Hey everyone - we're currently experiencing some intermittent connectivity issues with our servers. The app may be becoming unavailable for short periods of time and we are investigating.

Thanks for your patience.

Saw some cheapie little radio in the dollar store, bought 2 of them for reverse engineering. Powered it from my lab bench power supply as usual, and tested whether it actually works before doing anything else.. then I noticed that the tunes were actually quite catchy, so I just ended up listening.

Then I started to notice that the audio wire I was using (the one I've spent a couple of days building earlier) had intermittent audio issues where the right driver would drop out when the wire was held in certain positions. Oscilloscope probing showed that there was some sort of disconnect, with only the 50Hz noise from the power lines showing up. Opened up the connector and noticed that the ground wire had detached. An 28AWG electrical wire that was inside a jack that was meant for stress relief! Yet the copper strands must've detached one by one regardless. What do I need then, huh?! 18AWG which wouldn't even fit on the connector, only to see the strands in that eventually detach as well?! You know what, let's go fancy.. 1AWG which is meant for extremely high current applications!!

At that point I was literally shouting "FUCK!!! Why does this shit always happen to me?!! ONE FUCKING PROJECT THAT FINISHED SUCCESSFULLY, YET STILL BROKE?!!!! WHY!!!!!!"

Clearly I need some fresh air to cool down. On my way to the fast food restaurant to get some Bicky burgers. More shit, humans. One stupid driver who slowed down on me, which of all things I hate the most. GO FASTER ALREADY YOU SLOWFUCK, AND GET THE FUCK OUT OF MY SIGHT!!! Next a pedestrian with a dog.. I swear motherfucker, if that dog comes anywhere near me I'll personally turn it into fucking fricassee.

Ah and then comes the killer.. in this stupid fucking summer, all that's needed to fix any issue is a fucking stupid DESPACITO, right?! More like DeSPASTICo!! FUCK!!!!

.. Back home, rather tired. So essentially a wire that was specifically built to have high endurance broke on me. Back to Bluetooth I guess.

The stranges computer error I have seen.

When modems still was the way you communicated I worked with support.

We where the general distributor for TDC's PCCard modems.

One day we got a computer with modem with intermittent problems (worst kind).

After much trial and error we found that stroking the computer lightly in the top right corner made the modem work perfectly :)

Cause will be reviled later :p

Had a 1:1 with my boss last night and together we figured out a tricky bug related to my PR. However, either my PR or that bug patch broke a tangentially-related test. Queue my usual exhaustion, and I gave up trying to fix it.

This morning, I'm looking at it and nothing makes sense. My change should not have broken the test. So I reran the controller's tests, and... they all pass?

What is logic.

Good thing, though; that test leads to a few rabbit holes I haven't even begun exploring yet.

Oh, never mind. It broke again.
Ergh, here we go. 😔

I've told the same story multiple times but the subject of "painfully incompetent co-worker" just comes up so often.

I have one coworker who never really grew out of the mindset of a college student who just took "Intro to Programming". If a problem couldn't be solved with a textbook solution, then he would waste several weeks struggling with it until eventually someone else would pick up the ticket and finish it in a couple days. And if he found a janky workaround for a problem, he'd consider that problem "solved" and never think about it again.

He lasted less than a year before he quit and went off to get a job somewhere else, leaving the rest of our team to comb through his messy code and fix it. Unfortunately, our team is mostly split across multiple projects and our processes were kind of a mess until recently, so his work was a black box of code that had never been reviewed.

I opened the box and found only despair and regret. He was using deprecated features from older versions of the language to work around language bugs that no longer existed. He overused constants to a ridiculous degree (hundreds of constants, all of which are used exactly once in the entire codebase, stored in a single mutable map variable named "values" because why not). He didn't really seem to understand DRY at all. His code threw warnings in the IDE and had weird errors that were difficult to reproduce because there was just a whole pile of race conditions.

I ended up having to take a figurative hacksaw to it, ripping out huge sections of unnecessary crap and modernizing it to use recent language features to get rid of the deprecation warnings and intermittent errors. And then I went through the same process again for every other project he'd touched.

Good riddance.

