How has programming changed in the past decade besides ai? Everyone was doing react and electron in 2015 and it seems like they are just still doing it.

>>103889974
studying operating systems

>>103889974
>CL with batterie included
>even has a GUI library
>doesn't support windows
well, there goes my tinkering with GUI and a bit of gamdev

>>103889986
What has changed is the mindset.
Back in the day Sepples flat out refused to see how utterly garbage the code generation for their language was, which naturally turned them into the laughingstock of anyone even remotely competent. Nowadays they've turned defensive, spouting nonsense like "but muh game development" (you can do OOP in C too) or "I'll pay off the technical debt later" (lol no, you won't).

>>103889986
Programming hasn't changed at all. The people you're talking about aren't programming.

>>103889986
>Besides AI
so, Mrs. Lincoln, how was the play?

>>103889974
program a simpler danbooru-clone using drogon-framework

>>103890067
I love how you couldn't write "better" with a straight face. It had to be "simpler".

>>103890044
meds

>>103890082
less features -> simpler

>>103890044
... or just a simple "meds" (read and weep).

>>103890127
How can you care about that while using the satanic rube goldberg machine of pure inefficiency know as a web browser

>>103890146
By differentiating between "using" and "defending".

>>103889974
learning c
>When a fixed-length array a is used as input in c, is it possible to get sizeof(a) inside the function?
>Depends.

>tired of needlessly complicated programming languages
>want to try something new and fun
>discovered forth because of some fag here
>absolutely unreadable garbage what the fuck is all that?
>spent some time reading about it and gathering old literature
>toying around with gforth
>intense desire to take it further by writing my own interpreter on a dead console or cpu
Ah fuck it's the lisp mind-virus all over again

>>103890044
The abstractions are zero-cost

>>103890596
Called it: >>103890127
>your fellow sepples are violently disagreeing with your notions, hallucinating wildly about zero-cost abstractions and zero-cost exceptions and zero-cost templates and zero-cost everything.

>>103890409
You cannot provide arrays as arguments in C, they always decay to pointers. If you sizeof(a) the argument it will be the pointer object's size. You can provide a struct that has an array member and the entire object memory (including the array) will be copied "by value", so you could do sizeof(s.a). However, remember that you can only pass "by value" in C, which includes the pointers that arrays decay to as well.

>>103890747

void f(struct Array a)
{
    printf("%d\n", sizeof a.array);
}

void g(struct Array *a)
{
    printf("%d\n", sizeof a->array);
}

That's pretty cool, both work

>>103891061
You're not passing arrays here, no matter what names you're giving your structs.

>>103891097
Yeah, that's what anon said. I didn't know wrapping it in a struct was enough to preserve size info

>>103890747
>You cannot provide arrays as arguments in C
yes you can. however arrays as arguments are parametric, hence why you can substitute an array of any size or even a pointer.

>>103891218
No.
>t. very much an engineer

>>103890082
but simpler is better

>>103891311
Is it going to be simpler, or just easier? Are you going to end up with more or less memory allocations? How many cycles is your code going to spend in memcpy?

mmapfag what do you do besides spamming your schizo rants?

How hard is it to go from raw sockets to an HTTPS server

Bruising egos, and with pleasure.

>>103891407
HTTP? trivial. you don't even need a library, just raw syscalls.
HTTPS? non-trivial, you want a library at least for the TLS. Still easy after that.

>>103891407
You need at least to have some cipher negotiation logic, and on kernels that don't support kernel TLS (Linux does) also do encryption and decryption that you feed into an HTTP parser, which is probably the hardest part (because HTTP is a text protocol with optional encryption, chunk-wise transfers, encodings ...). And if you're talking server you probably want to do it non-blockingly, meaning you have to learn how epoll (or kqueue or completion ports or whatever the fuck exists on your machine) works, because you really don't want to have one pending thread per connection (due to mode switch costs), so you either have to use state machines or userspace threads.

>>103891525
>optional encryption
Errr, compression, not encryption.

>>103891248
No you can't, don't be confidently retarded.
>Since arrays decay immediately into pointers, an array is never actually passed to a function. You can pretend that a function receives an array as a parameter, and illustrate it by declaring the corresponding parameter as an array
And yes, a [static 1] parameter still decays to a pointer.
https://c-faq.com/aryptr/aryptrparam.html
https://yarchive.net/comp/c.html#5
https://www.bell-labs.com/usr/dmr/www/chist.html

>>103891342
it will be simpler and easier.

coming up with some design which will be easy to contribute to by others and which should also scale well into the future (extra features etc.); that could be difficult.

the problems you speak of are easy to solve, given good design

>>103891854
>not answering the questions
All anyone needs to know.

>>103891886
Nobody cares schizo go do something productive

>>103892177
i do
you got btfod. kek

I do something productive by destroying your ego.

>>103890409
using sizeof as a way to determine an arrays length is a common trick but I think a bad habit especially for a learner. You should expect to store the length of your arrays in a separate variable, and pass them around together. You can create a struct to store them alongside each other if you'd like. Using sizeof to compute the size should be considered an advanced trick only to be deployed once you've gotten skilled at the above.

>>103890747
This is correct, but missing an important note. Consider the following:

void f(int n, int (*x)[n]){
    printf("sizeof x = %zd\n", sizeof x);
    printf("sizeof *x = %zd\n", sizeof *x);
}

In advanced code, we can use this construct to pass arrays of runtime determined size while enabling use of the sizeof trick.

>>103892383
Semantically, the array is still never being passed to the function directly, and that's all I was saying. The object being provided "by value" to the function is a pointer to an array, in this case a VLA, which is an optional language feature.
>If the size is an integer constant expression and the element type has a known constant size, the array type is not a variable length array type; otherwise, the array type is a variable length array type. (Variable length arrays are a conditional feature that implementations need not support; see 6.10.8.3.)
Might as well also point out that GNU has an extension that allows parameter forward declaration

void f(int n; int (*x)[n], int n)
{
    // ...
}

https://gcc.gnu.org/onlinedocs/gcc/Variable-Length.html

>>103892743
NTA but depends entirely what you mean by "passing an array", passing the pointer is probably exactly what he meant by passing an array

>>103890549
Not him, but:
>absolutely unreadable garbage what the fuck is all that?
Once you've written a dozen of projects in it you will be able to read Forth. Keep at it, it has taken me longer than every other language to get comfortable in.
>intense desire to take it further by writing my own interpreter on a dead console or cpu
I've created a custom CPU arch emulator and a Forth for it to scratch that itch. It's a lot of fun.
>Ah fuck it's the lisp mind-virus all over again
The difference for me was that I actually stuck with Forth for some reason.

>>103892830
It does not depend on what you mean. "Passing an array" means passing an array and the distinction is important throughout the entire language.

>>103890549
sounds like you are interested by powerful evaluation models/compilation models/abtract machines and by expressive languages anon
it's time for you to rake your brain and make your own ;)

>should we focus on the development of good optimizers to improve the general quality of code out there, and thus making things a little bit better for everyone?
>no! we should make new languages that repeat all the mistakes others have already done in the past, thus adding to the pile of garbage that is software development
Autism apparently doesn't help with pattern recognition.

lol im going fucking nuts here
i wasted hours of trying to understand this shitty thing
so in YAML
if I have something like
somekey: "[a,b,c]"
this should be a string and not interpreted as a list right?
then why the fuck in Go Im getting the error when asserting this shit to be .(string)
error: panic: interface conversion: interface {} is []interface {}, not string

where the fuck did []interface{} come from?

