Join devRant
Do all the things like
++ or -- rants, post your own rants, comment on others' rants and build your customized dev avatar
Sign Up
Pipeless API
From the creators of devRant, Pipeless lets you power real-time personalized recommendations and activity feeds using a simple API
Learn More
Search - "corner cases"
-
In may this year, the new mass surveillance law in the Netherlands went into effect. Loads of people were against it with the arguments that everyone's privacy was not protected well enough, data gathered through dragnet surveillance might not be discarded quickly after the target data was filtered out and the dragnet surveillance wouldn't be that 'targeted'.
They were put into the 'paranoid' corner mostly and to assure enough support/votes, it was promised that:
- dragnet surveillance would be done as targeted as possible.
- target data would be filtered out soon and data of non-targets would be discarded automatically by systems designed for that (which would have to be out in place ASAP).
- data of non-targets would NOT be analyzed as that would be a major privacy breach.
- dragnet surveillance could only be done if enough proof would be delivered and if the urgency could justify the actions.
A month ago it was already revealed that there has been a relatively (in this context) high amount of cases where special measures (dragnet surveillance/non-target hacking to get to targets and so on) were used when/while there wasn't enough proof or the measures did not justify the urgency.
Privacy activists were anything but happy but this could be improved and the guarantees which were given to assure privacy of innocent people were in place according to the politicians... we'll see how this goes..
Today it was revealed that:
-there are no systems in place for automatic data discarding (data of innocent civilians) and there are hardly any protocols for how to handle not-needed or non-target data.
- in real life, the 'as targeted dragnet as possible' isn't really as targeted as possible. There aren't any/much checks in place to assure that the dragnets are aimed as targeted as possible.
- there isn't really any data filtering which filters out non-targers, mostly everything is analyzed.
Dear Dutch government and intelligence agency; not so kindly to fuck yourself.
Hardly any of the promised checks which made that this law could go through are actually in place (yet).
Fuck you.29 -
Well... I had in over 15 years of programming a lot of PHP / HTML projects where I asked myself: What psychopath could have written this?
(PHP haters: Just go trolling somewhere else...)
In my current project I've "inherited" a project which was running around ~ 15 years. Code Base looked solid to me... (Article system for ERP, huge company / branches system, lot of other modules for internal use... All in all: Not small.)
The original goal was to port to PHP 7 and to give it a fresh layout. Seemed doable...
The first days passed by - porting to an asset system, cleaning up the base system (login / logout / session & cookies... you know the drill).
And that was where it all went haywire.
I really have no clue how someone could have been so ignorant to not even think twice before setting cookies or doing other "header related" stuff without at least checking the result codes...
Basically the authentication / permission system was fully fucked up. It relied on redirecting the user via header modification to the login page with an error set in a GET variable...
Uh boy. That ain't funny.
Ported to session flash messages, checked if headers were sent, hard exit otherwise - redirect.
But then I got to the first layers of the whole "OOP class" related shit...
It's basically "whack a mole".
Whoever wrote this, was as dumb and as ignorant to build up a daisy chain of commands for fixing corner cases of corner cases of the regular command... If you don't understand what I mean, take the following example:
Permissions are based on group (accumulation of single permissions) and single permissions - to get all permissions from a user, you need to fetch both and build a unique array.
Well... The "names" for permissions are not unique. I'd never expected to be someone to be so stupid. Yes. You could have two permissions name "article_search" - while relying on uniqueness.
All in all all permissions are fetched once for lifetime of script and stored to a cache...
To fix this corner case… There is another function that fetches the results from the cache and returns simply "one" of the rights (getting permission array).
In case you need to get the ID of the other (yes... two identifiers used in the project for permissions - name and ID (auto increment key))...
Let's write another function on top of the function on top of the function.
My brain is seriously in deep fried mode.
Untangling this mess is basically like getting pumped up with pain killers and trying to solve logic riddles - it just doesn't work....
So... From redesigning and porting from PHP 7 I'm basically rewriting the whole base system to MVC, porting and touching every script, untangling this dumb shit of "functions" / "OOP" [or whatever you call this garbage] and then hoping everything works...
A huge thanks to AURA. http://auraphp.com/
It's incredibily useful in this case, as it has no dependencies and makes it very easy to get a solid ground without writing a whole framework by myself.
Amen.2 -
I think I'm falling in love. With TDD.
