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!!!

$work: Ey @Root, make this super simple thing.
$work: No, not like that.
$work: It also needs to do A, B, and C.
$work: No, not there. You should build it somewhere else, but I won't tell you where.
$work: You need to build out F and G, too.
$work: What do you mean you don't have the data? Just ask support drone #3. (who directs me to #2, and that one to #8 who doesn't know, and that one to #12 who won't answer)
$work: Why can't I do K, Y, or S? You should be able to infer these from the mind of whoever wrote the ticket by its wording, despite no mention of them whatsoever.
$work: Are you done yet? It's a super simple ask!

My Professor today, explaining data distribution in distributed systems:

"Imagine distributing Username data in subsets, such as A-C, D-F, G-I, J-L etc... And we have a lot of users with A .... *long pause* A bunch of assholes basically .."

Why dont people trust you?

I was hired to be an SQL developer, I don't actually get to do much development, normally doing something involving copying and pasting in Excel.

Some of our databases were running slow and we noticed some (a few hundred) indexes were in shit state.

I knocked up a couple of scripts, one to reorganise indexes that were up to a certain amount of fragmentation and one to rebuild the indexes

My boss wants them tested (they were several times in dev) we've had these for over 3 weeks, but she doesn't want to run them.

Instead of fixing hundred of indexes she decided I should contrate on fixing some historic data issues that are preventing 10 indexes from being rebuilt.

Now there are serious issues and the CTO is asking why the indexes haven't been fixed.

I could have done this nearly a month ago, but now it's turned into a huge fucki g deal, and no doubt they'll try and push it back on me

So I'm writing some multithreaded shit in C that is supposed to work cross-platform. MingW has Posix threads for Windows, so that saved already half of the platform dependency. The other half was that these threads need to run external programs.

Well, there's system(), right? Uhm yes, but it sucks. It's incredibly slow on Windows, and it looks like you can have only one system() call ongoing at the same time. Which kinda defeats the multithreaded driver. Ok, but there's CreateProcessA(), and that doesn't suck.

Fine, now for Linux. The fork/exec hack is quite ugly, but it works and is even fast. Just never use fork() without immediate exec(). First try under Cygwin... crap I fork bombed my system! What is this shit? Ah I fucked up the path names so that the external executable couldn't be run.

Lesson learnt: put an exit() right after the exec() in the path for child process. Should never be reached, but if it goes there, the exit() at least prevents a fork bomb.

Well yeah, sort of works under Cygwin, but only with up to 3 threads. Beyond that, it seems like fork() at some point gives two processes the same PID, and then shit hangs.

Even slapping a mutex around the fork and releasing it only in the parent process didn't help. Fork in Cygwin is like a fork in the ass. posix_spawn() should work better because it can be mapped more easily to the Windows model, but still no dice.

OK, testing under real Linux. Yeah, no issues with that one! But instead, I get some obscure "free(): invalid size" abort. What the fuck would that even mean?! Checking my free() calls: all fine.

Time to fire up GDB in the terminal! Put a catch on the abort signal, mh got just hex data. Shit I forgot to compile with -O0 and -g. Next try. Backtrace shows the full call trace, back to the originating line in my program - which is fclose() on a file.

Ahhh I remember! Under Linux, fclosing a file that is already closed makes the program crash. So probably I was closing it twice. Checking back.. yeah that's where it was.

Shit runs fast on several cores now!

My mom asked me to speed up her PC's startup process. I looked into it and and probably found the problem: G Data.

I told her she has to delete it, there is windows defender anyways (and virus scanners are just snake oil). She refused and said: But I bought it from the local IT shop. They said it is the best.

I said: Yeah because they cannot sell windows defender...

She still argued that she paid for it and wants to keep it.

That was were I said: then enjoy your slow startups.

She also said she feels more secure with it. 🙄

I would not even be able to work with that PC! Its not old at all, but the permanent scanning comes at a high cost and probably does not help much.

But she trusts that local it store guy more than me apparently.

