Checked all CMakeList.txt file but it puzzles me what is wrong.

Task #7336: Allow gen_LofarPackageList_cmake.sh to be run from any checkout, not just those named LOFAR

Task #7342: Added FindQPID support to CMake, and updated lhn001,2 variant files to include QPID install locations

Task #6526: Exported LOFAR_Scheduler from the ROD/trunk (r11938) and imported it into the LOFAR repository in the directory SAS/Scheduler. Minimal tweaks to integrate it into the LOFAR build environment.

Task #4936: remove FindLibSSH2 CMake file. It's no longer needed, since we implemented another way to start Cobalt outputProc months ago.

Task #6734: Added option USE_PYTHON_COMPILATION. It controls whether python byte-compilation will or won't be done during install. It defaults to ON.

Task #1254: Updated FindPVSS.cmake to also work with PVSS version 3.12. Needed to rewrite the logic to determine the version string from the VersInfo.mk file.

Task #1254: No longer need to use gcc-4.6 as CUDA host compiler after upgrading dop282 to Linux Mint 17 (Ubuntu 14.04 LTS).

Task #5921: Just ignore the target_link_libraries() command with LINK_INTERFACE_LIBRARIES argument completely, when using CMake >= 2.8.12. It will possibly lead to too many libraries on the link line, but modern linkers will not link in unused libraries anyway, so we're quite safe.

Task #5921: Reverted all changes on this branch. I'm going to solve this issue in a more pragmatic and less intrusive way.

Task #5921: Undid removal of argument ${${PACKAGE_NAME}_LINK_LIBRARIES} in call to target_link_libraries() in macro lofar_add_executable(). It *IS* needed when a package does not produce a library, like CEP/Calibration/BBSTools. Found out the hard way :(

Task #5921: Fixed target_link_libraries() call in macro lofar_add_executable() for CMake 2.8.12; removed unneeded linking to _LINK_LIBRARIES} in lofar_add_executable(), lofar_add_library() already takes care of this.

Task #5921: lofar_add_library() macro now works both with CMake 2.8.12 and earlier versions without policy warning. Also removed old 'if(NOT APPLE) .. endif()' code, which is probably no longer needed.

Task #5921: This fix seems to work for cmake 2.8.12.2, for both BUILD_SHARED_LIBS is OFF and ON. Of course this fix is not backward compatible, so I need to work on that.

Task #5781: Attempted reintegration of task branch into trunk (error has been reported: svn: E155023: Can't set properties on'/data1/schoenma/LOFAR_trunk/src/LOFAR/RTCP/Interface': invalid status for updating properties.

Task #5540: Apparently, the compile option -D_GLIBCXX_DEBUG causes all kinds of hard to track problems with third-party packages, causing random segmentation faults in, e.g., casacore and pqxx. So I reverted this change.

Task #5708: Initial implementation: move FinalMetaDataGatherer to Cobalt/BrokenAntennaInfo, and let GPUProc call its functions directly

Task #1254: Added timeout script for those (older) distros that don't have it in their coreutils package

Task #1254: Applied the same patch I applied to variants.lhn001 (r29075) to variants.lce072. It suffers from the same issue.

Task #1254: Override the now default C++ debug compile flag -D_GLIBCXX_DEBUG (issue #5540) for lhn001, because Python (or Boost.Python) barfs on it with a '*** glibc detected *** python: free(): invalid pointer'

Task #5632: move tManyPartTABOutput, tMultiPartTABOutput to Cobalt/CobaltTest/. Patch mostly by Marcel.

Task #5467: Added fix to de-duplicate the list of include directories. This is needed because cmake 2.8.8 doesn't do this anymore, and it causes all kinds of irritating problems ranging from 'argument list too long' to cmake eating up all available RAM memory.