Proper implementation of move constructors and move assignment operators for all module types (including tests for the assignment operator)

fixed connectTo() so it now uses the dynamic/virtual model and thus obeys e.g. eliminated hierarchies etc.

fix running into an assertion sometimes: getQualitifedName() should also be possible if the highest hierarchy is a Module (e.g. a VirtualModule)

Make Module::connectTo() no longer a template, since a virtual function can anyway not be overridden with a different return type... (c.f. the comment in the deleted code)

- removed the no-longer-necessary optional argument of findTag() - use flatten() instead

refactor ArrayAccessor and ScalarAccessor: split off common code and put into common base class InversionOfControlAccessor

Fixed some more problems with the move sematics of EntityOwner/Module. Re-registration at the owner was missing and some things were done in the wrong place. Also the move assignment is now back again.

put the move constructor and move assignment-operator into the EntitiOwner base-class and inherit them

introduced a ModuleGroup (grouping ApplicationModules and other ModuleGroups) and enforced a proper hierarchy of modules to exclude that an ApplicationModule can own (directly or indirectly) another ApplicationModule

update for interface change in DeviceAccess: TransferElement::write() has now the version number as an argument

allow determining a fully qualified name for application variables (i.e. including qualified name of owning module)

made Module::connectTo() compile-time polymorphic, to allow special modules with operator() returning a different type.

Fixed several bugs affecting poll-type variables being read by a poll-type application variable and distributed to push-type variables via a ConsumingFanOut.

Changed Module::readAll() to read all push-type variables first, to make sure that poll-type variables are up-to-date upon return

This reverts commit 7287c4b4.

revert last commit, cascading ApplicationModules should be avoided instead (and should be prevented in future)

Cache accessors obtained through TestFacility::getScalar() and TestFacility::getArray() by the application test code, to avoid the loss of the current value if a decorated accessor is obtained again.

also make sure for DeviceModules to always return the same sub-module with the [] operator when passing the same sub-module name

fixed not returning the same submodule on a second call of the ControlSystemModule::operator[] with the same module name

don't crash the application if Profiler::startMeasurement() is called while the measurement is already running, instead just ignore the second call