| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
5.4.0 (2022-12-13)
This bumps the minor version of liblzma because new features were
added. The API and ABI are still backward compatible with liblzma
5.2.x and 5.0.x.
Since 5.3.5beta:
* All fixes from 5.2.10.
* The ARM64 filter is now stable. The xz option is now --arm64.
Decompression requires XZ Utils 5.4.0. In the future the ARM64
filter will be supported by XZ for Java, XZ Embedded (including
the version in Linux), LZMA SDK, and 7-Zip.
* Translations:
- Updated Catalan, Croatian, German, Romanian, and Turkish
translations.
- Updated German man page translations.
- Added Romanian man page translations.
Summary of new features added in the 5.3.x development releases:
* liblzma:
- Added threaded .xz decompressor lzma_stream_decoder_mt().
It can use multiple threads with .xz files that have multiple
Blocks with size information in Block Headers. The threaded
encoder in xz has always created such files.
Single-threaded encoder cannot store the size information in
Block Headers even if one used LZMA_FULL_FLUSH to create
multiple Blocks, so this threaded decoder cannot use multiple
threads with such files.
If there are multiple Streams (concatenated .xz files), one
Stream will be decompressed completely before starting the
next Stream.
- A new decoder flag LZMA_FAIL_FAST was added. It makes the
threaded decompressor report errors soon instead of first
flushing all pending data before the error location.
- New Filter IDs:
* LZMA_FILTER_ARM64 is for ARM64 binaries.
* LZMA_FILTER_LZMA1EXT is for raw LZMA1 streams that don't
necessarily use the end marker.
- Added lzma_str_to_filters(), lzma_str_from_filters(), and
lzma_str_list_filters() to convert a preset or a filter chain
string to a lzma_filter[] and vice versa. These should make
it easier to write applications that allow users to specify
custom compression options.
- Added lzma_filters_free() which can be convenient for freeing
the filter options in a filter chain (an array of lzma_filter
structures).
- lzma_file_info_decoder() to makes it a little easier to get
the Index field from .xz files. This helps in getting the
uncompressed file size but an easy-to-use random access
API is still missing which has existed in XZ for Java for
a long time.
- Added lzma_microlzma_encoder() and lzma_microlzma_decoder().
It is used by erofs-utils and may be used by others too.
The MicroLZMA format is a raw LZMA stream (without end marker)
whose first byte (always 0x00) has been replaced with
bitwise-negation of the LZMA properties (lc/lp/pb). It was
created for use in EROFS but may be used in other contexts
as well where it is important to avoid wasting bytes for
stream headers or footers. The format is also supported by
XZ Embedded (the XZ Embedded version in Linux got MicroLZMA
support in Linux 5.16).
The MicroLZMA encoder API in liblzma can compress into a
fixed-sized output buffer so that as much data is compressed
as can be fit into the buffer while still creating a valid
MicroLZMA stream. This is needed for EROFS.
- Added lzma_lzip_decoder() to decompress the .lz (lzip) file
format version 0 and the original unextended version 1 files.
Also lzma_auto_decoder() supports .lz files.
- lzma_filters_update() can now be used with the multi-threaded
encoder (lzma_stream_encoder_mt()) to change the filter chain
after LZMA_FULL_BARRIER or LZMA_FULL_FLUSH.
- In lzma_options_lzma, allow nice_len = 2 and 3 with the match
finders that require at least 3 or 4. Now it is internally
rounded up if needed.
- CLMUL-based CRC64 on x86-64 and E2K with runtime processor
detection. On 32-bit x86 it currently isn't available unless
--disable-assembler is used which can make the non-CLMUL
CRC64 slower; this might be fixed in the future.
- Building with --disable-threads --enable-small
is now thread-safe if the compiler supports
__attribute__((__constructor__)).
* xz:
- Using -T0 (--threads=0) will now use multi-threaded encoder
even on a single-core system. This is to ensure that output
from the same xz binary is identical on both single-core and
multi-core systems.
- --threads=+1 or -T+1 is now a way to put xz into
multi-threaded mode while using only one worker thread.
The + is ignored if the number is not 1.
- A default soft memory usage limit is now used for compression
when -T0 is used and no explicit limit has been specified.
This soft limit is used to restrict the number of threads
but if the limit is exceeded with even one thread then xz
will continue with one thread using the multi-threaded
encoder and this limit is ignored. If the number of threads
is specified manually then no default limit will be used;
this affects only -T0.
This change helps on systems that have very many cores and
using all of them for xz makes no sense. Previously xz -T0
could run out of memory on such systems because it attempted
to reserve memory for too many threads.
This also helps with 32-bit builds which don't have a large
amount of address space that would be required for many
threads. The default soft limit for -T0 is at most 1400 MiB
on all 32-bit platforms.
- Previously a low value in --memlimit-compress wouldn't cause
xz to switch from multi-threaded mode to single-threaded mode
if the limit cannot otherwise be met; xz failed instead. Now
xz can switch to single-threaded mode and then, if needed,
scale down the LZMA2 dictionary size too just like it already
did when it was started in single-threaded mode.
- The option --no-adjust no longer prevents xz from scaling down
the number of threads as that doesn't affect the compressed
output (only performance). Now --no-adjust only prevents
adjustments that affect compressed output, that is, with
--no-adjust xz won't switch from multi-threaded mode to
single-threaded mode and won't scale down the LZMA2
dictionary size.
- Added a new option --memlimit-mt-decompress=LIMIT. This is
used to limit the number of decompressor threads (possibly
falling back to single-threaded mode) but it will never make
xz refuse to decompress a file. This has a system-specific
default value because without any limit xz could end up
allocating memory for the whole compressed input file, the
whole uncompressed output file, multiple thread-specific
decompressor instances and so on. Basically xz could
attempt to use an insane amount of memory even with fairly
common files. The system-specific default value is currently
the same as the one used for compression with -T0.
The new option works together with the existing option
--memlimit-decompress=LIMIT. The old option sets a hard limit
that must not be exceeded (xz will refuse to decompress)
while the new option only restricts the number of threads.
If the limit set with --memlimit-mt-decompress is greater
than the limit set with --memlimit-compress, then the latter
value is used also for --memlimit-mt-decompress.
- Added new information to the output of xz --info-memory and
new fields to the output of xz --robot --info-memory.
- In --lzma2=nice=NUMBER allow 2 and 3 with all match finders
now that liblzma handles it.
- Don't mention endianness for ARM and ARM-Thumb filters in
--long-help. The filters only work for little endian
instruction encoding but modern ARM processors using
big endian data access still use little endian
instruction encoding. So the help text was misleading.
In contrast, the PowerPC filter is only for big endian
32/64-bit PowerPC code. Little endian PowerPC would need
a separate filter.
- Added decompression support for the .lz (lzip) file format
version 0 and the original unextended version 1. It is
autodetected by default. See also the option --format on
the xz man page.
- Sandboxing enabled by default:
* Capsicum (FreeBSD)
* pledge(2) (OpenBSD)
* Scripts now support the .lz format using xz.
* A few new tests were added.
* The liblzma-specific tests are now supported in CMake-based
builds too ("make test").
|
|
5.2.9 (2022-11-30)
* liblzma:
- Fixed an infinite loop in LZMA encoder initialization
if dict_size >= 2 GiB. (The encoder only supports up
to 1536 MiB.)
- Fixed two cases of invalid free() that can happen if
a tiny allocation fails in encoder re-initialization
or in lzma_filters_update(). These bugs had some
similarities with the bug fixed in 5.2.7.
- Fixed lzma_block_encoder() not allowing the use of
LZMA_SYNC_FLUSH with lzma_code() even though it was
documented to be supported. The sync-flush code in
the Block encoder was already used internally via
lzma_stream_encoder(), so this was just a missing flag
in the lzma_block_encoder() API function.
- GNU/Linux only: Don't put symbol versions into static
liblzma as it breaks things in some cases (and even if
it didn't break anything, symbol versions in static
libraries are useless anyway). The downside of the fix
is that if the configure options --with-pic or --without-pic
are used then it's not possible to build both shared and
static liblzma at the same time on GNU/Linux anymore;
with those options --disable-static or --disable-shared
must be used too.
|
|
4.0.2
Fix a memory leak when frame decompression fails
4.0.1
Add flush method to LZ4FrameFile
4.0.0
This release relegates the stream bindings to experimental status and disables them in all wheel builds. This was necessary as the test suite was causing build failures on multiple architectures. These bindings are currently in an unmaintained state.
