23 posts
Cucumber's 'skipped' status hides three completely different problems. Here's how to diagnose which one is eating your test suite, and how to stop it.
Half your Cucumber scenarios are skipped. Not failed. Skipped. That yellow status is worse than a red build. It makes your suite look healthy while hiding…
Gherkin specs drift out of sync the moment you trust humans to update them manually. Here's how to automate the checks that keep feature files honest.
Your Gherkin specs are lying to you. Not intentionally. They started out faithful. But six sprints later, someone refactored the checkout flow and forgot to…
Most teams document architecture rules in wikis. Here's how to write them as executable tests that fail CI when your dependency graph drifts.
Your test suite verifies that returns 42 when given the right input. It does not verify that is allowed to import . The compiler is happy with both. Your unit…
Clean architecture diagrams look great on a whiteboard.
Someone on your team just imported into . The PR compiles. The tests pass. The code review is three hundred lines long, and nobody notices. Three months later,…
Architecture documentation rots the moment you save it. Here's how to keep docs honest using code-generated diagrams, ADRs, and automated architecture tests.
Every architecture diagram I've seen in a wiki was wrong. Not dramatically wrong. Just quietly, incrementally wrong. The service labeled "Auth" was split into…
A timeout or crash does not mean your API request was lost.
Your service crashes halfway through a request. The client sees a timeout and retries. You now have two charges. The customer is angry. The database is…
In-memory mutexes vanish when your server restarts. Here's how distributed leases with fencing tokens and TTLs prevent duplicate work across crashes, and...
Your does not survive a . It does not survive an OOM, a deployment rollout, or a node reboot. The instant the process exits, the lock is gone. If that lock was…
Most circuit breaker libraries spawn background threads to probe for recovery. You don't need them.
Every production circuit breaker I've reviewed eventually spawns a background thread. It might be a Go goroutine, a Java , or a Rust tokio task. The job is…
Crash-only software treats every failure as a crash and every startup as a recovery.
Your web service has a shutdown handler. It flushes buffers, closes connections, writes checkpoints. You tested it once, maybe. In production, it probably runs…
Mutation testing found a survivor and you have no idea what the mutation even does.
Your mutation testing report is full of survivors, and at least one of them makes no sense to you. The tool says it flipped a to a on line 47, or replaced an…
Why enforcing a single mutation score across your entire codebase is a mistake, and how to set per-module thresholds that match actual risk.
Enforcing a single mutation score across your entire codebase is a great way to make your team hate testing. Run PIT or Stryker against a typical repo and…
Code coverage says you're safe. Mutation testing says your tests are mostly decoration. Here's how surviving mutants expose the gap, and how to close it.
Your tests pass. Your coverage report says 87%. But your mutation score is 40%, and half your mutants are still alive. That 40% doesn't mean your code is…
cargo-mutants finds the tests that only pretend to verify your code.
You have 100% line coverage. Every branch is hit. Every function is called. Then someone changes a to a in your pricing logic, runs the tests, and they all…
Most teams don't run full mutation testing suites on every commit.
If your mutation testing suite takes four hours to run, congratulations. You've proven what everyone already suspected: your test suite has gaps. You are not…
Mocked database tests verify your SQL syntax, not whether rows survive crashes, concurrent writes, or schema mismatches.
If you mock your database in tests, you are testing that your repository layer calls the right methods. You are not testing that data survives a crash, that a…
Mocking every Redux action turns your tests into a changelog validator. Here's how to test your store with real state transitions instead.
If you have ever written a test that verifies was called with the exact payload shape, you have written a test that breaks every time someone renames a…
The default 100 examples in property-based testing is a social compromise, not a statistical strategy.
If you're running property-based tests with the default 100 examples, you're getting the worst of both worlds. Your CI is slower than it needs to be, and…
Example-based testing only covers the inputs you thought of.
You wrote a function. You tested it with and . It passes. You ship it. A user passes a single-element slice. Your function drops it on the floor. They open an…
Code coverage metrics create a false sense of security. Here's why unit tests miss the bugs that actually cost you sleep, and what to test instead.
You have 90% code coverage and still got paged at 2 a.m. The unit tests passed. CI was green. The bug made it to production anyway. Coverage didn't lie, but it…
Rust strips debug assertions automatically, but real design-by-contract needs more than debug_assert!.
Rust can enforce runtime contracts in development and erase them completely from release builds. The catch is that the language does not treat contracts as a…
Developers either scatter assertions like confetti or avoid them entirely.
Most production codebases fall into one of two camps. Camp A treats as a decorative seasoning, sprinkling it on every other line until the function reads like…
Manual validation bloats your codebase and still misses edge cases. Here's how to enforce runtime contracts with declarative schemas that stay out of your way.
Every time your API receives a request, you validate it. Every time a function receives an argument from an external system, you check it. Do this manually,…
strict mode catches the nulls you write, not the nulls that arrive at runtime from APIs, DOM queries, and JSON.parse.
You flipped in . You fixed every red squiggle. You shipped to production confident that and were solved problems. Then a backend response changed shape, a DOM…