>>103890549
Forth wasted an absolutely embarrassing amount of my life, with really nothing to show for it. The fastest way out is to follow the meme and implement your own.
If you want something ultra simple but less madness-inducing, try a soft APL like Ivy or https://beyondloom.com/decker/lil.html
Avoid K, it has a very similar effects, and atw has some similar charisma to Jeff Fox even if he's less of a writer.

>>103889986
Rust saved programming with the promotion of memory safe languages. C++ is on its death bed now

>>103893471
>no! we should make new languages that repeat all the mistakes others have already done in the past
Talk for yourself. The reason I'm being on making new language is precisely to fix the stupid limitaions of current languages.
I would put emphas on how much low level can the language be (pristine integration between the language and assembly, manually allocate or constraint registers), on how much control on over *every* step of the compilation and on how the code is optimized, which optimizations are turned on and off and more. Also be able to write or generate code in a high level language or IR and let the compiler optimize when you don't care or think he can do a better job.
If you think those things are mutually exclusive you have imagination issues.

>>103893695
>Forth wasted an absolutely embarrassing amount of my life
Why "wasted"? You can write programs in Forth to solve real problems. I write programs in Forth to solve problems. I could solve those problems in other languages but I choose to solve them in Forth.

>>103893903
See you in 2035 then, when you're done.

>>103893933
What's hard is to come up with a clean way to have hooks in every step of the compilation process, or being able to declare stuff about them, how to integrate several DSLs or IRs together. Basically how to reinvent a much fancier Lisp's eval function.

I'm too busy with my daily job and (paid) side projects to try to do anything fun, but I'd love to do something comfy with pyqt6 and working on my emacs packages.

>>103890030
Fucking sweet

>>103890030
>>doesn't support windows
I wish more languages had this feature

>>103890030
the gui library seems to be using TK and sbcl works on windows so why wouldn't it support windows?

>>103890030
Surprisingly in-depth, I'll look into it more.

Does LTO (Link Time Optimization) break LGPL?

>>103894926
does running code on your CPU break lgpl?

>>103895012
ESL

Holy fuck, debugging async SSE-related shit is so ridiculously hard

>>103890549
>>discovered forth because of some fag here
Probably me. My Tetris is going well, but I'm at another point now where I have a decent set of fun features and the next step is a score table and gravity curve and that doesn't sound very interesting.

I have an intense desire to share things I know and how things work. If you'll humor me.
https://paste.myst.rs/5n3aoq61
This program takes about 30 seconds to compile, or about 7 if I fastforward the C64 emulator. The attached image is a few versions ago, including different profiling colors.

(line 5) The word B/, (b-slash-comma) takes a colorcode C and compiles 3 instructions: LDA #C, EOR $D020, STA $D020, a routine for flipping the C64 border color.

(line 6) The word PROFILE compiles two of these routines with a JSR in the middle, then adjusts the latest word to point to this new code, instrumenting it with border-flipping behavior.

(line 9) IF\ does conditional compilation, allowing me to recompile without profiling by putting a sentinel 1234 on the stack.

(line 12) ELSE: (else-colon) is effectively a syntax macro, expanding to EXIT THEN. Here's the usual IF syntax in Forth and my new word, its expansion, and an explanation for each in equivalent C:

> a IF b ELSE c THEN d
> if (a()) {b();} else {c();} d();

> a IF b ELSE: c
> a IF b EXIT THEN c
> if (a()) {b(); return;} c();

(line 36) N: parses and evaluates a word. It's used in loops to define data-parsing words, like:

(line 39) P: parses 8 bytes $yx, expanding into $0y0x-2 with a shift and mask, then compiles with , (comma).

DurexForth has two important global pointers: HERE is free memory to compile code to, and LATEST is nametable metadata. Using dF version 4:
> hex here u. latest u.
> 3994 9642 ok

The colon word defines new words:
> : double dup + ;
> ok
> 4 double .
> 8 ok

Colon parses 'double' from the interpreter's input buffer, subtracts 9 from LATEST, and stores length, characters, and code pointer:
> latest dump
> 9639 06 44 4f 55 42 4c 45 94 .double.
> 9641 39 01 56 08 38 0c 2d 2d 9.v.8.--
> ( ... etc ... )
> ok

Straddling the two lines is the code address $3994, the above value of HERE. The interpreter switched to compile mode, compiled JSR DUP and JSR PLUS, then the semicolon word patched JSR PLUS into JMP PLUS and switched back to interpreter mode.
> hex 3994 dump here u.
> 3994 20 66 08 4c 97 08 30 30 F.l..30
> ( ... omitted ... )
> 399a ok

Maybe you'll recognize the opcodes $20 JSR and $4C JMP. We can verify the compiled code addresses by searching the nametable:
> hex ' dup u. ' + u.
> 866 897 ok

If you want you can dump these two routines and you'll see opcodes for the LDA CLC ADC STA INX DEX RTS you'd expect for 16-bit routines on an X-indexed zeropage datastack.

There's a few more details like extending the compiler with words like semicolon but this is most of the important bits for understanding the whole system.

>>103895801
I rearranged my drawing flags and now a line sweep no longer redraws the hold piece and meets the sync deadline one frame earlier! Score!

If a piece of software is made and hosted by a government agency on a non-indexed and non-advertised but otherwise unsecured website (I have no idea if this is intended), how likely is it that they'll sue my ass if I make a better derrivative of it? I don't intend to make money selling it, but making it readily available and easy to use for the public would put quite a few leech corps out of business.

>>103896415
They will offer you a job, and if you refuse they will take what you wrote and imprison you for having 10TB of CP on a 50 year old calculator

>>103893759
Strange how C++ has been "on its death bed" for like two decades

>>103896597
it is now they are pushing the (((safety profile))) (((contracts))) etc bullshit demanded by 5eyes

>>103896597
It's been in a death spiral since C++17 was a nothingburger and they didn't add anything decent since, just more complicated bullshit full of UB. Now Bjarne, Herb, and Gaby are really sticking a fork in it

>>103896597
The latest standard is the turning point, so yes it was doing poorly before.

i think honestly we need language turnover, its a good thing let languages die but also kinda live on nichely and resurface

what if c++ without ub

>>103896718
there's no C++ without ubab

>>103896718
rust

>>103895801
>(line 9) IF\ does conditional compilation
I want to call attention to this. Can you add conditional compilation to your language with one line of code?

: if\ ( f-) if postpone \ then ;

This works because comments are not built into the compiler, the word backslash \ is just another word in the dictionary, and the interpreter's input buffer is made available to all user code as it is executing it. Backslash just drops the rest of the interpreter's line when you call it, so I just have to call it within an if statement.

The trouble though is that it is marked 'immediate', meaning you cannot directly call it within a word definition, the interpreter always calls it immediately even when compiling. The word POSTPONE, however, parses the next word from the input buffer, and if it's immediate will compile a call to it instead of executing it immediately.

>legacy django app
>2500 database queries to load the home page
I am in pain

>>103897045
thank god i never got hired for a django job 13 years ago so i never had to deal with that shit lol

>>103897045
curious what that home page looks like

>>103890127
...what the fuck is wrong with this language?

