Hacker News Reader

> reproduce the regression with a failing test, then fix it

Not just regressions, but bugs in general.

If you can reproduce your bug in a test, then verify that your fix causes your test to pass, that's a great way to build confidence before trying out your code in production.

And in fact, this is mentioned in the article as part of maintenance:

> Writing a test first to prove out a bug before you fix it is pretty neat.

View on HN

That's the only type of test I know for sure has value, which makes them easy to write. At work I feel a certain pressure to test more or less everything, but personally I much prefer a "common sense"-approach: Writing tests is motivated by either bugs found (including any issues you run into during the initial implementation) and/or complex operations that are easy to get wrong and could need the extra documentation that tests provide. I dispise aiming for 100% code completion, so much of that is just idealism and wasteful.

It's also a great way to codify "hey we found this edge case once" and make sure it doesn't happen again.

Reply on news.ycombinator.com