Dude comes for an interview for a mobile position - one of the first things he says is "Adobe is killing Flash, don't know why, big mistake."

Never forget.

Let none say it wasn't a productive day.

They asked me if I could recommend any video streaming frameworks. I said no, but I could google around a little. I found one, sent it to them with a note that I hadn't used it but it seemed solid.

One afternoon, right before hometime, a month later and the day before go-live, I come in, to emails with _all_ the managers on it, demanding that I assist immediately. They'd finally tried testing it, and they had found an issue. No details.

I email back, asking for the actual issue they'd found - no response. I phone - that developer has now gone on leave for week, there's a new dev who'll help me. I email him, asking for "precise technical details" of what had gone wrong.

He replies, "when you try use it, it literally causes the apocalypse." and goes silent. I check the skies, no visible apocalypse yet.

Based on some keywords they'd mentioned, I google and find a known issue as well as a patch for their version. I email it over to them.

The response? "If I'd known he was just going to Google it, I would have tried that myself."

Forget StackOverflow profiles, recruiters should look for devRant profiles.

Yeah, I commented the code.
// ლ(ಠ益ಠლ)

Remember defragging?

Dev gets hold of me, says my service is down in QA. Works if he hits it locally, works via Postman, but via the QA app server it gives a 401.

I’m like, look, if it works everywhere else, there’s something wrong on your side in QA.

He insists, no, I must help him, and begins CCing all the managers telling them this system has been down for days.

So I eventually climb into his system, check the credentials they’re using in the QA environment, and sure enough, the password is wrong.

This developer keeps sending me issues he finds in his code. I think he wants to be a tester.

So, they've decided that they "don't need architecture."

They've told me I'm going to have to start doing timesheets.

When should I tell them I'm going to start producing work much more slowly?

Because nothing says "security" like some good ol' Base64 encoding. Bet whoever wrote that code was wearing mirror shades.

“Hi - we’ve done an analysis of the front end codebase and we’ve determined that it would be a good idea to start using Bootstrap. Are there any concerns with this?”

“Hey - no concerns, we actually already use Bootstrap in all those places.”

“I was not aware. So you’re saying there are concerns?”

"Hey business, we've done some style changes in a new test evnvironent, can you check to see everything looks fine?"

"Page Y (which you've never heard of before) looks terrible!"

"It does! Sorry" - go check the page on the other test environment.

That's just what it looks like.

The development principle behind this project was the infinite time/monkey/IDE approach.

Send the dev a request for a design before they do a delicate process. They send back some mess done in MSPaint and a ton of select statements, with no context or explanation.

Send back an example of what they need to work through. Get back a response saying that the previous spec was just a draft, and once they complete development they’ll complete the design.

Why would I want a design once it was too late? What would be the point?

Yes, come tell me there's a user experiencing an issue, but you don't know on which page or exactly what it was. I'll fix that right away.

Designing a framework: design the framework. Have a bunch of drinks. Try use the framework. Sober up and examine how you’ve butchered it.

Adjust the framework.

The last communication we had was me asking you if you would like a service for that, with no response? And now, ten days later, you’d like an update on when it will be available?

Well I raised a pull request on your side a week ago, your developers have just been ignoring it.

What? No, I don’t know when it will be merged and ready.

I get listening to music while you code, but I’m afraid I just can’t believe in watching a TV show while you code. You’re not coding, you’re watching a show. Or if you are coding, the level of code that’s being written is probably low enough to make it a waste of time.

When you've been without a clue for ages, life just isn't making sense anymore, and then you realise your IDE imported java.awt.List

Trying to work out an encoding bug after upload, I ask them to send me the original version of the file that is currently uploaded. I look at the files and they seem strangely similar, but strangely different.

I reply to them, "To confirm, this is the version you uploaded?"

"Nope."

‪And verily, he looked upon the codebase and cried out, "Who has done these terrible things?".

And there was wailing and gnashing of teeth.‬

These devs keep getting hold of me to tell me that what I told them to do worked.

Well, yeah. That's why I said to do it.

GODDAMNSIT when you have a perfectly good rant to rant but it’s long and it’s a pain to write on a mobile and devRant DOESN’T HAVE A REAL SITE

My favourite bug reports are the ones for where something is missing. No details on pages, or steps to recreate, but they do include a screenshot.