>>103897239
post your language? that's what i thought

Why is corecursion so underrated, bros?

>>103897282
lambda calculus is based on natural deduction, what we need are some big languages based on sequent calculus

just ordered a esp32, want to practise some cpp. I just got some noobie stemmaqt sensors then maybe I'll upgrade to a breadboard or something

>>103892743
>the array is still never being passed to the function directly
and nobody wants that either for anything much larger than a quaternion; it's horribly slow when the array gets larger

>>103897045
This is your brain on ORM. It's very easy to screw up and require 2k5 queries for something critical if you're busy trying to pretend there's no such thing as a proper query language

I hate C++ constexpr, it's so weak it can't even precompute an array.

>>103898055
for you >>103893471

>>103897273
Why would it require a new language to implement the basics of optimizations?
>what basics
I've already listed them.
>the compiler [is] too stupid to realize that RSP + 0x28 (char**) is never being passed to a callee
>the compiler has no ability to keep track of [object state which] can still cause changes to the buffer, so it just doesn't
>what is it with this miserable compiler's fascination with fucking std::string
>why is that exception handler still there?
>why are those calls to operator delete still there?
>why don't you fucking LEA an empty string into the register?

Although to be fair it's not like simpler languages have the ability to keep track of object state either, it's just that printf usually happens to have all relevant state in its arguments. But it's been more than thirty years since C++'s inception, you'd think that they had figured this shit out by now.
>inb4 but that's difficult
What, and Stroustrup didn't know that beforehand? Oh, let me guess, it simply wasn't a priority back in the day, and now it's too late and they'll continue to run around with those gunshot wounds until WW3? Guess the point of C++ was never to outspeed C in the first place, but to simply avoid the pains of memory management, and the rest can go to hell.

And somehow I don't think that adding a new turd to the pile of garbage that are programming languages would've helped with C's problems nonsense either.

>>103898055
Using G++? GCC has the same issue.

>>103898055
rust doesn't have this problem

>>103898433
Maybe, but Rust has Rust problems.

>>103898055
compiletimelets BTFO

>>103898528
Imagine not having your entire language available at compile time.

>>103897497
The alternative:
select *
from users
left join products
left join stores
left join history
left join comments

dumped to json then filtered

>>103898406
>but to simply avoid the pains of memory management
And as someone who's recently had to deal with old 20-bit code I totally get the sentiment. Having to deal with segments and paragraphs and FAR and NEAR pointers and interrupts and switches from RM to PM and vice versa and memory-based calling conventions with limited sets of registers was a major fucking pain in the ass. All of this has been done away with protected/long mode in userspace. C++ solves a problem that no longer exists.

It's like a worse C, but it's my worse C!

Anyone know where I can get the Qt C++ suite free? Is there a torrent?

>>103898850
nice undefined behavior in your add function

>>103898742
I hate this place, this prison, whatever you want to call it.
I can't stand it any longer.
I must get out of here.
I must get free.

>>103899023
>undefined behavior
>in anon's toy language
UB is a property of a language standard, bucko. His cool language is whatever his implementation says it is.

>>103899194
This is what I'll say when I sell my innovative language.

I know this is the programming thread but there's not many other places to ask this. Things keep getting more expensive so I might have to start working but being a NEET for years has ruined me. I can't drive and probably can't program but at least the programming I can relearn. Does anyone here have experience with freelancing or online only work? I'd appreciate tips to try and get some income.

>muh UB
yeah I don't give a shit UB is fast, I'm trying to go to the stars before I die not waste my life in academia with the pencil necks

imagine telling Tesla his electricity experiments were unsafe rofl

oh nooooo your tesla coil is too unsafe and has undefined behavior

>>103898876
why would you need to torrent it
it's already free just LGPL for most of the libraries (has to be open source if you static link) and i think most of if not all of the "enterprise" qt libraries are full GPL without a commercial license
so qt themselves? they've got some sort of faggy little installer that wants your email but you don't need to use that
most package managers ship it
you could build it from scratch from their repositories or the c++ package manager vcpkg can build it from source

>>103899516
the qt turotial Im using said to download the sdk but the online installer is a 10 day trial. are you saying just get the libraries and build with them instead?

>>103899535
no, you just downloaded the wrong installer, you do not need a commercial qt license, (most people don't that's why they push it so hard)
from the main page it's under qt "community edition" or on this page
https://www.qt.io/download-open-source
you should be able to pick the version(s) and libraries you want to install, and additional optional development tools and libraries

all of those other things are alternative ways to download the open source licensed version of the SDK

>>103899626
>>you should be able to pick the version(s) and libraries you want to install
*from the gay little installer after you log in
i think they did away with offline installers but as mentioned other people vendor it

>>103899626
I installed using chocolatey. thanks

>extract the code of some old BIOS
>see this
And there goes the reputation of the hardware industry.

>>103899368
we can already go to the stars and it's not about speed but safety

>>103900692
If that was the case we'd have space elevators.

matplotlib FUCKING SUCKS!!!!!!!!

>>103901044
too bad there is no better alternative

>>103901044
do you have ideas about better primitves for plot graphs like that?
I'm wondering if plot graphs are simple enough that lines and points are all you need

>>103899023
overflow is defined behavior

>>103900736
space elevators are a material science problem, not a programming problem unless you believe ai will solve all our problems then in which case it's a programming problem

>>103902014
Excuses.

>>103899626
so much for that. can't even get a simple program ran

>>103901044
time to switch to gnu octave

>>103904147
i'm not 100% sure that installing everything through chocolatey would have actually set everything up properly
when i said package managers i meant real ones like linux ones, or programming specific ones like vcpkg

>>103904529
Im on windows, so I'll try the vcpkg route

>>103904569
i also don't know about that, especially if you aren't 100% on how C++ dependency resolution works
their installer does integrate everything properly, the safest thing to do would be to just make a stupid account with a burner email
that's all they really want from you

I compiled Qt from source on Windows and then used it with CMake. Even got Qt Quick working.

Which ones do I want?

>>103904817
qtcreator and qt 5.12

>>103904905
why 5.12

>>103904147
he is telling you to use vcpkg but you are using a qt project, and vcpkg only works with visual studio projects OR cmake.
cmake is a world of hurt, and I think vcpkg does not make things as easy as it should be (it won't copy all the DLL's for you like every other vcpkg library, you need to access the DLL's in \vcpkg\installed\x64-windows\Qt6\plugins\platforms\ since QT doesn't use normal DLL's, they hotload the DLL so there is no "link" to automatically detect, so you need to manually install certain plugins using a cmake script, I think it's possible you don't need any plugins if you only include QCoreApplication (aka nothing), if you installed QT using the online installer it should make the plugins globally visible using enviroment variables, but that means if you copied your EXE to another PC, it wont run, QT's solution is to use static linking, which should work with vcpkg easily, but static linking breaks the LGPL license and now you need the commercial licence, unless your code is open source, like MIT/BSD or GPL).
And I think when it comes to cmake, there are NO tutorials or youtube videos that act as the best way of using QT. All the information is just scattered around and it's up to you to find the best approach, each with pros and cons. For example QT's official page has a deploy tutorial on cmake (without vcpkg) and the problem with it is that it only copies the plugins after you "install" (it depends on the global QT path when you are developing), that's not ideal in my opinion. On stackoverflow I have seen Qt6::windeployqt run in Cmake after build, I have seen people suggest manually copying each file manually. It's grim.
You don't need to use QT creator, i'm sure there are benefits like inspecting QT variables through the debugger but I don't know if that's useful.
Also cmake is a meta build system, unless it's a cache/syntax error, build / linking errors are more related to the build system (msvc), try building using command lines.

