Good Morning!, its time for practiseSafeHex's most incompetent co-worker!

Todays contestant is a very special one.

*sitcom audience: WHY?*

Glad you asked, you see if you were to look at his linkedin profile, you would see a job title unlike any you've seen before.

*sitcom audience oooooooohhhhhh*

were not talking software developer, engineer, tech lead, designer, CTO, CEO or anything like that, No No our new entrant "G" surpasses all of those with the title ..... "Software extraordinaire".

*sitcom audience laughs hysterically*

I KNOW!, wtf does that even mean! as a previous dev-ranter pointed out does this mean he IS quality code? I'd say he's more like a trash can ... where his code belongs

*ba dum tsssss*

Ok ok, lets get on with the show, heres some reasons why "G" is on the show:

One of G's tasks was to build an analytics gathering library for iOS, similar to google analytics where you track pages and events (we couldn't use google's). G was SO good at this job he implemented 2 features we didn't even ask for:

- If the library was unable to load its config file (for any reason) it would throw an uncatchable system integrity error, crashing the app.
- If anything was passed into any of the functions that wasn't expected (null, empty array etc.) it would crash the app as it was "more efficient" to not do any sanity checks inside the library.

This caused a lot of issues as some of the data needed to come from the clients server. The day we launched the app, within the first 3 hours we had over 40k crash logs and a VERY angry client.

Now, what makes this story important is not the bugs themselves, come on how many times have we all done something stupid? No the issue here was G defended all of this as the right thing to do!

.. and no he wasn't stoned or drunk!

G claimed if he couldn't get the right settings / params he wouldn't be able to track the event and then our CEO wouldn't have our usage data. To which I replied:

"So your solution was to not give the client an app instead? ... which also doesn't give the CEO his data".

He got very angry and asked me "what would you do then?". I offered a solution something like why not have a default tag for "error" or "unknown" where if theres an issue, we send up whatever we have, plus the file name and store it somewhere else. I was told I was being ridiculous as it wasn't built to track anything like that and that would never work ... his solution? ... pull the library out of the app and forget it.

... once again giving everyone no data.

G later moved onto another cross-platform style project. Backend team were particularly unhappy as they got no spec of what needed to be done. All they knew was it was a single endpoint dealing with very complex model. There was no Java classes, super classes, abstract classes or even interfaces, just this huge chunk of mocked data. So myself and the lead sat down with him, and asked where the interfaces for the backend where, or designs / architecture for them etc.

His response, to this day frightens me ... not makes me angry, not bewilders me ... scares the living shit out of me that people like this exist in the world and have successful careers.

G: "hhhmmm, I know how to build an interface, but i've never understood them ... Like lets say I have an interface, what now? how does that help me in any way? I can't physically use it, does it not just use up time building it for no reason?"

us: "... ... how are the backend team suppose to understand the model, its types, integrate it into the other systems?"

G: "Can I not just tell them and they can write it down?"

**
I'll just pause here for a moment, as you'll likely need to read that again out of sheer disbelief
**

I've never seen someone die inside the way the lead did. He started a syllable and his face just dropped, eyes glazed over and he instantly lost all the will to live. He replied:

" wel ............... it doesn't matter ... its not important ... I have to go, good luck with the project"

*killed the screen share and left the room*

now I know you are all dying in suspense to know what happened to that project, I can drop the shocking bombshell that it was in fact cancelled. Thankfully only ~350 man hours were spent on it

... yep, not a typo.

G's crowning achievement however will go down in history. VERY long story short, backend got deployed to the server and EVERYTHING broke. Lead investigated, found mistakes and config issues on every second line, load balancer wasn't even starting up. When asked had this been tested before it was deployed:

G: "Yeah I tested it on my machine, it worked fine"

lead: "... and on the server?"

G: "no, my machine will do the same thing"

lead: "do you have a load balancer and multiple VM's?"

G: "no, but Java is Java"

... and with that its time to end todays episode. Will G be our most incompetent? ... maybe.

Tune in later for more practiceSafeHex's most incompetent co-worker!!!

Welcome back to practiseSafeHex's new life as a manager.

