Autonoma AI’s Post

Trying to stay consistent... here's another Pytest article! A quick and easy addition to shave time off tests.

Everybody knows that clean code is important. Clean code helps with code maintainability, readability - it just makes a fun time going into the code again. But what about tests. Don't they deserve that same treatment? Why don't we have clean tests? Here's a video on why we should do that. https://lnkd.in/eZxJFi5q #automation #testautomation #CleanTests #cleancode

Everybody knows that clean code is important. Clean code helps with code maintainability, readability - it just makes a fun time going into the code again. But what about tests. Don't they deserve that same treatment? Why don't we have clean tests? Here's a video on why we should do that. https://lnkd.in/enwvAUbu #automation #testautomation #CleanTests #cleancode

Copilot has become a force ⚡ Dev teams are increasingly considering it not just as a good code assistant but also a test assistant. Devs hate writing unit tests and Copilot can keep it that way. But Copilot as a test assistant has a few misses that most teams are caught unaware of. Everyone talks about the hits not what it lacks. Here is what teams need to know. Copilot is great at [1] Boilerplating. If teams have not written UTs ever, Copilot can generate boilerplates for your tests that can cut 70% of the scripting effort. [2] Discovery. Can explore and cover edge cases that devs ignore. Removes code bias but can introduce bias of its own. But Copilot will [1] Not generate tests that covers integration scenarios. These teams might still be fixing production bugs when their Copilot UTs are happily passing [2] Not test application behavior instead just code paths. UTs could be so overfitted to code that they might need refactoring to test behaviour. We explain this and more in our webinar but the simple message is, Think about Copilot as best at generating just low level unit tests which can provide a good head start. Don't depend on them for complete sanity. More here - https://lnkd.in/giERH9DP

Sprint demo summary: - Day before: development and testing, seems to work ok - 4h before demo: test, ok - 2h before demo: test, ok - 20min before demo: test, ok - Demo: run exact same test, fail - 30s after mic is muted: test, ok WTF!

1️⃣ You click once to create a Jam 2️⃣ Engineers get all their debug tools in one easy link

You can use Jam's automatic repro steps to write automated tests

I'm sharing this directly again after watching it. I highly recommend it! Marcelo Chiaradia does a good job of discussing the long-term costs of mocks and benefits of fakes instead. Which I'm actually personally experiencing at the moment in some heavily mocked tests. There's some good Q&A at the end, including an excellent bit by Valentina Cupać (Валентина Цупаћ) on Continuous Delivery, where she points out that CD means software is *shippable* to production at any time. Whether or not software *is* shipped every minute, every day, every few months or a year depends on other concerns. So don't downplay CD just because you're not planning on shipping continuously. What if someone comes in and says, "We need to push a release out to <someConsumer> immediately?" for any of a variety of reasons? Whether you're a startup still building initial product, an embedded system not planning on fielding the next version for months, or an existing web server that needs to fix a bug immediately, this scenario is always possible. Critical bugs, critical demos ("The investors are thinking about pulling their money"), etc. may trigger it. This presentation also matches James Grenning's advice from his recent TDD demo for embedded systems where he says not to fall in love with mocking, because it makes tests fragile, ugly, and hard to read. And I'll mention, hard to maintain! As he says, it's a tool, don't overuse it. So it's a method of last resort.

🥁 drumroll 🥁 Experimental Retries is here! 🤯 You can now specify what level of test flakiness should pass or fail by setting fine-tuned thresholds for retries based on what’s most important to you. Allowing you to reduce noise, boost agility, and increase control over CI resources and time. Ready to get started? 👉 https://lnkd.in/e9q73hHZ

🚀 Exciting News! 🚀 Just had an incredible experience using my performance test lab for the first time! 🔬🖥️ Left: array of machines, 💻 Middle: performance tooling, 📜 Right: clear lab instructions. Everything came together seamlessly! 🌟 #PerformanceTesting #testing #softwareTesting