Recently I disassembled this Remington beard trimmer that I have, right. Thinking that I probably just depleted the battery or something like that, and that that's why the fucking thing would turn off after 10 seconds, with a miserable motor speed that could only be caused by what.. 2.5V or something like that? So I kept up with it for a while, and then a few days ago, I finally disassembled the fucking thing.

New 4.2V LiPo battery of my own attached, wicked! The thing has a 2Ah capacity and is fully charged, so fuck all in a big ship, I'ma shave my fucking hair without protection, who cares with 2Ah anyway, given that motors at 4.2V consume only 800-ish mA, right. So I shaved my head, and figured.. holy shit mate, I fucking beat this fucking cunt, fuck you Remington! Planned obsolescence my ass!!!

Just now @xewl sent me a request that made me reconsider this shaver. Some PS3 controller, should be easy enough to fix. But I referenced my shaver so I figured that I might actually check it beforehand.

The original fucking battery has 4V left on it. 4 fucking volts! That ain't gonna trip shit for an overdischarge protection circuit!!! WHY THE FUCK REMINGTON, WHY THE FUCK DID IT (IF ANYTHING) TRIP?! And more importantly, why did you make the motor turn for only 10-ish seconds after charging, 1 year after purchase, despite the fucking cell having 4 fucking volts in it, which is more than plenty to get the thing turning for at least a 60% charge?! Planned obsolescence perhaps, hmm?

But let's say that it was bad contacts or something like that. After all, my probing was a bit intermittent, showing 1.5V (if it's shorting, shouldn't it be in the mV range?) and then 4V at times, until I separated the nickel connectors from the 16500 cell. Why the fuck should that thing short, considering that it's unprotected?! The fucking cell is unprotected, and you Remington are going to let it short? What kind of fucking engineering is that, you fucking braindead morons?!!!

And why do you use a 16500 cell which isn't very standard over a 18650 cell, which is? Again, planned obsolescence? Fucking vile pieces of shit.

Now what's more.. I tried to power on the shaver for a picture for this fucking rant, using a LiPo cell of my own that I used to shave my hair with the other day, to get a nice picture of my little hack. But the fucking thing, despite having 3.89V left in it after that fucking shave, which is more than plenty for a lithium cell, the fucking unit doesn't power on. WHAT THE FUCK REMINGTON?!! That's planned obsolescence at its best, init?!! You vile fucking pieces of shit.

So I guess that this rant is no more one of "I beat these Remington cunts". It is one of "you can bet your fucking ass that planned obsolescence is a thing!!"

The worst kind of problem you can have while coding is an inconsistent or intermittent bug... For code you didn't originally write...

How can you finish you task if your workstation don't have internet access, your "internet" laptop - which your personal laptop - have limited/intermittent internet connection, and the db your testing is under restoration...

I've been compiling the project for about 5 hours now, still no successful build due to bad tests generating intermittent test failures...

All I wanted to do was to release the web project to the customer not fucking wrestle Cthulhu!
The worst part is that the release is set up so that you need to release the entire project internally before you can release one part.

"I'm gonna wake up early, and hammer out this feature, being super productive tomorrow"

Tomorrow comes:
- unexpected build errors
- unexpected runtime errors
- intermittent CI pipeline errors
- spends two hours trying to resolve errors
- literally hasn't touched the thing that's important

What else is going to go sideways? Watch Bitbucket or CircleCI fuck up and refuse to deploy this live for some stupid reason.

Rant!
Shifted to new office few days back.
Just had to put up with the shitty intermittent internet connection.
SSH failing and coming out for every 50 secs.
Trying to take Mongo Dump and restore for the past 2 hours. Finally! ufff.

When the network is disconnected SSH gets disconnected. Fine. Acceptable. Again, when the network gets connected, it should start from the same place right ? Dumb Enough. But i wanted that today.

I'm noticing that installing a new distro and making it your main distro tends to be very stressful to me.

Here's a couple of reasons why:

There's this anxiety I get that is just crazy, like I need to set this thing up perfectly, meanwhile I have work to do.
And the thing you estimated to take 2 hours at most ends up being a full week.
Work starts to pile up but can't just shake the obsession with setting this thing up.

