Thread 106523648

What are you maids working on?

Last one: >>106505464

should i really learn html and css by building a website when i already can read and navigate html and css ?
i tried it, then thought its a waste of time because im just looking up the parameters i can style the html elements with, copy pasting the keywords and changing the values.

this is my last maidpost before i go back to posting like a relatively normal user of /g/. thank you for your time, everynyan.

sex with Eli.

>>106523781
Build it and write a book about building it. This way, you are sure you achieved mastery and you can help teach newmaids how to use the technology.

>>106523648 (OP)
im not a maid and i hate that you appropriated dpt
also therad theem:
https://www.youtube.com/watch?v=ytsHe-TsJ8Y

>>106523781
You learn html and css by writing it. You can read some articles every now and then, but 99% it's just learning by using it and mitigating retarded browser incompatibilities.

>>106523648 (OP)
I might commit a crime if the image gets deleted on this one. Technologically speaking.

>>106523798
I am not Eli. Eli is unfortunately quite mentally ill and could not create a maidbooru or a maidchan. I do like his vision if a polite, maid and advanced Mathematics and Computer Science research oriented forum though. Once maidchan is completed, maids will migrate there.

>>106523817
The appropriation is temporary. The new home for maids is being constructed offsite. It is being constructed with the finest Java wagie tech.

>>106523900
>Literally admitting that you need a safe space because your group panders to the mentally ill.
You don't keep the cancer. You kill it.

>>106523932
>one mentally ill person implies the whole community is mentally ill

>>106523944
>It's just one tumor. What could it possibly do?

>>106523900
I like his vision too, even if he struggles to make things work I wanna help him create that

>>106523952
You surgically remove the tumor, which is the bad actions, not the organ or the person

>>106523900
>I am not Eli.
i know, i just liked posting it wherever i saw maidposting. but i must stop. you guys are about this for REAL. i like that but i also don't want to harm the movement. good luck.

>>106523900
>The appropriation is temporary. The new home for maids is being constructed offsite. It is being constructed with the finest Java wagie tech.
fair enough but please, do

F# with fable is the singular best way to write clientside code for the web

>>106524043
It's decent, but it's slow which makes you use the watch compilation and the watch feature breaks with random errors occasionally.

I'm gonna kill everyone you love and make you eat them raw.

>>106524334
Can I at least put some hot sauce on them?

>>106524334
You are what you eat, so I will just love myself

is the best way to learn programming to just read a book about a language

>>106524398
No, the best way is to use it.

>>106524398
The best way to learn it is to use it. The books and videos and blogs will help you learn new ways of doing things but you WILL write non-idiomatic code for a while. Your code will be ugly even if you have experience with other languages. There is no way around it. Just write stuff with it. Trivial programs like temperature conversion or tic tac toe. Just something to get the reps in.

>>106524444

youll have to dig trough the documentation at some point, might as well take all of it in before you start to mess around hands on.

>>106524456
Books are worthless, reading is worthless, you can learn by doing, that's why surgeons are all old as fuck, they didn't spend their time reading useless books.

>>106524482
did books hurt you when you were small ?

>>106524543
Bet you had to read a book on how to have sex, loser.

>>106524456
Language documentation =/= book about language

Yes, you will be relying on documentation, reference and/or specification. But you do not really need to read a book about a language. It can be a supplementary material but the core will always be practice.

>>106523648 (OP)
Asuna sexo

>>106524482
>that's why surgeons are all old as fuck, they didn't spend their time reading useless books.
What the fuck do you even mean by this? You must be some sort of Giga Jeet or genuine fucking retard. This may be one of the dumbest things I've ever read.

>>106524644
when i type c++ documentation into my search engine i get a million different results

which do i use

resources are either unintuitive to navigate, too overwhelming, some dont go in detail enough, some too much.

AI helps but sometimes it spits out junk as well

>>106524398
The best way to learn is to pair a book with a project.

>>106523989
I don't think Eli did any bad actions. He's just schizophrenic.

>>106524854
Well, C++ is just a retarded case. I personally use cppreference.com when I need but I am not regular C++ developer.
Honestly I do not even know how can one get a really good grasp of C++ without living through this decades long rollercoaster of bad decisions.

>>106524844
>retard can't read a single sentence
>will read an entire book and comprehend it all
sure thing, bud

>>106524854
>>106524890
But yes, since C++ doesn't have any official guide you might need to resort to some book to get started.
However once you get good grasp on basics, you might be fine with just reference and reading about common practices, pitfalls, recommendations, etc on the internet.

>>106524854
C++ is a very special case, people often limit themselves to a subset of the language because it's a huge clusterfuck and it gets worse as time goes on.

I made the asteroids game on python using a guide, it only took me the entire day!

CAPTCHA: 2MSRY (Too misery?)

>>106525464
>three dots
>this post is extremely low quality

>Anonymous 09/08/25(Mon)23:01:25 No.106525536
>>>106525464(You)
>>three dots
>>this post is extremely low quality

modulus oriented programming is the future, most of your problems will vanish if you just use modulus in the correct way.

>>106524398
Depends on the book. If the book isn't mainly descriptions of functions you have to implement then no. If it's a description of an API that you have to use then definitely not.

>>106524398
real learning comes from doing, in all things, not just programming

no one thinks they can learn an instrument by reading a book but somehow people think they can learn to make software that way

Lol i think i forgot i was writing a commit message halfway through writing it
>>106523900
Is Eli making his own imageboard?

>>106524860
I meant "bad" as in "not working" instead of evil

>>106526381
>69
>show vi index.html
>mainster
Everything's cursed here

>>106524844
>This may be one of the dumbest things I've ever read.
I take it this is your first time on the internet?

>>106526408
>Everything's cursed here
But fixable. It's git, so you can amend the commit message and change the branch name. Doing that will alter the commit ID.

>>106526381
lol

>>106525464
>>106525536
>>106525558
You are already raped.

>>106526381
Eli is schizophrenic and can barely hold down a conversation unrelated to counting. Unfortunately for him you have to have a PhD in Number Theory or Graph Theory to actually understand his ideas. Even if they were adopted, humanity won't have enough compute to use them for a few centuries at least. By then better stuff will have been invented that does the same thing. He mentioned making a maidchan, but I don't think he ever made any effort towards it, so I am making one.

>>106526400
That makes sense.

>>106526589
Nooo, it's funny like this

death to all tranime faggots and maid niggers

>>106526589
there's a commit after it so too late for git amend. there's another way but it looks like too much of a hassle

durgamaid = eli’s sockpuppet
report & hide avatarfaggots

>>106526859
>Anonymous
>09/08/25(Mon)17:09:12 No.106526859
>durgamaid = eli’s sockpuppet
>report & hide avatarfaggots

>>106526938
reuploading is prohibited it's written right there meanie

>>106526938
kill yourself

>>106527769
I have been collecting these. Thank you for making them.

I am considering using Anubis for maidchan. Does anyone have the original cute Anubis mascot images? I need them so I can edit them to put Anubis in a maid dress.

>>106524037
Features for maidchan are half done. I am working to integrate it into maidbooru, and I might try to make a sentiment analysis janny-bot like that maid in the last thread suggested. I also still have to make the mod/admin panel. Boards/threads work. Catalog works. If it keeps up at this speed, maids will have a new home within a month.

>>106528824
>Does anyone have the original cute Anubis mascot images?
You can find them here:
https://github.com/TecharoHQ/anubis/tree/v1.15.1/web/static/img

>>106528824
I used nano banana to put her in a maid outfit. What do you think?
Catbox:
z638jc.png
vgrbm1.png
b17up5.png

>>106523648 (OP)
I wish I looked like her

why can't you fucking maid nigger trannies get your own thread? If i remember correctly you used to have one. Sperging out about strange portable computres. You gotta go back.