I used to be very skeptic about it. You know, the usual reasons: it takes longer to deliver, constant "flow" interruptions, etc, etc. But ever since I've tried it I'm nothing but happy about my choice :)
I'm moving forward, I'm not making any regressions, I'm no longer afraid to make any changes in my code as I know tests will show what exactly I break,.. And most importanty, I have all use-cases with corner-cases defined and "explained" in the code... No more do I have to search in Confluence for how this exact scenario should behave. Everything is here. Everything's in the tests.
It's amazing!
Yeah, it DOES take longer to deliver so if you're hardcore Agile living by "Ship it as soon as it compiles" TDD might be too slow. But if you prefer knowing when your code is covering all the use cases w/o any errors -- TDD is the way.12 -
I fucking hate chained methods. Ok, not all of them. Query things like array.where.first... that stuff is ok.
Specially if it's part of the std lib of a lang, which would be probably written by a very competent coder and under scrutiny.
But if you're not that person, chances are you'll produce VASTLY inferior code.
I'm talking about things like:
expect(n).to.be(x).and.not(y)
And the reason I don't like it is because it's all fine and dandy at first.
But once you get to the corner cases, jesus christ, prepare to read some docpages.
You end up reading their entire fucking docs (which are suboptimal sometimes) trying to figure if this fucking dsl can do what you need.
Then you give up and ask in a github issue. And the dev first condescends you and then tells you that the beautiful eden of code he created doesn't let you do what you want.
The corner cases usually involve nesting or some very specific condition, albeit reasonable.
This kind of design is usually present in testing or validation js libraries. And I hate all of those for it.
If you want a modern js testing lib that doesn't suck ass, check avajs. It's as simple as testing should be.
No magic globals, no chaining, zero config. Fuck globals forced by libs.
But my favorite thing about it that is I can put a breakpoint wherever the fuck I want and the debugger stops right fucking there.
Code is basically lines of statements, that's it, and by overusing chaining, by encouraging the grouping of dozens of statements into one, you are preventing me from controlling these statements on MY code.
As an end dev, I only expect complexity increases to come from the problems themselves rather than from needlessly "beautified" apis.
When people create their own shitty dsl, an image comes to my mind of an incoherent rambling man that likes poetry a lot and creates his own martial art, which looks pretty but will get your ass kicked against the most basic styles of fighting.
I fucking hate esoteric code.
Even if I had to execute a list of functions, I'd rather send them in an array instead of being able to chain them because:
a) tree shaking would spare from all the functions i didn't import
b) that's what fucking arrays are for, to contain several things.
This bad style of coding is a result of how low the barrier to code in higher level langs are.
As a language or library gets easier to use you might think that's a positive thing. But at the same time it breeds laziness.
Js has such a low learning curve that it attacts the wrong kind of devs, the lazy, the uninspired, the medium.com reader, the "i just care about my paycheck" ones.
Someone might think that by bashing bad js devs I'm trying to elevate myself.
That'd be extremely stupid. That's like beating a retarded blind man in a game and then saying "look, I'm way better than this retarded blind man".
I'm not on a risky point of view, just take a stroll down npmjs.com. That place is a landfill. Not really npm's fault, in fact their search algorithm is good.
It's just the community.
Every lang has a ratio of competence. Of competent to incompetent devs.
You have the lang devs and most intelligent lib devs at the top. At the bottom you have the bottom.
Well js has a horrible ratio. I wouldn't be shocked to find out that most js devs still consider using import or await the future.
You could say that js improved a lot, that it was way worse beforr. But I hate chaining now, and i hated back then!
On top of this, you have these blog web companies, sucking the "js tutorial" business tit dry, pumping out the most obscenely unprofessional and bar lowering tutorials you can imagine, further capping the average intelligence of most js devs.
And abusing SEO while they're at it, littering the entire web with copy paste content.2 -
2 things I'm working on now:
#1 a personal project I am hoping to commercialize and turn it into my moneymaker. Hoping it'd at least be enough to pay the bills and put food on my table so I could forget 9/5 for good. But it has a potential of becoming a much, MUCH bigger thing. This would need the right twist tho, and I'm not sure if I am "the right twister" :) We'll see.
