We plan to release coreutils-9.8 early next week
so any testing you can do on various different systems
between now and then would be most welcome.
This release comes about 23 weeks after the 9.7 release.
--------------------------------------
You can download the coreutils snapshot in xz format (6.0 MiB) from:
https://pixelbeat.org/cu/coreutils-ss.tar.xz
And verify with gpg or md5sum with:
https://pixelbeat.org/cu/coreutils-ss.tar.xz.sig
MD5 (coreutils-ss.tar.xz) = c3ec7d90dd9aa715f41730c8777b3fc5
--------------------------------------
To test, follow this standard procedure:
tar -xf coreutils-ss.tar.xz
cd coreutils-9.7.321-dc30e/
./configure && make check VERBOSE=yes
Failures are shown, and details are in tests/test-suite.log
Please report/attach any issues to [email protected]
--------------------------------------
This snapshot was bootstrapped with:
Autoconf 2.72.70-9ff9
Automake 1.17
Gnulib v1.0-2279-gb9f485800d
Bison 3.8.2
Changes in coreutils since v9.7 are summarized at
https://git.sv.gnu.org/cgit/coreutils.git/tree/NEWS
and all commits grouped by author are as follows:
Bernhard Voelker (1):
maint: remove now-unused include of 'safe-read.h'
Bruno Haible (7):
build: fix compilation error on OpenBSD and Solaris
date: force a Gregorian calendar for options --iso-8601 and --rfc-3339
date: update documentation regarding non-Gregorian calendars
date: fix calendar-related test failures on macOS and OpenBSD
maint: date: refactor to be entirely bottom-up
basenc: Don't trigger undefined behaviour in mini-gmp
build: Update after gnulib changed
Collin Funk (56):
build: update gnulib submodule to latest
doc: suggest GNULIB_REFDIR instead of GNULIB_SRCDIR
build: support cross-compiling directly from git checkout
build: fix 'make distcheck'
build: generate src/crctab.c when building from git
maint: use _GL_CMP instead of handwriting three-valued comparisons
build: add src/termios.c to the tarball
stdbuf: support AIX
maint: prefer UINTMAX_MAX to TYPE_MAXIMUM (uintmax_t)
doc: mention that POSIX.1-2024 added dd iflag=fullblock
doc: list signals added by POSIX.1-2024
doc: use ``...'' instead of "..."
doc: remove outdated TODO/FIXME items
maint: add files generated by 'make ps' to .gitignore
maint: prefer endian.h macros to WORDS_BIGENDIAN
maint: prefer endian.h macros to WORDS_BIGENDIAN, part 2
doc: fix a dead link
build: update gnulib submodule to latest
doc: NEWS: mention that 'date' now supports non-Gregorian calendars
tests: date: add tests for the Thai solar calendar
tests: date: add tests for the Solar Hijri calendar
tests: date: add tests for the Ethiopian calendar
maint: avoid syntax-check failure from previous commit
tests: avoid a false failure on Cygwin
doc: update git version documented in HACKING
build: don't use '$<' in non-suffix rules
build: fix previous commit
build: update gnulib submodule to latest
readlink: emit errors when POSIXLY_CORRECT is set
readlink: improve documentation for POSIXLY_CORRECT in --help
maint: use consistent references to standard files in messages
maint: prefer attribute.h in .c files
realpath: support the -E option required by POSIX
maint: avoid syntax-check failure from previous commit
logname: list David MacKenzie as the author
maint: prefer https to http
fold: add the --characters option
maint: prefer STRUCT_UTMP to struct gl_utmp
fold: use fread instead of getline
fold: don't truncate multibyte characters at the end of the buffer
fold: fix handling of invalid multi-byte characters
maint: avoid syntax-check failure from previous commit
maint: prefer issymlink to readlink with a small buffer
fold: check that characters are not non-breaking spaces when -s is used
maint: avoid syntax-check failure from previous commit
cksum: add support for SHA-3
tests: fold: check if multi-byte spaces are treated as blank
maint: prefer c32isspace to iswspace
cksum: don't leak memory using -a sha3 with OpenSSL
maint: cleanup libraries unnecessarily added to fold
maint: document some functions used by expand and unexpand
basenc: fix an uninitialized index when decoding an empty file
fold: fix out of bounds write with zero width characters
doc: NEWS: mention fold can operate on very long lines
doc: NEWS: correct the previous commit
maint: remove unnecessary uintmaxtostr usage in printf
H. Peter Anvin (4):
stty: arbitrary or non-a priori known speed_t support
doc: NEWS: improve recent stty details
stty: support baud rate macro extraction on more compilers
maint: make it clear when to add speeds to the speedgen list
Harry Fellowes (1):
test: removed dead code for unrecognised binary operators
Paul Eggert (173):
ls: omit a static var
maint: fix Automake warning
maint: use MAYBE_UNUSED more consistently
maint: modernize GCC warning configuration
maint: minor followon touchup of GCC warnings