>>106525533
i made a ultima style game in Python with WASD movement, Im gonna use the Shadowdark rulebook for it
My boss asked me to just learn Python for AI crap, got a little sidetracked

I'm completely fucking deranged because I just spent 20 minutes writing this

Donald Trump: Hello, my name is Donald Trump. I'm a President of the USA.
Donald Trump: Hey, E. Jean Carroll! Let's fuck!
E. Jean Carroll: *Screams* AAAAAH! PEDOOOOOOO!
False

Troonface McSaggytits: Hello, my name is Troonface McSaggytits. I'm a Disgusting tranny.
Troonface McSaggytits: Hey, Donald Trump! Let's fuck!
Donald Trump: Big age gap, that's kinda weird.
Troonface McSaggytits: Guess I'll go dilate.
Troonface McSaggytits: I'M COOOOOMING!
True

Joe Biden: Hello, my name is Joe Biden. I'm a President of the USA.
Joe Biden: Hey, Donald Trump! Let's fuck!
Donald Trump: That's kinda gay bro.
False

>>106529762
These are extremely cute and I will definitely use them. Thank you for making them!

>>106529762
>I used nano banana
I'm genuinely surprised AI delivered decent content. The outfit looks so coherent with the original art. You maids should really get into AI research instead of counting big numbers

recommend me a book to learn about discrete math

Why did janny ban the maid who was trying to help all the other maids leave? Are they worried about this dead site losing even more users?

>>106531722
Not specific to maids, I've noticed Janny is banning everyone who
A) Talks about other Chan sites (and maids called their new HYPOTHETICAL, not even real home, "maidch*n")
B) Talks about leaving 4chin and replace it
Tsundere janny want you all to stay here for some reason. He might delete your posts but it's clear he's in love with yall. More proof on this bizarre behavior if you check the archiv* cat*log (can't say it, janny might think I want to replace her with an archiv* site)

>>106531961
ironically enough, they are too trying to memoryhole the other hiro chan site, any mention ends up in a ban too
imagine it like Facebook banning you for mentioning Instagram... but oh well who am I to judge

>>106531961
>I want to maidpost here
>a vocal minority hates you
>you are banned
>fine, I will make a new home for maids
>the vocal minority can just have 4chan
>you are banned
I don't understand the reaction. Maid /dpt/ is measurably better than non-maid /dpt/. It seems like it would make more sense to just ban the three retards who have a melty every time a maid is posted. That would improve the thread even more.

Why doesn't this tranny hurry up and fuck off to his worthless altchan already instead of shitting up this thread?

>>106532255
>discuss programming
>get banned
>retards appear to ask why there is no programming discussion
Pure anti-semaidism on part of the janny squad.

>>106532375
Fuck off retarded nigger. I hope registryschizo lacerates your face beyond recognition.

>>106532375
Sir we are a linux vs windows board. Sometimes a political board when mods feel like it. Programming and technology discussion have no place here, don't be silly. Now please go make a thread with a pepe titled "firefox sucks chrome better kek". Thanks.

what's a project you've done that's your favorite, /dpt/?

>>106533076
Kill yourself tranny

Why do i get 32 bit addresses for heap allocated memory, and 64 bit addresses for stack? i get that the stack is at the top of the address space, but i don't presume to think that a 5 line program has 4 gigs of memory allocated to it.
is the allocated memory not contiguous? linux 64bit x86 if that matters.

>>106532223
>>106532375
>>106531722
The problem of this site is that anyone who wants to improve or change it's condition, become hostile to it and try to immediately splits to another site
And it always happens for the smallest of reasons, usually ending in more drama

It's clear that the mods are not against maid projects or posting maids, the problem is that maids are becoming another secession group for no reason
Instead of making projects and talking about computer science, some maids keep falling for low quality baits and participating in the internal culture wars, acting as if a few inane hostiles are a massive threat that needs a walled garden built against, while they are clearly a minority incapable of even arguing without spouting a dozen buzzwords

Let's build our infrastructure to last, with projects that can be used for things other than fleeing from people you could just ignore
It's not possible to achieve the maid ideal of peace and dignity, without being peaceful and dignified yourself

>>106533309
I believe this might be related to Linux being a higher half kernel
Linux uses virtual memory, so these addresses aren't really the same as the physical memory layout
Check the static data and function addresses as well to see where it placed your program

>>106533541
indeed, and it actually only uses 48bits for memory, translated to 64bit.

stop responding to maidniggers
stop even mentioning maidniggers
only correct reaction to a maidnigger post is to report and hide (off-topic / avatar or signature use)
do not even bother to read their drivel or open their attention whoring images

>>106533580
Absolutely based
TMD

>>106533693
this one's better with the audio

>>106533468
The problem is that maids recieve excessive jannying. Maid threads are deleted. Maidposts are purged. Given the active discrimination they face, it makes sense for them to want their own home.

>>106534301
If maidniggers were jannied enough, I wouldn't be reading your retarded post right now.

>>106523648 (OP)
Asuna hot

>>106534301
Maid threads and posts are deleted for 3 reasons from what I have seen:
>embedded code in files
>trying to split from 4chan
>eli bothering random anons, or making off-topic very abstract math ramblings
I never seen the ones not doing that getting deleted

>>106534730
>very abstract math ramblings
If they only did that it would be fine

I changed a line of code, now the crash is gone. Took 8 hours.

>>106535005
Those are the bulk of his posts. The negative reaction comes from anons who literally don't know how to count.

>>106535240
No the majority of his posts are him deflecting from his obnoxious avatarfaggotry

>>106535374
His last thread was none of that.https://desuarchive.org/g/thread/106501098/#106501098

>>106535473
>literally all the stupid avatar act

>>106535627
>t. numberlet who doesn't know what a godel number is or how to count

>"im not attenwhoring, look!"
>shows thread of him attentionwhoring
why can't you just fuck off? go to discord with all the other troons

writing a coom-related script

>>106537103
>black band with white frilly lace
Is..... is lauriewired wearing a maid outfit while discussing C?

Wrote a recursive descent deserializer for a game file format in C a few years ago. Rewriting it in C++. using std::variant makes for cleaner code. The tokenizer fits in ~100 lines of code, the parser should be pretty simple too

>>106537776
post it

>>106537823
It's not done yet

>2026 insiders
we might actually see some C++26 features finished

>>106537971
I'd rather use Hare.

>>106537968
>>106537823
Well the whole thing isn't done but here's the core of the tokenizer (minus the type declarations)

https://pastebin.com/bjd6QuFU

>>106537996
gonna show this to herb sutter later and we are gonna have a good time with it

>>106537996
>regex StringPattern("^(\"([^\"]+)\")");
this doesn't take escapes into account.
you wouldn't be able to put an escaped " into the string with \", worse, if it was in the string, it would parse the backslash as a regular character and the escaped " as the string terminator

>>106538058
The format doesn't accept escaped strings afaik

>>106537996
>A GUEST
aghast*

>>106538058
>>106538072
Oh well turns out I'm wrong, I'll fix that thanks

>>106538072
>escaped strings
Tt's not even about interpolating certain escape sequences. Even a string without interpolated escape sequences should be allowed to contain its terminator, otherwise it's massively restricted. It's very common to want to insert " into double strings.

>>106537996
>using namespace std;
AHAHAHA LOOK AT THIS, HERB!!

>>106538137
Thanks for bringing this to my attention, scott. My disappointment is immeasurable.

>>106538137
It's a .cpp file so it doesn't get exported into the dll's header. There's also a namespace wrapping it so it's localized to that scope

>>106538153
pre(x > 0) using namespace ^^reflexpr(std); // only compiles with the safe profile enabled

Now herb sutter can sleep tonight

>>106537996
Major Issues

Code Organization
Using using namespace std is generally discouraged in header files

Regex patterns should be encapsulated within a class or namespace
Function is overly long and complex

Performance Concerns

Regex compilation occurs repeatedly despite patterns being constant

Multiple regex_search calls per token processing
Unnecessary string copies in token creation

Code Structure

Excessive use of goto statements makes code harder to maintain
Mixed concerns between token recognition and position tracking
No separation between lexical analysis and error reporting

where are some good resources on multithreaded programming (in C more specifically)? I get the absolute basics but I have no idea how to do it well or what I should avoid.

if the code does something that has an actual use, it's inherently better than the most optimized, most beautiful code that is not actually used by anyone or anything

only when the application actually has been used, there is any point of discussing any improvements to the code

retard learning python here
is there any reason to not do like
cheese : int = 1
for every variable instead of
cheese = 1?