!rant
I am on vacation from my full time job this week. I wanted to use this week to write a PoC for a potential customer of my side business. really interesting project for me.

potential customer is a window and door manufacturer and needs an application to manage their racks.

their ERP system already has a simple rack management but it is only useable in house.

they want the drivers to be able to scan racks they deliver to a customer with a native app and they want to have a webapp for the customers to see racks that are assigned to them as well as reporting a rack ready for collection. And that all needs to be in sync with their local ERP system.

as i am a .net guy i decided to go with the abp framework (because it got recommended to me) and xamarin for the native app part (because i have experience in this).

i have now spent 4 days implementing this and it has been so rewarding. the framework is so powerful and it's template saved me endless hours.

i even wrote a very basic connector service which synchronizes data between my app and the clients ERP system. Just one way until now because of time issue, but i learned to scaffold an ef core with db first. It is noticable that the ERP system is 2-tiered - meaning the clients directly talk to the db.

Tomorrow i will implement the xamarin client.

4 days just coding what i want to. choosi g my own velocity and making my own priorities without any interruptions or discussions and a bunch of new things to learn.

Probably wasted half a day because of stupidy (implemented some bugs) but fixing and learning is part of the journey and i lime that part, too.

i am so relaxed right now 😁 just wanted to share this without a real reason :P

An actual email thread:

G: I need this done

Me: okay. Whats the priority, when's the due date, who are the stakeholders, what's the sensitivity level of the data?

G: Priority is ASAP, due date is ASAP, stakeholders are ALL OF IT, and i want to be able to control the data sensitivity level

...

Those arent ANSWERS; they're REQUESTS!!!

How it started:

Need to replace in a lot of SQL files certain stuff...

find . -type f -iname '*.sql' -exec sed -i 's|new|old|g' {} \;

12 hours later that find executed a shell script containing roughly 120 lines of text pipelining.

The jolly of inconsistent workflows.

Different SQL format stylings... Makes fun when single line string replace needs to be extended to multiline RegEx handling. Or matching SQL comment configuration..

Different line endings. MacOS, Windows, Unix, Bukkake.

Different charsets / collations. Anyone wants latin1_swedish_ci... utf8... utf16... :/

Realizing some people even left sensitive data inside the SQL files (e.g. API Tokens..... Yayyyyyyy).

...

Ugh. It's never a one liner. It's never easy. -.-

I hate cleaning up messy shit.

Hello and welcome, to a presentation in which I will tell you my thoughts on the shortcomings of modern day computers and programming practices.

Computers are based on a very fundamental and old idea, folders, and files, a file is basically a concrete amount of data, whereas a folder is a group of files, and it comes from the real life concept of files and folders, now it might be quite obvious already that using a concept invented in 1898 by a guy called Edwin G. Seibels, might not be the best way for computers to function in the year 2020, but alas, it is.

Unless of course, you step into the world of a programmer.

A programmer’s world is much different, they use this idea of a data structure, or in simpler terms, an object. An Object is just like what you would think of as an object in your head, something with different properties that you can think about in different ways, for example your mobile phone, it has a battery percentage, it has a screen size, it has free space available. Programmers use these data structures to analyse data very quickly, like finding all phones with a screen size bigger than a certain size for example.

The problem is that programmers still use files and folders to create the programs that use these objects.

Consider this example.

Let’s say you want to create a virtual version of a drink bottle, consider what properties it will have, colour, volume, height, width, depth, material, etc..

As a programmer, you can leverage programming features and change the properties of a drink bottle directly, if you wanted to change the colour, you just say, drink bottle “dot” colour, equals blue, or red.

But if the drink bottle was represented as a file, all the drink bottles data would be inside the one file, so you would have to open the whole file, find the line or section of the file that has the colour data of the drink bottle, and select it, highlight it, delete what’s there, and type in your new value.

One way to explain this better is to imagine a folder that now represents the drink bottle, imagine adding a new file into that folder that represents each property I described before, colour, volume, etc.., well now, you could just open that folder, find the file for colour, either by looking with your eyes or you could do a file search in the folder for a file called colour, open it, and edit the value inside. This way of editing objects is the one that more closely represents the way programmers and a program itself interacts with objects inside a running programming language.

But the thing is, programmers don’t use the folder/file way of creating objects and putting them into programs, because it would be too cumbersome, they just create 1 file for an object, or have lots of objects in a file, and create all the objects in 1 file, and then run the program which creates the objects, then when they stop the program, it deletes the objects. So there is no actual link between the object in a file and the object that the program creates by reading the data from that file, if you change the object in your program, it does not get saved to the file.