>>103904941
jesus christ, so much for getting a Hello World of this going.

>>103904941
>try building using command lines.
*on windows this is using a visual studio project, my bad.
essentially you go into your project settings and add the include path and library objects.
the problem is that QT needs code generation if you use QT widgits CMAKE_AUTOMOC, but you don't need it for a "nothing" QT application.
It's possible to build using the command line on msvc, but it's not worth learning BEFORE gcc style commands, since gcc is much more simpler and you can sort of figure out what msvc is doing.

You know GUI's should be a solved problem.
You should not need to design how it looks just what's in some context places.
The system theme should determine how the application looks like and user should be able to customize that to their liking.

>>103906265
Thank fuck you're wrong.

is there any particular course i should augment learncpp dot com with or is it enough to just go through it and then do leetcode or advent of code or whatever

>>103906993
none of those things teach you c++

>>103906993
the only way you augment is to start building something. small game, small program, something

>>103905006
>the problem is that QT needs code generation if you use QT widgits CMAKE_AUTOMOC, but you don't need it for a "nothing" QT application.
Strictly, you need codegen if you're using signals and/or slots (usually together). Yes, that is Qt widgets, but a few other useful things besides (like the timers).
The codegen tends to be very quick and easy, as long as you keep class definitions in headers and use cmake in the officially recommended way. (You can do it with plain make if you handle the fiddly details, and qmake is too stupid for words.)

>>103907121
huh?

I used to love OOP but now I'm starting to not like it that much anymore.

>>103901044
I think I'll use pandas.plot from now on and never make a complex plot

>>103908368
congrats on your second year of programming

>>103908602
congrats on your first year of programming

>>103908626
good luck on your entry exam

>made the mistake of pop() a list while iterating through it in Python

fuck

>>103908714
Rust doesn't have this problem

>>103908457
you can't go wrong with pure functional design

I am writing a browser extension to simplify bookmark management on a new tab page a bit.
So far I just added bookmark management but I want to add history, reading list, groups and page archiving there.

I hope I will finish it this week.

>>103909336
I am also building an rss reader for the browser.
For whatever reason there are no news tickers for linux and I had to write my own.
it is kinda stupid to build an rss reader outside of browser so I do it in a form of extension.

>>103904302
>>103908149
Reminds me of jean pierre hebert

>>103897107
A regular page that display some stuff for users, like things they have to do or things they haven't seen and must be.
The thing is a critical govt application used in a specific sector, and it landed squarely in the N+1 problem because of retards.

>>103908989
well when I have time to learn something else besides java and python i'll take a look.
speedrunning a master's trying to finish it in <= a year

>>103908368
Doing pure OOP demands not breaking encapsulation.
See where that gets you?
Back to pure functional structures.

>>103911381
is it really that difficult to write code that operates just on an interface and doesn't modify another object's internal state directly?
and before you start with this stale shitpost, just using getters/setters for fields does not count as encapsulation

>>103912035
It's not difficult, just retarded, even *if* you don't incur mode switches (prologue and epilogue overheads, instruction caches, and so forth).

>>103912035
>is it really that difficult
No but if you don't enforce purity like Haskell you get
>just using getters/setters for fields does not count as encapsulation
getter/setter galore which break encapsulation and you let objects molest other objects which defeats the purpose.

>write py script to read a directory of a thousand text files and search for specific lines (no more than 100) to print out
>runs in like 20 seconds which isn't terrible but trying to make it faster
>figure out that 15 of the 20 seconds are just doing the print statements
>run it on command line
>5 seconds
Make IDLE great again.

>>103913649
The simple answer is to not print.

>>103913649
Does python have the equivalent of the C setbuf function?
If you're using line buffered stdout, fully buffered stdout might be noticeably more efficient.

lol

>Latest GHC broke modulo operator
Uhh fizzbuzz sisters? We were the kings

Why shouldn't Thread.Sleep() be used?

>>103914750
>intentionally giving up YOUR cpu cycles for no reason
cuck behavior

>>103909846
Don't listen to him he's a shill. Use whatever you like.

>>103914750
There are times where it's unavoidable or even appropriate, but having your code dependent on that usually means you're doing something dumb. You usually want to trigger a wakeup on something actually related to what you're waiting for, not just trying to time it in a very fragile way.

>>103893471
Can you link to the commits you've submitted to LLVM and GCC?

>>103915493
No, I can't. But nice try to make me self-dox anyway.

>>103898406
>picrel
they made C less shit, it's called Rust and it's the best

>>103916815
Eww, a noob programmer

>Rust is the best
Lol
Lmao
Thanks, I needed that laugh: >>103898464

Thoughts on this book? Is it really that good for network programming?
https://beej.us/guide/bgnet/

>>103914685
Haskell is such a kitchen sink of a language

>>103917374
it's fine. Just run through it and learn something about BSD sockets. It is not complete, particularly about TLS or newer and more performant APIs like io_uring, but its contents will not harm you.

>>103917634
to be fair you can fizzbuzz in hasklel without modulo

>>103920182
And how many gigabytes of garbage per second do you produce doing so?

>>103920198
at least 10 in salty replies

Oh boy, I have a gnarly background task related error that I need to debug before our user base ends up triggering that error path. The bug literally hangs the entire app and crashes it.

>t. FastAPI backend with a bunch of SSE async horseshit and background tasks to save said SSE

>>103920213
Didn't you learn not to include delusional projections in your numbers?

>>103901131
plotly is better imo

wrote some Python sloppa and it's working I think I'm providing value to shareholders

>have to learn pandas for an online class
ugh. I'd almost rather just do everything from the ground up instead of learning to deal with modules. but it is a very popular module...

I just read The Clean Coder. It was ok.

Any other good code or code-related or tech job related book you'd recommend?

>>103921297
None, they're all garbage written by fake programmers.

>>103921297
Mythical Man-Month
J for C Programmers

>>103921297
The Art of Unix Programming

Are habits or motivation the key to learn programming and start programming?

>>103922790
Yes

>>103922790
No, but Ritalin is, at least for ADHD scum like you.

>>103922407
>>103922726
thx bros

>>103912075
>getter/setter galore which break encapsulation
Leaving aside setters (which a pure OOFP language wouldn't have because all objects would be immutable) getters don't break encapsulation because they state that the presence of an attribute property is part of the public interface exposed: the object has a Thing and I can read it (whether that's done by a direct record field read or some other mechanism isn't so important).
The place where the FP crowd lose their minds is when they decide that OOP necessarily requires mutability. That's wrong and always was wrong, and no amount of whining will change that basic fact.

>>103923840
The best thing about solo dev is that I don't have to worry about none of this crap. I don't even use git on solo code anymore.

if AI is going to make programmers unnecessary then why would software companies be necessary? wouldn't everyone just start using AI to produce their own software on demand, or at least "open source" AI generated software that anyone can download for free

>>103925705
There's a difference between programmers and code monkeys. AI is going to make code monkey unnecessary.

>>103925913
It's not the future when it's happening right now. That's called "present".

