General observation about exceptions:
Geometry is such a vast field that exceptions are inevitable. They are inevitable in any software actually. We are trying to give some response and error report whenever it is possible, but generally we consider DGK as an “open” software in the sense that not cases are handled and not all errors are reported.
There are some mega algorithms like Boolean operations which would need many years to handle all weird cases and as the result it would be much slower
So please consider an exception as an error message. Often it means that the input has problems. Also, there are bugs of course. Please report them. We do not mind fixing bugs and would love to have the most robust software possible.
The problem is in priorities. Which is more important to fix a bug which happens rarely or add some new useful functionality? We often get it wrong.
Your feedback is the only way to get it right.
Nick