|
|
v3.11.0
=======
* Added support for new methods on ``Path``:
- ``match``
- ``glob`` and ``rglob``
- ``relative_to``
- ``is_symlink``
|
|
Change log:
### engrampa 1.26.0
* Translations update
* update copyright to 2021
### engrampa 1.25.1
* Translations update
* ar: fix filename in file list
* dlg-package-installer: Don't build the path to packages.match
* build: Don't install packages.match when packagekit is disabled
* packages.match: add and run update-packages-match script
* build: don't add twice the file engrampa.appdata.xml to CLEANFILES
* rar 6.00: fix listing archive content with encrypted file list
* i18n: use g_dngettext instead of ngettext
* build: allow users to disable gettext support (--disable-nls)
* build: Use PACKAGE_URL variable
* build: display package name and version in configure summary
* build: set GETTEXT_PACKAGE=AC_PACKAGE_NAME
* update resource for transifex
* Update copyright to 2021
* Remove cppcheck warnings about the variable scope can be reduced
* update issue template
* update issue template
* Remove USE_MATE2_MACROS from autogen.sh (legacy)
* fr-archive: Fix memory leak
* Add Electronic Publication (EPUB) support
* Remove unused macros
* update.ui: missing on_update_file_ok_button_clicked handler
* fr-command-unarchiver: Remove trailing spaces
* Avoid memory leak in java utils
* Do not use stock icons in ui files (#392)
* Update compilation instructions
### engrampa 1.25.0
* Translations update
* build: includes all required modules by libegg in PKG_CHECK_MODULES
* Update compilation instructions after use of git submodule
* add git.mk to generate .gitignore
* dlg-update: Fix callback names for callback symbol update_cb
* use git submodule of libegg
* dlg-batch-add: Set labels GtkSizeGroup in ui file
* Use callback symbols defined in ui files
* gtk-utils: Remove _gtk_builder_get_widget
* Remove fr-stock.c and fr-stock.h
* build: silent build warnings for distcheck
* help: fix and update archive format names
* actions: expand the comment on about dialog
* Terminate the application if the resource doesn't exists
* gtk-utils: Remove _gtk_count_selected
* engrampa.appdata.xml: recursive acronyms for .iso, .rpm filetypes
* glib-utils: remove g_ptr_array_free_full
* fr-window: duplicated code
* Add ARC archive support
* glib-utils: remove n_fields function
* Add dlg-add-files.ui and dlg-add-folder.ui
* Update authors
* dlg-new: Adding many callbacks at once
* new.ui: Stock icons are deprecated
* dlg-new: Use GET_WIDGET macro
* autoconf: get zstd mime type using libmagic if enabled
|
|
|
|
|
|
on archive.org, used by MacPorts
|
|
v3.10.0
=======
* ``zipp`` is now a package.
|
|
These are currently duplicated in mk/platform/SunOS.mk but the generic
removals will be removed soon in favour of per-package removals, due to
flags getting leaked into installed files.
|
|
5.2.8 (2022-11-13)
* xz:
- If xz cannot remove an input file when it should, this
is now treated as a warning (exit status 2) instead of
an error (exit status 1). This matches GNU gzip and it
is more logical as at that point the output file has
already been successfully closed.
- Fix handling of .xz files with an unsupported check type.
Previously such printed a warning message but then xz
behaved as if an error had occurred (didn't decompress,
exit status 1). Now a warning is printed, decompression
is done anyway, and exit status is 2. This used to work
slightly before 5.0.0. In practice this bug matters only
if xz has been built with some check types disabled. As
instructed in PACKAGERS, such builds should be done in
special situations only.
- Fix "xz -dc --single-stream tests/files/good-0-empty.xz"
which failed with "Internal error (bug)". That is,
--single-stream was broken if the first .xz stream in
the input file didn't contain any uncompressed data.
- Fix displaying file sizes in the progress indicator when
working in passthru mode and there are multiple input files.
Just like "gzip -cdf", "xz -cdf" works like "cat" when the
input file isn't a supported compressed file format. In
this case the file size counters weren't reset between
files so with multiple input files the progress indicator
displayed an incorrect (too large) value.
* liblzma:
- API docs in lzma/container.h:
* Update the list of decoder flags in the decoder
function docs.
* Explain LZMA_CONCATENATED behavior with .lzma files
in lzma_auto_decoder() docs.
- OpenBSD: Use HW_NCPUONLINE to detect the number of
available hardware threads in lzma_physmem().
- Fix use of wrong macro to detect x86 SSE2 support.
__SSE2_MATH__ was used with GCC/Clang but the correct
one is __SSE2__. The first one means that SSE2 is used
for floating point math which is irrelevant here.
The affected SSE2 code isn't used on x86-64 so this affects
only 32-bit x86 builds that use -msse2 without -mfpmath=sse
(there is no runtime detection for SSE2). It improves LZMA
compression speed (not decompression).
- Fix the build with Intel C compiler 2021 (ICC, not ICX)
on Linux. It defines __GNUC__ to 10 but doesn't support
the __symver__ attribute introduced in GCC 10.
* Scripts: Ignore warnings from xz by using --quiet --no-warn.
This is needed if the input .xz files use an unsupported
check type.
* Translations:
- Updated Croatian and Turkish translations.
- One new translations wasn't included because it needed
technical fixes. It will be in upcoming 5.4.0. No new
translations will be added to the 5.2.x branch anymore.
- Renamed the French man page translation file from
fr_FR.po to fr.po and thus also its install directory
(like /usr/share/man/fr_FR -> .../fr).
- Man page translations for upcoming 5.4.0 are now handled
in the Translation Project.
* Update doc/faq.txt a little so it's less out-of-date.
|
|
|
|
Changes not documented.
|
|
|
|
|
|
|
|
devel/libjit, devel/m17n-lib, devel/quilt, devel/treecc, emulators/simulavr,
fonts/jomolhari-ttf, graphics/dvipng, graphics/freetype2, graphics/libotf,
ham/xlog, misc/color-theme, misc/m17n-db, multimedia/flvstreamer,
net/mimms, print/chktex, security/oath-toolkit, sysutils/attr,
sysutils/pidof, sysutils/renameutils, textproc/lout:
change MASTER_SITES from http://download.savannah.gnu.org to https://download.savannah.gnu.org
|
|
Change log:
lxqt-archiver-0.7.0 / 2022-11-05
===================================
* Removed the redundant function `n_fields()` from `glib-utils`.
* Silenced some compilation warnings.
|
|
|
|
|
|
0.19.0
Bug Fixes
The C backend implementation of ZstdDecompressionObj.decompress() could
have raised an assertion in cases where the function was called multiple
times on an instance. In non-debug builds, calls to this method could have
leaked memory.
Changes
PyPy 3.6 support dropped; Pypy 3.8 and 3.9 support added.
Anaconda 3.6 support dropped.
Official support for Python 3.11. This did not require meaningful code changes
and previous release(s) likely worked with 3.11 without any changes.
CFFI's build system now respects distutils's compiler.preprocessor if it
is set.
The internal logic of ZstdDecompressionObj.decompress() was refactored.
This may have fixed unconfirmed issues where unused_data was set
prematurely. The new logic will also avoid an extra call to
ZSTD_decompressStream() in some scenarios, possibly improving performance.
ZstdDecompressor.decompress() how has a read_across_frames keyword
argument. It defaults to False. True is not yet implemented and will raise an
exception if used. The new argument will default to True in a future release
and is provided now so callers can start passing read_across_frames=False
to preserve the existing functionality during a future upgrade.
ZstdDecompressor.decompress() now has an allow_extra_data keyword
argument to control whether an exception is raised if input contains extra
data. It defaults to True, preserving existing behavior of ignoring extra
data. It will likely default to False in a future release. Callers desiring
the current behavior are encouraged to explicitly pass
allow_extra_data=True so behavior won't change during a future upgrade.
|
|
Changelog from AN-2022-10-16:
UPDATING:
- Various man pages have been renamed to follow BSD conventions.
Users may need to manually remove the following old man pages on
update: streamarchive.4, makefiles.4, makerules.4, changeset.4,
sccschangeset.4, sccsfile.4, and star.4.
NEW FEATURES:
- libfind/sfind: Fix incorrect error handling of -exec ... {} +
primaries. Previously in some cases the primary returned an
error even though it should always return true and in other
cases it completely disregarded the exit code. Now the primary
always returns true and sets the exit code of libfind and sfind
to a non-zero value in case one of the commands returned with a
non-zero exit code.
Reported by Tavian Barnes. Submitted by Nico Sonack.
- libfind/sfind: Report failures to query the password/group
database for -nouser/-nogroup. Previously failure to query
these databases was treated as if the user id/group id was not
found. We now print a diagnostic message and return a non-zero
value if the getpwuid() or getgrgid() call failed.
Reported by Tavian Barnes.
- libfind/sfind: Report presence of directory loops. Previously,
such directory loops were silently ignored. Now, upon
encountering a directory loop, sfind prints a diagnostic message
to standard error. The search is continued but sfind will
terminate with a nonzero exit status. This change improves
compatibility with POSIX (cf. PASC interpretation request 1606).
All libfind consumers calling the treewalk() function (mkisofs,
sccs, and star) are affected.
Reported by Tavian Barnes. Submitted by Nico Sonack.
- star: rename tests/compress archives to avoid issues on case
insensitive file systems. These file systems do not like having
multiple files whose names only differ in letter case. Rename
the archives to completely distinct file names to avoid this
issue.
Reported by tux <rosa.elefant@icloud.com>.
- *: With the switch of Solaris from SysV man page sections to BSD
man page sections, man pages are now installed according to BSD
conventions instead of SysV conventions. This affects the man
pages makefiles(4), makerules(4), streamarchive(4),
changeset(4), sccsfile(4), sccschangeset(4), and star(4).
Reported by Friedhelm Mehnert.
[pkgsrc note: Not yet, see Issue #40]
- *: All references to Solaris man pages have been adjusted to BSD
man page sections as used in Solaris 11.4.
- libschily/getargs.3: document surprising behaviour that the
the argument to a foo= parameter may be in the following
argument. Thus, shell scripts should write foo= "$bar" instead
of foo=$bar if $bar could possibly be empty.
Submitted by Nico Sonack.
- libschily/getargs.3: improve spelling and grammar.
|
|
Changes not found
|
|
|
|
Use latest zlib 1.2.13 sources.
|
|
|
|
Plzip is a massively parallel (multi-threaded) implementation of lzip, fully
compatible with lzip 1.4 or newer. Plzip uses the compression library lzlib.
Lzip is a lossless data compressor with a user interface similar to the one
of gzip or bzip2. Lzip uses a simplified form of the 'Lempel-Ziv-Markov
chain-Algorithm' (LZMA) stream format and provides a 3 factor integrity
checking to maximize interoperability and optimize safety. Lzip can compress
about as fast as gzip (lzip -0) or compress most files more than bzip2
(lzip -9). Decompression speed is intermediate between gzip and bzip2.