Is the only reason why DOS extenders bothered with PM because they were originally developed for the 286, which wasn't a 32-bit CPU? Because I can issue 32-bit memory accesses just fine while still being in RM as long as I'm running a 386 or newer. 8086 just stops executing anything with a 32-bit address override, and 286 flatout hangs the machine.

>>103925705
ai probably will be able to replace all glue code which is most of programming jobs but for a while we will still need humans to write libraries that do things

Is gleam a meme language? I was trying to find at least single video on youtube how someone implemented anything in it - OS kernel, TODO app, text editor, lisp, vidya, you know - shit ziggers and rustians did, but found only hype "version X.Y.Z of gleam has been dropped!" and "gleam is your next favorite language" and shit like this.

>>103920198
Just repeat "unused memory is wasted memory" until you believe it.

>>103928330
Isn't cleam an OTP language? So it's like a fancy Erlang, and you don't use Erlang to write kernels and games. (desu I don't know what you would use Erlang for outside legacy telecom systems, it's slow as shit, even with BEAM).

Math chads please help. If I have a list of numbers where each number is at max baseBits + (length - i - 1) bits long and numbers that come first in the list are less significant i can convert it to a single number like this:

let baseBits = 4
let numbers = [2, 3, 5]
let total, offset
let bigNumber = numbers.reduce((memo, number, i) => {
total = memo[0]
offset = memo[1]
return [
total + number * (2 ** ((i * baseBits) + offset)),
offset + numbers.length - 1 - i
]
},[0, 0])[0]


I can convert it back into a list like this:

let newNumbers = []
let devider;
for (let i = 0; bigNumber > 0; i++) {
console.log(bigNumber)
devider = (2 ** (baseBits + numbers.length - 1 - i))
newNumbers.push(bigNumber % devider)
bigNumber = Math.floor(bigNumber / devider)
}

But is it possible to convert it back into a list without knowing how long the list should be? I could do it if numbers that came first in the list had fewer bits but I can't figure out how to do it with numbers having more bits coming first in the list, and I need them to

>>103926454
The context switch to make a PM call while in RM is extremely painful, and extender would speed it up a lot by emulating as much as possible. The 286 cannot do the context switch without a reboot.

I'm writing a video file manager similar to music managers like Clementine, Amarok, Jukbox etc. The idea to scan a given path or paths for all video files, store them in a database along with their properties (title, duration, resolution, thumbnail etc.) and let the user browse, tag, filter and rate them.

I need to be able to uniquely identify a file even if it's renamed or moved, and I wonder what the best approach would be.

My first thought was keeping track of the file's inode, but that won't work if a file is copied or moved to another partition.

Another (better, I think) idea is just calculating a hash of the file (or a chunk of it, like first 1024 bytes to speed things up, since the files can be quite large). But some video files contain internal metadata that can be edited, which would obviously change the hash.

What other ideas are there?

Also, if I go with hashing, what would the optimal hashing algo and chunk size be, in terms of maximizing speed while minimizing risk of collisions?

>inb4 is it for porn?
What else.

>>103928515
That's not answering the question at all. There is no reason to switch from RM to PM in the first place, as long as you have 32-bit registers (which the 386 and subsequent processors had). The segment descriptor fuckery to enable unreal mode was a kludge once the CPU entered PM, which programmers only did to move beyond the 1 MB limit - but with 32-bit registers they could do so anyway without ever having to leave RM.

The only way this makes sense in my head is if DOS extenders were purposefully targeting a CPU that supported more than 1 MB of address space without 32-bit registers, which is just about the definition of the 286.

>>103928341
Who even came up with this mantra? When was the last time uncached memory accesses weren't black holes of inefficiency?

>>103928776
Protected mode is not just 32 bit addressing, and programs like DOS/4GW were for C programmers, not assembly programmers so much.

>>103928840
>Protected mode is not just 32 bit addressing
Yeah, that's the reason why programmers wanted to switch back from PM to RM, didn't they? Because PM required its own sets of drivers as BIOS interrupts and direct hardware accesses no longer worked; even in V86 this shit had to be emulated by the kernel through expensive mode switches (or invalid instructions, at least if you worked for MS: https://devblogs.microsoft.com/oldnewthing/20041215-00/?p=37003)

>programs like DOS/4GW were for C programmers
And? You can't tell me that compiler builders of the time were unable to make use of full 32-bit pointers. They already had to differentiate between 16-bit and 20-bit-through-2-16-bit pointers (because Intel, in their infinite wisdom, insisted on wasting 12 bits in the segment registers for some janky address wrapping, rather than doing a full 16-bit left-shift to maintain forwards compatibility with machines that came equipped with more memory). cdecl functions were usually called with two 16-bit arguments pushed to the stack anyway.

>>103928913
>And? You can't tell me that compiler builders of the time were unable to make use of full 32-bit pointers.
The problem wasn't if they were capable of trying, it is if they were capable of beating the solution that already existed and was dominating the market

>>103928989
Yeah, beating a solution that came with its own mini-kernel, additional mode switches, larger memory footprint, and loads of copies between conventional and extended memory ranges sure sounds hard to beat (lol). If you then consider that the 286 wasn't that popular to begin with I'm really struggling to see any reason why such a broken system became popular in the first place.

>>103898424
>C and C++ are LE BAD
>Rust is LE BAD
and what the fuck do you autist nigger propose then? Pure x86 assembly?

>>103912065
>File: cpp_is_garbage.png
>NtKillNiggersSlowSyscallWrittenInC
??? retard

>>103929086
C++ is unsalvageable unless you shoot Stroustrup in the head and cut all ties with previous standard versions, which is not gonna happen. Rust is in a slightly better position, but requires to jump through so many hoops that there's no real advantage to using it because you'll just end up with code generation that is on par with C's, which is still far from perfect.
>Pure x86 assembly
More like inline assembly in special cases, seeing as intrinsics in C still shit the bed, and will probably continue to do so in the future - but yes: https://en.wikipedia.org/wiki/Single_instruction,_multiple_data
>Currently, implementing an algorithm with SIMD instructions usually requires human labor; most compilers do not generate SIMD instructions from a typical C program, for instance. Automatic vectorization in compilers is an active area of computer science research.

>retard
And to you, nigger jew.

>>103928467
>But is it possible to convert it back into a list without knowing how long the list should be? I could do it if numbers that came first in the list had fewer bits but I can't figure out how to do it with numbers having more bits coming first in the list
If each element of a list has fewer bits than the remaining elements, then the list length can be at most log2(head), e.g. for a list (99, 56, 30, 12, 5, 3, 1) log2(99) = 7, not sure if that's what you're asking

Fuck, I spent the last month trying to figure out why I can't even follow beginner-level tutorials on how to make an inventory in Unity. It just refused to find the slot below the object I dragged.
Turns out I added a text field to my game objects and the scripts in the tutorials only set raycastTarget to false for the image itself. I also needed to set blocksRaycasts for the whole canvasGroup to false.
I could have figured this out so much faster if I had just followed the tutorials step by step.

finally a language for literal schizos, written by separate personalities.
manic: https://github.com/mlochbaum/Singeli/blob/master/doc/interpreter.md
>It prints fib{90} in hex, so quirky!
depressive: https://github.com/mlochbaum/Singeli/blob/master/doc/compiler.md
>in a fit of overengineering the authors have decided maybe you'd want something other than C operators and backend

>>103889974
A two way pager with tor layer for privacy. With address book, possibility to write messages and all that with a 4g lte capable esp32 to get lower costs.
Even uses an eink display

>>103930195
make it non-explosive and you've got a niche

>>103930200
It is indeed non explosive but I could add that feature any time if I ever decide to make this commercial. Which I probably won't because the material cost of a single one of these is 75 usd

>>103928330
It's a good language.

