Make test fail if .approved does not exist

My initial thought was to have the test pass if the .approved file wasn't found. I was thinking that this would only occur the first time you run the test and typically for legacy code you would just accept whatever it spits out. This is because you want to use the test as a software vise. You want to refactor and make sure the output doesn't change. So I thought just passing the test and automatically saving that data as the .approved file made sense, since that is what you would do in almost every case anyway. Why make it explicit?

However there is a flaw in this logic. If you push the code to GitLab and forget to add the .approved file, then when the test runs it will pass no matter what on the server and even worse, whatever it spits out will become the new .approved file - that is not what we want. better to make it explicit.