5
black-kite
252d

A colleague changes the location of a test helper file imported in a bunch of tests. Doesn’t bother to check where that file is imported (except where he’s specifically using it himself).

As a result good dozen tests fail later on. The culprit doesn’t realize. And the rest of us have no clue why.

Multiple people are asked to look into why the tests are failing.

"Ok, who’s working on what?
We’ll create a shared document to track who’s working on what test."

Document is created, people get assigned.

"Hold on, looks like it’s just a faulty import." "Oh yeah same here." "Yeah for me as well."

"Ok we could simply appoint one person to fix all the imports."

"Well I’ve already gone ahead and opened a pull request to fix the test I was appointed to."

"Ah ok, well who can take care of the rest?"

"Wait I also opened a PR."

"Ok so I guess we can each open a PR?" "No we can just have a single branch we can all push to".

"Sure, who’s branch are we using"

"You can use the branch of my PR"

"Guys let me handle this, it’s ridiculous for us to all be doing this separately."

"You’re right, go ahead".

——

The culprit? A senior dev.

What would have literally taken a minute to do (or even no time at all with proper use of the IDE) turned into hours of wasted time. People getting interrupted, having to drop what they were doing to fix the consequences of this guy’s laziness (seriously don’t know what else to call it).
Ok maybe our reaction could have been more efficient, but we never should’ve even gotten to that point in the first place.

Comments
  • 4
    CI for the win here. Tests start to fail after a commit, the commit author's problem.
  • 1
    @atheist These are e2e tests. They’re quite expensive to run, so don’t get triggered automatically when a PR is opened…
Add Comment