Adapt to changes in ChimeraTK/DeviceAccess#116

[killenb]

Child of ChimeraTK/ApplicationCore#123 Depends on ChimeraTK/DeviceAccess#116

Adapt to the changes in DeviceAccess:

• ChimeraTK/DeviceAccess#116 introduce API-breaking changes to postRead/Write. Add the new argument to the function overloads.

• Check whether the arguments should be used. For instance

  • don't generate a new version number in postRead if readTransferLatest did not get new data
  • when reading and the preRead() does not swap the user buffer, don't swap in postRead() becase the user buffer has to stay intact => no action if no new data
  • when the user buffer had been swapped in preWrite(), it always has to be swapped back in postWrite() => always action, even if dataLost is true

• [x] Extend the API so the code compiles again.

• [x] Check if you have to adapt functionality. No in this ticket where changes were made.

• Check that all tests are still OK.

  • [x] All existing tests should still pass.
  • [x] If you had to adapt the functionality check that there is a test for that behaviour

[ckampm]

Implementations to be adapted in this ticket:

• [x] BidirectionalProcessArray<T>::doPostRead: Version number update and user buffer swap
• [x] BidirectionalProcessArray<T>::doReadTransferAsync: Set hasNewData in postRead call
• [x] TypeChangingDecorator::doPostRead: User buffer swap in convertAndCopyFromImpl
• [x] testTypeChangingDecorator.cpp: Pass new parameter to postReadWrite in the testDecorator function (to be fixed in #27)
• [x] UnidirectionalProcessArray<T>::doPostRead: Version number update and user buffer swap

[killenb]

Please also check for side effects of ChimeraTK/DeviceAccess#117:

• [x] If a decorator delegates the postAction and does something after it in the override of doPostXxx(), it might need a try-catch block, to the stuff and re-throw

[killenb]

I removed the urgent flag. The urgent part was to make it compile again.