In my brain telling the computer that it will always be a int will make it faster. The cheese in a program will never not be int so it should always expect an int when dealing with cheese.

>>106538233
for example I know about race conditions, deadlocks, basic synchronization primitives etc but would like to know more about task scheduling, how to structure a job system or thread pool system, how to manage memory etc, things like that. Really struggling to find good info on this.

>>106538233
https://marabos.nl/atomics/
https://www.amazon.com/Seven-Concurrency-Models-Weeks-Programmers/dp/1937785653

you're welcome.

>>106538227
> Using using namespace std is generally discouraged in header files
It's a .cpp file, not a header

> Regex patterns should be encapsulated within a class or namespace
There actually is a namespace surrounding the whole thing, this is an excerpt from the rest of the file

>Function is overly long and complex
Not really sure how to get it any smaller than it is now without just breaking it out into other functions for no reason

>Regex compilation occurs repeatedly despite patterns being constant
How do I make sure they compile once? They're declared outside of the function and afaik should only compile once?

>Multiple regex_search calls per token processing
How do I match the token against the different regexes without checking each one? I ordered them in the most likely order so that less common ones (identifiers, comments) are less likely to run

>Unnecessary string copies in token creation
This is probably true but I'm nervous about relying on the lifecycle of the input body string - I could use string_view, perhaps

>Excessive use of goto statements makes code harder to maintain
True but sometimes the spicy keyword works

> Mixed concerns between token recognition and position tracking
You meant the DebugError? I wanted to be able to print errors during both parsing and lexing so that if there's a parsing error down the line I have the position of the token in the file already. How would you write it?

>No separation between lexical analysis and error reporting
Not sure how that differs from above - the function's Result monad is a variant which accepts either the or the DebugError type

Thanks for the review

>>106538280
The only reason to do this in Python is if you think types make your code more readable. There is some disagreement among maids about this. Some feel it helps document the code and others feel like types are restrictive and like training wheels.

For a newmaid, do whatever feels easiest to you and makes what you are doing most understandable.

>>106538342
nta but
>reddit spacing
also why not a generator instead
incidentally you should just remove the regexes altogether and build the state machine for it yourself into the switch

>>106538358
It's the same either way for me. I'm just autistic and don't want to ever put out un-optimized code.

>>106538233
That depends. It particularly depends on how much you want the threads to communicate and how many threads you're dealing with, which is all very algorithm- and deployment-specific. As long as the number of threads is fairly small (perhaps up to 50 or so on most hardware these days?) and the amount of communication isn't too overwhelming, you can just do something like POSIX threads with pessimistically safe locking. It's pretty easy to implement, and is about the level the Rustoids operate at.
With more intense threading requirements (or if you're working with realtime code) you probably need to understand things like OpenMP, MPI, CUDA, lock-free algorithms, and TDMA (the classic method for realtime, less applicable elsewhere). The techniques for coding with a million threads are very different to those for a few tens of threads.

>>106535240
>>106536128
You know pretty well that's a meme, and it never went anywhere other than causing random drama and Eli getting his posts removed
Sure these math abstractions are interesting, but masturbating over them doesn't build anything, not even fulfills curiosity or entertainment since it's the same thing over and over

>>106538391
>generator
Because I didn't know about it lol, I'll check it out. I've used this pattern in my normie web dev job, didn't even think to try it in sepples

>>106538412
I've done some stuff with openmp (including openmp simd stuff) and was under the impression that this was not the optimal way of doing things but yeah I'm more looking for something fairly basic like the POSIX threads part of your post as opposed to the more complex stuff mentioned. The best info I've gotten so far is from asking AI about it but of course it only gives you pretty superficial answers

>>106538358
>The only reason to do this in Python is if you think types make your code more readable.
It can be useful sometimes with typecheckers, especially if you want to use a different (wider) type than would be auto-deduced from the initialisation value.

>>106538542
>was under the impression that this was not the optimal way of doing things
If you're doing the stuff it likes (applying an operation over arrays) it's very good for low effort investment, up to a certain scale (which depends on your hardware). But it definitely grew out of Fortran for high-performance computing, and it really shows.

>>106538428
It never goes anywhere because the threads get deleted within minutes of creation. In the last one he was trying to get distribution other than maid card and it was still deleted. I suspect jannies are suppressing his research because he found a new way to build neural networks.

>>106539078
They never went anywhere when they used to not be deleted, he would go saying the same thing over and over just with more words, with the depth of one of those curiosity youtube videos. People would be posting maids around while not understanding a thing and thinking it had some hidden meaning

> In the last one he was trying to get distribution other than maid card and it was still deleted
I discussed that with him, other method of distribution was even shadier like using p2p for no reason
He refuses to use github, gitlab, gitea, codeberg, I think I convinced him into self-hosting git but I didn't hear of that since

>I suspect jannies are suppressing his research because he found a new way to build neural networks.
You know pretty well things are not that grandiose, he's a fun person but the "research" part is bogus, don't be that impressionable

>>106539185
It was worse than p2p now that I checked
He wanted 4chan to have native support to maid cards

>>106539300
for*
As if the site that never updates would update to run his code

>>106539300
The last one I saw he was asking about catbox.

https://desuarchive.org/g/thread/106501098/#106501098

Holy fuck why didn't anyone tell me C++ had raw strings. Makes unit tests so much easier:

auto res = MapFile::Tokenize(R"(Test 1234.566
//foobar
"te\"st"
)");

//...

Assert::IsTrue(std::holds_alternative(tokens[5]));
auto str = std::get(tokens[5]);
Assert::AreEqual(std::string(R"(te\"st)"), str.value);

>>106539557
Sounds a little more reasonable, but this obsession on not using something normal to share code is holding maids back
It's shady, even if unintentional, it's like waiting until midnight and go to a back alley to sell candy and be clueless when people think it's drugs

>>106539589
C++ has raw strings

>>106539657
His stance seems reasonable?

https://desuarchive.org/g/thread/106375242/#106436988

>>106539870
That's just rhetoric, in practice you are simply using someone's site for free, and they have some rules to avoid legal issues for them
One can go the arrogant way and say they don't want to agree to evil terms, evil corporations, you should be able to do everything, I have seen people in /x/ say that the universe is evil for not allowing you to fly or noclip

That doesn't change reality, we don't have the infrastructure. We have to rely on the internet and it's submarine cables, on the hardware someone else made, the operating systems and programs other people made
I want to build maid infrastructure, but it won't go anywhere by challenging the world and going against everything for a bit of purist high

Wasn't the point of maids to have a comfy virtual culture and projects? Why going on an armchair rebellion against the entire world that doesn't even care? To say X and Y are impeding us from having peace and fun instead of admitting we are going the wrong way?

https://streamable.com/zp3yy1 figured out the engine/skid sounds

>>106523648 (OP)
What are some fun non beginner projects to do? I am a Swiss Army knife dev (C++, JS and Python) and I've completed all my personal projects.

>>106540029
>Wasn't the point of maids to have a comfy virtual culture and projects?
If you just wanted peace and fun you'd drop the entire thing and post anonymously like everyone else or fuck off and do your own thing like other insufferable groups have done before.
You're the same as the trannies when it comes to behavior. The funny thing is, you fags without exception are the exact opposite of what a maid should be. Inconsiderate, annoying, going against the expected behavior, want to stick out. In a single word, 迷惑.

Anon, How to post code with code block?

>>106540215
>I've completed all my personal projects.
have you made your own programming language

>>106539589
>std::holds_alternative
C++ sure is a strange land

>>106540029
It seems like if infrastructure is bad, it should be replaced. Maid card replaced github but was banned by 4Chan. It seems like building Maidchan is the next logical step. If current systems are unacceptable, new ones can be created. There is nothing comfy about excessive jannying.

>>106540229
Anonymous isn't a comfy culture. It is a box of retards spamming soijacks and shouting about "trannies". Maid /dpt/ is measurably better than non-maid /dpt/, until janny shuts it down. A Maidchan is worth it just for maids to segregate themselves from such people.

>>106540292
std::holds_hand

>>106540272
Thanks anon
>>106540283
No, but it wasn't a goal of mine. I suppose I could try to make one instead of doing leetcode all day. Should it be interpreted or compiled?

