* resolves issues with newer CMake
-- Performing Test HAVE_MM_MALLOC
CMake Error: TRY_RUN() invoked in cross-compiling mode, please set the following cache variables appropriately:
HAVE_MM_MALLOC_EXITCODE (advanced)
HAVE_MM_MALLOC_EXITCODE__TRYRUN_OUTPUT (advanced)
For details see pcl/1.8.0-r0/build/TryRunResults.cmake
-- Performing Test HAVE_MM_MALLOC - Success
* should resolve the issues with memory usage during the build:
https://github.com/PointCloudLibrary/pcl/issues/2284#issuecomment-437584571
but it still triggers OOMK with -j8 on 32GB RAM builder, use at least
-j4 to improve the build time a bit
* use git fetcher for sources, github archives can be regenerated
from time to time with different checksums:
https://github.com/bmwcarit/meta-ros/issues/609
* 0001-make-the-pcl-library-compile-with-gcc6.patch is needed only
with thud and older, with warrior it builds OK without this
patch as well
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
* fixes:
ERROR: festival-2.4-r0 do_package_qa: QA Issue: festival: /festival/usr/share/festival/dicts/wsj.wp39.poslexR is owned by uid 1026, which is the same as the user running bitbake. This may be due to host contamination
festival: /festival/usr/share/festival/dicts/wsj.wp39.tri.ngrambin is owned by uid 1026, which is the same as the user running bitbake. This may be due to host contamination
festival: /festival/usr/share/festival/dicts/COPYING.poslex is owned by uid 1026, which is the same as the user running bitbake. This may be due to host contamination
festival: /festival/usr/share/festival/dicts/cmu/cmudict-0.4.diff is owned by uid 1026, which is the same as the user running bitbake. This may be due to host contamination
festival: /festival/usr/share/festival/dicts/cmu/cmu_lts_rules.scm is owned by uid 1026, which is the same as the user running bitbake. This may be due to host contamination
festival: /festival/usr/share/festival/dicts/cmu/cmudict_extensions.scm is owned by uid 1026, which is the same as the user running bitbake. This may be due to host contamination
festival: /festival/usr/share/festival/dicts/cmu/cmulex.scm is owned by uid 1026, which is the same as the user running bitbake. This may be due to host contamination
festival: /festival/usr/share/festival/dicts/cmu/cmudict-0.4.out is owned by uid 1026, which is the same as the user running bitbake. This may be due to host contamination
festival: /festival/usr/share/festival/dicts/cmu/Makefile is owned by uid 1026, which is the same as the user running bitbake. This may be due to host contamination
festival: /festival/usr/share/festival/dicts/cmu/cmudict_compile.scm is owned by uid 1026, which is the same as the user running bitbake. This may be due to host contamination
festival: /festival/usr/share/festival/dicts/cmu/allowables.scm is owned by uid 1026, which is the same as the user running bitbake. This may be due to host contamination
festival: /festival/usr/share/festival/dicts/cmu/cmudict-0.4.scm is owned by uid 1026, which is the same as the user running bitbake. This may be due to host contamination
festival: /festival/usr/share/festival/dicts/cmu/cmu2ft is owned by uid 1026, which is the same as the user running bitbake. This may be due to host contamination
festival: /festival/usr/share/festival/dicts/cmu/COPYING is owned by uid 1026, which is the same as the user running bitbake. This may be due to host contamination
festival: /festival/usr/share/festival/dicts/oald/cuvoald710-0.2.scm is owned by uid 1026, which is the same as the user running bitbake. This may be due to host contamination
festival: /festival/usr/share/festival/dicts/oald/oald-0.4.out is owned by uid 1026, which is the same as the user running bitbake. This may be due to host contamination
festival: /festival/usr/share/festival/dicts/oald/oald2ft.tar.gz is owned by uid 1026, which is the same as the user running bitbake. This may be due to host contamination
festival: /festival/usr/share/festival/dicts/oald/Makefile is owned by uid 1026, which is the same as the user running bitbake. This may be due to host contamination