#2 smth I'm thinking of opensourcing once finished -- a new form of TLS. This model could be unbreakable by even quantum computing once it's mature enough to crack conventional TLS. I'm probably gonna use md5 or smth even weakier - I'm leveraging the weakness of hashing functions to make my tool stronger :)
I mean how long can we be racing with more powerful computers, eh? Why not use our weakneses to make them our strengths?
Unittests are already passing, I just haven't polished all the corner-cases and haven't worked out a small piece of the initialization process yet. But it's very close6 -
Holy crap, I can't take it anymore.
I know that user acceptance testing is supposed to be done by the end user but it's as if they entirely skipped UNIT TESTING and QUALITY ENGINEERING.
Does their API work? Yes. It does.
Are their endpoints working? Sort of... why are query parameters required again?
Is it good overall? No, there are CORNER CASES ALL OVER THE PLACE (are they even still corner cases at this point?). It feels like it was made by amateurs!
Why am I doing quality testing on their services??? holy crap, they should pay ME for doing this1 -
Sorry, need to vent.
In my current project I'm using two main libraries [slack client and k8s client], both official. And they both suck!
Okay, okay, their code doesn't really suck [apart from k8s severely violating Liskov's principle!]. The sucky part is not really their fault. It's the commonly used 3rd-party library that's fucked up.
Okhttp3
yeah yeah, here come all the booos. Let them all out.
1. In websockets it hard-caps frame size to 16mb w/o an ability to change it. So.. Forget about unchunked file transfers there... What's even worse - they close the websocket if the frame size exceeds that limit. Yep, instead of failing to send it kills the conn.
2. In websockets they are writing data completely async. Without any control handles.. No clue when the write starts, completes or fails. No callbacks, no promises, no nothing other feedback
3. In http requests they are splitting my request into multiple buffers. This fucks up the slack cluent, as I cannot post messages over 4050 chars in size . Thanks to the okhttp these long texts get split into multiple messages. Which effectively fucks up formatting [bold, italic, codeblocks, links,...], as the formatted blocks get torn apart. [didn't investigate this deeper: it's friday evening and it's kotlin, not java, so I saved myself from the trouble of parsing yet unknown syntax]
yes, okhttp is probably a good library for the most of it. Yes, people like it, but hell, these corner cases and weird design decisions drive me mad!
And it's not like I could swap it with anynother lib.. I don't depend on it -- other libs I need do! -
So my colleagues don't proactively do PR reviews, even when they're assigned to them, unless asked for the n-th time by me or the PM.
In the meantime other PRs get merged, I need to merge master in my branch, and solve conflicts every time so that it is eligible for merging.
Additionally, while reviewing, the requierements engineer realizes that there were corner cases not initially covered in the ticket description, so I get review findings about covering these.3 -
Fuck you android studio, fedora, and my incompatibility with you. Maybe I'm just a stupid user or I'm forever treading on corner cases, but all I'm trying to do IS GET STUDIO TO SEE WHERE THE JDK IS WTF.4
-
Some common themes I have observed:
1. Why is this API using XML? It should be JSON.
2. Why are they using old shit? It should be new shit.
3. Why are they mixing new shit with old shit?
4. Why is the technology being used not fitting my sense of perfection and order?
What happens when JSON is the old shit? What happens when a new format is adopted that solves some corner case issues with JSON? What if those corner cases were solved by using XML, or a mix of XML and JSON? What happens when all the new shit is the old shit?
We will still be using the new old shit and listening to fucking noobs complain about mixing, matching, and abusing everything ever written. Oh JSON, how quaint and limited. Why did anyone ever use that? Fucking senior devs...10 -
* Gets handed additions to current software platform (web)
* Gives back estimte of time after meeting with everyone and making them understand that once the testing phase of the project is reached there will be no changes, tests should be exhaustive and focus on SAID FUNCTIONALITY of the new additions. NO CHANGES OR ADDITIONS AT THIS POINT IN TIME
* All directives, stakeholders, users etc agreed on my request and spend an additional hour thinking of different corner and edge cases as provided by me in case they can't think of them (they can't, because they are fucking stupid, but I provided everything)
* Boss looks irritated at their lack of understanding of the scope and the time needed, nods in approval after he sees my entire specification, testing cases, possible additions to the system etc
* All members of the committee decide on the requirements being correct, concrete and proper.