Another reason is that usually my main distro instance tends to be VERY customized.
So much that when you switch, you need to configure those things again, and troubleshoot a lot.

Troubleshooting isn't a problem in itself. The problem is that you're trying to troubleshoot with a machine that needs troubleshooting.

the shell, the editor, the terminal, the window manager, tmux, the notifications, the custom kb layout. it was all so neatly customized that fixing problems on the machine was manageable.

but without all that I feel like an idiot.

Even something as simple as keeping having youtube music to focus is now intermittent because you need to restart the X server.

Maybe my problem is that I don't do well for bootstrapping.

Anyhow, I'm really stressed right now, but I think it'll be worth it.

anyone else having this weird bug that when you scroll down under a rant, nothing happens and scroll bar grows to infinity? (only tested in firefox, rambox)
when i manually scroll very fast i might "skip" this behavior and actually be able to scroll through comments under a rant, but when i scroll up too much it jumps to the top again. ^^'
if required, i can upload a meaningless, blurry video filmed with my phone in catastrophe mode (portrait)

Monday AM. Instead of building images in the backend i'll just make them in HTML and use a headless browser to screenshot them ! Sweet. Doable in a day and saves the rest of my week !

Friday afternoon. Crashes, browser failure, zombie processes, intermittent failure, docker failure with headless browser. Still not working well.

Why am I such a dumb fuck ?

Working on a multi-year college project, going through tests from previous team.
Every test is not working quite right. They're almost all intermittent failures.
The reason? Every single test class extends some test class, which usually extends from some primary test class.
That primary test class opens up their whole UI, and outside of their UI test package, the only thing that gets used is a variable named session (a string), which isn't even specific.
WHY THE FUCK WOULDN'T YOU JUST MAKE THE SESSION NAME STRING A VARIABLE IN THE TEST FILES YOU DUMB FUCKS
THE ARGUMENT VALIDATION TESTS DO NOT NEED TO OPEN THE UI, LET ALONE CREATE THE WHOLE FUCKING DATABASE JUST TO VALIDATE ARGUMENTS, WHICH YOU DO APPLICATION SIDE
(Also they made it so every session has their own tables as opposed to having session IDs. E.g., "person_sessionID1" and "person_sessionID2" exist.)

Back when I was starting out in a full stack role, I worked on a fairly big chunk of functionality that would trigger off a few entry points. It was wonderful for a few months. As we approached go live, our QA team started reporting weird intermittent issues. The logic wouldn't "trigger" the first time, but would on subsequent saves. Worse yet, the state required resetting of data every time we needed to test. Three weeks later, it boiled down to a 2 second time difference between the database's GETDATE() values and the new Date() object we passed in from our application.

I'll never forget that one system should be the source of truth again.

We have had 2 days on intermittent internet literally we have 5minutes of down time then 15-25seconds of the internet working. How can a tech company be expected to function like this!

I vaguely remember some joke about how difficult networking is and how some Jeff Atwood blog post I think makes the comparison about analogy of sending actors in a taxi to somewhere being compared to a packet, anything can happen inbetween but you will get the packet or something indistinguishable from the original at the other end if all goes well.

Are occasional/intermittent 503 service unavailable or 504 gateway timeouts unavoidable for microservices calling another external microservice?

Like at that point isn't receiving a 503 or 504 from something else, somewhat outside my jurisdriction, albeit I am fucked if I am depending on them and need to fail gracefully.

I've been working on an extremely intermittent bug for the past week in my project that occurs during a stress testing between a PC based server and an embedded device that share files. When the crash happens, I analyze what happened by looking at a file as a result of an fwrite, look at a diff of it, look at the packets etc. For the past 3 days I had been lead to believe there was a bug in stdio.h's fwrite due to a file being written looking like it was truncated in the diff, but the packets telling a different story (X bytes sent to be written, on the result I report X bytes written). Today I noticed that there was either a bug or an issue in the diffing algorithm that led me to think my code was the issue. I spent 3 fucking days trying to figure why fwrite was truncating and lieing about its result when my diff tool was the culprit. FML.