Google sucks!
No, not as e-mail or for privacy reasons. Sure, that too, but it comes with "free" stuff.

It sucks because it's breaking every possible record in the worst, shittiest, most insanely stupid APIs and integrations out there on the entire fucking planet!

It is comically stupid!

Aside from their LOVE of hard-deprecating APIs every few months, requiring constant, time consuming maintenance of every tool that integrates deeply with Google services, some of their APIs, for expensive stuff, look like they've been written by Bobby McFartface from 7th grade.

Take a look at DoubleClick Search (their ad performance reporting tool, that sure does sound like one). To upload custom, additional data, you must pass in a ton of parameter, and they REQUIRE some of them to have a specific, hardcoded value. What's the point in passing that parameter then you dickheads?!
But fine, so you uploaded some stuff using the API. Now you want to delete everything and try again after you fixed a bug - well you fucking CAN'T! You can't delete stuff, you can only mark them as "deleted" using an update call.
Bulk operations? Fuck no!
Can I just add on top? Well of course not! That will raise a ton of exceptions. Same message should be transmitted using the PUT, not POST request, in order to edit.
Can I send everything to PUT? Of course not! You can't edit something that's not there, dummy!

Can I see what's there so that I can update it, and add what's missing?
Well of course not! Why on Earth would you need to see what information is in there after you uploaded it? Who needs that anyway?

Simply send, pray, and hope that everything will be fine (it will not).

Like holy fucking crap, it can't get any more stupid!
Google is a huge pile of idiots who feed on only a single cow - the search engine.

It's times like these when I think that Google right now is the worst thing that exists for everyone in tech. It's dragging everyone down with their monopolies everywhere and complete idiocy in managing them.

Once I had to do a 'hands on' pair programming session for a position I applied for... Together with the lead dev we would switch coding every 15 minutes It was somewhat of a horror story...

The assignment was to implement an password reset flow, connecting it to the api and then handling the entire password reset flow, in Angular becahs ye know has to be Angular...

After drafting the ui and setting up the click events, I wanted to hookup the api calls, but then it was time to switch around...

The fucktard dev first started to adjust my classmappings to be more in line with his preference, without touching the css classnames... Ok... Micro managing ... Check...

So after breaking the styles, he wrote the fetches to the api endpoints and that was his 15 minutes of shame...

I continued only to find out the endpoints we were using had errors in them and would not return anything workable...

The dev said he'd tested the endpoint before and it worked, but clearly it didn't...

After about an hour of going back and forth trying to get this to work he got a call from a client because server was down (surprise), he excused himself and had to prioritize on this, running out and leaving me there for the remaining morning ...

I just sat there waiting for the HR checkout talk, only to lean towards rejecting the position...

Fucking waste of time, and in the end the feedback was they doubted MY TECHNICAL SKILLS ... And wouldn't make me an offer 😂👍 nice story bro...

K THX BAI!

I just recently started my first job as a full stack programmer (still studying at university). I got assigned one month to code a complete front end to our api. Now, 4 days before release day the owner of the company makes breaking changes to the api.

Just. Beautiful.

Manager: "We can't have new releases breaking older versions of the mobile app!!!!! We'll lose all our customers!!!!"

fullStackChris: "That's fine, we can do API versioning, but it will take some time to implement, I'll have to be quite careful and write some tests to implement it. Probably 2-3 weeks..."

Manager: "NO WAY, THAT TIME ESTIMATE IS WAY TOO LONG, WE DON'T HAVE TIME FOR THAT!!!"

fullStackChris: "So how do you wanna support multiple versions of the app without doing any sort of versioning?"

Manager: "...we'll think of something!"

Dev: "..."

And with 99% certainty, I expect to hear this in a week or two:

Manager: "fullStackChris, we'd like to introduce you to the highly technical concept, API versioning. It's a way to version the API so we can support multiple versions of the application our customers use! It's amazing! Please implement this immediately so we can support multiple versions of the application!"

Sigh... each day managers learn a bit more how physical reality works... you can't have your cake and eat it too.