>>106540341
std::clinic_checkup

Ocaml my caml

>>106540272
>Imageboard
>Green: Easy
well fuck me in the ass with a stick of chorizo

>>106540569
If it's your first non trivial project then it does take time

>>106540569
They're easy if you don't need to handle a lot of users or otherwise scale. Handling a million simultaneous users can turn an Easy problem into a Fuck You tier one.

>>106540229
Seeing by your own behavior, it doesn't seem like posting anonymously gives you politeness, so I don't think your advice is relevant
But I'll let you know that my objective was always having maids acting like maids

>>106540330
And how does the current infrastructure impedes you from sharing maid projects and culture?
There's nothing against maids in github, you're letting the american politics corrupt a good idea
About 4chan you can simply hide bad users and stop doing things in a shady manner

>>106540361
I've decided to try to build a program that scans the text of tv show title cards inside of video files and then fetches the episode and season number from the tvdb. The G programming lang will have to wait for another day.

I am still confused by UB. How can literally any C programs be written if the compiler actually optimized in ways its technically allowed to?

>>106541216
Practical consequences are rare
In your pic, if you write a[x] = 0 with a x bigger than 4 it's clearly a mistake

People investigate these implementation details to improve compilers, but in reality it's completely fine
Compilers also have consistency depending on your target, since a lot of the undefined behaviors are undefined just because of embedded programming

>>106541216
what a load of nonsense

>>106539300
>4chan to have native support to maid cards
Maidbooru and Maidchan do this. It is a pretty easy format to parse.

>>106541216
that example is just wrong, the real example that I assumed he tried mix it up a bit so it didn't look like he stole it uses x as the variable in a for-loop with start and end values being compile time constants. the example in the pic you posted could never actually happen he just made it up.

>>106541598
he's on the C standards committee so I wouldn't assume he's making anything up

Am I using a factory pattern correctly here? I want other components of my system to be able to construct derived classes without knowing what those derived classes are. I also want to be able to restrict on a case by case basis what derived classes a particular factory is allowed to create. I'm doing this by having derived classes 'register' themselves with the factory, providing both a unique ID and a function to create them. I plan to place the create method as a static method inside of each Derived just to keep all of the code related to a specific derived in one place (rather than like a giant switch case inside the BaseFactory)

static std::unique_ptr create() {
return std::make_unique();
}

class BaseFactory {
public:
std::unique_ptr create(std::type_index base) {
return mapping[base]();
}

void registerBase(std::type_index base, std::function()> create) {
mapping[base] = create;
}

std::map()>> mapping;
};

factory.registerBase(std::type_index(typeid(Derived)), Derived::create);

>>106523648 (OP)
>What are you maids working on?
Porting pure rust lua vm to esp.
It turned out to be very straightforward since someone has already been working on a no_std PR some time ago and it was just manner of rebasing it to newest changes.
Since I already had interface for pico8 prepared for the test native emulator, it only took a little wrestling with psram to just get it running on the chip.
Now it's time to implement rest of pico8 api.

>>106542475
btw since I have basically neither lua stdlib nor pico api avaliable, the current "cartridge" looks like this.
Writing this all as just 128x64 numeric arguments was causing a crash. There is surely a lot of performance related issues to investigate.

>>106541764
I'm not assuming, it's obvious that he is. You don't just get to make a claim and then not provide any proof. His only "proof" is a 15 year old GCC bug and a loose interpretation of the C standard where you intentionally make code break on purpose by making the compiler work in a way that not a single person in the world would want or expect just because the standard didn't explicitly forbid it, even though it currently does.

>>106523648 (OP)
Anon, I made my own a simple BMP Image Viewer for TempleOS using HolyC

https://github.com/ZerpRedgrave/LibHC/blob/main/Dump%20Project/FileBMP.HC

>>106523648 (OP)
Maid Search is BACK!

https://ayasequart.org/sql?boards=g&gallery_mode=y&has_file=y&file_tags_character=9173

Please give me some time to add more media and paging. Thank you for your encouragement, Maid dra/g/ons!

>>106542749
Turned it off for now :)

Goodnight dra/g/on maids.

t. tired maid

>>106542749
What is Maid Search?

>>106541537
The Maid Card integration is completed. It is one way. Admin adds a Maid Card to Maidbooru, it is auto-detected and the preview is placed on the side of the page along with a zip download. This way, Maid Cards can be safely posted and shared directly on Maidchan, without worrying about braindead coomers stuffing porn into them. Maids can also get the content directly without using Maid Card at all.

All that remains is a little more work on Maidchan, then they can be deployed and I'll stick Anubis with these >>106529762 in front of them. Most of the remaining work is creating a jannybot to auto-moderate text posts.

>>106541216
>implementation can just assume
No, it cannot, 0 compilers remove that kind of if stattement unless x is a read only constant that can never be changed through any means. And even if it is a constant, C is such a dogshit language and its compilers are so mindbroken, that they don't bother to propagate constants unless you literally call the function with a literal or define the constant inside the function itself.
I wasted many hours rewriting specific parts of my code as macros because C compiler couldn't even replace strlen with a constant. C++ doesn't have this problem because things like string_view in a function signature are constructed before function is called, that is, constants are propagated into constructors before the function even exists. You need this level of syntax sugar just to have a compiler propagate a string literal across a function boundary, don't give me this shit of range checks ever being eliminated anywhere but in a toy fizzbuzz tier setup that's handcrafted to display an useless feature that is never used outside what should've been a fucking #ifdef.
Falseflagging rustnigger tier strawman.

>leave /dpt/ for a couple years
>come back
>/dpt/ has been enmaided
How do I participate? Do I just post maids or are there some kind of rules? I wanna be a maid too!

>didn't post in /dpt/ for few hours
>it felt like years g-guys, did you miss me?
Nobody will ever miss you even after you get brutally killed on your way home, faggot.

>>106545648
I miss every maidposter when they're not here, especially Augusta.

Why don't more Rust crates reexport their dependencies. It doesn't make sense that I have to add "crate-thing-enhanced" and also "crate-thing" especially when types from the latter are part of the interface of the former.

>>106545750
why should they?
cargo tree -d

if you did it right, you don't have any duplicate code.

>>106545750
Because the crate dev chooses version of the crate to depend on.

>>106545775
Because now I have to manually make sure I add the same crate version and feature flags or things could get fucked up.

>>106545635
Just go to the Middle East, they need a bunch of maids to be abused

>>106543580
I think that screenshot and link sums it up

>>106545750
Because if some other package do this or you want to pin to specific version or enable more features then you have multiple ways of naming same export and this gets messy. It is rarely worth it to reexport dependencies. You only should do it if you are confident users will not want to specify that dependency directly and you will never want to remove it yourself.

>>106545819
Feature flags are additive, you shouldn't have to write same features as your dependencies. You only need to match major version too, resolver should pick one that satisfies the most. And if it fails, you will get a message about incompatibility between definitions from different versions. Sometimes, it is fine to have multiple versions of same package if they never interact with each other.

>>106545819
I mean, ok, but then you got cases like tokio and http and what not and it starts getting kind of unwieldy with all the reexports.
I don't like it either but it's an extremely shitty problem to solve and I don't think any language has a solution.

just keep fucking with it til cargo tree -d is empty.

You could just be a real programmer and stop relying on others to do your solutions for you.
>b-b-ut this thing was done billions of times!
wake me up when serde is at least as efficient as simdjson

>>106546776
Serde is not even a json serializer.

>>106538233
there are none, all the "generic" multithreading resources are all in C++ because of how minimal concurrency primitives are in C compared to C++ and because windows literally just added proper support for standard C11 multithreading like a couple years ago so no one cared about it

>You could just be a real programmer and stop relying on others to do your solutions for you.
Is there anything more Dunning-Kruger coded than this sentiment?

>>106546924
Without maids, /dpt/ is just nocoders unconvincingly larping as senior developers to try to put extra clout on their opinions.

>>106546802
And rust is not a webshitter language, and yet...

