Rethink aborting sweeps with analysis functions

[t-b]

In #1234 the op uploaded some files. Both files have in the analysis function dashboard something like

And the odd thing here is that the dashboard code thinks that some sweeps are still ongoing because the set QC entries are still missing. Although they are definitly finished.

A bit of digging in the labnotebook revealed that the sweep was just stopped by button press (DAQ stop reason 0x1). And when stopping data acquisition prematurely we don't issue the POST_SWEEP_EVENT/POST_SET_EVENT so the analysis function can not write the expected set QC entry.

I think a few things need clarification:

• Is stopping a sweep during execution via a button press (or something else) to be expected?
• If yes, maybe we should issue the POST_SWEEP_EVENT/POST_SET_EVENT even in that case so that the analysis function can write expected entries? This is not trivial to do as this changes the mental model of the existing analysis functions quite a bit.

[t-b]

For now the dashboard view was improved see 3c106d37 (AD_GetResultMessage: Handle manually stopped sets better, 2021-12-21).

[t-b]

Conclusion after chat with @timjarsky:

• Make all PSQ/MSQ analysis functions fail the sweep/set QC if a sweep was terminated early
• Issue all POST_XXX events even when aborted
• Add tests

[t-b]

Another example is 2022-04-28-issue-ram-BUG-dashbaord-missing-BL-QC on the FTP.