The lead dev left the company two weeks ago. His last hurrah involved committing a bunch of complicated code into our API. The code he writes is generally overly coupled but this particular code is INSANE.
It is so coupled that the tests for it almost mock the entire application end to end. I only found this heap of garbage when the deployment tests were hung after I made a simple change in a totally unrelated file. THEY DIDNT FAIL. JEST GOT INTO A STATE WHERE IT CANT CONCLUDE AND HAS NO ERROR MESSAGES. We are taking about entirely different parts of the code. As far apart in the code as it can get. It took six hours of playing Sherlock Holmes to figure out what was breaking.

He got the most junior developer to approve the garbage PR as well.

So here I am, skrewing around with the Google Authenticator app and the dodgiest base32 code generator I've ever built and generating a 56 char unique ID, and a 8 digit time based code.

WTF, all these products, services and logins that use 6 digit codes... and this fucking thing can handle 8 without breaking 😑

Now... to hook it into a QR code class... and spit out an image I can actually scan, without calling google charts api.

I can't say I've written one of those before 🙃

Why is it that pretty much zero package & framework maintainers understand semantic versioning?

1. If you do a complete rewrite of your package, but the resulting API is identical, you don't need to bump to the next major version. As a user, I'm thankful for your increased performance or cleaner internal code, but it doesn't really affect my update process.

2. If your package required some-framework 6.0.0, and now ALSO supports some-framework 7.0.0 but is still compatible with 6.0.0, you don't need to bump to the next major version. As a user, I can now upgrade the framework, and know that the package will keep working, but otherwise it doesn't really affect me.

3. Following your versioning along with the framework/language version is super annoying, especially if your library really doesn't need to differentiate between framework versions because it's not actually utilizing new framework functionality.

4. On the other hand, if you stop supporting a certain language, framework or shared library version, or change the public methods, exceptions, fields, etc, you MUST bump to a new major version.

Yet everyone gets this wrong.

For example, many of Laravel's underlying subpackages (for collections, filesystem, database, config, http, mail, etc) do not change their code in a breaking way, or do not even change at all between major framework versions.

Yet they follow along with the major framework version.

Now if someone makes a library "laravel-elasticsearch" which uses the support libraries and collections from laravel, they need to update their package to move along with the versions as well, and often they choose to number their library along with the framework in turn.

This means that to update the framework, you also need to update over 9000 dependencies.

FOR NO FUCKING REASON. THE ONLY CHANGE IN THOSE FUCKING DEPENDENCIES IS TO UPDATE COMPOSER.JSON TO BE COMPATIBLE WITH THE FUCKING FRAMEWORK.

Meanwhile, Laravel itself breaks repeatedly on minor/patch version updates, because breaking changes slip through their review process.

Ugh.

I've been so pissed off so many times, I thought I should divide them into categories.
- Pissed off at a fellow dev: I told him to use a constant instead of a hardcoded number.
He changed this: obj.method(3);
to this:
public static final int three = 3;
obj.method(three);
- Pissed off at management: I once got a $10 yearly raise.
- Pissed off at a client: They rejected our design proposal because the text was in spanish and they didn't speak spanish. It was lorem ipsum text.
- Pissed off at code: I once had to refactor a 500 line legacy jsp script with HTML, CSS, JS and Java completely intertwined.
- Pissed off at Twitter: They changed their API the day of our go-live, breaking all of their widgets, forcing us to move the go-live date and making me work an additional 8 hours after a week with almost no breaks.
- Pissed off at travel and logistics: They sent me to a hotel in Mexico City 2.5 hours away from the client's office.

Fun times...

A word of advice to framework authors:
If I am currently using v2.5.5 of a library, and I update to 2.8.6, I would expect to maybe have to update a few deprecated method calls here and there.
I do not expect the entire API to be completely different, with half the classes totally renamed and restructured. Breaking changes should go into a new major version plz

Facebook API...

Facebooks "graph" or API's in general fucking stink donkey dick.

Their implementation of oAuth is horrible.. 3 different tokens, which can be either short or long lived, for fetching a facebook page feed (the clients own facebook page)
To that you add a clientID and a ClientSecret.

Great... after painstakingly reading confusing documentation and itching your head... You get it to work.

Then they, without notice, makes a breaking change of deprecate an endpoint you were using.. Jesus..

And all the support you can get comes from a "community group" which may or may not reply with a generic link to their documentation...