Episode 2: Why automate when you can spend all day doing it by hand

This is a particularly special episode for me, as these problems are taking up so much of my time with non-sensical bullshit, that i'm delayed with everything else. Some badly require tooling or new products. Some are just unnecessary processes or annoyances that should not need to be handled by another human. So lets jump right in, in no particular order:

- Jira ... nuff said? not quite because somehow some blue moon, planets aligning, act of god style set of circumstances lined up to allow this team to somehow make Jira worse. On one hand we have a gigantic Jira project containing 7 separate sub teams, a million different labels / epics and 4.2 million possible assignees, all making sure the loading page takes as long as possible to open. But the new country we've added support for in the app gets a separate project. So we have product, backend, mobile, design, management etc on one, and mobile-country2 on another. This delightfully means a lot of duplication and copy pasting from one to the other, for literally no reason what so ever.

- Everything on Jira is found through a label. Every time something happens, a new one is created. So I need to check for "iOS", "Android", "iOS-country2", "Android-country2", "mobile-<feature>", "mobile-<feature>-issues", "mobile-<feature>-prod-issues", "mobile-<feature>-existing-issues" and "<project>-July31" ... why July31? Because some fucking moron decided to do a round of testing, and tag all the issues with the current date (despite the fact Jira does that anyway), which somehow still gets used from time to time because nobody pays attention to what they are doing. This means creating and modifying filters on a daily basis ... after spending time trying to figure out what its not in the first one.

- One of my favourite morning rituals I like to call "Jira dumpster diving". This involves me removing all the filters and reading all the tickets. Why would I do such a thing? oh remember the 9000 labels I mentioned earlier? right well its very likely that they actually won't use any of them ... or the wrong ones ... or assign to the wrong person, so I have to go find them and fix them. If I don't, i'll get yelled at, because clearly it's my fault.

- Moving on from Jira. As some of you might have seen in your companies, if you use things like TestFlight, HockeyApp, AppCenter, BuddyBuild etc. that when you release a new app version for testing, each version comes with an automated change-log, listing ticket numbers addressed ...... yeah we don't do that. No we use this shitty service, which is effectively an FTP server and a webpage, that only allows you to host the new versions. Sending out those emails is all manual ... distribution groups?? ... whats that?

- Moving back to Jira. Can't even automate the changelog with a script, because I can't even make sense of the tickets, in order to translate that to a script.

- Moving on from Jira. Me and one of the remote testers play this great game I like to call "tag team ticketing". It's so much fun. Right heres how to play, you'll need a QA and a PM.

*QA creates a ticket, and puts nothing of any use inside it, and assigns to the PM.
*PM fires it back asking for clarification.
*QA adds in what he feels is clarification (hes wrong) and assigns it back to the PM.
*PM sends detailed instructions, with examples as to what is needed and assigns it back.
*QA adds 1 of the 3 things required and assigns it back.
*PM assigns it back saying the one thing added is from the wrong day, and reminds him about the other 2 items.
*QA adds some random piece of unrelated info to the ticket instead, forgetting about the 3 things and assigns it back.

and you just continue doing this for the whole dev / release cycle hahaha. Oh you guys have no idea how much fun it is, seriously give it a go, you'll thank me later ... or kill yourselves, each to their own.

- Moving back to Jira. I decided to take an action of creating a new project for my team (the mobile team) and set it up the way we want and just ignore everything going on around us. Use proper automation, and a kanban board. Maybe only give product a slack bot interface that won't allow them to create a ticket without what we need etc. Spent 25 minutes looking for the "create new project" button before finding the link which says I need to open a ticket with support and wait ... 5 ... fucking ... long ... painful ... unnecessary ... business days.

... Heres hoping my head continues to not have a bullet hole in it by then.

Id love to talk more, but those filters ain't gonna fix themselves. So we'll have to leave it here for today. Tune in again for another episode soon.

And remember to always practiseSafeHex

*wrestling commentator voice*
"In this weeks episode of encoding hell:
The iiiinnnfamous UTF-8 Byte Order Mark veeeersus PHP!"

For an online shop we developed, there is currently a CSV upload feature in review by our client. Before we developed this feature, we created together with the client a very precise specification, including the file format and encoding (UTF-8).

