aslkfjasf. i've spent 12 hours today (and lots more over the past two days) trying to reproduce a bug that my [sort of] coworker insists is present. I haven't seen any proof of it anywhere, let alone steps to reproduce it.

I've poured through the code, following all of its tangled noodles of madness from start to fuck-this-shit. I've read and reread the pile of demon excrement so many times i can still read the code when i close my eyes. so. not. kidding.

anyway, the coworker person is getting mad because i haven't fixed the bug after days, and haven't even reproduced it yet. This feature is already taking way too fucking long so I totally don't blame him. but urghh it's like trying to unwind a string someone tied into a tight little ball of knots because they were bored.

but i just figured out why I haven't been able to reproduce it.

the stupid fucking unreliable dipshit ex-"i'm a rockstar and my code rocks"-CTO buffoon (aka API Guy, aka the `a=b if a!=b`loody pointless waste of mixed spaces and tabs) that wrote the original APIs ... 'kay, i need to stop for breath.

The dumbfuck wrote the APIs (which I based the new ones on mostly wholesale because wtf messy?), but he never implemented a very fucking important feature for a specific merchant type. It works for literally every type except the (soon-to-be) most common one. and it just so happens that i need that very specific feature to reproduce this bug.

Why is that one specific merchant type handled so differently? No fucking idea.

But exactly how they're handled differently is why I'm so fking pissed off. It's his error checking. (Some) of his functions return different object types (hash, database object, string, nullable bool, ...) depending on what happened. like, when creating a new gift, it (eventually...) either returns a new Gift object or a string error basically saying "ahhh everything's broken again!" -- which is never displayed, compared against, or recorded anywhere, ofc. Here, the API expects a Hash. That particular function call *always* returns a Hash, no matter what happens in the myriad, twisting, and interwoven branches the code could take. So the check is completely pointless.

EXCEPT. if an object associated with another object associated with the passed object (yep) has a type of 8. in which case, one of the methods in the chain returns a PrintQueue that gets passed back up the call stack. implicitly, and nested three levels in. ofc.

And if the API doesn't get its precious Hash, it exclaims that the merchant itself is broken, and tells the user to contact support. despite, you know, the PrintQueue showing that everything worked perfectly. In fact, that merchant's printer will be happily printing away in the background.

All because type checking is this guy's preferred method of detecting errors. (Raise? what's that? OOP? Nah, let's do diverging splintered-monolithic with some Ruby objects thrown in.)

just.
what the crap.

people should keep their mental diarrhea away from their keyboards.

Anyway. the summary of this long-winded, exhaustion-fueled tirade is that our second-most-loved feature doesn't work on our second-most-common merchant type.

and ofc that was the type of merchant i've been testing on. for days. while having both a [semi] coworker and my boss growing increasingly angry at me for my lack of progress.

It's also a huge feature, and the boss doesn't understand that. (can't or won't, idk)

So.
yep.
that's been my week.

...... WHAT A FUCKING BUFFOON!

!!pointless story

Bug report comes in from a coworker. "Cloudinary uploads aren't working. I can't sign up new customers."

"I'll look into it" I say.

I go to one of our sites, and lo! No Cloudinary image loads. Well that can't be good.

I check out mobile app -- our only customer-facing platform. None of the images load! Multiple "Oops!" snackbars from 500 errors on every screen / after every action.

"None of our Cloudinary images load, even in the mobile app," I report.

Nobody seems to notice, but they're probably busy.

I go to log into the Cloudinary site, and realize I don't have the credentials.

"What are the Cloudinary credentials, @ceo?" I ask.

I'm met with more silence. I use this opportunity to look through the logs, try different URLs/transforms directly. Oddly, everything seems fine except on our site.

I check Slack again, and see nothing's changed, so I set about trying to guess the credentials.

Let's see... the ceo is basically illiterate when it come to tech, so it's probably not his email. It's a startup, and custom emails for things cost money, and haven't been a thing here forever, so it's probably oen of the CTO's email aliases. he likes dots and full names so that narrows it down. Now for the password.... his are always crappy (so they're "easy to remember") and usually have the abbreviated company name in them. He also likes adding numbers, generally two-digit numbers, and has a thing for 7s and 9s. Mix in some caps, spaces, order...

