Feb. 20th, 2024
I was getting on the nerves of anyone who did not refuse to listen: what exactly do I have to do to make tests run. Finally, Esteban writes to me: you need docker login. Hmm, wow, really? Ok, I do docker login. It does log in. Tests don't pass. Oh, I need to write a url where to do docker login. Ok, I add the url And I need to log in every time before launching tests. And? It does not log in. Credentials? Ok, I need to find credentials. Well, but I almost remember these credentials by heart.
Eventually, since docker shows the error, mentioning on which url it fails (it's artifactory again, but not the url I gave it). So I try it via curl. It works via curl. Ok, how about doing it via docker? Passing a parameter, -u user:password. It tells me that this is not secure enough. Well, wtf? What's secure? Manual input is secure. Ok, I do manual input. It digests it ok. And it logs in. And it loads tons of files into itself. And?
And it doesn't work.
Now, after the meeting, after scrum (imagine, people do work, seriously) everybody stayed and started helping me with advices. Saying some really bizarre words, like, I have to set environment variables. Which environment variables? Silence. Instead they write, in chat, something like `prlsl. mzbaz`. What's that? - I ask them.
And suddenly now Esteban writes that these tests don't work for him either. Wow, how interesting!
Eventually we all came to a conclusion that, since my computer is M2, not Intel, docker here should be configured differently, meaning, the VM.
And now they found how should I set tis `prlsl`. Then they told me what environment variables to set, and which values to assign. Good, good!
And you know what? The tests started passing. Except one, which requires k8s. I'll think about it tomorrow. Meanwhile, I have to document all these hacks and achievements. No, nothing like that was ever documented in this company.
I'll put it into a README file. I would also add it to their "wiki", but I already tried. First, it's not a wiki. It's the usual shit, doc files that go through "code review". And I tried, and I was told that I write it all wrong: it can be done the way I do, but can be done differently, and besides, there must be some parameters. In short, my doc is not good enough.
Why should I care, it's not my company. I have my branch of that wiki, and I'll keep my docs there. BTW, I can keep other docs in my own branch, I'm not their documenter or something. More, I'm not paid for that. I can even keep the stuff in my local files. It's not my company! How convenient! (And how unusual.)