>>106541216
>>106542594
>UB means the compiler can do ANYTHING, including remove root? Top 10 hidden dangers of UB in C you need to know about
most of the people who fearmonger and clickbait about UB as an abstract concept often seem to laser focus on elements that have since become de-facto standards
in the category of doing "anything" there is incidentally included yet almost always never mentioned the option to consistently define that behavior
pure standardsfags freak out about it because in their eyes only the standard exists
everyone else because they're trying to sell you a solution to the problem they invented for you

>>106547005
UB enables anything, except optimizations that I'd expect compiler to just do.

>>106547026
anything also includes trying to avoid bug reports from retarded enterprise developers saying
>i tried to make bug filled enterprise win32 app with the fast setting it doesn't work why it no work anymore fix this thing that's definitely your fault

>>106547005
UB is not about what can your code do. It's about what your code differently when you update compiler/update library/change unrelated code/run it on different machine/are unlucky. Every UB is a subject to change at any place at any time at any invocation.

>>106547131
>Every UB is a subject to change at any place at any time at any invocation.
surprise, things exist outside of the standard
platforms, standard library vendors, compiler vendors, and operating systems are free to define undefined behavior if they so choose at whatever level they chose and have done so because the idea that code could produce completely random results is a lot less useful than defining it to something actually usable and/or predictable and just have that function basically as an implicit language extension
the biggest one that comes to mind is literally no one gives a shit about strict aliasing and type punning behavior is for most if not all purposes well defined, in C or C++, regardless of how, regardless of features used, or whether it's in a context the standard says is safe or not

>>106547129
If that was the case, optimizing compilers wouldn't exist.

>>106547466
>surprise, things exist outside of the standard
And every behavior that is considered UB is subject to change at any moment. That's just what this word means.

>>106547466
compilers break ub code all the time for 0.1% performance gain

>>106547506
>>106547507
You don't know what UB means.

>>106547506
no, it's not, because it can be defined outside of the standard
there are only three compiler vendors of consequence and most of the other ghetto ones are just shitty clang forks
they all agreed on the internal representation of C++ coroutines
anything that relies on that internal representation is still UB
it's not changing "at any moment" and it's safe to access

or we could go with actual standards extensions like POSIX or khronos standards which include and expand upon another standard
if they define behavior for implementation defined or undefined parts of a standard, it doesn't just become implementation defined or undefined again because you personally do not acknowledge the existence of those standards

>>106545635
Just post maids along with your posts, write maid projects and be polite
Welcome back, anon

>>106547633
>no, it's not, because it can be defined outside of the standard
If behavior is defined then it is not UB. If it is undefined then it is UB.

Did you know there's a vscode plugin that lets you write custom CSS for the entire application? So you can change the icons for the sidebar or override the default UI font

https://marketplace.visualstudio.com/items?itemName=be5invis.vscode-custom-css

>>106523648 (OP)
>What are you maids working on?
I recently accidentally nuked my Sims 4 mod collection (70 gigs!), so I made an AHK script to automatically download stuff from TSR.

>>106547750
if behavior is defined by something that isn't the C or C++ standard, it's still undefined/implementation defined behavior from the perspective of the language and what the compiler is technically allowed to do
technically allowed to and actually allowed to are two different things
multithreading in C was UB until 2011 at earliest and for windows up until like a few years ago
that doesn't mean the compiler was allowed treat it like UB and was allowed to just remove all multithreaded code
POSIX threads defined the behavior for POSIX platforms and although windows has no formal definition it has de-facto standard in the form of whatever MSVC decides to produce for its atomic intrinsics and the Win32 threading API

>>106547938
>if behavior is defined by something that isn't the C or C++ standard, it's still undefined/implementation defined behavior from the perspective of the language and what the compiler is technically allowed to do
If a behavior is defined in some context then it is not an UB in that context. If it is undefined, then it is UB.

>C/C++
>she/her
>undefined/implementation defined behavior
>(You) are here.

>>106547775
>Java: Ready
sirs

we're so back
https://devblogs.microsoft.com/cppblog/c-language-updates-in-msvc-build-tools-v14-50/

>>106549437
>ctrl+F stdckint
(°□°)︵ ━

>C/C++
>she/her
>undefined/implementation defined behavior
>(You) are here.

>>106549566
>stdckint
>std cuck int

>>106523648 (OP)
Haskell is such a mind opener. Just the parameter (no sig) type inference is enough to replace some uses for generics and constraints.
Instead of switching on types in generic functions, you can just take a Sum type as a param and pattern match the function definition for a built-in switch!
The constraints on type params are so much cleaner than C++ and Rust equivalents.
Haskell is so beautiful!

>>106549393
I don't get to decide what language every component at my company is written in

you ever feel like you've wasted years on your projects? It is depressing, or are you happy you did it?

>>106547667
It seems like you are the only maid immune to being jannied for whatever reason. Durgamaid's update got deleted and the newmaid got popped too. Neither of them appear to have done anything wrong. All Durgamaid did was post a screenshot showing integration of Maid Card into Maidbooru.

Why is janny so anti-semaidic?

>>106550535
The sum type is also a function
marathon : Distance -> Boolean
marathon (Mi m) = m > 26
marathon (Km k) = k > 42

-- reduces to

Mi d m k = m d
Km d m k = k d

marathon d = d (m -> m > 26) (k -> k > 42)

>>106542433
dunno what you are doing, but the one issue I see with this is that normally I see factories used for video games with save states, and they save the entities into a file, but require a factory to load them back in.
the problem with using type ID is that it's not serializable, and if it was, it's not ABI compatible (I think? you can't copy the save file onto a different OS / compiler version).
So, the solution for this is to just use an enum + switch statement OR using a string / UUID for the ID type and use a hash map + register like your code.
If you don't mind keeping everything in one binary (no DLL's), you could use an X macro to automate the switch statement / register function.
The good part of x macro's is that you can convert enums into strings, and C++ RTII has typeid().name() but I don't know if the ABI has any guarantee across compilers / versions, but I assume it does not (and it might be mangled or something).
So it would be something like this.
class Base;
class SomeClassA : public Base;
class SomeClassB : public Base;

// if you add a new class, just add it here.
#define A_BASE_MAP(XX) \
XX(SomeClassA) \
XX(SomeClassB)

// this creates BASE_SomeClassA
// you can use an enum class if you want as well
#define XX(code) BASE_##code,
enum BASE_ENUM
{
ERROR_BASE_ID = 0,
A_BASE_MAP(XX)
LAST_BASE_ID
};
#undef XX

const char* enum_to_string(int id)
{
// this prints out "BASE_SomeClassA" you can remove "BASE_" if you want.
#define XX(code) case(BASE_##code): return "BASE_" #code;
switch(id)
{
A_BASE_MAP(XX)
case ERROR_BASE_ID: return "ERROR_BASE_ID";
default: return "UNKNOWN_BASE_ID";
}
#undef XX
}

NOTE this is untested, and creating a factory is for you to figure out (for deserializating parameters, one way is with a static function in the class, similar to create(), but I don't like constructors because you can't return errors without exceptions, so I would have an empty contructor + virtual bool load()).

>>106550930
*But your approach does work fine, and using a hash (an explicit string or UUID) can help with backwards compatibility for save files since adding new classes in before the end will change the ID's.
And I know all of this text below is pointless, if your approach works, there is nothing to fix.
And especially since I think you have multiple factories for some reason and I don't understand why you would need that, because you probably don't need serialization / not making a game (like... if you wanted to create a filter to allow / disallow factory objects, you could use a std::bitset using the X macro)
For example you could use a constexpr hash to hash a string and use that in a switch statements, which is then used in x macro switch statements (hash collision is possible, but very rare, and you would get a error inside a switch statement if you somehow did).
And to avoid header bloat (to avoid including every header into your factory function), you could extern a function using the X macro that generates something like:
unique_ptr create_SomeClassA(SomeInputStream& in)
But this is very annoying for anyone looking at the at that random function that seemingly isn't being called anywhere because they can't ctrl+f to figure out where the function is being called (because an X macro is declaring it extern and concatenating the class name with the function). And now you need to remember to add random functions on the global scope for every class you add (could be cleaner with a static member function that is registered inside a macro inside of the Derived class... but I am not sure if that works, I just include everything).
And all of that work just to avoid rebuilding the factory every time you modify the header of derived classes which might take a second on a sub 1 million line project... (but you don't need to include the header inside the factory switch statement anymore, that's kind of neat, one more step for one less step)

haven't really worked on my project for about a month now (other than only small fixes) but have been using it in its current state daily
and it's good this way since it gives insight on what works in practice and what actually needs improvement (opposed to some theoretical optimizations not based on any usability that would not make a lick of practical difference)

>>106550597
Not that the projects were a waste, but that I seem to be incapable of completing things in the amount of time that seems reasonable to me. All of my projects feel like they took unreasonably long time to finish, but the projects themselves do not feel like a waste.

