The build_detect_platform-Check-for-__SSE4_2__.patch tries to use fgrep
with pipe to get the SSE flag. But actually it would fail with error:
fgrep: /tmp/leveldb_build_detect_platform-cxx.25886: No such file or
directory
That means it can not get the correct SSE flag.
We should make sure the fgrep command runs after g++ command has
finished and redirect the stdout and stderr to /dev/null to avoid make
error:
Makefile:18: *** missing separator. Stop.
Signed-off-by: Yi Zhao <yi.zhao@windriver.com>
Signed-off-by: Armin Kuster <akuster808@gmail.com>
Use libhwloc instead of lib${PN} to fix problem in case of multilib.
Otherwise, when installing lib32-hwloc into image, we would have the
following error at do_rootfs.
- nothing provides lib32-liblib32-hwloc = 1.11.5-r0 needed by lib32-hwloc-1.11.5-r0.corei7_32
Also, fix FILES variable for libhwloc to not contain .so link file.
This would fix the QA issue, and the INSANE_SKIP line is no longer needed.
Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Armin Kuster <akuster808@gmail.com>
Add a new recipe libyui-ncurses, which is a rdepend for dnfdragora.
Signed-off-by: Zheng Ruoqin <zhengrq.fnst@cn.fujitsu.com>
fixed patch to apply
Signed-off-by: Armin Kuster <akuster808@gmail.com>
They are installed properly by plymouth itself. With buildhistory no change
in packed files could be detected.
Signed-off-by: Andreas Müller <schnitzeltony@gmail.com>
Signed-off-by: Armin Kuster <akuster808@gmail.com>
In pre recipe specific sysroot times AC_PATH_PROG detected host's
systemd-tty-ask-password-agent. In my case that path was wrong and
systemd-ask-password-plymouth.service failed silently.
Now with recipe specific sysroot and no link in PATH to buildhost,
SYSTEMD_ASK_PASSWORD_AGENT remains empty and that causes:
| systemd[1]: /lib/systemd/system/systemd-ask-password-plymouth.service:12: Executable path is not absolute, ignoring: --watch --plymouth
| systemd[1]: systemd-ask-password-plymouth.service: Service lacks both ExecStart= and ExecStop= setting. Refusing.
For both environments this patch fixes log-off from X-session to
display-manager. Before this patch screen went dark and it seemed that X-server
was restarted.
Signed-off-by: Andreas Müller <schnitzeltony@gmail.com>
Signed-off-by: Armin Kuster <akuster808@gmail.com>
This avoids unwanted linking against static libraries as happened for yoshimi.
Signed-off-by: Andreas Müller <schnitzeltony@gmail.com>
Signed-off-by: Armin Kuster <akuster808@gmail.com>
A project could need to have LUA in SDK, an example is
to compile LUA scripts which checks its syntax before embedded
them into a project compiled by Yocto.
Change-Id: I3c36acd58d45121c3344b5322244bdc0b12061b9
Signed-off-by: Romain Forlot <romain.forlot@iot.bzh>
Signed-off-by: Armin Kuster <akuster808@gmail.com>
Since glib-2.0 updated to 2.54.2, build glade 3.20.0 failed
...
|../../glade-3.20.0/gladeui/glade-signal-model.c:68:1:
error: conflicting types for 'g_ptr_array_find'
| g_ptr_array_find (GPtrArray *array, gpointer data)
...
The following commit in the new version fixed the issue.
...
commit cb87c89745c93969c3742e7a8a11d546c00107c8
Author: Juan Pablo Ugarte <juanpablougarte@gmail.com>
Date: Fri Aug 11 18:17:18 2017 -0300
GladeSignalModel: rename g_ptr_array_find() to prt_array_find()
GLib 2.54 introduced g_ptr_array_find() which clashes with our own
function
gladeui/glade-signal-model.c | 8 ++++----
1 file changed, 4 insertions(+), 4 deletions(-)
...
Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Armin Kuster <akuster808@gmail.com>
Some basic function such as password change invoked
by programs such as passwd depends on pam_passwdqc.so
which is provided by subpackage pam-passwdqc.
Add pam-passwdqc to RDEPENDS for passwdqc.