festival: /festival/usr/share/festival/dicts/oald/oald_extensions.scm is owned by uid 1026, which is the same as the user running bitbake. This may be due to host contamination
festival: /festival/usr/share/festival/dicts/oald/oald_lts_rules.scm is owned by uid 1026, which is the same as the user running bitbake. This may be due to host contamination
festival: /festival/usr/share/festival/dicts/oald/oaldlex.scm is owned by uid 1026, which is the same as the user running bitbake. This may be due to host contamination
festival: /festival/usr/share/festival/dicts/oald/README.oald is owned by uid 1026, which is the same as the user running bitbake. This may be due to host contamination
festival: /festival/usr/share/festival/dicts/oald/COPYING is owned by uid 1026, which is the same as the user running bitbake. This may be due to host contamination
festival: /festival/usr/share/festival/voices/english/rab_diphone/COPYING is owned by uid 1026, which is the same as the user running bitbake. This may be due to host contamination
festival: /festival/usr/share/festival/voices/english/rab_diphone/festvox/rab_diphone.scm is owned by uid 1026, which is the same as the user running bitbake. This may be due to host contamination
festival: /festival/usr/share/festival/voices/english/rab_diphone/group/rablpc16k.group is owned by uid 1026, which is the same as the user running bitbake. This may be due to host contamination
festival: /festival/usr/share/festival/voices/english/kal_diphone/COPYING is owned by uid 1026, which is the same as the user running bitbake. This may be due to host contamination
festival: /festival/usr/share/festival/voices/english/kal_diphone/festvox/kal_diphone.scm is owned by uid 1026, which is the same as the user running bitbake. This may be due to host contamination
festival: /festival/usr/share/festival/voices/english/kal_diphone/festvox/kaldurtreeZ.scm is owned by uid 1026, which is the same as the user running bitbake. This may be due to host contamination
festival: /festival/usr/share/festival/voices/english/kal_diphone/group/kallpc16k.group is owned by uid 1026, which is the same as the user running bitbake. This may be due to host contamination [host-user-contaminated]
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
* using PN here won't work with multilib
* fixes QA warning:
WARNING: festival-2.4-r0 do_package_qa: QA Issue: festival: SRC_URI uses PN not BPN [src-uri-bad]
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
* resolves following error in world builds without everybody needing
to blacklist it manually for builds which don't include
meta-intel-realsense layer, we could use dynamic layers for this, but
as long as this is the only recipe depending on optional layer, we
can just use this DISTRO_FEATURE (like thud branch of meta-ros is
using now).
NOTE: Resolving any missing task queue dependencies
ERROR: Nothing PROVIDES 'librealsense' (but /OE/build/oe-core/meta-ros/recipes-extended/realsense-camera/realsense-camera_1.8.0.bb DEPENDS on or otherwise requires it)
ERROR: Required build target 'meta-world-pkgdata' has no buildable providers.
Missing or unbuildable dependency chain was: ['meta-world-pkgdata', 'realsense-camera', 'librealsense']
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
* fixes:
| checking for bison... no
| configure: error: You need to install bison
| NOTE: The following config.log files may provide further information.
| NOTE: /data/cibot/deleteme/build/tmp-glibc/work/i586-oe-linux/sphinxbase/0.8-r0/build/config.log
* needed since oe-core change:
http://git.openembedded.org/openembedded-core/commit/?id=b7edc20cc2dd82989bd9561f860cb25478a40f69
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
The initial commit from Johannes Schrimpf provided the geographiclib
recipe in recipes-ros. The general policy is to only have ROS
packages in recipes-ros, i.e., packages that are tracked and updated
in the rosdistro/distribution.yaml file.
Hence, I then moved the geographiclib recipe to recipes-extended,
as it is a supporting library, not a ROS package.