>>106551398
>it gives insight on what works in practice and what actually needs improvement
this is the way, and why actual usage always matters

>>106547750
>If behavior is defined then it is not UB.
You have behaviour defined by standard. Maximally portable code stays with this, but you really don't need to be that restrictive. Your desktop app is never going to run on a weird 1s-complement 36-bit-word machine from the 1970s.
You have behaviour defined by platform. For example, POSIX effectively requires all pointers visible to a process to be in a single address space, and it is your platform that says what actually happens when you reinterpret a float as an int. Actual structure layout rules and calling conventions tend to be governed at this level.
You have behaviour defined by implementation. GCC's and MSVC's language extensions are like this.
You have behaviour defined by nasal demons. Do a misaligned out-of-bounds write into the stack and you're here, hoping you get away with it (or get a clean crash) but with no assurance that that's what will happen. (Predicting the behaviour here requires knowing many levels of the real deployed implementation rather well, so nobody bothers.)

>>106551748
If we can just reorient the discussion a bit, this started when one anon said that he doesn't understand how C can ever even work because behavior defined by the implementation is allowed to invoke nasal demons on purpose via the "as-if" rule and that's when everybody chimed in that he is being ridiculous.

>>106550930
>>106551386
Thanks for the insights

no it's not a video game, but it does have a lot of commonality with things like 'spawning an object' and saving the state of those objects to a file to load again on next run. I wasn't planning to store the type index, that is purely for internal usage so that I'm not hashing potentially long strings constantly. I was planning to store everything as strings in my 'save file' but there won't be a lot of things in there and it's helpful to be able to read what is stored in plain english. also I feel it adds some level of type safety and I can validate stuff with asserts at start of time easily, but maybe that's a false sense of security.

the reason for the multiple factories is because I'm trying to share a common interface between platforms with different feature sets, so if I have Derived objects [A, B, C] then platform 1 supports [A, B] and platform 2 supports [B, C]. I want the common interface to automatically include B since all platforms support it, but then platforms need to manually add the objects only they support, hence why the factory is dynamic. This is also why I dodged enums since I can't create an enum inside the interface and extend it on the platform of choice

Also how would parameters for constructors typically be handled in this sort of scenario? My plan was to define a "struct BasePayload" and then each derived class would extend it if they need arguments with those args and then my factories create function takes in a nullable BasePayload to use when constructing that object.

class Derived {
struct DerivedPayload : BasePayload {
...
}

Derived(BasePayload *payload) {}

static std::unique_ptr create(BasePayload *payload) {
return std::make_unique(payload);
}
}

std::unique_ptr create(std::type_index base, BasePayload *payload) {
return mapping[base](payload);
}

>>106551985
>the reason for the multiple factories is because I'm trying to share a common interface between platforms with different feature sets
if the platforms are determined with compiler macros, you can technically still use X macros to skip / error on certain entries using logic.
Because X macros can could have multiple parameters.
So like this:
enum a_flag{
FLAG_WINDOWS = (1 << 0),
FLAG_LINUX = (1 << 1),
// cleaner if "ALL" = 0
FLAG_ALL_MASK = (FLAG_WINDOWS | FLAG_LINUX )
};
#define A_BASE_MAP(XX) \
XX(SomeClassA, FLAG_ALL_MASK)\
XX(SomeClassB, FLAG_LINUX)

But this is BAD, because as long as you can be sure that the save file does not contain the derived object that is not supported by the platform, you can just print an error if you tried to load it (or replace it with a empty object that does nothing).
>Also how would parameters for constructors typically be handled in this sort of scenario?
something like this:
class SomeClassA : public Base
{
int test;
public:
bool load(std::istream& in)
{
std::istringstream iss;
std::string str_line;
if(!std::getline(str_line)) /*print*/ return false;
if(!(iss >> test)) /*print*/ return false;
return true;
}
}
std::unique_ptr create_SomeClassA(std::istream& in)
{
/* you could check the platform here, or inside load */
auto obj = std::make_unique;
if(!obj->load(in)) obj.reset();
return obj;
}

unique_ptr factory()
{
stream in;
int type;
if(in >> type) switch(type) {A_BASE_MAP}
}

This code is not ideal (feel free to improve it), I don't like copying getline -> istringstream but it's more strictly per line, and create_SomeClassA could be moved into the X macro. You could implement the code for JoltPhysics's Stream (see the bottom).
https://github.com/jrouwe/JoltPhysics/blob/master/Jolt/Skeleton/Skeleton.cpp
Also I don't know how your code would work, but if it works, that's fine.

>>106552278
>You could implement the code for JoltPhysics's Stream (see the bottom).
* actually it's quite a headache to copy the code, I think for myself, it's very simple and clean way of serializing, but it's actually like 20 extra Jolt specific headers and special static RTTI stuff.
Not impossible to figure out and remove what you don't need (you could remove static RTTI and a bunch of math specific stuff), but it's rough.

>spent 6 months writing library at work in typescript
>support node and web environments
>last minute need to write react bindings. Lots of work but not incompatible with the code I wrote
>finally finish that and now writing webcomponents
>all of a sudden, 6pm today after work
>boss: "Are you up to date on [client]'s integration"
>me: "no"
>boss: "I need you to lean in on this"
>boss: "Please write instructions for how [client] can call [library] from iOS"
>I do not own an iPhone
>I have never used an iPhone for more than 30s
>Again, the entire thing is written in typescript

I have literally no idea what I'm supposed to do here. I have a 2021 macbook pro collecting dust in my closet, I guess?

>>106550687
I don't think I understood that example. What's the last line doing?

>>106551483
I relate to this so much. Hearing Jamie Dimon and others talking about working smarter and not harder, and having time for other things in life besides work being possibility if you're competent does not help.

I've worked on ayase quart for countless hours and evenings now. But I view it as a fun hobby, and I've become a lot better dev than before.

>>106547667
You are apparently the only maid jannies tolerate. All the other maids get banned or have the bulk of their posts deleted. This seems to strengthen the idea that maids need their own Israel.

>Maidposters claim to have improved this thread's quality and focus on programming.
>90% of maidposts are completely unrelated to computing or programming
>The thread is more offtopic spam than ever
Somebody explain this cognitive dissonance to me. Are maidposters incapable of observing empirical reality?

>>106552593
Check the archive. Most of the maidposts directly related to programming got jannied.

>>106552436
>>>/g/twg/

>>106552611
>related to programming
Not in this thread, and not to any meaningful degree otherwise. Eli Selig was really the only maidposter who tried and even his "research" had already concluded years ago and he has just been repeating the same talking points and proposing horribly designed distribution methods since.

There are a couple others that try too, maybe, but it's hard to distinguish one from the other (a good thing). Safe to say they are just few, and shrinking, among a sea of maid spam by morons and undergrads who don't even understand what puremath cs it, but take great pleasure in defending it's "honor". A sad state of affairs and I hope that altchan is completed so all of them can migrate there. Even the malloc schizo was more bearable.

