Turns out that a point initialization run has a pretty robust time range check, but the fetching does not. The getTimeList()/generateSingleTimeObject() functions also do not, or if they do, it needs to be made more straightforward.
The error checking within getTimeList()/generateSingleTimeObject() also need to be revised, to better match what is required for the GUI ninjaCom error handling. The current method is to use CPLDebug() and return values, causing the error messages to get dropped by the GUI.