Trace in dev comment

First panic is in (*BatchCursor).clearBatch(...) (inlined)
Second is after we panic there and in a defer try to close the cursor (*BatchCursor).Close at bc.currentBatch.Data = nil
Seems like currentBatch somehow became nil and panicked twice accessing it

I'm thinking there's some condition where (cs *ChangeStream) executeOperation( calls NewBatchCursor but there's no error which puts it in a bad state for the next Next/TryNext.