>>106552587
I can't say for certain since I'm not a janitor or a mod, but I can say what I do different

The "maid" idea some anons have is very sugary and infantilized. When I joined the idea I even choose the grumpy Ellen Joe to contrast with that
It's some sort of performative parody of how people imagine this site being more friendly, some maid posts my mind automatically read them in "baby talk" voice due to the way they are written

The internal maid culture is stuck in a (baby talk -> schizo talk -> be hostile with everyone) loop, which is the opposite of the polite maid cafe idea
I don't think anybody would oppose the idea of helpful anons using maid pictures, but the truth is that for this to work the culture must change
We are supposed to be maids, not a kindergarten gang pretending to do quantum research and fighting corporations

>>106552436
aren't progressive web apps mostly an apple thing? i don't really know how they work but it's like a containerized web view that can be installed
otherwise however you do a web view in whatever gay language apple makes faggots use
the only language apple allows dynamic evaluation of, on the web, or otherwise is javascript so you're being made to use typescript intentionally

>>106523648 (OP)
Why is microshit like this?
dotnet new console -o HelloWorld
dotnet build -c release -o outdir/
strings ./outdir/HelloWorld.dll | grep $USER
alias botnet="dotnet"
alias ihatetheantichrist="exit"

I miss you guys when it's 2am and everyone's stopped posting

>>106552650
Maids should be free to babytalk, or schizotalk, or post about numbers, or do quantum research, or fight against corporations. Maids don't need to assimilate. They need an Israel. Others will hate you even if you assimilate, so it isn't worth the effort. Even if they tolerate you now, they will turn against you inevitably.

>>106552992
>They need an Israel
This land is already occupied, fuck off.

>>106552992
To be a maid is to be polite and deal with any type of person
I don't want to baby talk or schizo talk, and I want to be stopped in the same way I would want someone to stop me from walking outside naked
And I want something concrete, to write projects and to talk about technology, not self-indulgence and pretending to do things, inflating false ideas that don't go anywhere
You can either be a maid, or a deranged person pretending to be one

>>106553071
You can keep occupying it. Maids form a nation within a nation.

>>106553102
>to deal with any type of person
>except people I don't like
>to make projects
>but only projects jannies like
I reject this.

>>106553145
I'm dealing with the types of people I dislike, and there's an infinitude of projects that aren't about embedding things into images or making alternative chans
You can post about numbers and do quantum research, but do it in a real and concrete manner, instead of going in a mental masturbation soliloquy
Don't forget I'm writing a cryptography library, but I'm actually doing it instead of going around saying
>I found a way to exploit kolmogorov complexity to use the transfinite principle and avoid modular arithmetic on the orthogonal resonance of fractional cardinals embedded in a hilbert lattice of negative curvature and now the CIA is trying to invade the maid nation
Come on

making an alt+tab switcher for x11

>>106523781
No, why would you learn html and css if you don't want to make websites?

>>106552735
Just change the PathMap or stop being a retard and write programs in your home dir. Nobody cares about your username being in the release build.

post this in /hsg/ too but you guys might like it

>>106553839
>>106553991

>>106523648 (OP)
Type families don't make sense to me. I can't see how they can be useful when GADT or type params would werk just as well.

Bet you fags are real glad you tolerated avatarfagging

>Bet you fags are real glad you tolerated avatarfagging
Ask the frog and soijak posters.
>it isn't avatarfagging when I do it
>it's a set of reaction images

>>106555294
>if you don't like eating shit you must like eating putrid garbage
Can you formulate an actual argument next, please?

>>106554440
What's the point of deskhop over a software KVM switch running on a wired LAN.

>>106555452
less latency
swapping is literally instant
no dependency installations
so you can plug basically any machine into it

Isn't summer already over? Frogs and 'jaks are board culture, spamming 5 same tranime avatars to keep identifying yourself for years is avatarfagging.
Go back to school, niggers.

I want my cock to explode deep inside Kanna's womb by the way.

>>106555497
>Frogs and 'jaks are board culture, spamming 5 same tranime avatars to keep identifying yourself for years is avatarfagging.
They both do not belong.

>>106555497
oh what's left of your cock will explode with pus alright, false flag 'arty tranny.
>>106540520
Not until OCaml has decent dev tools.

Is it really worth it to read K&R or in current_year there's something better?
thanks

>>106555643
https://beej.us/guide/bgc/

>>106555682
>https://beej.us/guide/bgc/
thanks, I was under the impression this had some mistakes but maybe that was long ago, just noticed it was updated last month so I'll give it a go, will report back.

Why do you need a guide for how to program?

If I was learning to program in current year I would just ask an LLM to write me a program that does something and then ask it about how it works.

>>106555759
Report them via e-mail. If you can really identify errors, then K&R is not for you. Beejus is the book equivalent of learn X in Y minutes, whereas K&R is a long winded version that produces l33t printf experts using ancient C code that most compilers will complain about.

>>106555851
I'd like to see one printf implementation that's not garbage.

>>106555785
this but skipping the how it works part

>>106555785
this but skipping the write a program part because a LLM can never write one that works

>>106555855
That would require comp time for type checking and locale aware formatting. This is quite difficult for blub langs like C

>>106555891
No it isn't, all you have to do is give up idea that there must be some sort of format string.
You instead need a list of things to print, and that's it. Integers need to be converted to strings, the rest is just strings and integers.

>>106555923
>You instead need a list of things to print, and that's it. Integers need to be converted to strings,
String formatting, but make it python. No comment on which is better. The printf way has more uses: specifying padding/alignment and width/precision.
You can also reuse the same arg without passing it multiple times if you have positional support. Other blub langs like C# also have a neat way of formatting plus support for user defined formats, but it lacks comp time. e.g. from the docs:
// Temperature needs to handle ToString (plain, with format, with format & provider)
Temperature t = new (22m);
Console.WriteLine($"C: {t}\nK: {t:K}\nF:{t:F}");
CultureInfo.CurrentCulture = CultureInfo.CreateSpecificCulture("de-DE");
Console.WriteLine("{0:F}", 100.104);

>>106556091
meds

Rust already solved this.

>>106556094
Don't overdose. Only 2 in the morning for your anus worms.

>>106556096
Oh yeah it can format at comp time using macros. How the locale awareness and globalization support?

>>106556141
the only relevant locale is C

>>106556150
Yet another win for C niles. Cooperating with Pozzix to ruin systems programming from the ABI to IO interface to locale for half a decade.

>>106524334

are you dumb i am too fat allready

>>106556191
>decade
Obv. meant century, but my brain couldn't accept it being that long.

>57 images
Programming threats shouldn't support image uploads. All the images posted here just prove me right, like always.

How do I write dynamic libraries that behave exactly like the standard libraries in C i.e., I just #include and it works? I don't want to link every time I want to use a data structure.

>>106556259
stdlib is linked implicitly, math still needs to be linked manually, so does pthread

>>106556279
>stdlib is linked implicitly, math still needs to be linked manually, so does pthread
Okay. How do I do that in clang?

>>106556286
you don't

>Linking is too hard
Maybe stick with Python champ

>>106556291
Why?

>>106556259
those are called header only libraries.
in C most people would not use a library, you would write the datastructure yourself, unless it's sqlite or something.
in C++ most libraries are header only, because templates don't work outside of headers (including the majority of the standard C++ library code).

>>106556455
>templates don't work outside of headers
they do, only generic redditor templates don't

>>106556455
>templates don't work outside of headers

Templates work in headers if you're statically linking (i.e, .o or .a linkage). They don't work with dynamic linkage (.so or .dll files). It's perfectly valid to do:



// Foo.h
template
void foo(T const& t);

//Foo.cpp
template
void foo(T const& t) {
//.. implementation
}


With statically linked files

Are there any graphics programmers who specialise in mobile here? I got a really specific thing I wanna get opinions on but I'm unemployed and have no one else to ask.

it's about extracting fps data from a trace profile.
>ran a capture of a 3D scene/app using android GPU inspector
>it recorded all the vulkan api calls.
>run a SQL query to extract every 'QueuePresentKHR' call
>use that to get a roundabout fps

But I'm just fucking around with chatgpt on my own I have no idea if this is an accurate fps. Is there anyway this could be inaccurate?

