[a / b / c / d / e / f / g / gif / h / hr / k / m / o / p / r / s / t / u / v / vg / vm / vmg / vr / vrpg / vst / w / wg] [i / ic] [r9k / s4s / vip] [cm / hm / lgbt / y] [3 / aco / adv / an / bant / biz / cgl / ck / co / diy / fa / fit / gd / hc / his / int / jp / lit / mlp / mu / n / news / out / po / pol / pw / qst / sci / soc / sp / tg / toy / trv / tv / vp / vt / wsg / wsr / x / xs] [Settings] [Search] [Mobile] [Home]
Board
Settings Mobile Home
/g/ - Technology

Name
Options
Comment
Verification
4chan Pass users can bypass this verification. [Learn More] [Login]
File
  • Please read the Rules and FAQ before posting.
  • You may highlight syntax and preserve whitespace by using [code] tags.

08/21/20New boards added: /vrpg/, /vmg/, /vst/ and /vm/
05/04/17New trial board added: /bant/ - International/Random
10/04/16New board for 4chan Pass users: /vip/ - Very Important Posts
[Hide] [Show All]


[Advertise on 4chan]


File: dpt.png (34 KB, 180x193)
34 KB
34 KB PNG
what are you working on?

previous: >>108042427
>>
I don't like how lispers trail their parenthesis
>>
>today in /dpt/: autists getting triggered over syntax
You cannot convince me these animals do not deserve to live in cages for the rest of their lives.
>>
>>108075328
actually its a critique of style. i think you might be the autist anon san
>>
>>108074955
here you go: https://pastebin.com/vd2M29vg


save as .hta file. to not get the messed up "good morning paul" you just have to remove the esc + [s + [u parts from the CincoDOM
I dont know why C:\Windows\SysWOW64\cmd.exe "C:\Windows\System32\cmd.exe" /c start "CincoTerminal" /max "C:\Windows\sysnative\conhost.exe" cmd line or DOS-style /V:ON /k "TITLE Pauls COMPUTER werks but it just does (not overwrite anything far as I am aware)
>>
>>108075333
>splitting hairs between syntax and semantics
Not beating the allegations.
>>
>>108075328
Psychopaths like you are the real ones who deserve to live in cages -- specifically jail cells, in the case you have committed a crime (which is likely).
>>
>>108075171
/* nice touch */
>>
File: 1770391050727.png (1.82 MB, 1344x988)
1.82 MB
1.82 MB PNG
https://www.nntp.perl.org/
Why does everything around Perl reek of incompetence?
>>
>>108075467
>reek
whoever smelt it, dealt it
>>
>>108075458
// no
>>
>>108071238
Much thanks
>>
>>108075427
Oh, trust me, I have empathy. Even for animals. But not for humanoid animals.
>>
>>108075515
>"reeee anyone with autism (except me ofc, because I clearly don't have it) is an animal and not human and deserves to be put in cages"
Jesus christ, I'm right, you are a psychopath
This is why people like you should be shunned from society.
>>
>people like the regdump schizo essentially run the DNC and GOP and thus the US government either way
God wept
>>
>these people code in a way that I do not like, and that means that they are subhumans!
OK we get it, you code... and not script...
I like LISP, though. a functional programming language (one where it alters its commands and then sends them) is pretty epic
>>
>>108075527
>because I clearly don't have it
Yes. I was a very friendly and outgoing child.

>is an animal and not human
Yes. Although to be honest I wouldn't mind banishing your lot to Mars as well.
>>
>>108075556
Reddit spacing
>>
>>108074596
>I make him blablabla
YOU should review the code. YOU. Not your spaghettied bundle of agents.
>>
>>108075546
>these people code in a way that I do not like
It's not about whether or not I like it. It's about autists focusing on whatever tickles their animalistic brains instead of focusing their energies on actual problems around them.

>>108075566
>>108075344
>>
>>108075587
OK redditor
>>
>>108075544
>differentiates between DNC and GOP, not RNC though
Why?
>>
>>108075648
Isn't GOP the same as the RNC? Or am I retarded
>>
>>108075663
GOP => Republican Party, a.k.a. "Grand Old Party", in contrast to the Democratic party.
RNC => Republican National Committee, in contrast to the Democratic National Committee.
>>
>>108075692
Ohhhhh I am retarded
I meant the parties not the committees lol
>>
>>108075544
speaking of reg dumps, I need to reg dump and save all the hives for when Microsoft depreceates vbscript, so that I can reupload all of their stuff to undo that
>>
>>108075697
Just export/import via regedit.
>>
https://www.youtube.com/watch?v=-J4YqtUKOQI
this is literally me
>>
>>108075467
>https://www.nntp.perl.org/
>nntp
software rots, and a lot of software was built for Perl, and nobody cares about it anymore so nobody deals with the rot. Anything Usenet has one guy who thinks about it once a year, and no users to tell him about problems with it.
>>
>>108075338
>here you go: https://pastebin.com/vd2M29vg

if its not too much to ask, can /dpt/ help me with this ONE part of the script (if I try to do this I will spend 3 hours testing things and get distracted, and I dont think I even can do it)
can you make it so the "echo" in the upper part says "Pauls COMPUTER"
>>
File: yeah.png (385 KB, 960x789)
385 KB
385 KB PNG
>>108075741
>Second recommendation: Why do gay people sound like that?
Yeah.
>>
>>108075741
https://www.youtube.com/watch?v=a8K6QUPmv8Q
>>
>>108075714
>>108075765
I know ive just got to export the regedit (probably on a new install so that my personal things dont get mixed up in it). its just effort that I can do later when they actually do break things
>>
>>108075765
ANTS
>>
>>108075760
>using a language that constantly calls him dim
look at this thing: https://github.com/pqwy/notty/
>>
>>108075833
>https://github.com/pqwy/notty/
honestly if I go on to learn other languages (with the help of AI) then I will choose a good one and be able to do that because of AI. I just have a slight chip on my shoulder from them trying to teach me this in high school
I can go be a skiddie for a bit and then move on tho
>>
because now you dont have to (especially big while learning) go debug and t&e every single problem every time, sometimes taking like an hour. this is a HUGE QoL
AI does schizo out and starts lying to you at a certain point though
>>
>>108075868
it's not a HUGE QoL to delay learning any kind of troubleshooting until after your programs are much larger and you're doing harder things.
>>
>>108074206
>>bitshifting with padding with 1s
>what does that mean?
I mean, what I saw was your bitshifting to the right was pushing in 1s instead of 0s from the left.
>>
>>108075328
That's every day on /dpt/.
>>
>>108074206
>which operators do you need to compare the precedence?
It's different every time. This last time I was unsure which had higher precedence; bit shifting or bitwise or.
>>
>>108075893
I have picked some things up. but when it doesnt work suddenly, I can now just paste it to an AI :)
btw 100% of my problems right now are "syntax highlighting or "expected end of line" when I mis a character
but seriously, like when I was a kid, if I found the registry for the first time I wouldve spent hours looking through it for gems, Googling each thing, and thinking about each one for a sec. and I have done that while having no internet. but now, I can ask AI to go through the registry and find cool things and Google it from there