Signed-off-by: Mingli Yu <Mingli.Yu@windriver.com>
Signed-off-by: Armin Kuster <akuster808@gmail.com>
While building openldap, there was a warning, like below:
WARNING: QA Issue: openldap: Files/directories were installed but not shipped in any package:
/usr/libexec
Please set FILES such that these items are packaged. Alternatively if they are unneeded, avoid installing them or delete them within do_install.
openldap: 1 installed and not shipped files. [installed-vs-shipped]
This is because that after openldap all contents of /usr/libexec are split
into subsets or moved into other directory and no one makes use of it.
So, to fix this QA warning, delete /usr/libexec.
Signed-off-by: Qi Hou <qi.hou@windriver.com>
Signed-off-by: Armin Kuster <akuster808@gmail.com>
It is still required by libmusicbrainz, but is no longer needed
by oe-core. Switch the upstream to debian as webdav.org is down.
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Armin Kuster <akuster808@gmail.com>
Without libraries installed dev-package is useless.
Tested by buildung an image with fftw-dev included.
Signed-off-by: Andreas Müller <schnitzeltony@gmail.com>
Signed-off-by: Armin Kuster <akuster808@gmail.com>
We need to install to the sstate input dir and not the output dir, so
sstate handling will work correctly.
Signed-off-by: Stefan Müller-Klieser <s.mueller-klieser@phytec.de>
Signed-off-by: Armin Kuster <akuster808@gmail.com>
Add PACKAGECONFIG for autoconf-archive. Default to enable it to avoid
m4 file conflicts between gnome-common and autoconf-archive.
e.g.
IMAGE_INSTALL_append = " gnome-common autoconf-archive-dev"
bitbake <image>
We enable autoconf-archive PACKAGECONFIG to avoid the following error.
Error: Transaction check error:
file /usr/share/aclocal/ax_check_enable_debug.m4 conflicts between attempted installs of gnome-common-3.18.0-r0.noarch and autoconf-archive-dev-2016.09.16-r0.core2_64
file /usr/share/aclocal/ax_code_coverage.m4 conflicts between attempted installs of gnome-common-3.18.0-r0.noarch and autoconf-archive-dev-2016.09.16-r0.core2_64
Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Armin Kuster <akuster808@gmail.com>
The closest thing to a license is found in the http://metadata.ftp-master.debian.org/changelogs/main/d/debootstrap/debootstrap_1.0.92_copyright
The verbage looks very close to MIT so seems reasonable make to change the license to MIT instead of
creating a custom entry for debootstrap-custom-license. This follows what Redhat, Mandriva and Suse have for this package. (see rpmfind)
the difference between the MIT and the copyright files are:
-This package was debianized by Anthony Towns <ajt@debian.org> on
-Tue, 30 Jan 2001 10:54:45 +1000.
-
-It was written from scratch for Debian by Anthony Towns <ajt@debian.org>
-based loosely on the code for constructing base tarballs as part of the
-boot-floppies package.
-
-Copyright:
-
-Copyright (c) 2001-2005 Anthony Towns
+Copyright <YEAR> <COPYRIGHT HOLDER>
will fix: WARNING: debootstrap-1.0.67-r0 do_populate_lic: debootstrap: No generic license file exists for: debootstrap-custom-license in any provider
Signed-off-by: Armin Kuster <akuster808@gmail.com>
cdrkit only install genisoimage by default,but most tools use
mkisofs to generate an ISO9660/JOLIET/HFS hybrid filesystem,
so add a link to genisoimage with the name mkisofs to support
these tools.
Signed-off-by: Jibin Xu <jibin.xu@windriver.com>
Signed-off-by: Armin Kuster <akuster808@gmail.com>
If opencv was imported in python or dynamic linked in gstreamer it was
producing this output and dying:
[libprotobuf ERROR google/protobuf/descriptor_database.cc:58] File already
exists in database: caffe.proto
[libprotobuf FATAL google/protobuf/descriptor.cc:1315] CHECK failed:
generated_database_->Add(encoded_file_descriptor, size):
terminate called after throwing an instance of
'google::protobuf::FatalException'
This patch backports a patch that fixes this issue.
Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com>
Signed-off-by: Armin Kuster <akuster808@gmail.com>
This service allows gpsd to automatically recognize USB devices but is
currently missing from our packaging. Without this service and the
matching udev rules from upstream, gpsd will fail to recognize USB
devices when they are plugged in.