This is a short tale that can be summed up as "oh fuck meee".

After finishing an API the night before I settled in for a day of bug fixes and tidy ups. Until slack went off.

The front end dev was getting an error, a code breaking error. After doing the standard process of request checking i went okay must be me. I find the script that is has the error and the line that it is failing at.

Que 2 hours of the full cycle of anger, sadness, pleading, and finally acepting that it had finally happened I had gone insane. The code was to documentation best practise correct and it still had the same error.

I the cheaked the DB on a whim and I found that my code was not wrong and it was doing exactly what I wanted the data however had a single record that was old and the schema had change juuussstt enoigh to break everything at that record. One 3 secound deletion later code ran perfectly.

A friend came to me whether i want to do a project on c++(someone asked him to find a c++ guy).
Me needing money didn’t refuse. Even though i am a Java developer with 0 skills on c++, but wanted to give it a try.
So project started, and it was about a plugin for rhinoceros app(3d graphics app).
The plugin was simple, had some views and some services to upload a file into s3 and some api calls, not something complex..
So i ended up working on the project together with my friend(web dev).
So long story short, we had a lot of issues, but considering we both had no knowledge on c++, we were really lucky to finish the product almost on time(3 days after).
Did no memory management even though i’ve read that we have to do that by our selfs and that c++ doesn’t have garbage collector.
But the plugin worked great even without garbage collector.
Had a lot issues with string manipulation, which almost drive me crazy.

PS: did a post here before taking the project, to ask whether it is a good idea to take the project or not, had some positive and some negative replies, but i deleted the post since i thought i was breaking the NDA i signed 😂😂

PS2: just finished OCAJP 8 last week with a great score😃

Our project at work goes live in 3 weeks.

The code base has no automated tests, breaks very often, has never had any level of manual testing

will not be releasing with any form of enforced roles or permissions in our first release now due to no time to enforce, however there is a whole admin api where you can literally change anything in our database including roles.

We also have teams in various countries all working separately on the same solution using microservices with shared nuget packages and they aren't using them properly.

Our pull requests are so big - as much as, 75 file changes - in our fe app that I can't keep up with it and I honestly have no idea if it even works or not due to no automated tests and no time to manually test.

We have no testing team, or qa team of any sort.

Every request into the system has to hit a minimum of 3 different databases via 3 different microservices so 1 request = 4 requests with the load on the servers.

We don't use any file streams so everything is just shoved in the buffer on the server.

Most of the people working on the angular apps cba to learn angular, no one across 2 teams cba to learn git. We use git so they constantly face problems. The guy in charge has 0 experience in angular but makes me do things how he wants architecturally so half the patterns make no sense.

No one looks at the pull requests, they just click approve so they may as well push directly to master.

Unfinished work gets put in for pull request so we don't know if the app is in a release state since aall teams are working independently, but on the same code base.

I sat down and tested the app myself for an hour and found 25 fe only issues, and 5 breaking cross browser issues.

Most of our databases are not normalised. Most of our databases make no sense. 99% of our tables have no indexing since there is no expertise with free time to do it.

No one there understands css properly. Or javascript.

Our. Net core microservices all directly use ef in the controller actions so there is no shared code there.

Our customer facing fe app is not dry because no tests so it was decided it was better this way.

Management has no idea on code state, it seems team lead is lieing to them about things like having any level of tests.

Management hire devs that claim to be experts but then it turns out they have basically no knowledge of what they were hired to do, even don't know what json is or the framework or language they are hired for, but we just leave them to get on with it and again make prs too big to review.

Honestly I have no hope that this will go well now but I am morbidly curious to watch. I've never seen anything like the train wreck that we are about to get experience.

-click-

MySQL: 'the table does not exist'

I just fucking made the temp table dude, that's what you literally do in step 1 .... how could it not....

-click-

MySQL 'Records: 1 Duplicates: 0 Warnings: 0'

Me: "Oh there we g---"

-click-

MySQL: 'the table does not exist'

Me: "Hey you just worked!"

-click-

MySQL: 'the table does not exist'

GOD DAMN IT

-click-
-click-
-click-
-click-
-click-
-click-

MySQL 'Records: 1 Duplicates: 0 Warnings: 0'

Me: Uh you're working now?