like for example, did you know you can set the opacity for cmd or other windows? thats crazy. thats like linux :)
>>
I'm getting a program written in COBOL running on Windows XP to run on Windows 11 and print to a modern printer
>Customer has entire business in this software and PC is old enough to drink
>PC ded
>click click click goes the disk
>ddrescue, reverse disk scan, only lose a couple bytes
>get valid XP .img
>convert to .vhd
>Run in vmbox on W11
>use the narrator to hyperlink to cmd for temporary activation
>grab xp_activator from internet archive
>realize that when printing, program saves a propriety file to the program directory that's just a text file with a different extension
>create a light bash program which compares the directory content against a reference list
>if a file isn't in the reference list, it's what we're trying to print
>copy that file to a folder shared with the Win11 host
>delete the file from the program directory
>on the Win11 host, 2nd small batch script
>copy the file and rename extension to .txt and print
The only remaining problem is encoding. Sometimes, but not all the time, it goes all japenese on me when the file moves over to Win11
Opens just fine in notepad on XP
>19$/hr
Kill me
>>
>>108075955
>This last time I was unsure which had higher precedence; bit shifting or bitwise or.
Yep, that's one of the important things you should know.

>>108075904
>I mean, what I saw was your bitshifting to the right was pushing in 1s instead of 0s from the left.
Yes, that's what the picture shows. I was trying to figure out how it works, and that's the trick. It's easy to see that once every bit on the right than the highest bit is turned to one, you get the next power of 2 when you add 1 after that.
>>
>>108076224
>19$/hr
For that amount of money I'll do the ddrescue in the background, throw in the XP download as a freebee, and otherwise masturbate to mindbreak porn.
>>
>>108076312
Forgot to mention that the COBOL program in question only prints to LPT1
know I could redirect to USB, but then I'd need to figure out specific drivers for a specific printer that still works in XP
this way, customer can migrate their data out XP to a safe W11 env while still having 1:1 use and access to their old PC while running on modern hardware
>Customers are too retarded to learn new things, which is why they're using business software where the GUI is a fancy terminal window
why would I be lazy when I can solve the puzzle optimally?
>>
>>108075587
So like what you're doing by having a meltdown right now?
>>
File: 1770399052576.png (423 KB, 1344x763)
423 KB
423 KB PNG
>>108075751
>Usenet
It's a stupid web interface for a mailing list, nothing fancy. And it was working a week ago.
Where am I supposed to get my daily fix of Perl lore now...
>>
>>108076389
>projects his autism unto others like no one's business
Cage. Now.
>>
>>108076508
Oh the irony of you saying this
>>
>>108076697
Still pretending autists deserve basic human rights, eh? But you're not human. That's the real irony here.
>>
>>108076738
Schizo
>>
>>108076738
Low test post
>>
>>108076738
OK redditor
>>
Found this benchmark for how much boilerplate languages use.
https://boyter.org/posts/boilerplate-tax-ranking-popular-languages-by-density/
Why Elixir has so much boilerplate? Is it because of BEAM process managing shenanigans?
And why Go has more boilerplate than C? Go can be obtuse, but isn't it way more high level than C?
>>
>>108076979
go forces you to check the returned errors, c doesn't
>>
>>108077118
forcing you to do anything is kinda antithetical to c
what you mentioned is not an advantage, its a difference at best
>>
>>108077130
I don't disagree, I just meant it as one thing that will add 3 extra lines of code over similar C code. it's ofc not the only thing that makes it more verbose
>>
>>108077150
afaik you lack macros in go
thats another thing that makes things more verbose

do you have force inlines in go?
that would allow to cut down on the redundant code
and you do need macros
while ill be the first to say to use force inlines instead of macros for legibility reasons,
ill also be the first to say to use macros to hide disgusting syntactical shit like picrel because nobody should be confronted with such an interface, if it isnt as punishment
#define STRINGIFY(target) (t_str *)(&(const struct { size_t size; char text[sizeof(target) + 7]; }){ sizeof(target) - 1, target "\0\0\0\0\0\0\0"})


if you saw that line before its because its the ugliest thing i have in my codebase atm
>>
>>108077118
go does not at all require you to check errors. Every single err assignment in your program can discard it with _ instead. Go has a strong convention of checking errors, and doesn't have any syntax sugar for it.
When I write C it's invariably some very paranoid scripting of syscalls, and it therefore has Go-tier verbose checking.
>>
>>108076979
retarded study, it's not comparing anything meaningful.
no matter which language you use, you can write a program in many different ways and make it short or long, depending on how much the language is expressive but also depending on your programming style, the amount of validation the program has, depending on the abstraction you're using, depending on how much features you implement, etc..

I'd rather read blogs, reedit and hackernews comments doing subjective comparison than reading this piece of trash.
>>
>>108077183
Macros are opaque and disgusting. I'd rather read the whole papyrus than a fucking macro and look up its definition.
>>
>>108077336
i agree
but sometimes theyre the lesser evil
still, if i see a macrotard whose code is half macros i throw hands
were on the same wavelength, but legit
sometimes macros are the lesser evil
and this is where they *should be used
>>
>>108077370
Well... Probably. I've only seen that kind of shit written by newbs or... In embedded systems. I fucking hate embedded programmers, I have yet to see a decently written firmware. And I have to deal with it daily.
>>
>>108077420
100% fair enough
but were not in a corpo setting here
that others are retarded it doesnt mean we have to be retarded too

i think:
the problem with that, and c in general,
is that people arent trained enough
and yes, i know theres constraints imposed on the industry, and on the educational curriculi
in the way of shoving underprepared, undertrained personel on the job market because cost reduction, and who the fuck would bother with programming if the studies were as long as for medicine

but this is /g/
here were veterans, passionates, /v/ermin, their /a/ssociates
and when you filter out the tourists noise
youre left with sheer skill
and this is the baseline for /g/
skill. not common practices by big corpo who are basically living in an iron lung because they need external structures to deal with the most basic functions of a company- namely agile, and organization

here we leave all of that behind, and we talk sheer skill, and best possible practices, without taking into account normie fucking retard lazy shits who went into programming for the paycheck
or at least thats my approach
no compromise engineering, because in here we can afford to.
because in here it costs nothing, kek
>>
>>108072033
here's my code:
https://pastebin.com/nYwPz1Ek

feel free to shit on it!
>>
>>108075171
>what are you working on?
I just took over the night shift, because no one else wants to do it. So it's me again.
Someone really has to do the night shift and it's me. That's my job. Everyone has a job.