Lzip is better than gzip and bzip2 from a data recovery perspective. Lzip
has been designed, written, and tested with great care to replace gzip and
bzip2 as the standard general-purpose compressed format for unix-like
systems.
|
|
|
|
Tarlz is a massively parallel (multi-threaded) combined implementation
of the tar archiver and the lzip compressor.
Originally packaged in pkgsrc-wip by Aleksej Lebedev, thanks!
|
|
|
|
Lzlib is a data compression library providing in-memory LZMA compression and
decompression functions, including integrity checking of the decompressed data.
The compressed data format used by the library is the lzip format.
Lzlib is written in C.
Originally packaged in pkgsrc-wip by Aleksej Lebedev, thanks!
|
|
1.0.6
Changed
- Improved cmake for mingw by Miloš Komarčević
1.0.5
Changed
- Updated documentation to new 121.0-B-3 standard. The new standard
mainly clarifies and explicitly defines special cases which could be
ambiguous or misleading in previous revisions.
These changes did *not* require any substantial changes to libaec.
Existing compressed data is still compatible with this version of
the library and compressed data produced by this version can be
uncompressed with previous versions.
- Modernized CMake
- Better CMake integration with HDF5 by Jan-Willem Blokland
|
|
v3.9.0
Path objects are now pickleable if they've been constructed from pickleable objects. Any restored objects will re-construct the zip file with the original arguments.
|
|
rarfile v4.0
Main goals are:
Increased zipfile-compatibility, thus also achieving smaller difference between RAR3 and RAR5 archives.
Implement RarFile.extract on top of RarFile.open instead using unrar x directly, thus making maintenance of alternative backends more manageable. Negative aspect of that is that there are features that internal extract code does not support - hard links, NTFS streams and junctions.
Breaking changes:
Directory names will have "/" appended.
RarFile.extract operates only on single entry, so when used on directory it will create directory but not extract files under it.
RarFile.extract/RarFile.extractall/RarFile.testrar will not launch special unrar command line, instead they are implemented on top of RarFile.open.
Keyword args in top-level APIs were renamed to match zipfile:
RarFile(rarfile) -> RarFile(file)
RarFile.setpassword(password) -> .setpassword(pwd)
RarFile.getinfo(fname) -> .getinfo(name)
RarFile.open(fname, mode, psw) -> .open(name, mode, pwd)
RarFile.read(fname, psw) -> .read(name, pwd)
PATH_SEP cannot be changed from "/".
New features:
RarFile.extract will return final sanitized filename for target file.
RarInfo.is_dir is now preferred spelling of isdir(). Old method kept as alias.
New RarInfo.is_file and RarInfo.is_symlink methods. Only one of ~RarInfo.is_file, ~RarInfo.is_dir or ~RarInfo.is_symlink can be True.
RarFile.printdir has file argument for output.
RarFile.__iter__ loops over RarInfo entries.
RAR3: throw NeedFirstVolume exception with current volume number, like RAR5 does.
Nanosecond timestamp support. Visible as nsdatetime instance.
Minimal CLI when run as script: python3 -m rarfile
Skip old file versions in versioned archive.
Cleanups:
Use PBKDF2 implementation from hashlib.
Improve test coverage.
|
|
5.2.7 (2022-09-30)
* liblzma:
- Made lzma_filters_copy() to never modify the destination
array if an error occurs. lzma_stream_encoder() and
lzma_stream_encoder_mt() already assumed this. Before this
change, if a tiny memory allocation in lzma_filters_copy()
failed it would lead to a crash (invalid free() or invalid
memory reads) in the cleanup paths of these two encoder
initialization functions.
- Added missing integer overflow check to lzma_index_append().
This affects xz --list and other applications that decode
the Index field from .xz files using lzma_index_decoder().
Normal decompression of .xz files doesn't call this code
and thus most applications using liblzma aren't affected
by this bug.
- Single-threaded .xz decoder (lzma_stream_decoder()): If
lzma_code() returns LZMA_MEMLIMIT_ERROR it is now possible
to use lzma_memlimit_set() to increase the limit and continue
decoding. This was supposed to work from the beginning
but there was a bug. With other decoders (.lzma or
threaded .xz decoder) this already worked correctly.
- Fixed accumulation of integrity check type statistics in
lzma_index_cat(). This bug made lzma_index_checks() return
only the type of the integrity check of the last Stream
when multiple lzma_indexes were concatenated. Most
applications don't use these APIs but in xz it made
xz --list not list all check types from concatenated .xz
files. In xz --list --verbose only the per-file "Check:"
lines were affected and in xz --robot --list only the "file"
line was affected.
- Added ABI compatibility with executables that were linked
against liblzma in RHEL/CentOS 7 or other liblzma builds
that had copied the problematic patch from RHEL/CentOS 7
(xz-5.2.2-compat-libs.patch). For the details, see the
comment at the top of src/liblzma/validate_map.sh.
WARNING: This uses __symver__ attribute with GCC >= 10.
In other cases the traditional __asm__(".symver ...")
is used. Using link-time optimization (LTO, -flto) with
GCC versions older than 10 can silently result in
broken liblzma.so.5 (incorrect symbol versions)! If you
want to use -flto with GCC, you must use GCC >= 10.
LTO with Clang seems to work even with the traditional
__asm__(".symver ...") method.
* xzgrep: Fixed compatibility with old shells that break if
comments inside command substitutions have apostrophes (').
This problem was introduced in 5.2.6.
* Build systems:
- New #define in config.h: HAVE_SYMBOL_VERSIONS_LINUX
- Windows: Fixed liblzma.dll build with Visual Studio project
files. It broke in 5.2.6 due to a change that was made to
improve CMake support.
- Windows: Building liblzma with UNICODE defined should now
work.
- CMake files are now actually included in the release tarball.
They should have been in 5.2.5 already.
- Minor CMake fixes and improvements.
* Added a new translation: Turkish
|
|
1.21.1 (2022-09-16)
* PHP 8.2 compatibility
|
|
|
|
|
|
Removed patch-a* to use the new
naming convention.
This version requires c++14 so
the patch concerning vector.assign()
missing in sunpro has been removed.
src/libdar/mask_list.hpp could not
be patched and the my_char method
does not exist anywhere in the
sources so the patch has been removed.
This version includes autotools
changes that have AM_LDFLAGS and
looks for getopt in unistd.h.
Added extra buildlink files so that
all deps are found.
Changes from Changelog:
from 2.7.6 to 2.7.7
- added support for sequential reading more of sliced backup, to
accommodate tape support used with slices (at the opposite of dar_split)
- fixing few typos in doc
- making libdar more tolerant when calls to fadvise fail
from 2.7.5 to 2.7.6
- adding -f option to dar_cp
- adding static version of dar_cp (dar_cp_static) as compilation outcome
- added FAQ for tape usage with dar
- fixing error in libdar header file installation
- fixed bug met when interrupting the creation of a block compressed
backup (always used by lzo compression and by other algorithm only
when performing multi-threaded compression)
- typo fixes in documentation
- fixed message in lax mode used to obtain from the user the archive format
when this information is corrupted in the archive.
- fixing lax mode condition that popped up without being requested
- fixing bug met when reading slice an special block device by mean of
a symlink
- adapting sanity checks to the case of a backup read from a special
device in sequential-read mode.
- fixed bug that lead dar to report CRC error while reading a backup
from a pipe with the help of an isolated catalogue
- adding -V option to dar_split (was using -v) for homogeneity with
other commands
from 2.7.4 to 2.7.5
- fixed double free error met when deciphering an archive with a
wrong password/passphrase and when multi-threading is used.
from 2.7.3 to 2.7.4
- fixed excessive context control that led libdar to report a bug
when an file-system I/O error was returned by the operating system
- fixed mini-digits auto-detection, which only worked when slice number 1
was present, even if subsequent slices could be used to detect its
value
- fixed typos and minor incoherence in documentation
- update version information to display libthreadar barrier implementation
used, info available since its release 1.4.0
from 2.7.2 to 2.7.3
- fixed bug met when restoring files in sequential-read mode and feeding
the backup to dar on its stdin: When dar had to remove a file that had
been deleted at the time of the backup since the backup of reference
was made, dar aborted the restoration reporting it could not skip
backward on a pipe.
- adding call to kill() then join() in destructor of class slave_thread
and fichier_libcurl to avoid the risk of SEGFAULT in rare conditions
under Cygwin (Windows)
- fixed several typos in bug, messages and comments
- fixed script used to build windows binary to match recent cygwin dll
- fix spelling and improved clarity in dar_split messages
from 2.7.1 to 2.7.2
- fixed bug met when a user command returns an error while dar is
saving fsa attributes of inodes (this conjunction make it an infrequent
situation)
- fixed typo in documentation
- fixed remaining bug from 2.7.1 met when compiling dar/libdar with a
compiler (clang-5.0 here) that requires a flag to activate C++14 support
while complaining (for a good reason) when this flag is passed too while
compiling C code.
- fixed self reported bug escape.cpp line 858 met when using lzo compression
or multi-threaded compression while creating a backup to stdout.
- fixed bug met when creating a backup to stdout that lead libdar to corrupt
data when trying to re-save a file uncompressed due to poor compression
result
- fixed minor bug met when setting --mincompr to zero and using
block_compressor (lzo algo or any algo in block compression): empty files
were reported as truncated due to the lack of block header (compressed
empty file being stored as an empty file)
from 2.7.0 to 2.7.1
- fixed compilation script to require support for C++14 due to new features
introduces in 2.7.0 that rely on modern C++ constructions. Updated
documentation accordingly about this updated requirement.
- fixed missing included header for compilation to suceed under MacOS X
- fixed typo in man page
- adding minor feature: storing the backup root (-R option) in the archive
to be able to restore 'in place' thanks to the new -ap option that
sets the -R path to the one stored in the archive.