Took me a few minutes, but I managed to figured it out.

"Nevermind, I guessed them." I reported.

After getting into Cloudinary, I couldn't find anything amiss. Everything looked great. No outage warnings, metrics looked fine, images all loaded. Ex-cto didn't revoke payment or cancel the account.

I checked our app; everything started loading -- albeit slowly.

I checked the aforementioned site; after a few minutes, everything loaded there, too.

Not sure what else to do, and with everything appearing to work, I said "Fixed!" and closed the issue.

About 20 minutes later, the original person said "thanks" -- never did hear anything from the ceo. I've heard him chatting away in the other room the entire time.

Regardless, good thing for crappy passwords, eh?

My manager had someone else manage me for my whole time at the company so far. Nearly two years now. Anything I’d come to him with, he’d direct me to this other person.

Fair enough, dude’s really good and I learn a lot from him. I see why they trust him with so much. I think he’s a genius. I’ll never be that good. Embarrassed I’m only a few years his junior. Wonder why he’s okay with being a manager for employee pay. Don’t think about it much, normal corporate BS.

Well it got way more “normal” when his ass got laid off without notice. Feel terrible. Him and 70% of my branch’s full timers. Wonder how I got so lucky. Everyone’s gone. We barely have enough people to do a standup. They all had 5+ years on their belts minimum. Only the contractors are left.

Manager emergency meets with me. Tells me all his best staff are gone and I am now the only front end guy on the team. He tells me he is not confident in the fact I am responsible for all of the old guys work and he is worried. He thinks I can’t do it cause he thinks I suck. Fuck me man.

My manager is pissing himself realizing he has lost the only people keeping HIS job for him. He has no clue my skill level. He sees my PR’s take a bit longer to merge, yet doesn’t realize I asked that friend of mine who was managing me to critique my code a bit harder, mentorship if you will, so we’d often chat about how to make the code better or different ways of approaching problems from his brain, which I appreciated. He has seen non-blocking errors come through in our build pipelines, like a quota being reached for our kube cluster (some server BS idfk, all I know is I message this Chinese man on slack when I get this error and he refreshes the pods for me) which means we can only run a build 8x in one day before we are capped. Of all people, he should be aware of this error message and what is involved with fixing it but he sees it and nope, he reaches out to me (after the other guy had logged out already, of course) stating my merged code changes broke the build and reverts it before EOD. Next day, build works fine. He has the other guy review my PR and approve, goes on assuming he helped me fix my broken code.

Additionally, he’s been off the editor for so long this fool wouldn’t even pass an intro to JavaScript course if he tried. He doesn’t know what I’m doing because HE just doesn’t know what I’m doing. Fuck me twice man.

I feel awful.

The dude who got fired has been called in for pointless meetings TO REVIEW MY CODE still. Like a few a week since he was laid off. When I ask my manager to approve my proposals, or check to verify the sanity of something (lots of new stuff, considering I’m the new manager *coughs*) he tells me he will check with him and get back to me (doesn’t) or he tells me to literally email him myself, but not to make any changes until he signs off on them.

It’s crazy cause he still gets on me about the speed of stuff. Bro we got NOTHING coming from top down because we just fired the whole damn corp and you have me emailing an ex-employee to verify PATCH LEVEL CHANGES TO OUR FUCKING CODE.

GET ME OUT

Why do we still use floating-point numbers? Why not use fixed-point?

Floating-point has precision errors, and for some reason each language has a different level of error, despite all running on the same processor.

Fixed-point numbers don't have precision issues (unless you get way too big, but then you have another problem), and while they might be a bit slower, I don't think there is enough of a difference in speed to justify the (imho) stupid, continued use of floating-point numbers.

Did you know some (low power) processors don't have a floating-point processor? That effectively makes it pointless to use floating-point, it offers no advantage over fixed-point.

Please, use a type like Decimal, or suggest that your language of choice adds support for it, if it doesn't yet.

There's no need to suffer from floating-point accuracy issues.