Signed-off-by: Lukas Bulwahn <lukas.bulwahn@gmail.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
[lukas.bulwahn@gmail.com: include reference to meta-ros issue tracker]
Signed-off-by: Lukas Bulwahn <lukas.bulwahn@gmail.com>
* fixes:
http://errors.yoctoproject.org/Errors/Details/174716/
qhull/2012.1-r0/qhull-2012.1/src/testqset/testqset.c:217:9: error: format not a string literal and no format arguments [-Werror=format-security]
printf(promptstr);
^~~~~~
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Commit openembedded/openembedded-core@0391fca ("classes/utils: remove
compatibility functions") forces us to update to the proper oe.utils
functions.
Signed-off-by: Lukas Bulwahn <lukas.bulwahn@gmail.com>
Rather than solving #526 in assimp 3.1.1, this commit updates to
the latest assimp version 3.3.1 and applies a backported patch, as
all packages then compile on x86 and mips.
Signed-off-by: Lukas Bulwahn <lukas.bulwahn@gmail.com>
This commit implements Christian Ege's suggestion in the meta-ros
issue tracker to make gcc only warn on ignored template attributes.
Signed-off-by: Lukas Bulwahn <lukas.bulwahn@gmail.com>
Project depending on libeigen's headers can't be compiled with
gcc 7.1 because of the int-in-bool-context warning.
As suggested in
https://listengine.tuxfamily.org/lists.tuxfamily.org/eigen/2017/06/msg00013.html
it's easier to disable the warning for now until libeigen
fixes them.
Signed-off-by: Dmitry Rozhkov <dmitry.rozhkov@linux.intel.com>
This version includes important fixes for those working with point
cloud.
The remaining patch was already merged upstream on intel-ros/realsense
and will probably make through v1.8.1. So it will be possible to drop
it soon.
Signed-off-by: Murilo Belluzzo <murilo.belluzzo@intel.com>
This node always install launch setups that uses rgbd-launch, so it
needs to be listed as a dependency otherwise those configs won't work.
Signed-off-by: Murilo Belluzzo <murilo.belluzzo@intel.com>
As the main repository for the librealsense recipe now contains
fixes required for linking catkin-based packages against it
we can stop maintaining our own version of librealsense recipe.
Also this patch adapts realsense-camera recipe to link against
the non-catkin version of librealsense.
Signed-off-by: Dmitry Rozhkov <dmitry.rozhkov@linux.intel.com>
sphinxbase is linked to libsndfile1, libasound from alsa-lib and
libsamplerate0, but these build dependencies are missing in the
recipe.
This update adds the missing dependencies.
Signed-off-by: Dmitry Rozhkov <dmitry.rozhkov@linux.intel.com>
The fix for build type detection in librealsense's CMakeLists.txt
turned out to be non-functional in the official ROS build farm as
explained in https://github.com/IntelRealSense/librealsense/pull/404
Since this recipe is specific to meta-ros set the build type
explicitly to ROS to guaranty the catkin-related configs for
librealsense get created and installed properly.
Signed-off-by: Dmitry Rozhkov <dmitry.rozhkov@linux.intel.com>
Due to the update to 1.8.0, the upstream-accepted patch is not
anymore needed to be applied.
Signed-off-by: Lukas Bulwahn <lukas.bulwahn@oss.bmw-carit.de>
Compiling libflann 1.8.4 with gcc6 fails with:
```
| /home/lukas/work/repositories/openembedded.org/openembedded-core/build/tmp-glibc/work/i586-oe-linux/libflann/1.8.4-r0/flann-1.8.4-src/src/cpp/flann/algorithms/kdtree_index.h:666:39: error: call of overloaded 'abs(flann::KDTreeIndex<flann::KL_Divergence<double> >::ElementType)' is ambiguous
| ElementType span = abs(point[i]-leaf_point[i]);
| ~~~^~~~~~~~~~~~~~~~~~~~~~~~
```
This issue has been already addressed in the latest version 1.9.1,
hence the commit simply updates the recipe to the version 1.9.1,
which is available on github.