>>103928330
go to the website and scroll down.

why can't I do this if a function requires a double pointer?

int a = 1;
foo(&&a);

>>103932802
Because that assumes that there's a directory of addresses where your address is being stored at, which doesn't exist. Use a fucking pointer and provide its address to foo, or stop programming altogether.

>>103932802
&a does not have an address

>>103932802
>>103932842 is correct, &a is stored in a register, not memory
If you wanted to make your intent explicit, you should do this:

int a = 1;
int* a_ptr = &a;
foo(&a_ptr);

>>103932802
The others are right, but if you want to learn about some of the more precise details of how the language thinks about this, it's the difference between an lvalue and an rvalue, which are some terms you can look up.

>Addresses
What's the point?

>>103914750
dont listen to the other anons. it's helpful to make sure a heavy process at least allows other processes to work. I had to use that in my multithreaded c# minecraft project years back

>>103933218
Storage.

Anyone else smoke crack before coding

>>103933366
>Dude drugs lmao
That does not make you an interesting person.

>>103929178
you keep posting this and saying everything is shit
so what do you use? what do you even work on?

>>103936026
>so what do you use? what do you even work on?
Right now I'm wrapping my mind around 32-bit addressing in real mode. I know that my values are correct, despite people telling me that they shouldn't as I never bothered setting up unreal mode.

>>103933863
anon I am inclined to believe that person doesn't actually smoke crack before coding

>>103936118
Oh, the assembler I'm using is FASM. The fact that it cross-compiles for COM files means I don't have to use a build environment in my emulator, because that shit sucks.

>>103936118
damn so you just masturbate to working on useless bullshit. but at least it makes you feel cool

>>103936154
Yeah, and the best part is that it's more than you'll ever achieve in your life.

>>103897406
update

This thing is pretty cool. I'm considering going through a E&M and an electronics course while I look for a job

>>103936148
tsoding?

>>103936981
Nah. I don't wear glasses for one.

SDL2 is a gigachad library.

SFML and raylib are like: "hey, what codepoints from this giant ass ttf you want. I'm too weak to be useful".
SDL: "I got these details covered bro, just find me a font"

And not only output. It can handle input through IME. I'm in love.
I hope SDL3 doesn't break shit.

>>103937177
>gigachad with shitty memory management
Doesn't compute, you're full of shit.

Listen to my commentary on the tech industry:
https://taoistchina.bandcamp.com/track/programming-history-2005-2025

>>103937300
>Why did MIT drop Scheme for Python?
https://www.youtube.com/watch?v=OgRFOjVzvm0
>How much programming the economy needs?
All of it.
There is not enough.
Retards can't center a div or debug a null pointer.
>Does academia follow industry or viceversa?
Academia is a business. It follows industry.

>>103937617
>617
BWAHAHAHAHAHAHAHAHAHA

>>103937300
Talk about the Dao and programming I think it would be cool. Also
>uh

>>103936118
I love by the way how no one is competent enough to comment on the code. Just goes to show how pathetic /dpt/ truly is.

>>103938291
Taoism is basically Chinese Buddhism, and Buddhism is a reaction Zoroastrianism (up-down yes-no voting religion and mythological basis for democratic activity in a republic), which was a reaction to the founding of Rome in 750 BC and the problem of authority that is associated with any democratic republic. It is possible to criticize historical materialism along these lines: history is mostly authority, and when there are problems with authority (as there were when Rome annexed the Kingdom of Egypt in 30 BC) there is historical political upheaval that is remembered for millennia.
It isn't an exaggeration to characterize Taoism as "fuck this, I'm outta here" religion. If we take the legislative decisions of democratic republics as a matter of "social programming" then Taoism is about getting as far away as possible from the raging fire of political nerds arguing about the laws and rules for society. Taoism is about refraining from getting sucked into a political battle that might piss off the mob and cost you your life.
In fact, Zoroastrianism, Buddhism, and Taoism are all united in that they provide support for people influenced by democratic republics and the raging war of political conflict they produce.
Computer programming is an advanced form of political / policy / operational control developed in the most advanced democratic republic in history, the USA. With the advent of AI, it is trendy to think of cognition or thought as "human programming."
So, to sum up, Taoism says, "don't do it!" and provides you with a toolkit for criticizing and untangling all of the problems and knots programmers get themselves tied up in. Taoism is an excellent foundation for debugging and understanding the psychology of debugging.
Comment theme: https://www.youtube.com/watch?v=7bCdrDhUjPo "White Lines (Don't Don't Do It) Original Long Version - Grandmaster Flash & Melle Mel"

>>103938591
>Taoism
>Buddhism
I thought this was the Touhou thread for a second

>>103938591
So, for example, we can look at Julius Caesar, the annexation of the Kingdom of Egypt at the hands of the Roman Republic, and the conversion of the Roman Republic to the Roman Empire through the lens of the history of authority. In fact, we can do the same with Jesus Christ, Christianity, and the decision of the Roman Empire to make Christianity the official religion. From the point of view of programming, we can massively oversimplify and say that the fundamental problem of this era of ancient history, from 750BC to 500AD, the founding and fall of Rome, is that people didn't know who to listen to as a source of authority. You've heard of too many cooks in the kitchen, well this is too many politicians in the forum. The result is confusion, suffering, secret codes, secret societies, people lying systematically, political movements, the sort of stuff you might read about in the Illuminatus! trilogy.
What's so curious is that all of the meta-physical theories that developed to support people, heal them, give them hope and encourage them, feed them, make them strong, and so on are still very much relevant today, but some parts have been forgotten such as: Christianity wasn't originally intended as a universal world religion, and it took the prophet Mani to retrofit such qualities onto Christianity around the 3rd century AD.
What we might ask is, "how much meta-physics do the people need?" and the answer is, as this anon says
>>103937617
All of it
There is not enough
Retards can't into calm or the uncarved block.

Hey, /dpt/, how does it feel to have an actual schizo rambling about off-topic garbage? Or are you not ready for this discussion yet?

>>103938881
a registry dumper made with assembly is not that off-topic

>>103938881
Talk about codata

>>103889986
Not much changed. Most web applications are in SpringBoot and React. Many imperative languages are borrowing functional programming parts but never really go all the way. Rust continues to gain track. I guess the biggest difference is some devs use some form of AI.

In general the trend is to use safer and less verbose programming languages but most people don't even know what that means, so we end up with half assed attempts introduced in mainstream programming languages.

>>103938915
Is that registry dumper in the room with us now?

>>103937177
The first time I was experimenting with SDL2 there was a bug that made lines not pixel perfect. I got annoyed and never thouched again. Actually I just had other priorities but the non-pixel perfect part did happen.

>>103938881
You must be new here. Before it was registry dumping nonsense it was some kid spamming "python rocks because christ is king" or whatever for three months and before that it was a couple of people going back and forth about "clean code" for months. Maybe not in that order, I don't remember.

>>103938967
Yes, and it just dumped an uncarved block or Pu in moon runes.