-click-

MySQL 'Records: 1 Duplicates: 0 Warnings: 0'

-click-

MySQL 'Records: 1 Duplicates: 0 Warnings: 0'

-click-

MySQL 'Records: 1 Duplicates: 0 Warnings: 0'

Guess that API just needed breaking in....

Finally finished the longest ticket I've ever worked on in my life. The ticket title and description was a pretty simple and straightforward one: "Upgrade from PHP 7.4 to 8".

If it was only so simple in real life. Our application is mostly done with API Platform framework, which is based on top of Symfony framework which is based on top of PHP language.

Once I did PHP 7 => 8 upgrade I needed to upgrade API Platform 2 => 3. But of-course that couldn't have been done as before that I needed to upgrade from Symfony 5 => 6.

This all was literally an equivalent of touching into a wasp nest - it took me a bit over 5 months and 800 hours of work and there was literally not a single source file left untouched.

In the process of all of this I've ran into literally dozen undocumented feature-breaking changes, broken backwards-compatibility promises and inside out architectural changes - from both the frameworks and the language itself.

Upgrading just one major version of anything SHOULD NOT be so hard. And to top it all up just to think I will need to do this again in a year or two..

Experiences like these really set my hate for time-based model of releases and the state of today's development in general.

We use a third party paid company to produce a service and give ongoing support for it, which all our revenue streams depend upon. They are shit and their service is shit. Here's how my conversation about testing went today.

Me: 'hey X wrote an integration test project for the service. It shows the service is broken 50% of the time. We should give their team access to it and have them run it as part of CI'

Colleague: 'They are too shit to setup CI'

PM: 'we are stuck with them so there is no point. It is what it is'

Boss: just ignores me. Not even a reply.

Some days later

Head of QA: 'Hey Dev and QA are broken'

Me: 'because their service is broken. I made so and so suggestion before but it was rejected. We will just have to accept Dev and QA are broken 50% of the time'

Head of QA: 'no we cant'

Me: 'ok so we should setup the tests to run by giving them access'

Head of QA: 'No we shouldn't. The tests can only be used by us and if they break it tells us so we can act on it, or choose not to'

Me: 'We would not want to act immediately on all our revenue streams breaking? Yes we can reverse engineer their client and fix errors as they occur, or we could just have them run the tests and a team our company pays for can stop adding breaking changes to their own API every other day. Right now it has been broken for 2 weeks.'

Head of QA: 'in an ideal world we would have an internal team so you're wrong'

Me: :)

I really don't understand how they can come to such a conclusion. Am I missing something or am I surrounded by total fucking idiots?

Colleagues of mine, even though they show signs of versioning their api by having "v1" in the URL, keep committing breaking changes to that version. 😂

I had the oppertunity to join a non profit organization to help them automate stuff instead of serving the army. One of their core applications got rewritten like a year ago from a terrible and very old Symfony stack to Laravel / React.

The guys who were in charge for the rewrite didn't really adapt the mindset of either MVC for Laravel nor the component idea behind React. There are a few controllers in the backend, but they sometimes have functions defined which would clearly belong in a model or service class. They rarely defined relationships on models, instead they're joining the tables together for the same effect. The frontend rendering mostly happens in for loops over the returned array from the API instead of breaking things down into little components. This ends in components which have sometimes over 1000 lines with super-nested logic in it.

But I did find my favorite piece of code today in of the controllers. Some many questions ...

For all the hate that Java gets, this *not rant* is to appreciate the Spring Boot/Cloud & Netty for without them I would not be half as productive as I am at my job.

Just to highlight a few of these life savers:
- Spring security: many features but I will just mention robust authorization out of the box
- Netflix Feign & Hystrix: easy circuit breaking & fallback pattern.
- Spring Data: consistent data access patterns & out of the box functionality regardless of the data source: eg relational & document dbs, redis etc with managed offerings integrations as well. The abstraction here is something to marvel at.
- Spring Boot Actuator: Out of the box health checks that check all integrations: Db, Redis, Mail,Disk, RabbitMQ etc which are crucial for Kubernetes readiness/liveness health checks.
- Spring Cloud Stream: Another abstraction for the messaging layer that decouples application logic from the binder ie could be kafka, rabbitmq etc
- SpringFox Swagger - Fantastic swagger documentation integration that allows always up to date API docs via annotations that can be converted to a swagger.yml if need be.
- Last but not least - Netty: Implementing secure non-blocking network applications is not trivial. This framework has made it easier for us to implement a protocol server on top of UDP using Java & all the support that comes with Spring.