Signed-off-by: Martin Kelly <mkelly@xevo.com>
Signed-off-by: Armin Kuster <akuster808@gmail.com>
We have our own versions of the udev, hotplug, and systemd scripts,
but at this point they are several years old and were added to fix bugs
at the time. As far as I can tell, the relevant bugs are now fixed
upstream, so we can move back to the upstream versions.
Signed-off-by: Martin Kelly <mkelly@xevo.com>
Signed-off-by: Armin Kuster <akuster808@gmail.com>
1.8 is a minor update from 1.7 and is planned to be the last version before
2.0.
Signed-off-by: Pascal Bach <pascal.bach@siemens.com>
Signed-off-by: Armin Kuster <akuster808@gmail.com>
Set PREFIX via EXTRA_OEMAKE instead of exporting it in do_install_prepend():
than ensuring the correct prefix gets into hiredis.pc, which is generated by
upstream makefile at do_compile() time.
Signed-off-by: Andrea Galbusera <gizero@gmail.com>
Signed-off-by: Armin Kuster <akuster808@gmail.com>
This service is essential for making hotplugging work correctly on a
systemd system. Without it, USB devices get recognized by systemd
instead of the usb-modeswitch udev rule, so hotplugging breaks. Once the
service is added, hotplugging works properly again.
Signed-off-by: Martin Kelly <mkelly@xevo.com>
Signed-off-by: Armin Kuster <akuster808@gmail.com>
consolekit depends on virtual/libx11 then it requires x11
distro feature, so add consolekit option only when x11
is in DISTRO_FEATURES.
Signed-off-by: Jackie Huang <jackie.huang@windriver.com>
Signed-off-by: Armin Kuster <akuster808@gmail.com>
The ALTERNATIVES logic for the /etc/default/gpsd file is broken,
specifying ALTERNATIVE_LINK without brackets. In addition, it refers to
ALTERNATIVE_PATH, which update-alternatives.bbclass doesn't use.
With this patch, /etc/default/gpsd correctly gets a default link
pointing to /etc/default/gpsd.default. Currently, no such link gets
created, and gpsd starts off disabled by default since it's missing this
config file.
Signed-off-by: Martin Kelly <mkelly@xevo.com>
Signed-off-by: Armin Kuster <akuster808@gmail.com>
1) Upgrade freerdp from 1.2.5 to 2.0.0.
2) Delete two patches for they are included in 2.0.0.
0001-FindGStreamer_1_0-fix-build-failure-for-new-gstreame.patch
0003-add-missing-define.patch
3) Add one patch 0001-Fix-gstreamer-1.0-detection.patch to fix the do_configure problem:
| -- Finding required feature XRandR for X11 randr (X11 randr extension)
| -- Found XRANDR: /yocto/work001/fnst/zrq/community/build_master/tmp/work/i586-poky-linux/freerdp/2.0.0+gitrAUTOINC+1648deb435-r0/recipe-sysroot/usr/lib/libXrandr.so
| CMake Error at channels/tsmf/client/gstreamer/CMakeLists.txt:21 (message):
| GStreamer library not found, but required for TSMF module.
4) Modify the do_install directory from winpr/tools/makecert/cli/ to winpr/tools/makecert-cli/
Signed-off-by: Zheng Ruoqin <zhengrq.fnst@cn.fujitsu.com>
Signed-off-by: Armin Kuster <akuster808@gmail.com>
libc-bench is a set of time- and memory-efficiency tests to compare
implementations of various C/POSIX standard library functions.
Coverage includes:
- Malloc throughput with varying degrees and forms of contention between
threads.
- Malloc overhead and ability to return freed memory to the OS.
- Literal string and regular expression searches.
- Thread creation and joining throughput.
- UTF-8 decoding.
- Stdio buffered reads and writes.
Signed-off-by: Ming Liu <liu.ming50@gmail.com>
Signed-off-by: Armin Kuster <akuster808@gmail.com>
Per issue MDEV-14044 on https://jira.mariadb.org/browse/MDEV-14044, mysqld_safe_helper
needs to be installed with the server, but was absent from the mariadb.inc file feeding
the recipes. This resulted in a broken mariadb installation. Installing
mysqld_safe_helper alongside mysqld_safe resolves this problem.
Signed-off-by: Jonathan Haws <jhaws@sdl.usu.edu>
Signed-off-by: Armin Kuster <akuster808@gmail.com>
