A recent blog post from Jane Street explores the concept of "expect tests," a testing methodology designed to make test writing more enjoyable. Unlike traditional unit tests, expect tests use a pattern/library that provides rapid feedback, much like a REPL session or a Jupyter notebook. This approach allows developers to quickly iterate and visualize the behavior of their code, leading to faster development cycles.

The core idea behind expect tests is to minimize the initial effort required to set up a test. Instead of manually asserting expected values, developers can use a blank `[%expect]` block, letting the system figure out the output. The system then provides a diff showing the exact changes needed to pass the test. This workflow stands in contrast to the more cumbersome process of traditional unit testing.

This methodology, inspired by tools like Mercurial's unified testing format, has become a core part of Jane Street's testing suite. The blog post argues that expect tests offer benefits such as better documentation and regression detection. It promotes a testing style that can evolve to be more targeted and concise, while still allowing for rapid experimentation.

Expect tests offer a compelling alternative to conventional testing frameworks. They promise to reduce the friction associated with writing tests. As developers seek ways to improve productivity and code quality, expect tests are a tool worth considering. They enable faster feedback cycles and greater ease of use, making testing a more integral and less burdensome part of the development process.