem reported by Roland (Bug#48106).
* src/touch.c (touch): Take more care when deciding whether
to use open_errno or utime_errno in the diagnostic.
Stop worrying about SunOS 4 (which as part of the problem),
as it’s long obsolete. For Solaris 10, verify that EINVAL
really means the file was a directory.
* configure.ac: Use AC_PROG_CC, not AC_PROG_CC_STDC.
* gl/modules/smack (configure.ac):
* m4/jm-macros.m4 (coreutils_MACROS):
* m4/xattr.m4 (gl_FUNC_XATTR):
Use AS_HELP_STRING, not AC_HELP_STRING.
* m4/check-decl.m4 (gl_CHECK_DECLS):
Do not require AC_HEADER_TIME; we no longer care about it directly.
* m4/jm-macros.m4 (coreutils_MACROS):
Do not require AC_ISC_POSIX, which became obsolete in 2006.
Use AC_LINK_IFELSE instead of AC_TRY_LINK.
* src/csplit.c (load_buffer):
* src/pinky.c (create_fullname):
Use intprops-based checks rather than xalloc_oversized,
since Gnulib xalloc.h no longer includes xalloc-oversized.h.
* src/copy.c (sparse_copy): Fallback to read() if copy_file_range()
fails with ETXTBSY. Otherwise it would be impossible to copy files
that are being used as swap. This used to work before introducing
the support for copy_file_range() in coreutils. (Bug#48036)
On newer systems like Fedora 34 and openSUSE Tumbleweed, ls(1) calls
newfstatat(STDOUT_FILENO, ...), but only when there is something to
output.
* tests/ls/stat-free-color.sh: Add -a option to the reference invocation
of ls, thus enforcing something gets output.
* doc/coreutils.texi (ln invocation): State --symbolic is required.
* src/ln.c (usage): Explicitly state -s is not implied.
Fixes https://bugs.gnu.org/47703
* src/wc.c (usage): State that only printable characters are considered
when counting words. This also disambiguates wether we're talking
about bytes or characters in this context.
* doc/coreutils.texi (wc invocation): Likewise. Also clarify
that --characters counts valid locale aware characters,
and that --lines does not count a trailing "line" unless
it ends with a newline character.
Fixes https://bugs.gnu.org/47702
This is especially important now for --sort=width,
as that can greatly increase how often this
expensive quote_name_width() function is called per file.
This also helps the default invocation of ls,
or specifically the --format={across,vertical} cases
(when --width is not set to 0),
to avoid two calls to this function per file.
Note the only case where we later compute the width,
is for --format=commas. That's only done once though,
so we leave the computation close to use to
maximize hardware caching.
* src/ls.c (struct fileinfo): Add a WIDTH member to cache
the screen width of the file name.
(update_current_files_info): Set the WIDTH members for cases
they're needed multiple times. Note we do this explicitly here,
rather than caching at use, so that the fileinfo
structures can remain const in the sorting and presentation functions.
(sort_files): Call the new update_current_files_info() in this
initialization function.
(fileinfo_name_width): Renamed from fileinfo_width,
and adjusted to return the cached value if available.
This helps identify the outliers for long filenames, and also produces
a more compact display of columns when listing a directory with many
entries of various widths.
* src/ls.c (sort_type, sort_types, sort_width): New sort_width sort
type.
(sort_args): Add "width" sort arg.
(cmp_width, fileinfo_width): New sort function and helper for file name
width.
(quote_name_width): Add function prototype declaration.
(usage): Document --sort=width option.
* doc/coreutils.texi: Document --sort=width option.
* tests/ls/sort-width-option.sh: New test for --sort=width option.
* tests/local.mk: Reference new test.
* NEWS: Mention the new feature.
* bootstrap.conf (gnulib_modules): Add idx.
* src/env.c: Include idx.h, minmax.h.
Prefer idx_t to ptrdiff_t when values are nonnegative.
(valid_escape_sequence, escape_char, validate_split_str)
(CHECK_START_NEW_ARG):
Remove; no longer needed now that we validate as we go.
(struct splitbuf): New type.
(splitbuf_grow, splitbuf_append_byte, check_start_new_arg)
(splitbuf_finishup): New functions.
(build_argv): New arg ARGC. Validate and process in one go, using
the new functions; this is simpler and more reliable than the old
approach (as witness the recent bug). Avoid integer overflow in
the unlikely case where the string contains more than INT_MAX
arguments.
(parse_split_string): Simplify by exploiting the new build_argv.
The assertions didn’t help catch the most recent bug which
was in their area, and kind of get in the way.
* src/env.c: Do not include <assert.h>, and remove all assertions.
These seem to have been put in to pacify gcov, but surely there’s
a better way.
(escape_char): Pacify GCC with 'assume' instead.
Problem reported by Frank Busse (Bug#47412).
* src/env.c (C_ISSPACE_CHARS): New macro.
(shortopts, build_argv, main): Treate all C-locale space
characters like space and tab, for compatibility with FreeBSD.
(validate_split_str, build_argv, parse_split_string):
Use the C locale, not the current locale, to determine whether a
byte is a space character.
* src/cksum.c (cksum_slice8): Fix bug on little-endian
platforms lacking __bswap_32: the SWAP macro evaluates
its argument multiple times, but the macro has a side effect.
The diacrit module is obsolete, and ptx’s use of it is obsolete
too; it assumes an 8-bit locale (not that common these days) and
that TeX cannot process the 8-bit characters (nowadays, it can).
* NEWS, doc/coreutils.texi (Charset selection in ptx): Document this.
* bootstrap.conf (gnulib_modules): Remove diacrit.
* src/ptx.c: Do not include diacrit.h.
(print_field, fix_output_parameters): Remove obsolete support
for 8-bit diacritics.
This behavior was introduced in commit FILEUTILS-4_0_44-4-g519b707b4.
* src/cksum.c (cksum_slice8): Only report the overflow, and continue.
* src/cksum_pclmul.c (cksum_pclmul): Likewise.
* src/cksum.c: (main): Use getopt_long to parse options,
and handle the new --debug option.
(pclmul_supported): Diagnose various failures and attempts.
* NEWS: Mention the new option.
Use cpuid to detect CPU support for hardware instruction.
Fall back to slice by 8 algorithm if not supported.
A 500MiB file improves from 1.40s to 0.67s on an i3-2310M
* configure.ac [USE_PCLMUL_CRC32]: A new conditional,
set when __get_cpuid() and clmul compiler intrinsics are supported.
* src/cksum.c (pclmul_supported): A new function using __get_cpuid()
to determine if pclmul instructions are supported.
(cksum): A new function refactored from cksum_slice8(),
which calls pclmul_supported() and then cksum_slice8()
or cksum_pclmul() as appropriate.
* src/cksum.h: Export the crctab array for use in the new module.
* src/cksum_pclmul.c: A new module to implement using pclmul intrinsics.
* po/POTFILES.in: Reference the new cksum_pclmul module.
* src/local.mk: Likewise. Note we build it as a separate library
so that it can be portably built with separate -mavx etc. flags.
* tests/misc/cksum.sh: Add new test modes for pertinent buffer sizes.
build-aux/gen-single-binary.sh (override_single): A new function
to refactor the existing mappings for dir, vdir, and arch.
This function now also sets the DEPENDENCIES variable so that these
dependencies can be maintained later in the script, where
we now propagate the automake generated $(src_$cmd_DEPENDENCIES)
to our equivalent src_libsinglebin_$cmd_a_DEPENDENCIES.
This will ensure that any required libs are built,
which we require in a following change to cksum that
builds part of it as a separate library.
GNU/Linux is unusual here in that rmdir("symlink/") returns ENOTDIR,
whereas Solaris and FreeBSD at least, will follow the symlink
and remove the target directory. We don't make the behavior
on Linux kernels consistent, but at least clarify
the confusing error message.
* src/rmdir (main): Output a specific error message for the above case.
(remove_parents): In the error message, don't assume intermediate paths
are directories, as they could be symlinks.
* tests/rmdir/symlink-errors.sh: Add a new test.
* tests/local.mk: Reference the new test.
* NEWS: Mention the improvement.
This regression was introduced in commit COREUTILS-6_8-58-g553d347d3
* src/pr.c (init_parameters): Process tabs for multiple columns.
* tests/pr/pr-tests.pl: Add test cases.
* NEWS: Mention the bug fix.
Fixes https://bugs.gnu.org/46422
- \r\n is common a line end combination
- catting such a file without options causes it to display normally
- overwriting the first char with $, loses info
* src/cat.c (cat): Convert \r preceeding a \n to ^M.
* tests/misc/cat-E.sh: New test.
* tests/local.mk: Reference new test.
* tests/misc/cat-proc.sh: Fix typo.
* doc/coreutils.texi (cat invocation): Mention the new behavior.
* NEWS: Mention the improvement.
This functionality regressed with the adjustments
in commit v8.25-4-g62e7af032
* src/split.c (bytes_chunk_extract): Account for already read data
when seeking into the file.
* tests/split/b-chunk.sh: Use the hidden ---io-blksize option,
to test this functionality.
* NEWS: Mention the bug fix.
Fixes https://bugs.gnu.org/46048
Problem reported by David McCall (Bug#45886).
I introduced this problem when fixing Bug#14371.
* NEWS: Mention the fix.
* src/mkdir.c (struct mkdir_options): New members umask_ancestor,
umask_self, replacing umask_value.
(make_ancestor): Use them when temporarily adjusting umask.
(main): Set them, and set the umask to umask_self instead
of leaving it alone.
* tests/mkdir/perm.sh (tests): Add test case for bug.
* doc/coreutils.texi: Fix regexp cross-reference that had become
out-of-date (Bug#45749). Also, fix some obsolete references to
SunOS and to /usr/dict/words, and change “Linux” to “GNU/Linux”
where appropriate. Unfortunately the pipeline example gets more
complicated since /usr/share/dict/words is not sorted the way that
‘comm’ wants.
None of the coreutils man pages - but the two above - are using bold
setting for the references to other man pages in the SEE ALSO section.
* man/cat.x (SEE ALSO): Remove '\fB...\fP' setting.
* man/tac.x: Likewise, and add a reference to cat(1).
This file is a copy from gnulib and therefore should not get changed
by the yearly update.
* .x-update-copyright: Add pattern for the above file.
* doc/fdl.texi: Revert the previous change.
Run "make update-copyright" and then...
* gnulib: Update to latest with copyright year adjusted.
* tests/init.sh: Sync with gnulib to pick up copyright year.
* bootstrap: Likewise.
* tests/sample-test: Adjust to use the single most recent year.
A 100MB file improves from 2.50s to 1.80s on a Sparc T5220
A 100MB file improves from 0.54s to 0.13s on an i3-2310M
* bootstrap.conf: Explicitly depend on byteswap,
since now used directly by coreutils.
* src/cksum.c (cksum): Process in multiples of 8 bytes.
(main): Adjust for generation of expanded crctab.
* src/cksum.h: Split now larger crctab to separate header.
* src/local.mk: Reference the new header.
* NEWS: Mention the improvement.
* doc/coreutils.texi (seq invocation): Mention "inf" is supported,
and describe that it's handled specially to generate infinite
whole integer sequences. Also mention that such infinite generation
is supported for integer steps up to 200.
(sleep invocation): Give `sleep inf` as an example to sleep forever.
* src/seq.c: Add a comment on SEQ_FAST_STEP_LIMIT, to say it's
reflected in the texinfo description.
Chris Colohan wrote that the man page did not do enough to dispel
a common misunderstanding that “contributed to one of the scariest
outages Google has ever seen” (Bug#45258).
* doc/coreutils.texi (mkdir invocation):
* src/mkdir.c (usage): Document -m vs -p better.
* src/nl.c (main): Enforce the POSIX specified
behavior of assuming ':' is specified after a single
character argument to -d.
* tests/misc/nl.sh: Add a test case.
* NEWS: Mention the bug fix.
* doc/coreutils.texi (nl invocation): Mention the GNU extensions
of allowing arbitrary length and empty delimiter strings.
* src/nl.c (usage): Likewise.
* tests/misc/nl.sh: Add test cases for the GNU extensions.
* src/nl.c (main): Update the default delimiter characters
when passed two characters with --section-delimiter.
Avoid redundant copies for the body and footer delimiter strings,
and instead, just offset into the header string.
(check_section): Avoid redundant comparing of 2 bytes of memory
for an empty delimiter.
The format can be determined from --options or the locale,
so it's useful to output the format string being used.
* src/date.c (show_date): Show the output format
along with the date being shown.
* tests/misc/date-debug.sh: Adjust accordingly.
Addresses https://bugs.gnu.org/44960
When removing a directory fails for some reason, and that directory
is empty, the rm_fts code gets the return value of the excise call
confused with the return value of its earlier call to prompt,
causing fts_skip_tree to be called again and the next file
that rm would otherwise have deleted to survive.
* src/remove.c (rm_fts): Ensure we only skip a single fts entry,
when processing empty dirs. I.e. only skip the entry
having successfully removed it.
* tests/rm/empty-immutable-skip.sh: New root-only test.
* tests/local.mk: Add it.
* NEWS: Mention the bug fix.
Fixes https://bugs.gnu.org/44883
This was needed before libselinux-2.3 (May 2014),
but modern releases have the correct const declarations.
* src/chcon.c: Remove se_const() wrapper.
* src/cp.c: Likewise.
* src/install.c: Likewise.
* src/mkdir.c: Likewise.
* src/mkfifo.c: Likewise.
* src/mknod.c: Likewise.
* src/system.h: Likewise.
* gnulib: update to pick up const correctness fixes in selinux stubs.
At least, I *think* they are false alarms. An SELinux expert eye
would be welcome.
* src/install.c (setdefaultfilecon): If selabel_lookup fails
due to either ENOTSUP or ENODATA, don’t diagnose the issue.
Problem reported by Kamil Dudka in:
https://lists.gnu.org/r/coreutils/2020-11/msg00050.html
* src/selinux.c: Don’t include die.h; no longer needed.
(computecon, defaultcon, restorecon): Propagate errno.
(defaultcon, restorecon): Don’t diagnose errors or exit, as that’s
the caller’s responsibility.
* src/selinux.c: selabel_lookup requires absolute paths
(while only older matchpathcon before libselinux < 2.1.5 2011-0826 did).
* po/POTFILES.in: Readd src/selinux.c since we now have
a translatable error message.
The previous commit introduced a couple of syntax-check failures.
* .gitignore (/lib/se-label.h): Add entry to silence the
sc_gitignore_missing check. Sort entries in C locale.
* po/POTFILES.in (src/selinux.c): Remove entry as this source doesn't
contain any translatable strings anymore; avoids a sc_po_check failure.
* src/mv.c: Replace tabs by spaces to avoid complaints by
sc_prohibit_tab_based_indentation.
Ubuntu 20.10 is using a newer version of libselinux that
complains that matchpathcon is obsolete. Rewrite the code
that it uses the recommended selabel_lookup instead.
* m4/jm-macros.m4 (coreutils_MACROS): Do not check for
matchpathcon_init_prefix, as it is no longer used.
* src/copy.c (set_file_security_ctx): Omit process_local arg,
as it is equivalent to !x->set_security_context. All callers changed.
* src/copy.h (struct cp_options): set_security_context is now of
type struct selabel_handle *, not bool. All uses changed.
* src/cp.c, src/install.c, src/mkdir.c, src/mkfifo.c, src/mknod.c:
* src/mv.c: Include selinux/label.h.
(main): Use selabel_open for set_security context.
* src/install.c (matchpathcon_init_prefix): Remove; now unused.
(get_labeling_handle): New static function.
(setdefaultfilecon, main): Use it.
(setdefaultfilecon): Do something regardless of
ENABLE_MATCHPATHCON, which seems to be a revenant macro.
(setdefaultfilecon): Use selabel_lookup instead of the obsolescent
matchpathcon. Report an error unless it fails due to ENOENT.
* src/local.mk (src_ginstall_CPPFLAGS): Remove.
* src/selinux.c: Include selinux/label.h
Do not include die.h, error.h, canonicalize.h.
(defaultcon, restorecon_private, restorecon):
New arg HANDLE. All callers changed.
Use selabel_lookup rather than matchpathcon.
(restorecon_private, restorecon): Don’t lose track of errno.
* src/selinux.c, src/selinux.h:
(restorecon): Don’t call ‘error’; that’s the caller’s job.
Use HAVE_SELINUX_LABEL_H, not HAVE_SELINUX_SELINUX_H,
in case there is some weird system with the former but not the latter.
* src/selinux.h (struct selinux_handle): Add forward decl.
* src/local.mk (src_ln_LDADD, src_mktemp_LDADD, src_tac_LDADD):
Add $(LIB_CLOCK_GETTIME), since these use tempname which uses
clock_gettime if getrandom fails. On platforms like Solaris 10,
clock_gettime is not in the standard C library.
* doc/coreutils.texi (sort invocation): Mention explicitly
that --general-numeric-sort supports arbitrary format hex numbers,
but also mention that consistent case/width hex numbers can
be sorted faster with a standard sort.
This crash was identified by Cyber Independent Testing Lab:
https://cyber-itl.org/2020/10/28/citl-7000-defects.html
and was introduced with commit v8.5-163-g3f48829c2
* src/tr.c (validate_case_classes): Don't apply these
extra case alignment checks in the --complement case,
which is even more restrictive as to the contents of SET2.
* tests/misc/tr-case-class.sh: Add a test case,
for a large SET1, which caused the length adjustment
in validate_case_classes to underflow and trigger the assert.
* NEWS: Mention the bug fix.
This crash was identified by Cyber Independent Testing Lab:
https://cyber-itl.org/2020/10/28/citl-7000-defects.html
and was introduced with commit v6.9.90-11-g4245876e2
* src/ls.c (gobble_file): Ensure scontext is initialized
in the case where files are not statable.
* tests/ls/selinux-segfault.sh: Renamed from proc-selinux-segfault.sh,
and added test case for broken symlinks.
* tests/local.mk: Adjust for the renamed test.
* NEWS: Mention the bug fix.
GCC 10.1.1 without optimization gives:
error: ‘strncat’ argument 2 declared attribute ‘nonstring’
[-Werror=stringop-overflow=]
strncat (comment, UT_ID (utmp_ent), utmpsize);
Note the strncat man page says that:
"src does not need to be null-terminated
if it contains n or more bytes."
And the POSIX spec says that the second (source) parameter
is an array not a string.
So I think it's incorrect for strncat to require src be a string type.
This constraint seems to be being added to the gcc builtin strncat,
as specifiying -fno-builtin also avoids the warning.
Note specifying any optimization level also avoids the warning.
* src/who.c (make_id_equals_comment): Avoid the issue by using
stpcpy + stzncpy, instead of strcpy + strncat.
This pattern is used elsewhere in who.c
* src/local.mk: Ensure we map 2 hex digits to 4,
so that we don't output already handled Z3FOLD file system (0x33).
Also hide the generation command for src/fs.h.
* src/basenc.c (usage): Remove extraneous blank line,
to be consistent with other utilities that have options.
* src/realpath.c: Likewise.
* src/runcon.c: Likewise.
Addresses https://bugs.gnu.org/44248
gcc version 11.0.0 20201025 (experimental) warns that
src/sort.c:1655:1: warning: function might be candidate for attribute \
'pure' if it is known to return normally [-Wsuggest-attribute=pure]
* src/sort.c (limfield): Mark as pure.
The workaround triggers warnings from newer kernel versions in case
a user does not have sufficient privileges for the MTIOCGET ioctl.
* src/dd.c (skip_via_lseek): Drop wrapper function no longer needed.
(skip): Use lseek() directly.
(advance_input_after_read_error): Likewise.
Reported-by: Nir Soffer at https://bugzilla.redhat.com/1876840
Fixes https://bugs.gnu.org/44235
Previously we would have failed immediately upon internal overflow,
which didn't output the full line being processed, and assumed
there would be another numbered line.
* src/nl.c (line_no_overflow): A new global to track overflow.
(print_lineno): Only fail if about to output an overflowed number.
(reset_lineno): A new function to refactor resetting of the number,
and which also clears line_no_overflow.
* tests/misc/nl.sh: Add a test case.
* man/help2man: sync to changes from version 1.47.16.
Note this doesn't materially change the generated man pages.
Addresses https://bugs.gnu.org/44105
* gl/lib/randperm.c, src/cp-hash.c, src/ls.c, src/sort.c, src/tail.c:
Change all instaces of hash_delete to hash_remove to accommodate
change to Gnulib API.
* src/tail.c: Remove FIXME to follow a file name in a recreated
directory. The comment was added in commit v8.5-191-g61b77891c
while the fix (albeit not using inotify) was added in
commit v8.27-21-gba5fe2d4b
* src/stat.c (usage): Replace a mistaken semicolon with a colon,
and replace mistaken backticks with single quotes. Also reorder
some words, for clarity.
Fixes https://bugs.gnu.org/43707
* doc/coreutils.texi (timeout invocation): Avoid any implication
that `timeout --foreground` could be used to retroactively
timeout commands not already invoked by timeout(1).
Fixes bug https://bugs.gnu.org/42831
* src/csplit.c (process_regexp): Process the line suppression
in all invocations so that the last match is suppressed.
Previously with a non infinite match count,
the last regex pattern was not suppressed.
* NEWS: Mention the bug fix.
* tests/misc/csplit-suppress-matched.pl: Add a test case.
Fixes https://bugs.gnu.org/42764
Parts of this test expect that the rmdir syscall returns with EPERM,
but the root user does not see that.
* tests/rmdir/ignore.sh: Add uid_is_privileged_ guards around parts
of the test which expect rmdir() to fail with EPERM.
Reported by Nick Alcock <nix@esperi.org.uk> in
https://bugs.gnu.org/42633
* doc/coreutils.texi (timeout invocation): Document that the the
duration of --kill-after=DURATION begins when sending the initial
signal. Also mention that -k does not have any effect if timeout's
duration is 0.
Suggested by Jonny Grant <jg@jguk.org>.
* src/factor.c (mp_factor_using_division): Use mpz_fdiv_q_2exp
instead of its no-longer-documented mpz_div_2exp alias.
(print_factors): Use mpz_out_str instead of gmp_printf.
* src/factor.c (strto2uintmax): Instead of here ...
(print_factors): ... skip spaces and '+' here, so that
bignums are treated like non-bignums.
* tests/misc/factor.pl (bug-gmp-plus_2_sup_128_plus_1): New test.
This lets use assume multiple-precision arithmetic on all
platforms, simplifying the code.
* bootstrap.conf (gnulib_modules): Add libgmp.
* configure.ac: Don’t call cu_GMP, as this is now done by Gnulib.
* m4/gmp.m4: Remove.
* src/expr.c, src/factor.c: Use gmp.h unconditionally.
* src/factor.c: Use the simpler ‘#ifndef mpz_inits’ to
determine whether there is an mpz_inits macro.
* tests/cp/fiemap-FMR.sh: Avoid FICLONE ioctl,
which would avoid the point of the test (fiemap testing).
Also it avoids a valgrind bug with this ioctl:
https://bugs.kde.org/show_bug.cgi?id=397605
If it works, prefer lseek with SEEK_DATA and SEEK_HOLE to FIEMAP,
as lseek is simpler and more portable (will be in next POSIX).
Problem reported in 2011 by Jeff Liu (Bug#8061).
* NEWS: Mention this.
* src/copy.c (lseek_copy) [SEEK_HOLE]: New function.
(enum scantype): New constants ERROR_SCANTYPE, LSEEK_SCANTYPE.
(union scan_inference): New type.
(infer_scantype): Last arg is now union scan_inference *,
not struct extent_scan *. All callers changed.
Prefer SEEK_HOLE to FIEMAP if both work, since
SEEK_HOLE is simpler and more portable.
(copy_reg): Do the fdadvise after initial scan, in case the scan
fails. Report an error if the initial scan fails.
(copy_reg) [SEEK_HOLE]: Use lseek_copy if scantype says so.
* src/copy.c (extent_copy): New arg SCAN, replacing
REQUIRE_NORMAL_COPY. All callers changed.
(enum scantype): New type.
(infer_scantype): Rename from is_probably_sparse and return
the new type. Add args FD and SCAN. All callers changed.
* src/stat.c (human_fstype): Add case for the 'vboxsf' file system type
which is used for VirtualBox Shared Folders mounted in VirtualBox guest
VMs.
* NEWS: Mention the Improvement.
Fixes https://bugs.gnu.org/41935
Since -LONG_MIN results in LONG_MIN again, the operation itself is
a signed integer overflow.
This can be observed with the following calls (best if compiled
with -ftrapv or -fsanitize=undefined):
$ numfmt --padding=-9223372036854775808
$ seq 1e-9223372036854775808
Technically, the change in seq "reduces" the precision, but a double
or long double that small would be represented as 0 anyway.
* src/numfmt.c: Explicitly disallow --padding=LONG_MIN.
* src/seq.c: Treat 1e$LONG_MIN as 1e-$LONG_MAX.
* tests/misc/numfmt.pl: Add a test case.
* tests/misc/seq-precision.sh: Likewise.
Fixes https://bugs.gnu.org/41850
* doc/coreutils.texi (timeout invocation): Document that the exit
status is 137 when the KILL signal is used, regardless of whether that
signal is sent to COMMAND or timeout.
* src/timeout.c (usage): Likewise. Also split out and expand
on the possible exit status values to a separate table.
Discussed at https://bugs.gnu.org/41634
This makes for one Gnulib module less, and at runtime there’s
typically just one getrandom syscall instead of several for large
nonces.
* gl/lib/randread.c: Include sys/random.h instead of sys/time.h
and unistd.h.
(get_nonce): Use getrandom, not getentropy.
* gl/modules/randread (Depends-on):
Depend on getrandom, not getentropy.
* src/shred.c (main):
* src/shuf.c (main):
* src/sort.c (random_md5_state_init):
Say "getrandom" rather than "getentropy" in (unlikely) diagnostic.
Update gnulib submodule to latest and use its new features.
Gnulib’s new getentropy module means coreutils can now assume
getentropy instead of approximating it, badly in some cases.
Gnulib’s improvements to the tempname module mean coreutils no
longer needs to maintain private patches.
* bootstrap.conf (gnulib_modules): Remove gettimeofday.
* gl/lib/randread.c (NAME_OF_NONCE_DEVICE): Remove.
(get_nonce): Return success indicator. Remove bytes_bound arg.
All callers changed. Rewrite by using getentropy instead of
reading the nonce device and falling back on gettimeofday.
Fail if getentropy fails.
(randread_new): Return NULL (setting errno) if get_nonce fails.
All callers changed.
* gl/lib/tempname.c.diff, gl/lib/tempname.h.diff:
* gl/modules/tempname.diff: Remove.
* gl/modules/randread (Depends-on):
Depend on getentropy, not gettimeofday.
* src/ptx.c (swallow_file_in_memory):
* src/shuf.c (read_input):
Adjust to read_file changes in Gnulib.
* src/shred.c (main):
* src/shuf.c (main):
* src/sort.c (random_md5_state_init):
Diagnose the new form of randread_new failures: randread_new can
fail now when !random_source, meaning getentropy failed.
Since the previous gnulib update, bootstrap outputs this warning:
Notice from module fdl:
Don't use this module! Instead, copy the referenced license file \
into your version control repository.
See gnulib commit:
https://git.sv.gnu.org/cgit/gnulib.git/commit/?id=88fc5afbccc9
* bootstrap.conf (gnulib_modules): Remove 'fdl'.
* doc/fdl.texi: Add file as a copy of 'gnulib/doc/fdl.texi'.
* doc/.gitignore (/fdl.texi): Remove entry.
* cfg.mk (FILTER_LONG_LINES): Add pattern for the 'fdl.texi' file.
The previous attempt to skip that test on NFS (commit 4181fc5183)
made the test fail; it introduced two problems:
a) In the good case, i.e., when the subshell returns with exit status 0,
the test ran into framework_failure_.
b) As the subshell also runs with 'set -x', the later comparison of
/dev/null with 'err' would fail.
