I jump on an existing scala project.

git pull && sbt compile test

Tests are failing.

Me: "Hey team, the tests are failing."

Team member: "That cannot be. They were passing for the the last run."

Me: "Did you run them locally?"

Team member: "No, on Jenkins. It was fine."

I check Jenkins.

Me: "What do you mean it's fine. The last successful deployment was on the end of May."

Team member: "The Pull Request checker always went through successfully."

I check how our Jenkins tasks are configured. It's true that the Pull Request Checker runs successfully yet due to a "minor misconfiguration" (aka "major fuckup") the Pull Request Checker only tests a tiny subset of the entire test suite.

Team members were were fine if their Pull Request got the "Success" notification on bitbucket's pull request page. And reviewers trusted that icon as well.

They never checked the master run of the Jenkins task. Where the tests were also failing for over a month.

I'm also highely confused how they did TDD. You know, writing a test first, making it green. (I hope they were just one specific test at a time assuming the others were green. The cynic in me assumes they outsourced running the tests to the Jenkins.)

Gnarf!

Team member having run the tests locally finally realizes: "The tests are broken. Gonna fix them."

Wow. Please, dear fellow developers: It does not kill you to run the entire test suite locally. Just do it. Treat the external test runners as a safety net. Yet always run the test suite locally first.

So I have that custom-made wifi router I've built. And it uses a USB wifi adapter with AC (wifi5) capability - the fastest one I could find in AliExpress.

I set it up a while ago - the internet access works fine, although speeds are somewhat sluggish. But hey, what to expect from a cheapo on Ali! Not to mention it's USB, not a PCIe...

A few days ago I ran a few speedtest.net tests with my actual AC router and the one I've built. Results were so different I wanted to cry :( some pathetic 23Mbps with my custom router :(

This evening I had some time on my hands and finally decided to have an umpteenth look.

nmcli d wifi
this is what caught my eye first. The RATE column listed my custom router as 54Mbps, whereas the actual router had 195Mbps.

I have reviewed the hostapd configuration sooo many times - this time nothing caught my eye as well.

Googling did not give anything obvious as well.

What do we do next? Yes, that's right - enable debug and read the logs.

> VHT (IEEE 802.11ac) with WPA/WPA2 requires CCMP/GCMP to be enabled, disabling VHT capabilities

This is one of the lines at the top of the log. Waaaaiiitttt.. VHT is something I definitely want with ac -- why does it disable that??? Sounds like a configuration fuckup rather than the HW limitation! And config fuckups CAN be fixed!

Turns out, an innocently looking
`wpa_pairwise=TKIP`
change into
`wpa_pairwise=TKIP CCMP`

made a world of a difference!

:wq
!hostapd
connect to the hostapd hotspot and run that iperf3 test again, and... Oh my. Oh boi! My pants fell off -- the speed increased >3x times!

A quick speedtest.net test deems my custom router's download speeds hardly any worse than the speeds obtained using my LInksys!!

The moral of the story: no matter how innocent some configurations look, they might make a huge difference. And RTFL [read the fucking logs]

In the pic -- left - my actual router, right - my custom-built router with a USB wifi adapter. Not too shabby!

OMFG I don't even know where to start..
Probably should start with last week (as this is the first time I had to deal with this problem directly)..

Also please note that all packages, procedure/function names, tables etc have fictional names, so every similarity between this story and reality is just a coincidence!!

Here it goes..

Lat week we implemented a new feature for the customer on production, everything was working fine.. After a day or two, the customer notices the audit logs are not complete aka missing user_id or have the wrong user_id inserted.
Hm.. ok.. I check logs (disk + database).. WTF, parameters are being sent in as they should, meaning they are there, so no idea what is with the missing ids.

OK, logs look fine, but I notice user_id have some weird values (I already memorized most frequent users and their ids). So I go check what is happening in the code, as the procedures/functions are called ok.