>>103939041
Yeah, no, at least registry dumping and Python and clean code are on topic.

>>103939071
Can I see it?

>>103939085
Clearly, you shouldn't suggest that it's not part of the 4chan. Haven't you seen it in the /b/? It says, "His log expounds Dharma, the microbes that live inside his logs expound it, throughout the three flushes, everything expounds it." ' " The Master thus completed his narration.

>cannot show the registry dumper in assembly
Yeah, thought so.

>>103939072
You wouldn't know because you're underage but weekend /dpt/ has always had a tendency for drifting off topic

>>103939072
Dongshan is clearly discussing a codec in cases #3 and #4.
https://terebess.hu/zen/dongshan-eng.html
>"The National Teacher said, 'You yourself haven't installed the codec for it, but this can't hinder those who have installed it'

>>103889974
If I learnt assembly programming from current sources that were written with modern architecture in mind, how much of that knowledge could be used for reverse-engineering older games (NES, Sega Mega Drive, etc.)? How about C and C++ for 5th generation consoles like N64 and PSX?

>>103939124
I just find it funny that your parents never taught you to not project.

>>103939159
It will teach you that there is a CPU that interacts with both memory and hardware, and that's about it. The details are all hardware-dependent.

why isnt C more popular than C++?
isnt a simpler language be better for maintainability in the long term?

>>103940343
You will never find a reason to justify the popularity of c++. Its shit and everyone knows it's shit, any other language would be better. But everyone uses it anyway

>>103940364
so C++ caught on when OOP was all the rage and sticks around because of inertia?

>>103940343
C needs strong typing and templates

>>103940343
>why isnt C more popular than C++?
C++ allows sloppy OOP practices that simplify resource handling which does the job and spawned a quadrillion lines of middleware code. Thus, it stays.
>isnt a simpler language be better for maintainability in the long term?
It's better for maintainability to have a language without millions of lines of code to maintain. Nobody wants to rewrite them and painstakingly find out all edge cases, again.

We are all buried under mountains and mountains of slop.

>>103939159
Related, what's a good book for learning x86 assembly?

>>103899265
geh arbeiten?

>>103940387
>Nobody wants to rewrite them and painstakingly find out all edge cases, again.
Sounds bloody sutff for everybody then, because where else but on the software level can you get decent performance gains these days? I even hear normies complain all day about how long everything takes with modern software.

>>103889974
How do you guys approach super big APIs that you're seeing for the first time? Assume that the source is readable but there's virtually no documentation because while the docs do exist they're also like 25 months out of date.

day idk of learning programming and teaching myself c# through youtube and books... assignment from book required me to make a console tic tac toe game.
took me 2 days to finish and the code is hideous, but at least it works.

>>103941463
Death threats.

>>103941500
"it works" is always the top priority
once you have a working version, save that somewhere (either git, or just make a copy of the files) and try refactoring, ie. improving code without changing behavior so it still works

>>103937191
Please fap to GC somewhere. sdt::unique_ptr takes care of memory.

>>103941845
What if I want to reuse that memory for later? Preferably without the nonsense that is new/delete/malloc/free?

Yeah, thought so. And now do the world a favor and kill yourself.

>>103941900
enjoy you're shitty performance

>he thinks avoiding malloc/free yields shitty performance
/dpt/ is truly gutter trash these days. And here I thought programmers were supposed to be smart.

>>103942289
Noone asked
Noone cares
See you in 5 minutes where you post this again for the millionth time

>implying I need your permission to post anything
You sound so dumb that killing you off would improve the national average IQ by 30.

>>103942330
Your discord-troon style of not replying to the post you're referring to doesn't do your cause any favours.

>>103942289
post code

Oh, no. It doesn't cause me any "favours".

Oh well.

>post code
I already have. /dpt/ is just too stupid to see and/or understand it.

>>103942368
>Schizo Discord troon is also a burger
Like pottery.

>dumb Yuropoor is also a nocodeshitter
Likewise.

I love how much my primers trigger you. I must be hitting very sensitive nerves.

>>103942289
my nigga no one is reading these and no one cares about your nothingburger assembly hobbyproject snippet where 80% is comments, go make something that actually has a usecase instead of being a fucking faggot
>le pathetic... i am ze azzemblykike and you niggers are NOT on muh level
fucking kek and holy cringe, now gtfo

>no one cares
Considering the delicious seethe my uploads trigger I'd say you've a big, fat liar. And there is already software out there that covers most use cases big, fat liars like you would be interested in: https://x.com

>>103940343
> isnt a simpler language be better for maintainability in the long term?

Not necessarily. Python is just simpler than Java but it's horrible for maintenance because it's not strongly typed. Good luck refactoring anything in Python.

In C vs C++ is issue is C is too simple so you end up requiring a lot more code and that ends up needing the reader to pay a lot more attention which is not good for maintenance. Neither is particularly good for maintenance. Zig and Rust are better than both. They're just as powerful, but easier to read, build, write tests and safer.

>>103942622
>In C vs C++ is issue is C is too simple so you end up requiring a lot more code
Not really. Just use a library you fucking retard. When you are doing the business logic, struct + functions end up being less code compared classes defined in fucking header files you dumb retarded nigger monkey.

>>103942622
>In C vs C++ is issue is C is too simple
I'll still take C over C++ because at the very least the C compiler knows what it's doing most of the time. In C++ it's absolute guesswork.

>>103942674
Like you can't use libraries in C++. Your still inevitably end up with more code in C regardless of your opinion with classes. It's not like I personally like C++, it's horrible, but that's just how it is.

>>103942719
That's a different problem and if you need assembly to be predictable yes C is obviously better.

Why do people compare generated assembly WITHOUT optimizations?

Like this image >>103890127

>>103942808
>g++
>clang++
Spotted the problem, you should be using GHC

>>103942841
Right, because generating one GB of garbage per second is so much better.

What do you prefer for scoped enums and why?
>enum class Foo { A, B, C }
OR
>enum class foo { a, b, c }

>>103942808
In fact picrel is the nonsense clang++ generates with -O0. Your idea that I was showcasing unoptimized code is just an ego defense mechanism to avoid the inconvenient truth that C++ is garbage and Stroustrup's existence has been a detriment towards human development.

>>103942908
I was going off of the defendant's own admission: https://wiki.haskell.org/GHC/Memory_Management
>it's not uncommon to produce 1gb of data per second

>>103942931
static enum Type { A, B, C }, where A is zero and C is count

>>103942931
struct foo { enum type {..., N}; };

>>103942954
That was 10 years ago, computers are faster now. Haskell isn't limited to 1GB/s

>>103943005
Surely you have a source for your claim?
>inb4 I made it the fuck up

>>103942973
>>103942995
t. never programmed professionally

>>103943048
t. only ever worked on trash codebases

>>103943048
when they introduced enum struct they fucked it up by doing two orthogonal things, so there's no nice scoped numeric enum

>>103943015
Even if he doesn't have a source GHC is like what? Version 9 something? It's always being updated. It's likely the GC was also updated. Even stack that is supposed to be the stable LSP thing is constantly being updated.

>>103942973
>static
deprecated, use ANS

>>103942995
>struct foo { enum type {..., N}; };
deprecated, use enum class

>>103943109
So why don't you show us the numbers then, rather than making assumptions that are potentially unfounded?

>>103943048
True, do guys in the private sector only use lowercase?