* tests/ls/removed-directory.sh: Revert to the style without subshell,
and add 'test -d .' to verify that 'ls' can read the removed dir.
* src/chown-core.c, src/comm.c:
* src/tsort.c (record_relation):
Remove GCC 10 pragmas that are not needed in GCC 10.1.0 (the first
public GCC 10 release) and that in some cases cause diagnostics
with GCC 10.1.0. The tsort.c change fixes a bug that was
inadvertantly introduced when these pragmas were added.
* src/env.c (build_argv): Add an assert() to avoid:
warning: use of NULL 'n' where non-null expected
[CWE-690] [-Wanalyzer-null-argument]
note: argument 1 of 'getenv' must be non-null
* src/dd.c (alloc_ibuf): Don't discard the allocated pointer, to avoid:
[CWE-401] [-Wanalyzer-malloc-leak]
(alloc_obuf): Likewise.
(cleanup): Deallocate the now tracked buffers which
also avoids "possibly lost" warnings from valgrind.
* src/tsort.c (search_item): Add asserts to avoid:
[CWE-690] [-Wanalyzer-null-dereference]
(record_relation): An assert doesn't suffice here,
so disable the warning for this function.
* src/comm.c: Suppress the following false positive for the whole file:
[CWE-457] [-Wanalyzer-use-of-uninitialized-value]
* src/chown-core.c: Suppress the following false positive for the file:
[CWE-415] [-Wanalyzer-double-free]
Have the `ls` `--classify` option take an optional argument for when to
classify ("always", "auto", "never"), just like the optional argument
for `--color`. When the optional argument is not specified, default to
"always" for backwards compatibility.
* src/ls.c (usage): Update help text.
(decode_switches): Support an optional argument for --classify.
* tests/ls/classify.sh: Add a new test.
* tests/local.mk: Reference the new test.
* NEWS: Mention the new feature.
Pull in a fix for FTS to avoid a crash when traversing a heavily
changed XFS file system:
> fts: remove NOSTAT_LEAF_OPTIMIZATION
* NEWS (Bug fixes): Mention the fix.
* gnulib: Update to latest.
* bootstrap: Sync from gnulib/build-aux/bootstrap.
Discussed at:
<https://lists.gnu.org/r/bug-gnulib/2020-04/msg00068.html>
* tests/ls/removed-directory.sh: Remove host_triplet test.
Skip this test if one cannot remove the working directory.
From a suggestion by Bernhard Voelker (Bug#39929).
* NEWS: Mention this.
* src/ls.c: Do not include <sys/sycall.h>
(print_dir): Don't worry about whether the directory is removed.
* tests/ls/removed-directory.sh: Adjust to match new (i.e., old)
behavior.
* tests/misc/env-S.pl: `env -i env` will call the system env
due to the path being cleared, so pass the absolute path
of our env binary under test to avoid that. This was seen
to be an issue on Guix where /usr/bin/env was not available.
* src/basenc.c (z85_decode_ctx_init): Ensure we're working
with unsigned, as otherwise ubsan triggers with:
src/basenc.c:767:18: runtime error: signed integer overflow:
43 * 52200625 cannot be represented in type 'int'
(z85_encode): Likewise to avoid the usban error:
src/basenc.c:630:26: runtime error:
left shift of 134 by 24 places cannot be represented in type 'int'
* tests/misc/timeout-parameters.sh: Split the large timeout
handling to ...
* tests/misc/timeout-large-parameters.sh: ... here, so that
the 3 second delay is contained in its own test, and if
the test is skipped due invalid handling within timeout(1),
it will be more apparent.
Also adjust the check so we skip whenever the kernel timer
fires immediately, to handle the buggy OpenIndiana 11 kernel also.
Reported by Bruno Haible.
* init.cfg (require_bash_as_SHELL_): A new function to replace
SHELL for the current test, with bash if available.
This is useful on OpenIndiana 11 where /bin/sh was seen
to have races in handling of SIGPIPE.
* tests/misc/seq-epipe.sh: Use the new function to enforce bash.
* tests/misc/env-signal-handler.sh: Likewise.
Reported by Bruno Haible
* tests/ls/stat-free-color.sh: Check for the availability
of various stat calls individually, and add statx() and fstatat64()
to the list to check. Fix the stat counting logic to
ignore lines like "+++ exited with 0 +++".
* tests/ls/stat-free-symlinks.sh: Check syscalls other than stat().
* init.cfg (gcc_shared_libs_): New variable.
(gcc_shared_): Use it, instead of hardcoding -ldl.
(require_gcc_shared_): Determine the suitable value
for gcc_shared_libs_.
With these adjustments, all tests pass on macOS Catalina.
* tests/dd/sparse.sh: Adjust so that systems like apfs that
don't create holes < 16 MiB do not fail erroneously.
* tests/touch/trailing-slash.sh: Darwin was seen to dereference
symlinks to files when given a trailing slash, so avoid
that particular case.
* configure.ac: Reenable distribution of gzip-compressed
tarballs, for Guix bootstrapping reasons as discussed at:
https://lists.gnu.org/r/coreutils/2020-02/msg00042.html
* THANKS.in: Remove me, as now a committer.
* NEWS (Build-related): Mention this.
If the current directory has been removed, then "ls" confusingly
produced no output and no error message, indistinguishable from
running on an empty directory.
* src/ls.c (print_dir): Report ENOENT on GNU/Linux if readdir
finds no directory entries at all, not even "." or "..",
and a recheck with the getdents syscall returns ENOENT.
We recheck with getdents() as POSIX states that
"The directory entries for dot and dot-dot are optional".
* tests/ls/removed-directory.sh: New file.
* tests/local.mk (all_tests): Add new test.
* NEWS: Mention the change in behavior.
Reported by Owen Thomas.
* src/blake2/blake2-impl.h: Sync load16() implementation,
which doesn't change code generation.
Also leverage (builtin) memcpy to more efficiently
move data on little endian systems,
giving a 2% win with GCC 9.2.1 on an i3-2310M.
* src/longlong.h: Sync changes from:
https://gmplib.org/repo/gmp/log/tip/longlong.h
mips64: Provide r6 asm code as default expression yields.
arm32: Define sub_ddmmss separately for non-thumb (no rsc instruction).
powerpc: Add "CLOBBER" descriptions for some registers.
x86: Fix criterion for when to use mulx in umul_ppmm.
strcoll() is only significant to uniq(1) if it returns 0,
and it generally only does so with buggy locales or mismatched
locales and data. Some systems may have strcoll()
return 0 for equivalent normalized unicode forms,
but for consistency across platforms strcoll() is avoided.
The various cases are defined in the new test.
This is consistent with newer POSIX standards as discussed at:
https://www.austingroupbugs.net/view.php?id=963
* src/uniq.c: s/xstrcoll/memcmp/.
* tests/local.mk: Reference the new test.
* tests/misc/uniq-collate.sh: Add a new test.
* NEWS: Mention the change in behavior.
Fixes https://bugs.gnu.org/38627
* src/stat.c (usage): Mention permission bits rather than
"access" so there is no confusion with ACLs etc.
Also indicate we output the file type with '%A'.
* doc/coreutils.texi (stat invocation): Likewise.
Also indicate '%A' is similar to `ls -ld` output.
Addresses https://bugs.gnu.org/39613
Pick up recent build fixes to avoid sysctl.h inclusion on glibc systems,
restrict the max file size supported by read-file to PTRDIFF_MAX,
and to avoid a -Werror=unused failure in test-canonicalize.
* tests/cp/proc-short-read.sh: Switch to using /proc/cpuinfo,
rather than /proc/kallsyms which was seen to vary in some cases.
Fixes https://bugs.gnu.org/39357
Since v6.10-21-ged5c4e7 `rmdir --ignore-fail-on-non-empty`
had reversed the failure status for directories that failed
to be removed for permissions reasons. I.E. it would have
returned a failure status for such non empty dirs, and vice versa.
* src/rmdir.c (errno_may_be_non_empty): Rename from the
more confusing errno_may_be_empty(), and remove the EEXIST
case (specific to Solaris), which is moot here since
handled in errno_rmdir_non_empty().
(ignorable_failure): Fix the logic error so that
_non_ empty dirs are deemed to have ignorable failures.
(main): Fix clobbering of errno by is_empty_dir().
(remove_parents): Likewise.
* tests/rmdir/ignore.sh: Add a test case.
* THANKS.in: Add reporter who fixed the errno handling.
* NEWS: Mention the bug fix.
Fixes https://bugs.gnu.org/39364
* configure.ac: Add -Wno-vector-operation-performance to suppress the
following gcc-9.2 error in gl/lib/randperm.c:
error: vector operation will be expanded piecewise
* src/ls.c (usage): Reorganize help for --time,
and add description for --time=birth.
(do_statx): Store btime in mtime if available.
(get_stat_btime): A new function to read the creation time
from the appropriate stat structure member.
(cmp_btime): A new function to compare birth time.
(print_long_format): Output '?' when birth time unavailable.
* doc/coreutils.texi: Document --time={birth,creation}.
* tests/local.mk: Reference the new test.
* tests/ls/birthtime.sh: Add a new test.
* NEWS: Mention the new feature.
* src/yes.c (main): Convert for loop to do-while in order to indicate
that the loop will be run at least once.
This avoids the following warning after the second loop:
src/yes.c:110:20: error: writing 1 byte into a region of size 0
* configure.ac: Set --with-openssl=auto-gpl-compat as the default,
so that openssl is used for md5sum etc., with openssl >= 3,
which is newly licensed under ASL v2.
* gnulib: Update to include "auto-gpl-compat" support.
Run "make update-copyright" and then...
* gnulib: Update to latest with copyright year adjusted.
* tests/init.sh: Sync with gnulib to pick up copyright year.
* bootstrap: Likewise.
* tests/sample-test: Adjust to use the single most recent year.
* doc/coreutils.texi (sleep invocation): Add an example to demonstrate
how to use the floating-point and the scientific notation to sleep
for sub-second times, e.g. milli-, micro- and nanoseconds.
Inspired by Stephane Chazelas in:
https://lists.gnu.org/r/coreutils/2019-12/msg00005.html
* src/chcon.c (main): Skip call of security_check_context()
in case SELinux is disabled to avoid unnecessary failure.
Bug: https://bugzilla.redhat.com/1777831
* doc/sort-version.texi (Minus/Hyphen and Colon characters):
Rename from “Minus/Hyphen @samp{-} and Colon @samp{:} characters”,
as texi2any 6.6 complains about colons in node names.
* doc/coreutils.texi (shred invocation):
Modernize discussion to today’s technology (Bug#38168).
* src/shred.c (usage): Omit lengthy duplication of the manual’s
discussion of file systems and storage devices, as that became out
of sync with the manual. Instead, just cite the manual.
This addresses a longstanding "update all callers" FIXME in
lib/xstrtol.c, by having programs check that numbers do not
have unknown suffixes. The problem was also reported for
'shuf' by my student Maggie Huang while reimplementing a shuf
subset in Python as an exercise in UCLA Computer Science 35L:
https://web.cs.ucla.edu/classes/fall19/cs35L/assign/assign3.html
This patch also improves the portability of the code to unusual
platforms where ULONG_MAX < SIZE_MAX.
* NEWS: Mention user-visible changes.
* src/chgrp.c (parse_group):
* src/chroot.c (parse_additional_groups):
* src/du.c (main):
* src/install.c (get_ids):
* src/join.c (string_to_join_field):
* src/ls.c (decode_switches):
* src/md5sum.c (split_3):
* src/shuf.c (main):
* src/sort.c (specify_nthreads):
* src/uniq.c (size_opt, main):
Use uintmax_t instead of unsigned long, for portability
to oddball platforms where unsigned long is not wide enough.
* src/du.c (main):
* src/expr.c (mpz_init_set_str) [!HAVE_GMP]:
* src/install.c (get_ids):
* src/ls.c (decode_switches):
* src/mknod.c (main):
* src/ptx.c (main):
* src/shuf.c (main):
* src/sort.c (specify_nmerge, specify_nthreads):
Reject numbers with suffixes.
* src/md5sum.c (split_3): Simplify.
* gl/lib/randperm.c: Include randperm.h first, since it’s the API.
Include stdint.h, count-leading-zeros.h, verify.h.
(floor_lg): Rename from ceil_log (which was not actually
implementing the ceiling!) and implement the floor using
count_leading_zeros.
(randperm_bound): Use floor_lg, not ceil_log. Use uintmax_t
instead of size_t in case the size gets large on a 32-bit host.
* gl/modules/randperm (Depends-on): Add count-leading-zeros, stdint.
* configure.ac: When --enable-gcc-warnings is used, omit
-Wno-type-limits. The need for -Wno-type-limits has passed, now
that intprops.h uses builtin primitives for GCC 5 and later, given
that recent GCCs issue type-limits warnings only for non-constant
expressions. --enable-gcc-warnings is not intended for use with
old compilers, so we can drop -Wno-type-limits now.
‘shuf -r -n 0 file’ would mistakenly read from standard input.
Problem reported by my student Jingnong Qu while reimplementing a
shuf subset in Python as an exercise in UCLA Computer Science 35L:
https://web.cs.ucla.edu/classes/fall19/cs35L/assign/assign3.html
* NEWS: Mention the fix. Also, ASCIIfy a previous item.
* src/shuf.c (main): Fix bug.
* tests/misc/shuf.sh: Add a test case for the bug.
statx allows ls to indicate interest in only certain inode metadata.
This is potentially a win on networked/clustered/distributed
file systems. In cases where we'd have to do a full, heavyweight stat()
call we can now do a much lighter statx() call.
As a real-world example, consider a file system like CephFS where one
client is actively writing to a file and another client does an
ls --color in the same directory. --color means that we need to fetch
the mode of the file.
Doing that with a stat() call means that we have to fetch the size and
mtime in addition to the mode. The MDS in that situation will have to
revoke caps in order to ensure that it has up-to-date values to report,
which disrupts the writer.
This has a measurable affect on performance. I ran a fio sequential
write test on one cephfs client and had a second client do "ls --color"
in a tight loop on the directory that held the file:
Baseline -- no activity on the second client:
WRITE: bw=76.7MiB/s (80.4MB/s), 76.7MiB/s-76.7MiB/s (80.4MB/s-80.4MB/s),
io=4600MiB (4824MB), run=60016-60016msec
Without this patch series, we see a noticable performance hit:
WRITE: bw=70.4MiB/s (73.9MB/s), 70.4MiB/s-70.4MiB/s (73.9MB/s-73.9MB/s),
io=4228MiB (4433MB), run=60012-60012msec
With this patch series, we gain most of that ground back:
WRITE: bw=75.9MiB/s (79.6MB/s), 75.9MiB/s-75.9MiB/s (79.6MB/s-79.6MB/s),
io=4555MiB (4776MB), run=60019-60019msec
* src/stat.c: move statx to stat struct conversion to new header...
* src/statx.h: ...here.
* src/ls.c: Add wrapper functions for stat/lstat/fstat calls,
and add variants for when we are only interested in specific info.
Add statx-enabled functions and set the request mask based on the
output format and what values are needed.
* NEWS: Mention the Improvement.
* src/truncate.c (do_ftruncate): Simplify overflow checking,
and don’t rely on theoretically-nonportable assumptions
like assuming that OFF_MAX < UINTMAX_MAX.
* src/seq.c: (seq_fast): Accept STEP as a parameter and use that
to skip the output of generated numbers.
(main): Relax to using seq_fast for integer steps between 1 and 200.
For larger steps the throughput was faster using the standard
incrementing procedure.
(cmp): Use the equivalent but faster memcmp for equal len strings.
* tests/misc/seq.pl: Update fast path cases.
Addresses https://bugs.gnu.org/37241
The previous commit introduced a bug into the following syntax-check,
and thus effectively turned it off:
$ make sc_prohibit_test_calls_print_ver_with_irrelevant_argument; \
echo $?
prohibit_test_calls_print_ver_with_irrelevant_argument
fatal: cannot change to 'grep': No such file or directory
0
* cfg.mk (sc_prohibit_test_calls_print_ver_with_irrelevant_argument):
Remove changing directory, and pass $(srcdir) as argument to 'git -C'.
* cfg.mk (sc_prohibit_colon_redirection): Don't expect `|` to denote
the pipe character in git grep.
(sc_tests_executable)
(sc_case_insensitive_file_names)
(sc_some_programs_must_avoid_exit_failure)
(sc_prohibit_test_background_without_cleanup_)
(sc_prohibit_test_calls_print_ver_with_irrelevant_argument)
(sc_prohibit_test_ulimit_without_require_)
(sc_prohibit_test_background_without_cleanup_)
(sc_THANKS_in_duplicates)
*sc_prohibit_test_calls_print_ver_with_irrelevant_argument):
Don't expect builddir to be a descendant of srcdir.
(sc_strftime_check): Don't check file size against 0 when "N\nq\n" was
already put in the file.
* THANKS.in: Remove me.
Under certain circumstances seq prints an extra line when the output
format has custom format with characters following the printed numbers:
$ seq -f "%g " 1000000 1000000
1e+06
1e+06
This is due to the "print_extra_number" logic using strings to determine
whether a 'extra number' is needed, but only one string was trimmed
when using a custom printf format.
Prompted by https://lists.gnu.org/r/coreutils/2019-08/msg00001.html
* NEWS: Mention fix.
* src/seq.c (print_numbers): Trim the 'x0_str' string before comparing
it to the previous 'x_str' string.
* tests/misc/seq-extra-number.sh: Add this scenario.
* tests/local.mk (all_tests): Add new test.
* doc/sort-version.texi: Fix some typos, avoid overly long lines in
the generated PDF, enclose some sample strings in @samp{...} for better
readability, etc. This also avoids an sc-avoid-builtin error:
s/builtin/built-in/
* doc/sort-version.texi: New file.
* doc/local.mk (doc_coreutils_TEXINFOS): Add new file.
* doc/coreutils.texi: @include new file, replace previous "Details about
version sort" section.
When calling 'stat -c %N' to print the filename, don't explicitly
request the size of the file via statx(), as it may add overhead on
some filesystems. The size is only needed to optimize an allocation
for the relatively rare case of reading a symlink name, and the worst
effect is a somewhat-too-large temporary buffer may be allocated for
areadlink_with_size(), or internal retries if buffer is too small.
The file size will be returned by statx() on most filesystems, even
if not requested, unless the filesystem considers this to be too
expensive for that file, in which case the tradeoff is worthwhile.
* src/stat.c: Don't explicitly request STATX_SIZE for filenames.
Problem reported by Szőts Ákos (Bug#36291).
* NEWS: Mention this.
* src/od.c (skip): Try fseek even on files that do not have usable
sizes, falling back on fread if fseek fails.
When debugging an invalid date due to DST switching, the intermediate
'normalized time' should not be checked - its value can differ between
systems (e.g. glibc vs musl).
Reported by Niklas Hambüchen in
https://lists.gnu.org/r/coreutils/2019-05/msg00031.html
Analyzed by Rich Felker in
https://lists.gnu.org/r/coreutils/2019-05/msg00039.html
* tests/misc/date-debug.sh: Replace the exact normalized time
with 'XX:XX:XX' so different values would not trigger test failure.
* src/stat.c: Drop statbuf argument from out_epoch_sec().
Use statx() rather than [lf]stat() where available,
so a separate call is not required to get birth time.
Set STATX_* mask bits only for things we want to print,
which can be more efficient on some file systems.
Add a new --cache= command-line option that sets the appropriate hint
flags in the statx call. These are primarily used with network
file systems to indicate what level of cache coherency is desired.
The new option is available unconditionally for better portability,
and ignored where not implemented.
* doc/coreutils.texi: Add documention for --cached.
* man/stat.x (SEE ALSO): Mention statx().
* NEWS: Mention the new feature.
* src/split.c (set_suffix_length): Use a more standard
zero based logN calculation for the number of units.
* tests/split/suffix-auto-length.sh: Add a test case.
* THANKS.in: Mention the reporter.
* NEWS: Mention the fix.
Fixes https://bugs.gnu.org/35291
Problem reported by Hans Henrik Bergan (Bug#36007).
* NEWS: Mention this.
* src/dd.c (iclose, ifdatasync, ifstat, ifsync):
New functions, which are more careful about SIGINT.
(cleanup): Use iclose instead of close.
(finish_up): Process signals first.
(skip, dd_copy, main): Use ifstat instead of fstat.
(dd_copy): Use ifdatasync and ifsync instead of fdatasync and fsync.
Problem reported by Jakub Kulik (Bug#35713).
* NEWS: Mention this.
* configure.ac (DEV_FD_MIGHT_BE_CHR): New macro.
* src/copy.c (DEV_FD_MIGHT_BE_CHR): Default to false.
(follow_fstatat): New function.
(copy_internal): Use it.
* src/copy.h (XSTAT): Remove; no longer used.
The wording of the dd --help text suggests that output will be skipped
for sparse *input* blocks (i.e. that NUL-checking is done on input
blocks) while the code actually checks/skips all-NUL *output* blocks.[1]
* src/dd.c (usage): Update the --help text to clarify the above.
* tests/dd/sparse.sh: Ensure sparseness is controlled with obs.
[1]: https://superuser.com/a/1136358
Its support for the -include option is flaky. Problem reported by
Michael Osipov (Bug#35650). Plus, we could run into other
compilers that don’t support any option like -include. Change the
code so that -include is not needed. Although this causes us to
depart from the upstream version, we’re already doing that for
other reasons.
* configure.ac (USE_XLC_INCLUDE): Remove, as there’s no
guarantee a compiler will support something like -include.
* src/blake2/b2sum.c [HAVE_CONFIG_H]: Include <config.h>.
* src/local.mk (src_b2sum_CPPFLAGS): Add -DHAVE_CONFIG_H.
Do not use -include or a substitute.
Problem reported by Michael Osipov (Bug#35650).
* configure.ac: Use AC_LANG_WERROR to pay attention to compiler
and linker warnings when testing whether stdbuf will work.
* src/blake2/blake2.h (BLAKE2_PACKED):
Don’t assume __attribute__ ((packed)) works on non-Microsoft
compilers. Instead, assume it works only if we have good
reason to assume so, and fall back on Microsoft (or not packing)
otherwise. In practice, not packing is good enough and the
BLAKE2_PACKED macro is mostly just for documentation.
* src/basenc.c: Various minor style cleanups.
(struct base_decode_context): Do not use anonymous unions, as
they’re not in C99. Use a named union instead. All uses changed.
* src/system.h (X2NREALLOC, X2REALLOC, DECIMAL_DIGIT_ACCUMULATE):
Use verify_expr instead of verify_true, which has been removed.
(DECIMAL_DIGIT_ACCUMULATE): Remove unnecessary size check.
Problem reported by John Marino (Bug#34894).
* src/ln.c (main): Port ln -s to Solaris symlink function,
where symlink ("x", ".") fails with errno == EINVAL.
Some files are physically copied from gnulib, and should get sync'ed
after each update to latest gnulib. This was forgotten during recent
updates.
* COPYING: Merge from gnulib/doc/COPYINGv3.
* tests/init.sh: Merge from gnulib/tests/init.sh.
* tests/misc/test-N.sh: The subsecond values for atime and mtime
were potentially seen to differ on newlyl created files.
So we include the subsecond portion when comparing stat values.
* tests/misc/wc-nbsp.sh: Add gating checks for all characters,
as there are disparate classifications on various systems:
SunOS 5.10 treats \u202F, \u2060 as !iswprint()
SunOS 5.10 treats \u00A0, \u2007 as iswspace()
AIX 7.2, Darwin 17.4.0, NetBSD 7.1 treat \u2060 as !iswprint()
* tests/id/zero.sh: sed on OSX will output a \n even
if the input doesn't have a \n on the last "line".
So ensure we always have a trailing '\n' to avoid the disparity.
Testing by Assaf Gordon on OSX showed the atime wasn't
being updated when explicitly set back in time.
Also Debian 8.11 / mips64 was seen to not update the
mtime when truncating an empty file.
* tests/misc/test-N.sh: Isolate from different timestamping
behaviors of various (file) systems, by correlating
the timestamps with stat(1) before using `test -N`.
Very old makeinfo-4.13 fails with:
./doc/coreutils.texi:2286: Unknown command `hashchar'.
./doc/coreutils.texi:2286: Misplaced {.
./doc/coreutils.texi:2286: Misplaced }.
Reported Bernhard Voelker in
https://lists.gnu.org/r/coreutils/2019-03/msg00016.html .
* doc/coreutils.texi (basenc invocation): Replace @hashchar{} with
actual hash character. The special syntax is only required
when referring to #line directives.
* tests/misc/wc-nbsp.sh: FreeBSD and OS X don't
treat non breaking space as printable characters.
So use wc -L to determine printability before
testing non breaking space functionality.
* src/env.c (initialize_signals): A new function to initialize
the signals array on the heap, to avoid a build failure on
opensolaris, where SIGNUM_BOUND is not a constant.
* man/local.mk: commit f114495e added an extra check to ensure
a binary was working before using it to generate the man page.
However this was not working for the false(1) command,
and also one can generally specify that one should not
be using generated commands on the current system by passing
'cross_compiling=yes' to the configure invocation.
* src/env.c (main): Output blocked or ignored signals
before a command is executed.
* doc/coreutils.texi (env invocation): Add the option.
* tests/misc/env-signal-handler.sh: Add a test case.
* NEWS: Mention the new feature.
New options to set signal handlers for the command being executed.
--block-signal suggested by Paul Eggert in http://bugs.gnu.org/34488#71
--default-signal is useful to overcome the POSIX limitation that shell
must not override inherited signal state, e.g. the second 'trap' here is
a no-op:
trap '' PIPE && sh -c 'trap - PIPE ; seq inf | head -n1'
Instead use:
trap '' PIPE && sh -c 'env --default-signal=PIPE seq inf | head -n1'
Similarly, the following will prevent CTRL-C from terminating the
program:
env --ignore-signal=INT seq inf > /dev/null
See https://bugs.gnu.org/34488#8
* NEWS: Mention new options.
* doc/coreutils.texi (env invocation): Document new options.
* man/env.x: Add example of --default-signal=SIG usage.
(SEE ALSO): Mention sigprocmask.
* src/env.c (signals): New global variable.
(longopts): Add new options.
(usage): Print new options.
(parse_signal_params): Parse comma-separated list of signals, store in
signals variable.
(reset_signal_handlers): Set each signal to SIG_DFL/SIG_IGN.
(parse_block_signal_params): Parse command-line options.
(set_signal_proc_mask): Call sigprocmask to block/unblock signals.
(main): Process new options.
* src/local.mk (src_env_SOURCES): Add operand2sig.c.
* tests/misc/env-signal-handler.sh: New test.
* tests/local.mk (all_tests): Add new test.
* configure.ac: Check for statx(), available on glibc >= 2.28.
* src/stat.c (get_birthtime): Call statx() when available.
* NEWS: Mention the improvement.
* src/df.c (replace_problematic_chars): A new wrapper to be
more conservative in our replacement when not connected to a tty.
* tests/df/problematic-chars.sh: Add a test case.
* doc/coreutils.texi (node seq invocation): Clarify to use the tool
'yes'; otherwise the reader may interpret the sentence as if one
could pass 'yes' as the INCREMENT value.
* src/wc.c (iswnbspace): A new function to match
characters in this class.
(isnbspace): Likewise for single byte charsets.
(main): Initialize posixly_correct from the environment,
to allow disabling honoring NBSP in non C locales.
(wc): Call is[w]nbspace() along with is[w]space.
* bootstrap.conf: Ensure btowc is available.
* tests/misc/wc-nbsp.sh: A new test.
* tests/local.mk: Reference the new test.
* NEWS: Mention the change in behavior.
The recent Gnulib update fixed Bug#34608; document and test this.
* NEWS: Mention the change.
* doc/coreutils.texi (Padding and other flags):
Update doc to cover new flag and other POSIX.1-2017 changes.