Essentially, they go "Here's a picture of something that isn't there."

Great, thanks, I'll get right on that.

"Okay, we'll try to explain again why you being able to execute your test cases against prod is a bad thing..."

2005 called. It wants its numbered file names back.

While I am mostly satisfied with "celluloid" as a worthy successor to xplayer, the first major disappointment I stumbled upon is `celluloid-shot0001.jpg`. Are we in 2005?

Just like xplayer, Celluloid, the new default media player of Linux Mint, should use proper, i.e. time-stamped names such as `celluloid-2023-04-10T00-47-42.jpg` or `celluloid-video_file_name-2023-04-10T00-47-42.jpg` for screenshots taken from videos, to eliminate the possibility of file name conflicts if files are moved into other directories, to make screenshots searchable by video file name, and to retain the date and time information if the files are moved to a device that does not support date and time stamp retention such as MTP (Media Transfer Protocol), and to allow for date range selection using wildcards in the terminal (e.g. `celluloid-2023-04*` for all screenshots from April 2023). Besides, PNG screenshots should be supported too, but that's out of scope here.

As a reference, the gnome and mate screenshot tools also pre-fill time stamps into the file name field.

Numbered file names were useful in an era when there was no VFAT and file names needed to have 8.3 file names that could impossibly fit a date and a time, and compact cameras used such names, but those times are long over. Just like the useless and annoying pull-to-refresh gesture on mobile apps and the Media Transfer Protocol, numbered file names belong to the technological graveyard.

If numbers are really desirable, at least `celluloid-shot0001.2023-04-10T00-47-42.jpg` should be used, to include both a number and a date. The command to get this date format is `date +"%Y-%m-%dT%H-%M-%S"`. For compatibility across operating systems, dashes instead of colons have to be used to separate hours and minutes and seconds.

Numbered file names are a thing of the past. Use time stamps.

So I'm dealing with an Apache Tomahawk schedule component that's started acting up.

It turns out that when you click on a time slot, instead of just posting up the time slot, it posts back to the server the component you clicked on (the whole calendar), and the y position of the click.

Now, imagine the user has zoomed in... yeah, shit stops working.

"Hi, I was wondering if we have a way to <do a standard thing that we have support for, and has already been done in their project>"

Also, of course, let's CC every team lead and manager we can find so they all think it's something dire and we can end up in tons of meetings about it. Great.

This guy keeps insisting there’s a bug in my code, on a specific line. The stack trace shows that a NPE is thrown in his code, before that line is reached, but he won’t be persuaded by this argument and won’t send me the class.

Somehow he’s certain that Java would throw a NPE on trying to iterate through an empty list, as if his code was even returning an empty list. Can you imagine the chaos.

"When I try make a change, system XYZ is not happy do you know why?"

Really? That's all you're giving me to go on?

They did the worst thing you could do to a self-respecting dev.

They asked him to pop around to HR to have a head-shot taken.

We regularly have deadlines set before they've worked out anything about what's in the release, and business takes their time, so we're literally told "we don't know what business is going to want, but we have to start coding now to meet the deadline"

Ok so that's my plan, find a kernel with HUGE amout of drivers and , high version.

I built a small os based on linux
-- kernel version 5.0.2 from Plop Linux,
many libraries added 'by hand' -- packages from apts of Debian&Ubuntu, and unpacked packages into system with ArchiveManager,
has GUI but it's called xfree86 ( looks strange when a very old app running on Kernel5 )
So, without compiling, i can make a os.

But i found that Plop didn't compile rtl8188eu module which makes linux support some specific network cards.

I have no professional compiler but a tiny C/Cpp compiler called TinyCC (aka. tcc), but for my pc ( CPU freq = 800MHz ), it seems not possible to compile the module by myself.

And then i downloaded a 5.2 kernel with modules from kernel.ubuntu.com, but when i tried to mount my disk ( part. vfat ), i got some errors like IO charset not found, and then i replaced it with Xanmod kernel but also reported an error said Invalid Arguments, but i checked /proc/filesystems, it supports.

So what can i do? Are there any pre-compiled kernel & modules with 'full common supports'?

I tried kernel 4.4 ( from Ubuntu 16.04 LTS ) just now but the driver crashed when wpa_supplicant tried to initialize the device.