- merging fixes an enhancements brought by release 2.6.15
from 2.6.x to 2.7.0
- using the truncate system call whenever possible in place of skipping
back in the archive, when file need to re-save uncompressed,
or when file has changed while it was read for backup
- improved slice name versus base name error, now substituting (rather than
just asking the user) the first by the later when else it would lead
to an execution error.
- auto-detecting min-digits at reading time in complement of the feature
listed just above
- added the possibility to specify a gnupg key by mean of its key-id in
addition to the email address it may be associated to, both for
asymmetrical encryption and archive signing.
- added -b option to dar_split to set the I/O block size
- added -r option to dar_split to limit the transfer rate
- added -c option to dar_split to limit the number of tape to read
- new feature: zstd compression algorithm added
- replaced old and experimental multi-threaded implementation by
production grade one. -G option may now receive an argument on
command line for fine tuning. libdar API has been updated
accordingly.
- added multi-threaded compression when using the new per block
compression mode, the legacy streaming compression mode is still
available (see both -G and -z option extended syntax).
- added lz4 compression algorithm.
- removed some deprecated/duplicated old options (--gzip,...)
- enhanced the delta signature calculation providing mean for the user
to adapt the signature block size in regard to the file's size to
delta-sig. "--delta sig" option received an extended syntax.
- increased timestamp precision from microsecond to nanosecond when the
operating system supports it. It is still possible to use configure
--enable-limit-time-accuracy=us at compilation time to use microsecond
or even second accuracy even if the OS can support finer precision.
- added argon2 hashing algorithm for key derivation function (kdf) which
becomes the default if libargon2 is available, else it defaults to sha1
as for 2.6.x. When argon2 is used, kdf default iteration count is
reduced to 10,000 (and stays 200,000 with sha1). This can be tuned
as usual with --kdf-param option.
- adding support for statx() under Linux which let libdar save file's
birthtime. Unfortunately unlike under BSD systems (FreeBSD, MACoS X,
and so on), the utime/utimes/timensat call do not set birthtime, so
you cannot restore birthtime on Linux's unlike under BSD systems
- AES becomes the default when gnupg is used without additional
algorithm.
- new implementation of the libcurl API use for more efficient reuse of
established sessions.
- You can now exclude/include per filesystem rather than just sticking
to the current filesystem or ignoring filesystem boundary. -M option
can now receive arguments.
- new feature: dar_manager can now change the compression algorithm of
an existing database.
- Added a benchmark of dar versus tar and rsync for different use cases.
- documentation review, update, cleanup, restructured and beautification.
from 2.6.15 to 2.6.16
- fixed bug met when restoring files in sequential-read mode and feeding
the backup to dar on its stdin: When dar had to remove a file that had
been deleted at the time of the backup since the backup of reference
was made, dar aborted the restoration reporting it could not skip
backward on a pipe.
- adding call to kill() then join() in destructor of class slave_thread
and fichier_libcurl to avoid the risk of SEGFAULT in rare conditions
under Cygwin (Windows)
- fixed bug met when removing tape marks (-at option) and due to poor
compression, dar skips back to re-save file uncompressed leading to
self reported bug (due to an error in the sanity check).
- fixing error message display by dar when -y option is used with another
command (-c/-t/-l/-d/-x/-+/-C)
from 2.6.14 to 2.6.15
- fixed error message formatting error leading message to contain garbage
in place of system error information.
- fixing bug (internal error) met while trying restoring files and dirs
without sufficient write permission on the destination directory tree
to perform the operation.
- adding minor feature to avoid restoring Unix sockets (-au option)
- fixing dar-catalogue.dtd
from 2.6.13 to 2.6.14
- script used to build dar windows binary has been fixed to have the
official default etc/darrc config file usable and used out of the
box.
- fixed bug met when removing slices of an old backup located on
a remote sftp server
- fixed bug in cache layer met when writing sliced backup to a remote
ftp or sftp repository
- enhancement to the -[ and -] options to work as expected when "DOS"
formatted text file is provided as a file listing.
from 2.6.12 to 2.6.13
- fixed compilation warning in testing routine (outside libdar and dar)
- due to change in autoconf, the --sysconfdir path (which defaults to
${prefix}/etc) was read as an empty string, leading dar to look for
darrc system file at the root of the filesystem (/darrc)
- fixed bug that should occur in extremely rare conditions (it has been
discover during 2.7.0 validation process): compression must be used,
no ciphering, no hashing, file changed at backup time or had a poor
compression ratio, was not saved at slice boundary, the previous
entry had an EA saved but no FSA or an unchanged FSA. In such
conditions are all met, dar tries to resave the file in place, but
partially or totally overwites the EAs of the previous entry leading
to archive testing to fail for these EA (though the archive could be
finished without error).
- fixed bug met when case insensitive mask is requested (-an option)
and locale of file to restore or backup is not the one the dar binary
is run with.
from 2.6.11 to 2.6.12
- fixed regression met in 2.6.11 when generating encrypted archives
from 2.6.10 to 2.6.11
- fixing bug in dar_manager libdar part, met when the two oldest entries
for a file are recorded as unchanged (differential backup).
- fixed typo in dar_manager man page
- updated lax mode to ignore encryption flag found in header and trailer
- fixed two opposite bugs in strong encryption code that annihilated
each other, by chance
- fixing bug met when merging an archive an re-compressing the data
with another algorithm that gives a less good result, this condition
lead the merging operation to fail reporting a CRC mismatch
- improving archive header code to cope with unknown flags
from 2.6.9 to 2.6.10
- update the configure script to handle some undocumented --enable-*
options that existed but were not expected to be used.
- fixed spelling in darrc comments
- fixed bug in dar_split that could occur in very rare conditions
- fixed EA support build failure due to what seems to be a change
in Linux kernel header
- fixed symbol conflict with s_host of in.h on omniOS platform
from 2.6.8 to 2.6.9
- fixed some obvious bug when running doxygen (inlined documentation)
- fixing configure.ac to detect xattr.h system header when it is
located in /usr/include/sys like under Alpine Linux distro (musl libc)
- fixed typo in symbol name "libdar.archive_summary" in python binding
- fixed bug met when testing an archive in sequential-read mode leading
dar to skip back to test deleted inode which is useless and may lead
to failure if the archive is read from a pipe
- adding *.zst files as excluded from compression when using the
predefined target "compress-exclusion"
- fixed text diagram alignment in documentation and spelling errors
- moving date_past_N_days script to doc/sample with other scripts
from 2.6.7 to 2.6.8
- fixing bug leading binary delta failed to be read from an archive in
some quite rare condition.
- fixed bug that was not listing file with delta path when filtering out
unsaved inodes
- updated source package for the python binding tutorial document be
installed with the rest of the documentation
- adding date_past_N_days helper script to backup only files later than
"today minus N days"
- incorporated the "args" support built script in dar source package
from 2.6.6 to 2.6.7
- fixing shell_interaction_emulator class declaration to avoid compilation
errors and warning under MacOS
- fixed bug: dar failed creating an archive on its standard output
reporting the error message "Skipping backward is not possible on a pipe"
- new feature: added python binding to libdar!
from 2.6.5 to 2.6.6
- fixing script that builds windows binary packages to include missing
cygwin libraries
- fixing bug: dar_manager batch command (-@ option) contains a inverted
test in a sanity check, leading the execution to systematically abort
reporting an internal error message.
- fixed message error type when asymmetrical encryption is requested and
gpgme has not been activated at compilation time
- fixed dar/libdar behavior when gpg binary is not available and gpgme
has been activated at compilation time. Instead of aborting, dar now
signal the gpgme error and proposes to retry initialization without
gpgme support. This makes sense for dar_static binary which stays
usable in that context when all options have been activated
from 2.6.4 to 2.6.5
- fixed bug: dar crashed when the HOME environment variable was not
defined (for example running dar from crontab)
- removed useless skip() in cache layer
- cache layer improvement, flushing write pending data before asking
lower layer about skippability
- fixed bug met when several consecutive compressed files were asked to
be compressed and failed getting reduced in size by libdar. In that
situation as expected, libdar tries to skip backward and stores the
file uncompressed. However the cache layer was introducing an offset
of a few bytes leading the next file to be written over the end of the
previous one, which dar reported as data corruption when testing the
archive.
- updating licensing information with the new address of the FSF
- clarifying message about possibly truncated filename returned by the
operating system
from 2.6.3 to 2.6.4
- fixed display bug indicating delta signatures were about to be
calculated even when this was not the case.
- updating dar man page about the fact aes256 replaced blowfish as
the default strong encryption algorithm
- bug fix: -D option used at creation time was not adding escape mark
of skipped directories. This lead the empty directories that would
replace each skipped one to be inaccessible and unable to be restored
only in sequential read mode (it worked as expected in direct mode)
from 2.6.2 to 2.6.3
- feature enhancement: added option to specify the block size used to
create delta signatures.
- feature enhancement: added the ability to provide login for sftp/ftp
remote access, that contain @ and other special characters.
- fixed bug in dar_xform, leading dar not finding source archive if
destination was not placed in the same directory as source
from 2.6.1 to 2.6.2
- fixed incoherence in documentation
- updating in-lined help information (-h option)
- fixed unexpected behavior of the dar command-line filtering mechanism
met when the provided path to -P or -g options was ending with a slash
- renaming 'path operator + (std::string)' as method append() to avoid
compiler using it when a std::string need first to be converted to path
before adding it to an existing path.
- adding check test to detect when path::append() is used to add a path
instead of a filename to an existing path object.