After the first test day, the client informed us, that there were invalid characters after processing the uploaded file.
We checked the code and compared the customer's file with our template.
The file was encoded in ISO-8859-1 and NOT as specified UTF-8.
But what ever, we had to add an encoding check, thus allowing both encodings from now on.

Well well well welly welly fucking well...

Test day 2: We receive an email from said client, that the CSV is not working, again.
This time: UTF-8 encoding, but some fields had more colums with different values than specified.
Fucking hell.
We tell the customer that.
(I was about to write a nice death threat novel to them, but my boss held me back)

Testing day 3, today:
"The uploading feature is not working with our file, please fix it."
I tried to debug it, but only got misleading errors. After about 30 minutes, at 20 stacks of hatered, I finally had an idea to check the file in a hex editor:

God fucking what!?!!?!11?!1!!!?2!!

The encoding was valid UTF-8, all columns and fields were correct, but this time the file contained somthing different.
Something the world does not need.
Something nearly as wasteful as driving a monster truck in first gear from NYC to LA.

It was the UTF-8 Byte Order Mark.
3 bytes of pure hell.
Fucking 0xEFBBBF.
The archenemy of PHP and sane people.

If the devil had sex with the ethernet port of a rusty Mac OS X Server, then 9 microseconds later a UTF-8 BOM would have been born.

OK, maybe if PHP would actually cope with these bytes of death without crashing, that would be great.

Template for every Silicon Valley episode be like:

1. Start solving problem from previous episode.
2. Creating some shit that solves the problem.
3. Initialize new problem.

Damn it. I'm still hooked!!

I usually find Fridays really exciting 'cause they mean the end of a long week of work and a nice weekend where I can just relax and chill or do whatever the fuck I want, And also because nothing really major happens regarding work happens on Fridays.
Till this Friday, my boss who I really respect and who I find a nice boss to work for starting complaining about the speed of an app we developed and comparing its speed with 2 other versions of that were built using different stack, different architecture and another environments. I explained that it's absurd to compare these and expect the same performance from 3 differents implementations.
He was not convinced and I just kept my mouth shut 'cause I don't want to explode in anger. Because of all Friday night sucked, felt all depressed, wanted to distract myself by watching a movie, but I didn't find anything that I liked, I remembered that a new episode of this series I watching will be coming out that night, when I went to my usual streaming website I didn't find it, and discovered that it'll be coming out on March 1st 😣.
I had no video games to play, didn't feel like coding. By then i realised that tonight will be another nigh where I would be crying myself to sleep... which happened.
I woke up this morning with a resolution that I will go out and do something fun.
Little did I know, my depression was still there, now it's 8pm, I spent the whole day in bed. I wish I had someone to talk to, I friends are all busy living and I didn't want to disturb them.
I have another chance to save this weekend by doing something on Sunday, otherwise next week will be a hard one with my current mental state.

Excuse any typos in my rant. I have no energy left.

The ticket system blokes - episode 3

So we always had and have very awful performance with our ticket system. You can't get anything to load in under ~4s normally. Now since it has gotten worse over the last weeks i decided to set aside a few hours to closely watch our SQL server.
After i identified a culprit that was hogging the CPU almost every 2 minutes i looked at other long running queries in the server and found out where exactly the 4s come from.

6 tables from various DBs. Sure, no problem.

Left Outer Join. Sure, why not.

Querying every fucking column in every fucking table explicitly adding up to a whopping 160 columns which they need not even 10% of. We're talking about session IDs, passwords, stock count, IBANs and all that stuff to show the work done on a ticket. Absolutely not.

So i extracted the query and reduced it to the stuff we need and the execution time went from 4 seconds to almost instant.

The funny thing is that their idea of performance optimization is throwing LIMIT around everywhere to get these monstrous queries under control.

So in the next few days I'll have an appointment with their lead programmer. I'm looking forwards to it.

So out of curiosity: does anyone know an SQL builder or toolset that does shit like
SELECT X AS [t0_c0],
SELECT Y AS [t0_c1],
SELECT Z AS [t1_c0],
and so on? I'd like to know how they got to this point.