USE_KDE4=kdehier component has been deprecated, new components added:
baloo - Baloo core libraries
baloo-widgets - Baloo widgets library
kfilemetadata - KDE library for extracting file metadata
New ports:
graphics/kqtquickcharts - QtQuick plugin to render interactive charts
misc/artikulate - Pronunciation trainer for KDE
(not usable currently, links to both
GStreamer 1.x and 0.10.x via dependencies)
sysutils/baloo[-widgets] - KDE framework for searching and
managing user metadata
sysutils/kfilemetadata - Library for extracting file metadata
l10n ports:
- Farsi (Persian) and Indonesian translations has been readded
- Vietnamese didn't pass threshold for inclusion into release
astro/kstars:
- switch dependency from math/eigen2 to math/eigen3
- add PYKDE option for updating supernovae data
deskutils/kdepim4:
- update dependencies: add libkgapi and baloo,
remove now needless clucene, link-grammar, strigi
- add patch to fix build with gcc42
- update COMMENT and description for all KDE PIM ports
devel/ruby-krossruby:
- remove BROKEN, it builds with ruby 2.x now
editors/kate:
- add patch to disable memory-hungry build of the kate tests [1]
graphics/okular:
- add dependency on graphics/libkscreen
math/cantor:
- add optional dependency on lang/luajit for LuaJIT backend
- fix gfortran detection [2]
misc/kdehier4:
- adapt to new pkg world. Now the purpose of kdehier4 only
to link some stuff between KDE4_PREFIX and LOCALBASE.
science/kalzium:
- switch dependency from math/eigen2 to math/eigen3
- add dependence on science/chemical-mime-data
x11-themes/kdeartwork4
- switch dependency from math/eigen2 to math/eigen3
among other changes:
- drop deprecated USE_KDE4=kdehier
- drop @dirrm from plist
- clean up pkg-descr
- convert to options helpers
- other portlint fixes
The area51 repository features commits by alonso, rakuco and myself.
PR: 187150 [1]
Reported by: pe.freethread@live.com
Patch by: Tobias Berner <tcberner@gmail.com>
PR: 180674 [2]
Reported by: torsten.eichstaedt@web.de
PR: 194316
Exp-run: antoine
On FreeBSD 11, the wpa_supplicant couldn't find the readline headers.
It turns out that Mk/Uses/readline.mk only augments LDFLAGS and CPPFLAGS,
but wpa_supplicant needs -I${LOCALBASE}/include add to CFLAGS. To fix
FreeBSD build, augment CFLAGS with value of CPPFLAGS in the makefile.
This wasn't an issue on DragonFly. Apparently wpa_supplicant found the
readline compatibility headers of libedit and used those instead.
reported by: pkg-fallout
is being removed. No portrev bump needed since the kernel will continue to
accept this flag for old binaries even after support for it in new binaries
is removed.
PR: 193961
Approved by: portmgr (antoine)
1.11 remains a maintenance release.
- Update security/krb5 1.12.2 --> 1.13
- Copy the old security/krb5 1.12.2 to security/krb5-112
(now a maintenance release supported by MIT)
- Move the old krb5-maint (1.11.5: old maintenance release) to
security/krb5-111 (the old maintenance release still supported by MIT)
- Update Firefox ESR to 31.2.0
- Update NSS to 3.17.2
- Update Thunderbird to 31.2.0
- Update libxul to 31.2.0 (and mark as BROKEN)
- Disable SSL 3.0 with pref (Upstream bug 1076983)
- (workaround) replace USE_GCC=yes with USES=compiler:gcc-c++11-lib in
order to fix runtime for PGO and powerpc/powerpc64 on libc++ systems
- Add OSS audio fallback for HTML5 audio from upstream bug;
not exposed yet because WebRTC still needs ALSA or PulseAudio
- Kill @dirrm from gecko@ ports per CHANGES from 20140922
- Drop workaround for LLVM PR 19007: base and lang/clang34 have the fix
- Improve workaround comment for LLVM PR 15840, partially rejecting
r348851 by marino@ until bug 193555
PR: 194356
Submitted by: Jan Beich
Security: http://www.vuxml.org/freebsd/9c1495ac-8d8c-4789-a0f3-8ca6b476619c.html
DragonFly adopted it's "new" IEEE 802.11 infrastructure from FreeBSD.
