Golden File Testing

Golden File Testing is also known as Snapshot Testing. There’s no Wikipedia article under either name, so here’s a summary for anyone who might not know this neat testing strategy.

An ordinary test case takes the form test_assert(actual, expected), and if actual != expected then it fails with a message giving the value for actual and for expected and saying that they’re not equal.

A snapshot test takes the form test_snapshot(actual). The first time you run it, it succeeds and records the result (typically in a file). Every subsequent time you run it, it compares actual to the recorded result and fails if they’re different, showing the diff. You can then either (i) realize this diff is bad and go fix your code, or (ii) explicitly tell the test framework that the new output is good and should replace the recorded result.

There are a lot of details that can vary. Like, if the snapshot is of text then there’s a question of how the text diff is shown. Or if it’s of an image, how do you show that? Side by side comparison? But that’s the basic idea.

I’m sad it doesn’t have a Wikipedia article. It’s a little niche (you should almost always use an ordinary test if it’s feasible) but snapshot testing is really nice when you need it.