- adding a warning when restoring a Unix socket if the path to that
socket is larger than what the sockaddr_un system structure can handle
- fixing bug due to Linux system removing file capabilities (stored as
EA) when file ownership is changed. Though restoring EA after ownership
may lead to the impossibility to restore them due to lack of permission
when dar/libdar is not run as root. Thus we try restoring EA a second
time after ownership restoration. This is not efficient but restores
the file as close as possible to their original state whatever
permission dar has been granted for a restoration operation.
from 2.6.0 to 2.6.1
- fixed error in man page
- fixing bug in the routine removing files for local filesystem, used at
archive creation time to remove an existing archive (after user
confirmation), or at restoration time used to remove file that had been
removed since the archive of reference was done. The file to remove was
always removed from the current directory (missing the path part), most
of the time this was leading to the error message "Error removing file
...: Not such file or directory". It could also lead to incorrectly
removing files (not directory) located in the directory from which
dar was run.
- fixing bug met while repairing an archive containing delta signature
for unsaved files
- merging patch from ballsystemlord updating list of file extension
not to compress (see compress-exclusion defined in /etc/darrc)
- review cat_delta_signature implementation in order to be able to fix
memory consumption problem when delta signature are used
- fixed missing mark for data CRC when the data is a delta patch, leading
sequential reading to fail when a delta patch was encountered
- fixed bug in XML output about deleted entries
- fixed XML output to be identical to the one of dar 2.5.x for deleted
entries.
- Adding the deleted date in 'mtime' field for deleted entries in XML
output
- fixing bug in xz/lzma routine wrongly reporting internal error when
corrupted data was met
- fixed code for compilation with clang to succeed (it concerns MAC OS X
in particular)
- fixed inconsistencies in libdar API that avoided gdar to compile with
libdar released in 2.6.0
from 2.5.x to 2.6.0
- new feature: support for binary delta in incremental/differential backups
(relying on librsync)
- new feature: support ftp/sftp to read an archive from a cloud storage.
(relying on libcurl)
reading is optimized to not transfer a whole slice but only the needed
part to proceed to the operation (restoration, listing, and so on)
- new feature: support ftp/sftp to write an archive eventually with hash
files to a remote cloud storage (relying on libcurl)
- modified behavior: While creating a single sliced archive, DUC file is
now executed unless user interrupted dar/libdar. This to stay coherent
with multi sliced archive behavior
- new feature: display filters nature (-vmasks option)
- new feature: follow some symlinks as defined by the --ignored-as-symlink
option
- new feature: one can define the compression algorithm a dar_manager
database will use. This choice is only available at database creation
using the new dar_manager's -z option. In particular "-z none" can be
used to avoid using compression at all
- repair mode added to re-create a completed archive (suitable for direct
access mode and merging) from an interrupted one due to lack of disk
space, power outage or other reasons leading to similar problem.
- Dar can now only save metadata inode change without re-saving the whole
file if its data has not changed. Dar_manager also handle this by
restoring the full backup and then the inode metadata only when
necessary.
- In regard to previous point, if you want to keep having dar saving the
data when only metadata has changed use --modified-data-detection option
- moved dar_slave code into libdar as class libdar::libdar_slave
- moved dar_xform code into libdar as class libdar::libdar_xform
- added libdar_slave and libdar_xform in libdar API
- modified dar_xform and dar_slave to rely on new libdar API
- API: simplified user_interface class
- API: using std::shared_ptr and std::unique_ptr to explicitly show the
ownership of the given pointed objects (C++11 standard)
- API: simplified class archive to only require user_interaction at
object construction time
- API: simplified class database to only require user_interaction at
object construction time
- API: making enum crypto_algo an C++11 "enum class" type
- security refresh: default crypto algo is now AES256. As you do not
need anymore since 2.5.0 to specify the -K option when reading an
archive this should not bring any backward compatibility issue
- security refresh: adding salt per archive (one is still present per
block inside an archive)
- security refresh/new feature: adding option --kdf-param to define
the iteration count for key derivation, which now defaults to 200,000
and hash algorithm used to derived key, still using sha1 by default
- slide effect of previous feature due to starvation of free letters
to add a new command, the -T option with argument is no more
available, one need to provide explicitly the desired argument
- security refresh: improving seed randomization for the pseudo-random
generator used in elastic buffers
- feature enhancement: activate needed Linux capabilities in the
"effective" set if it is permitted but not effective. This concerns
cap_chown at restoration time, cap_fchown for furtive read mode,
cap_linux_immutable to restore the immutable flag, and cap_sys_
resource to set some linux FSA. This let one set the capabilities
for dar binary only in the "permitted" set, capabilities will then be
allowed only for users having them in the "inheritable" set of their
calling process (usually a shell), without root privilege need.
- the ./configure --enable-mode option now defaults to 64, which
will setup a libdar64 in place of infinint based libdar by default.
You can still build a infinint based libdar by passing
--enable-mode=infinint to the ./configure script.
from 2.5.21 to 2.5.22
- removed useless skip() in cache layer
- cache layer improvement, flushing write pending data before asking
lower layer about skippability
- fixed bug met when several consecutive compressed files were asked to
be compressed and failed getting reduced in size by libdar. In that
situation as expected, libdar tries to skip backward and stores the
file uncompressed. However the cache layer was introducing an offset
of a few bytes leading the next file to be written over the end of the
previous one, which dar reported as data corruption when testing the
archive.
- updating licensing information with the new address of the FSF
- fixing bug met when restoring file having FSA but EA and overwriting
an existing file in filesystem
- clarifying message about possibly truncated filename returned by the
operating system
from 2.5.20 to 2.5.21
- bug fix: -D option used at creation time was not adding escape mark
of skipped directories. This lead the empty directories that would
replace each skipped one to be inaccessible and unable to be restored
only in sequential read mode (it worked as expected in direct mode)
from 2.5.19 to 2.5.20
- adding a warning when restoring a unix socket if the path to that
socket is larger than what the sockaddr_un system structure can handle
- fixing bug due to Linux system removing file capabilities (stored as
EA) when file ownership is changed. Though restoring EA after ownership
may lead to the impossibility to restore them due to lack of permission
when dar/libdar is not run as root. Thus we try restoring EA a second
time after ownership restoration. This is not efficient but restores
the file as close as possible to their original state whatever
permission dar has been granted for a restoration operation.
- fixing compilation problem with recent clang++ compiler
from 2.5.18 to 2.5.19
- fixed compilation issue on system that to not have ENOATTR defined
- fixed compilation warning about deprecated dynamic exception
specifications in C++11
- fixing bug in xz/lzma routine wrongly reporting internal error when
corrupted data was met
- fixed compilation warning with gcc about deprecated readdir_r system
call
from 2.5.17 to 2.5.18
- fixed compilation issue in context where EA are not supported
- fixed typo in dar man page (--sequential-mode in place of
--sequential-read)
- moved the "no EA support warning" trigger when restoring an archive
later in the EA restoration process, in order to have the possibility
thanks to the -u "*" option to restore an archive containing EA using a
dar/libdar without EA support activated at compilation time,
- at restoration time, avoiding issuing an "EA are about to be
overwritten" warning when the in place file has in fact not only
one EA set.
from 2.5.16 to 2.5.17
- bug fix: dar failed to restore EA when file permission to restore
did not included user write access. Fix consists in temporarily
adding user write access in order to restore EA and removing this
extra permission afterward if necessary
- updated FAQ
- fixed typos in dar man page
- fixed bug met when writing slices to a read-only filesystem
- fixed compilation problem under Solaris
- fixed typos in dar man page
- bug fix: self reporting bug in filtre.cpp line 2932 or 2925 depending
or dar's version (report occurs in a normal but rare condition that
was not imagined by developer, leading dar to abort the backup)
- bug fix: wrong evaluation of possibility to seek backward in the
escape layer (layer managing tape marks) which lead to useless but
harlmess skip trials in some rare conditions.
from 2.5.15 to 2.5.16
- bug fix: while rechecking sparse file (-ah option) during a merging
operation, dar wrongly reported CRC mismatch for saved plain files
- fixed man page about sparse-file handling while merging: To remove
sparse file datastructure during a merging operation you need to
set --sparse-file-min-size to a valuer larger than all file sizes
contained in the archive (for example 1E for one exabyte)
- bug fix: met when using compression and creating the archive to
dar's standard output (ssh) and leading files to be corrupted in the
archive and reported as such.
- optimisation of escape_sequence skippability (avoids trying skipping
and failing for some corner cases, when we can detect it does even
not worth trying)
from 2.5.14-bis to 2.5.15
- fixing self report bug message met when trying to create an isolated
catalogue into a directory that does not exist
- adding slice overwriting verification before creating a isolated
catalogue, to be coherent with other operations creating an archive
(backup and merging)
- storage size of compressed files was often wrongly stored in archive
(shorter than reality), the only impact took place at archive listing
time where the compression ratio displayed was better than reality
- fixed auto-detected bug condition triggered when -Tslicing is used
with --sequential-read. Both options are not compatible and have been
excluded by a nicer message than this auto-detection bug message.
from 2.5.14 to 2.5.14-bis
- avoiding using the syncfs() system call in dar_split when the
platform does not support it (replacing it by sync() in that case
for compilation to be successful)
from 2.5.13 to 2.5.14
- made libgcrypt built-in memory guard be initialized before obtaining
ligcrypt version, to respect libgcrypt usage (but no problem was seen
nor reported about this inconsistency)
- fixed syntax error in XML listing output (EA_entry and Attributes
tags)
- fixed typos in dar man page
- Updating Tutorial for restoration
- fixed bugs in dar_split: cygwin support, filedescriptors were not
explicitly closed at end of execution, allocating buffer on heap
rather than in the stack for better size flexibility, avoiding buffer
size to be greater than SSIZE_MAX.
- added -s option to dar_split in order to disable the by default SYNC
write that was used and which caused poor performance. To keep the
same behavior as the older dar_split (and its poor performances) you
need now using -s option.
- dar_split enhancement: added call to syncfs before closing the file
descriptor in split_output mode
- fixed bug in dar_split that was did not lead dar_split to completely
fulfill an device before asking for user to change the media when
used in split_output mode, this was sometimes leading dar reporting
file as corrupted at dar_split at media boundary.