This introduced an additional isr_meshid_len field in the
ieee80211req_scan_result structure. It is necessary to include this
additional offset when calculating the address of the IE data buffer.
Imre Vadasz introduced this fix to DragonFly on 01 Sept 2014, but a
similar fix doesn't appear to be present in FreeBSD's base wpa_supplicant.
It turns out the driver list was only showing "null" for NDIS when -h
(help) or an unknown switch was passed to wpa_supplicant. The cause is
that the NDIS driver is the only that that has to be initialized, and
the initialization wasn't occurring in the above case.
The fix is the move the NDIS driver initialization before the command
switches are examined in case wpa_supplicant aborts to usage early.
The NDIS driver didn't build when privsep option was turned off. Moveover,
it didn't display the driver name correctly, so I figured out a patch
based on base's ndis driver code. Thirdly, when privsep option is turned
off, wpa_priv is not built, so let's handle that condition.
Finally, remove privsep from default options because it removes the
standard list of drivers and replaces it with privsep driver.
This port was retired at version 0.3.8 because wpa_supplicant is
part of FreeBSD base. However, the last few releases have had a period
of only a few months, so the base is always going to be behind. DragonFly
is also affected, so I'm bringing the port back at the latest version.
It features the same patches as FreeBSD including the conversion to use
libutil's pidfile routines. There are some additional patches for
DragonFly support and to fix some bugs from the 9 Oct 2014 release.
The WPA Supplicant build system has been converted to ports options, and
there are dozens of them. I've set the defaults to match the
configuration in base and verified that it builds with all options
selected at once.
It seems that USES=libtool doesn't rewrite ${LIBTOOL} in
a configure script to use ${LOCALBASE}/bin/libtool like
bsd.autotools.mk did.
So, rewrite it explicitly for workaround.
PR: 194176
Submitted by: jhujhiti__at__adjectivism.org
The Rekall Framework is a completely open collection of tools, implemented in
Python under the GNU General Public License, for the extraction of digital
artifacts from volatile memory (RAM) samples. The extraction techniques are
performed completely independent of the system being investigated but offer
visibility into the runtime state of the system. The framework is intended to
introduce people to the techniques and complexities associated with extracting
digital artifacts from volatile memory samples and provide a platform for
further work into this exciting area of research.
WWW: http://www.rekall-forensic.com/
- Patch libtool so it uses the same library version specification as on
Darwin, Linux and other systems. Given the version current:revision:age
a library will be given the extension .so.major.age.revision with major
equal to current-age. Before libtool would use .so.current on FreeBSD.
- Patch libtoolize to remove two cases of umask 0 that caused libltdl
files to be copied world writable (--ltdl option)
- Let USES=libtool patch this new version correctly
- Adjust all ports with USES=libtool:build and bump PORTREVISION on their
dependent ports if a library version changed
PR: 194068
Exp-run by: antoine
Approved by: portmgr (antoine)
These will be removed on January 1 2015.
Really ports should not be touching the base system at all.
This option is a big foot-shoot problem:
1. Recent versions of FreeBSD such as 9.3, 10.0, 10.1+, now remove all ssh
files from /usr if you 'make delete-old' with WITHOUT_SSH. This results in
removing the overwrite base files.
2. Uninstalling the package leaves the system with no ssh.
3. Running installworld without WITHOUT_SSH results in overwriting the
package, or giving false-positive 'pkg check -s' errors.
4. The port fails to pass QA checks because it removes system files.
and did not match the documentation. Users should tune to their needs
instead of relying on the defaults, but if they are this will be an
improvement.
Submitted by: John Vinopal
The existing patch removed "<malloc.h>" but didn't provide "<stdlib.h>".
Adjust to compound patch to fix that oversight.
FWIW, the extra work to partially fix a patch is one reason I don't like
one-patch-many-files. It's not the only reason.
for verifying whether a certificate is valid for the intended
purposes.
In the simplest case, this means host name verification.
However, service_identity implements RFC 6125 fully and
plans to add other relevant RFCs too.
WWW: https://github.com/pyca/service_identity
PR: 193930
Submitted by: Axel Rau <axel.rau@chaos1.de>