* tests/misc/date.pl (date-century-plus): New test.
For select programs which accept only --help and --version options
(in addition to non-option arguments), process these options before
any other options.
Before:
$ dd bs=1 --help
dd: unrecognized option '--help'
Try 'dd --help' for more information.
$ yes me --help
me --help
me --help
...
After:
Any occurrence of '--help' in the arguments (prior to '--') will
show the help screen.
Discussed in https://bugs.gnu.org/33468 .
* NEWS: Mention change.
* src/cksum.c, src/dd.c, src/hostid.c, src/hostname.c, src/link.c,
src/logname.c, src/nohup.c, src/sleep.c, src/tsort.c, src/unlink.c,
src/uptime.c, src/users.c, src/whoami.c, src/yes.c (main): Replace
parse_long_options() + getopt_long() calls with
parse_gnu_standard_options_only(); Remove <getopt.h> inclusion;
Remove empty 'struct long_options' variable;
* tests/misc/help-version-getopt.sh: Add test.
* tests/local.mk (all_tests): Reference it.
* src/sort.c (main): Adjust the debug info regarding locales,
to clarify that only textual comparisons are affected.
* tests/misc/sort-debug-warn.sh: Adjust accordingly.
Fixes https://bugs.gnu.org/34490
* src/comm.c (main): Output a warning right before exit,
in case previous errors have scrolled from view.
* src/join.c (main): Likewise.
* tests/misc/comm.pl: Addjust accordingly.
* tests/misc/join.pl: Likewise.
Fixes https://bugs.gnu.org/34347
* gnulib: Update to make the new strtold module available.
* bootstrap.conf: strtod is now a dependency of c-strtod,
which in turn is a dependency of cl-strtod. This treats
strtold and strtod similarly.
* gl/lib/cl-strtod.c: Adjust to assume strtold is available.
* tests/misc/sort-float.sh: Likewise.
* src/sort.c: Likewise.
(nan_compare): Adjust comment to indicate
we still have to init padding bits as per
https://sourceware.org/bugzilla/show_bug.cgi?id=13246
* src/seq.c (print_numbers): Only reset the locale if it
was successfully set originally.
* tests/misc/seq-locale.sh: Add a new test.
* tests/local.mk: Reference the new test.
* NEWS: Mention the fix.
bootstrap.conf: Explicitly depend on select, rather than transitively.
* src/tail.c: Unconditionally include select.h as we use select()
outside inotify contexts now.
* src/extract-magic: Treat android like linux,
which fixes the build by ensuring the constants are defined.
* src/stat.c: Support all constants on android, including
the android specific "sdcardfs".
* src/tail.c: Fix inclusion of statfs headers to be independent
of inotify availability, as fremote() is used on linux even
if inotify has been disabled. Also enable fremote() on android.
* NEWS: Mention the improvment.
Fixes https://bugs.gnu.org/34239
These commands now accept floating-point numbers in the
current locale, as well as in the C locale.
Compatibility problem reported by Robert Elz.
* NEWS: Document this.
* bootstrap.conf (gnulib_modules): Add cl-strtod, cl-strtold.
Remove c-strtold.
* doc/coreutils.texi (Floating point, tail invocation)
(printf invocation, timeout invocation, sleep invocation)
(seq invocation): Document this.
* gl/lib/cl-strtod.c, gl/lib/cl-strtod.h, gl/lib/cl-strtold.c:
* gl/modules/cl-strtod, gl/modules/cl-strtold: New files.
* src/printf.c, src/seq.c, src/sleep.c, src/tail.c, src/timeout.c:
Include cl-strtod.h instead of c-strtod.
* src/printf.c (vstrtold):
* src/seq.c (scan_arg, print_numbers):
* src/sleep.c (main):
* src/tail.c (parse_options):
* src/timeout.c (parse_duration):
Use cl_strtold instead of c_strtold.
Problem reported by Robert Elz.
* doc/coreutils.texi (sleep invocation):
Say that arguments must be non-negative, which means they cannot
be arbitrary floating-point numbers. Mention POSIX, not
“historical implementations” that are no longer of practical
interest. List the extensions to POSIX.
* src/sleep.c (usage): Omit needless words, removing dubious
commentary about “most implementations” and incorrect commentary
about “arbitrary”. Details about exactly which numbers are
allowed can be found in the documentation.
* init.cfg (trap_sigpipe_or_skip_): A new function refactored from...
* tests/misc/printf-surprise.sh: ...here.
* tests/misc/seq-epipe.sh. Likewise.
* src/tail.c (die_pipe): Ensure we exit upon sending SIGPIPE.
* tests/tail-2/pipe-f.sh: Ensure we exit even if SIGPIPE is ignored.
* NEWS: Mention the bug fix.
Fail developer builds if VLAs are used,
as there are portability concerns to consider with them.
* configure.ac: Enable -Wvla which is implicit in the full list added.
* m4/jm-macros.m4: Define GNULIB_NO_VLA which disables use of
VLAs within gnulib code.
When building against an incompatible GLIBC version compared to that
on the build host, then running the just-built binary might fail
although it is the same platform - thus CROSS_COMPILING is false.
As a result, generating the man pages fails.
* man/local.mk (.x.1): Add a check to verify that running the utility
with --help succeeds, otherwise falling back to using 'dummy-man'.
* src/ls.c (is_linked_directory): A new function to
also consider symlinked directories.
(main): Rename check_symlink_color to check_symlink_mode,
and enable that with --group-directories-first.
(DIRFIRST_CHECK): Adjust to use is_linked_directory,
rather than just is_directory.
(gobble_file): Simplify to always update f->linkmode
if the stat() succeeds.
* tests/ls/group-dirs.sh: A new test.
* tests/local.mk: Reference the new test.
* NEWS: Mention the change in behavior.
Suggested by Amin Bandali in
https://lists.gnu.org/r/coreutils/2018-12/msg00017.html
* src/tail.c: Fix the check_output_available check on AIX.
Note we don't use poll for all systems as the overhead
of adding the gnulib poll module wouldn't be worth it
just for this single use.
* tests/tail-2/pipe-f.sh: Fix the test which always passed
due to only the exit code of sleep being checked.
* NEWS: Mention the bug fix and rearrange alphabetically.
Fixes http://bugs.gnu.org/33946
Allocate the encoding/decoding buffers dynamically on the heap instead
of using variable-length-array (VLA) on the stack.
Discussed in https://lists.gnu.org/r/coreutils/2019-01/msg00004.html .
* src/basenc.c (do_encode,do_decode): Allocate inbuf/outbuf using
xmalloc, and free if using LINT.
* src/system.h: Adjust lines containing URLs so that
they don't wrap on 80 column terminals. One could also
use .UR macros, but these aren't universally available.
Note the adjustments here need to be compatible with
the pattern matching done in help2man.
Addresses https://bugs.gnu.org/33914
Run "make update-copyright" and then...
* gnulib: Update to latest with copyright year adjusted.
* tests/init.sh: Sync with gnulib to pick up copyright year.
* bootstrap: Likewise.
* tests/sample-test: Adjust to use the single most recent year.
In the following invocation, 'a' is the input file, and 'b' is the extra
operand:
$ base64 a b
Report 'b' in the error message instead of 'a':
$ base64 a b
base64: extra operand 'b'
Discussed in https://lists.gnu.org/r/coreutils/2018-12/msg00008.html .
* src/basenc.c (main): If there is more than one non-option operand,
report the second one (assuming the first is a the input file name).
* tests/misc/base64.pl: Add tests.
* tests/misc/basenc.pl: Adjust expectedc error message in tests.
* NEWS: Mention bugfix.
Encodes/decodes data in various common formats:
base64,base64url,base32,base32,base16,base2,z85.
Discussed here:
https://lists.gnu.org/r/coreutils/2018-11/msg00014.htmlhttps://lists.gnu.org/r/coreutils/2018-12/msg00019.html
* AUTHORS: Add basenc.
* README: Reference the new program.
* NEWS: Mention the new program.
* build-aux/gen-lists-of-programs.sh: Add basenc.
* doc/coreutils.texi: (basenc invocation): Document the new command.
* man/.gitignore: Ignore the generated man page.
* man/basenc.x: A new template, with few examples.
* man/local.mk: Reference the new man page.
* scripts/git-hooks/commit-msg: Allow basenc as program prefix.
* src/.gitignore: Ignore the new binary.
* src/basenc.c:
(usage): Mention new options.
(main): Handle new options.
(isbase*, base*_length, base*_encode, base*_decode_ctx): Implement new
encoding/decoding formats.
* src/local.mk: Add new program.
* tests/local.mk: Add new test.
* tests/misc/basenc.pl: New tests.
* tests/misc/help-version.sh (basenc_setup): use '--version' for default
invocation (basenc errors with no parameters).
Problem reported for split by Scott Worley (Bug#33761):
* src/shred.c (do_wipefd):
Also report an error if ftruncate fails on a shared memory object.
* src/sort.c (get_outstatus): New function.
(stream_open, avoid_trashing_input): Use it.
* src/sort.c (stream_open):
* src/split.c (create):
If ftruncate fails, do not report an error
unless it is a regular file or a shared memory object.
* NEWS (Bug fixes): Mention the fix in commit 94d364f157.
While at it, remove duplicate "Changes in behavior" heading.
* tests/misc/sync.sh: Add a test with a write-only file for the fix.
Open a target directory and use its file descriptor in linkat,
symlinkat, etc. syscalls, instead of constructing long file names
by concatenating the target directory name to a basename.
This avoids O(N²) behavior with ‘ln F1 F2 ... Fn DIR’ when DIR is
a long file name with many slashes. It also avoids some races if
DIR is renamed while ln is running.
* bootstrap.conf (gnulib_modules): Add openat-safer.
* src/ln.c: Include fcntl-safer.h.
(O_PATHSEARCH): New constant.
(errno_nonexisting, target_directory_operand): Remove; no longer used.
(atomic_link, do_link): New arg DESTDIR_FD. All uses changed.
(do_link): New arg DEST_BASE. All uses changed.
(main): Open target directory and use its file descriptor
as DESTDIR_FD.
* src/echo.c (usage): Assert that STATUS is always EXIT_SUCCESS.
* tests/misc/echo.sh: Add further tests for all hex and escape and
escape characters.
To get coverage statistics, run:
make coverage -j 4 TESTS=tests/misc/echo.sh SUBDIRS=.
xdg-open doc/coverage/src/echo.c.gcov.frameset.html
* src/echo.c (main): Always enable backslash processing if
POSIXLY_CORRECT is set.
* tests/misc/echo.sh: Add (the first) test for the echo command.
* tests/local.mk: Reference the new test.
* tests/misc/printf.sh: Update a stale comment.
* doc/coreutils.texi (echo invocation). Mention that POSIXLY_CORRECT
now always enables backslash processing.
* NEWS: Mention the change in behavior.
Fixes https://bugs.gnu.org/32703
Issue identified by Eric Blake.
Bash knows 'test -N FILE'. Add it to GNU 'test' as well.
* src/test.c (unary_operator): Add a case for 'N'.
(usage): Document it.
* doc/coreutils.texi (node File characteristic tests): Likewise.
* NEWS (New features): Likewise.
* tests/misc/test-N.sh: Add a test.
* tests/local.mk (all_tests): Reference it.
Remove the function 'test_unop', as the cases therein are redundant to
those handled by 'unary_operator'; exception: the cases 'o' and 'N':
they had been present in test_unop and handling the commands
test -N STR
test -o STR
and
test x = x -a -N STR
test x = x -a -o STR
which ran into an error later on anyway.
With this commit, the error diagnostic will change from ...
$ /usr/bin/test -N STR
/usr/bin/test: extra argument '-N'
$ /usr/bin/test -o STR
/usr/bin/test: extra argument '-o'
... to ...
$ src/test -N STR
src/test: '-N': unary operator expected
$ src/test -o STR
src/test: '-o': unary operator expected
* src/test.c (test_unop): Remove.
(unary_operator): Fail with test_syntax_error in the default case.
(term): Directly call unary_operator.
(two_arguments): Likewise.
* tests/misc/test-diag.pl: Adjust error diagnostic.
* src/test.c (unary_operator): Remove case 'a'.
(test_unop): Likewise.
* NEWS (Changes in behavior): Document the change.
Discussed at https://bugs.gnu.org/33097
On my openSUSE:Tumbleweed system, I get a false positive test failure
in the above 'check-root' test because the group lists inside and
outside the chroot have a different order:
++ chroot --userspec=berny / id -G
++ id -G berny
+ test '100 454 457 480 492' = '100 480 492 457 454'
+ fail=1
* tests/misc/chroot-credentials.sh (num_sort): Add function to sort
group lists, and use it in the test cases which test multiple groups.
Previously, 'ln A B' did 'stat("B"), lstat("A"), link("A","B")'
where the stat and lstat were necessary to avoid hard-linking
directories on systems that can hard-link directories.
Now, in situations that prohibit hard links to directories,
'ln A B' merely does 'link("A","B")'. The new behavior
avoids some races and should be more efficient.
This patch was inspired by Bug#10020, which was about 'ln'.
* bootstrap.conf (gnulib_modules): Add unlinkdir.
* src/force-link.c (force_linkat, force_symlinkat): New arg for
error number of previous try. Return error number, 0, or -1 if
error, success, or success after removal. All callers changed.
* src/ln.c: Include priv-set.h, unlinkdir.h.
(beware_hard_dir_link): New static var.
(errnoize, atomic_link): New functions.
(target_directory_operand): Use errnoize for simplicity.
(do_link): New arg for error number of previous try. All callers
changed. Do each link atomically if possible.
(main): Do -r check earlier. Remove linkdir privileges so we can
use a single linkat/symlinkat instead of a racy substitute for the
common case of 'ln A B' and 'ln -s A B'. Set beware_hard_dir_link
to disable this optimization.
$ id root nobody
uid=0(root) gid=0(root) groups=0(root)
uid=99(nobody) gid=99(nobody) groups=99(nobody)
* src/id.c (main): Make variables opt_zero, just_group_list,
just_group, use_real, just_user global to be used in a new
function.
(print_stuff): New function that will print user and group
information for the specified USER.
When using -G option delimit each record with two NULs.
Restructure the code in the file to have global variables
followed by functions.
* tests/id/zero.sh: Add test cases to check the usage
of -z option with multiple users.
* tests/id/uid.sh: Add a test case to ensure all users
are queried in the presence of errors.
* doc/coreutils.texi: Document the interface changes.
* NEWS: Mention the new feature.
* doc/coreutils.texi (csplit invocation): Detail the behavior
with regexp patterns and negative offsets, which differs from
line number patterns, to avoid looping on the input. For example:
$ seq 50 | csplit -s - /15/-5 /12/
csplit: ‘/12/’: match not found
* doc/coreutils.texi (csplit invocation): Clarify that
portions of the input may be skipped and thus the input
may not be reproducible by just concatenating the output files.
Fixes https://bugs.gnu.org/32317
This improves on the earlier fix for the problem reported by
Chih-Hsuan Yen (Bug#32236), by also looking for other control
characters and for encoding errors.
* src/df.c: Include wchar.h and wctype.h instead of c-ctype.h.
(hide_problematic_chars): Process the string as multibyte.
Use iswcntrl, not c_iscntrl.
* src/df.c (hide_problematic_chars): Use c_iscntrl() as
passing 8 bit characters to iscntrl() is not supported on macOS.
* NEWS: Mention the bug fix.
Fixes https://bugs.gnu.org/32236
* tests/rm/rm-readdir-fail.sh: Skip the test entirely on 32 bit,
so we avoid conflating the 32bit and 64 bit types, as that
triggers alignment issues (SIGBUS) on Gentoo sparc.
Fixes https://bugs.gnu.org/29886
* tests/cp/cp-a-selinux.sh: Use 'skip_' rather than the probably
undefined 'skip'.
* tests/du/2g.sh: Likewise.
* tests/install/install-Z-selinux.sh: Likewise.
* tests/misc/chcon.sh: Likewise.
* tests/misc/selinux.sh: Likewise.
* tests/mkdir/restorecon.sh: Likewise.
* cfg.mk (sc_prohibit-skip): A new syntax check to catch the issue.
* init.cfg (require_membership_in_two_groups_): This fixes a bug
introduced by me in v8.15-8-gdd0e4c562. Luckily, the consequence
of low-probability triggering the bug was the mere added backslash
in the diagnostic: "...but running id -G\ either...". It would be
triggered in a test failure for one who is a member of only one or
fewer groups.
* src/copy.c (copy_internal): Use the lint protected src_mode,
rather than accessing the src_sb again. Also unconditionally
populate src_sb when !x->move_mode and in lint mode.
Reported by Kamil Dudka with coverity and clang analyzer.
* src/env.c (main): Don't process '-' specially since
that causes an issue on the openbsd getopt implementation
where a lone '-' is now processed as an option, and anyway
it doesn't particuarly help diagnosing common shebang
usage issues. Also don't restrict the extra diagnostics
for shebang usage to the case with 3 arguments, as
further arguments can be passed to a script.
* tests/misc/env-S.pl: Adjust accordingly.
On OpenBSD 6.2, invalid single options produce error messages
without single quotes:
$ ./src/chroot -/
chroot: unknown option -- /
As opposed to other systems:
./src/chroot: invalid option -- '/'
Modify the grep search to accept this.
* tests/misc/usage_vs_getopt.sh (checkprg): Change the grep pattern
to accomodate no-single-quotes cases.
The module is not needed anymore (was used during development).
Despite being a Perl core module, platforms like CentOS don't install
it by default. Reported by Bruno Haible at
https://lists.gnu.org/r/coreutils/2018-06/msg00093.html.
* tests/misc/csplit-suppress-matched.pl: Remove Data::Dumper.
Add attribute 'malloc' to mpz_get_str to prevent
the following on GCC 8.1.1
src/expr.c:117:1: error: function might be candidate for attribute
'malloc' if it is known to return normally
[-Werror=suggest-attribute=malloc]
mpz_get_str (char const *str, int base, mpz_t z)
^~~~~~~~~~~
cc1: all warnings being treated as errors
* src/expr.c (mpz_get_str): Add _GL_ATTRIBUTE_MALLOC.
* gnulib: Update to latest.
* .gitignore: Add new entries.
* bootstrap.conf: Enable wchar-single, which will enable more
efficient replacements of wcwidth and mbrtowc, as we indicate
that the charset will no change between invocations of these functions.
* src/local.mk (fs_normalize_perl_subst): `make src/fs-magic-compare`
was reporting incorrectly that AFS was not being handled.
Add a mapping to our KAFS identifier.
* .gitignore: Add intermediate files from `make src/fs-magic-compare`
With an uninitialized variable 'fail', the unquoted use like
test $fail = 1
lead to the shell error
"unary operator expected".
The uninitialized 'fail' variable was a side effect of
https://git.sv.gnu.org/cgit/gnulib.git/commit/?id=e91c0d4f9
which was pulled into coreutils-v8.26 with
https://git.sv.gnu.org/cgit/coreutils.git/commit/?id=ef9650170
Coreutils test code relied and relies on 'fail' to be initialized,
so initialize that variable here.
* tests/local.mk (TESTS_ENVIRONMENT): Initialize fail=0.
* src/remove.c (rm_fts): With the --preserve-root=all extension,
reject command line arguments that are mount points.
* src/remove.h (rm_options): Add preserve_all_root to store config.
* src/mv.c (rm_option_init): Init preserve_all_root to false.
* src/rm.c (main): Init preserve_all_root as per option.
(usage): Describe the new option.
* src/remove.c (rm_fts): Lookup the parent device id,
and reject the cli argument if a separate file system.
* tests/rm/one-file-system.sh: Add a test case.
* NEWS: Mention the new feature.
This mode is currently the default, but most if not all users of
reflink-capable filesystems want --reflink=auto, which is often
encapsulated into an alias. Adding --reflink=never allows overriding
such an alias.
* doc/coreutils.texi (cp invocation): Describe the new option.
* src/cp.c: Support --reflink=never.
* tests/cp/reflink-auto.sh: Add a test case.
* NEWS: Mention the new feature.
Adopted from FreeBSD's env(1), useful for specifing multiple
parameters on a shebang (#!) script line, e.g:
#!/usr/bin/env -S perl -w -T
Discussed in https://lists.gnu.org/r/coreutils/2018-04/msg00011.html
* src/env.c (valid_escape_sequence,escape_char,scan_varname,
extract_varname,validate_split_str,build_argv,
parse_split_string): New functions.
(main): Process new option and call parse_split_string.
(usage): Mention new option.
* tests/misc/env-S.pl: Test new option from the command line.
* tests/misc/env-S-script.sh: Test new option from shebang scripts.
* tests/local.mk (all_tests): Add new tests.
* man/env.x (OPTIONS): Show a brief example of -S usage and point to
the full documentation for more information.
* doc/coreutils.texi (env invocation): Detail usage of -S/--split-string
option.
* NEWS: Mention new option.
Keep unset envvars (-uFOO) in an array for later deletion,
instead of reiterating over argv. Done in preparation for
'-S string' feature. Related to '-u' discussion in
https://lists.gnu.org/r/coreutils/2018-04/msg00013.html
* src/env.c (append_unset_var,unset_envvars): New functions.
(main): Use new functions.
* src/ls.c (get_color_indicator): s/STREQ_LEN/c_strncasecmp/
* src/dircolors.hin: Remove a now redundant entry.
* tests/ls/color-ext.sh: Add a new test.
* tests/local.mk: Reference the new test.
* NEWS: Mention the change in behavior.
* doc/coreutils.texi (md5sum invocation): Describe the new option,
and how it's not supported by --check, and how it disables escaping.
* src/md5sum.c (delim): A new global to parmeterize the out delimiter.
(main): Don't enable file name escaping with -z, and output '\0'.
* tests/misc/md5sum-newline.pl: Add a test case.
* NEWS: Mention the new feature.
===== Benchmark setup (on GNU/Linux) ====
$ yes áááááááááááááááááááá | head -n100000 > mbc.txt
$ yes 12345678901234567890 | head -n100000 > num.txt
===== Before ====
$ time src/wc -Lm < mbc.txt
real 0m0.186s
$ time src/wc -m < mbc.txt
real 0m0.186s
$ time src/wc -Lm < num.txt
real 0m0.055s
$ time src/wc -m < num.txt
real 0m0.056s
==== After ====
$ time src/wc -Lm < mbc.txt
real 0m0.196s
$ time src/wc -m < mbc.txt
real 0m0.173s
$ time src/wc -Lm < num.txt
real 0m0.031s
$ time src/wc -m < num.txt
real 0m0.028s
* src/wc.c (wc): Only call wide variant functions like
iswprint() and wcwidth() for non is_basic() characters.
I.E. non ISO C "basic character set" characters.
This is especially significant on OSX where wcwidth()
is very expensive (about 10x in tests).
* NEWS: Mention the improvement.
Suggested by Eric Fischer.
Problem reported by Eric S. Raymond (Bug#31803).
* man/test.x: Add SYNOPSIS section, since help2man
understandably gets confused by the square brackets.
* src/ln.c (usage): Omit parenthetical "(Nth form)" in usage,
as it confuses doclifter.
This issue was introduced in commit v8.19-145-g24ebca6
* src/copy.c (copy_internal): With --no-preserve=mode,
only reset permissions for newly created files.
(copy_reg): Likewise.
* NEWS: Mention the fix.
* tests/cp/preserve-mode.sh: Add a test case.
Fixes https://bugs.gnu.org/31675
* tests/ls/abmon-align.sh: Base relative month adjustment
from the middle of the month, to avoid failures due
to months being repeated.
Fixes https://bugs.gnu.org/31644
Avoid warnings from: groff -b -e -mandoc -T utf8 -rF0 -t -w w -z
* man/du.x: Change ".BR" to ".B" if there is only one argument.
Protect an end-of-sentence indicator (.?!) with '\&'
if it does not mean an end of a sentence.
Change '--' to '\-\-' if it indicates an option.
* man/rm.x: Change '\=' to '='.
* src/copy.c (copy_internal): Don't fail immediately upon
getting ELOOP when running stat() on the destination,
rather proceeding if -f specified, allowing the link
to be removed. If the loop is not in the final component
of the destination path, we still fail but at the
subsequent unlink() stage.
* doc/coreutils.texi (cp invocation): Adjust wording to say
that --force doesn't work with dangling links, rather than
all links that can't be traversed.
* tests/cp/thru-dangling.sh: Add a test case.
* NEWS: Mention the change in behavior.
Discussed in https://bugs.gnu.org/31335
Ensure this _does_ recreate the symlink
Given "path1" and "path2" are on different devices.
$ touch "path1/file"
$ cd path2/; ln -s path1/file
$ cp -dsf path1/file .
Ensure this does _not_ overwrite file
$ touch file
$ ln -s file l1
$ cp -sf l1 file
* src/copy.c (same_file_ok): Remove device ids from consideration,
instead deferring to future EXDEV with --link or allowing
the first case above to work.
Also ensure that we do not exist this function too early,
when the destination file is not a symlink, which protects
against the second case.
* tests/cp/cross-dev-symlink.sh: Add a test for the first case.
* tests/cp/same-file.sh: Add a test for the second case above.
* NEWS: Mention the bug fixes.
* THANKS.in: Mention the reporters who also analyzed the code.
Fixes https://bugs.gnu.org/31364
* src/cp.c (target_directory_operand): Allow through inaccessible
arguments with -f or --remove.
* doc/coreutils.texi (cp invocation): Clarify that -f doesn't directly
impact the removal of non-traversable symlinks.
* tests/cp/dir-rm-dest.sh: Test the new behavior.
* tests/cp/thru-dangling.sh: Enforce -f behavior wrt symlinks.
* NEWS: Mention the bug fix.
Fixes https://bugs.gnu.org/31335
* src/chmod.c: Deallocate the mode change array in dev mode.
* src/chown.c: Make chopt_free() actually deallocate, but
only call in dev mode.
* src/chgrp.c: Likewise.
* src/chown-core.h (chopt_free, gid_to_name, uid_to_name):
No longer const.
* src/make-prime-list.c (xalloc): Add malloc attribute.
* src/who.c (make_id_equals_comment): Work around GCC bug 85602
by using mempcpy rather than strncat. Although the old code
was correct, strncat raises so many hackles that it’s not
worth maintaining its use here.
* NEWS: Expand on the 8.28 description of how tail more
responsively reacts to closed output, and move from "Improvements"
to "Changed behavior".
* cfg.mk (old_NEWS_hash): Regenerate.
Fixes https://bugs.gnu.org/31225
* src/timeout.c (usage): Mention that a duration of 0 disables
the associated timeout, which is both concise info and useful
functionality as timeouts are frequently configured.
Increase max range from SIZE_MAX to UINTMAX_MAX, which will
allow cut to support line lengths up to the max file size
on all systems. The inherent SIZE_MAX limitation in cut was
removed with the enhancements in https://bugs.gnu.org/13127.
Also numfmt gets similarly increased --field ranges due to
shared code.
* src/cut.c: s/size_t/uintmax_t/.
* src/numfmt.c: Likewise.
* src/set-fields.c: Likewise.
* src/set-fields.h: Likewise.
* tests/misc/cut-huge-range.sh: Adjust accordingly.
* tests/misc/numfmt.pl: Likewise.
* NEWS: Mention the improvement.
Problem reported by Karl Berry (Bug#30963).
* NEWS: Mention this.
* src/ls.c (decode_switches): Implement this.
* tests/ls/a-option.sh: New file.
* tests/local.mk (all_tests): Add it.
This will impact relatively few languages,
and will make Arabic or Catalan etc.
output unambiguous abbreviated month names.
* src/ls.c (MAX_MON_WIDTH): Increase from 5 to 12.
* NEWS: Mention the bug fix.
* tests/ls/abmon-align.sh: Augment to check for ambiguous output.
Fixes https://bugs.gnu.org/30814
Enhanced XFS (EXFS) is a version of XFS maintained by HPE.
EXFS uses a unique magic number to allow the use of community
XFS, and EXFS filesystems at the same time.
