* POSIX extended format headers do not include PID by default
* --delay-directory-restore works for archives with reversed member ordering
* Fix extraction of a symbolic link hardlinked to another symbolic link
* Wildcards in exclude-vcs-ignore mode don't match slash
* Fix the --no-overwrite-dir option
* Fix handling of chained renames in incremental backups
* Link counting works for file names supplied with -T
* Accept only position-sensitive (file-selection) options in file list files
Stefan Hagen provided analysis of a regression that has been fixed
by upstream (cherrypicked).
The -Wl,--as-needed linker flag causes the std::unique_ptr configure
test to fail due to undefined reference errors. A fallback exists, using
std::auto_ptr, but it has been removed from the C++17 standard, used by
upstream.
Remove the use of -Wl,--as-needed on powerpc, so we use the same code on
all archs, instead of reenabling std::auto_ptr by force.
OK thfr@ (maintainer)
Changes: https://github.com/lz4/lz4/releases/tag/v1.9.3
Minor of shared lib has been bumped because of the addition of a symbol.
While here update HOMEPAGE and take MAINTAINER.
OK sthen@
- remove patch from upstream fixing build failure with gcc
- patch CMakeLists.txt to use pkg-config to find system zstd (instead of
trying to fetch zstd source from github..)
- enable zstd support
Minizip was originally developed in 1998. It was first included in the zlib
distribution as an additional code contribution starting in zlib 1.1.2. Since
that time, it has been continually improved upon and contributed to by many
people.
+ Creating and extracting zip archives.
+ Adding and removing entries from zip archives.
+ Read and write raw zip entry data.
+ Reading and writing zip archives from memory.
+ Zlib, BZIP2, LZMA, and ZSTD compression methods.
+ Password protection through Traditional PKWARE and WinZIP AES encryption.
+ Buffered streaming for improved I/O performance.
with fixes and tweaks from cwen@:
- add patch to fix build on gcc
- disable -Os
- build & enable tests
some ports bundle it, some ports allow building against a systemwide
one, some ports bundle/rely on an incompatible version - all those will be
fixed in following commits.
required by an upcoming geo/spatialite/libspatialite update.
ok bentley@ cwen@
algorithm that compresses data using a combination of a modern
variant of the LZ77 algorithm, Huffman coding and 2nd order context
modeling, with a compression ratio comparable to the best currently
available general-purpose compression methods. It is similar in
speed with deflate but offers more dense compression.
Original version by thfr. Moved to current python port standards by
me. Typo fix pointed out by sthen.
ok sthen
The bindings provided in this package cover the frame format, the
block format, and the streaming format specifications. The frame
format bindings are the recommended ones to use, as this guarantees
interoperability with other implementations and language bindings.
Needed for thfr's pending py-unitypy port and another port I have in progress
Some details picked up from thfr's version. Tweak from sthen.
ok sthen
- 'COMPILER=base-clang ports-gcc' is needed because of -Wvla in a couple
of Makefiles, which is easily patched out
- switch ALL_TARGET to 'default' as 'allmost' builds stuff that we do
not package. This saves a couple of seconds while building, plus the
build is C-only
- No need for WANTLIB+=atomic/LDFLAGS+=-latomic (hppa only). There is
nothing related to atomics in the stuff that is build (related to
ALL_TARGET)
- No need to set CXX* in MAKE_ENV as the build is C-only
- No need to set CFLAGS. Upstream's (GNU)Makefile properly amends CFLAGS
on a per-target basis. We already provide CFLAGS through MAKE_ENV, let
the Makefile do its job
- Update CPPFLAGS
OK sthen@
archivers/zstd has tripled. This is caused by zstd building everything
in examples and contrib, which we do not package. Setting ALL_TARGET
fixes this. Build time, on my x230, dropped from 20m41.61s to 5m47.06s.
No REVISION bump required.
OK naddy@
I highly suspect this is not needed but it's too tempting to just bump
three ports and forget about it.
No objection from tb@ (py-tlslite-ng maintainer)
LASzip quickly turns bulky LAS files into compact LAZ files without
information loss and is the de-facto standard for LiDAR compression.
LASzip compression can be many times smaller and many times faster than
generic compressors like bz2, gzip, and rar because it knows what the
different bytes in a LAS file represent.
with tweaks from & ok sthen@ jca@
CVE-2018-1000035 (heap overflow in processing password-protected archives)
CVE-2019-13232 (mishandles the overlapping of files inside a ZIP container)
From Moritz Buhl
revert to using MASTER_SITES again
successfully extracts a few test archives from GOG.com
While here, add comments with reason for COMPILER=base-clang ports-gcc
(C++17), and reason for WANTLIBs boost_system-mt and pthread which
portcheck lists as "Extra"
Changelog: https://constexpr.org/innoextract/changelog
This is used for -T0 which auto-detects the number of cores for parallel
compression. xz prefers sysctl over sysconf (see m4/tuklib_cpucores.m4
and src/common/tuklib_cpucores.c for reasons) but this doesn't work for
us; just set an autoconf cache variable to force sysconf which works
better for us.
Remove the qt5 FLAVOR; qt5 is the only supported/used version these days.
Disable the tests; they've not been updated for qt5 and aren't accessible
from the standard cmake build system.
Update all dependent ports to cope with the FLAVOR removal.
Requested by rsadowski@; improvements by sthen@
Python 3.3 onwards includes module lzma in the standard library,
providing support for working with LZMA and XZ compressed files via the
XZ Utils C library (XZ Utils is in a sense LZMA v2).
ok sthen@
Blosc is a high performance compressor optimized for binary data. It has been
designed to transmit data to the processor cache faster than the traditional,
non-compressed, direct memory fetch approach via a memcpy() OS call. Blosc is
meant not only to reduce the size of large datasets on-disk or in-memory, but
also to accelerate memory-bound computations.