build: update gnulib submodule to latest
maint: update NEWS from Gnulib update
factor: omit incorrect affirms in mulredc2
maint: sync longlong.h from gmp
factor: be more cautious about C compiler versions
maint: pacify GCC when selinux is missing
factor: check overflow with ckd_add etc
maint: --help now outputs bug reporting address
factor: refactor to allow non-uintmax_t ints
factor: fix bug with 128-bit uintmax_t
maint: omit -Wmissing-variable-declarations from tests
stdbuf: port better to Mac OS X 10.4.11
stdbuf: simplify apply_mode + getenv
stdbuf: don’t allocate more than SIZE_MAX / 2
build: update gnulib submodule to latest
maint: assume O_DIRECTORY
cksum: port to Apple LLVM version 10.0.0
build: update gnulib submodule to latest
date: port test to NetBSD 10
tty: better fix for Bug#26371
od: fix theoretical size_t malloc overflow
od: fix another off-by-one issue with --strings
maint: assume long long int
od: don’t assume no holes in wide unsigned
od: prefer idx_t to size_t
od: fix some unlikely integer overflows
od: fix '+N.' bug
od: prefer intmax_t to uintmax_t
od: initialize type-size tables statically
od: support uintmax_t too
od: replace lookup tables with simple arithmetic
od: omit some duplicate code
od: minor lcm tuning
od: simpler static initialization
od: simplify away one loop copy
od: check sign bit more often
od: speed up -S
od: fix integer overflow with large pseudos
od: be more consistent re sizeof
od: pacify gcc -Wduplicated-cond
maint: fix version in NEWS
od: refactor parse_old_offset
od: more minor fixes for offsets
od: add test cases for recent fix
od: port to Apple clang 14
factor: don’t used uninitialized uu[0]
factor: squfof cleanup
factor: add platform sanity check
factor: remove wide_int
factor: generalize BIG_POWER_OF_10
factor: fix test case name
factor: unsigned long → mp_bitcnt_t
factor: use same word size as GMP
factor: remove wide_uuint
factor: FALLTHROUGH instead of break
factor: prefer uuint to two words in a couple of places
factor: simplify longlong.h setup
factor: paren cleanup
factor: fix comment
factor: check unsigned char counts
factor: factor insertion simplifications
factor: add comments
factor: let builder specify MR_REPS
factor: prefer non-macros
factor: mp insert multiplicity too
factor: omit q from mp_factor_using_division
factor: primes_diff idx type consistency
factor: switch from mp to single when doable
factor: refactor to for later performance speedup
factor: use primes_diff more consistently
factor: omit unnecessary divisions by small primes
factor: use single vector in struct mp_factors
factor: use a more functional style
factor: use function for binv
factor: speed up multiprecision Pollard’s rho
factor: port back to mini-gmp
factor: move prime_p local in
factor: don’t give up before last prime in table
factor: use mpz_probab_prime_p
build: update gnulib submodule to latest
factor: always use Baillie-PSW
factor: add comments to factor.c
factor: don’t prove primality
factor: decrement Miller-Rabin tests from 25 to 24
factor: use 1-word code only when tested
factor: put FACTORS first
factor: improve millerrabin2 API
factor: speed up ge2, gt2
factor: simplify add_ssaaaa default
factor: define SQUARE_OF_FIRST_OMITTED_PRIME
factor: no ferror+fclose
factor: prime comments
factor: simplify by assuming USE_BAILLIE_PSW
factor: simplify by assuming !PROVE_PRIMALITY
factor: DELIM and DIVBLOCK no longer macros
factor: tune submod2
factor: prefer exact division
doc: update ‘factor’ benchmarks
factor: faster gcd_odd since 2nd is odd
factor: streamline gcd2_odd
factor: speed up umul_ppmm when !USE_LONGLONG_H
factor: redo ge2 in terms of lt2
factor: Pollard-rho a is now mp_limb_t
factor: avoid an mpz init+clear
factor: speed up Pollard-rho loop counters
maint: port speedgen to FreeBSD 15
factor: prefer false for boolean 0
factor: simplify primes table
factor: speed up converting strings to uuint
maint: fix apostrophe in old NEWS
factor: fix mp_factor_using_pollard_rho aliasing
factor: use 64-bit internal counters
factor: remove SINGLE_WORKS
tests: fix integer overflow in sort-float
tests: fix fraction comparison in sort-float
test: simplify for clarity
doc: mention 2025 speedup on large composites
build: update gnulib submodule to latest
head: omit unnecessary lseek
head: simplify head_bytes
wc: don’t assume opening a file puts you at start
maint: don’t lseek memory objects
tail: simplify -c+ on regular files
tail: fix SEEK_END typo
tail: check OFF_T_MAX vs COPY_A_BUFFER
tail: allow >=2**64 in traditional form
tail: speed up -c N for huge N
tail: refactor to skip stat call on failure
tail: refactor SEEK_END and lines
tail: optimize tail -n +2**63
tail: prettyname cleanup
tail: xlseek switch → table
tail: don’t output more lines than requested
tail: prefer intmax_t to uintmax_t