* src/stat.c (human_fstype): Add file system ID definition,
and use "exfs" as the name.
* NEWS: Mention the Improvement.
Problem reported by John Wiersba (Bug#30718)
* src/stat.c (human_time): Avoid giving an integer constant
expression a name, as it runs afoul of a bug in IBM XL C/C++ for
AIX 12.01.0000.0002.
This issue was introduced in commit v8.19-145-g24ebca6
* src/copy.c (copy_internal): When setting default permissions
to use with --no-preserve=mode, only set executable bits for
directories or sockets.
* NEWS: Mention the fix.
* tests/cp/preserve-mode.sh: Add a test case.
Fixes https://bugs.gnu.org/30534
* doc/coreutils.texi: Use @example consistently
as we don't need the smaller or fixed width representation.
This is especially true for the synopsis of commands.
@smallexample is rendered left aligned for HTML
which is awkward to read with the center aligned main content.
This builds on a previous patch for mv atomicity (Bug#29961).
It merely improves performance; it does not fix bugs.
* src/copy.h (struct cp_options): New members last_file, rename_errno.
* src/copy.c (copy_internal): Support new rename_errno member
for the copy options. Avoid calling stat when new members
suggest it’s not needed.
(cp_options_default): Initialize new members.
* src/mv.c: Include renameat2.h.
(main): With two arguments, first call ‘renamat2 (AT_FDCWD, "a",
AT_FDCWD, "b", RENAME_NOREPLACE)’. Use its results to skip
remaining processing if possible; for example, if it succeeds
there is no need to stat either "a" or "b". Also, set
x.last_file when it is the last file to rename.
Problem reported by Kamil Dudka (Bug#29961).
* NEWS: Mention this.
* src/copy.c: Include renameat2.h.
(copy_internal): If mv, try renameat2 first thing, with
RENAME_NOREPLACE. If this works, skip most of the remaining code.
Also, fail quickly if it fails with EEXIST, and we are using -n.
In both chown and chgrp (which shares its code with chown), operating
on symlinks recursively has a window of vulnerability where the
destination user or group can change the target of the operation.
Warn about combining the --dereference, --recursive, and -L flags.
* doc/coreutils.texi (warnOptDerefWithRec): Add macro.
(node chown invocation): Add it to --dereference and -L.
(node chgrp invocation): Likewise.
See also: CVE-2017-18018
* NEWS: Mention these fixes.
* doc/coreutils.texi (cp invocation, mv invocation):
Mention that -n is silent, and that it overrides -u.
* src/cp.c, src/mv.c (main): -n overrides -u.
* doc/coreutils.texi: the documentation for the --dereference
flag of chown/chgrp states that it is the default mode of
operation. Document that this is only the case when operating
non-recursively.
* tests/misc/shred-remove.sh: AIX xargs defaults to using
'_' to indicate end of input, thus ignoring it.
Rather than specifying -E to avoid this behavior, simplify
by removing sed and xargs usage.
Run "make update-copyright" and then...
* gnulib: Update to latest with copyright year adjusted.
* tests/init.sh: Sync with gnulib to pick up copyright year.
* bootstrap: Likewise.
* tests/sample-test: Adjust to use the single most recent year.
* tests/tail-2/pipe-f.sh: Close stdout in a subshell
to ensure the current shell isn't impacted. Subsequent
piped commands like `echo foo | blah` were seen to fail
due to the previous closing of stdout.
Reported by Assaf Gordon.
* README: Document the new handling of 32 bit time_t,
with examples of how to build in 64 bit mode on AIX.
Also mention that GNU make is desired on AIX
due to its mishandling of the "[" target.
Suggested by Assaf Gordon.
* tests/misc/ptx.pl: Escape the '^' character which is
otherwise considered as a line continuation character.
* tests/misc/shred-remove.sh: sed doesn't support \n.
* tests/misc/timeout.sh: dash outputs the "Killed"
message to stderr rather than the terminal.
* tests/misc/usage_vs_getopt.sh: dash doesn't yet
support the POSIX proposed $'...' shell quoting syntax.
* src/ptx.c (fix_output_parameters): GCC 6.3.1 with
./configure --enable-single-binary would give:
error: assuming signed overflow does not occur
when simplifying conditional to constant [-Werror=strict-overflow]
if (file_index > 0)
So change the type of file_index to signed (size_t).
* src/tail.c (tail_bytes): On systems were blksize_t is unsigned
and the same size or wider than off_t (android for example),
our initialized (off_t) -1 would be promoted to unsigned before
comparison, and thus fail to follow the appropriate path.
* tests/tail-2/tail-c.sh: Add a test case.
* NEWS: Mention the fix.
This issue was introduced in commit v8.23-47-g2662702
Reported at https://github.com/termux/termux-app/issues/233
* m4/jm-macros.m4: Check for the header.
* src/dd.c: Avoid the workaround where the header
is not available (on non glibc systems).
* src/shred.c: Likewise.
* src/ls.c (usage): Clarify -k only applies to -s usage
and directory 'total' lines. Move the description
of TIME_STYLE out of the option section as it was awkward
to read and write there within 80 columns.
* man/chmod.x: Update the information to state one can
clear the setuid and setgid bits for directories numerically
using an additional leading '0' or a leading '='.
That has been supported since v8.15-64-g8931cdb.
Fixes https://bugs.gnu.org/29390
* doc/coreutils.texi (cp invocation): The language used
to describe recreating the file was a little confusing
as it mentioned opening a removed file.
Fixes https://bugs.gnu.org/29315
Verify that all options mentioned in usage are actually recognized
by the program.
* tests/misc/usage_vs_getopt.sh: Add test.
* tests/local.mk (all_tests): Reference it.
Co-authored-by: Pádraig Brady <P@draigBrady.com>
* tests/cp/preserve-mode.sh: This was the only use of awk,
which may not be available on the system resulting
in an ineffective test. Also the permissions bits for
directories were not being checked at all.
* src/timeout.c (main): Add short option character 'v' to getopt_long
call.
* tests/misc/timeout.sh: Run the test both for the long and the short
option.
* src/dd.c (iread): Handle read error with a non-aligned
file offset in the O_DIRECT case. This is not an issue
on XFS at least, but on EXT4 the final read will return
EINVAL rather than the expected 0 to indicate EOF.
* tests/dd/direct.sh: Test the iflag=direct case also.
* NEWS: Mention the improvement.
This is useful as handling in shell is complicated
with the varying exit status in the --kill-after case.
* src/timeout.c (main): Handle '-v' and store
COMMAND for the diagnostic.
(cleanup): Diagnose the signal name before sending.
(usage): Document -v, --verbose.
* doc/coreutils.texi (timeout invocation): Likewise.
* tests/misc/timeout.sh: Add a test case.
* NEWS: Mention the new feature
Fixes https://bugs.gnu.org/21760
* src/tail.c (tail_bytes): Try lseek(..., SEEK_END) when
we can't determine the file size.
* tests/tail-2/end-of-device.sh: Add a new root only test.
* tests/local.mk: Reference the new test.
* NEWS: Mention the improvement.
Paul Eggert suggested using lseek() (rather than ioctl(BLKGETSIZE64)).
Fixes https://bugs.gnu.org/29259
These templates instruct contributors not to use github, and instead
use the upstream GNU development resources. Discussed in
http://lists.gnu.org/archive/html/coreutils/2017-11/msg00007.html .
* .github/ISSUE_TEMPLATE.txt,
.github/PULL_REQUEST_TEMPLATE.txt: New files.
Older versions of 'makeinfo' choke on a missing reference:
./doc/coreutils.texi:14177: `Realpath usage examples' has no Up field\
(perhaps incorrect sectioning?).
makeinfo: Removing output file `doc/coreutils.info' due to errors; \
use --force to preserve.
* doc/coreutils.texi (realpath invocation): Add a menu referencing
the usage examples - introduced in v8.27-91-g7449f0d.
Suggested by L A Walsh in https://bugs.gnu.org/28763 .
* src/stat.c (fmt_terse_fs): Define format for --terse -f here.
(fmt_terse_regular): Define format for --terse here.
(fmt_terse_selinux): Likewise for when SELinux is enabled.
(default_format): Use the above constants.
(usage): Output the formats for the terse modes.
* src/df.c (main): stat() before open(), and avoid
the optional open when given a fifo argument.
* tests/df/unreadable.sh: Add a test case.
* NEWS: Mention the fix.
Fixes https://bugs.gnu.org/29038
* src/ls.c (DEFINE_SORT_FUNCTIONS): Apply _GL_ATTRIBUTE_PURE
to each strcmp-derived function definition, since GCC8 with
-Wsuggest-attribute=pure now warns it is needed.
* man/stat.x (SEE ALSO): Mention statfs(2) in addition to stat(2).
Note statfs() is generally used rather than statvfs(),
so we'll defer that reference to the SEE ALSO section of statfs(2).
Fixes https://bugs.gnu.org/28989
* tests/dd/nocache_eof.sh: Only run the O_DIRECT tests
when 512 byte alignment is supported. Otherwise with older
XFS on systems with > 1MiB pages, or on file systems not
supporting O_DIRECT, there would have been false failures.
* tests/dd/direct.sh: Clarify the skip message.
Previously with oflag=direct the call to invalidate_cache()
was not passed to the kernel, as it was less than a page size,
and a subsequent call was not made to invalidate the pending space.
Similarly with oflag=nocache the pending space at EOF was
not invalidated. Even though these amount to only a single page
in the page cache it can be significant. For example on
XFS before kernel patch v4.9-rc1-4-g0ee7a3f, O_DIRECT files
would have been read inefficiently if any pages were cached,
even if they were already synced to storage.
* src/dd.c (i_nocache_eof, o_nocache_eof): New bools used
to control when we want invalidate_cache(,0) to clear to EOF.
(cache_round): Use IO_BUFSIZE (currently 132KiB) to minimize
calls to the relatively expensive advise function, rather
than page_size. This also makes it clear that while the
kernel function operates on pages, this size is chosen for
performance reasons.
(invalidate_cache): Refactor to share more code between
input and output paths. Use i_nocache_eof and o_nocache_eof
rather than proxying off max_records. Ensure we
invalidate full pages when clearing to EOF as the kernel
will ignore any non complete pages. Fix the offset used
for the output path.
(dd_copy): Invalidate the cache of the input after the
offset is updated, for consistency and so we don't try to
invalidate before the start of the file. When we read
EOF on input, set flags so that we invalidate to EOF.
(main): Invalidate to EOF in more cases, by depending
on the i_nocache_eof and o_nocache_eof flags.
* doc/coreutils.texi (dd invocation): Clarify the alignment
and persisted caveats on the example applying "nocache"
to part of a file.
* tests/dd/nocache_eof.sh: A new test.
* tests/local.mk: Reference the new test.
* NEWS: Mention the bug fix.
Issue reported by Eric Bergen.
This should have been part of commit v8.28-17-gf926f7c
* src/stty.c (check_argument): Align line continuation chars,
and ensure the function macro is immune to usage with if/else.
Suggested by Jim Meyering and Paul Eggert.
This was a latent issue that became significant with
the addition of the -F option in FILEUTILS-3_16n-56-ge46a424
* src/stty.c (apply_settings): Refactor argument checking
to a function macro. Augment the argument check to ignore
NULLed out arguments (already processed -F).
* NEWS: Mention the fix.
* tests/misc/stty-invalid.sh: Add a test case.
Fixes https://bugs.gnu.org/28859
This fixes a regression from commit v8.26-39-g2f69dba
* src/timeout.c (block_cleanup_and_chld): Rename from block_cleanup
to indicate we also block SIGCHLD to avoid the race where SIGCHLD
fires between waitpid() polling and sigsuspend() waiting for a signal.
* NEWS: Mention the fix.
We inherit the signal mask from our parent process,
therefore ensure SIGCHLD is not blocked.
If SIGCHLD is blocked, sigsuspend() won't be interrupted
when the child process exits and we hang until the timeout (SIGALRM).
This fixes a regression from commit v8.26-39-g2f69dba
* src/timeout.c (install_sigchld): Ensure SIGCHLD is unblocked.
* NEWS: Mention the issue.
man pages change little between systems,
so falling back to distributed pages make sense
when cross compiling or lacking perl.
* man/local.mk: Add all man pages to EXTRA_DIST
so that they're distributed in the generated tarball.
Use the dummy-man page generator if cross compiling.
Set TZ to avoid a distcheck failure where man pages
used a diffent month than those rebuilt (with a .timestamp).
* man/dummy-man: Only fall back to generating a stub
if copying an existing man page fails.
* man/help2man: Sync portable TZ=UTC0 specification
from upstream help2man.
* NEWS: Mention the build-related change.
Fixes https://bugs.gnu.org/28574
* tests/split/filter.sh: Due to an invalid 'FILE = zero.in'
construct trying to initialize a FILE variable, it would
instead try to run the FILE command which is present on
macOS 10.13 with APFS.
We also remove a redundant duplicate test clause introduced
during a rebase, and simplify the piped timeout command,
to avoid requiring a subshell and associated quoting.
* THANKS.in: Add the reporter Jack Howarth.
Fixes https://bugs.gnu.org/28506
* tests/misc/expr.pl: Skip the quote varying tests in
the multi-byte locales as these tests aren't that interesting
in those locales. Also ERR_SUBST is already defined for
some tests so awkward to redefine to munge UTF8 quotes to ASCII.
Show offending argument instead of a generic 'syntax error' message.
Suggested by Bernhard Voelker in https://bugs.gnu.org/28461#13 .
* src/expr.c (syntax_error): Remove.
(required_more_args): New function.
(eval7, main): Replace syntax_error call with detailed die message.
* tests/misc/expr.pl: Add tests for new messages.
* doc/coreutils.texi: Remove doubled word.
* src/targetdir.c: Explicitly mark exported function.
* tests/local.mk: This is not a root only test.
* tests/mv/vulnerable-target.sh: Use returns_.
Introduced in v8.28-3-g44ccd1c
This was unintentionally removed in v8.27-60-g2ae1460
* src/shred.c (wipename): Interate through all name lengths.
* tests/misc/shred-remove.sh: Add test cases.
* NEWS: Mention the bug fix.
Fixes https://bugs.gnu.org/28507
* NEWS, doc/coreutils.texi (Target directory): Document this.
* src/cp.c, src/install.c, src/ln.c, src/mv.c: Include targetdir.h.
(target_directory_operand): Use the new targetdir_operand_type
function to check for vulnerable target directories.
* src/cp.c (stat_target_operand): New function.
(target_directory_operand, do_copy): Use it.
* src/local.mk (noinst_HEADERS): Add src/targetdir.h.
(src_ginstall_SOURCES, src_cp_SOURCES, src_ln_SOURCES)
(src_mv_SOURCES): Add src/targetdir.c.
* src/targetdir.c, src/targetdir.h: New files.
* tests/mv/vulnerable-target.sh: New test.
* tests/local.mk (all_root_tests): Add it.
* src/ptx.c (find_occurs_in_text): Die with an appropriate error
diagnostic when the given regular expression returns a match of
length 0.
* tests/misc/ptx.pl (S-infloop): Add a test.
* NEWS (Bug fixes): Mention the fix.
Fixes https://bugs.gnu.org/28417 which was detected using
Symbolic Execution techniques developed in the course of the
SYMBIOSYS research project at COMSYS, RWTH Aachen University.
* tests/ls/hyperlink.sh: If the hostname or any part of
the absolute path would be changed due to URL encoding,
the test would fail. Therefore simplify to remove
these components of the URL from consideration.
* gnulib: The only change in this gnulib update
is the tagging of the fts-tests module as longrunning,
which gnulib-tool currently implicitly excludes.
This test was seen to take about 20s and 285MB.
Reported by Assaf Gordon on space restricted VMs.
* src/tty.c (main): Don't distinguish ENOTTY from other errors,
because isatty() doesn't portably distinguish errors.
Solaris returns ENOENT for all input errors for example.
Musl also returns ENOENT, and ENODEV may be returned as disscussed at:
http://openwall.com/lists/musl/2017/04/06/6
* tests/misc/tty.sh: Adjust accordingly.
On some setups the root:object_r:tmp_t context is invalid.
This does indicate a limitation in the test framework,
but for now we'll relax this to skipping the tests.
The tests still run on a Fedora 25 system for example.
* tests/cp/cp-a-selinux.sh: Upon chcon error, skip rather than ERROR.
* tests/install/install-Z-selinux.sh: Likewise.
* tests/misc/chcon.sh: Likewise.
* tests/misc/runcon-no-reorder.sh: Likewise.
* tests/misc/selinux.sh: Likewise.
* tests/mkdir/restorecon.sh: Likewise.
* src/tty.c (main): All systems mention that isatty()
man return EINVAL as well as (the POSIX compliant) ENOTTY.
Also Centos 6 was seen to return EINVAL from ttyname().
* tests/misc/tty.sh: Fix a test issue where we assume
standard input is always a valid tty.
Reported by Assaf Gordon on OpenSolaris 5.10 and 5.11,
and Centos 6.5
This reverts commit v8.27-97-g8cb06d4 because
the setsid() fallback was not implemented correctly
and disabling the ioctl was not a complete solution
to the security issue of the child being passed
the tty of the parent.
Given runcon is not really a sandbox command,
the advice is to use `runcon ... setsid ...`
to avoid this particular issue.
We only use one of statfs or statvfs for `stat -f`
and on the BSDs we use statfs which doesn't have the
f_namelen member. However on OpenBSD and later FreeBSD
systems statfs does provide f_namemax, so use that.
* NEWS: Mention the improvement for OpenBSD and FreeBSD.
* m4/stat-prog.m4: Check for f_namemax in the statfs struct.
* src/stat.c: Return '?' rather than '*' when we can't
determine the max length of the file system.
* tests/ln/sf-1.sh: This test was failing on all BSDs
due to '*' being returned for the max length which
caused the test to attempt to create 1Mi+1 names.
The test now uses a short name when we can't determine
the max name length to use.
Reported by Assaf Gordon on various BSD based systems.
Similar to the issue with SELinux sandbox (CVE-2016-7545),
children of runcon can inject arbitrary input to the terminal
that would be run at the originating terminal privileges.
The new libseccomp dependency is widely available and used
on modern SELinux systems, but is not available by default
on older systems like RHEL6 etc.
* m4/jm-macros.m4: Check for libseccomp and
warn if unavailable on selinux supporting systems.
* src/local.mk: Link runcon with -lseccomp.
* src/runcon.c (disable_tty_inject): A new function to
disable use of the TIOCSTI using libseccomp, or with setsid()
where libseccomp is unavailable.
* tests/misc/runcon-no-inject.sh: A new test that uses
python to make the TIOCSTI call, and ensure that doesn't succeed.
* tests/local.mk: Reference the new test
* NEWS: Mention the fix.
Addresses http://bugs.gnu.org/24541
Terminals such as iTerm2 and VTE based terminals
(as of version 0.49.1), support hyperlinks when
passed terminals codes as described at:
https://gist.github.com/egmontkob/eb114294efbcd5adb1944c9f3cb5feda
* src/ls.c (gobble_file): Allocate an absolute file name to output.
(quote_name): Output the absolute name with the appropriate codes.
(file_escape): A new function to encode file names as per rfc8089.
(main): Handle the new option and call the file_escape_init() helper.
Disable --dired when --hyperlink is specified.
(print_dir): Get the absolute file name here too, so that the
directory name can be linkified.
* NEWS: Mention the new feature.
* tests/ls/hyperlink.sh: Add a new test.
* tests/local.mk: Reference the new test.
* doc/coreutils.texi (ls invocation): Describe --hyperlink.
This saves about 0.5MB uncompressed from the tarball.
* Makefile.am: Following on from v8.26-34-g2c64bc8
update the oldest documented version to 8.18 which
is now about 5 years old. Also remove older ChangeLogs
that were previously thought to be for changes not
in the git history, but are adequately recorded upon review.
* build-aux/ChangeLog-2007: Remove file.
* lib/ChangeLog-2007: Likewise.
* m4/ChangeLog-2007: Likewise.
This is useful when chaining with other commands that run commands in a
different context, while avoiding using the shell to cd, and thus
having to consider shell quoting the chained command.
* NEWS (New features): Document the new option.
* doc/coreutils.texi (env invocation): Likewise.
* src/env.c (usage): Likewise.
(main): Implement the new option.
* tests/misc/env.sh: Test the new option.
When systemd is configured to automount a remote file system - see
'man systemd.automount(5)', then the mount point is initially
mounted by systemd with the file system type "autofs".
When the resource is used later on, then the wanted file system is
mounted over that mount point on demand.
'df -l' triggered systemd to mount the file system because it called
stat() on the mount point.
Instead of single-casing "autofs" targets, we can avoid stat()ing
all dummy file systems (which includes "autofs"), because those are
skipped later on in get_dev() anyway.
*src/df.c (filter_mount_list): Also skip dummy file systems unless
the -a option or a specific target are given.
* NEWS: Mention the fix.
Co-authored-by: Bernhard Voelker <mail@bernhard-voelker.de>
Fixes http://bugzilla.suse.com/show_bug.cgi?id=1043059
The 'readlink' node has '@findex realpath' in it. This results in
info doc/coreutils.info realpath
incorrectly jumping to the 'readlink' node (instead of the 'realpath'
node). Change it to @cindex instead.
* doc/coreutils.texi (readlink): Change '@findex realpath' to @cindex.
* src/realpath.c (usage): Explicitly say 'DIR' instead of 'FILE' for
--relative-{to,base} parameters, to avoid giving the impression
that regular files can be used as relative base.
* doc/coreutils.texi (realpath): Same.
* src/ls.c (gobble_file): Disable the optimization to avoid quoting
if the symlink target itself needs quoting. This was introduced
with the quoting alignment adjustments in v8.25-106-g01971c0
* tests/ls/symlink-quote.sh: Add a test.
* tests/local.mk: Reference the test.
* NEWS: Mention the fix.
commit v8.27-44-g18f6b22 was too aggressive in
only allowing inotify use with regular files. This will
support responsive processing of `tail -f fifo | ...`
* src/tail.c (any_non_regular): Adjust to allow FIFOs
since inotify supports these well.
* tests/tail-2/inotify-only-regular.sh: Adjust comment.
* tests/misc/seq-epipe.sh: Remove stale comment.
* tests/misc/sort-debug-warn.sh: musl doesn't indicate a set_locale()
failure with missing locales, so avoid a test portion in that case.
* tests/misc/wc-files0.sh: Avoid a bug on older ash implementations.
Addresses http://bugs.gnu.org/28054
Problem reported by Lukas Zachar at:
http://bugzilla.redhat.com/1482445
* src/ptx.c (line_width, gap_size, maximum_word_length)
(reference_max_width, half_line_width, before_max_width)
(keyafter_max_width, truncation_string_length, compare_words)
(compare_occurs, search_table, find_occurs_in_text, print_spaces)
(fix_output_parameters, define_all_fields):
Use ptrdiff_t, not int, for object offsets and sizes.
(WORD, OCCURS): Use ptrdiff_t, not short int.
(WORD_TABLE, number_of_occurs, generate_all_output):
Prefer ptrdiff_t to size_t where either will do.
(total_line_count, file_line_count, OCCURS, fix_output_parameters)
(define_all_fields):
Use intmax_t, not int, for line counts.
(DELTA): Remove. All uses changed.
(OCCURS, find_occurs_in_text, fix_output_parameters):
Use int, not size_t, for file indexes.
(tail_truncation, before_truncation, keyafter_truncation)
(head_truncation, search_table, define_all_fields)
(generate_all_output):
Use bool for booleans.
(digest_word_file, find_occurs_in_text):
Use x2nrealloc instead of checking for overflow by hand.
(find_occurs_in_text, fix_output_parameters, define_all_fields):
Omit unnecessary cast.
(fix_output_parameters): Don’t assume integers fit in 11 digits.
(fix_output_parameters, define_all_fields):
Use sprintf return value rather than calling strlen.
(define_all_fields): Do not rely on sprintf to generate a string
that may contain more than INT_MAX bytes.
(main): Use xstrtoimax, not xstrtoul.
Use xnmalloc to catch integer overflow.
POSIX says sigprocmask has unspecified behavior in a multithreaded
program like ‘sort’.
* src/sort.c (pthread_sigmask) [GNULIB_defined_pthread_functions]:
New macro, for use when ‘sort’ is not multithreaded.
(cs_enter, cs_leave): Use it. Pass address, not value, as
this is typically a tad faster. All callers changed.
Use O_CLOEXEC when creating file descriptors, so that subsidiary
processes do not inherit file descriptors that they do not need.
This is helpful for ‘sort’, as it is a multithreaded program that
forks and execs.
* bootstrap.conf (gnulib_modules): Add mkostemp, open, pipe2.
* src/sort.c (create_temp_file): Open temporary file with O_CLOEXEC.
(stream_open): Open the stream with O_CLOEXEC.
(pipe_fork): Create the pipe with O_CLOEXEC.
(check_output): Open the output file with O_CLOEXEC.
(main): Use xfopen/xfclose to handle --files0-from, so that
O_CLOEXEC is used properly. This is simpler anyway.
* tests/misc/sort-files0-from.pl: Adjust to change in diagnostic
wording.
* src/operand2sig.c (operand2sig): AIX uses a different bit pattern
in the returned status from the wait() functions and from shells.
Therefore hardcode the selection of the lower bits of the number.
* NEWS: Mention the fix.
* configure.ac: Set USE_XLC_INCLUDE when __xlc__ is defined.
* src/local.mk: Use it to select the appropriate include option.
Reported by Michael Felt.
* tests/ln/sf-1.sh: Limit the symlink size to 1MiB
to avoid memory exhaustion seen on NFS on AIX, giving:
+ printf '%0*d' 4294967296 0
+ ./tests/ln/sf-1.sh: line 38: printf: warning: 0: Result too large
* tests/id/setgid.sh: Skip the test when the adjusted gid
would equal 4294967295, as that's reserved on AIX.
Reported by Michael Felt.
* src/sort.c (main): Don't assume hard_LC_COLLATE implies
a successful setting of the locale as musl defaults to
UTF8 when failing to set the specified locale.
* tests/misc/sort-debug-warn.sh: Adjust for the now
separated locale debug info and map the musl specific
message back to the common case.
Addresses https://bugs.gnu.org/28054
* src/seq.c (io_error): Use the same error message as would
be generated at exit time when closing the stdout stream.
The inconsistency was added with commit v8.25-26-gc92585b.
This was noticed due to an inconsistency in the expected
error message generated by seq on musl libc.
Addresses https://bugs.gnu.org/28054
* tests/misc/printf-surprise.sh: With musl libc the
large printf format does succeed, outputting data.
To avoid SIGPIPE being generated we ignore that signal
and then handle the subsequent EPIPE error.
Addresses https://bugs.gnu.org/28054
* configure.ac: Disable some new warnings to avoid false positives.
Building with warnings enabled and latest gcc would evoke build
failure without these changes. Disable the following in coreutils
proper: -Wformat-overflow=2 -Wformat-truncation=2, and
disable these for gnulib: -Wformat-truncation=2 -Wduplicated-branches
* src/shred.c (wipename): As per the comment, the arguments
to error() are sufficiently quoted, so split the call over
multiple lines to avoid the syntax-check.
Suggested by Kamil Dudka in:
http://lists.gnu.org/archive/html/coreutils/2017-07/msg00072.html
* NEWS: Document the changed nature of the fix.
* doc/coreutils.texi, tests/cp/backup-is-src.sh:
* tests/mv/backup-is-src.sh: Revert previous change.
* src/copy.c (source_is_dst_backup): New function.
(copy_internal): Use it. Fail instead of falling back on numbered
backups when it looks like the backup will overwrite the source.
Although this reintroduces a race, it's more compatible with
previous behavior.
* NEWS, doc/coreutils.texi (Backup options): Document the change.
* bootstrap.conf (gnulib_modules): Add backup-rename.