Wow, boy was I surprised.. many many times..
In the code, we actually check for user in this apps db or in case of using SSO (which we were) in the main db schema..
The user gets returned & logged ok, but that is it. Used only for authentication. When sending stuff to the db to log, old user Id is used, meaning that ofc userid was missing or wrong.

Anyhow, I fix that crap, take care of some other audit logs, so that proper user id was sent in. Test locally, cool. Works. Update customer's test servers. Works. Cool..

I still notice something off.. even though I fixed the audit_dbtable_2, audit_dbtable_1 still doesn't show proper user ids.. This was last week. I left it as is, as I had more urgent tasks waiting for me..

Anyhow, now it came the time for this fuckup to be fixed. Ok, I think to myself I can do this with a bit more hacking, but it leaves the original database and all other apps as is, so they won't break.
I crate another pck for api alone copy the calls, add user_id as param and from that on, I call other standard functions like usual, just leave out the user_id I am now explicitly sending with every call.
Ok this might work.

I prepare package, add user_id param to the calls.. great, time to test this code and my knowledge..

I made changes for api to incude the current user id (+ log it in the disk logs + audit_dbtable_1), test it, and check db..
Disk logs fine, debugging fine (user_id has proper value) but audit_dbtable_1 still userid = 0.

WTF?! I go check the code, where I forgot to include user id.. noup, it's all there. OK, I go check the logging, maybe I fucked up some parameters on db level. Nope, user is there in the friggin description ON THE SAME FUCKING TABLE!!
Just not in the column user_id...

WTF..Ok, cig break to let me think..

I come back and check the original auditing procedure on the db.. It is usually used/called with null as the user id. OK, I have replaced those with actual user ids I sent in the procedures/functions. Recheck every call!! TWICE!! Great.. no fuckups. Let's test it again!
OFC nothing changes, value in the db is still 0. WTF?! HOW!?

So I open the auditing pck, to look the insides of that bloody procedure.. WHAT THE ACTUAL FUCK?!
Instead of logging the p_user_sth_sth that is sent to that procedure, it just inserts the variable declared in the main package..
WHAT THE ACTUAL FUCK?! Did the 'new guy' made changes to this because he couldn't figure out what is wrong?! Nope, not him. I asked the CEO if he knows anything.. Noup.. I checked all customers dbs (different customers).. ALL HAD THIS HARDOCED IN!!! FORM THE FREAKING YEAR 2016!!! O.o

Unfuckin believable.. How did this ever work?!

Looks like at the begining, someone tried to implement this, but gave up mid implementation.. Decided it is enough to log current user id into BLABLA variable on some pck..
Which might have been ok 10+ years ago, but not today, not when you use connection pooling.. FFS!!

So yeah, I found easter eggs from years ago.. Almost went crazy when trying to figure out where I fucked this up. It was such a plan, simple, straight-forward solution to auditing..

If only the original procedure was working as it should.. bloddy hell!!

When will I fuckin learn that
a) customers lie
b) customers are sloppy
c) customers are wrong
d) customers do not do their work (properly)
e) customers want us to do their (dirty) work
f) possibly all of the freakinly above?! + khm....

They will fuckin aaaalwaaaays say sth is not working after the update..
And I will alwaaaays assume I fucked up something..even if I didn't touch that part of the code/data..
And almost aaaaalways it turns out that the bug they complain about is how the system worked (or didn't work) before the update and/or some fuckup from their side..

Anyhow, I rushed over, grabbed the files went testing in dev..wtf, output is different, mine is ok, theirs is..wtf is that shit?!

Transfer newly built dll to test..same shit as on prod..wtf?! How?!

I assumed they have thing A correctly linked to thing B.. ofc thing A was linked to thing C in their case and in another case (our test) to correct thing B..
I got chillies when grabbing files, that
I should have tripple checked that they didn't fuck up something on the link part, but I just assumed they know what they were doing & that they checked they linked correct files with correct content already, before being pissy that the update fucked up things.. riiiight!! :/

I wanted to find solutions to this fuckup asap so I disregarded my gut feeling..yet again!! Fuuuck!
I've spent too much time trying to find ways to fix a bug that wasn't even a real bug to begin with.. :/

Fuuuuuck!!