- added feature in dar_split to show the amount of data written since
the last media change
from 2.5.12 to 2.5.13
- added -az option to automatically nullify negative dates returned from
the system in the archive under creation (filesystem is not modified)
- included the birthtime (HFS FSA) into the negative dates handling
- modified behavior: dar now fails upon unknown option instead of warning
the option is unknown and thus ignored
- bug fix: dar 2.5.12 and below in 2.5.x branch could not read archive
generated by dar 2.4.x and below (unless in infinint compilation mode)
when the old archive included a file which date(s) was returned by the
system as a negative integer at the time of the backup. Note that if
dar can now read old archive in that particular case, such date stay
recorded in the dar archive as very far in the future and not in the
past, because 2.4.x and below blindly assumed the system would always
return a positive integer as number of second since 1970. Since 2.5.12
release, when the system provides a negative date the date is assumed
as zero (Jan 1970) with user agreement.
- fixed missing throw in tools.cpp (exception condition was not reported)
from 2.5.11 to 2.5.12
- documenting in man page the limitation of -[ and -] options concerning
the maximum line length that could be used in a listing file. This
limitation was only described in doc/Limitations.html
- dar now aborts if a line exceeding 20479 bytes is met in a listing file
- improved error message issued when a file listing (-[ or -] option) is
missing for it provides the missing filename in the error message
- improved error message issued when a line of a file listing exceeds
20479 characters for it display the start of that line
- fixed bug in file listing (-[ option) leading some directories and their
content to be excluded in a somehow rare condition
- improved behavior when dar reads a negative date. Instead of aborting
it now asks the user if it can substitute such value by zero
- improved behavior when dar is asked to read an archive located in a
directory that does not exist. DUC file passed to -E option is now
properly run in that case too and has the possibility for example to
create that directory and download requested file
from 2.5.10 to 2.5.11
- minor feature: displays the archive header which is never ciphered and
aborts. This feature is activated while listing archive content and
adding the -aheader option. This brings the side effect to invert two
lines in the archive summary (dar -l archive -q) "catalogue size" and
"user comment".
- adding date format info for -w option in "dar_manager -h" usage help
- fixed several mistakes in tools.cpp leading compilation to fail under
certain environments
- fixed a typo in filesystem.cpp and portability issue that lead
compilation to fail under openbsd 6.0
- fixed bug in the filtering mechanism relying on file listing (-[ and
-] options) that could not find an entry in the listing upon certain
condition leading a file not being excluded as requested or not
included as requested
from 2.5.9 to 2.5.10
- fixed bug: -r option (only more recent overwriting policy) was
considering a file to be more recent when it had the exact same date as
the file in place.
- updating documentation about requirements for compiling dar from sources
- fixed bug: bug met when restoring of a file that has the immutable
flag set. Dar/libdar failed restoring such file in the context of
differential/incremental backup. The fix consists of the removal of the
immutable flag from filesystem before restoring the new version of the
file's data, then setting back the immutable flag afterward.
- updating FAQ with description of the way dar uses lzo compression
compared to the lzop program
- fixed bug: aborting an archive was leading to an unreadable archive in
direct mode, most of the time when strong encryption was used
- minor new feature: added two flavors of lzo algorithm: lzop-1 and lzop-3
in order to match compression levels 1 and 3 of the lzop command
from 2.5.8 to 2.5.9
- fixed typos in documentation about dar internal use of symmetric
encryption
- fixed bug: merging operation could wrongly melt different unrelated hard
linked inodes when merging using an archive which results from a previous
merging operation.
- fixed bug: aborting an archive was sometimes leading to an unreadable
archive in direct mode (was readable only in --sequential-read mode)
- fixed bug: libgpgme was only present at linking time of final binaries
(dar, dar_slave, dar_xform, dar_manager, dar_cp, dar_split), not at
linking time of libdar, which caused problem under Linux Rosa distro
where the "no-undefined" flag is passed to the linker.
- minor new feature: -ay option has been added to display sizes in bytes
instead of the default which uses the largest possible unit (Kio, Mio,
and so on.)
from 2.5.7 to 2.5.8
- fixed double memory release occurring in a particular case of read error
- improving robustness of FSA code against data corruption
- fixed bug: DAR_DUC_PATH was not used with -F and -~ options
- new feature: added -aduc option to combine several -E options using the
shell '&&' operator rather than the shell ';' operator. The consequence
is that with -aduc option a non zero exist status of any script (and not
only of the script given to the last -E option) will lead dar to report
the error.
- man page updated about combination of several -E options
- fixed bug: merging partial FSA led to self reported bug in cat_inode.cpp
at line 615
from 2.5.6 to 2.5.7
- fixed bug leading dar to not include directories given to -g option nor
to exclude directories given to -P option when at the same time the
directory given to -R option starts by a dot ("-R ./here" in place of
"-R here")
- bug fix and speed improvement: under certain circumstances dar was
reading several times the data at slice boundary, leading dar to ask for
slice N then N-1 then again N, this caused sub-optimal performance and
was triggering user script unnecessarily
from 2.5.5 to 2.5.6
- added speed optimization when comparing dates with hourshift flexibility
(-H option)
- fixed bug met when using as reference an archive generated by dar 2.5.4
or older, bug that lead dar saving almost all file even those that did
not change.
from 2.5.4 to 2.5.5
- fixed message displayed when reading old archives
- fixed bug that avoided dar-2.5.x code to read old archive format when
special allocation was set (by defaut) at compilation time
- disabling special-alloc by default reducing memory footprint
- fixed error in FAQ about the way ctime/atime/mtime are modified during
normal operating system life.
- new implementation of class datetime with better memory footprint
- avoding storing sub-microsecond part of date to preserve limitint
ability to store large dates
- moving field cat_inode::last_cha from pointer-to-field to plain field
of the class, this slightly reduce catalogue memory footprint.
- fixing bug in the returned exit status when dar failed executing DUC
command due to system error (now returning the expected code 6 in that
case too)
from 2.5.3 to 2.5.4
- fixing missing included files for libdar API
- removed extra try/catch block introduced by commit
72da5cad5e52f959414b3163a2e2a320c2bc721e
- removed sanity check that caused problem when writing an archive to a
FUSE based filesystem.
- fixing non call of the -E script/command after last slice creation,
when encryption or slice hashing was used
- fixed bug in dar_manager: archive permutation in database lead libdar
to check an archive number of range under certain circumstances
- fixed inversion of the condition triggering a warning about archive
date order in a dar_manager database while moving an archive within a
database
- fixed typos in documentation
- catalogue memory optimization, with the drawback to limit the number of
entry in an archive to the max integer supported by the libdar flavor
(32 bits/64 bits/infinint).
- fix configure script to temporarily rely on LIBS rather LDFLAGS to
check for gpgme availability
- removed order dependency between -A and -9 options of dar_manager: -9
can now be specified before or after -A option.
- resetting to "false" the "inode_wrote" flag of hard link data-structure
before testing and merging. Merging a previously tested archive or
testing a second time would not include hard linked inode in the
operation. This situation does not occurs with dar but could succeed
with some external tools that keep the catalogue in memory to perform
different operations on it.
- fixed bug in the routine that detects existing slices to warn the user
and/or avoid overwriting, bug that lead dar to "bark" when an archive
base name started by a + character.
- avoiding to use AM_PATH_GPGME in configure script when gpgme.m4 is not
available
- adding new methods in libdar API to obtain the archive offset and
storage size of saved files (class list_entry)
- adding new method in libdar API to translate archive offset to file
offset (class archive)
- reporting a specific error message when filename returned by the system
has the maximum length supported by the system itself, assuming
filename has been truncated
from 2.5.2 to 2.5.3
- Fixing a 2.5.x build issue met when a 2.4.x libdar library is already
installed in an FreeBSD system.
- Improving message and behavior of libdar in lax mod when a truncated
archive is read
- Fixing self reported bug at "tronconneuse.cpp line 561" met while
reading truncated/corrupted archive
- Fixed not closed filedescriptors, met when saving a filesystem that has
not ExtX FSA available
- Fixing configure script to be more robust in front of system where
gpgme.h is installed in a non standard path and user did not provide
coherent CPPFLAGS, LDFLAGS before calling ./configure
- Displaying CRC values when listing isolated catalog as XML output
- Fixing compilation issue when system does not provide strerror_r() call
- Avoiding warning about FSA absence when fsa-scope is set to "none"
- Adding --disable-fadvise option to configure script for those that want
back full pressure from dar on the system cache (same behavior as 2.4.x)
- Fixing bug, fadvise() called a wrong time making it having no effect
- updating FAQ about comparative performance from 2.4.x to 2.5.x
- optimization: reduced the number of call to dup() at libdar startup
- improvement: printing file type on verbose output
- new feature: added %t macro reflecting the inode type in dar's
--backup-hook-execute option
from 2.5.1 to 2.5.2
- fixed bug met when permission is denied while reading or writing slices
- fixing bug that avoided creating an archive at the root of the filesystem
- fixing bug met in rare situation while reading in sequential-read mode
an archive encrypted using gnupg encryption. In that situation libdar
may fail reading the archive (but succeeds in normal read mode) issuing
an obscure message (message has also been fixed).
- code simplification, removing field reading_verion from class crypto_sym
as its parent class tronconneuse already have such information
- removed extra newline displayed by dar at end of execution
- fixed bug avoiding dar to properly read an entry (reporting CRC error)
when specific sequence of character (start of escape sequence) fall at
end of the read buffer of the escape layer.