* src/copy.c (copy_internal): Silently switch to numbered backups
if a simple backup might lose data. Use backup_file_rename
to avoid races with numbered backups.
* tests/cp/backup-is-src.sh, tests/mv/backup-is-src.sh:
Adjust to match new behavior.
Use renameat2 to avoid a rename race condition, on recent-enough
GNU/Linux.
* bootstrap.conf (gnulib_modules): Add renameat2.
* src/shred.c: Include renameat2.h.
(wipename): Use renameat2 instead of rename.
* src/shred.c (dopass): shred used to read the input file,
and so needed to ensure internal memory was cleared.
This is no longer the case since SH-UTILS-1_16f-260-gf381610
so avoid this redundant clearing.
(do_wipefd): Likewise.
* NEWS: Remove the recent mention of this issue.
* tests/misc/sync.sh: Normalize the error messages
when syncing a non read/write directory, as AIX
gives the "Is a directory" error.
Also ensure that sync(1) returns an error for this
case on all systems.
Most programs take care to operate on all command-line-specified
operands before exiting. That is an important feature that allows
to identify all problems with the first run. However, groups would
exit upon the first problematic user name.
Bug introduced via commit v6.10-56-g167b8025ac.
* src/groups.c (main): Do not exit immediately upon error.
* tests/misc/groups-process-all.sh: New file. Test for this.
* tests/local.mk (all_tests): Add it.
* NEWS (Bug fixes): Mention this.
* tests/misc/groups-dash.sh: Avoid false failure on a system for which
"none" is a valid user name. The first invocation would succeed, and
the second would fail with "groups: ‘--’: no such user".
Use a user name that cannot exist.
Discussed in https://bugs.gnu.org/26779 .
* NEWS: Mention the improvement.
* bootstrap.conf: Add gnulib modules mbslen,mbschr.
* src/expr.c (mbs_logical_substr): New function to return a substring
based on logical character positions (instead of bytes).
(mbs_logical_cspn): Similar to strcspn/mbscspn, but returns number of
logical characters instead of byte offset.
(mbs_offset_to_chars): New function to return number of logical
characters fitting in a given byte offset.
(docolon): Report matched logical characters instead of bytes.
(eval6): For length/substr/index operations, use logical characters
instead of bytes by calling the above new functions.
* tests/misc/expr.pl: Repeat all tests with non-C locale to detect any
regressions.
* tests/misc/expr-multibyte.pl: New tests with multibyte input.
* tests/local.mk: Add new test file.
When the generated file, doc/constants.texi, happens to be older than
doc/coreutils.info, it will not be updated until/unless its generated
contents change. This is due to way that rule is careful to update
the file, to avoid provoking a pointless rerunning of makeinfo.
Note that this does not happen when one first runs "make distclean",
as recommended in README-release. However, I sometimes run it as
a more-rigorous "make check", and shouldn't have to manually run
"make distclean" first, in that case.
Before this change, one could reproduce the failure by running
`touch -dyesterday doc/constants.texi && make distcheck`. It would
fail with "makeinfo: could not open ../../doc/coreutils.info-t
for writing: Permission denied"
* Makefile.am (dist-hook): Touch the two generated files, so that
they cannot be out of date wrt doc/coreutils.texi.
* src/tail.c (any_non_regular): A new function to check passed files.
(main): Use the above to skip inotify if any non regular files passed
like /dev/tty or /dev/ttyUSB0 etc.
* tests/tail-2/inotify-only-regular.sh: A new test.
* tests/local.mk: Reference the new test.
* NEWS: Mention the bug fix.
Fixes http://bugs.gnu.org/21265 and http://bugs.gnu.org/27368
* src/tail.c: (main): Only issue the warning about -f being
ineffective when we're not going into simple blocking mode.
* tests/tail-2/follow-stdin.sh: Ensure the warning is output correctly.
Fixes http://bugs.gnu.org/27368
This will support use cases like:
tail -f file.log | grep -q trigger &&
process_immediately
* src/tail.c (check_output_alive): A new function that
uses select on fifos or pipes to detect if they're broken.
(tail_forever): Call check_output_alive() periodically.
(tail_forever_inotify): Merge the select() call from
check_output_alive() into the select() originally present
for the --pid case, and adjust accordingly.
* tests/tail-2/pipe-f.sh: Add test cases.
* NEWS: Mention the improvement.
* NEWS: Mention the bug fix.
* src/tail.c (tail_forever_inotify): With --pid, avoid waiting
for new events if there are still events to process.
* tests/tail-2/inotify-dir-recreate.sh: Adjust to trigger.
* src/copy.c (copy_internal): Honor the x->require_preserve flag
for symlinks as we do for ordinary files, so we don't exit with
failure upon failure to chown a symbolic link.
* NEWS: Mention the bug fix.
* src/uptime.c (main): 00-23 was always used for the hour component
of the current time, so remove the AM/PM output (which was only
present in some locales anyway). Also add seconds to the time
to be more consistent with the usual procps-ng uptime implementation
on GNU/Linux.
* NEWS: Mention the fix.
Fixes http://bugs.gnu.org/26783
* HACKING: Change from explicit instructions to using gnulib
provided coverage testing targets. Also include instructions
for adding root only tests to the report.
Fixes http://bugs.gnu.org/26709
Add a test for CVE-2017-7476 which was fixed in gnulib at:
http://git.sv.gnu.org/gitweb/?p=gnulib.git;a=commitdiff;h=94e01571
* tests/misc/date-tz.sh: Add a new test which overwrites enough
of the heap to trigger a segfault, even without ASAN enabled.
* tests/local.mk: Reference the new test.
* NEWS: Mention the bug fix.
Problem reported by Benno Schulenberg (Bug#26621).
* NEWS: Document this.
* src/dd.c (print_xfer_stats): With status=progress,
format times with %.0f rather than %g. Improve
translator comments.
* doc/coreutils.texi (Options for date): Capitalize a sentence.
* tests/misc/date-debug.sh: Adjust --debug output to match
recent changes to Gnulib’s parse-datetime module.
* src/shred.c (fillpattern): Fix the "off by one" issue when
testing whether we have enough space to copy the already
written portion of the buffer to the remainder of the buffer.
Specifically for buffer sizes that are (3*(2^x))+1, i.e. 7,13,...
we both use an uninitialized byte and invoke undefined
behavior in memcpy() operation on overlapping memory regions.
* tests/misc/shred-passes.sh: Add an invocation that will
trigger either valgrind UMR, or ASAN like:
ERROR: AddressSanitizer: memcpy-param-overlap: memory ranges
#1 0x403065 in fillpattern src/shred.c:293
A direct test is awkward due to the random writes surrounding
the problematic pattern writes.
Fixes http://bugs.gnu.org/26545
* src/tail.c (tail_forever_inotify): Add the IN_DELETE_SELF flag when
creating watch for the parent directory. After the parent directory
is removed, an event is caught and then we switch from inotify to
polling mode. Till now, inotify has always frozen because it waited for
an event from a watched dir, which has been already deleted and was not
added again.
* tests/tail-2/inotify-dir-recreate.sh: Add a test case.
* tests/local.mk: Reference the new test.
* NEWS: Mention the bug fix.
Fixes http://bugs.gnu.org/26363
Reported at https://bugzilla.redhat.com/1283760
On some platforms, isatty succeeds but ttyname fails.
POSIX does not seem to allow this, but there it is.
Problem reported by Christian Brauner (Bug#26371).
While we’re at it, check for errors more carefully and return a
new exit status 4 if stdin is closed or a similar error occurs.
* doc/coreutils.texi (tty invocation): Document new behavior.
* init.cfg (stderr_fileno_):
Don't assume have_input_tty is not in the environment.
* src/tty.c (TTY_STDIN_ERROR): New constant.
(main): Exit with nonzero status if there is a usage error,
like other coreutils programs.
Check for error in getting stdin type.
* tests/misc/tty.sh: New file.
* tests/local.mk (all_tests): Add it.
* src/expand-common.c (emit_tab_list_info): A new function to
output the extended info on --tab=LIST, including the new
'+' and '/' prefixes.
* src/expand-common.h: Declare the above.
* src/expand.c (usage:): Call emit_tab_list_info and
match alignment with that used in unexpand --help.
* src/unexpand.c (usage): Likewise.
Support --tabs="1,+8" which is equivalent to --tabs="1,9,17,..."
useful for viewing unified diff output with its 1 character
gutter for example.
* doc/coreutils.texi ({expand,unexpand} invocation): Document,
using diff processing as the example.
* src/expand-common.c (set_increment_size): Update the new
increment_size global.
(parse_tab_stops): Handle the new '+' prefix.
(finalize_tab_stops): Verify both '+' and '/' prefixes
are not used together.
* tests/misc/expand.pl: Add test cases.
* NEWS: Mention the new feature.
* tests/misc/cut-huge-range.sh: On some systems returns_ may
use more memory, so incorporate that in the determination
of the ulimit value to use. Noticed on ARMv7 with bash-4.4.12,
and x86_64 with bash-4.2.37.
Fixes http://bugs.gnu.org/26253
* doc/coreutils.texi (split invocation): Document the new option.
* src/split.c (usage): Likewise.
(main): Process the new option much like --numeric-suffixes,
but with an adjusted alphabet.
* tests/split/numeric.sh: Refactor to support --hex mode.
* NEWS: Mention the new feature.
* src/md5sum.c (split_3): Verify hex digits internally before
triggering the global bsd_reversed mode flag.
(bsd_split_3): Likewise.
* tests/misc/md5sum-bsd.sh: Add a test case.
* NEWS: Mention the bug fix.
Fixes http://bugs.gnu.org/26263
* src/df.c (filter_mount_list): Avoid stat() on
explicitly excluded file systems, which is especially
significant in cases like `-x nfs` which may hang.
* NEWS: Mention the bug fix.
* src/expand-common.c (next_file): We're dependent on calling
this function with NULL to initialize things appropriately.
So enforce this with assert(), which avoids a warning from
clang-anaylzer.
* src/split.c (bytes_split): Don't write to an existing filter
if it has exited. When filters exit early, skip input data if
possible. Refactor out 2 redundant variables.
* tests/split/filter.sh: Improve test coverage given the
new more efficient processing. Also use a 10TB file to
expand the file systems tested on.
commit v8.25-4-g62e7af0 introduced the issue as it
broke out of the processing loop irrespective of
the value of new_file_flag which was used to indicate
a finite number of filters or not.
For example, this ran forever (as it should):
$ yes | split --filter="head -c1 >/dev/null" -b 1000
However this exited immediately due to EPIPE being propagated
back through cwrite and the loop not considering new filters:
$ yes | split --filter="head -c1 >/dev/null" -b 100000
Similarly processing would exit early for a bounded number of
output files, resulting in empty data sent to all but the first:
$ truncate -s10T big.in
$ split --filter='head -c1 >$FILE' -n 2 big.in
$ echo $(stat -c%s x??)
1 0
I was alerted to this code by clang-analyzer,
which indicated dead assigments, which is often
an indication of code that hasn't considered all cases.
* src/split.c (bytes_split): Change the last condition in
the processing loop to also consider the number of files
before breaking out of the processing loop.
* tests/split/filter.sh: Add a test case.
* NEWS: Mention the bug fix.
* tests/misc/sort-debug-keys.sh: Disparate LC_CTYPE and LC_MESSAGES
are not supported, with the result LC_MESSAGES=C is used throughout.
Therefore just set LC_ALL in the test, and normalize the message
variants with sed.
Reported and tested by J Rogowsky.
* src/local.mk [check-duplicate-no-install]: Depend on the
single-binary tr, or the system tr, as the edge case
where these are not available only result in the sanity
check being effectively ignored.
Switching to non-recursive makefiles broke the 'install-html' target:
The gettext plumbing requires an 'install-html' target in po/Makefile.
This was fixed in gettext v0.19.8.1-41-ge5a008a, but packages using
older gettext need to manually patch po/Makefile.in.in.
Reported (for 'sed') and suggested fix by Eric Blake in
https://bugs.gnu.org/25690 .
* bootstrap.conf (bootstrap_epilogue): Add 'install-{html,pdf,dvi,ps}'
targets to po/Makefile.in.in (if needed).
* bootstrap.conf: s/4.13/6.1/ as versions previous to that
generated invalid html with interspersed <span> tags that
were visible to the user. Version 6.1 is available for a
year now, and is available in most distros.
makeinfo issues the following:
doc/coreutils.texi:6568: warning: @sc argument all uppercase,\
thus no effect.
* doc/coreutils.texi (join invocation): Remove the @sc macro around
the all uppercase "GNU".
* src/timeout.c (install_sigchld): A new function to
install the SIGCHLD handler using sigaction() rather
than signal(), because with the latter on solaris
the signal handler is reset to default and thus
sigsuspend() only returns for the first finished child.
Reported by Assaf Gordon.
* src/libstdbuf.c: undef malloc so as libstdbuf is
not linked with gnulib, and anyway the replacement is
never needed since we never malloc(0).
Reported by Assaf Gordon.
* man/md5sum.x: Give a more direct warning againt the use
of this hash algorithm for security purposes.
* man/sha1sum.x: Likewise.
Suggested by Jim Meyering.
* src/expand-common (parse-tab-stops): Exit earlier upon overflow
so another warning isn't issued (on 32 bit) in add_tab_stop().
Flagged in https://hydra.nixos.org/build/49499970
* doc/coreutils.texi (sha1sum invocation): Given that a SHA-1
preimage attack has occurred as documented at http://shattered.io/,
document sha1sum as having the same limitations as md5sum.
(md5sum): Parameterize the warning for use in both cases.
* man/md5sum.x: Mention b2sum(1) as a more secure alternative.
* man/sha1sum.x: Give the same warning as done for md5sum(1).
* doc/coreutils.texi (expand invocation): Document the feature.
(unexpand invocation): Likewise.
* src/expand-common.c (extend_size): A new global to use
when the last tab stop is prefixed by '/'.
(set_extend_size): A new function to validate and set
the new extend_size global.
(parse_tab_stops): Call set_extend_size() for '/' prefixes.
(finalize_tab_stops): Ensure a single specified '/' is
treated like a standard tabsize, but also ensure that
when '/' is specified with a single other entry that
we process as a list rather than a tab size.
(get_next_tab_stop): Use the tab size if set,
for items after the user specified tab position list.
* tests/misc/expand.pl: Add test cases
* NEWS: Mention the new feature.
Fixes http://bugs.gnu.org/25540
This comes from the latest gnulib.
Also handling of OMP_NUM_THREADS has been adjusted
to support comma separated values indicating a nesting level,
in which case the first value is taken. Also OMP_NUM_THREADS=0
is now ignored instead of being treated as 1, to match
the behavior of libgomp.
* NEWS: Mention the OMP_THREAD_LIMIT improvement,
and OMP_NUM_THREADS now handling nested values.
* doc/coreutils.texi (nproc invocation): Describe OMP_THREAD_LIMIT
as a way to set the max value, with OMP_THREAD_LIMIT setting the min.
* tests/misc/nproc-override.sh: A new test to exercise the
updated gnulib code with all combinations of these OMP variables.
* tests/local.mk: Reference the new test.
* src/copy.c (set_process_security_ctx, set_file_security_ctx):
Export for use in cp.c.
* src/copy.h: Likewise.
* src/cp.c (make_dir_parents_private): Call the exported functions
to set the security context for new and updated directories.
* tests/cp/cp-a-selinux.sh: Add a test case.
Fixes http://bugs.gnu.org/25378
* bootstrap.conf, src/base64.c, src/cat.c, src/cksum.c:
* src/head.c, src/md5sum.c, src/od.c, src/split.c, src/sum.c:
* src/tac.c, src/tail.c, src/tee.c, src/tr.c, src/wc.c:
Adjust to renaming of the xsetmode module to xbinary-io,
and of the xsetmode function to xset_binary_mode.
This fixes a bug noted by Eric Blake. Code was using xfreopen to
change files to binary mode, but this fails for stdout when in
append mode. Such code should use xsetmode instead. This affects
only the port on platforms like MS-Windows which distiguish text
from binary I/O.
* bootstrap.conf (gnulib_modules):
Remove xfreopen and add xsetmode. Sort.
* src/base64.c (main):
* src/cat.c (main):
* src/cksum.c (cksum):
* src/head.c (head_file, main):
* src/md5sum.c (digest_file):
* src/od.c (open_next_file):
* src/split.c (main):
* src/sum.c (bsd_sum_file, sysv_sum_file):
* src/tac.c (tac_file, main):
* src/tail.c (tail_file):
* src/tee.c (tee_files):
* src/tr.c (main):
* src/wc.c (wc_file): Use xsetmode, not xfreopen.
* src/force-link.h: Don't include headers already included by system.h
* src/force-link.c: Likewise. Also include system.h and
explicitly mark extern functions as such.
If the file B already exists, commands like 'ln -f A B' and
'cp -fl A B' no longer remove B before creating the new link.
Instead, they arrange for the new link to replace B atomically.
This should fix a race condition reported by Mike Crowe (Bug#25680).
* NEWS, doc/coreutils.texi (cp invocation, ln invocation):
Document this.
* bootstrap.conf (gnulib_modules): Add symlinkat.
* src/copy.c, src/ln.c: Include force-link.h.
* src/copy.c (same_file_ok): It's also OK to remove a destination
symlink when creating symbolic links, or when the source and
destination are on the same file system and when creating hard links.
* src/copy.c (create_hard_link, copy_internal):
* src/ln.c (do_link):
Rewrite using force_linkat and force_symlinkat, to close a window
where the destination temporarily does not exist.
* src/cp.c (main): Do not set x.unlink_dest_before_opening
merely because we are in link-creation mode.
* src/force-link.c, src/force-link.h: New files.
* src/local.mk (copy_sources, src_ln_SOURCES): Add them.
* tests/cp/same-file.sh: Adjust test case to match fixed behavior.
The race is unlikely, as timeout(1) needs to receive a signal
in the few operations between waitpid() returning and exit().
Also the system needs to have reallocated the just released pid
in this time window.
Previously we never disabled the signal handler that sent
the termination signal to the "child" pid. However once waitpid()
has reaped the child, the system is free to allocate that pid,
so we must ensure we don't process any further signals.
* build-aux/gen-lists-of-programs.sh: Build timeout(1) optionally...
* configure.ac: ...predicated on sigsuspend() being available.
* src/timeout.c (block_cleanup): A new function to ensure the
cleanup() handler is disabled after waitpid has returned.
(main): Use sigsuspend() to wait with cleanup() enabled but
disabled once it returns, and thus disabled for the waitpid() call.
(monitored_pid): Change to the more accurate pid_t.
* NEWS: Mention the fix.
Fixes http://bugs.gnu.org/25624
* doc/coreutils.texi (realpath invocation): Mention that realpath
is the preferred command for canonicalization.
(readlink invocation): Likewise.
* man/readlink.x: Likewise.
* src/tail.c (check_fspec): Only enable printing of the file header
if we've actually read some data and this is a new file. Also
move printing of the file header to...
(dump_remainder): ...here, to allow printing only when data read.
* tests/tail-2/overlay-headers.sh: A new test for suspension
and resumption of tail.
* tests/local.mk: Reference the new test.
* NEWS: Mention the fix.
Fixes http://bugs.gnu.org/23539
* tests/tail-2/retry.sh: The replacement of the "missing" directory
is not atomic, and therefore tail(1) can take a different path,
especially if there is a delay between the rmdir(2) and creat(2).
This is noticeable for example with `make coverage` because in
that case the coverage files written by rmdir(1) on exit,
induce a significant delay thus triggering the issue.
* src/tail.c (recheck): Set f->ignore before we
use it to show the appropriate error.
* tests/tail-2/retry.sh: Ensure the "giving up" message
is not presented.
Remove old log files that have corresponding entries
in the source code repository.
This saves about 2.5MB uncompressed, 0.5M compressed.
* Makefile.am (gen-ChangeLog): Adjust to taking all
logs since a particular version (8.15 in this case).
Also mention in the truncated log where to get older entries.
(changelog_etc): Remove the no longer distributed files.
* build-aux/git-log-fix: Remove now unused entries.
* ChangeLog-200[5-8]: Delete.
* doc/ChangeLog-2007: Likewise.
* po/ChangeLog-2007: Likewise.
* old/*: Likewise.
* src/copy.c (punch_hole): Work around an empty definition
of HAVE_FALLOCATE which leads to a build error of:
"error: #if with no expression"
That was triggered by the inclusion of <linux/fs.h> in
commit v8.25-68-g89e1fef with kernel-headers-2.6.18.
Reported by Nelson H. F. Beebe
Problem reported by Paul Wise for Debian, in:
https://bugs.debian.org/851934
This is fallout from the fix for GNU Bug#23035.
* src/date.c (batch_convert): New args TZ and TZSTRING.
All uses changed.
(batch_convert, main): Adjust to parse_datetime2 API change.
(main): Allocate time zone object.
* tests/misc/date-debug.sh: Fix incorrect test case,
caught by the fix.
* tests/misc/date.pl: Test the fix.
* NEWS: Document this behavior, which comes with recent Gnulib.
* doc/coreutils.texi (Formatting file timestamps, du invocation)
(Time conversion specifiers, Setting the time, Options for date):
Mention when -00 is output for numeric time zones.
Be more careful about Internet RFC numbers, ISO 8601, etc.
A lot of this is converting http: to https:.
Also, gmane went away, so remove URLs that no longer work and
are not easy to figure out what they were.
Some of this stuff is so old that it no longer matters anyway.
The above new section looked a bit odd as the only general documentation
in between the utility chapters.
* doc/coreutils.texi (File timestamps): Move to a separate chapter.
Commit 4f650aad was incomplete; it changed NEWS but not the hash,
and introduced a grammar error.
* cfg.mk (old_NEWS_hash): Update via 'make update-NEWS-hash'.
* doc/coreutils.texi (File timestamps): Fix doubled word.
Prompted by a bug report from Scott Deerwester (Bug#25407).
* doc/coreutils.texi (File timestamps): New section.
Revamp other sections to use this new section, and
use more-consistent terminology.
* src/stty.c (apply_settings): A new function refactored
from main() that is used to both check and apply options.
(main): Call apply_settings before we open the device,
so all validation is done before interacting with a device.
* NEWS: Mention the improvement.
* tests/misc/stty.sh: Add a test case.
Run "make update-copyright" and then...
* gnulib: Update to latest with copyright year adjusted.
* tests/init.sh: Sync with gnulib to pick up copyright year.
* bootstrap: Likewise.
* tests/sample-test: Adjust to use the single most recent year.
* tests/chgrp/basic.sh: On some NFS setups a user is
not allowed to set a group on a file even if a member
of that group. Therefore skip this test on remote file systems.
* tests/chgrp/default-no-deref.sh: Likewise.
* tests/chgrp/no-x.sh: Likewise.
* tests/chgrp/posix-H.sh: Likewise.
* tests/chgrp/recurse.sh: Likewise.
* tests/tail-2/inotify-rotate-resources.sh: Change to skipping
on remote file systems in the standard way.
* src/wc.c (wc): Avoid reading the end of the file
when the size is not a multiple of PAGE_SIZE,
as the special case handling for files in /proc and /sys
is only required when st_size is 0 or a multiple of PAGE_SIZE.
* tests/misc/wc-proc.sh: Add a test case.
While st_size would have been incorrect for subsequent
files since v7.1, it was only used since v8.24.
* tests/misc/wc-files0.sh: s/7.1/8.24/
* NEWS: Likewise.
Reported by Bernhard Voelker
* src/wc.c (main): Reset fstatus[0].failed between files when reusing
the fstatus[0] entry in --files-from0 mode. This ensures a stat() is
done for each file, avoiding incorrect counts and redundant reading.
* NEWS: Mention the bug fix.
* tests/misc/wc-files0.sh: Add a test case.
Fixes http://bugs.gnu.org/23073
When running "make check" on a Linux system running SELinux with a
non-MLS policy, tests/mkdir/restorecon.sh test fails with:
chcon: invalid context: root:object_r:tmp_t:s0: Invalid argument
Indeed in such a configuration, contexts cannot have ":s0" suffix.
* init.cfg (get_selinux_type): Refactor this function to here
from various tests. Update to work with a non-MLS policy.
(mls_enabled_): A new function to detect if MLS is enabled.
(skip_if_mcstransd_is_running_): Update to not skip when
MLS is not enabled.
* tests/mkdir/restorecon.sh: Use a valid non-MLS context when needed.
* tests/install/install-Z-selinux.sh: Likewise.
* tests/cp/cp-a-selinux.sh: Likewise.
* tests/misc/selinux.sh: Likewise.
* tests/misc/chcon.sh: Skip if non-MLS as --range used throughout.
Fixes http://bugs.gnu.org/22631
* tests/misc/ls-time.sh: Skip the test rather than ERROR
when `touch -m -d ...` fails (Hurd).
* tests/tail-2/follow-stdin.sh: Avoid false FAILs by ignoring
the variances in sterror output.
* tests/rm/rm-readdir-fail.sh: Likewise. Also avoid ERRORs
on systems that don't define _D_EXACT_NAMELEN.
* src/head.c (elide_tail_bytes_file): Ensure we don't use
st_size unless we've previously used seek() to determine
the CURRENT_POS in the seekable file.
This was seen to cause issue on FreeBSD 11 when the pipe
buffer was filled with `yes | head --lines=-0`, in which
case st_size was 64KiB while ST_BLKSIZE() was 4KiB.
Reported by Assaf Gordon.
... in the case where two or more directories nested in each other are
created and each of them defaults to a different SELinux context.
* src/install.c (make_ancestor): When calling defaultcon(), give it the
same path that is given to mkdir(). The other path is not always valid
wrt. current working directory.
* src/mkdir.c (make_ancestor): Likewise.
* NEWS: Mention the bug fix.
Reported at https://bugzilla.redhat.com/1398913
This was detected with ASAN, but can also be seen without ASAN with:
$ tac - - <&-
tac: standard input: read error: Bad file descriptor
*** Error in `tac': malloc(): memory corruption: 0x...
* src/tac.c (copy_to_temp): Don't close our output stream on
(possibly transient) output error, or on input error.
(temp_stream): clearerr() on the stream about to be reused,
to ensure future stream use is not impacted by transient errors.
* tests/misc/tac-2-nonseekable.sh: Add a test case.
* NEWS: Mention the bug fix.
Fixes http://bugs.gnu.org/25041
Reproduced under UBSAN with `tail -f <&-` giving:
tail.c:2220:18: runtime error: load of value 190,
which is not a valid value for type ‘_Bool'
* src/tail.c (tail_file): Ensure f->ignore is initialized
in all cases where we can't tail the specified file.
* tests/tail-2/follow-stdin.sh: Add a test case which
checks stderr has no UBSAN warnings.
Fixes http://bugs.gnu.org/25041
* NEWS: Mention in improvements about the workaround for
the glibc issue with closed stdin, and the new supported file systems.
* tests/misc/b2sum.sh: Spelling fix.
* src/longlong.h: Sync from gmp repo incorporating:
Protect umul_ppmm with do ... while (0)
Replace obsolete ARC asm 'J' constraints with 'Cal'
Provide umul_ppmm for riscv64
glibc has changed the public define
from _STRING_ARCH_unaligned to _STRING_INLINE_unaligned as per
https://sourceware.org/bugzilla/show_bug.cgi?id=19462
* gl/lib/rand-isaac.c: Cater for both defines.
* gl/lib/randread.c: Likewise.
* src/system.h: Update commented out code.
* tests/ls/quote-align.sh: Remove "total" line
which can vary per file system depending on allocation.
Reported by Assaf Gordon on OpenSolaris (5.11/5.10).
This was detected with:
echo a > a; pr "-S$(printf "\t\t\t")" a -m a > /dev/null
Resulting in ASAN triggering:
====================================================
ERROR: AddressSanitizer: global-buffer-overflow
READ of size 1 at 0x00000041b622 thread T0
#0 0x40506a in print_sep_string ../src/pr.c:2241
#1 0x407ec4 in read_line ../src/pr.c:2493
#2 0x40985c in print_page ../src/pr.c:1802
#3 0x40985c in print_files ../src/pr.c:1618
#4 0x4036e0 in main ../src/pr.c:1136
* src/pr.c (init_parameters): Ensure we only override the
specified separator when it's a single tab, thus matching
the calculated separator length.