So yeah, always treat the customers like they are 3yrs old & have no clue what they are doing & check exactly wtf they were indeed trying to do..it will save you time & nerves..

And note to self: reread this shit daily!! And imprint it in your brain that everything is not always your fault!!

Fuck...
I'm not getting that job then.

So I just had one of those interview coding tests on hacker rank and screwed it up big time.

I'm a C# guy and it was a Java position. I worked with Java, like 10 years ago, and they're pretty similar so I brushed up over the last week when I had free time.

Absolutely blew it. It's not like it was hard, I just got into one question (of 6) and it ate up all of my time. The task was simple, make a JSON call, read the data, check if you need more calls, pull out a data field from all the concatenated results and return it in a sorted list. ONE HOUR it took me. A combination of not knowing the API well enough, simple syntax errors and relatively slow compilation.

Godammit.

The next question was implement an Object hierarchy but since I'd run out of time, all I got was the class declarations before the timer ran out.

fuck, fuck, fuck.

I guess the test did it's job and weeded out someone who can't contribute to the team...

Language Warning

There year is 2019, and down here in Australia we have a fuckup called the NBN (National Broadband Network). We all have to switch to it, from ADSL2. Promised dreams and joys, only to be fucked again and again.

Paying $89 AUD per month for 25mbps, however only getting 0.61mBITSps at times. (Literally can see it at home.danferg.com)

Call up support. "Have you rebooted your router?" Yes. "Have you done a test?" Yes. "Can you send me the screenshot?" No, I did it through the CLI. "... What's a CLI?" Fuck.

"So.. have you got a test?" Yeah, I have 7000 of them, from each minute of the day for the past while. *Silence* "oh.... Ok well, we'll have to see if you qualify for a discount" DISCOUNT?! GIVE ME MY FUCKING MONEY BACK.

gradle is infuriating.

firstly there are so limited resources to understand how it's building a java/android code. everything happens by magic and hit+trial

secondly the plugins and the tasks works in mysterious ways. sometime they work when applied in the project root's gradle file, other times they work when applied in module's gradle file, nd other times they need configuration at both levels.

then there are gradle tasks like build ,test, assemble , clean etc. these are less of an action and more of an alias to run a bundle of actions.

then we have 3rd party plugins which attach themselves to these "fat-actions" and run before/after them

and finally we have the fuckup from the java world where the only available code coverage plugin is jacoco and IT FUCKING SUCKS!!! it is a test environment plugin, it should impact test tasks , but somehow it's fucking with the assemble taskin such a manner, that the jars ans aar files generated via plugin are giving runtime errrors. yes , runtime! as if we are back in the messed up js world of "everything is good unless running live"

even if it was a compile time eeror, i would have considered. but runtime?!! fucking runtime error?! i barely understand this shit, there is absolutely no info available as to which classes are being used to create a build and how, and i am supposed to fix this? wtf?!

This is a continuation of my previous rant about admob being not very informative when it comes to invalid traffic and the resulting restriction in ad delivery.

I then wanted to use admob mediation to hang in facebook ads. My app is written with Xamarin.Forms.

So first I needed to make some facebook configuration - create an account, let my app review, create some ad placements and other shit. I came to the point where I had to put in a link to my privacy policy and the link could not be accepted due to some SSL fuckup -.-'

I then found out that there is an issue with my SSL Chain. With the help of whatsmychaincert.com I solved that issue. Little side note here: I have limited knowledge of that stuff and my cousin helped me set up my homepage so I had no idea what I was doing. Did a snapshot and luckily I did not needed that as everything worked :)
This took me around half an hour just so I can paste the fucking link to activate my app in facebook developer portal.

After that I made the whole mediation configuration shit - not an issue as google documented this quite well but it took some time.