For these & many more am grateful for Java & the big big community of devs that love & support it.

The moment when your ecommerce site relies completely in 3rd party vendor, and their API is changing without any heads up or announcement

Breaking out the big guns to wrap my head around a buggy API

So I made an update to my React Native app. I changed UI of a couple of screen, added a few animations here and there, refactored how my graphQL resolvers work in the backend(no breaking changes), changed how data gets loaded into the database etc.

It worked in dev so I figured hey let's deploy it. Today is(was because it's now 3am but more on that later) a national holiday so no one goes to work so no one will use my app so I have an entire day to deploy.

I started at 15:00(because i woke up at 13:00 lol). I tested the update once again in dev and proceeded to deploy it to prod. I merged backend to master, built docker images, did migrations on the db, restarted docker-compose with new images. And now for the app. I run ./gradlew assembleRelease and it starts complaining that react-native-gesture-handler is not installed. Ugh, rm -rf node_modules && yarn install. It worked. But now gradlew crashes and logs don't tell me anything. Google tells me to change a bunch of gradle settings but none of them work. Fast forward 5h, it's around 20:00 and I isolated the issue to, again, react-native-gesture-handler. They updated from 2.2.4 to 2.3.0 which didn't fucking compile. 2 more hours passed (now 22:00) and I got v2.3.1 working which fixed the problem in 2.3.0 but made my app crash on startup. YOUR FUCKING LIBRARY GETS 250K WEEKLY DOWNLOADS AND YOU DONT EVEN BOTHER CHECKING IF IT COMPILES IN PROD ON ANDROID?! WHAT THE FUCK software-mansion?

After I solved that, my app didn't crash. Now it threw an error "Type errors: Network Request Failed" every time I fetch my legacy REST API(older parts use rest and newer use graphql. I'll refactor that in the next update). I'll spare you the debugging hell i went through but another 5h passed. Its 3am. My config had misspelled url to prod but good for dev... I hate myself and even more so react-native-gesture-handler.

Motherfucker, do you even review your own code, never mind getting anyone else to do it?

"Hi" randomly added on a new line in the middle of a switch block, a syntax error, as the only change in a file?

Breaking two methods by misunderstanding which database object a variable identifies- but making no other change to those functions? And not adding permissions checks to the new API methods you added in that file?

Overwriting the email template that goes out to users who were added straight to the CRM, by reusing the same file for a template for users that have been invited to an event?

Adding your new fields to the old CRM sync code, again leaving me to figure it out, thereby leaving users' changes likely to be overwritten every morning?

And pushing this to master, supposedly tested, without a heads-up?

How often does your mum need to buy you a new box of crayons? Because these ones are chewed to pieces.

Suck my balls. Or rather don't, you probably don't know you're not meant to use your teeth.

I can work with Angular, even though it's pain in the but.
My current Angular job is actually the job with the first manager that had decent human values and ethics, I like my team, and yeah, what we building is shit. But it's only 30% shit because of Angular, another 30% are due to SAFe, and the rest is the usual stuff.
Still enjoy my job and respect my team.

But please do not expect me to pretend Angular is on a comparable level to React. Angular hasn't brought any actual innovation in most major versions but releases those breaking major updates still at least twice a year.
Ivy might be awesome, but only because Angular told the world 3 years ago also to have Ivy compatible compile targets for their libs/packages doesn't mean everybody cared.
And the ngcc, the awesome compatibility compiler, mutates node modules in place. So ne parallel stuff, no using yarn2 or pnpm.
At the same time, React brought so many innovations into the frontend world but is basically backwards compatible.

Not sure how the Angular partial compilation and whatever needs to go on works, but it seems like there's hardly anyone that really knows, so you can't use Vite or whatever other new tool.
And sure, if you're really good, you can write Angular without producing memory leaks.
But it's really hard. Do you know what's also quite hard: Producing memory leaks with React!