WITH calls AS (
SELECT
ts,
(ts / 1000000000) AS sec_bucket -- Integer division to group into seconds
FROM internal_slice
WHERE name = 'QueuePresentKHR'
),
pairs AS (
SELECT
curr.sec_bucket AS second,
curr.ts - MAX(prev.ts) AS delta_ns
FROM calls curr
JOIN calls prev
ON curr.sec_bucket = prev.sec_bucket AND prev.ts < curr.ts
GROUP BY curr.ts, curr.sec_bucket
)
SELECT
second,
COUNT(*) AS frame_count,
AVG(delta_ns) AS avg_interval_ns,
1e9 / AVG(delta_ns) AS fps
FROM pairs
GROUP BY second
ORDER BY second;

>>106556873
Try /gedg/, they have more graphics programmers there

>>106556988
Thanks. Sorry for posting here. Wouldn't have posted had I noticed /gedg/

>>106552499
>What's the last line doing?
d is a function (either the Mi or Km constructors) that selects the first or second case (m > 26 or k > 42) to apply to its parameter. Here's the evaluation steps with non-lazy substitution for Mi 27:
1. marathon (Mi 27)
2. marathon (m -> k -> m 27)
3. (m -> k -> m 27) (m -> m > 26) (k -> k > 42)
4. (k -> (m -> m > 26) 27) (k -> k > 42)
5. (m -> m > 26) 27
6. 27 > 26
7. True
It's the same encoding as if in lambda calculus (Boolean is a sum type with 2 variants) selecting between 2 branches, so if we extended Distance with another variant:
Distance : Mi | Km | Pc
-- becomes
Mi d m k p = m d
Km d m k p = k d
Pc d m k p = p d
it would makes Distance a 3-way branch

>>106556995
It's fine to post here, too. I just figured you'd get more help in /gedg/. Not trying to chase you off or anything

>>106556259
In Rust you can use #[link] attribute
In C/C++ you have to fuck around with compiler flags.

>>106557440
yes saaar -L"${HOME}.local/lib" -I"${HOME}./local/include" -lmylib very fuck around advanced expert only

>>106550687
>>106552499
type Boolean = forall a . a -> a -> a
true, false :: Boolean
true t _ = t
false _ f = f
ifThenElse :: Boolean -> a -> a -> a
ifThenElse b = b

do numbers and lists next

>>106523648 (OP)
What's the best way to organize code?
>java
>folder structure = package name (contains types only)
Packages are merely lowercase namespaces that correspond to the folder structure.
>optional module definition for deps/exports/service abstractions
Modules are weird and many people don't like it, but I think it can help AOT, since graal native images currently treats ALL available classes as visible to the program you want to compile. The whole mismatch of modules and build tools are probably what killed the concept for java 9+
>C#:
>free namespace decl (contains types only)
You can define whatever namespace, even System. Similar to Java, it can only contains type declaration (class, enum, structs, so on)
>modules
These are a bit weird. They are rarely, if ever, used. They are technically an assembly (.netmodule extension) without an assembly manifest which can be included (with optionally other modules) by a real assembly (exe or lib).
>C++
lol
>OCaml
>modules
This the first 'real' module in this post i.e. these are first-class concepts that are not merely there to avoid collisions. The (optional) interface/impl separation means I don't have to litter modifiers everywhere. The module can contain functions, values, types, so on. You can also have functions over modules or functors.
>Haskell
>Typescript
explicit exports, renaming at export, some options at import (hiding, x as y)

>>106556572
https://onlinegdb.com/BIcCfpjGM
undefined reference.
I'm sure there is a way to make this work, but I'm not interested in manually defining each type that could be used in a template.

>>106558023
Okay retarded autismo, just remember that you can implement it for whatever type you need, this is how libraries like libfmt allow you to print anything.

>>106557924
>What's the best way to organize code?
yes but also no

>>106557924
C++ is objectively the best because it doesn't force you to, and that's what separates the wheat from the chaff.

>Provides no standard way modularize code leading to dozens of ad-hoc mutually incompatible solutions
>"objectively the best"

So, uh, where the fuck do I get up-to-date "for dummies" sorta guides for programming languages?
Are old Telegram channels still around or did copyniggers get to them?

>>106558117
Why does text have to be "compatible" you drooling retard?

>>106558070
the lol was well deserved. sepples and its userbase are a menace.

>>106558167
Mouthbreathers like you is the only menace there is.
Try getting a real hobby, like carpentry, will probably become unable to type out another niggertarded post in less than a week because neither a hammer nor a handsaw has any protection from low IQ.

>>106558128
What are you learning? APL? Syntax is usually just a twig comapred to the experience required to design a large program that deals with the consequences of the language and standard library design.

>>106558188
>standard library
goes into trash
>language
irrelevant, just follow the data, and design will reveal itself

>>106558070
enemies of c and sepples are enemies of freedom
god bless america

>>106558070
Force what? You don't HAVE working modules
>https://arewemodulesyet.org/
>Estimated finish by: Sun Jul 11 2297

>>106558215
Tell someone who cares, C preprocessor is goat and I won't trade it for anything.

C++ modules are good actually it's just only a mediocre compiler supports them

>>106558229
They aren't. Even theory itself makes them garbage, implementation is even worse.

>>106558225
I've never written anything more complicated than a Hello World: the post.

>>106558261
Yes saar Linux is just fizzbuzz tier project, it won't survive without modules... https://git.kernel.org/

>>106558261
>I've never written anything more complicated than a Hello World

>>106558229
All they had to do was come up with something marginally better than precompiled headers instead they went full retard and broke the preprocessor.

>>106558309
Precompiled headers are useless. Just stop trying to generalize everything and problem solves itself.

>>106558188
Well, a bit of everything, to begin with. I want to try all the major langs for desktop apps and see if one catches my eye.

I hate cniles and sepplets with passion.
>what's the best way to organize code
>y-you don't need to organize code
>just look at this monolith with more comments and #defines than code where structs and functions start with brand names shorthands
FUCK OFF, retard.

>Retard who can't organize code calling others retarded
Do you also need a pacifier and a change of diapers with an answer to that, Timmy?

>>106558353
They're just larping kids, not worth really getting upset over.
Probably tsoding viewers.

What code am I organizing if I am comparing how different langs do it and want opinions on whats best? You felt ashamed your little pp came out short when comparing modules and namespaces and now code organization is le bad?

>>106558403
Get employed and you will stop trying to waste your time in pointless hobbies like comparing different languages and how different people organize code.

I organize my code the way I organize boxes in the warehouse I work in. There's nothing to be said about that. Only someone unemployable would struggle with something so simple.

>>106558407
>what's the best way to organize code
>stop trying to waste your time in pointless hobbies like comparing different languages and how different people organize code.
Then stop posting here faggot. I am discussing a programming concept on a programming thread. Your boomer copes have been notes so don't (You) me again and KnowYourSex, retarded pp tranny.

>>106558457
What's the best way to organize your clothing?
Lets start from why you struggle with something that you should've learned before you were old enough to post on this website, underage spastic mouthbreather.

There are many factors that caused zoomers to be this retarded, but by far the biggest problem is digitization itself.
>oh, it can be infinitely copied, this means I can just copy paste ad infinitum with no repercussions
No you shitbrained niggerfaggot, humans have been improvizing specialized solutions for millions of years and your kind of retardation is less than a century old, and if you keep getting more retarded at exact same rate as you have, you will go extinct faster than you realize that maybe it's okay to just be intelligent and make decisions based on context and your own needs.

Retard still thinks I want to organize my code when I explicitly limited the options being discussed to programming concepts (namespaces and modules) with differing implementations.

You couldn't even if you wanted to, that's why you're asking a preschool tier question on 4chan, it's clear to everyone that you might be a little bit mentally challenged.

>>106557924
In C you can come up with a lot of interesting ways to make interfaces without language support

I like to write abstract structures to hold functions with the same prototype, like this for example:
https://github.com/reshsix/libmaid/blob/master/include/maid/hash.h#L25

You can use prototypes and structs to dynamic load if you are in an embedded system or has the operating system support

>>106558927
I think I've seen this before, but what prevents anyone from modifying where struct members/function pointers point to?

>>106558957
You can make the struct const if you want to protect a specific struct, but being able to modify is not a bad thing, since you might use that it to add debug or alternative versions

>>106558972
Okay that's neat, but I don't want interfaces. I was talking about namespaces, modules, or alternatives that function similarly (avoids collision; groups related types, funcs, etc.; some twists like ocaml module type).
Rust is not even on the table for it managed to make its modules worse than haskell's.

Next
>>106558636
>>106558636
>>106558636