This fixes together with another change in DeviceAccess the LNM
MathPlugin with push-type variables as parameters.

Introduce StatusWithMessage and StatusWithMessageReader which handle this.
It is used in DeviceModule error reporting.
This is a preparation in order that we can report errors consistently to DOOCS, in DoocsAdapter

make StatusAggregator use StatusWithMessage

it should output always status code + string
on the input side, connects to either status or status + string
corresponding string input is found automatically
NOTE: DeviceModule API change, the status output variables are
renamed from (status, message) to (status, status_message)!

update testing of DeviceModule outputs

because of API change of DeviceModule  (renamed message ->
status_message)

fix crash and improve message and test

- fix crash due to incorrect handling of cast to Derived
- improve message output
- make existing tests pass
- add test for status message of StatusAggregator

move StatusWithMessageReaderBase and Status def to ControlSystemAdapter

since it is required from ApplicationCore as well as from the adapters

Slashes are not allowed in module names, but the DeviceModule crates a status module using the CDD (or alias) as a name. The slashes are now replaced by underscores.

See #7425

The previous implementation using the mutex had two issues:

1. The mutex was locked in a different thread than unlocked (DeviceModule constructor vs. handleException() thread). This results in undefined behaviour.
2. The boost::shared_mutex::lock_shared() function is not an interruption point, so applications did not terminate while waiting for initial values. (boost::latch::wait() is an interruption point.)

- testCircularDependencyFaultyFlags now uses modern AC without defining CS module and overriding defineConnections()

# Please enter the commit message for your changes. Lines starting
# with '#' will be ignored, and an empty message aborts the commit.
#
# On branch master
# Your branch is up to date with 'origin/master'.
#
# Changes to be committed:
#	modified:   include/DeviceModule.h
#	modified:   src/DeviceModule.cc
#	modified:   tests/executables_src/testCircularDependencyFaultyFlags.cc
#
# Untracked files:
#	build-example-Desktop-Debug/
#	build-example2-Desktop-Debug/
#	build-example_status_monitor-Desktop-Debug/
#	demoBidirectionalVariables/
#	example2b/
#	include/ApplicationModule_BACKUP_1938963.h
#	include/ApplicationModule_BASE_1938963.h
#	include/ApplicationModule_LOCAL_1938963.h
#	include/ApplicationModule_REMOTE_1938963.h
#	message.txt
#	tests/testVirtualHierarchy.xlmap
#

- Implemented but not tested yet
- All existing tests passing

This might have led to test stalled exceptions despite the test did not yet even start properly.

This greatly helps debugging real setups, as the messages will usually end up in the server log.
Note: Ideally, a proper logging mechanism should be used instead, but this does not yet exist!

	calling device.getOneDRegisterAccessor<double> causses an error in server,
	when having uder type string

fix tsan complaining about incrementing/decrementing testableMode_deviceInitialisationCounter outside testable mode

should not make a real difference, it is just more logical like this

[wip #173] move clearing error status flag in control system after the point where operations are allowed again (fixes race condition)

A corresponding update of the specification will follow.

- Initial values are sometimetimes only propagated in the first run of the main loops, not in prepare. If other modules are waiting for it they cannot enter their main loop. That's why all threads have to be started first, and only the it is waited for all of them to enter testable mode.

testProcessVariableRecovery: waiting for device to be opened avoids race conditin with runtime error

- recoveryLock now protects the whole critical section
- activateAsyncRead() is called on the device

- removed DeviceModule::waitForRecovery()
- ExceptionHandlingDecorator is throwing to skip transfers