- speed optimization for datetime class
- fixed bug that avoided dar reading archives in sequential read mode
while reading from a pipe
- fixed bug in non regression test routine provided beside dar/libdar
- fixing display message showing not always in the correct context
- fixing case inversion leading the cache layer not to be used when
necessary and used when useless while reading an archive
- improved heuristic in dar_manager to determine the date a file has been
deleted.
from 2.5.0 to 2.5.1
- fixed display bug in dar_manager met when using -o option and adding
options for dar that does not exist for dar_manager (like -R option)
- reactivating disabled (by mistake) optimization for some read-only dar
manager database operations
- fixing compilation issue with dar against gcc 4.9.2
- fixing syntax error in dar_manager message
- fixed bug that forbade dar_manager to write down modified database
when only database header was modified (-o, -b, -p switches).
- adding dar_manager database format version information with -l option
- fixed libdar inability to read dar_manager's database format version 4
- adapting code to build under cygwin environment, where thread_local seems
broken
- fixed output to stderr in place of stdout for Licensing information
- fixed bug met when permission is denied while reading or writing slices
- fixing bug that avoided creating an archive at the root of the filesystem
from 2.4.x to 2.5.0
- added support for posix_fadvise()
- added entrepot class hierarchy to support in the future other storage
types than local filesystem for slices
- API: added access to the entrepot through the API
- modified class hash_fichier for it becomes entrepot independent
- API: extended libdar API with an additional and more simple way to read
an archive: archive::get_children_in_table() method, see
doc/API_tutorial.html for details
- added support for extX (see lsattr(1)) and HFS+ (birthtime date)
Filesystem Specific Attributes (FSA).
- dar is now able to skip backward when a file is found to be "dirty" at
backup time. This avoids wasting space in archive but is only possible if
the backward position is located in the current slice and no slice
hashing nor strong encryption is used. Of course if the archive is
written to a pipe or to stdout, skipping back to retry saving data at the
same place is neither possible, --retry-on-change option stays possible
in that cases at the cost of data duplication (wasted bytes amount, see
--retry-on-change man page).
- by default dar now performs up to 3 retries but do not allow for wasting
bytes if file has changed at the time it was read for backup,
this can be modied using --retry-on-change option.
- With the same constraints as for a changing file, if a file is saved
compressed but its compressed data uses more space than uncompressed,
the file's data is resaved as uncompressed. However, if skipping backward
is not possible, data is kept compressed.
- if system provides it, dar uses "Linux capabilities" to check for the
ability to set file ownership when dar is not run as root. This allows dar
to restore ownership when allowed even when it is not run as superuser.
- removing dar-help tool used to build dar -h messages. That tool became
useless for a long time now.
- added several more specific verbosity options: -vm, -vf and -vt
- added support for microsecond timestamps (atime, mtime, ctime, birthtime)
- Using lutime() to restore atime/mtimes of symlink on systems that support
it.
- API: removed backward compatible API for old libdar 4.4.x
- API: simplified implementation of archive isolation thanks to isolation
evolution features brought by release 2.4.0. Memory requirement is now
devided by two compared to releases of previous branch (2.4.x).
- dar has been updated to use this new API for archive isolation
- added exclude-by-ea feature to avoid saving inodes that have a particular
user defined EA set.
- added comparison of an isolated catalogue with a filesystem, relying on
embedded data CRC and inode metadata in absence of the saved data.
- The new archive format (version 9) holds the ciphering algorithm used
at creation time, only the passphrase is now required at reading time and
-K option may be ignored which will lead dar to prompt for passphrase.
- Adding support for public key encryption (GnuPG) supporting several
encryption keys/recipients for a given archive
- Adding support for public key signature when public key encryption is used
- While listing archive contents, directories now show the size and average
compression ratio of the data they contain
- Archive summary (-l with -q options) now reports the global compression
ratio
- added the -vd switch to only display current directory under process for
creation, diff, test, extraction and merging operations
- added xz/lzma compression support
- added -Tslicing listing option to show slice location of files inside
an archive archive.
- isolated catalogues now keep a record of the slicing layout of their
archive of reference in order to provide -Tslicing feature when used
on the isolated catalogue alone.
- However if an archive has been resliced (using dar_xform) after its
isolated catalogue has been generated, using -Tslicing option with the
isolated catalogue would give wrong information. To overcome that, it
is possible to specify what is the new slicing of the archive of
reference by using the -s and -S options in conjunction with -Tslicing
- added dar_split command to provide on-fly multi-volume archive support
for tape media
- experimental feature to have libdar using several threads (not activated
by default due to poor performance gain)
- dar now aborts when a given user target cannot be found in included file
- added sha512 hashing algorithm beside already available md5 and sha1, the
generated hash file can be used with 'sha512sum -c <file>' command
- removed useless --jog option for memory management
- removed previously deprecated -y/--bzip2 command, bzip2 compression
remains available using -z option (-zbzip2 or --compression=bzip2)
- replaced SHA1 by SHA224 to generate IV for encryption blocks, this
slightly improves randomness of IV and stay available when libgcrypt is
run in FIPS mode
from 2.4.23 to 2.4.24
- fixed bug: merging operation could wrongly melt different unrelated hard
linked inodes when merging using an archive which results from a previous
merging operation.
from 2.4.22 to 2.4.23
- fixed bug leading dar to not include directories given to -g option nor
to exclude directories given to -P option when at the same time the
directory given to -R option starts by a dot ("-R ./here" in place of
"-R here")
from 2.4.21 to 2.4.22
- fixing bug in the returned exit status when dar failed executing DUC
command due to system error (now returning the expected code 6 in that
case too)
from 2.4.20 to 2.4.21
- removed sanity check that caused problem when writing an archive to a
FUSE based filesystem.
- fixed bug in dar_manager: archive permutation in database lead libdar
to check an archive number out of range under certain circumstances
- fixed inversion of the condition triggering a warning about archive
date order in a dar_manager database while moving an archive within a
database
- removed order dependency between -A and -9 options of dar_manager: -9
can now be specified before or after -A option.
- resetting to "false" the "inode_wrote" flag of hard link datastructure
before testing and merging. Merging a previously tested archive or
testing a second time would not include hard linked inode in the
operation. This situation does not occurs with dar but could succeed
with some external tools that keep the catalogue in memory to perform
different operations on it.
- fixed bug in the routine that detects existing slices to warn the user
and/or avoid overwriting, bug that lead dar to "bark" when an archive
base name started by a + character.
from 2.4.19 to 2.4.20
- fixed display bug in dar_manager met when using -o option and adding
options for dar that does not exist for dar_manager (like -R option)
- reactivating disabled (by mistake) optimization for some read-only dar
manager database operations
- fixing compilation issue with dar against gcc 4.9.2
- fixing syntax error in dar_manager message
- fixing bug that avoided creating an archive at the root of the filesystem
from 2.4.18 to 2.4.19
- fixed missing quote in dar_par.dcf which is called by the par2 directive
- fixed bug in dar_manager's -u option, not displaying most recent files of
an archive when they have been marked as removed in a more recent archive
of the same dar_manager database.
- fixed bug met while restoring in sequential read mode a file having
several copies (was modified at the time it was saved and retry-on-change
was set).
from 2.4.17 to 2.4.18
- Initial Vector used for strong encryption was set with pseudo-random data
generated using SHA1 message digest and blowfish cipher, which are not
available when ligcrypt is running in FIPS mode. Since 2.4.18 we now use
SHA256 and AES256 for IV assignment in order to have libdar compatible
with FIPS mode. For data encryption nothing changes: the cipher specified
(-K, -J, -$ options on CLI) are used as before.
- fixing bug met when performing archive isolation in sequential-read mode,
If an archive corruption or truncated archive leads an inode to not have
its CRC readable, dar aborts and issues a BUG report.
- updating list of project relying on dar/libdar
from 2.4.16 to 2.4.17
- fixing issue when case insensitive comparison was requested and invalid
wide char for the current local was met in a filename. In such situation
the corresponding file was never saved before (considering a filesystem
error for that file), while now the ASCII case insensitivity is used
as fallback.
from 2.4.15 to 2.4.16
- fixing archive listing displayed information for catalogue size when
archive is read in --sequential-read mode
- fixing bug that avoided dar releases 2.4.x up to 2.4.15 to read encrypted
archive generated by dar release 2.3.x and below
- adding informational note at the end of ./configure script execution
when --enable-mode has not been used.
- adding support for case sensitivity in filename comparison (-an option)
for other character sets than POSIX/C locale like Cyrillic for example.
- fixing bashisms in doc/samples scripts
from 2.4.14 to 2.4.15
- fixing bug met when reading an encrypted archive in sequential mode
- fixing bug met when reading an encrypted archive in sequential mode from
an anonymous pipe
- changed option '-;' to -9 as '-;' does not work on all systems with getopt
(only long option equivalent --min-digits worked) for dar, dar_cp,
dar_manager, dar_xform and dar_slave commands.
- fixing bug met when restoring deleted files in sequential read mode and
some directory where they should be "restored" are not readable or
could not be restored earlier
- adding extra buffer to handle sequential read of encrypted archive
when the last crypto block contains some but not all clear data after
encrypted one (the archive trailer).
- fixing compilation issue using clang
- fixing bug that prevents using -~ option with on-fly catalogue isolation
in order to execute an user command once on-fly isolation has completed
- added some autoconf magic to determine the correct (BSD/GNU) flag to
use with sed in order to activate regular expression parsing
- new implementation of mask_list class which is compatible with libc++
- fixed bug met on FreeBSD with dar_xform where the system provides a
standard input file descriptor in read-write instead of read-only mode.
from 2.4.13 to 2.4.14
- limiting memory consumption of the cache layer to stay below 10 MiB, under
certain circumstances (very large archive), it could grow up to an insane
value like 50% or the available RAM. reducing to 10 MiB does not impact
performance in a noticeable manner while it avoids system to swap out due
to the libdar cache layer becoming huge.