So programmers created databases to house these objects, but there is still a flaw in databases, they are hard to interface with, and mostly databases are just used to send data or retrieve data from, programmatically, you can’t really browse a database the way you can browse the files on your computer. You can, but database interfaces are not made to be easily navigated the way files and folders are.

As it stands, there is no way to store objects instead of files on your computer and interact with them in complex ways the way programmers can inside the programs they create.

If the idea of an object became standard the way a file and folder is standard, I think it would empower human’s a great deal to express things far more easily and fluidly than they can today.

Thanks for reading.

Someone tell me should I just give up because I'm stupid and simple shit escapes me or tell me bro calm the fuck down the guy is full of shit...

Dude says he can't verify 3rd statement in a nested IF - elseif logic because the third check for a false condition is the True condition in the first 2 statements.

So

If (mode) = manual and then
Data(g) /= Status1
Or else Data(g) = Invalid
Then
Do this thing that sounds cool
Elsif
Data(g) = Status1
And then Data(g) /= Invalid
Then
Do something else equally cool
Elsif (mode) /= manual
and then Data(g) /= Invalid
and then Data(g) /= Status1
Then
Do some less cool stuff
end if

short: The admin with enough xp is ill, there is no one with xp with varnish is and after 1 restart varnish outputs only 503.
long: there original admin is ill but he gave me an project to migrate an typo3 installation to a new server. Thats ok.
Plan: I move 150 GB of data with rsync to the new server, let specialists do something and switch ips between the new and old and clear varnish with a restart.
Reality: +2 hours to migrate the data, because of false infos from the admin, 7 hours preparing the switch, 5 minutes switch, 3 hours to find out the F*****G varnish is the single point of failure. I and the t3 guys agree to see the next day what went wrong.

ALL HAPPENED TODAY!

Plan for tomorrow: speak with the boss to account the extra hours to that day so i dont get over 10 hours and debug that fucking varnish and delete some servers from another project from the backupsystem and monitoring.

!dev
I've finally been so agitated at G+ I need somewhere to just vent.
So for context. What I'm talking about is Google+, or more specifically, the Android app. The website is bad in its own way, but that's not here nor there. No opinions on the iOS version, as I simply REFUSE to touch iOS.

So anyways. The platform itself honestly is not bad. With competent developers behind it, and them actually listening to their dwindling fucking userbase, they could easily turn it into something successful, but the issue is that they just aren't
You see, it's almost like they change dev staff every 6 or so months. Why do I believe this? Because the GUI changes about that fucking often. They also have a history of forcing updates, but allowing you to use an older version, just horrifically slapping on a new and unwelcome skin. This isn't an isolated practice by any means, but it's by far the most prevalent here.

So, now a list of some of the issues the current version has:
-After about a week, the app becomes unstably slow, to the point of it taking about a minute to refresh your home feed, or an individual page.
-Searching is never good, always being slow and rarely giving you who you asked for.
-Transparency is non fucking existent. There isn't a development roadmap to speak of, and when something happens we get it second hand from staff in a "G+ help" community.

There is a solution for the first one, going and clearing the data/cache, but really, the end user shouldn't have to regularly do that. Not to mention the storage space Google apps IN GENERAL fucking take up. Why does Google Play Services regularly use 250MB? (For most people, this really isn't much. But when you only get to fucking use 4 GB of internal storage it's a giant fuck you.)

Bah, back to the topic at hand.
There isn't a good solution to searching, or for transparency at the moment.
The spam filter is awful as well. REGULARLY letting obvious spam pass, regularly blocking and filtering genuine users. It's real annoying that the Android app itself doesn't have support for seeing these flags outside of rooting through the settings a bit, but still. The web and iOS versions have this already.

Oh, it also completely lacks a dark mode like most Google apps for some fuckin reason.

That concludes my random 1:30 AM rant about something I have no ability to change, except hope in vain that someone who has the ability to change this forwards this to the developers of G+.

I need a better sleep schedule.

So I've been working with heavily data centric applications for 15 years or so. And I must say moving our data warehouse into the G Cloud was the worst idea ever. No tooling, everything is barely working, debugging is a nightmare.

If you are thinking about it, just don't.