Set of various fixes

[pierrewillenbrockdfki]

Take 2: fixed the memory leak fix crash, dropped the build system stuff for now.

The build system stuff needs changes elsewhere(base/cmake currently breaking knowledge_reasoning/gqr) to work right.

[pierrewillenbrockdfki]

@skasperski this probably does no longer crash rock-runtime.

[skasperski]

====================================================
                  rock-runtime
----------------------------------------------------

Initializing ROCK...                            [OK]    
Initializing process server local...            [OK]    

Starting process servers:
  process_server_local                          free(): double free detected in tcache 2
Stack trace (most recent call last):
#15   Object "/usr/lib/x86_64-linux-gnu/ld-linux-x86-64.so.2", at 0xffffffffffffffff, in 
#14   Object "/home/dfki.uni-bremen.de/skasperski/Robotics/rock-local/install/bin/rock-runtime", at 0x62c8a3570e54, in _start
#13   Source "../csu/libc-start.c", line 360, in __libc_start_main_impl [0x717d1722a28a]
#12   Source "../sysdeps/nptl/libc_start_call_main.h", line 58, in __libc_start_call_main [0x717d1722a1c9]
#11   Object "/home/dfki.uni-bremen.de/skasperski/Robotics/rock-local/install/bin/rock-runtime", at 0x62c8a356fe93, in main
#10   Object "/home/dfki.uni-bremen.de/skasperski/Robotics/rock-local/install/lib/liborocos_cpp.so", at 0x717d1836c97b, in orocos_cpp::ConfigurationHelper::applyConfToProperty(RTT::TaskContext*, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, libConfig::ConfigValue const&)
#9    Object "/home/dfki.uni-bremen.de/skasperski/Robotics/rock-local/install/lib/liborocos_cpp.so", at 0x717d1836c83d, in orocos_cpp::ConfigurationHelper::applyConfigValueOnDSB(boost::intrusive_ptr<RTT::base::DataSourceBase>, RTT::types::TypeInfo const*, libConfig::ConfigValue const&)
#8    Object "/home/dfki.uni-bremen.de/skasperski/Robotics/rock-local/install/lib/librtt-typelib-gnulinux.so", at 0x717d17e07df2, in orogen_transports::TypelibMarshallerBase::deleteHandle(orogen_transports::TypelibMarshallerBase::Handle*)
#7    Object "/home/dfki.uni-bremen.de/skasperski/Robotics/rock-local/install/lib/orocos/types/libstd-transport-typelib-gnulinux.so", at 0x717d1476d11f, in orogen_transports::TypelibMarshaller<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > >::deleteTypelibSample(orogen_transports::TypelibMarshallerBase::Handle*)
#6    Source "./malloc/malloc.c", line 3398, in __libc_free [0x717d172addad]
#5    Source "./malloc/malloc.c", line 4541, in _int_free [0x717d172ab55e]
#4    Source "./malloc/malloc.c", line 5772, in malloc_printerr [0x717d172a8ff4]
#3    Source "../sysdeps/posix/libc_fatal.c", line 134, in __libc_message_impl [0x717d172297b5]
#2    Source "./stdlib/abort.c", line 79, in abort [0x717d172288fe]
#1    Source "../sysdeps/posix/raise.c", line 26, in raise [0x717d1724527d]
#0  | Source "./nptl/pthread_kill.c", line 89, in __pthread_kill_internal
    | Source "./nptl/pthread_kill.c", line 78, in __pthread_kill_implementation
      Source "./nptl/pthread_kill.c", line 44, in __pthread_kill [0x717d1729eb2c]
Aborted (Signal sent by tkill() 19194 1545606974)
Aborted (core dumped)

[pierrewillenbrockdfki]

The data actually was freed twice. The right thing to do is to have the rtt-typelib deallocate and free it(as was done before), it does call the destructor.

And then i went on a hunt for a leak inside std::strings, first thinking they'd still do copy-on-write(they don't), guessing the deallocation should happen far away. Then wondering why the CNDHandlerConfig's destructor was not destroying its strings(it did), only to find that the temporary string used to push the data into its parent vector was not destroyed.

This leaked one copy of the contents of all containers(typelib speak for strings and vectors) that reside inside another container, at each stage where a container is built.