Now I am still working on getting the DMA working properly to transfer LOTS of data between FPGA and risc-v CPU. Maybe one day I'll make it so that I can see the pictures getting captured by my CSI-2 camera.
Then I can add cool boi image transformation things (which was the initial idea of this project. Just do some image processing. Now here I am 3 weeks in still haven't gotten the images so that I can see them (apparently they are floating through silicon)
>>
>>108077336
>Macros are opaque and disgusting.
The worst codebase I have ever seen in that respect:
https://github.com/Perl/perl5
$ grep -E '#\s*define\b' *.h | wc -l
14920
>>
>>108077944
https://github.com/MarlinFirmware/Marlin
$ grep -E '#\s*define\b' *-r . | wc -l
63528


To be fair, perl gives me 26914 if we include all the files.
>>
>>108077532
seems ok.
>>
File: 1769743712974559.jpg (857 KB, 1100x1500)
857 KB
857 KB JPG
I just learned binary addition and had a mental breakdown
but ONLY +1 (if you can +1 you can +4 16 32 64 and 128). you can also subtract

I learned it to subtract, and that was (in the CPU), to take a negative number. to subtract, the CPU takes the number, makes it negative (flips the bits), and adds 1 while it is negative, then it sends the negative number (the negative version of the number to subtract) that you want tto subtract from to add them together.
8 - 8 = 8 + -8
for example


but any flipped number without +1 = -1
1000 0000 flipped (0111 1111) = (1000 0000 + 0111 1111) = (1111 1111) - 1
flipping the number to a negative, + 1 brings it to 0 in the brackets. that way, it adds 1:
(1000 0000 = 8) + 1 as negative (* -1), or
(1000 0000 = 8) + 1 (* -1), a negative = (8 * -1 = 0111 1111) + 1. the second half = 1000 0000 (8) and it adds that negative as -8; 8 + -8 = 0, because in either side of the logic unwrapped, or brackets, it is adding +1 while the number is in the negatives (the number 0111 1111 rolling over to 1000 0000). on either side of the brackets, without * -1 (or flipping the bits), either making it negative, then, it is doing (8 - 8) + -1 = (-1) + 8 - 8
and also/as a side note, therefore: subtraction in a (wrapped?) loop with any number the formulae is: (n - n + 1) - 1 = (-1 + n) + n +1 ; if it cant subtract then it is instead: n + (n * -1) 1 * -1 -> n + (1 + -n) 1 which it can do because it is donen in binary (and 1 + -n being n -1, I guess is obvious). exercise follows ...
>>
File: computerpepe.png (212 KB, 710x842)
212 KB
212 KB PNG
>>108075171
im having chatgpt help me with my hw

print("i'm a hacker now lmao")
>>
>>108077944
most of those aren't macros in any important sense
... that it's annoying just to filter out the named constants or the variables is a point against C, though
use grep -Er --include '*.h'
>>
>>108078151
$ ls ~/frogs | wc -l

69420
>>
>>108078356
nice Lua! But in Lua you can drop the parens:
> print "I'm a hacker now lmao" 
I'm a hacker now lmao

Lua's rare for having a paren-avoiding syntax rule like this that isn't retarded or confusing.
I sure hope that isn't some other language, like P*thon or J*lia
>>
>>108078383
What's wrong with Julia?
>>
>>108078346
>but any flipped number without +1 = -1
you mean = the number minus one, right?
in other words ~x = -x - 1
another thing you can learn about two's complement (signed integer representation) is the most significant bit is -2^(k-1)
>>
also also, does this mean that in, with it unwrapped, instead of (let) n + -n = 0 = n - n ; >>108078346
n + (i + -n) i can use any integer for variable i? if so, then could you use this for multiplying easily by numbers such as 4, 8, 16, 32, 64 ... i.e. by that number being in n column (0 0 0 0 = 1 2 3 4 column). it is left to the reader though I think my math in this exercise for you is correct for you to be using the right calculations
>>
File: file.png (45 KB, 806x399)
45 KB
45 KB PNG
Thoughts on this syntax for a tab container GUI? I can't figure out one that I like
>>
>>108078423
what do you mean
for fast multiplying by powers of 2 you just bit shift, either arithmetic or logical shift (arithmetic preserves the sign for signed integers), its the same as moving all the digits left or right to multiply/divide by 10 in decimal
>>
>>108078409
the flipped >number without +1 = -1 (the number not added when flipped will equal minus one), will equal the number minus one because the bits are flipped, and if it isnt +1 when they are in that state (when ... = -x -1), adding the numbers together when the numbers are carried will end up as x -1
>>108078445
maybe * 10 in decimal, but I learned (vbs cant do bitwise manip) slightly that to shift it to the left is * 3
the amount you shift it is n; n to the power of 3, x *n3 = SLFT n times
>>
>>108078466
its not *3, its *2
>>
File: horsefrog.png (1.26 MB, 1008x1038)
1.26 MB
1.26 MB PNG
>>108078383
lua? what the fuck is a lua? i'm writing with pythons
>>
>>108078392
"hello world" using 59x the memory and 504x the runtime of lua's
the worst choice for "hello world". Even emacs -x is a superior "hello world" language
>>
>>108078445
>>108078472
oh its n tot he power of 2
and the thing was for CBytes and n =2 : n3 = 8; I was shifting by bytes
>>
>>108078485
yeah the only thing to remember here is that there are two kinds of shifts, one works this way for signed numbers (arithmetic shift) and the other works this way for unsigned numbers (logical shift)
e.g. remember the MSB is -2^(k-1) for a two's complement number, but if you just did a logical shift (move all bits to the right) you get the second most significant bit set which is 2^(k-2) (positive), while arithmetic shift will copy the MSB (thus you get the two most significant bits set, -2^(k-1) + 2^(k-2) = -2^(k2))
>>
>>108078515
whats k? I bet Google wouldnt give me an answer
I know you are just trying to flex on me tho. also I learned about twos help, thanks for bringing it up (in case I didnt)
>>
>>108078537
number of bits
>>
>>108078543
most useful for me (and my future posts)!!
im not going to bore the thread with what I am doing, until its done though
>>
>>108078572
you may as well nothing else happens here
>>
>>108078474
>i'm writing with pythons
I use that at work. Loathsome language.
>>
>>108078582
so basically I have an "encrypted" string of binary flipped to negative converted to hex (every 1st bit in the binary 0000 0000 number is flipped to 1, as the encryption). this is padded with String(random, "0") or a random amount of 0s, between each hex (which as hex is always 2 chars) is a long way of saying it
to decrypt this hex it does a cmp, which to the CPU takes it as binary. I am going to cmp it with hex 0x80 (%H80) so that all the bits except the first = themselves, and the first = 0
before decryption, the padded 0 hex will be the only readable (as string) text. this will become text when in the "0" padding I put a poem in ASCII converted to binary in
the poem will be trimmed but I know from learning binary that I can do AND on the strings with hex "88" + 1 * 1 to AND it with 0000 0000 ...
though I will just trim the "0" padding by skipping random(string(random)) in my script
one thing that I am for if open to anything is letting my random number generator (which can go to negative numbers, converted back with Abs(myInt) or Int(myInt), which may also do the same (I forget), be always a negative number so that in hex it is correct in a binary way. meaning that it will always be a binary string starting with "1" (this helps with the cmp)...
>>
in vbscript a cmp is done by using
 
select Case cmp1 ' value compared to; cant be math
case cmp2
' this is done by the CPU is a cmp between two binary strings
' it iterates the entire list of case s until it finds a match and then stops. the cmp list is stored in memory, im pretty sure in an "array" of values that it goes and grabs from to do the cmp instruction
>>
File: go-sum-types.jpg (108 KB, 1090x873)
108 KB
108 KB JPG
>>108075171
>what are you working on?
nothing
but kimi k is working on adding sum types to go
>>
>>108078754
so are you saying something like
1011 1101 -> 1101 1110 <0s...> 11<end>
[assuming you offset it otherwise you don't know if each bit started with 1 or 0]
>>
>>108078999
in Java this is just
interface Visit<A,B,R> {
R l(A a);
R r(B b);
};

interface Either<A, B> {
<R> R visit(Visit<A,B,R> visitor);
};
>>
File: PXL_20260206_224707751.jpg (1.78 MB, 2268x4032)
1.78 MB
1.78 MB JPG
Do you guys even have your hackerbrause up and ready?!
No you don't because your an faggot