* tests/pr/pr-tests.pl: Add a test case.
* NEWS: Mention the fix.
* src/ptx.c (fix_output_parameters): Ensure line_width doesn't
go negative, which can happen when the --width is less
than the --gap-size.
* tests/misc/ptx-overrun.sh: Add a test case that triggers
with ASAN. (Note the longer filename is needed to trigger).
Fixes http://bugs.gnu.org/25011
* tests/rm/rm-readdir-fail.sh: ASAN correctly indicated
that fts was writing to freed memory. This was because
we reused a single dirent in our readdir() test wrapper.
Since fts was deallocating those dirents, we now get
a new dirent for each call to our readdir wrapper.
ASAN reported this error for: split -n2/3 /dev/null
ERROR: AddressSanitizer: negative-size-param: (size=-1)
#0 0x7f0d4c36951d in __asan_memmove (/lib64/libasan.so.2+0x8d51d)
#1 0x404e06 in memmove /usr/include/bits/string3.h:59
#2 0x404e06 in bytes_chunk_extract src/split.c:988
#3 0x404e06 in main src/split.c:1626
Specifically there would be invalid memory access
and subsequent processing if the chunk to be extracted
was beyond the initial amount read from file (which is
currently capped at 128KiB). This issue is not in a
released version, only being introduced in commit v8.25-4-g62e7af0
* src/split.c (bytes_chunk_extract): The initial_read != SIZE_MAX
should have been combined with && rather than ||, but also this
condition is always true in this function so remove entirely.
* tests/split/b-chunk.sh: Add a test case.
Fixes http://bugs.gnu.org/25003
In recent commit v8.25-93-g7fc7206 we used the f->remote flag
which wasn't set in all cases. This was detected with
ASAN giving this error when reading f->remote;
runtime error: load of value 190,
which is not a valid value for type '_Bool'
* src/tail.c (fremote): Query the system even without inotify.
(recheck): Always set f->fremote for valid files.
Installing with -j2 or greater could result in a failure like
/bin/sh: line 29: /P/bin/install: Permission denied
when /P/bin/install specifies your PATH-selected install program.
This would arise because we're using "install" to install all
man/*.1 files, and that command would run concurrently with the one
that installs "/P/bin/install" itself. We would run this command:
"src/ginstall src/ginstall /P/bin/install", and it would result
in intervals during which the destination file does not exist,
is empty or incomplete and not executable. We addressed this
problem long ago for installation of actual binaries by telling the
installation rules to use our just-built bin/ginstall (only when not
cross-compiling) rather than the PATH-resolved "install" program.
This change is to do the same for those .1 files.
* src/local.mk (INSTALL): Override automake's default of something
like "INSTALL = /P/bin/install -c".
(INSTALL_PROGRAM): Now that we set INSTALL, there is no longer any
need to set this derived variable. Its default definition,
"INSTALL_PROGRAM = ${INSTALL}" does what we require.
Improved by Eric Blake.
* src/comm.c (total_option): Add bool variable for the new option.
(TOTAL_OPTION): Add enum value.
(long_options): Add array element for the new option.
(usage): Document the new option here.
(compare_files): Count the lines in total[3], and output the summary at
the end.
(main): Accept the new option.
* doc/coreutils.texi (comm invocation): Document it.
* tests/misc/comm.pl: Test it. While at it, improve the test data
to have 1 unique line in the first file, 2 unique lines in the second
file, and 3 common lines.
* NEWS (New Features): Mention the new option.
Fixes http://bugs.gnu.org/24929
* tests/misc/printf-quote.sh: FreeBSD 11 was seen to treat
\u0378 as a printable character. Therefore change to
using the \u0081 C1 control character. We use the UTF-8
representation because our printf implementation explicitly
disallows \u0081 as input.
Reported by Assaf Gordon
On BSD /bin/sh it was seen that unexported env vars passed to
returns_() would not be propagated to the wrapped command.
* cfg.mk (sc_prohibit_env_returns): Add a syntax check to disallow.
* tests/misc/csplit-io-err.sh: Rearrange to export vars in a subshell.
* tests/rm/rm-readdir-fail.sh: Likewise.
* tests/misc/nohup.sh: Export and unset vars around returns_.
* tests/misc/printenv.sh: Likewise.
Reported by Assaf Gordon
This provides better alignment when some names are quoted,
which also provides better indication that quotes are not
part of the name.
* src/ls.c (align_variable_outer_quotes): A new variable
set when ls is aligning columns (not using -m, non-zero -w),
and has a variable quoting style (shell, shell-escape, c-maybe).
(quote_name_buf): Writes to buffer rather than FILE,
taking care to avoid data copying if possible. Refactored from...
(quote_name): ...here. This now manages the buffer passed
to quote_name_buf() and outputs the padding, colors and name
in the appropriate order, while managing the --dired offsets.
(get_color_indicator): A new function to return the color sequence,
refactored from...
(print_color_indicator): ...here. This now simply outputs.
(print_dir): Refactor common parts to quote_name().
(clear_files): Reset the flag indicating at least one
file is quoted in the current directory.
(needs_quoting): A new function to indicate at the scan stage
whether a name needs quoting. Called from...
(gobble_file): ...here, until we find the first quoted file.
(print_name_with_quoting): Mostly refactored to quote_name().
* tests/ls/quote-align.sh: A new test for various output formats.
* tests/local.mk: Reference the new test.
* NEWS: Mention the improvement.
Note we don't support the --algorithm option of the
b2sum command in the external BLAKE2 project, as that
was deemed too confusing for users.
"BLAKE2b" was chosen as the default algorithm to use,
which is single threaded but performs well on 64 bit.
* src/blake2: CC0 source copied from external project.
* cfg.mk[VC_LIST_ALWAYS_EXCLUDE_REGEX]: Exclude blake2/
from syntax checks, make update-copyright, etc.
* src/local.mk: Reference the sources for b2sum,
and set the compilation flags.
* doc/coreutils.texi (b2sum invocation): Reference the
md5sum invocation node, and add descriptions of -l.
* tests/misc/b2sum.sh: Add new test.
* tests/local.mk: Reference new test.
* AUTHORS: Add new binary.
* README: Likewise.
* build-aux/gen-lists-of-programs.sh: Likewise.
* man/.gitignore: Likewise.
* scripts/git-hooks/commit-msg: Likewise.
* man/b2sum.x: New man page template.
* man/local.mk: Reference new template.
* src/.gitignore: Ignore new binaries.
* src/blake2/.gitignore: Ignore new build atrifacts.
* src/md5sum.c (usage): Describe the new -l option.
* NEWS: Mention the new program.
* Makefile.am (my-distcheck): Remove all .deps directories before the
recursive diff that searches for left-behind files. Otherwise, with
automake master (some time after v1.15), "make distcheck" would fail
due to those directories being left behind after "make distclean".
* src/copy.h (cp_options): Add a new flag for install(1).
* src/copy.c (copy_internal): For cp, also output remediation
advice which also indicates why directories aren't copied by default.
The message is unchanged for install(1).
* src/cp.c (cp_option_init): Init install_mode to false.
* src/mv.c (cp_option_init): Likewise.
* src/install.c (cp_option_init): Init install_mode to true.
* tests/install/basic-1.sh: Add a test case.
* tests/cp/link-deref.sh: Adjust test case.
Fixes http://bugs.gnu.org/24958
* tests/misc/sort-h-thousands-sep.sh: The sv_SE locale
may be available while sv_SE.utf8 is not. Therefore
test and use the same locale, now definfed in a variable.
Reported by Jim Meyering.
The previous "returns_"-using change failed to convert many
uses of "$?". Convert all but two of the remaining ones.
* tests/ls/stat-vs-dirent.sh: Likewise.
* tests/misc/head-write-error.sh: Likewise.
* tests/misc/nice.sh: Likewise.
* tests/misc/nohup.sh: Likewise.
* tests/misc/stdbuf.sh: Likewise.
* tests/misc/sync.sh: Likewise.
* tests/tail-2/pid.sh: Likewise.
* tests/tail-2/wait.sh: Likewise.
Thanks to Bernhard Volker for spotting this.
This reduces a standard coreutils install size by about 160K.
* src/cat.c: Change to proper_name() which removes about 18K text.
* src/cp.c: Likewise.
* src/df.c: Likewise.
* src/du.c: Likewise.
* src/getlimits.c: Likewise.
* src/realpath.c: Likewise.
* src/split.c: Likewise.
* src/stdbuf.c: Likewise.
* src/timeout.c: Likewise.
* src/truncate.c: Likewise.
* src/local.mk: Remove -llibiconv from the above programs.
* cfg.mk (sc_check-AUTHORS): Adjust to use factor(1).
* AUTHORS: Adjust to use ASCII to satisfy sc_check-AUTHORS.
The previous commit v8.25-96-g22063c8 lets 'make syntax-check' fail,
because the above check falsely detects a case where 'returns_' is
already in use.
* cfg.mk (sc_prohibit_and_fail_1): Add 'returns_' to exclude list.
* src/tail.c (tail_file): On failure to open a file,
set ignore=true when --retry is not specified.
* tests/tail-2/assert-2.sh: Adjust to the new behavior.
* tests/tail-2/retry.sh: Add a test case. Also change
from `tail ... && fail=1` to the more robust `returns_ 1 ...`
construct which detects segfaults etc.
* NEWS: Document the fix.
* src/tail.c (tail_forever): The BLOCKING optimization is only
enabled for non regular files (which can't be truncated), so ensure
we don't enable that unless we've a valid file descriptor.
* tests/tail-2/retry.sh: Add a test case.
* NEWS: Mention the bug fix.
* src/tail.c (tail_forever): Only read up to st_size on network
file systems to avoid the issue with a stale attribute cache
returning a smaller st_size than we have already read().
The was seen with glusterfs at least and caused the complete
file to be repeatedly output due to assuming the file was
truncated in this case.
* NEWS: Mention the fix.
* src/tail.c (ignore_pipe_or_fifo): Mark the descriptor as -1
for pipes so that any_live_files() detects correctly that
the entry is no longer live.
* tests/tail-2/pipe-f.sh: Add a test case.
* NEWS: Mention the fix.
Fixes http://bugs.gnu.org/24903 which was detected
using Symbolic Execution techniques developed in
the course of the SYMBIOSYS research project at
COMSYS, RWTH Aachen University.
* doc/coreutils.texi(cut invocation): Give a more accurate description
of cut field handling limitations.
* tests/misc/stat-fmt.sh: Fix the test header to be more general.
* tests/tail-2/retry.sh: Spellings.
* src/system.h (emit_backup_suffix_note): A new function to
output the backup suffix info. The strings are unchanged,
so translations are not impacted.
* src/cp.c (usage): Use the new function.
* src/ln.c (usage): Likewise.
* src/mv.c (usage): Likewise.
* src/install.c (usage): Likewise.
* src/cp.c (main): Avoid the getenv("SIMPLE_BACKUP_SUFFIX") call,
which is now done if needed in the gnulib backupfile module.
Also avoid the redundant strdup, as we don't modify this suffix.
* src/install.c (main): Likewise.
* src/ln.c (main): Likewise.
* src/mv.c (main): Likewise.
Fixes http://bugs.gnu.org/23153
Incorporating these coreutils specific changes:
backupfile: initialize default suffix within the implementation
strftime,strptime: support %q to represent the quarter
Also sync these copies with gnulib:
* bootstrap: Now supports AIX.
* gnulib: Update to the latest.
* tests/init.sh: dash(1) is now disabled.
* doc/coreutils.texi (stat invocation): Describe the
QUOTING_STYLE values now supported.
* src/stat.c (getenv_quoting_style): A new function called
from main, that sets the default quoting style for quotearg.
(main): Call getenv_quoting_style() when %N specified.
* tests/misc/stat-fmt.sh: Add a test case.
* NEWS: Mention the improvement.
Fixes http://bugs.gnu.org/23422
* src/dd.c (parse_integer): Suggest to use "00x" instead of "0x",
which is significant for the "count", "seek", and "skip" operands.
* tests/dd/misc.sh: Add a test case.
Fixes http://bugs.gnu.org/24874
* NEWS: Mention the fix.
* src/md5sum.c (digest_file): Add a new MISSING parameter to
return whether the file was missing, separately from the digest.
* tests/misc/md5sum.pl: Add a test case.
Fixes http://bugs.gnu.org/24795
* src/expr.c (integer_overflow): Remove an abort-after-die.
* src/paste.c (write_error): Likewise.
* src/sort.c (badfieldspec): Likewise.
* src/tr.c (string2_extend): Likewise. Also remove a few
break statements, each after an abort.
die() has the advantage of being apparent to the compiler
that it doesn't return, which will avoid warnings in some cases,
and possibly generate better code.
* cfg.mk (sc_die_EXIT_FAILURE): A new syntax check rule to
catch any new uses of error (CONSTANT, ...);
* src/die.h (die): New file/function from grep.
Note: we expect this file to migrate to gnulib.
* src/csplit.c: Include die.h.
(check_format_conv_type): Use die in place of error-nonzero;break;
* src/install.c (strip): Likewise.
* src/nl.c (proc_text): Likewise. This also suppresses a new warning
from GCC 7's -Werror=strict-overflow.
* src/tail.c (parse_options): Likewise.
* src/basename.c (main): Adjust "fall through" comment
so that GCC 7's -Wimplicit-fallthrough honors it.
* src/cp.c (main): Add a "fall through" comment.
* src/ls.c (gobble_file): Likewise.
(get_funky_string): Adjust a "fall through" comment so it is
recognized.
* cfg.mk (exclude_file_name_regexp--sc_system_h_headers): Add die.h
to this list of exempt src/*.h files.
* src/pr.c (main): Avoid this warning from GCC 7:
src/pr.c:1119:6: error: assuming signed overflow does not occur when \
simplifying conditional to constant [-Werror=strict-overflow]
if (n_files == 0)
* src/rm.c (main): Ensure the full --no-preserve-root
option is specified, rather than allowing --n etc.
* tests/rm/r-root.sh: Add a test case.
* NEWS: Mention the change in behavior.
Improved by Jim Meyering.
Fixes http://bugs.gnu.org/24604
which was not the case when inotify was not available.
* src/tail.c (any_live_files): Simplify, since the IGNORE
flag is now only set when a file should be ignored indefinitely.
(recheck): Only output the "giving up on name" message
when that's actually the case. Only set the IGNORE flag
when ignoring a file indefinitely.
(tail_file): Likewise.
* tests/tail-2/retry.sh: Add a test case. Also run
all existing test cases with and without inotify.
NEWS: Mention the fix.
THANKS.in: Add the reporter.
Fixes http://bugs.gnu.org/24495 which was detected
using Symbolic Execution techniques developed in
the course of the SYMBIOSYS research project at
COMSYS, RWTH Aachen University.
* src/who.c (idle_string): This function would fail to compile
with -Werror and today's built-from-git gcc due to this warning:
src/who.c: In function 'print_user':
src/who.c:201:36: error: may write format character ':' at offset 4 \
past the end of the destination [-Werror=format-length=]
sprintf (idle_hhmm, "%02d:%02d",
^~~~~
The fix is to use an assertion to inform gcc of the existing
invariant that guarantees the number of hours is less than 24.
The old code mishandled --time-spec='+%%b', as it misinterpreted
the '%b' as being the month abbreviation. Also, it mishandled
the extremely unlikely case of a month abbreviation containing '%'.
The performance part of this patch sped up 'ls' by about 1% on my
little benchmark of 'ls -lR' on the source directory in the
en_US.UTF-8 locale (Fedora 24 x86-64).
* NEWS: Document the bug fix.
* src/ls.c (first_percent_b, abformat_init): New static functions.
(ABFORMAT_SIZE): New constant.
(use_abformat): New static var.
(abmon, required_mon_width): Remove these static vars.
(abmon_init): Now accepts a pointer to abmon, and returns a boolean.
All callers changed. Reject month abbrs containing '%', as these
would mess up strftime. Simplify mbsalign result checking,
since (size_t) -1 exceeds ABFORMAT_SIZE.
(abformat_init, align_nstrftime): Precompute all 24 formats at
startup, rather than computing a format for each time stamp.
(decode_switches): Call abformat_init instead of abmon_init.
(align_nstrftime): Accept recentness bool instead of format.
All callers changed.
* tests/misc/time-style.sh: Test for format with '%%b'.
Given that `ttyname` is already conditionally declared in src/system.h,
other declarations are redundant and problematic for example in ChromeOS
which has a new FORTIFY implementation that, as an artifact of how it's
implemented, causes the compiler to complain if certain standard library
functions are redeclared without special compiler-specific attributes.
* src/pinky.c: Remove declaration (which was unused anyway).
* src/who.c: Remove declaration.
* src/system.h (getlogin, getuid, geteuid, getgrgid, getpwuid, ttyname):
Add the parameter type to the declarations to avoid warnings when these
backup declarations are used.
This doesn't affect the generated code on my system; it's merely
a cleaner way to use the recently-introduced Linux-specific API.
* m4/jm-macros.m4 (gl_CHECK_ALL_HEADERS): Check for linux/fs.h.
* src/copy.c: Include <linux.fs.h> if available.
(FICLONE) [__linux__]: Define if not already defined.
(clone_file): Use FICLONE instead of the older BTRFS_IOC_CLONE,
as this ioctl is no longer btrfs-specific.
* bootstrap.conf (gnulib_modules): Add flexmember.
* src/sort.c: Include flexmember.h.
(struct tempnode): Make the last member flexible.
(create_temp_file): Port to strict C11/C99 rules for
allocation alignment with flexible array members.
* src/factor.c (lbuf_putc): Only buffer more than a line
when not using the tool interactively.
* NEWS: Mention the bug fix.
Fixes http://pad.lv/1620139
Postpone installation of signal handlers until they're needed.
That is right before the first escape sequence is printed.
* src/ls.c (signal_setup): A new function refactored from main()
to set and restore signal handlers.
(main): Move signal handler setup to put_indicator()
so that the default signal handling is untouched as long as possible.
Adjusted condition for restoring signal handlers to reflect the change.
(put_indicator): Install signal handlers if called for the very first
time. It uses the same code that was in main() prior to this commit.
* NEWS: Mention the improvement.
See https://bugzilla.redhat.com/1365933
Fixes http://bugs.gnu.org/24232
* src/tr.c (unquote): Use indices of type "unsigned int", rather
than a mix or "size_t" and "int", presuming that no command-line-
specified string will have length longer than UINT_MAX.
* src/ptx.c (fix_output_parameters): Switch to an unsigned type that
matches the OCCURS.file_index type. This avoids the following error
from gcc-7.0.0 20160829 (experimental):
src/ptx.c:1220:14: error: assuming signed overflow does not occur \
when simplifying conditional to constant [-Werror=strict-overflow]
if (file_index > 0)
* src/cp.c (make_dir_parents_private): Use default permissions for
created directories when --no-preserve=mode is specified.
* tests/cp/cp-parents.sh: Add a test case.
* NEWS: Mention the fix.
Fixes http://bugs.gnu.org/24251
* src/expand.c, src/unexpand.c: Move global variables from here...
* src/expand-common.h, src/expand-common.c: ... to here.
* src/expand.c, src/unexpand.c: (parse_tab_stops, validate_tab_stops,
next_file): Move identical functions to new module.
(add_tab_stop): Move to new module, including additional code from
'unexpand' (keeping max_column_width) which will have no effect in when
used in 'expand'. Refactor common next-column calculation code
into a new function 'get_next_tab_column'.
* src/local.mk: (src_expand_SOURCES, src_unexpand_SOURCES): Add
'expand-common.c'; (noinst_HEADERS): Add 'expand-common.h'.
* po/POTFILES.in: Add 'expand-common.c'.
* tests/misc/expand.pl: Add more tests.
* tests/misc/unexpand.pl: Likewise.
* TODO: Move conclusions to above test after investigation.
Use hash table for seaching in filter_mount_list() and get_dev()
This improves performance for 20K mount entries from:
real 0m1.731s
user 0m0.532s
sys 0m1.188s
to:
real 0m1.066s
user 0m0.028s
sys 0m1.032s
* src/df.c (devlist_table): Define hash table.
(devlist_hash): Add hash function.
(devlist_compare): Add hash comparison function.
(devlist_for_dev): Add lookup function.
(devlist_free): Add cleanup function.
(filter_mount_list): Use the above hash table.
While at it, rename the variable 'devlist' to 'seen_dev' for
better readability.
(me_for_dev): Use the above lookup function.
NEWS: Mention the improvement.
THANKS.in: Remove the committer; add original submitter Josef Cejka.
* src/date.c (usage): This was thought to introduce other ambiguities,
and was inconsistent with the format presented in the touch(1) docs.
* doc/coreutils.texi (date invocation): Likewise.
See http://bugs.gnu.org/24077
* cfg.mk (sc_THANKS_in_sorted): This check would fail on systems
for which "." is not ignored. Add a quick sort-based check for
that error, and skip the check on any broken system.
* src/date.c (usage): Make the characters used to summarize
the input format, match the output +FORMAT characters.
* doc/coreutils.texi (date invocation): Likewise.
Fixes http://bugs.gnu.org/24077
The following is a reproducer for the wrong dependency:
$ ./configure --enable-install-program=arch
$ make
$ rm -f src/arch man/arch.1
$ make man/arch.1
GEN man/arch.1
help2man: can't get `--help' info from man/arch.td/arch
Try `--no-discard-stderr' if option outputs to stderr
Makefile:14378: recipe for target 'man/arch.1' failed
make: *** [man/arch.1] Error 127
* man/local.mk (man/arch.1): Change to depend on src/arch rather than
src/uname: while the arch binary depends on uname.c and uname-arch.c,
its man page depends on the arch binary.
Reported downstream by Rodrigues Goldwyn <rgoldwyn@suse.com> in
https://build.opensuse.org/request/show/415172
* src/sort.c (traverse_raw_number): Accept thousands separator only
if it is immediately followed by a digit.
* tests/misc/sort-h-thousands-sep.sh: Cover the fix for this bug.
Suggested by Pádraig Brady in http://bugs.gnu.org/24015
* src/sort.c (traverse_raw_number): Allow to skip only one occurrence
of thousands_sep to avoid finding the unit in the next column in case
thousands_sep matches as blank and is used as column delimiter.
* tests/misc/sort-h-thousands-sep.sh: Add regression test for this bug.
* tests/local.mk: Reference the test.
* NEWS: Mention the bug fix.
Reported at https://bugzilla.redhat.com/1355780
Fixes http://bugs.gnu.org/24015
* src/sort.c (traverse_raw_number): New function for traversing numbers.
(find_unit_order): Use traverse_raw_number() instead of open-coding it.
(debug_key): Likewise.
* doc/coreutils.texi (install invocation): Update -Z documentation.
* src/install.c (make_ancestor): Set default security context before
calling mkdir() if the -Z option is given.
(process_dir): Call restorecon() on the destination directory if the
-Z option is given.
(usage): Update -Z documentation.
* tests/install/install-Z-selinux.sh: A new test for 'install -Z -D'
and 'install -Z -d' based on tests/mkdir/restorecon.sh.
* tests/local.mk: Reference the test.
* NEWS: Mention the improvement.
Reported at https://bugzilla.redhat.com/1339135
Fixes http://bugs.gnu.org/23868
* tests/rm/rm-readdir-fail.sh: Since we use the returned dirent
from the readdir wrapper it must be the correct type and not
just cast. Therefore setup so that we only have to define a
wrapper for readdir(), which works appropriately on 32 and 64 bit.
Issue reported by Bernhard Voelker, where rm was seen to invoke
rmdir() on invalid file names.
Follow-up to commit c92585b1: improve epipe test script, with
suggestions from Bernhard Voelker and Pádraig Brady.
see: http://lists.gnu.org/archive/html/coreutils/2016-04/msg00067.html .
* tests/misc/seq-epipe.sh: Avoid too many subshells,
stricter validation of errors and output from seq,
skip if trapping SIGPIPE is not supported.
* src/stty.c: Disable setting of "swtch" to ^z as that
conflicts with and disables using ^z as "susp".
* NEWS: Mention the bug fix.
Reported and tested by Rich Burridge at:
http://bugs.gnu.org/23866
* tests/rm/rm-readdir-fail.sh: A new test to simulate readdir()
failing immediately or after returning a few entries, and verifying
that rm does the appropriate thing.
This was initially reported at:
http://bugzilla.opensuse.org/show_bug.cgi?id=984910
where it was mentioned that readdir() may fail
when an NFS server has a poor readdir cookie implementation.
* gnulib: Update to latest.
* NEWS: Specifically mention the fts readdir() fix
and reindent to standard indentation.
* tests/init.sh: Update from gnulib.
* src/yes.c: Include full-write.h.
(main): Use full_write, not write, to simplify handling of partial
writes. Don't bother using stdio to output data; just use
full_write with a buffer as large as needed. Reuse operand
strings if possible, and if the buffer would otherwise be large.
* src/yes.c (main): Loop over the write buffer to
handle the case where write may write less than requested.
* NEWS: Mention the bug fix.
Reported by Paul Eggert.
GCC 7 warned about undefined behavior in this unlikely case.
Problem reported by Jim Meyering in: http://bugs.gnu.org/23825
* src/md5sum.c (main):
* src/paste.c (main):
* src/yes.c (main):
Avoid undefined behavior when argc == INT_MAX.
* doc/coreutils.texi (sort invocation): Mention in the summary
dicussion that --key is used to specify fields. Give a summary
in the --key description, of the most common use case of specifying
a field, and that by default those fields include the blank separators
at the start of each field in the comparisons.
Print warning regardless of locale, avoid warning if key is zero width.
Problem reported by Karl Berry in http://bugs.gnu.org/23665 .
* src/sort.c: (key_warnings): change conditions for 'leading spaces'
warning.
* tests/misc/sort-debug-warn.sh: adjust tests accordingly.
Easier troubleshooting of individual 'sort --debug' messages.
Discussed in http://bugs.gnu.org/23665 .
* tests/misc/sort-debug-warn.sh: add progress number before each sort
invocation.
* src/dircolors.hin: Reduce the list by replacing all specific
mentions of "color" with the pattern "*color*". This will also
cater for other entries like "konsole-256color".
Fixes http://bugs.gnu.org/23542
* stc/stat.c (human_fstype): Add file system ID definition,
and use "smb2" as the name.
* NEWS (Improvements): Mention the change.
Fixes http://bugs.gnu.org/23516
* src/id.c (main): When configured with --enable-gcc-warnings and using
the very latest gcc built from git, building would fail with this:
src/id.c:200:8: error: assuming signed overflow does not occur when \
simplifying conditional to constant [-Werror=strict-overflow]
bool default_format = (just_user + just_group + just_group_list
^~~~~~~~~~~~~~
Rewrite to use bool-appropriate operators.
Ensure I/O errors are detected (and terminate seq), preventing seq
from infloop (or running for long time with a large
range) upon write errors or ignored SIGPIPE. Examples:
seq 1 inf > /dev/full (seq_fast)
seq 1.1 0.1 inf >/dev/full (print_numbers)
* src/seq.c (io_error): A new function to diagnose appropriate
stdio errors and exit the program with failure status.
(seq_fast, print_numbers): Explicitly check for write errors
and terminate the program with diagnostic.
* tests/misc/seq-io-errors.sh: Test error detection with /dev/full.
* tests/misc/seq-epipe.sh: Test error detection with broken pipes.
* tests/local.mk: Add new tests.
* NEWS: Mention the fix.
* src/stat.c (human_fstype): Add file system ID definition,
and use "wslfs" as the name.
* NEWS (Improvements): Mention the change.
Fixes http://bugs.gnu.org/23273
* src/sleep.c (main): Allow ERANGE since we allow "inf" values.
