This looks correct to me and nothing obvious is jumping out. I remixed it and tried to run it myself, getting the same failed test. Below is a console log I get (chrome) … which I don’t see rerunning my project.
I believe this is an issue with the test and not my code.
I have found the issue. The test is looking for a different response for invalid dates. The boilerplate project has a different user story from the curse-arrow project.
Point 5 in boilerplate has return {“unix” : null, “utc” : “Invalid Date” } but the curse-arrow has {“error”: “Invalid Date”}. The test is expecting the latter response. Once I fixed this it started passing.
Probably easiest for FCC to fix the user story on the boilerplate project.