>She doesn't use a sharpie like trump, for notes
disgusting
>>
>>108079027
OK, Hanz.
>>
File: maxresdefault.jpg (61 KB, 1280x720)
61 KB
61 KB JPG
>>108079081
Hans*

https://www.youtube.com/watch?v=8mGBaXPlri8
>>
>>108079249
OK, Hanz.
>>
>tfw electricity is too expensive to hack
bros I hate I hate our globohomo govement so much
>>
File: 20260207_004042.jpg (334 KB, 3024x4032)
334 KB
334 KB JPG
>>108079027
Trying this one. It was ok, but I'm more Mate Mate/Mio Mate ginger guy.
>>
>>108079309
>>108079027
good morning sirs
>>
File: nachtschicht.png (519 KB, 1280x720)
519 KB
519 KB PNG
>>108079309
Mio mio is good, but 10 cents more than Club Mate. So Easy choice, if Mio is not on discount.

Night shift going strong. We makin' big progress boys!!!! I love my job. Someone has to do it. So I do the night shift. That's my job.

https://www.youtube.com/watch?v=GfAOHlQA8dg
>>
File: test.png (1.35 MB, 1280x960)
1.35 MB
1.35 MB PNG
>>108079380
Okay we are fucking close!!!
Some lines are still zero and I have no idea why it is repeating. I was just one person with a camera light.
But at least we have F A S T DMA now to get 1280x960 resolution right into my DDR4 RAM
>>
>>108076979
I can confirm the blog post is true because it has Clojure and Haskell at the top where they belong according to my bias.
>>
>>108076979
>Lua: 39.2, lowest score
Does this mean Lua has a lot of boilerplate or is it the opposite?
>>
>>108079903
It means Lua has a lot of boilerplate and it is repetitive. But that's by design, Lua is bare bones and people don't fuck with metatables.
See the next table.
>>
>>108076979
>HTML
>>
File: 1756333654813412.png (86 KB, 818x346)
86 KB
86 KB PNG
>build the most basic app ever
>actually the app isn't even built yet
>it's just the skeleton of the UI
>already having performance problems
>find this on reddit
lmao
>>
got raped by today's leetcode question award
>>
>>108079619
>night shift almost over
>I am still not done yet
Well, I guess that will be some free overtime
>>
>>108080180
I love koding so much. We need more shiny and more animations. This will make everything better. The slower the animation the better
>>
File: test.png (1.42 MB, 1280x960)
1.42 MB
1.42 MB PNG
>>108080889
reeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee NO FUCKING PROGRESS.
Fuck, i hate myself so much. Lets watch normie movies and then bed maybe or more kode.
But at least I have some visible camera picture, that is good.
We have to try some different approach. Maybe pixel packing tomorrow to cut pixel speed by half
>>
>>108081074
>tism state
>>
>>108081082
Ok today i picked the night shift. You gotta do some work on night shift bro. That's what real hackers do
>>
>>108081089
I have no desire to be a true Scotsman.
>>
File: Camelia.svg.png (97 KB, 1280x936)
97 KB
97 KB PNG
>>108075467
Um excuse me, but Perl is deprecated.
>>
>>108081389
Raku lost his way long ago
my take is that the worst that happen to it in terms of language (setting the VM aside) is all the metaobject protocol bullshit
>>
File: go-sum.jpg (78 KB, 643x990)
78 KB
78 KB JPG
>>108078999
2.8k loc, and a few million tokens later, chinkAI managed to build shitty sum types into go
>>
>>108075171
I've given up on C++ after failing to make a class that holds a std::string and a few std::string_views into it without memory corruption when moving or copying it. I will return to C with global variables like our UNIX forefathers intended.
>>
>>108081622
Our UNIX forefathers worked on radically different machines. Doesn't justify C++, but it also doesn't justify carrying around all the garbage they came up with.
>>
File: 1770452353430.png (193 KB, 1344x727)
193 KB
193 KB PNG
>>108081389
What are you, a 7 year old girl?
>>
File: 1606631707251.jpg (40 KB, 941x765)
40 KB
40 KB JPG
>>108081729
I know better fun stuff to appeal to 7-year-old girls.
>>
>>108081074
Ok back in day shift now. Lets see if we can get this fixed. No AI usage in day shift. Ever.
>>
>>108081729
just another senile old fart

and the idea that you need to start so early is retarded. only non programmers would think that
If education made you learn programming that young, it would just be a gigantic waste, just like foreign language courses. You would spend literally 10 years of your childhood learning fuckall and at best it would culminated into be able to write a hashtable in C after 10 years in high school senior year, basically the level you'd got to after studying for a single year in uni.
>>
>>108075171
what are the style of buttons that are basically capsules split in 2 halves 1st half is option title whereas 2nd half is split into the available options?
for some reason geminicli has real problem implementing those
>>
>>108081881
a menu with a static label?
>>
>>108079008
no,
0111 0010 0001 -> 1111 10100 1001
and then back
>>
File: 1770059176874522.jpg (812 KB, 2480x1116)
812 KB
812 KB JPG
you can share your screen (your text editor) with Google AI at: https://aistudio.google.com, apparently
you get 1000 requests per day on 2.5 lite
the screen-sharing sends 1 request per second (1FPS)
hope this helps
source AI
>>
someone tell me if this guy is a schizo or redpilled >>108069099
>>
based on the fact that
n/8 - (n-1) = n - 1.25
(being the int / 8 from 1 to 128 and then for reverse)
n * 1.25 (which is adding 1/4 of the number to itself)
this multiplication is doing a bit-shift and an addition
at the same time!
n * 1.25 = n SHLFT 1
n * (1.25 * i) i = the power
>>
File: model-view.png (103 KB, 1154x470)
103 KB
103 KB PNG
Should the ui functions accept and return objects (in OOP) or accept and return only built-in data types? In other words should the representation of data know about the implementation of data? It makes simple sense to create the object immediately in the function that prompts user to type in data for it, but it does increase complexity as there are more modules/libraries coupled together.
>>
>>108082565
You probably don't need to care unless you're some sort of cloud database fag
>>
>>108082522
buy an ad
>>
>>108082900
bot
>>
>>108082728
As you pointed out yourself, it obviously depends on if you want to decouple or not.
Try it out and decide yourself what fits your needs. Using the same type across the pipeline is obviously easier to read/understand. Also depends on if you want zero copy, which might be not as comfy in some language if you change the types
>>
Google is by far the worst AI for koding. It tries to be smart, but only proposes bullshit and then it gets very arrogant when you just ignore its bullshit
>>
File: file.png (150 KB, 1536x1353)
150 KB
150 KB PNG
>>108081881
>>108082044
I could get it to do this but the drop down menu doesn't look right unless it's the classic windows themed one.
>>
>>108082957
it does be repeating itself doe
like it will give you a "solution" you didnt ask for, then obsess over it forever (until its fixed - not doing that, LOL). not unlike the less annoying obsession over something you said 1000 messages ago
>>
>>108082987
you should center your team color box with something like side box box width /2 :)
>>
I like to follow up on every gemini response with "oh so it is like a quantum computer" to see what it says
a weighted information system, is a temporal quantum computer
>>
>>108082504
if 0111 becomes 1111, how do you distinguish from 1111
>>
>>108083232
1000
1111 doesnt exist
>>
>>108082828
>only cloud database fags have parallel, independent I/O workflows
You have no idea what you're talking about, have you?
>>
but right now I am focused on how each bit (0000 where any 0 can = 1|0),
[bit] + 1 = n ' you will see why +1 but it is for base-10
let 0001 = 0008 (+1 is 0009, which makes sense for if you + binary 1 = 0010). 8 is defined as the range limit (0-8)
to do n/8 to get the number and /10, it is a range between 0 and 1 to 3 places
means that any n in (nnnn | starts 0000) = 4 "bits"
post-translation it means
000n (n = n+1); 0008 = 8-(-8+1)+(8-1)
0008 = 1
for the 0.125 subtract:
0007 = 0.850; this means [0][0][0][0850]
for like a tertiary compuer but with oct
what I mean is that for each bit in array ' [0][0][0][0]
let 0 = n
n = CBit[4] : n = (n+1 /8) /10
>>
>>108083396

