• 1 Post
  • 163 Comments
Joined 3 years ago
cake
Cake day: July 3rd, 2023

help-circle
  • Oof. I’ve had places that the pipeline was getting long. At one of my previous jobs I made it so all the tests could run locally, and we were keeping the full build as slow as possible.

    We also didn’t do any browser tests (eg: selenium) because those tend to be slow and most people are bad at making them stable.

    It’s important to know whats worth testing.



  • There’s a lot of fear at my job about changing code. I’ve been trying to tell them to start writing automated tests. Or at least a linter to check for syntax errors. They’re all like “ooh that sounds hard maybe next quarter”

    Meanwhile, a trivial change requires a whole day because the developer has to manually test everything.

    I just unilaterally added checks to code I have ownership over, but anything shared I’m getting “maybe in two quarters we can prioritize this” from management.


  • My job has a “scrum master”. She’s nice, I guess, but as far as I can tell her entire job is sharing her screen so we can look at tickets. Then people tell her what to click on and what text to change. It’s excruciating because it would just be faster for the person talking to change it, instead of being like “remove the second bullet point. No, not that one”

    On top of that they have all these tasks for “unit testing” but they don’t actually do unit testing. Someone just said, in the distant past, we should do testing so it’s there.


  • Lol, yeah. If I saw an account labeled “American Nazi Party” with a blue check mark, I wouldn’t think “wow, Bluesky endorses Nazis” - I’d think “wow, this isn’t a satire account, these are actual Nazis, imma block them.”

    I’d think “wow they let Nazis on here. Like they know about them and are cool with that. This place is trash”















  • I’ve wasted entire days with people like that because they couldn’t be fucking arsed reading error messages and figuring things out by themselves.

    I’ve had a couple interview tasks that are like “clone this repo and run it. Try to do [action]. Tell us any errors you find and how to fix them”

    One of them was some sort of redux app, and the problem was a state mutation. Another one, the CSS had some weird so stuff rendered crazy. Both were pretty easy to track down and fix. You could probably also do something that’s like an error thrown, but people would probably just feed that into an AI now.