* Finish the additions in a couple of weeks due to the increased demand for other projects, this directly affects the user base, so my VP and Director make it a top priority, I agree with their sentiment, since my Director knows what he is doing (real OG)
* I make the changes, test inside of my department and then stage for the testing environment. Everything is ready, all migrations are in order, the functionality is working as proper and the pipeline for the project, albeit somewhat lacking in elegance is good to go.
* Testing days arrive
* First couple of hours of test: Oh, you know what, we should add these two additional fields, and it would be good if the reporting generated by the system would contain this OTHER FORMAT rather than this one.
* ME: We stated that no additions would be done during the testing environment, testing is for functionality, not to see if you can all think of something else, even then, on June 10 I provided a initial demo and no one bothered to check on it on say something.
Them: Well, we are doing it now, this is what testing is for.
Me: Out of this room, the software engineer is me, and I can assure you, testing is not for that. I repeatedly stated that previously, I set the requirements, added corner cases, tables charts everything and not one single one of you decided to pay attention or add something, actually, said functionality you are requesting was part of one of my detailed list of corner cases, why did you not add it there and then before everything went up?
Them: Well I didn't read it at the time (think of the I in plural form since all of these dumb fucks stated the same)
Then my boss went on a rampage on their dumbasses.
I fucking hate software development sometimes.
Oh well. Bunch of fucking retards.4 -
When an interviewer asks, what are the steps that you take usually when given a task to do something, what do you guys say?
I said, I devote 10-20% of the time to understand the given problem - sit and explore all possible scenarios to handle, then develop a brute Force approach, improve the approach to make it more efficient, see if it handles all edge and corner cases, then write test cases for it.
( I'm thinking, the process doesn't vary a lot for most of the people?, Except in TDD - one'll write the test cases first)
How would you answer this question?
I have this feeling that I messed up something 🤔8 -
i often do tech support in chat rooms in my free time (because i like spreading good will,) so here's a tech horror story
"""
"hey, can you help me fix something?"
sure?
"so i dug my old XP machine out of my closet and replaced the bad Ethernet card with a different one and when i plug in the ethernet cable the PC bluescreens."
# oboi
did you install the drivers? Sounds like it needs drivers
"no"
then install them
"no"
why not?
"it doesn't need any"
why do you say that?
"it said \"This device is set up and ready to use.\" in the balloon in the corner"
it has generic drivers to deal with devices before the real drivers can be found
"shouldn't they work?"
some devices need the extra support provided by the intended drivers, so the generic ones cause issues in those cases
"ok, well, where do I find them?"
do you have a model number?
"yes, it's " # scrubbed for... privacy? i dunno
gimme a few minutes
<insert 45 minutes of aggressive Googling for (str(DEVICE_MODEL_NUMBER) + " xp drivers")>
alright i have the drivers, go here:
# again, removed for... idk.
"they don't work"
# oh here we go
why not?
"These drivers are not compatible with your system architecture."
what version of XP are you using?
"XP Pro"
x86 or x64?
"x64"
# fucking...
ok so this is gonna get real complicated real fast: use x86 XP or I can't help you, none exist for x64 XP.
"oh ok"
<User left the IRC channel.>
"""4 -
The reason I don't use Linux on my desktop is its hobby of saying “fuck off” spontaneously and without warning when I need it the most.
A designer friend shares his After Effects project and asks to export that to Lottie? Fuck off.
Your Android phone decided to brick itself with an OTA update (yep, happened to me, thanks Sony), and you need to unbrick it? Fuck off!
A musician friend wants to connect his audio card (that of course has no Linux drivers) and record some bass riffs? Tell him to fuck right off mate.
Your boss suddenly asks you to check an MS Access file for him as he's en route to an important meeting? Yep, you guessed it — fuck off.
Your government now requires your tax papers to have digital signatures? Fuck off, it only works for Mac and Windows.
Want to connect an old digital camera? Would you please fuck off?
I know I'm gonna get heat from Linux fanboys, especially on this platform. After all, a designer should know how to export to Lottie if he's a real designer, you should've bought a better phone, your friend should've had his laptop with him, your boss should've used open source tools instead of MS Access… Wait, he was tasked that from above? Then his boss should've used open source tools! Government mandates digital signatures? Well, tell them to port that to Linux! Start a riot! Get a better government! Move to a better country! Digital cameras? Who uses them in 2024, especially old ones! Are you some kind of hipster?
