1. 16
  1.  

  2. 4

    Testing small spaces is a great idea. It’s especially useful for string algorithms, since using a small alphabet finds tricky cases with repeating letters. I think the same concept would apply to any algorithm that operates on lists of non-unique elements.

    The post mentions the worry that you’ll lose out on bugs that only happen with a big search space. I thought about that, and my solution was just to do both. In a few places, I run tests with a small alphabet, and if those pass, I rerun the test with a large alphabet. The ordering matters, because test failures that happen with the small alphabet are usually more pleasant to reason about. I’ve actually never found a bug that only appeared with the larger alphabet, but it’s painless enough to keep both tests.

    1. 1

      Thanks for this, @bwr! Quick note: my RSS client says it can’t find your feed. I tried both the root URL and the /feed/ path.

      1. 1

        Oh, weird! It seems to work fine in my reader. Note that the URL is https://, and doesn’t have www. prepended. Can i ask what client you’re using?

        1. 2

          I tried it again, and it worked. 🤷‍♀️ Subscribed!

          1. 1

            Of course! I use Reeder 4 on MacOS and iOS. I’ll try again, making sure the URL doesn’t have the www.