And for sure, Angular Universal, which isn't used by anyone, it feels like, will still be on a comparable level to an open source product that's used all over the world, builds the basis for an open source company, and is improved by thousand of issues day by day.

And sure, two kinds of change detection are a great idea. And yeah, pretending Angular comes with all included makes it worth it that the API is fucking huge and you're better of knowing nothing, because you have to read up things, than knowing quite a lot, since making assumptions and believing apis work in a similar way and follow similar contentions...

Whatever... I work with it. Like the time. Like the company, even my poss. But please don't expect my lying to you this was a good idea, or Angular is even remotely the same level of React.

Today, I have installed/uninstalled a combination of [windows 7, arch linux, dual-boot] a total of 9 times...

I wouldn't be surprised if my 120G SSD fails next week

It all started when I had to whip up a GUI-wrapped youtube-dl based program for a windows machine.

Thinking a handy GUI python library will get it done in no time, I started right away with the Kivy quick-start page in front of me.

Everything seemed to be going fine, until I decided it would be "wise" to first check if I can run Kivy on said windows machine.

Here I spent what felt like a day (5 hours) trying to install core pip modules for kivy.. only before realizing my innocent cygwin64 setup was the reason everything was failing, and that sys.platform was NOT set to "win32" (a requirement later discovered when unpacking .whl files)

"Okay.. you know what? Fuck........ This."
In a haze of frustration, I decided it was my fault for ever deciding to do Python on windows, and that "none of this would've happened if I were installing pip modules on a Linux terminal"...

I then had the "brilliant" idea of "Why don't I just use Linux, and make windows a virtual machine within, for testing."

And so I spent the next hour getting everything set up correctly for me get back to programming.... And so I did.

But uh... you're doing GUI stuff, right? -> Yeah...

And you uh.. Kivy uses OpenGL on windows, doesn't it? -> Yeah..?

OpenGL... 2.

-> Fuck.

That's when I realized my "brilliant" idea, was actually a really bad prank. Turns out.. I needed a native windows environment with up-to-date non-virtual graphics drivers that supported at least OpenGL2 for Kivy GUI programs!

Something I already had from square 1.

And at this point, it hurts to even sigh knowing I wasted hours just... making... poor decisions, my very first one being cygwin64 as a substitution for windows cmd.

But persistent as any programmer should be in order to succeed, I dragged my sorry ass back to the computer to reinstall windows on the actual hardware... again.

While the windows installer was busy jacking off all over my precious gigabytes (why does it need that much spaaace for a base install??? fuck.). I had "yet another brilliant idea" YABI™

Why not just do a dual-boot? That way, you have the best of both worlds, you do python stuff in Linux, and when it's time to build and test on the target OS, you have a native windows environment!

This synthetic harmony sounded amazing to the desperate, exhausted, shell of a man that I had become after such a back-breaking experience with cygwin

Now that my windows platter with a side of linux was all set-up and ready-to-go, I once again booted up windows to test if Kivy even worked.

And... It did!

And just as I began raising my victory flags, I suddenly realized there was one more thing I had to do, something trivial, should take me "no time" to do, being in a native windows environment and all.................... -.- (sigh)

I had to make sure it compiles to a traditional exe...

Not a biggy, right? Just find one of those py2exe—sounding modules or something, and surprisingly enough, there was indeed a py2exe—sounding module, conveniently named... py2exe.

Not a second thought given, I thought surely this was a good enough way of doing it, just gonna look up the py2exe guide and...

-> 3 hours later + 1 extra coffee

What do you meeeeean "module not found"? Do I need to install more dependencies? Why doesn't it say so in the DAMN guide? Wait I don't? Why are you showing me that error message then????

-------------------------------

No. I'm not doing this.
I shut off my computer and took a long... long.. break.

Only to return sometime the next day and end up making no progress, beating my SSD with more OS installs (sometimes with no obvious reason to do so).

Wondering whether I should give up Kivy itself as it didn't seem compatible with py2exe.. I discovered pyInstaller, which seemed to be the way Kivy wants exe's to be made on windows..