I know preparing for corner cases is bona fide premature optimization, but that's the whole point — with Mac or Windows, you don't have to prepare at all. You always have options. With Linux, your number one option is to have Windows handy if need be.
Linux works perfectly on my server, but not on my laptop.24 -
The Odyssey of the Tenacious Tester:
Once upon a time in the digital kingdom of Binaryburg, there lived a diligent software tester named Alice. Alice was on a mission to ensure the flawless functionality of the kingdom's latest creation – the Grand Software Citadel.
The Grand Software Citadel was a marvel, built by the brilliant developers of Binaryburg to serve as the backbone of all digital endeavors. However, with great complexity came an even greater need for meticulous testing.
Alice, armed with her trusty testing toolkit, embarked on a journey through the intricate corridors of the Citadel. Her first challenge was the Maze of Edge Cases, where unexpected scenarios lurked at every turn. With a keen eye and a knack for uncovering hidden bugs, Alice navigated the maze, leaving no corner untested.
As she progressed, Alice encountered the Chamber of Compatibility, a place where the Citadel's code had to dance harmoniously with various browsers and devices. With each compatibility test, she waltzed through the intricacies of cross-browser compatibility, ensuring that the Citadel would shine on every screen.
But the true test awaited Alice in the Abyss of Load and Performance. Here, the Citadel's resilience was put to the test under the weight of simulated user hordes. Alice, undeterred by the mounting pressure, unleashed her army of virtual users upon the software, monitoring performance metrics like a hawk.
In the end, after days and nights of relentless testing, Alice emerged victorious. The Grand Software Citadel stood strong, its code fortified against the perils of bugs and glitches.
To honor her dedication, the software gods bestowed upon Alice the coveted title of Bug Slayer and a badge of distinction for her testing prowess. The testing community of Binaryburg celebrated her success, and her story became a legend shared around digital campfires.
And so, dear software testers, let the tale of Alice inspire you in your testing quests. May your test cases be thorough, your bug reports clear, and your software resilient against the challenges of the digital realm.
In the world of software testing, every diligent tester is a hero in their own right, ensuring that the digital kingdoms stand tall and bug-free. -
About to go on crunch to release a feature that is late. I have my own blame to put on it, as I wasted a lot of time, but goddamn.
Every time I said we'd need to take time to test for corner cases and check for errors here and there, my boss told me I need not worry about it, it's just an MVP. Then the marketing people see the feature half-ready and start suggesting their own changes. Then the idea of the project is refined and changed, a new subfeature is added, new backend business logic is added, right as I'm about to finish the original core features. They have the full product in their heads and are already selling it to people while I'm still catching up with quite a significant number of tasks. Now I have to crunch to launch tomorrow morning.
I do mainly the backend parts, but while a frontend guy who knows his CSS does components and pages, I'm the one to figure out pretty much all logic, and how to stitch said components and pages together and how to make the frontend interact with the backend. I'm supposed to do this whole thing and also deploy it all. Hell yeah.2 -
Holy f-ing hell!
Why do the small things have such fucked up corner cases?
This is very likely a giant bug with Qt, but how does this even happen?
I am using Qt with QML and sending data to a database on the backend. I call functions in QML from a Date JS object (property actually, but it calls functions) to set the date as a QDate. This is stored in the database as yyyy/MM/dd. This is fine. When I read the date out I convert it back from string to QDate and send this object to QML. Which then converts this to a Date object in JS in QML.
But at the point where it converts from a QDate to a Date object it loses an entire day. Seriously? You didn't gain a day going from Date -> QDate, but you lose a day going from QDate -> Date?
How long has QML and Qt been around? At least 5 to 10 years. How has this bug lasted this long? I don't want to do a bug report. I will, but I don't want to .6 -
Original class has 185 LoC, unit test class have 315.
But it's one of the most important classes, with most states and corner cases.
And, thanks to unit tests, one of the few classes I'm almost proud of. -
Damn, I hate Spring and most CoC frameworks. We’re using it from 2 years, I read most of the docs and it still gives issue for basic taks which takes hours to fix because I have to guess the exact convention which allows the program to process corner cases which aren’t explained in the docs. Yes, in the end the code is more elegant but it’s worth it when issues which could be solved in minutes by a trivial if/else statement takes hours (rigorously on pair programming™️ since being one of the few which actually read the docs I end up with lots of calls for help for Spring related issues by other teammates) and huge headaches to fix following the framework’s way?2