* src/timeout.c (parse_duration): Likewise.
* tests/misc/sleep.sh: New file. Tests for sleep(1).
* tests/misc/timeout-parameters.sh: Add case for newly allowed
$LDBL_MAX. Also use returns_ throughout the file. Also avoid
small timeout values which might give false failures under load.
* cfg.mk (sc_prohibit_expr_unsigned): expr uses intmax_t internally
when GMP is not available, so flag any uses of unsigned limits.
* tests/misc/timeout-parameters.sh: Remove the overflow related
to UINT_MAX as it's handled by the following case. Change the
following case from 99... to $TIME_T_OFLOW to be more expressive
in what it's actually testing. Directly check that commands succeed,
rather than inspecting $? afterwards.
* tests/dd/skip-seek-past-dev.sh: Guard against large device sizes.
* tests/id/setgid.sh: Protect: Guard against large group IDs.
* tests/misc/cut-huge-range.sh: Tweak comment to avoid syntax check.
It's tempting to simplify to just skip the test if expr fails,
but SIZE_MAX > INTMAX_MAX is the common case.
* tests/cp/parent-perm-race.sh: This new race introduced in
commit v8.25-5-g632eda5 is quite hard to hit, but is due to
`ls > fifo` doing write()/close()/exit() once `cp` has
open() the source fifo. Then the subsequent comparison of the
destination file may fail due to the file being missing or empty.
Previously `ls` generated output that was independent of `cp`.
Now we must wait for `cp` to finish before inspecting the
destination file that it wrote.
* src/seq.c (isnan): Define macro.
(scan_arg): Add check if the given argument is NaN, and exit with
a proper error diagnostic in such a case.
(usage): Document it.
* tests/misc/seq.pl: Add tests.
* doc/coreutils.texi (seq invocation): Document the change.
* NEWS (Changes in behavior): Mention the change.
* src/seq.c (main): Exit with an error diagnostic when the given
step value is Zero.
(usage): Document it.
* doc/coreutils.texi (seq invocation): Likewise.
* tests/misc/seq.pl: Add tests.
* NEWS (Changes in behavior): Mention the change.
Reported by Маренков Евгений in:
http://bugs.gnu.org/23110
* src/stat.c (human_fstype): Add file system ID definition.
* NEWS (Improvements): Mention the change, moving the previously added
"prl_fs" change note from "Changes in behavior" to here.
Fixes http://bugs.gnu.org/23283
* src/stat.c (human_fstype): Identify the parallels file system.
Also tag as remote so that tail(1) doesn't use inotify, which
fails to detect changes made outside a VM.
Fixes http://bugs.gnu.org/23143
The problematic code computed a struct tm in one time zone, and
then printed it or converted it to a string in another. To be
portable the same time zone needs to be used for both operations.
On GNU platforms this is not an issue, but incorrect output can be
generated on System V style platforms like AIX where time zone
abbreviations are available only in the 'tzname' global variable.
Problem reported by Assaf Gordon in: http://bugs.gnu.org/23035
* NEWS: Document the bug.
* src/date.c (show_date):
* src/ls.c (long_time_expected_width, print_long_format):
* src/pr.c (init_header):
* src/stat.c (human_time): Use localtime_rz instead of localtime,
so that the time zone information is consistent for both localtime
and time-formatting functions like fprintftime and nstrftime. For
'stat' this change is mostly just a code cleanup but it also
causes stat to also print nanoseconds when printing time stamps
that are out of localtime range, as this is more consistent with
what other programs do. For programs other than 'stat' this fixes
bugs with time zone formats that use %Z.
* src/du.c, src/pr.c (localtz): New static var.
(main): Initialize it.
* src/du.c (show_date): New time zone argument, so that localtime
and fprintftime use the same time zone information. All callers
changed.
* tests/misc/time-style.sh: New file.
* tests/local.mk (all_tests): Add it.
* tests/misc/date.pl: Test alphabetic time zone abbreviations.
* gl/lib/regcomp.c.diff: Remove file, now that gnulib's
regcomp.c compiles regex.c with -Wno-unused-parameter.
* gl/lib/regex_internal.h.diff: Likewise.
* gl/lib/regex_internal.c.diff: This file induced a change to ensure
that the "Idx" type was unsigned and to remove a few "VAR < 0"
comparisons. These days, it is probably fine to stay in sync with
gnulib/glibc's copies
of these files, so remove these patches, too.
* gl/lib/regexec.c.diff: Likewise.
Prompted by a report by Assaf Gordon and a suggestion from Paul Eggert.
* tests/split/filter.sh: Use OFF_T_MAX-1 rather than OFF_T_MAX
as the size of a test file, to avoid false failure on an XFS file
system (or any file system permitting a file of size OFF_T_MAX).
Reported as http://bugs.gnu.org/22931
POSIX recommends avoiding -a and -o, for good reason.
src/test.c (usage): Mention that inherent ambiguities exist with
binary -a and -o.
Problem reported by Martin Gebert in: http://bugs.gnu.org/22909
sort, tail, and uniq now support traditional usage like 'sort +2'
and 'tail +10' on systems conforming to POSIX 1003.1-2008 and later.
* NEWS: Document this.
* doc/coreutils.texi (Standards conformance, tail invocation)
(sort invocation, uniq invocation, touch invocation):
Document new behavior, or behavior's dependence on POSIX 1003.1-2001.
* src/sort.c (struct keyfield.traditional_used):
Rename from obsolete_used, since implementations are now allowed
to support it. All uses changed.
(main): Allow traditional usage if _POSIX2_VERSION is 200809.
* src/tail.c (parse_obsolete_option): Distinguish between
traditional usage (which POSIX 2008 and later allows) and obsolete
(which it still does not).
* src/uniq.c (strict_posix2): New function.
(main): Allow traditional usage if _POSIX2_VERSION is 200809.
* tests/misc/tail.pl: Test for new behavior.
Problem reported by Yanyan Jiang 蒋炎岩 in: http://bugs.gnu.org/22769
* doc/coreutils.texi (sort invocation):
Mention possibility of 'sort -o F F' data loss during crashes.
* src/split.c (lines_rr): Reinstate the conditional
setting of the WROTE boolean, as otherwise split -n r/1 would
consume all input when all --filter commands are stopped.
There was a test in place to check for this, but
it was incorrect as detailed below.
(input_file_size): Immediately disallow --number with
non seekable inputs, as such an invocation is not currently
generally supported and will fail as the data overflows
the internal buffer.
* tests/split/l-chunk.sh: Adjust to again disallow -n /dev/zero.
Also change all '&& fail=1' checks to use the 'returns_ 1' form.
* tests/split/filter.sh: Change the no longer supported /dev/zero
case to a regular $OFF_T_MAX file (supported on XFS for example).
Also fix the timeout(1) commands so they're not subject to
pipefail issues.
* src/stty.c (usage): Remove an erroneous call to translate an
empty string, added in commit v8.23-112-g564f84a, which results
in the gettext header being printed for translated languages.
* THANKS.in: Remove the now committer.
* NEWS: Mention the bug fix.
* tests/cp/parent-perm-race.sh: Add timeouts so that the test does
not wait forever on GNU/Hurd. This does not fix the underlying
bug but at least lets the tests make progress.
Problem reported by Nelson H.F. Beebe in: http://bugs.gnu.org/22624
Also, check that 'cp' actually copies the data; this is a better
test anyway, and simplifies the test code.
Also, avoid unlikely race if 'pid' is set in the environment.
Problem reported by Nelson H.F. Beebe in: http://bugs.gnu.org/22624
Other problems also fixed: basically, the code got confused because
GNU/Linux reports that /dev/zero has size zero.
* src/split.c (input_file_size): Now takes struct stat *, not just
size. Always store the first buffer. All callers changed. Treat
/dev/zero as an infinitely-large file, both on GNU/Linux where
fstat and lseek say its size is zero, and on GNU/Hurd where they
say the size is OFF_T_MAX.
(cwrite): Return true on success.
(bytes_split): Don't try to read past EOF, and stop if a write fails.
(lines_rr): Omit stray check for ignorable errno.
(main): Get file size only when n_units > 1, since that's the only
time it is needed. Defer most of the work to input_file_size.
* tests/split/l-chunk.sh: Adjust tests to match new behavior
on oddball inputs.
* src/date.c (usage): The colon is used in the timezone offset
since commit v8.24-64-g17bbf6c.
* THANKS.in: Remove committer.
Fixes http://bugs.gnu.org/22491
Solaris Studio 12 on sparc (not x86) will not remove
unused functions, thus leaving a reference to an
undefined program_name symbol from emit_try_help().
* src/system.h (emit_try_help): Change from an inline function
to a macro, so that the inline function is not actually
defined in libstdbuf.c.
Fixes http://bugs.gnu.org/22430
* tests/tail-2/wait.sh: Restrict inotify specific test portion
to local file systems and also not with ---disable-inotify specified.
Failure noticed on NFS.
* tests/rm/rm1.sh: Also remove the group write bit which
was required on one NFS setup at least. Note u-w was
enough to deny file creation, g-w was also required to
deny file removal.
* tests/rm/cycle.sh: Likewise.
* tests/mv/perm-1.sh: Likewise.
* tests/cp/preserve-slink-time.sh: Add a delay between the
ln and the cp so that there is enough difference between
the timestamps so GPFS won't discard the update.
Reported by Assaf Gordon.
* tests/misc/shred-passes.sh: Specify an exact amount to shred,
to avoid running out of simulated random data on file systems
with a large st_blksize like GPFS for example.
Reported by Assaf Gordon.
tests/tail-2/F-headers.sh and test/tail-2/retry.sh fail on
on remote file systems due to tail going into inotify mode
due to not being able to determine the remoteness of the
non existent files.
* src/tail.c (any_non_remote_file): A new function used
to disable inotify when there are no open files, as
we can't determine remoteness in that case.
* NEWS: Mention the bug fix.
* gnulib: Update to latest where the only change is to
not unconditionally enable leaf optimization for fts on NFS,
as it was seen to abort() with some NFS servers as per:
https://bugzilla.redhat.com/1299169
This affects utilities that traverse directories like
cp, rm, chmod etc.
* NEWS: Adjust the improvement message to leave only XFS.
* tests/rm/dangling-symlink.sh: The effective delay was only
1.5s. So delay before the operation, and increase the iteration
count by 1 to increase the delay to 6.3s. This failure was
noticed once on a FreeBSD 10.2 x86_64 virtual machine.
* tests/dd/stats.sh: Remove quotes on sleep argument
to ensure arguments are accumulated appropriately.
* tests/du/move-dir-while-traversing.sh: sleep before operation,
to increase the effective delay from 1.6s to 3.2s.
* tests/tail-2/flush-initial.sh: Likewise.
* tests/tail-2/pipe-f2.sh: Sleep first so that the effective delay
matches the commented value.
* tests/rm/r-root.sh: This test was seen to ERROR due to
sed input not having a terminating '\n'. Therefore just skip
with diagnostics upon failure to adjust the error output with sed.
Reported by Assaf Gordon.
* init.cfg (dump_mount_list_): A new function to output the
system mount list.
* tests/df/df-symlink.sh: Call dump_mount_list_ upon failure.
* tests/df/over-mount-device.sh: Likewise.
* tests/df/problematic-chars.sh: Likewise.
* tests/df/skip-rootfs.sh: Likewise.
* tests/df/total-verify.sh: Likewise.
* tests/df/unreadable.sh: Likewise.
* tests/df/df-symlink.sh: Only check the symmetry of
the source <-> target lookup, when the source is only mounted once,
which isn't the case if '.' is a BTRFS subvolume for example.
Reported by Assaf Gordon on a default OpenSUSE 42.1 install.
* src/copy.c (copy internal): Remember directories irrespective
of their link count, because on some file systems like BTRFS,
directories always have a link count of 1.
* tests/misc/date.pl: This test was dependent on perl
unconditionally calling the system localtime() as date(1) does.
However perl has its own configure checks in place to avoid buggy
localtime implementations. Therefore comment out this test
as a FIXME for now. This test was seen to fail on FreeBSD 10
where localtime() returns nonsense values instead of failing
for arguments >= 2^56.
* tests/misc/head-c.sh: Further increase the limit, to account
for the fact that head(1) needs at least 1.576MB over the base test
command on FreeBSD 10. 4MB should be enough headroom while
still providing over allocation protection.
Reported by Assaf Gordon.
* src/operand2sig.c (operand2sig): Add a detailed comment explaining
why we validate even very large shell exit status values.
* tests/misc/kill.sh: Add a test case for the ksh scheme.
Simplify the INVALID signal number determination which also avoids
a false failure on systems like FreeBSD 10 with incomplete
signal list (caused by inaccurate NSIG).
* tests/cp/fiemap-extents.sh: Support RHEL6 fallocate
which doesn't support IEC suffixes like "MiB" on numbers.
Also add some extra framework_failure_ protections.
Reported by Assaf Gordon.
Locale categories are not equivalent on OpenBSD,
and LC_COLLATE only supports "C" for example.
Now LC_ALL is supported to set multiple other categories
on OpenBSD, so setlocale(LC_ALL, "") returns a string
indicating which categories were updated and which ignored.
Therefore...
* src/sort.c (main): ...Call setlocale(LC_COLLATE, "")
to explicitly check whether a specified LC_ALL or
LC_COLLATE environment variable value is supported
for the LC_COLLATE category. Also use !! to explicitly
convert to bool to support c89 systems where bool is an int,
and thus would get values > 1.
Reported by Assaf Gordon.
doc/coreutils.texi (dd invocation): Add oflag=sync to the streaming
example. Also reference the "direct" flag. Mention this is only
a request to the system.
* src/dd.c (usage): Mention the "sync" flag along with "nocache".
Also mention that it's only a request to drop the cache.
* THANKS.in: Add reporter Francois Rigault.
* src/stat.c (usage): Mention the '#' and '0' flags are useful with %a.
* doc/coreutils.texi (stat invocation): Likewise. Also give an example
printing unambiguous octal output.
Reported at http://bugs.debian.org/810539
commit v8.23-31-g90aa291 failed to consider this case,
where the previous rename has failed, thus causing the
following to remove the specified directory:
mv dir dir dir
* src/copy.c (copy_internal): Assume this rename attempt has
succeeded, as a previous failure will already have been handled,
and we don't want to remove the source directory in this case.
* tests/cp/duplicate-sources.sh: Consolidate this test file to...
* tests/mv/dup-source.sh: ...here. Add test cases for same
source and dest.
* tests/local.mk: Remove the consolidated test.
* NEWS: Mention the bug fix.
Reported at https://bugzilla.redhat.com/1297464
* src/wc.c (write_counts): Shell escape the file name
if it contains '\n' so only a single line per file is output.
* tests/misc/wc-files0.sh: Add a test case.
* NEWS: Mention the improvement.
* doc/coreutils.texi (numfmt invocation): Reference the description.
* src/numfmt.c: Parameterize '\n' references.
* tests/misc/numfmt.pl: Add tests for character and field processing.
* NEWS: Mention the new feature.
* doc/coreutils.texi (paste invocation): Reference -z description.
* src/paste.c (main): Parameterize the use of '\n'.
* tests/misc/paste.pl: Add test cases.
* NEWS: Mention the new feature.
* src/comm.c (main): Track the output delimiter length,
so that it can be adjusted to 1 for the NUL delimiter.
Also rename the global variable from "delimiter" to
"col_sep" so its use is more obvious, and to distinguish
from the recently added "delim" global variable.
* tests/misc/comm.pl: Adjust accordingly.
* doc/coreutils.texi (comm invocation): Reference option description.
* src/comm.c (main): Use readlinebuffer_delim() to support
a parameterized delimiter.
* tests/misc/comm.pl: Add test cases.
* NEWS: Mention the new feature.
* doc/coreutils.texi (tac invocation): Mention the
NUL delineation with an empty --separator.
* src/tac.c (main): Allow an empty separator when -r not specified.
* tests/misc/tac.pl: Add test cases.
* NEWS: Mention the new feature.
Fixes http://bugs.gnu.org/8103
* doc/coreutils.texi (cut invocation): Reference the description.
* src/cut.c: Parameterize '\n' references.
* tests/misc/cut.pl: Add tests for character and field processing.
* NEWS: Mention the new feature.
* NEWS: Mention the change in behavior.
* doc/coreutils.texi (newlineFieldSeparator): A new description,
referenced from ({join,sort,uniq} invocation).
* src/system.h (field_sep): A new inline function to determine
if a character is a field separator.
* src/join.c (usage): s/whitespace/blank/ to be more accurate
wrt which characters are field separators.
(xfields): s/isblank/field_sep/.
* src/sort.c (inittables): Likewise.
* src/uniq.c (find_field): Likewise.
* tests/misc/join.pl: Adjust -z test, and add a test/example
for processing the whole record with field processing.
* tests/misc/sort.pl: Add -z test cases, including case with '\n'.
* tests/misc/uniq.pl: Add -z -f test case with \n.
At least the false positive in tail-2/follow-stdin.sh could be seen
on a 4-core i5 system with -j8. Fix similar cases. too.
* tests/tail-2/follow-stdin.sh: Empty the 'out' file in each iteration.
Otherwise, under heavy system load, 'check_tail_output' would see the
expected output from the previous round before tail would have the
chance to come up in the background.
While at it, move the creation of the 'exp' file out of the loop.
* tests/dd/stats.sh: Empty the output file of the background process
here, too.
* tests/misc/cat-buf.sh: Likewise.
* tests/misc/stdbuf.sh: Likewise.
* tests/tail-2/wait.sh: Likewise.
This is used to discard terminal output on FreeBSD based systems.
* src/stty.c (mode_info): Add FLUSHO where supported.
(usage): Document for "local" and "combination" settings.
* doc/coreutils.texi (local stty settings): Document,
mentioning that this option is supported but ignored on GNU/Linux.
Instead of commit v8.24-132-g5171bef which only provides
control to disable this behavior (with -I), provide
the symmetrical "[-]drain" special setting.
* src/stty.c (main): Parse the [-]drain setting instead of -I,
and treat like a global option.
(usage): Adjust accordingly.
* tests/misc/stty.sh: Test "drain" with and without options.
* NEWS: Mention the new feature.
In some cases an initial drain may block indefinitely as discussed at:
http://lists.gnu.org/archive/html/coreutils/2016-01/msg00007.html
* src/stty.c (main): Use TCSANOW rather than TCSADRAIN if -I specified.
(usage): Document the new option.
* doc/coreutils.texi (stty invocation): Likewise.
* tests/misc/stty.sh: Ensure -I is supported.
* NEWS: Mention the new feature.
* doc/coreutils.texi (coreutils): Remove this direntry
so that `info coreutils` will display the table of contents,
rather than the unlikely intended coreutils Multi-call invocation.
This also supports lookups like `info coreutils readlink`,
but note that only works with the standard info viewer,
and not with pinfo for example.
* src/install.c (usage): Mention this commonly required functionality
in the -D option description.
* doc/coreutils.texi (install invocation): Likewise for the
--target-directory description.
* src/install.c (main): As an optimization, when calling
install_file_in_dir() for each file, only attempt to create
the target directory once, as this is invariant over the loop.
* src/install.c (mkancesdirs_safe_wd): Unconditionally
restore the current working directory when possibly called
multiple times (from install_file_in_dir()).
* tests/install/create-leading.sh: Add a test case.
* NEWS: Mention the fix.
Fixes http://bugs.gnu.org/21497
Since commit v8.23-60-g414a8cf, the above check failed to find sources
where the length of the longest line is a 3 or more digit number.
* cfg.mk (sc_long_lines): Remove the '\' escape character before '{'
and '}', because this would mean literal '{' and '}' characters in
an extended regular expression in sed(1).
Run "make update-copyright" and then...
* gnulib: Update to latest with copyright year adjusted.
* tests/init.sh: Sync with gnulib to pick up copyright year.
* bootstrap: Likewise.
* tests/sample-test: Adjust to use the single most recent year.
Problem noted by Pádraig Brady in: http://bugs.gnu.org/22277#8
Also, make the output a bit more precise while we're at it.
* NEWS: Document this.
* src/dd.c (previous_time): Remove, replacing with ...
(next_time): New var. All uses changed.
This avoids some rounding errors, and should be a bit faster.
(newline_pending): Remove, replacing with ...
(progress_len): New var. All uses changed.
This lets us keep track of how many trailing spaces to append.
(print_xfer_stats): Get the time first thing, so that it's
closer to being correct. Count the bytes output, and append
trailing spaces if needed. Add remarks to translators about
translation lengths.
Problem reported by Linda Walsh in: http://bugs.gnu.org/17505
* NEWS: Document this.
* doc/coreutils.texi (dd invocation): Use a simpler script.
Adjust output example to match new behavior.
* src/dd.c (human_size): Remove.
All uses changed to use human_readable and ...
(human_opts): ... this new constant.
(abbreviation_lacks_prefix): New function.
(print_xfer_stats): Use it. Output both --si and --human-readable
summaries, but only if they have prefixes.
* tests/dd/reblock.sh, tests/dd/stats.sh: Test new behavior.
Mainly for these changes:
- freadptr: fix to work with ungetc on all uClibc configs
- fts: enable leaf optimization for XFS
- fts: ensure leaf optimization used for NFS
- strftime-tests: avoid false failure on OS X
- intprops-tests: avoid warnings (causing CI failures)
* NEWS: Update with items from above that are
significant from the previous coreutils release.
* doc/coreutils.texi (Connectives for test): Add notes
on precedence and associativity. Also mention the
portability caveats with these operators.
* cfg.mk: Avoid sc_prohibit_test_minus_ao for coreutils.texi.
Fixes http://bugs.gnu.org/22216
Commit v8.24-116-g896006b changed the file permissions for the tarball
using ug+rw. Better to let the umask decide whether the file should
be group-writeable or not.
* Makefile.am (dist-hook): Change chmod call from 'ug+rw' to '+rw'.
* init.cfg (get_min_ulimit_v_): Increase the determined
virtual memory limit for the command by a page size
to avoid alignment issues causing false failures for
subsequent runs.
* tests/misc/cut-huge-range.sh: Be defensive and match
the cut invocations under test, more closely with the
form used to determine the VM memory limit.
This was noticed on i686 linux at:
http://hydra.nixos.org/build/28990456
* Makefile.am (dist-hook): Ensure all files are writable
so that extracted tarballs are easy to remove, without
getting prompts about removing read-only files.
Reported by Benno Schulenberg.
We were erroneously skipping blanks in the marked comparison
_after_ the key start offset was applied.
* src/sort.c (debug_keys): Don't skip starting blanks
if already handled by begfield().
* tests/misc/sort-debug-keys.sh: Add a test case.
* NEWS: Mention the bug fix.
Fixes http://bugs.gnu.org/22155
* src/stat.c (human_fstype): Add "overlayfs", and tag it as "remote"
to ensure that tail continues to use the more conservative polling mode.
* README-release: Remove a stale comment about updating fremote()
in tail.c. Also give a link to *_SUPER_MAGIC definitions
not in the standard linux/magic.h location.
* NEWS: Mention that this file system is recognized.
procps, bash >= 4.4, and util-linux >= 2.23
all have the -L option to output a table.
* src/kill.c (main): Have -L synonymous with -t.
Luckily there is no signal that begins with 'L'.
* tests/misc/kill.sh: Update the test.
The character 'K' representing numeric user input can be confusing,
especially in the context where it can be suffixed by the multipliers
K, M, G, etc.
Use NUM instead.
* doc/coreutils.texi (head invocation): Replace @var{k} by @var{num}.
Also change @var{count} in the paragraph about the obsolete option
syntax.
(tail invocation): Likewise. Furthermore, adjust the option order to
alphabetical sorting.
* src/head.c (usage): s/K/NUM/ in as placeholder in the -c and -n
options. Furthermore, print the actual DEFAULT_NUMBER instead of the
hard-coded 10.
* src/tail.c (usage): s/K/NUM/ in as placeholder in the -c and -n
options. Add "[+]" to clarify that NUM may optionally be preceded by
this sign. Remove the redundant explanation of the optional '+' prefix
for NUM, as that is now clear from the option description above.
Suggested by Dan Jacobson in http://bugs.gnu.org/22042
* man/head.x: Add "SEE ALSO" section referencing tail(1).
* man/tail.x: Add "SEE ALSO" section referencing head(1).
Suggested by Dan Jacobson in http://bugs.gnu.org/22041
Use a consistent date in the generated man pages
even if SOURCE_DATE_EPOCH is not set.
* Makefile.am: Generate .timestamp in the tarball.
* man/local.mk: Pass .timestamp if available to help2man.
* man/dummy-man: Don't bother with the year in the stub.
Reported in http://bugs.debian.org/806321
Sync with version 1.47.3 which provides these significant changes:
- support for reproducible builds by using $SOURCE_DATE_EPOCH.
- Adjust spacing of italic text at roman/italic boundaries.
* doc/local.mk (constants.texi): Avoid calling makeinfo unless
the constants derived from tail and shred have changed.
This avoids a dependence on makeinfo from patched sources,
and avoids an expensive makeinfo call for developers.
* doc/coreutils.texi (uniq invocation): Fix the description of
problematic input to say "blank lines" rather than "two or more
blank lines".
(tr invocation): Clarify that -s only works with SET1 when
not translating (when SET2 not specified). Also explicitly
state in examples where blank lines are deleted.
Also add "deleting" to the menu item.
* src/tr.c (usage): Improve the -s summary to say it always
operates on the last specified SET.
* doc/coreutils.texi (tail invocation): Provide an example using awk
to convert tail ==> file <== headers to file: prefixes on each line.
Suggested by Stephen Shirley.
fallocate can allocate extents beyond EOF via FALLOC_FL_KEEP_SIZE.
Where there is a gap (hole) between the extents, and EOF is within
that gap, the final hole wasn't reproduced, resulting in silent
data corruption in the copied file (size too small).
* src/copy.c (extent_copy): Ensure we don't process extents
beyond the apparent file size, since processing and allocating
those is not currently supported.
* tests/cp/fiemap-extents.sh: Renamed from tests/cp/fiemap-empty.sh
and re-enable parts checking the extents at and beyond EOF.
* tests/local.mk: Reference the renamed test.
* NEWS: Mention the bug fix.
Fixes http://bugs.gnu.org/21790
* src/md5sum.c (digest_check): Update a matched_checksums bool upon
matched checksum, and fail (loudly unless --status is specified)
if there were no matches. Also change properly_formatted_lines
to a bool while at it since we don't need to track the plurality.
* tests/misc/md5sum.pl: Add a test case.
Suggested by Jim Meyering.
* doc/coreutils.texi (md5sum invocation): Document the new option.
* src/md5sum.c (digest_file): Return an empty digest to indicate
a missing file.
(digest_check): Don't fail or output status given an empty checksum.
(usage): Document the new option.
(main): Process and validate the new option.
* tests/misc/md5sum.pl: Add new test cases.
* NEWS: Mention the new feature.
Fixes http://bugs.gnu.org/15604
This can be useful if you want to further process data
from process substitutions. For example:
datagen | tee >(md5sum --tag) > >(sha256sum --tag) | sort
* doc/coreutils.texi (tee invocation): Mention that -p is
useful with pipes that may not consume all data.
Add an example, similar to the one above.
* THANKS.in: Add Jirka Hladky.
* tests/ls/stat-failed.sh: Skip the test if 'd' is returned as the type,
and document where this was seen. Also flag failure to write small
temp files during the test as an error rather than a failure.
Fixes http://bugs.gnu.org/21130
* src/local.mk: Run `mkdir -p src` in all our explicit rules,
as in a VPATH build the src/ dir is only created at configure time
as a side effect of dependency tracking generation.
discussed in:
http://lists.gnu.org/archive/html/coreutils/2015-10/msg00091.html
* src/csplit.c: (save_line_to_file): check fwrite failures, report
and exit immediately instead of deferring to 'close_output'.
* tests/misc/csplit-io-err.sh: test fwrite failure using LD_PRELOAD.