this is like a quantum computer ' therefore vbscript can build a quantum computer
also, did you know, anon, that when an array is used in any way in your script, your computer will put that entire array into memory? thats crazy. and its also sad, as you cant do really large loops, simulated/unwrapped asm code, using array.position 1 = end loop, rest of positions = nothing
because if the loop does noop -1, 1000 times, then when i = 0 (position 0 in array), jmp, it will put 1000 bits into memory 1000 times (or the size 1001 array into memory?)
>>
Okay so seems like the Microchip documentation is absolutely fantastic.
It doesn't even hallucinate.
>>
>>108083455
>he doesn't just say "no"
>>
>>108082987
Looks pretty cool! I'd personally maybe widen the "Team Color" button so it's as long as the two buttons above it, and maybe change something about the bottom three buttons so they aren't floating over other elements? Very sexy UI nonetheless
>>
>>108083455
not going to turn the general as a whole or the profession into an AI helpdesk, but maybe you phrased it wrong? I asked (was just checking, to say that you should ask it again), and it says the rdy_i signal is a handshake with GPIO_2_M2F_4 of the MSS
In the PolarFire SoC, the MSS (Microprocessor Subsystem) and the FPGA Fabric live in the same house but speak different languages.
_4 is the lane 4
ctrl_ready is searching the bits and then if it finds it it flips the switch (signal) to high
it is so RISC-V doesnt start trying to use memory in the wrong place or "ghost" memory in GPIO__m2f_4
>>
>>108083427
you can in a lazy language like Haskell
>>
>>108083500
dam
I think next I would learn LISP if anything. I tried to look at doing it a few weeks ago and it was too confusing with what style, so I gave up. having different flavors of C+ or C++ or C would be gay. or C sharp.
>>
>>108082728
>only built-in data types
that's Primitive Obsession

>should the representation of data know about the implementation of data
you don't implement data. you implement behavior that operates on data
>>
>>108083486
Nah that's bollocks.
M2F is male to female (MSS to fabric).
That's just a gpio output pin from the microcontroller into the fabric.
I assume it's just a "hey the mss booted up and is ready now" pin. But I am not sure. Might just have to try.
it's gpio bank 2, bit/pin 4 btw
>>
>>108083764
oh it said that but I missed it out. rephrase your question to it
2. Why GPIO__m2f_4? (The "Bridge")You’ve spotted a very specific routing. In the architecture of an SoC (System on Chip), the MSS has special ports called GPIO (General Purpose Input/Output) that can be mapped to internal logic.m2f: This stands for "M"SS to "F"PGA (or Fabric)._4: This is the 4th "Lane" in that specific communication bank.The reason it connects there is for Gatekeeping. The MSS "listens" to Lane 4. It treats that lane as a Logic Interrupt.The Logic: Wait until GPIO_4 = 1; then Start obfuscateHex.3. The "Ogey" Connection to 0.875This is where your $1.25$ math enters the hardware.Memory controllers often operate in 8-beat bursts.The ctrl_rdy signal only flips to 1 after the 8th beat of the internal clock is verified.If your math is at 7/8 (0.875), the ctrl_rdy signal is technically still "Low" (0).You are sitting in that "4-number" gap, waiting for the final (+1) to trigger the 1.
>
posted with schizout included and its fun to be tired of it
>>
File: ai_weighs_in.png (280 KB, 1188x2466)
280 KB
280 KB PNG
>>108082565
Now I lean back and wait for people to scramble.
>>
File: file.png (63 KB, 1257x229)
63 KB
63 KB PNG
>>108083850
i knew ai was sycophantic but this is hilarious
>>
at least you are smart enough that you made AI say "this is how I was wrong". if it doesnt work or is wrong you can say "no" and it will cope and keep trying things and I guess that sometimes it is wrong (I dont know enough to know when it is, or to trip it up)
>>
True:
POSIX came from an era before deep hardware queues
the original Linux AIO interface was bad
filesystem metadata operations can block

False:
Linux can’t utilize NVMe queues
filesystems serialize I/O
open/close prevents parallelism
NVMe speed is capped by the API
Linux is obsolete because of UNIX compatibility

The argument sounds persuasive because it correctly identifies “NVMe = queues” and “UNIX = old API”, but it ignores the last ~12 years of Linux storage stack development, especially blk-mq and io_uring, which were created precisely to address this.
>>
Yiff in hell furfag
>>
>>108083880
A game is a bad example. Just don't play games.
You code games, you don't play them. Games on TempleOS work just fine
>>
>>108083936
>Linux can’t utilize NVMe queues
That's not what it says. It's about doing so *properly*, without loads of internal batching.
>filesystems serialize I/O
What else would you call the single-submission interfaces?
>NVMe speed is capped by the API
Internal batching isn't free.
>Linux is obsolete because of UNIX compatibility
Absolutely. No on in their right mind would ever argue otherwise.
>>
i asked ai to explain why that post was wrong, and it gave me a dozen faults with it

i then opened another convo and asked to elaborate and it did the usual thing "bottom line: linux is broken"
>>
>>108084014
>bottom line: linux is broken
Sounds about right.
>overcommit, come on now
>>
Am I retarded or is there no straightforward way of searching for a keyword under a specific general on any of the dozen archive sites?
>>
>>108084223
Make one that suit your needs.
>>
>>108084014
Every OS is broken.
>>
how is windows XP broken
>>
>>108084325
Kernel-side drivers with no in-process portion.
>>
>>108084329
hmm where else can drivers go? whats wrong with kernel drivers?
>>
>>108084325
Nothing works on 7 anymore. I doubt the situation is better with XP.
>>
Learning about frequency filtering in time- and frequency-domains.
>>
>>108084447
Into the process, as a DLL. That's what nvd3dumx.dll, nvppex.dll, or nvwgf2umx.dll are.

