It isn't usually appropriate to continue testing after a runtime error has occurred. The student code is probably badly broken and the same error is likely to occur on all subsequent tests too. There's not much point in repeating the same error over and over - the student already has 0 marks, at least with an all-or-nothing grading scheme, and will have to resubmit anyway. Worse still, if it's a timeout error, say because the student code is looping, it can lead to very long run times if there are lots of tests and each one has to run until the timeout limit is reached.
In the example you show, the student has pasted in the test code as well as the function they were asked to write, so will get the same error on every test.
I guess the short answer is: no, testing is always aborted if a runtime error occurs. However, the longer answer is that you can if you want write a template grader for a question or question type; it can catch the runtime error and simply return a 'wrong answer' result (i.e. a mark of 0). Combinator template graders give you total control of the entire grading process and the feedback delivered to the student.