- added --with-pkgconfigdir to define an alternative path for libdar
pkgconfig file (to ease portability to FreeBSD)
- modified some Makefile.am for better FreeBSD support
- fixed display bug in XML listing output concerning hard linked inodes
- fixing typo in man page
- fixing bug met while isolating a catalogue in --sequential-read mode.
Using such isolated catalogue lead dar report an error about inaccessible
EA.
- displaying compression rate for sparse files even when they are
uncompressed, sparse file detection also leads to compress files
- fixing bug that lead libdar to fail comparing an inode having EA when
comparison is done in --sequential-read mode
- fixing display bug in in ligcrypt check of configure script for minimum
required version
- fixing 'make clean' to remove some forgotten files generated by 'make'
|
|
|
|
Bump PKGREVISION.
Changelog from AN-2022-09-18:
With great sadness we report the death of schilytools main author and
long time maintainer Jörg Schilling (1955-03-27--2021-10-10). Jörg, we
miss you! After his passing, development has been picked up by a group
of volunteers who would like to keep the project alive and well. Our
goal is to carefully preserve the portability and compatibility of this
project while extending it to new platforms and implementing new tools
and features that fit the tool set.
As a part of the change in management, the schilytools project has moved
from Sourceforge to Codeberg.
- general: Explain how to contribute and update README
- general: Add list of contributors
- all: Misc spelling fixes
Submitted by Friedhelm Mehnert.
- all: Update E-Mail address and project URL in many places.
Jörg's old joerg@schily.net address is defunct and so is his
postal address. As a part of an ongoing documentation cleanup,
these addresses have been replaced by our new project pages in
many locations.
Submitted by Nico Sonack.
- man pages: work around buggy syntax highlighting by adding
\" " comments into various places.
Submitted by Nico Sonack.
- RULES: Fix build on macOS 12.4
- RULES: add i386-openbsd-clang.rul
This should permit compilation on modern OpenBSD using clang as
the system compiler.
- autoconf: Work around clang bug in #pragma weak detection.
A clang compiler bug causing a crash in the configure test for
#pragma weak has been worked around, permitting use of weak
symbols when compiling with clang.
- tests: fix arm64/aarch64 architecture confusion on OpenBSD
Submitted by Nico Sonack.
|
|
Changelog from AN-2022-09-18:
With great sadness we report the death of schilytools main author and
long time maintainer Jörg Schilling (1955-03-27--2021-10-10). Jörg, we
miss you! After his passing, development has been picked up by a group
of volunteers who would like to keep the project alive and well. Our
goal is to carefully preserve the portability and compatibility of this
project while extending it to new platforms and implementing new tools
and features that fit the tool set.
As a part of the change in management, the schilytools project has moved
from Sourceforge to Codeberg.
- general: Explain how to contribute and update README
- general: Add list of contributors
- all: Misc spelling fixes
Submitted by Friedhelm Mehnert.
- all: Update E-Mail address and project URL in many places.
Jörg's old joerg@schily.net address is defunct and so is his
postal address. As a part of an ongoing documentation cleanup,
these addresses have been replaced by our new project pages in
many locations.
Submitted by Nico Sonack.
- man pages: work around buggy syntax highlighting by adding
\" " comments into various places.
Submitted by Nico Sonack.
- RULES: Fix build on macOS 12.4
- RULES: add i386-openbsd-clang.rul
This should permit compilation on modern OpenBSD using clang as
the system compiler.
- autoconf: Work around clang bug in #pragma weak detection.
A clang compiler bug causing a crash in the configure test for
#pragma weak has been worked around, permitting use of weak
symbols when compiling with clang.
- star: convert star_fat.c generation to MKLINKS
- star: fix wrong handling of pre-1970/post-2036 dates in gnutar
archives on platforms with sizeof(time_t) > sizeof(long).
- tests: fix arm64/aarch64 architecture confusion on OpenBSD
Submitted by Nico Sonack.
|
|
builds the same object files for different targets
e.g.
ld: build/temp.netbsd-9.0-amd64-cpython-310/lz4libs/lz4.o: file not recognized: file truncated
|
|
16 months of development.
|
|
5 months of development.
|
|
|
|
V0.10.0
--------
- Considered the modes of emblemized icons. This is especially relevant to Qt
icon engines that colorize selected symbolic icons.
- Get the mimetypes of symlinks directly, without consulting their targets
(because their targets may be symlinks themselves and so, they aren't
reliable).
- First try the file manager DBus interface when opening the containing folder.
- Removed the action for opening the containing folder when running as root
(because there are DBus complications with most file managers).
- Added read-only support for VirtualBox Extension Pack.
- With 7z, don't show password prompt when the archive can't be read and
there's no mention of encryption.
- Enabled compilation against Qt6. With cmake, "ENABLE_QT5" should be set to
"OFF" (as in `cmake .. -DENABLE_QT5=OFF`) — it's "ON" by default.
|
|
5.2.6 (2022-08-12)
* xz:
- The --keep option now accepts symlinks, hardlinks, and
setuid, setgid, and sticky files. Previously this required
using --force.
- When copying metadata from the source file to the destination
file, don't try to set the group (GID) if it is already set
correctly. This avoids a failure on OpenBSD (and possibly on
a few other OSes) where files may get created so that their
group doesn't belong to the user, and fchown(2) can fail even
if it needs to do nothing.
- Cap --memlimit-compress to 2000 MiB instead of 4020 MiB on
MIPS32 because on MIPS32 userspace processes are limited
to 2 GiB of address space.
* liblzma:
- Fixed a missing error-check in the threaded encoder. If a
small memory allocation fails, a .xz file with an invalid
Index field would be created. Decompressing such a file would
produce the correct output but result in an error at the end.
Thus this is a "mild" data corruption bug. Note that while
a failed memory allocation can trigger the bug, it cannot
cause invalid memory access.
- The decoder for .lzma files now supports files that have
uncompressed size stored in the header and still use the
end of payload marker (end of stream marker) at the end
of the LZMA stream. Such files are rare but, according to
the documentation in LZMA SDK, they are valid.
doc/lzma-file-format.txt was updated too.
- Improved 32-bit x86 assembly files:
* Support Intel Control-flow Enforcement Technology (CET)
* Use non-executable stack on FreeBSD.
- Visual Studio: Use non-standard _MSVC_LANG to detect C++
standard version in the lzma.h API header. It's used to
detect when "noexcept" can be used.
* xzgrep:
- Fixed arbitrary command injection via a malicious filename
(CVE-2022-1271, ZDI-CAN-16587). A standalone patch for
this was released to the public on 2022-04-07. A slight
robustness improvement has been made since then and, if
using GNU or *BSD grep, a new faster method is now used
that doesn't use the old sed-based construct at all. This
also fixes bad output with GNU grep >= 3.5 (2020-09-27)
when xzgrepping binary files.
This vulnerability was discovered by:
cleemy desu wayo working with Trend Micro Zero Day Initiative
- Fixed detection of corrupt .bz2 files.
- Improved error handling to fix exit status in some situations
and to fix handling of signals: in some situations a signal
didn't make xzgrep exit when it clearly should have. It's
possible that the signal handling still isn't quite perfect
but hopefully it's good enough.
- Documented exit statuses on the man page.
- xzegrep and xzfgrep now use "grep -E" and "grep -F" instead
of the deprecated egrep and fgrep commands.
- Fixed parsing of the options -E, -F, -G, -P, and -X. The
problem occurred when multiple options were specied in
a single argument, for example,
echo foo | xzgrep -Fe foo
treated foo as a filename because -Fe wasn't correctly
split into -F -e.
- Added zstd support.
* xzdiff/xzcmp:
- Fixed wrong exit status. Exit status could be 2 when the
correct value is 1.
- Documented on the man page that exit status of 2 is used
for decompression errors.
- Added zstd support.
* xzless:
- Fix less(1) version detection. It failed if the version number
from "less -V" contained a dot.
* Translations:
- Added new translations: Catalan, Croatian, Esperanto,
Korean, Portuguese, Romanian, Serbian, Spanish, Swedish,
and Ukrainian
- Updated the Brazilian Portuguese translation.
- Added French man page translation. This and the existing
German translation aren't complete anymore because the
English man pages got a few updates and the translators
weren't reached so that they could update their work.
* Build systems:
- Windows: Fix building of resource files when config.h isn't
used. CMake + Visual Studio can now build liblzma.dll.
- Various fixes to the CMake support. Building static or shared
liblzma should work fine in most cases. In contrast, building
the command line tools with CMake is still clearly incomplete
and experimental and should be used for testing only.
|
|
v1.9.4:
perf : faster decoding speed (~+20%) on Apple Silicon platforms, by @zeux
perf : faster decoding speed (~+70%) for -BD4 setting in CLI
api : new function LZ4_decompress_safe_partial_usingDict() by @yawqi
api : lz4frame: ability to provide custom allocators at state creation
api : can skip checksum validation for improved decoding speed
api : new experimental unit lz4file for file i/o API, by @anjiahao1
api : new experimental function LZ4F_uncompressedUpdate(), by @alexmohr
cli : --list works on stdin input, by @Low-power
cli : --no-crc does not produce (compression) nor check (decompression) checksums
cli : fix: --test and --list produce an error code when parsing invalid input
cli : fix: support skippable frames when passed via stdin, reported by @davidmankin
build: fix: Makefile respects CFLAGS directives passed via environment variable
build: LZ4_FREESTANDING, new build macro for freestanding environments, by @t-mat
build: make and make test are compatible with -j parallel run
build: AS/400 compatibility, by @jonrumsey
build: Solaris 10 compatibility, by @pekdon
build: MSVC 2022 support, by @t-mat
build: improved meson script, by @eli-schwartz
doc : Updated LZ4 block format, provide an "implementation notes" section
|
|
The package changed with the addition of its libepoll-shim dependency.
Otherwise, we can get:
ERROR: libepoll-shim>=0.0.20210418 is not installed; can't buildlink files.
|