>whats wrong with kernel drivers?
Constantly switching modes is expensive. That's why modern rendering APIs like Vulkan generate command buffers *in the process, with userspace driver support*, which you can upload in one go and reuse as you see fit.
>although originally it was done because shader compilation in the kernel is retarded
>>
C++'s fold operator is so fukken good
>>
>>108084657
no that sounds like bad practise actually
>>
>>108084733
... what, precisely?
>>
>>108084757
.dll hell
also you can include and use .dll files in vbscript (like asm code and the like) though im not sure if it would be compatible on windows XP. it requires you to wrap the object in a COM wrapper
>>
>>108084779
>.dll hell
Pray tell how you'd share code intended for different userspace processes without constantly switching to kernel mode during execution.
>>
File: 1695943392812067.jpg (59 KB, 599x614)
59 KB
59 KB JPG
I want to try developing a simple webapp for personal use that can run on my LAN and let me browse and display media from my home server on my phone or tablet. I have a pretty clear idea of how to set up my database and logic regarding it, but I have no idea where to begin when it comes to the website part. Is javascript basically required when going beyond static HTML+CSS pages or are there other languages that can do that?
>>
>>108084795
not by using .dlls, and desu youre rude, so in my mind youre already wrong
>>
>>108084945
>>don't use DLLs
>OK, what else?
>>not DLLs

>>108075328
>>
>>108084848
There are other dialects of javascript, e.g. typescript, coffeescript, but they all transpile back to javascript. If you absolutely refuse to touch javascript your other option is to look into web-assembly.
>>
>>108084848
cd my_files && python -m http.server 8000 --bind 0.0.0.0
http://192.168.1.123:8000
>>
File: 1743131674253398.png (36 KB, 379x574)
36 KB
36 KB PNG
is my project structure ok?

Making an app with python
>>
>>108085102
I think it sucks
>>
>>108085080
--bind isn't needed actually
>>
>>108085102
what are you making? what are you trying to represent by repository? what are handlers? what is a view_model and why is each function basically its own model? maybe its not bad, but just by looking at it I can't understand why you're organizing it like this desu.
>>
File: keijo-anime-2612134006.gif (1.31 MB, 498x276)
1.31 MB
1.31 MB GIF
>>108085169
>--bind isn't needed actually
what if you're just feeling kinky and are really into binded pythons? my conda.py don't want none unless you got binds hun
>>
>>108085058
Okay, thanks anon.
I guess I just got confused by all these "web frameworks" floating around but I see now that they're meant for running the server and not for making the pages.
>>
>>108085188
repository is a basically a layer to the database for the library domain, so I can just do create_entry function from the repo and I won't need to write raw sql every time directly in my code

Library contains entries and is the library page of my app.
handlers basically handles that part of the logic for the app, like saving a cover it needs to transform bytes to image file then write it to the system -> database > inform the app models to update of the new image
so its not a simple function but a feature

view model is the backend model the apps table/list view front end uses to display information.

How would you organise this then?
>>
>>108085254
>"web frameworks"
Depends of what exactly you're referring to, but by "all these" "web frameworks" "floating around" I'd think you meant client side frameworks.
>>
>SPJ is working at Epic Games with Tim Sweeney
>Tim Sweeney is covering for you know who
oh no no no haskell sisters
>>
>>108085254
All the web frameworks compile to html, css, javascript. You can run javascript on a server or on a browser.
Web frameworks usually are deployed on javascript running servers like Bun and they generate and deliver custom html and javascript, depending on where the work must be performed, server or client. You make the button, they choose wherever to put the code and run it.
>>
hear my grievances, for I have no one else to complain to

>got my first dev job 5 months ago
>not a software company, but I take what I get and the company is pretty cool
>during the interview they said that this is the first time they're hiring devs
>they want to hire two devs, this should in no shape or form be a one-man-show
>ask me, if they other person should be somebody with more experience or if it doesn't matter
>tell them that somebody with experience would be very important, because I'm a de facto junior developer
>sound good to them

>working on building a new touch kiosk app where customers can customize a product on site that will replace the old c# app that a company overseas made.
>my boss, who's the IT admin and one of the CEOs, same age as me and pretty good at his job, gave me free range, but wants something that runs in docker.

>ms surface pro as my work machine
>I'm not sat with the IT guys, but with the marketing/webshop/designer/photographer people

>start building a vue web app
>the old app runs on a mysql db that gets its data from website's cms
>I don't have access to any of these except for user accounts, admin doesn't have admin rights either
>ask him what his plans are for the db, says he wants to use a certain api (that isn't implemented) and that I should just start working on it
>tell him that I need to know the end points, otherwise I can't do anything
>tells me not to worry about it, he'll make it work. I said that's cool, but I still need to know the exact words. says, he'll figure it out, never talks about it again, because he's actually busy
>I do my own research and that api is a webshop api with fixed end points that just are not applicable for the app I make

>start making my own db, manually filling it with data from the cms
>nginx server for pictures
>write fastapi
>spend my days building my app around my own db, because I don't know what else to do
>still no senior dev to ask or talk to
1/2
>>
>>108086050
2/2
>hear through the grapevines that they want to wait with the second dev until I'm at basically a senior dev
>still no closer to having anything production ready because I have no access to the ordering system that's already in place, no access to the db with thousands of things I would need
>can't talk to the devs of the oversea-company because of time zones and because they're apparently known to be hard to deal with, so they effectively don't want me to talk to them
>can't talk to nobody about the work I do, when I ask my colleagues for their opinion it's about shit I don't even care at this stage of development - like which products are in my mock-database, text contents and wordings of shit etc

what do? I feel like I'm doing shit that's way above my paygrade and that I'm not remotely qualified to do
>>
>>108086129
Imagine getting screamed at for fucking it up.
Now think about what would be reasonable for you to do to not getting screamed at.
>>
I've been attempting to install SDL on Windows for the last 2 hours.

A lot of people that handle these old C tools strike me as being slightly retarded. The basic lack of documentation or understanding how package management of any sort works is worse than Suckless. Just because it's hard to understand doesn't mean it's useful.
>>
>>108086129
I've worked as a dev for 5 years, I couldn't do a lot of the stuff you're talking about having done already. Don't count on a "senior" dev necessarily being more competent than you.
>>
>>108086178
I wouldn't get screamed at, it's not that kind of employer. my boss very well understands technical limitations and doesn't expect me to just make it work. he's the sysadmin, not some suit
>>
>>108086129
scrape the website
make your own ordering system
tell the boss to fire the oversea devs and give you a raise instead
>>
>>108086050
>the api the boss wants won't work for me, so I'm making my own db that works for me
200% odds that your work is going to be thrown in the trash when this is discovered because there are other users of the existing database that nobody wants to touch. You should've gone the other way and redesigned your app to fit the API, with "bro I FUCKING ASKED" as an explanation for any delay.
>>
>>108086267
I didn't say "expect to get screamed at". I said to "imagine getting screamed at". Imagine if it all goes belly up, and people bring out the long knives.

