I had to read this three times before understanding that you're just using it as an example of a computationally expensive function. The fact that you explicitly state the return value of the function definition, and that you're explicitly attaching it to the global scope, makes no sense without knowledge of doctest. It seemed to me at first as if you're trying to demonstrate something here, whereas in fact you're really just working around the semantics of your linter.

I think it might be better to use something native to JavaScript (or in fact exposed by Sanctuary) to demonstrate. Maybe even something like /*perform expensive calculation*/ 1 + 1. Or if you really feel fib demonstrates it best, it would've helped me to see a comment above fibs definition explaining why it's being defined there.

Ah. I see you've already taken the same approach for documenting maybe_. I suppose multiple instances justify the learning curve.

Perhaps if you wouldn't have to assert the return value of the function definition, it would read more easily, and it wouldn't look like a "stand-alone" example.

Thanks for highlighting this, Aldwin. We are indeed working around a limitation of the tool. I just opened davidchambers/doctest#73. At some point I will address that issue, publish a new version of doctest, and update these fib definitions. This needn't all happen before this pull request is merged, though. :)

Here you go, Aldwin. :)

\o/