It's not surprising that a Markov chain could put together plausible sounding class names, because yeah, there are definitely patterns that class names fall into.
What's surprising is that one third of the names it's showing me are actually real. I mean, "MetaMetaContextHierarchyConfig" actually exists?
I got "ConfigClassesAndProfileResolverWithCustomDefaultsMetaConfig.ResolverConfig" as real (which I failed to pick... I mean, really?) and had to doublecheck it really is.
Sadly, it's true. What the hell, Spring Framework?
Don't forget you can't test the ConfigClassesAndProfileResolverWithCustomDefaultsMetaConfig without a set of ConfigClassesAndProfileResolverWithCustomDefaultsMetaConfigTests.
And of course, what about testing the custom overrides of that config?
Well just use the ConfigClassesAndProfileResolverWithCustomDefaultsMetaConfigWithOverridesTests!
What's surprising is that one third of the names it's showing me are actually real. I mean, "MetaMetaContextHierarchyConfig" actually exists?