* tests/local.mk: add new test.
programs may not be built due to missing system dependencies,
or any program can be excluded at configure time with
--enable-no-install-program. So ensure we're not testing the
system versions in these cases.
* init.cfg (print_ver_): Call require_built_ first.
* tests/misc/tty-eof.pl: Skip programs not built.
* tests/Coreutils.pm (run_tests): Likewise.
* tests/misc/ls-misc.pl: Use 'env test' rather than abs path.
* tests/misc/test-diag.pl: Likewise.
* tests/local.mk: Adjust include order for dependencies.
* tests/misc/arch.sh: Remove redundant calls to require_built_.
* tests/misc/chroot-fail.sh: Likewise.
* tests/misc/groups-dash.sh: Likewise.
* tests/misc/groups-version.sh: Likewise.
* tests/misc/stdbuf.sh: Likewise.
* tests/cp/acl.sh: Remove problematic call to print_ver_ [gs]etfacl.
* tests/mv/acl.sh: Likewise.
* cfg.mk (sc_env_test_dependencies): A new syntax check to enforce
specifying dependencies with print_ver_ for programs
specified through the env command.
* du/bigtime.sh: Add new print_ver_ dependencies.
* du/max-depth.sh: Likewise.
* dd/ascii.sh: Likewise.
* tests/ls/capability.sh: Likewise.
* tests/ls/root-rel-symlink-color.sh: Likewise.
* tests/misc/chroot-fail.sh: Likewise.
* tests/misc/readlink-fp-loop.sh: Likewise.
* tests/misc/sort-debug-keys.sh: Likewise.
* tests/readlink/can-e.sh: Likewise.
* tests/readlink/can-f.sh: Likewise.
* tests/readlink/can-m.sh: Likewise.
* tests/tail-2/inotify-race.sh: Likewise.
* tests/tail-2/inotify-race2.sh: Likewise.
* tests/touch/no-create-missing.sh: Likewise.
* tests/touch/no-dereference.sh: Likewise.
* tests/misc/printenv.sh: Tweak to avoid syntax check trigger.
* tests/misc/help-version.sh: Likewise.
* tests/misc/yes.sh: Likewise.
* tests/misc/printf-quote.sh: Use previously unused $prog.
* configure.ac (EXTRA_MANS): Add $gl_no_install_prog to the list
so that check-x-vs-1 syntax check is satisfied.
A side effect of this cleanup is we no longer
depend on our own kill command being built.
* init.cfg (require_trap_signame_): A new function to verify
that the shell supports specifying traps by signal name.
(require_kill_group_): A new function to ensure the shell
supports sending a signal to a group.
* tests/du/move-dir-while-traversing.sh: Ensure trap supports
signal names.
* tests/misc/stty-invalid.sh: Likewise.
* tests/misc/stty-pairs.sh: Likewise.
* tests/misc/stty-row-col.sh: Likewise.
* tests/misc/stty.sh: Likewise.
* tests/misc/sort-compress.sh: Likewise. Also simplify trap call.
* tests/install/trap.sh: Likewise.
* tests/misc/timeout.sh: Likewise.
* tests/dd/stats.sh: Likewise. Also use default kill command.
* tests/misc/timeout-group.sh: Likewise.
* init.cfg (require_dirent_d_type_): Don't use df -x
to exclude XFS, since this depends on a correct mtab
which is brittle and often not correct within chroots.
* tests/d_type-check: Check also the d_type of files,
which excludes XFS appropriately. Specify all argument
and return types to avoid truncated pointers being passed,
which skipped the test due to crashes on x86_64 at least.
Simplify the C library lookup by reusing the interpreter's.
chroot issue reported at https://bugzilla.redhat.com/1263341
* src/paste.c (main): Use our styled wrapper for quotearg_colon().
* cfg.mk (sc_prohibit-quotearg): A new syntax check to avoid
future uses of unstyled quotearg to one of the internal slots,
and thus destined for diagnostic output.
* src/paste.c (main): Setting the quoting style to "escape"
went against the intent of the comment about presenting
doubled backslashes to the user. Instead use "c-maybe"
which is the only mode which avoids doubled backslashes,
and provides protection against arbitrary control characters.
* tests/misc/paste.pl: Adjust accordingly.
* src/ls.c (decode_switches): Set "shell-escape" if isatty().
* doc/coreutils.texi (ls invocation): Update the defaults description.
* NEWS: Mention the change in behavior. It should not have
backwards compat issues, but mentioning here just in case.
* src/printf.c (usage): Mention the new format.
(print_formatted): Handle the quoting by calling
out to the quotearg module with "shell-escape" mode.
* doc/coreutils.texi (printf invocation): Document %q.
* tests/misc/printf-quote.sh: New test.
* tests/local.mk: Reference new test.
* NEWS: Mention the new feature.
This is mainly noticeable when the multi-byte code
within ls.c is triggered by multi-byte quotes.
$ seq 200000 | xargs touch
$ time ls-old -U --quoting=locale --hide-control-chars >/dev/null
real 0m0.483s
$ time ls-new -U --quoting=locale --hide-control-chars >/dev/null
real 0m0.430s
* src/ls.c (quote_name): Avoid rescanning the output looking for
unprintable chars when we know the quoting mode already escapes them.
* tests/misc/ls-misc.pl: Add tests for all quoting modes, with and
without -q, to verify this assumption.
* doc/coreutils.texi (ls invocation): Describe the new
'shell-escape' and 'shell-escape-always' quoting options.
* src/ls.c (usage): Mention the new quoting options.
* tests/misc/ls-misc.pl: Add a test for 'shell-escape'
Quote file names using the "shell-escape" or "shell-escape-always"
methods, which quote as appropriate for most shells,
and better support copy and paste of presented names.
The "always" variant is used when the file name is
embedded in an error message with surrounding spaces.
* cfg.mk (sc_error_shell_quotes): A new syntax check rule
to suggest quotef() where appropriate.
(sc_error_shell_always_quotes): Likewise for quoteaf().
* src/system.h (quotef): A new define to apply shell quoting
when needed. I.E. when shell character or ':' is present.
(quoteaf): Likewise, but always quote.
* src/*.c: Use quotef() and quoteaf() rather than quote()
where appropriate.
* tests/: Adjust accordingly.
* src/md5sum.c: Use the same file name escaping method used
when generating and checking checksums. I.E. ensure a single line
per file by starting the line with '\' for any file name containing '\n'
and replacing those with "\\n".
* NEWS: Move the item from changes in behavior to improvements,
since this is no longer a backwards incompat change when
processing stdout status messages.
* tests/misc/md5sum.pl: Remove quotes from expected status output.
* tests/misc/sha1sum.pl: Likewise.
Related to commit v8.24-61-g6796698 this provides
more consistent quoting, as quotearg_colon() defaults
to "literal" quoting by default, while quote()
provides appropriate quoting for diagnostics by default.
* gl/modules/randread: Depend on quote module rather than quotearg.
* gl/lib/randread.c: Used quote() not quotearg_colon().
* src/: Likewise.
* src/shred.c: Likewise. Also avoid unnecessary quoting
introducing overhead when wiping names.
* cfg.mk: Relax the matching expression to allow
"qname" variables as used in shred.c to satisfy the check.
* tests/: Adjust accordingly.
* doc/coreutils.texi: (tail invocation): Add missing -s,
along with the existing --sleep-interval description.
(csplit invocation): s/--suffix/--suffix-format/.
(head invocation): Use same variable (COUNT) for -n and --head-count.
(seq invocation): Add opindex items for all options.
(ptx invocation): Likewise.
Fix typo s/--flac-truncation/--flag-truncation/.
(touch invocation): State explicitly that -d takes a parameter,
which also indicates that an '=' is not to be used
for the short option syntax.
(ls invocation): Likewise for the -w option.
Fixes http://bugs.gnu.org/21809
* doc/coreutils.texi (du invocation): Remove 'like' from
mentions of ISO-8601 as the components are individually conformant.
(ls invocation): Likewise, except for --time-style=iso for recent
files, where the MM-DD component is not a valid ISO-8601 timestamp.
* src/date.c (main): Use %:z rather than %z with --iso-8601
as the standard states to consistently use extended format.
Note either format can be parsed by date.
* tests/misc/date.pl: Adjust accordingly.
* doc/coreutils.texi (du invocation): Clarify that "iso"
time styles are only similar to ISO-8601.
(ls invocation): Likewise.
(date invocation): Adjust the comment stating
that only --rfc-3339 output can be parsed by date(1).
* NEWS: Mention the change in behavior.
Reported at http://bugs.debian.org/799479
* src/factor.c: Move LIKELY() definition to...
* src/system.h: ...here.
(is_nul): Reimplement with a version that doesn't
require a sentinel after the buffer, and which calls
down to (the system optimized) memcmp.
Performance analyzed at http://rusty.ozlabs.org/?p=560
* src/dd.c (alloc_obuf): Simplify the is_nul() call by
not needing to write the sentinel.
* src/copy.c (sparse_copy): Likewise.
(copy_reg): Simplify the buffer allocation by avoiding
consideration of the sentinel in the buffer size calculation.
These strings are often file names or other user specified
parameters, which can give confusing errors in
the presence of unexpected characters for example.
* cfg.mk (sc_error_quotes): A new syntax check rule.
* src/*.c: Wrap error() string arguments with quote().
* tests/: Adjust accordingly.
* NEWS: Mention the improvement.
This is especially significant when using --check
with files generated on a windows system, where the \r
characters produce corrupted and confusing error messages.
This also ensures status messages are output on a single line.
* src/md5sum.c: Use quote() for printed file names.
* tests/misc/md5sum.pl: Adjust accordingly.
* NEWS: Mention the change in behavior.
Fixes http://bugs.gnu.org/21757
* src/ls.c (main): Account for the first column not including
a separator when calculating max_idx.
* tests/ls/w-option.sh: Add a test case.
* NEWS: Mention the bug fix.
* src/ls.c (print_with_separator): Renamed from print_with_commas,
and parameterized to accept the separator to print.
Also fix an edge case where '\n' not printed when
the POS variable overflows SIZE_MAX.
(print_current_files): Degenerate -x and -C to using the
cheaper print_with_separator() in the -w0 case.
* doc/coreutils.texi (ls invocation): Document the new feature.
* tests/ls/w-option.sh: A new test.
* tests/local.mk: Reference the new test.
* NEWS: Mention the improvement.
Fixes http://bugs.gnu.org/21325
* src/ls.c (know_term_type): Corresponding to commit v8.24-48-gc249a5a,
use fnmatch to inspect the dircolors database. Noticed due to
failing tests/ls/color-{dtype-dir,term}.sh tests.
The gnulib provided sc_tight_scope target was ineffective,
as it was checking against an invalid blank regular expression,
and thus ignoring any extern function issues. This is now
fixed up in gnulib, and so we need to fix our scoping issues
before the next gnulib update.
* cfg.mk: Setup and document the tight_scope config variables
appropriately.
* src/selinux.h: Since declared in *_SOURCES, use the two line
form for the extern function declarations.
* src/set-fields.h: Add the extern declarations, and since declared
in noinst_HEADERS use the single line form.
It was a little confusing as to whether the SQUFOF algorithm was
enabled, and in fact there were no options available to enable it.
Therefore clarify the 3 configurable behaviors for the code to
3 defines at the top of the program, and only include the SQUFOF
code if enabled at compile time.
$ size src/factor-before
text data bss
93997 1412 2504
$ size src/factor-after
text data bss
87885 1404 2504
* src/factor.c: Only include the SQUFOF factor code
when enabled via the USE_SQUFOF define.
* doc/coreutils.texi (factor invocation): Update note about
factor limits, as we can factor 128 bit numbers without GMP.
* tests/local.mk (all_tests): Remove the tests 'tests/id/setgid.sh' and
'tests/mkdir/smack-root.sh' because they are mentioned in the
'all_root_tests' list; these tests are skipped anyway during a non-root
run because flagged with 'require_root_'.
sha512sum can be faster than sha256sum.
E.g., ‘dd if=/dev/zero bs=1024k count=1024 | time sha256sum’
reports 8.16 user CPU seconds on my host, whereas sha512sum
consumes 5.45 seconds (Fedora x86-64 on an AMD Phenom II X4 910e).
Although sha512sum is still considerably slower on x86, a good
chunk of uses are on 64-bit hosts and anyway there’s little point
to scaring people away from sha512sum nowadays.
* doc/coreutils.texi (sha2 utilities): Remove obsolete comment.
* src/uniq.c (main): Make the type of "nfiles" unsigned,
to avoid a brand new warning from a gcc I built from today's
sources (gcc version 6.0.0 20151015 (experimental) (GCC)):
src/uniq.c:523:14: error: assuming signed overflow does not occur \
when simplifying conditional to constant [-Werror=strict-overflow]
if (nfiles == 2)
^
* src/dircolors.c (dc_parse_stream): Support globbing of
TERM entries, to allow entries like "TERM *256color*" for example.
* src/dircolors.hin: Reduce the internal list with globbing.
* tests/misc/dircolors.pl: New test cases.
* NEWS: Mention the improvement.
* tests/rm/r-root.sh: Skip the test if there are gdb warnings
that will impact further stderr checks. For example some
buggy gdb versions may report "Got object file from memory
but can't read symbols: File truncated". Also fix an incorrect
stderr check from the previous change.
Reported by Bernhard Voelker.
* tests/rm/r-root.sh: Use gdb rather than timeout(1) as the
last resort protection against unlinkat() calls. The timeout
of 2s was susceptible to false positives under load, and
gdb is stronger protection in any case. We remove the
"expensive" tag on this test also since it should be robust.
Reported by Jim Meyering.
* tests/tail-2/follow-stdin.sh: Use the standard tail
testing framework to avoid the race seen under very high load,
and also test the non inotify case.
Reported by Jim Meyering
To reproduce:
setfacl -dm group::rx .
setfacl -dm other::rx .
make check
* init.cfg (require_no_default_acl_): A new function to skip
when default ACLs are detected, or if the getfacl utility is
not available then skip if any non LSM ACLs detected.
* tests/cp/existing-perm-race.sh: Call require_no_default_acl_.
* tests/mkdir/parents.sh: Likewise.
* tests/mkdir/perm.sh: Likewise.
* src/tail.c (fremote): No longer prompt the user to email
with the unrecognized file system constant, since we have
process in place to sync periodically with the latest Linux
constants, and the fall back polling mode is still fully functional.
* tests/cp/cp-a-selinux.sh: Ensure we skip the portion of the test
depending on restorecon to be effective. I.E. also skip when restorecon
warns, as it doesn't exit with error status when matchpathcon fails to
find a match for a file. This is the case in /tmp on Fedora for
example, in which case the new destination that cp creates will have the
default security context of the process, rather than the explicit
context we set on the source file.
Details at: http://bugzilla.redhat.com/1247641
* tests/misc/sort-compress-hang.sh: Use --foreground with the
timeout(1) command (noting the caveats), to run the sort command
in the foreground program group, and thus be responsive to Ctrl-C.
This very_expensive_ test takes over a minute on a i3-2310M,
with RAM backed /tmp.
* tests/dd/no-allocate.sh: Account for timeout(1) when
determining the required mem, as timeout has additional shared libs.
This avoids the need for the hardcoded 4M addition to the limit.
* tests/misc/head-c.sh: Increase the base limit, to account for
the fact that head(1) will allocate some additional mem in this case.
* tests/misc/cut-huge-range.sh: Remove mention of specific limits.
* tests/misc/printf-surprise.sh: Likewise.
Reported by Dmitry V. Levin
* src/tee.c (tee_files): Last arg is now char ** instead of char
const **, as that is a bit simpler. All callers changed. Modify
files[-1], not files[nfiles], as that is a bit faster and simpler.
Latter problem pointed out by Rainer Deyke in:
http://bugs.gnu.org/21611
This pacifies GCC 5 in a better way, without disabling diagnostics.
* src/df.c (main): Tell compiler that optind is positive.
* src/shred.c (known): New function.
(dopass): Go back to off_t for file sizes.
Avoid integer overflow if we run off the end of the file.
Tell compiler that a write cannot write more bytes than requested.
Avoid the intermittent loss of "... has become inaccessible" messages.
That would cause tests/tail-2/assert.sh to fail sometimes,
mainly on uniprocessor systems.
* src/tail.c (tail_forever_inotify): Also monitor IN_DELETE
events on the directory, to avoid a dentry unlink()..open() race,
where the open() on the deleted file was seen to succeed after an,
unlink() and a subsequent IN_ATTRIB, was sent to tail. Note an
IN_ATTRIB is sent on the monitored file to indicate the change in
number of links, and we can't just use a decrease in the number of
links to determine the file being unlinked, due to the possibility
of the file having multiple links.
Reported by Assaf Gordon and Ludovic Courtès.
Fixes http://bugs.gnu.org/21460
* src/shred.c (dopass): With -O2, GCC 5.1 gives "assuming signed
overflow does not occur when simplifying conditional to constant"
warnings, in regard to the signed (off_t) variables. Therefore
use unsigned (uintmax_t) instead, and a separate boolean to cater
for the special meaning of the negative part of the integer range.
Noticed at http://hydra.nixos.org/build/24983447
Includes a change to xalloc.h to avoid -Wstrict-overflow warnings
with GCC 5.1 on 32 bit with optimization enabled. A subsequent
commit will fix similar issues in shred.
* src/dircolors.hin: Add "xterm-termite" as this VTE based terminal
emulator is quite different from xterm, despite the name.
For example "Termite supports italic text and it won't work if TERM
is set to xterm. Even the backspace key won't work properly anymore
for applications relying on terminfo".
Reported also by Lukas Sabota and Sven-Hendrik Haase.
In the presence of bind mounts of a device, the 4th "mount root" field
from /proc/self/mountinfo is now considered, so as to prefer mount
points closer to the root of the device. Note on older systems with
an /etc/mtab file, the source device was listed as the originating
directory, and so this was not an issue.
Details at http://pad.lv/1432871
* src/df.c (filter_mount_list): When deduplicating mount entries,
only prefer sources nearer or at the root of the device, when the
target is nearer the root of the device.
* NEWS: Mention the change in behavior.
This includes a change to propagate the 4th "mount root"
field from /proc/self/mountinfo from the mountlist module,
which is needed in a subsequent commit in df.
* gl/lib/regcomp.c.diff: Regenerate against latest gnulib.
* gl/lib/regex_internal.c.diff: Likewise.
* gl/lib/regex_internal.h.diff: Likewise.
* cfg.mk: Exclude diffs from trailing whitespace check,
which is generally correct, and now needed.
When configured with either 'symlinks' or 'shebangs' as value for
the --enable-single-binary option, tests based on `ulimit -v` are
skipped. The reason is that the multicall 'coreutils' binary requires
much more memory due to shared libraries being loaded, and the size of
the 'date' binary (~290KiB) compared to the multicall binary (~5MiB),
of course. Finally, in the case of 'shebangs', the starting shell
requires more memory, too
Instead of using hard-coded values for the memory limit, use an
adaptive approach: first determine the amount of memory for a similar,
yet more trivial invocation of the command, and then do the real test
run using that limit (plus some buffer in some cases).
* init.cfg (require_ulimit_v_): Remove function.
(get_min_ulimit_v_): Add function to determine the minimum memory limit
required for a given command in an adaptive way.
* cfg.mk (sc_prohibit_test_ulimit_without_require_): Change the name
of the above function in the syntax-check rule.
* tests/cp/link-heap.sh: Use the above function to determine the
minimum memory required to run a command simpler than in the real test
run. Use that limit plus a buffer there. While at it, change to list
of commands in the subshell to fail also if the beginning `ulimit -v`
fails.
* tests/dd/no-allocate.sh: Likewise.
* tests/misc/csplit-heap.sh: Likewise.
* tests/misc/cut-huge-range.sh: Likewise.
* tests/misc/head-c.sh: Likewise.
* tests/misc/printf-surprise.sh: Likewise.
* tests/split/line-bytes.sh: Likewise.
* tests/rm/many-dir-entries-vs-OOM.sh: Likewise - doing it separately
for each program under test.
* src/runcon.c (main): As per the compile time warning from
libselinux-2.4-3, lookup the class with string_to_security_class(),
rather than using defines from flask.h.
* src/sort.c (main): With --debug, warn upon setlocale() failure,
which can happen due to incorrectly specified environment variables,
or due to memory exhaustion (simulated with ulimit -v), etc.
* tests/misc/sort-debug-warn.sh: Add a test case.
See also http://savannah.gnu.org/bugs/11004
du calls stat for each mount point at startup. This would block or
even make du fail if stat for an unrelated mount point hangs.
The result is not needed in the normal case anyway and therefore
should be avoided. Issue introduced in commit v8.19-2-gcf7e1b5.
* src/du.c (fill_mount_table): Move function up as it's not used ...
(mount_point_in_fts_cycle): ... here, i.e., the DI_MNT set is
initialized and filled only iff FTS has detected a directory cycle.
(main): Remove the initialization and filling of the DI_MNT set here,
and free the DI_MNT set only if it was used.
The -NUMBER option was removed from ‘shred’ in 1999, but the
manual wasn’t updated to match. Problem reported by Nick Rose in:
http://bugs.gnu.org/21502
* doc/coreutils.texi (shred invocation):
Remove documentation for -NUMBER option.
numfmt --field=LIST can accept the same options as cut.
* bootstrap.conf: remove xlist, linked-list
* src/local.mk: link numfmt with set-fields
* src/numfmt.c: use set-fields.c instead of custom field parsing code.
(include_field): adapt to new code.
* tests/misc/numfmt.pl: add new tests, adapt current tests to new
error message wording from set-fields.c
Extract the functionality of parsing --field=LIST into a separate
module, to be used by other programs.
* src/cut.c: move field parsing code from here ...
* src/set-fields.{c,h}: ... to here.
(set_fields): generalize by supporting multiple parsing/reporting
options.
(struct range_pair): rename to field_range_pair.
* src/local.mk: link cut with set-field.
* po/POTFILES.in: add set-field.c
* tests/misc/cut.pl: update wording of error messages
* src/sort.c (main): Ensure we don't free() and invalid
pointer when reading implicit stdin. Also avoid
"definitely lost" valgrind warnings in the --files0-from case.
Since commit v8.23-19-g8defcee, main() will return,
rather than call exit(), this inducing "definitely lost"
warnings in valgrind's leak checker. That precludes using
the following to flag memory leaks:
valgrind --leak-check=full --error-exitcode=1 \
--errors-for-leak-kinds=definite
* src/pr.c (main): In dev builds, explicitly free memory allocated.
* src/sort.c (main): Likewise.
* src/tail.c (main): Likewise.
* src/tsort.c (tsort): Likewise.
* src/ls.c (getenv_quoting_style, decode_switches, parse_ls_color):
Use quote() rather than quotearg(), as the latter defaults to
outputting the input unquoted.
* src/ptx.c (main): Likewise.
* src/base64.c (main): Support decimal numbers with leading zeros,
by disabling the auto detection of octal and hex. It's not
envisaged that base conversion is needed for --wrap parameters,
and in the edge case it is, $((0x0)) shell constructs can be used.
* tests/misc/base64.pl: Adjust accordingly.
* NEWS: Mention the change in behavior.
* src/base64.c (main): Use the higher level xnumtoumax()
rather than xstrtoumax(), which is simpler and improves
validation of input. Also pass the _empty_ rather than NULL
string as the suffixes parameter so that invalid trailing
characters are not allowed. For example -w08 is now
flagged as an error, rather than being interpreted as 0.
A subsequent commit will further improve verification
of numbers with leading zeros by dropping backwards compatibility
wrt auto parsing oct and hex numbers.
* tests/misc/base64.pl: Add tests for invalid wrap values.
Suggested in https://bugzilla.redhat.com/1250113
* AUTHORS: Add base32.
* THANKS.in: Add suggester.
* README: Reference the new program.
* NEWS: Mention the new program.
* src/.gitignore: Ignore the new binary.
* bootstrap.conf: Reference the gnulib base32 module.
* build-aux/gen-lists-of-programs.sh: Add base32.
* man/base32.x: A new template.
* man/.gitignore: Ignore the new man page.
* man/local.mk: Reference the new man page.
* doc/coreutils.texi (base32 invocation): Document the new command.
* src/local.mk: Adjust to build base32 based on base64.c.
* src/base64.c: Parameterize to use the correct headers,
functions and buffer sizes, depending on which binary
is being built.
* tests/misc/base64.pl: Adjust to test both base32 and base64.
* tests/misc/tty-eof.pl: Add base32 as a program that
accepts input on stdin without any options specified.
* scripts/git-hooks/commit-msg: Add base32 to the template.
* src/shred.c (usage): For -u, separate the decscription
of the short and long option, to clarify that the short option
takes no parameter.
* src/split.c (usage): Likewise for -d.
* src/tee.c (usage): Likewise for -p.
* src/uniq.c (usage): Likewise for -D.
Suggested by Stephane Chazelas
Since glibc-2.22, specifically commit [0], the opendir() implementation
implicitly makes an additional stat call thus leading to a FP.
Seen on openSUSE:Tumbleweed since snapshot 20150821.
[0]
https://sourceware.org/git/?p=glibc.git;a=commitdiff;h=46f894d8c60a
* tests/ls/stat-free-color.sh: Change the test to verify that ls(1)
needs the same number of stat-like calls for a single, empty directory
argument as for one with a few directory entries (sub-directory,
regular file, symlink, etc.).
* doc/coreutils.texi (split invocation): Clarify that -d takes no param.
(uniq invocation): Likewise for -D.
(shred invocation): Likewise for -u.
(tee invocation): Likewise for -p.
* cfg.mk (sc_man_check_x_copyright): Add rule to ensure that non-trivial
.x files in the 'man/' subdirectory, i.e., files exceeding a line count
of 20 or a byte count of 1000, contain a proper Copyright notice.
* bootstrap.conf (gnulib_modules): Add time_rz,
since the main source code now uses timezone_t.
* src/date.c (batch_convert, main, show_date):
* src/ls.c (align_nstrftime, long_time_expected_width)
(print_long_format):
* src/stat.c (human_time):
Use timezone_t rather than boolean to specify which time zone
is wanted.
* src/ls.c (localtz): New static var.
(main): Initialize it.
* src/uname.c (usage): State that the non POSIX -i and -p options
are non-portable.
* doc/coreutils.texi (uname invocation): Mention the discrepancies
even across GNU/Linux distros, and that the results should
be used as informational only, rather than impacting any
logic decisions.
Fixes http://bugs.gnu.org/13001
* src/realpath.c (usage): Mention 'directory' in the --help
output, so that ENOTDIR errors may be more easily investigated,
by inspecting the man page.
Reported at http://pad.lv/1474519
This was detected in about 25% of runs with gcc -fsanitize=address
ERROR: AddressSanitizer: global-buffer-overflow on address ...
READ of size 4 at 0x000000416628 thread T0
#0 0x40479f in genpattern src/shred.c:782
#1 0x4050d9 in do_wipefd src/shred.c:921
#2 0x406203 in wipefile src/shred.c:1175
#3 0x406b84 in main src/shred.c:1316
#4 0x7f3454a1ef9f in __libc_start_main (/lib64/libc.so.6+0x1ff9f)
#5 0x4025d8 (/tmp/coreutils-8.23/src/shred+0x4025d8)
0x000000416628 is located 56 bytes to the left of
global variable '*.LC49' from 'src/shred.c' (0x416660) of size 17
0x000000416628 is located 12 bytes to the right of
global variable 'patterns' from 'src/shred.c' (0x416540) of size 220
SUMMARY: AddressSanitizer: global-buffer-overflow src/shred.c:782
* src/shred.c (gen_patterns): Restrict pattern selection
to the K available, which regressed due to v5.92-1462-g65533e1.
* tests/misc/shred-passes.sh: Add a deterministic test case.
* NEWS: Mention the bug fix.
Fixes http://bugs.gnu.org/20998
Some files were not shown because too many files have changed in this diff
Show More
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
