Are #FlakyTests causing trouble for your team's productivity & morale?
Trisha Gee, Java Champion and Developer Productivity Advocate, dives deep into the challenges of Flaky Tests, their costly impact on developers, and shares best practices for testing and boosting team efficiency.
Watch the #InfoQ #videopodcast on YouTube: https://youtu.be/xh_ur8VkO1g
The awesome team at @gradle.com have taken my blog posts and Opinions about Flaky Tests and put them all together in a very wonderful white paper!
Whether you care about flaky tests or think you don't have to worry about them, there's something in here for you.
https://gradle.com/resources/ebooks-and-whitepapers/flaky-tests/
Are #FlakyTests disrupting your CI/CD pipeline & draining Developer Productivity?
Trisha Gee dives deep into the hidden impact of these non-deterministic tests.
Learn why ignoring them is a costly mistake for code quality, build times & your team's sanity.
Watch the full #InfoQ #podcast video: https://bit.ly/3Gf20lX
#timestamps included in the description
I bumped into this again on an asymc (frontend) test suite and I have bad news for everyone:
If you don't run your tests in (pseudo)random order, you have no clue if your test suite is stable or not. No, none, nil, even if you've never seen it fail.
Do run your tests in random sequence. Always.
Today I’m going to delete some flaky tests.
Flaky tests are worse than no tests.
If there is more than one call to `faker.seed` in your entire test suite then you are *very* likely to hit many problems once the code grows big.
One of the easiest way to show almost guaranteed free lying rake is the concept of IDs that in most cases *need* to be unique for different instances of the same model. For that faker has cool function `uuid()` that promise to give you exactly that ... unless you use `seed`:
```
faker.seed(1);
const uuid1 = faker.datatype.uuid();
...
faker.seed(1);
const uuid2 = faker.datatype.uuid();
console.log(uuid1 === uuid2); // true
```
There are many others and even if you think this does not apply to your case the unnecessary risk is IMO way too high to instil this in your code.
Step into the life of a Google #DPE engineer with Simona Bateman's #DPESummit 2022 session! Watch “#FlakyTests & #BinaryRepositories: A Year in the Life of a Google DPE Engineer” now - https://dpesummit.com/flaky-tests-and-binary-repositories/https://www.facebook.com/photo.php?fbid=630405715753054&set=a.528530279273932&type=3
Can you run your test suite successfully 1000 times in a row?
