#4010 closed task (fixed)
coreutils-8.26
Reported by: | Douglas R. Reno | Owned by: | DJ Lucas |
---|---|---|---|
Priority: | normal | Milestone: | 8.0 |
Component: | Book | Version: | SVN |
Severity: | normal | Keywords: | |
Cc: |
Description
New minor version
* NEW PROGRAM: b2sum *
NEWS * Noteworthy changes in release 8.26 (2016-11-30) [stable] ** Bug fixes cp, mv, and install no longer run into undefined behavior when handling ACLs on Cygwin and Solaris platforms. [bug introduced in coreutils-8.24] cp --parents --no-preserve=mode, no longer copies permissions from source directories, instead using default permissions for created directories. [bug introduced in coreutils-5.93] chcon, chgrp, chmod, chown, du, and rm, or specifically utilities using the FTS interface, now diagnose failures returned by readdir(). [this bug was inherent in the use of fts: thus, for rm the bug was introduced in coreutils-8.0. du, chmod, chgrp and chown started using fts in 6.0. chcon was added in coreutils-6.9.91 with fts support. ] date, du, ls, and pr no longer mishandle time zone abbreviations on System V style platforms where this information is available only in the global variable 'tzname'. [bug introduced in coreutils-8.24] factor again outputs immediately when numbers are input interactively. [bug introduced in coreutils-8.24] head no longer tries to process non-seekable input as seekable, which resulted in failures on FreeBSD 11 at least. [bug introduced in coreutils-8.24] install -DZ and mkdir -pZ now set default SELinux context correctly even if two or more directories nested in each other are created and each of them defaults to a different SELinux context. ls --time-style no longer mishandles '%%b' in formats. [bug introduced in coreutils-7.2] md5sum --check --ignore-missing no longer treats files with checksums starting with "00" as missing. This also affects sha*sum. [bug introduced with the --ignore-missing feature in coreutils-8.25] nl now resets numbering for each page section rather than just for each page. [This bug was present in "the beginning".] pr now handles specified separator strings containing tabs correctly. Previously it would have output random data from memory. [This bug was detected with ASAN and present in "the beginning".] sort -h -k now works even in locales that use blank as thousands separator. stty --help no longer outputs extraneous gettext header lines for translated languages. [bug introduced in coreutils-8.24] stty "sane" again sets "susp" to ^z on Solaris, and leaves "swtch" undefined. [This bug previously fixed only on some older Solaris systems] seq now immediately exits upon write errors. [This bug was present in "the beginning".] tac no longer crashes when there are issues reading from non-seekable inputs. [bug introduced in coreutils-8.15] tail -F now continues to process initially untailable files that are replaced by a tailable file. This was handled correctly when inotify was available, and is now handled correctly in all cases. [bug introduced in fileutils-4.0h] tail -f - 'untailable file' will now terminate when there is no more data to read from stdin. Previously it behaved as if --retry was specified. [This bug was present in "the beginning".] tail -f 'remote file' will now avoid outputting repeated data on network file systems that misreport file sizes through stale metadata. [This bug was present in "the beginning" but exacerbated in coreutils-8.24] tail -f --retry 'missing file' will now process truncations of that file. Previously truncation was ignored thus not outputting new data in the file. [bug introduced in coreutils-5.3.0] tail -f will no longer continually try to open inaccessible files, only doing so if --retry is specified. [This bug was present in "the beginning".] yes now handles short writes, rather than assuming all writes complete. [bug introduced in coreutils-8.24] ** Changes in behavior rm no longer accepts shortened variants of the --no-preserve-root option. seq no longer accepts 0 value as increment, and now also rejects NaN values for any argument. stat now outputs nanosecond information for time stamps even if they are out of localtime range. sort, tail, and uniq now support traditional usage like 'sort +2' and 'tail +10' on systems conforming to POSIX 1003.1-2008 and later. The 2008 edition of POSIX dropped the requirement that arguments like '+2' must be treated as file names. ** Improvements df now filters the system mount list more efficiently, with 20000 mount entries now being processed in about 1.1s compared to 1.7s. du, shuf, sort, and uniq no longer fail to process a specified file when their stdin is closed, which would have happened with glibc >= 2.14. install -Z now also sets the default SELinux context for created directories. ls is now fully responsive to signals until the first escape sequence is written to a terminal. ls now aligns quoted items with non quoted items, which is easier to read, and also better indicates that the quote is not part of the actual name. stat and tail now know about these file systems: "balloon-kvm-fs" KVM dynamic RAM allocation support, "cgroup2" Linux Control Groups V2 support, "daxfs" Optical media file system, "m1fs" A Plexistor file system, "prl_fs" A parallels file system, "smb2" Samba for SMB protocol V2, "wslfs" Windows Subsystem for Linux, "zsmalloc" Linux compressed swap support, stat -f --format=%T now reports the file system type, and tail -f uses polling for "prl_fs" and "smb2", and inotify for others. stat --format=%N for quoting file names now honors the same QUOTING_STYLE environment variable values as ls. ** New programs b2sum is added to support the BLAKE2 digest algorithm with a similar interface to the existing md5sum and sha1sum, etc. commands. ** New Features comm now accepts the --total option to output a summary at the end. date now accepts the --debug option, to annotate the parsed date string, display timezone information, and warn about potential misuse. date now accepts the %q format to output the quarter of the year.
Change History (4)
comment:1 by , 8 years ago
comment:2 by , 8 years ago
Owner: | changed from | to
---|---|
Status: | new → assigned |
So, SUSE updated their patch with the commit message: Sync I18N patch from Fedora, as the diff for the old i18n implementation of expand/unexpand has become unmaintainable. Not that I didn't try (as I found the same when reworking the old patch), but I'm afraid we're going to have to do the same for the time being. This means that we do not use the original {,un}expand functions in non-multibyte locals, and we'll have to run autotools - at least until somebody (not me) rewrites the multibyte version of those functions. I'll keep an eye out for this at SUSE, but for now -1 patch is being tested. The rest of it uses the separate *_multibyte() functions as before (where applicable).
comment:3 by , 8 years ago
Resolution: | → fixed |
---|---|
Status: | assigned → closed |
Resolved in r11149. Autotools were not required.
Quick review of i18n patch will require some manual finagling, but not too much. Mostly 's/error (/die (/' with the exception of sort (which will also apply mostly as is error/die|sort_die) and a couple of the tests. Upstream has not modified the patch in 10 mos. If nobody beats me to it, I'll try and get at it Saturday if SUSE doesn't have it.