Lock is only released for accessors with wait_for_new_data on blocking
reads. This fixes usage of readLatest() in testable mode.

This reverts commit b7fb1f68.

fix: TestableModeAccessorDecorator::preRead() always released the lock and did not delegate the call to the target

This partly reverts d5ab1447, where the existing doReadTransfer overrides
were removed. This commit gets the existing implmentations back, but
changes the function name to doReadTransferSynchronously.

Remove functions that do not exist anymore

Fixes #151

Replaces calls to doXxxTransferYyy of the TransferElement by the
corresponding xxxTransferYyy call

- treat the two directions of bidirectional variables properly in testable mode (i.e. as individual variables with each one counter)
- fix a data race in testable mode in case a bidirectional variable rejected a value (was not holding a required lock)

bidirectional variables: pass on version numbers in modules from the last blocking read on to any write operations, to resolve conflicts of concurrent variable updates properly

fixed testable mode for bidirectional variables: the return channel of control system to application variables must not increment the testableMode_counter

necessary changes for the latest DeviceAccess update (readAsync() return type no longer a reference, and other related changes)

fixed bug that using poll-type transfers in testable mode could sometimes lead to stalled tests or assert failures.

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

- use this return type in testable mode to detect when a queue overflow happened (testable mode counter should not increase in that case)

use the new readLatest(), which has been added to the latest DeviceAccess version, to implement the naturally expected behaviour when using a push-type variable with a poll-type accessor