Awesome! I should've looked up how Kivy developers make exe's instead of jumping straight into py2exe land, (I guess "py2exe" just sounded more effective to me then)

More hours pass, and you'd think I'd have eliminated all of my build environment problems by now... but oh, how wrong you'd be...

pyInstaller was failing, and half the solutions I found online were to download some windows update KB32946..whatever...

The other half telling me to downgrade from Python 3.8.1 to Python 3.8.0000.009 (exaggeration! But you get the point)

At the end of all that mess, I decided it wasn't worth some of my lifespan, and that maybe.. just maybe.. it would've been better to create WINDOWS GUI with the mother fuc*ing WINDOWS API.

Alright, step 1: Get Visual Studio..
Step 2: kys
Step 3: kys again.

Facepalm Monday...
My collegue denies to provide breaking changes in our login API in a separate version to the other teams depending on it.

What is the reason for his stubborn rejection?
It's scrum. We haven't planned the effort for realising a versioning concept for our API.
Let's build it in the next sprint as a part of live deployment strategy.

The point he miss is that the ProductOwner wants his API change deployed during the next sprint.
Additionally, it is best practice, having a compatible, deployable product after each sprint, without any risks.
Furthermore, another best practice to provide your API is one URI without a version part holding the current development of the API. And URIs with a version part in it to keep a specific request/response structure and behavior.

What really grind my gears are sayings like 'if the other teams had well programmed their software, modifying our API won't have any effect on them'

C'mon dude. That's far from reality, as anybody knows.

I can't accept, we provide unprofessional API builds, as he is going to do.

So, i have to spend my time and energy to change his mind, together with other software-architects, planning the big thing API-Gateway *sigh*

Question:

in teams with backenders and frontenders, how do you guys coordinate introducing breaking changes to the API while the system is live ?

I learnt a lot of my java background from Minecraft modding. I even went back a couple years ago when I was learning python and used the Minecraft python api to make my learning a little more interactive.

I'm​ always trying to find cool ways to apply programming to my hobbies, and hacking / modding / breaking the games I play it's a hell of a lot of fun and drives me to learn more

"Here is a minor update to our library from 0.8.x to 0.9.0. - Oh, and it is breaking everything, because we changed the API."

o'rly?!

What on earth will happen on a major update?

Really, what is it with the fear of major release numbers? Just name it 1.0 and 2.0 if you break 1.0.

Better yet: just give it a new name. Thingy2 would totally do the job.

When depression is hovering your back, it sucks when a potentially project breaking API update from a third party appears with shiny new features and your team glamour that we should implement it soon, but internally you scream that will break so much stuff and clients will be mad at you

What the hell kind of tool is Gitlab? I just want to automatically backmerge hotfixes from master to development. Even fucking Bitbucket had a checkbox to enable this. But not Gitlab, no, you better create a pipeline job in your already unreadable, overcrowded pipeline yml, but oh, the checked out repo in the pipeline is a detached head and you cant push with the user that checks out there. So what, just use a project acess token which revokes after a year breaking your task and then switch origin amd branch manually. But your token-user can't push to protected branches, so create a merge request instead, which requires approvals, making the automated step no longer automated.
But dont worry, you can just use the gitlab api to overwrite the approval rules for this MR so it requires 0 approvals. But to do so you must allow everyone to be able to overwrite approval rules therefor compromising security.
And so you made a feature that should effectively be a checkbox a 40+ line CI job which compromises your repo security.

which nuthead of an architect is responsible for the way gitlab (and its CI) is designed?

When you're using openapi generators and stuff for generating SDK code and let "the architect" handle the data structure and nomenclature, don't you hate having to add 33 (I counted) models, most of which are just the same class with different name or one property apart from each other, serialization of which gives request body overhead 56-132x (actual calculated results depending on the model complexity) the size of actual data you want to send, just to add support for one endpoint that needs just one model that started this whole madness?
I just had to add this one top level model reference and this happened to me. Those 33 models are not including the ones I already had included in my project so they didn't have to import them again.
For the love of <your_belief_here /> and all that's holy, never ever agree on generating code based on openapi if the person responsible for that is unexperienced. It will do more harm than good, trust me.
Before we decided to go with generated SDK my compiled product was a bit over 30KB, and worked just fine, but required a bit of work on each breaking API change. Every change in the API requires now 75% of that work and the compiled package is now over 8MB (750KB of which is probably my code and actually needed dependencies).
Adding an endpoint handler before? Add url, set method and construct the body with the bare minimum accepted by the server
Now? Add 33 models (or more), run full-project find&replace and hope it will work with the method supplied by the generated code, because it's not a mature tech and it's not always guaranteed it will work.