Now comes the shitty part. To use admob mediation you need adapters to the other ad network. I found a nuget package with exactly what I needed just to find out that it is outdated. So I pulled the repo and saw that this thing is an aar binding library. Never did that stuff so I read some docs again. Updated the package and consumed it in my app.

The google docs then said "Use this mediation test shit to check if you did everything correct before going prod" - aar binding nr. 2 (but I am now familiar with that :P). This thing then told me that facebook ads could not be loaded because the SDK version is outdated -.-' SDK version comes from another nuget package which is referenced by the first aar thingie. I tracked that thing back to a repo where I found out that they are indeed totally behind. So I downloaded the aar, made a binding lib and bound that to my first aar binding lib as that depends on this.

Put that all back in my app - tested mediation and fucking finally after 6 hours everything comes together! all lights are green and things work.

Sorry if this is not quite a rant but it was quite a journey and I just had to share it.

My answer to their survey -->

What, if anything, do you most _dislike_ about Firebase In-App Messaging?

Come on, have you sit a normal dev, completely new to this push notification thing and ask him to make run a simple app like the flutter firebase_messaging plugin example? For sure you did not oh dear brain dead moron that found his college degree in a Linux magazine 'Ruby special edition'.

Every-f**kin thing about that Firebase is loose end. I read all Medium articles, your utterly soporific documentation that never ends, I am actually running the flutter plugin example firebase_messaging. Nothing works or is referenced correctly: nothing. You really go blind eyes in life... you guys; right? Oh, there is a flimsy workaround in the 100th post under the Github issue number 10 thousand... lets close the crash report. If I did not change 50 meaningless lines in gradle-what-not files to make your brick-of-puke to work, I did not changed a single one.

I dream of you, looking at all those nonsense config files, with cross side eyes and some small but constant sweat, sweat that stinks piss btw, leaving your eyes because you see the end, the absolute total fuckup coming. The day where all that thick stinky shit will become beyond salvation; blurred by infinite uncontrolled and skewed complexity; your creation, your pathetic brain exposed for us all.

For sure I am not the first one to complain... your whole thing, from the first to last quark that constitute it, is irrelevant; a never ending pile of non sense. Someone with all the world contained sabotage determination would not have done lower. Thank you for making me loose hours down deep your shit show. So appreciated.

The setup is: servers, your crap-as-a-service and some mobile devices. For Christ sake, sending 100 bytes as a little [ beep beep + 'hello kitty' ] is not fucking rocket science. Yet you fuckin push it to be a grinding task ... for eternity!!!

You know what, you should invent and require another, new, useless key-value called 'Registration API Key Plugin ID Service' that we have to generate and sync on two machines, everyday, using something obscure shit like a 'Gradle terminal'. Maybe also you could deprecate another key, rename another one to make things worst and I propose to choose a new hash function that we have to compile ourselves. A good candidate would be a C buggy source code from some random Github hacker... who has injected some platform dependent SIMD code (he works on PowerPC and have not test on x64); you know, the guy you admire because he is so much more lowlife that you and has all the Pokemon on his desk. Well that guy just finished a really really rapid hash function... over GPU in a server less fashion... we have an API for it. Every new user will gain 3ms for every new key. WOW, Imagine the gain over millions of users!!! Push that in the official pipe fucktard!.. What are you waiting for? Wait, no, change the whole service name and infrastructure. Move everything to CLSG (cloud lambda service ... by Google); that is it, brilliant!

And Oh, yeah, to secure the whole void, bury the doc for the new hash under 3000 words, lost between v2, v1 and some other deprecated doc that also have 3000 and are still first result on Google. Finally I think about it, let go the doc, fuck it... a tutorial, for 'weak ass' right.
One last thing, rewrite all your tech in the latest new in house language, split everything in 'femto services' => ( one assembly operation by OS process ) and finally cramp all those in containers... Agile, for sure it has to be Agile. Users will really appreciate the improvements of your mandatory service.