tail: prefer signed types to size_t, blksize_t
maint: prefer 'read' to 'safe_read'
tail: record file offset more carefully
tail: prefer < 0 to == -1
tail: fix unlikely races with >=2 --pids
tail: prefer readlink to lstat+S_ISLNK
tail: omit redundant assignment
tail: track errno more accurately
tail: fix race between read and fstat
tail: refactor ‘failable’
maint: prefer same-inode.h
cp: --sparse=always was missing some holes
cp: omit some needless lseek calls
tsort: add do-nothing -w option
realpath: improve doc and NEWS
cp: always punch holes that we make
cp: go back to copy_file_range optimization
cp: improve hole handling on squashfs
df: pacify static analysis
build: update gnulib submodule to latest
factor: port to strict C
cp: improve umask caching
cp: refactor src/copy.c
cp: refactor out data copying
cp: port better to old limited hosts
cp: prefer signed type for file byte count
cp: copy_file_data now supports ibytes
cp: refactor copying to return bytes copied
cp: don’t allocate a separate zero buffer
cp: prefer signed types in copy-file-data.c
cp: create_hole now returns off_t
build: update gnulib submodule to latest
maint: STREQ → streq
maint: prefer memeq to memcmp
Pádraig Brady (79):
maint: post-release administrivia
maint: map Jim's new email address
tests: fix false failure with constrained setfacl
tests: add a test case for negative relative date
build: update to latest gnulib
tests: avoid false failures in user namespaces
maint: avoid sc_GPL_version failure
doc: NEWS: fix introduced version of a bug
maint: avoid sc_gitignore_missing failure
maint: avoid sc_preprocessor_indentation failure
sort: fix buffer under-read (CWE-127)
tests: avoid hung tests on older macOS
maint: avoid a sc_system_h_headers failure
tests: sort: expand multiple output testing
doc: make command synopsis more consistent
doc: cp --help: mention --reflink=auto is now the default
tests: stty: adjust tests for arbitary speeds
tty: translate new error
doc: stty: adjust description of supported speeds
stty: stricter floating point parsing
od: fix various off-by-one issues with --strings with -N
od: output standard diagnostics for invalid -w arguments
doc: cksum: --check: improve supported formats description
maint: avoid sc_prohibit-quotearg failure
od: reinstate half float validation check
build: fix VPATH builds with --enable-single-binary
tests: avoid false failure due to varied ERANGE errors
doc: remove extraneous double quotes within @samp{}
tests: cat: add test for O_APPEND bug fix
build: fix non parallel VPATH builds
build: avoid make distcheck error
tests: sort: add a test for equivalent floats
install: with -d, use more accurate diagnostics
maint: avoid :> pattern in tests
doc: use consistent references to standard files
du: improve diagnostics for --time-style
doc: mention copy_file_range data corruption avoidance
basenc: fix stripping of '=' chars in some encodings
basenc: add base58 support
doc: rearrange NEWS to more standard ordering
doc: --base58: add example usage to info
maint: use short form bug URLs
nproc: honor cgroup v2 CPU quotas
tests: nproc: add a test for cgroup quotas
tests: cp: ensure copy offload is not disabled for sparse files
tests: nproc: fix false failure on some systems
tests: fold: add a memory constraint test
tests: fold: consolidate all fold tests in tests/fold
build: fold: fix build failure with C99
tests: parameterize IO_BUFSIZE
tests: fold: copy i18n patch tests
tests: fold: add tests for multi-byte width
b2sum: --length: fix upper bound check
ls: fix alignment with locale formatted --size
seq: be more accurate with large integer start values
cksum: prefer -a sha2 -l ###, to -a sha###
doc: NEWS: expand on the fold multi-byte enhancements
tests: cksum: check more length variants
doc: update the md5/sha1 "weak hash" advisory
date: support overriding named formats
tests: ensure option aliases are supported
basenc: ensure partial padding with newlines induces an error
nohup: avoid FORTIFY runtime failure on Bionic libc
maint: basenc: refactor all encodings to use finalize
build: update gnulib submodule to latest
cksum,wc: support disabling hardware acceleration at runtime
maint: randperm: remove stale dependency
maint: avoid syntax-check failures from recent commits
doc: NEWS: fold: clarify mem exhaustion trigger
tests: fold/fold-zero-width.sh: avoid false failure with ENOSPC
maint: basenc: refactor overloaded use of ctx->i member
fold: exit promptly upon write errors
fold: fix write error checks with invalid multi-byte input
doc: add hardware acceleration configuration docs
doc: remove older ChangeLog items
tests: fix issues with fold-characters.sh test
cksum,wc: inspect GLIBC_TUNABLES on all architectures
tests: write-errors.sh: avoid portability issue with dash
build: update gnulib for stringeq adjustment
thank you!
Pádraig