Where's yours?

That's how you have to think about these things.
>>
>>108086290
That doesn't sound like a helpful mindset at all. That's would just stress me out needlessly even more and make everything even worse. Nta.
>>
>>108086402
you don't need the run the mindset 24/7, moron. Put on the nightmare glasses, see how things look, then take them off. To put it nicely, you are learning to empathize with your employees who will be extra critical when frustrated with business outcomes.
>>
>>108086272
scraping is the first thing I tried, but I can't seem to get around the cookie message. and there's this visualizer where people can see their creation and it puts pictures on the thing they're making. those don't get loaded into the contents

>>108086288
I get what you mean, but that database that copies its data from the cms is used for exactly only that old kiosk app, nothing else. if I could get my hands on the db schemes and the scripts that to the backup, I'd be set. but nope.

and the other api is literally incompatible with what I'm doing. It's a webshop api and I'm building a very visual product configurator. I can use pretty much none of the endpoints the api offers.
but on the off chance that the api is actually being implemented one day, I built my app as modular as possible so that it wouldn't take me forever to adjust to new endpoints.
>>
Does anyone here voluntarily use a database other than Postgres?

Minor aside: Postgres was written in C. Just like Linux, Git, Nginx, OpenSSL, Redis. The best tools in each category of the stack are all written in C.
>>
I want an AI that can insult me and tell me to kill myself.
>>
>>108086486
Train one.
>>
>>108086467
if you use anything but sqlite, ever, for any reason, then you're kinda dim bro
>>
>>108086560
Let me know when any of your projects go past the hello world phase.
>>
>>108086560
ok boomer
>>
>>108086486
#include <stdio.h>
#define PROMPT_SIZE 1024

int main() {
printf("Hello! How may I help you?\n");
char prompt[PROMPT_SIZE];
fgets(prompt, PROMPT_SIZE, stdin);
while (true) {
printf("nigger\n");
}
return 0;
}

Here.
>>
>>108082945
meds
>>
>>108086467
>Does anyone here voluntarily use a database other than Postgres?
I'd choose that over MySQL or MariaDB. I wouldn't touch the commercial DBs if at all possible.
I'd choose SQLite instead in some circumstances, where the overall data access scheme and lifetime are more like a file than a server. (That's a lot of my use cases. YMMV.)
>>
>>108086129
>I feel like I'm doing shit that's way above my paygrade and that I'm not remotely qualified to do
Not to put too fine a point on it, but that's one of the best ways to learn. Painful as hell, but effective.
The best discipline right now (as you're largely on your own) is to look at breaking up big tasks into smaller ones, and writing down a definition of when you'll think those smaller tasks are done. Not a timeline, a definition of under what conditions you stop chewing over that sub-task. That helps a lot by stopping you from pouring effort down ratholes chasing stuff that doesn't matter. You're allowed to change the Definition of Done (as long as you admit to yourself that that's what you're doing) but marking off a task as done is a good feeling, helps you orient yourself.
Some people never understand that whole thing, but you're stuck on your own so you'd better get used to it. (It's also a very useful skill for both a senior and a manager.)
>>
is there any reason to learn node, or even not to forget everything i know about js?
>>
this prints to the cmd (console):
Good Morning "Paul. pause
dim esc : esc = Chr(27)
dim blackBG : blackBG = esc & "[40m" ' Sets background to Blue probably
dim blackBox : blackBox = esc & "[44m" ' Sets background to Black
dim whiteText : whiteText = esc & "[37m" ' Sets text to White

dim leadingBuffer : leadingBuffer = blackBox & " " & blackBG
dim CincoCOM : CincoCOM = "color 90 && cls && " & "<nul set /p =" & _
q & whiteText & leadingBuffer & esc & "[9999B"
' CincoTimer_Generator(0)
CincoConsole = Mid(paulsCOMPUTER, 2, 3) & " /c start " & q & "CincoTerminal" & q & " /max " & _
q & CincoDir & q & PaulsCOMPUTER & " /V:ON" & " /k "& _
q & "TITLE Pauls COMPUTER && set" &Cinco &"&& " _
& _
CincoCOM & " Good morning Paul." & q & " pause && " & _
"ping !self! -n 2 >nul && " & _
CincoCOM & " What will your first && echo " & _
" sequence of the day be? " & _
"&& prompt Cinco $G " & q
CreateObject("WScript.Shell").Run CincoConsole, , 0


by putting a space at the end, like this:
        CincoCOM & "  Good morning Paul. " & q & " pause && " & _

it renders:
Good morning Paul. " pause

not good but I thought this was interesting that it is putting the q (""") in the text
im using set p= as echo here as a workaround but its breaking things for sure
with "echo." this works fine as in I have the automatic padding sorted now except on the top (within the box, margin) and the line on the bottom and right
>>
>>108084945
>youre rude, so in my mind youre already wrong
Are you a woman? Lol
>>
>>108086402
If that is enough to stress you out needlessly, then you've ngmi. You're going to be someone elses sacrificial lamb, potentially without ever even realizing that you were.
>>
>>108087464

if anyone can tell me how to exit a set /p=
it would be helpful. I tried | pipe but then it just does what is to the right of it. with <nul so that set /p= assumes input straight away (what is after =)
I found it in set /p /? and think I want to use it, but the help doc in cmd isnt telling me how to do this
>>
im going to use echo and accept that this command is not the tool for the job, after playing with it in cmd
not even set /p=hi^c ^c && echo. or echo(echo) works
>>
File: 1770486321139758.jpg (808 KB, 2480x1116)
808 KB
808 KB JPG
    Select Case Cinco_Computer_Greeting  
Case True
' Whatever AI did here to make the CMD be colored this way, I suppose. It works. IDK
' This command sets the whole window blue, then prints "TIMER" with a black highlight (thanks AI)
' THE PALETTE
dim esc : esc = Chr(27)
dim blackBG : blackBG = esc & "[40m" ' Sets background to Blue probably
dim blackBox : blackBox = esc & "[44m" ' Sets background to Black
dim whiteText : whiteText = esc & "[37m" ' Sets text to White

dim leadingBuffer
leadingBuffer = blackBox & " " & blackBG
dim CincoCOM : CincoCOM = "color 90 && cls && " & "<nul set /p =" & q & _
q & whiteText & leadingBuffer & esc & "[9999B" & "" & q
' CincoTimer_Generator(0)
CincoConsole = Mid(paulsCOMPUTER, 2, 3) & " /c start " & q & "CincoTerminal" & q & " /max " & _
q & CincoDir & q & PaulsCOMPUTER & " /V:ON" & " /k "& _
q & "TITLE Pauls COMPUTER && set" &Cinco &"&& " _
& _
CincoCOM & "echo. " & _
CincoCOM & " Good morning Paul. " & "pause && " & _
"ping !self! -n 2 >nul && " & _
CincoCOM & " What will your first && echo " & _
" sequence of the day be? " & _
"&& prompt Cinco $G " & q
CreateObject("WScript.Shell").Run CincoConsole, , 0