Signed-off-by: Lukas Bulwahn <lukas.bulwahn@oss.bmw-carit.de>
Since 3739aaf644d5a0f28e85fd07c67cecbfa5b8fb8a on openembedded-core,
the build of the bfl recipe fails with:
ERROR: QA Issue: orocos-bfl.pc failed sanity test (tmpdir) in path /[...]/bfl/0.7.0-6-r0/sysroot-destdir//usr/lib/pkgconfig [pkgconfig]
ERROR: QA Issue: bfl.pc failed sanity test (tmpdir) in path /[...]/bfl/0.7.0-6-r0/sysroot-destdir//usr/lib/pkgconfig [pkgconfig]
ERROR: QA staging was broken by the package built above
ERROR: Function failed: do_qa_staging
ERROR: Logfile of failure stored in: /[...]/bfl/0.7.0-6-r0/temp/log.do_populate_sysroot.12063
ERROR: Task 4361 (/[...]/meta-ros/recipes-extended/bfl/bfl_0.7.0-6.bb, do_populate_sysroot) failed with exit code '1'
The openembedded-core recipe suggests that the bfl recipe was broken
since its initial commit, but it remained unnoticed. This commit now
adds a patch to drop the unneeded and wrong paths from the pkgconfig
file.
Signed-off-by: Lukas Bulwahn <lukas.bulwahn@oss.bmw-carit.de>
The update to the latest version 1.7.2 was triggered by compilation
errors in pcl 1.7.0 (i.e., in file common/include/pcl/point_types.h)
due to the update of boost to 1.60.0 in openembedded-core [1].
This failure was detected by the internal CI build on 2016-01-15.
Due to the version update, the commit drops patch
`Fix incompatibility with Boost 1.56`, as the change is included
upstream in version 1.7.2. This commit also adds the upstream-accepted
patch `Add "WITH_OPENGL" option (...)` to make the dependency to OpenGL
libraries optional.
The commit configures package dependencies and compiles with the
least possible packages, so that the pcl-ros recipe compiles. As a
result, all optional packages in PCL, except qhull, can be set to FALSE.
If qhull is deactivated, the do_configure task of pcl-ros reports:
| -- Could NOT find qhull (missing: QHULL_LIBRARY)
| ** WARNING ** surface features related to qhull will be disabled
However, in the do_compile task of pcl-ros, the build process assumes
that the surface features related to qhull are actually never disabled,
i.e., if qhull is deactivated, compiling pcl-ros fails with:
| In file included from [...]/pcl_ros/src/pcl_ros/surface/convex_hull.cpp:40:0:
| [...]/pcl_ros/include/pcl_ros/surface/convex_hull.h:44:37: fatal error: pcl/surface/convex_hull.h: No such file or directory
| compilation terminated.
| CMakeFiles/pcl_ros_surface.dir/build.make:89: recipe for target 'CMakeFiles/pcl_ros_surface.dir/src/pcl_ros/surface/convex_hull.cpp.o' failed
| make[2]: *** [CMakeFiles/pcl_ros_surface.dir/src/pcl_ros/surface/convex_hull.cpp.o] Error 1
| make[2]: Leaving directory '[...]/build'
| CMakeFiles/Makefile2:1173: recipe for target 'CMakeFiles/pcl_ros_surface.dir/all' failed
| make[1]: *** [CMakeFiles/pcl_ros_surface.dir/all] Error 2
[1] http://cgit.openembedded.org/openembedded-core/commit/?id=17d77db5e1203e18cdd178e7d9a457310425ebdc
Signed-off-by: Lukas Bulwahn <lukas.bulwahn@oss.bmw-carit.de>
Martin Jansa