I try to change some code in a VERY obvious way, no code that might be interfering. Compile, open browser, no change. Recompile, check, that the compiled output indeed contains the changed code, refresh, no change ... what??
Delete browser cache, also reopen with new tab, clear IDE cache, restart, rerun with npm i. Still no change... ??? What is this black magic?

So fucking exhausting when you have to change many things in a codebase in a repetitive manner with search and replace and it's not worth the time to write a rool to automate it. One of those special cases where you wish search and replace would support some kind of templating or be variable depending on the search result

Nobody in the department knows what an ancient program really does, not even people that have been in the department for 20 years. Yet it seems to do important things.Time to shut it down and see what it was responsible for. Maybe we'll see the consequence in the next year(s), lovely 🙃

This fucking Integration test with h2 is driving me insane as it just doesn't work for what I need it to! 😤
The setup seems correct, the sql files are "processed" according to logs. Everything seems to be set up as supposed, as AI, colleagues or internet recommend ... yet no insert statements are shown and no data is found on retrieve. No schemas or tables exist in h2 console although log says all these millions of tables are apparently created.
WTF am I doing?

Presenting your work and disabling all breakpoints sometimes feels like a driving license test but the brakes are disabled. Just pray nothing unexpected happens.

As other people have already stated in the past, safari is the new Internet explorer with its quirks. A feature that works on chrome/edge/chromium and Firefox does not work with safari. If only all browsers could build on one baseline of code as rendering engine and build their stuff on top...
Guess there is always a reason not to.

Thought I would be clever to batch insert a bunch of data into the db with sql, with the catch that the data of a specific column should have a zero-padded number as prefix, counting up from 0. Ask AI for suggestion. Generates some SQL beyond my basic understanding. No way I'm gonna run that! Can't be bothered for an explanation, don't have the time. I realize my assumption to SQL my way through failed right from the start. Instead I let AI generate a python script to create a CSV with the desired data. A few iterations of testing, I take the result file and import it into the db. Done.

Curious about AI assisted TDD. Assisted writing of specs and then assisted implementation. Any experience so far? I kind of like the idea but I'm afraid it's like doing TDD with a toddler, helping less than 50% of the time.

What to think about an OSS library which has basically thrown away all testing? If a group of contributors maintains a big library used by millions of people, wouldn't you expect some kind of QA in form of tests in order to not let people break existing behavior? Ideally new contributions should provide code with tests too, right?

Tried to run ollama with a small model on a laptop that's like 10 years old and use that inside vs code. Weak CPU, GPU not to mention. I didn't really expect it to work. I was still a bit disappointed, though as expected. It was crying for help.
Are there any laptops powerful enough?

How hard can it be to reference a file on a mounted windows network drive with UNC path to the java command? Seems like everyone uses the same env variable but always different syntax. Sometimes it's
-Dsmth=file://\\net.work.path\sub.xml, then it's -Dsmth=file:///\\net.work.path\\sub.xml, then
-Dsmth="file:///\\\\net...xml"
and none of them work?! 😤

Angular standalone migration breaks the project on the first step and I cannot find the culprit of the error. So it has to be done manually. Any other idea before doing all by hand? 😩
I guess nuking it all is also an option.

ExpressionChangedAfterContentCheckedError my ass.

For me the DX with Kotlin on VS Code is mostly trash. Never thought I'd give Intellij another shot.

So many open tickets and they keep becoming more and more. In the end they will never be worked on, as there's always smth more important, although they are reprioritized on fixed schedules. They will eventually be discarded because they were forgotten and the ticket was already done, just by a ticket with similar naming or similar intent.
How would you solve this? Can this be solved?

How can you build an oidc library almost completely on observables? Cannot complain about it, as it's open source, but what if you want to access some token data synchronously?

My network is somehow fucked up. VPN Client on laptop to my workplace is not connecting via wifi but instead connects when my phone creates a hotspot, while the phone is connected to the same router via wifi? Second laptop this occurrence.
When the work laptop is outside of VPN but connected to the router with wifi, about 50% of web pages are not reachable, same with my own PC and another laptop. Didn't touch my router in ages. Nice fucking conditions for wfh.

What a fucking weekend. Tried to clone an existing windows hard drive onto a new unused one with the same size. Oh the adventure that I got myself into. 😭
Old hard drive corrupted in the middle of the process (how???). Not bootable anymore (until at the very end).
Then an odyssey started to try to rescue the old data using my spare linux drive onto a USB stick. Fucking with uefi settings and apparently a boot sequence with incorrectly named drives (?).
Had to reflash a usb stick I had for linux installation to windows installation.
I had multiple drives attached to my PC and didn't want my own ones to be overridden, so I needed to detach all drives except for the target one.
The target one had already fucked up partitions from the failed cloning, so I had to research the tools to fix it manually, not knowing what I was doing.
At the end it worked out, windows installed, data copied onto the target.
The old drive was able to recover, but I didn't give a shit about it anymore.

Was a sleepless night, spent wayy too much time with this...

Safari shows a blank page and no errors in the console. How the fuck are you supposed to debug that??

Every time I try to write a unit test I seem to write an integration test instead. 🤦
I'm just awful at it.

TDD shows you just how much "junk" you've added to the codebase before without TDD. Then it becomes TDD refactoring. Wtf, what a mess. 😑

Rant/question
Seems like I cannot run any GPT model on my GPU as they require more gpu memory than my 6GB card can offer. 😪 Wanted to try out self hosted. Couldn't solve Cudas outofmemoryerror yet with given solutions.
Any recommendation for a smaller model?

Seems like spring boot kotlin doesn't want to cooperate in any way. Tried to set up a backend with it, but it keeps showing an error regarding wrong jvm-target. The project runs nevertheless. I changed the maven settings according to the docs, but no change. Whatever, still runs. However a newly created "hello world" rest endpoint just doesn't work, even following a beginner tutorial. localhost:8080/hello ... error fallback page.
I really wanted to give kotlin a try. Doing the same with java, instantly working.
Fuck spring boot kotlin. Or fuck me for not knowing how to handle it.

Thought I'd give kotlin spring boot a shot. I assumed it would work out of the box like with java. It didn't. Apparently with jdk 17 I chose an incompatible version with the gradle version provided. Downgrade gradle. 'runApplication()' still marked with an error, which I cannot seem to solve. Answers from the internet are no solution.
But I can run the project ... but I cannot reach my dummy address, same with maven ... wtf, which part of rtfm did I miss? Wasn't kotlin supposed to be the better java?

My work laptop didn't connect to our in-house vpn when connected to my home wifi. Then again, connected the laptop to my phone as a hotspot, my phone connected to my home wifi. That worked flawlessly? How??

Fuck npm and its mess with dependencies 🙄

Is there a (F)OSS solution for a self hosted document management system that includes ocr, text based search on all documents and a web ui? So far my research hasn't been very successful, maybe someone has a hint? I've thought about building one myself. Then again someone else must've already built something similar, right?

Setting up a task with windows task scheduler is fucking confusing if you need time-based repetition. At least if you're not doing it in english. Some options seem like the same thing, some are probably badly translated.

The logic in a method reflected a truth table not obvious to understand. The first time I implemented tests for coverage as I didn't trust anybody to remember and understand the logic after a few weeks.

Did a big refactor with no tests and the code works first try 🤔🤔