This behaviour is also present in the new JS curriculum we've been working on (which is where it was really an issue).

Javascript will not run if there are syntax errors in the code. The browser or JavaScript engine will stop executing the script and display an error message indicating where the syntax error occurred.

We are running the tests in the same eval() function as the executed code. I do not think we can prevent it with the current test runner.

Instead we will probably need to create a separate environment for the test to run in. A JS test runner, perhaps?

If we have two separate environments can they still reference one another?

This is trickier to do for HTML challenges than pure JS ones

Yeah, the pure JS ones should just be a matter of running each test individually and wrapping each evaluation in a try/catch block

To me, if there is an HTML error I think it's sufficient enough to just tell the user something roughly like "this test failed because the HTML is invalid". The catch though is how we can distinguish between an HTML error and a JavaScript error. Especially when the JavaScript referenced is not an external script

is this still open
/assign