It uses the blocking technique so as to reduce activity in the memory bus as
much as possible. In short, this technique works by dividing datasets in blocks
that are small enough to fit in caches of modern processors and perform
compression/decompression there. It also leverages, if available, SIMD
instructions (SSE2, AVX2) and multi-threading capabilities of CPUs, in order to
accelerate the compression/decompression process to a maximum.
From martin@; input and OK from bcallah@, feinerer@, and sthen@
* Fix CVE-2015-1197
* Fix CVE-2016-2037
* Fix CVE-2019-14866
* Remove --extract-over-symlinks option again, which was part of an earlier
non-upstream fix for CVE-2015-1197.
Version 3.1
DESCR:
rarfile is a Python module for RAR archive reading. The interface
is made as zipfile like as possible.
Supports both RAR3 and RAR5 format archives, multi volume archives,
Unicode filenames, password-protected archives, archive and file
comments, and archive parsing and non-compressed files handled with
pure Python code, Compressed files are extracted by executing either
unrar from RARLAB or bsdtar from libarchive. Works with both Python
2.7 and 3.x.
ok landry
ghc and the hs-packages now simply include the necessary (haskell)
package description files in lib/ghc/package.conf.d and update the
package.cache by running ghc-pkg recache at the end. register and
unregister scripts are no longer needed.
* relaxes the fix for CVE-2019-12900
* cleanup of bzdiff and bzgrep scripts so they don't use bash extensions
There's also a homepage again.
ok kmos@ cwen@
Maintenance release, which includes a few bug fixes, including a fix for
a rare data corruption bug, which could only be triggered in a niche use
case.
List of changes can be found at
https://github.com/facebook/zstd/releases/tag/v1.4.1
internal replacement function. Following the changes to make realpath(3) use the
__realpath() syscall these no longer detect broken realpath i.e. produce different
code. Bump REVISION to ensure that users get the new version.
only available in the directory of one of their committers who's been
inactive since 2001. It's not licensed such that it can be packaged.
Time to delete it.
OK rsadowski@ jca@
* Fixes for reading Android APK and JAR archives
* Support for non-recursive list and extract
* New tar option: --exclude-vcs
* Important fixes for storing file attributes and flags
* Support for xz, lzma, ppmd8 and bzip2 decompression in ZIP files
* RAR 5.0 reader
solution to fix incompatibility with some downloads from GOG.com made
with Inno Setup version that is incompatible with innoextract 1.7
release. Confirmed to fix extraction of Ultimate Doom, no regression on
testing with about a dozen random GOG.com Inno archives.
Take MAINTAINER while here.
hint to use pl versioning and ok bentley@
ok jca@
Follow the upstream recommendations for packagers and switch to
multi-packages:
devel/gettext -> devel/gettext,-runtime
devel/gettext-tools -> devel/gettext,-tools
(new) devel/gettext,-textstyle
* Support for zstd compression (if zstd is installed).
* The -K option interacts properly with member names given on the command line.
* Fix handling of file shrinkage during read access when --sparse is used.
Bump major of SHARED_LIBS as symbols have been removed. Consumers have
been build tested. Run tested in combination with libarchive and
borgbackup.
OK sthen@
post-install is pretty much a do-install, so simply pass the proper MANDIR
and just uncompress the manual.
The make symlink is only needed for tests/Makefile since $(MAKE) is already
used elsewhere, so turn do-configure into do-test.
OK bket
Quantum-compressed block of exactly 38912 bytes will write one byte
beyond the end of the input buffer", also improves the --fix option.
ok jasper@ and "Steve" (maintainer) "OK'd my update earlier this week
which was the same as your diff below except for the HOMEPAGE change."
if a CAB file has a Quantum-compressed datablock with exactly 38912
compresswed bytes, also rejects blank CHM filenames that are blank
because they have embedded null bytes, and chmextract now protects
against absolute/relative pathnames in CHM files.
some existing COMPILER lines with arch restrictions etc. In the usual
case this is now using "COMPILER = base-clang ports-gcc base-gcc" on
ports with c++ libraries in WANTLIB.
This is basically intended to be a noop on architectures using clang
as the system compiler, but help with other architectures where we
currently have many ports knocked out due to building with an unsuitable
compiler -
- some ports require c++11/newer so the GCC version in base that is used
on these archirtectures is too old.
- some ports have conflicts where an executable is built with one compiler
(e.g. gcc from base) but a library dependency is built with a different
one (e.g. gcc from ports), resulted in mixing incompatible libraries in the
same address space.
devel/gmp is intentionally skipped as it's on the path to building gcc -
the c++ library there is unused in ports (and not built by default upstream)
so intending to disable building gmpcxx in a future commit.
Bump minor of shared lib as symbols have been added. Drop patch as it
has been committed upstream. Switch from GitHub auto-generated tarballs
to MASTER_SITES.
OK sthen@
archivers/zstd (devel/llvm -> devel/cmake -> archivers/libarchive ->
archivers/zstd), so ports-clang should not be used in zstd's COMPILER.
Otherwise, zstd build is blocked if clang is not readily available,
as is the case with bulk builds on gcc architectures.
The dependency archivers/libarchive -> archivers/zstd was added
recently.
OK bket@ sthen@
* Avoid super-linear slowdown on malformed mtree files
* NO_OVERWRITE doesn't change existing directory attributes
* New support for Zstandard read and write filters
... plus unmentioned bug fixes.