Sometimes people want to be too smart. If you want to consume a handful different restful API, it might make sense to abstract away some common functionality in your client implementation — yet to assume they follow the same convention in how their URI is built is borderline insane.

All I wanted to do was to change one API to a newer version, and now the implementation breaks for at least two other because it was done in an Abstract class and now I have to untangle that mess.

In some cases code duplication wouldn't be that bad. Even if an otherwise unrelated API seemingly share the same contract, still assume it has its own contract. You never know how those API evolve and I proclaim they will evolve towards breaking your assumptions.

Why the fuck does jsoncpp change their file name with every version? Somebody who knows about this, please enlighten me.

TMW you make something good with much efforts and dedication, and then the API gets deprecated with breaking changes.

Is having breaking changes in an API at a very early stage normal? We have like every sprint at least 3 endpoints who have breaking changes 🤔 I mean, is having refactored code better then stability?

Don't you just hate it when there seems to be nothing but in some ways lacking solutions to a definite task in your capability arsenal? Or rather, I don't really know how I should feel about it... I've been developing this solution to receive a 3DES encrypted Azure Service Bus message, decrypting it and chewing the output XML down so as to be digestible to the PHP application whose API the message gets delegated to... but there just seems to be no perfect solution: subscribing to the event topic straight from the target app just... doesn't seem to work properly, a Python implementation.... well, let's just leave it at that... a Node.js implementation would require TS and completely rewriting a proprietary library with 100+ complex types - also, there's some hiccups with both the subscription and the decryption...

I started with an F# implementation (after deeming the PHP one flawed), and it seems it's still the best. But goddamn it I had problems with it on the dotnet core side of thing (decryption output incorrect), so I had to switch to dotnet framework... Now finally everything crucial is peachy, but I can't seem to be able to implement a working serialized domain model pipeline to validate the decrypted message and convert it to something easier to digest for the target application (so that I could use the existing API endpoint instead of writing a new one / heavily modifying the existing implementation and fear breaking something in the process...). I probably could do it in C#, I don't know, but for the love of Linus I'm not going to do it if I can avoid it, when implementing the same functionality I have now without the Dto and Domain type modules would take 3x LoC than the current F# implementation incl. the currently unused modules!

And then there's the problem of deployment... I have no idea what's the best way to deploy a dotnet framework module to an app completely based on MAMP running on a mostly 10yo AWS cloud solution. If I implemented a PHP or Node.js solution, it'd be a piece of cake, but... Phew, I don't know. This is both frustrating, overwhelming and exciting at the same time.

I just realized i didnt push my BREAKING api changes for at least 2 Weeks, tomorrow is the initial deployment and either my code or the teams code is not gonna work, still deciding which is more important <.<

It’s a huge nightmare to develop a React front-end when:
- you have to adapt Bootstrap 3/jQuery based components to React
- the “back-end” is a sparse collection of micro services with cryptic URLs and finding the correct name means searching on a laggy WSO2 API manager
- the documentation of said micro services can be outdated and that means wasting a lot of time trying requests on cURL rather than in doing actual development and continuously breaking your concentration
- sometimes the micro services just become unavailable altogether
- the back-end shuts down at
6PM everyday, usually when after I finally achieved a flow and I’m doing meaningful progress

Fellow Ansible developers. I'm talking to you.
Are you freaking high or simply your morning pills have some serious side effects?
How do you manage to introduce a number of regressions in every fucking major release? How on earth you feel comfortable in breaking API in a minor and even bug fix releases?
You need to get me right. I really like Ansible project but those things... I imaging you every other day as a bunch of hamsters trying to find an exit in a shitty labyrinth which you call the codebase.
If you will not stop to eat and smoke those things this would became a lot worse indeed.