>>103943155
I don't have the tools for that right now.

>>103942289
I might have tried to read at least some of what you write but if you expect me to horizontal scroll back and forth a million times my friend you are delusional. Stop posting images full of text.

>>103943438
You can hold shift while using the mouse wheel to scroll horizontally, or just get better glasses.

>>103942931
Second. Always snake case except for template arguments.

>>103943729
What about free standing functions and class names?

>>103943769
Snake case. Exactly like the standard lib.

>>103943838
but c++hudda, what about...

>>103943144
>deprecated, use enum class
Doesn't solve the same problem retard

What language do you guys enjoy programming in the most and why?
>inb4 data mining

>>103944220
Assembly, because it makes me aware just how much compilers suck.

>>103944220
C++. Because it's like C but with strong typing and templates.

>>103944220
Haskell

>>103944332
>C++ is like C
Yeah, except the RAII and the exceptions and the code generation and ...

>>103944355
nobody cares schizo

>>103944220
Probably a lisp, the interactivity makes it fun, and html/json are basically sexps already. I go back and forth on whether CL or Scheme is more fun

>>103944299
what a fucking idiot

>>103941242
>because where else but on the software level can you get decent performance gains these days?
You don't. You don't want to reimplement for Linux, Windows, Mac, AND mobile (tablets, phones, android and iphones). You don't have people or time for 6 native platforms.
>I even hear normies complain all day about how long everything takes with modern software.
Another issue here is that if something takes 10 seconds, reducing it by 50% to 5 seconds to complete, it's still a long ass noticeable time and for people (rightfully), it's slow: 10 or 5 seconds don't make a difference in perception. You need to go below an order of magnitude, from 10 to 1. It's not happening.

>>103942622
Python is strongly typed, as doing

x = "num is " + 10

is a type error.
Javascript turns the number into a string, no questions asked, with implicit casting.
Python is strong, dynamic. Javascript is weak, dynamic. Java is strong, static. C is weak, static.

>>103944220
Iversonian langs. Code feels like code - cryptic, information-dense, worth studying, satisfying when understood. You can see more of your program at a time and it's worth seeing. And array-oriented programming is data-oriented programming: https://news.ycombinator.com/item?id=41756917
>Hard to write, hard to extend, hard to refactor. I generally recommend against writing this way in [1]. But... it's noticeably easy to debug. There's no control flow, I mean, with very few exceptions every line is just run once, in order. So when the output is wrong I skim the comments and/or work backwards through the code to find which variable was computed wrong, print stuff (possibly comparing to similar input without the bug) to see how it differs from expectations, and at that point can easily see how it got that way.
the way AI-assisted programming is going, you'll have AI generating the slop, and also AI summarizing the slop, and in the most advanced summarizing visualizations you might be rediscovering APL anyway.

>>103944220
I don't really care much about which programming language I use but in general if I want it to be very clean, very easy to maintain, very readable, easy to add concurrency, I go for Haskell. That's usually my default.

If really don't care I just use Bash. If I need more power, Python or Lua. If I need to be half assed and still easy to maintain, Java. If I need more power, Scala. If I need to mess around with low specs hardware, C. If it's a bit more powerful, Rust. Alternatives like Kotlin, Go, C#, Zig, etc. aren't very different from my go to languages.

It depends on what I'm doing. I don't usually do web front end in personal projects, but usually I use Typescript or
Purescript, depending on how much I care about the project. Typescript is when I don't care, I'm just doing an experiment.

>>103944512
>strong
>weak
this is decades-old dumb as shit propaganda, and the only point of it was to distinguish Python from Perl a little more. Well, Perl's dead. You can stop pretending that 'strength' is a property of type systems and not the standard library, or that Python's standard library is particularly strong:

>>> True + True
2
>>> True * 5
5
>>> "num is " * 3
'num is num is num is '

>you can't get an order of magnitude faster
Actually benchmark idiomatic Python vs. D or Nim or some language that not the slightest bit less convenient of a language, but not made out of piles of garbage, and you'll see several orders of magnitude speedups. Not 10x, but hundreds of times to thousands of times faster.

>>103944512
Fair, but being dynamic is a problem for maintenance.

>>103944607
>Well, Perl's dead.
Which is absolutely ridiculous by the way: https://www.fiverr.com/resources/guides/programming-tech/python-vs-perl#perl-is-the-winner-for-text-processing

>>103944607
There are degrees of strength. Javascript overloads everything to the point where nothing is an error.

var x = null;
var y = x + 10;

Strength is not "how powerful the type system is" but "how hard it slaps your ass if you do a type mismatch".
>Actually benchmark idiomatic Python vs. D or Nim or some language that not the slightest bit less convenient of a language
You can't use D or Nim because they don't have frameworks for brainlets. So in practice it is impossible: people have to refactor Python and JS because there is simply too much of them. The sunk cost is too big.

>>103944641
Perl is horribly designed. Very difficult to read. I tried my best to make it readable and after a while I didn't have a clue what the fuck my code was supposed to do. Granted I was noobish at the time but the syntax doesn't help.

>>103941900
Nice dubs. Unfortunately last time I've used new was around the time of

auto_ptr

was alive(back then you were not even born). By mentioning "new" you've shown your programming incompetence to anyone who had a doubt.

>>103942931
First. Foo::BAR to be precise, not Foo::Bar

>>103940343
C++ has better standard library with vectors, hash tables and beautiful RAII.
Which is a shock as C++ has a very shitty standard library, so having even shittier library than that deserves an achievement. Don't mention C's _Generic it's so bad, people look away instead of using it.

The only thing C is good at is compiler speed: it creates illusion that you work faster. You don't. What you can do in C++ in 30 seconds and compile in 1 minute would require 5 minutes to implement in C, 2 seconds to compile, 4 hours to debug after you run valgrind and see a memory leak.

If C committee was thrown away and replaced with GCC people only we'd have a nicer language with nested functions and ({expressions}) of several statements so MIN and MAX would be possible to define as a macro where args would be evaluated only once. As of right now C will never be good.

>>103944770
>Very difficult to read.
Sounds like a skill issue then. I've been writing Perl scripts for years and never had anyone complain about my code being unreadable.

>>103944778
>he doesn't even know what his constructors and destructors are calling internally
What I find hilarious is that you've just revealed your own earth-shattering incompetence for everyone to see. If we lived in a just society you'd be carried off to a death camp by now.

>>103944607
>True + True
>2
>True * 5
>5
/sci/ is gonna freak

Can any of you terminal fags explain why I would run a script like this
>. ./util_functions.sh
Instead of just
>./util_functions.sh

For the record the top line worked but I have no idea why the bottom line doesnt. They both run the script from the same directory, wtf

>>103944863
. runs from the current shell, in a subshell otherwise. Are you testing for interactivity at the top of the script or something?

>>103944863
https://superuser.com/questions/176783/what-is-the-difference-between-executing-a-bash-script-vs-sourcing-it

>>103944907
>>103944897
Thanks bruhs

>>103944858
What's your font and color scheme?

>>103944641
Perl is very good at its niche, but nobody really cares. People prefer general purpose languages. It's why we had faggots trying to shoehorn Javascript into everything for a while.

>>103944982
consoulless

New bread: >>103945044

>>103945020
>People prefer general purpose languages.
Yeah, well, a lot of people are borderline retarded.