How do we want to handle Whiteboard usage errors? (was ACTSFW-66)
Original author Hadrien Benjamin Grasland @hgraslan.
While migrating algorithms to the new APIs, the acts-test-fw!57 MR also has removed a number of warnings regarding bad Whiteboard writes and introduced new warning-less whiteboard writes in DigitizationAlgorithm, ExtrapolationAlgorithm, ParticleGun and ReadEvgenAlgorithm.
This got me thinking about whether this was the write things to do, and I realized that I don't have a clear answer to this, so I think we should discuss this.
Here's the heart of the issue: the ACTS test framework handles errors using status codes, rather than exceptions. One major implication of this is that debuggers can't catch errors and print backtraces of them, so manual test framework error logging becomes of paramount importance to fast problem solving.
By removing in-Algorithm error reporting, a test framework user does not know in which Algorithm the test framework failed to write into the WhiteBoard. All he knows, from the Whiteboard's warning, is that a certain object with a certain name failed to be written, most likely because two unrelated Algorithms are hitting the same location in the WhiteBoard.
Do we consider this satisfactory, or do we want to be more specific about which Algorithm failed to write data?