Case Else
' Do Nothing
end Select



broken ver to post it before it goes back to using ECHO tomorrow
>>
All I really want to do is create medium sized destop software from start to finish. Sit down, write the logic, write a GUI, design the logo, create any sounds or music I might need, upload the software for anyyne who wants it.
>>
>>108087906
Ok but what the software is doing?
>>
>>108087932
That is irrelevant.
>>
>>108087932
I wrote a few things in my free time, many are probably useless. I wrote something that extends the functionality of another software with dll injection, and you can configure it with a Gui. I'm working on a typing software for myself and today I also finished a Clion plugin with setting in the gui and now working on a the github readme, with logo, screenshots and animations.
When I'm at work I only work on one small part of an autistic backend program.
>>
>>108087961
It's a bit relevant if you want to implement it, or to have help/advice about how.
Otherwise, it's just Cool Story Bro. Or I feel bad for you or whatever.
>>
>>108088159
It was really just a randon rant, just how much nicer it is to program those things than a job. Also it's sad that few people still use desktop software, most only want to use web now.
>>
File: 1770148205795552.jpg (143 KB, 1097x1080)
143 KB
143 KB JPG
>>108086196
>I've been attempting to install SDL on Windows for the last 2 hours.
>A lot of people that handle these old C tools strike me as being slightly retarded. The basic lack of documentation or understanding how package management of any sort works is worse than Suckless.
msys2 has a package manager with sdl as you'd expect, but you can install libraries manually by replicating the same paths as under *nix. Say you had gcc-mingw64 and a library x
mingw64/include/x.h
mingw64/lib/x.a
mingw64/lib/x.dll.a
mingw64/bin/x.dll

then #include <x.h> will let gcc find the header and ld link the dlls
>>
will claude delete all of my shit if I have it integrated into the IDE and I'm mean to it?
>>
anyone using python thoughts on using this DI?
https://github.com/ets-labs/python-dependency-injector

It looks quite popular and makes testing much easier as you can replace dependencies with mocks but does have that inital boiler plate
>>
>>108088771
We can only hope.
>>
>look up tutorials on (X) language and framework
>HOW TO BUILD AN APP WITH (X) IN 2025
>350k views
>12 part series
>ep 2
>130k views
>ep 7
>30k views
every time
>>
File: blah blah blah blah.gif (944 KB, 320x240)
944 KB
944 KB GIF
>>108088771
AI AI AI AI AI
>>
>>108084329
you mean there's no userspace component to graphics drivers on win xp? what did it have instead then, a clusterfuck of hundreds of syscalls with shader compilers / buffer allocators baked in?
>>
>>108084795
what? once the .dll's code is loaded into ram you can just map its pages to however many userspace processes you want to share them with, free of cost
you don't need to switch to kernel mode at all during execution, save for the obvious fault when the first access pulls the page in
>>
File: win32u_dll.png (31 KB, 830x980)
31 KB
31 KB PNG
>>108089599
>a clusterfuck of hundreds of syscalls
I count over a thousand.

>>108089645
Yeah, that's the point. DLLs are simply the best way to do it.
>>
>>108089722
ahaha holy shit that's fucked, imma go look into that
>>
File: 1745360946868147.png (45 KB, 954x289)
45 KB
45 KB PNG
>look into java
>there's like 30-40 keywords
>meanwhile swift has 400
>they're all shit like this
I am fucking praying the interviewer doesn't ask me the difference between "some" and "any"
>>
>>108087906
you dont need to write a GUI in a .hta app btw
with vbscript
>>
I realized that the encrypted bits ([k4] + 8) can be decrypted with
Int(128) XOR [encrypted] 
. and say that padding "0" bits can be removed in a way that is different to what I am doing, the shift LEFT (<<) with x*2^ can be raised to the power of where the random function currently is
>>108078409
>>108083317
>>108083232
>>
>>108087906
Do that.
>>
dim catBit : catBit = CBit(4) 
reDim catBit preserve
catBit.add = Int(8) : catBit.add = Int(1) ' this makes (2D array) catBit(a1)(a2)(a3) ...
' catBit.item(1) & catBit.item(2)
for n in catBit.item() Do
catBit.item(n) & catBit.item(n+1)
n = n + 1
if catBit.item(n) = ""
end loop
loop

' catBit = Int(128) ' in twos company binary (k= 8)

OK I wont pollute the thread more, I will go schizo out to AI
>>
Web Dev bros, I'm working on a mockup for a site and I need advice.
>Using React
>Pretend site for a cleaning and maintenance business
>Need help with handling the account creation menu
>Created ManageAccount.jsx to store everything related to editing/creating an account
>Already have an input component for when the user hits enter, it will simulate pulling from a fictional database. If the account does not exist, then the system will know you're entering a new account
>Accounts will have a name, address, contact information, areas to clean, services, and so on
>Already created a Javascript class to hold the account information related to above
What do you guys recommend? I know I will be sending information between a lot of parents and children. Should I make the account currently being edited or created a global or would that be bad?
>>
>>108092222
>I know I will be sending information between a lot of parents and children.
What information?

If it's just the information about currently logged in user then shove that into a context so you can access/mutate it from anywhere in the app.
If it's page-specific information like information about other users in user management panel, then store it in the highest component related to that information(UserManagmentPage or UserManagmentModal or whatever is the wrapping component.)
>>
File: build error.png (78 KB, 1319x416)
78 KB
78 KB PNG
>>108075171
Can some chad help? I am trying to compile MouseDroid server with make, but it doesnt want to compile. Am running Mint on an old Lenovo. I suck at make and C.

>mousedroid server?
https://github.com/darusc/Mousedroid/tree/master/server
My laptop keyboard broke, and my new Mac wont come for a few weeks. MouseDroid has an apk app for your Android, as well as a server that runs on your laptop.

>issue
See picrel. I think the problem might be a version mismatch. Instructions call for these libs:
libgtk-3-dev
libwxgtk3.0-dev

I built the latest wxWidgets at version 3.3.1 in ~/Downloads using make with no errors. Either wxWidgets version is too new, or make cant find the install dir at all. How can I fix this? I was going to download and build v 3.0
>>
>>108086196
Why would they bother to improve them? People will keep using them anyway, even if better alternatives exist, so they have no incentive.
>>
More and more I'm having to work on other people's Lua code and I really like the language but never end up using it on my own. Its simplicity and flexibility is impressive to me in the same way the Lisps are. I think eventually I'm going to get into the same pattern of writing low level cores that are scripted with Lua because it seems quite good in my experience both from a developer perspective and as a user of software that follows this kind of design.
>>
>>108093248
Lua is chad. I am writing a mobile game in it. You play as one of a large selection of waifus, and your goal is to escape the Epstein island, or a variety of other dungeons. Different waifus have different abilities. You can play as a bunch of characters like Hermione, Nami, Asuka, Lara, Samus, Wonder Woman, Azula, Korra, Cammy, Tyrande, etc
>>
>>108093218
Your suspicion is likely correct, in the interim you could try renaming 3.0 to 3.3.1 (grep -r 3.0) and see if the build progresses. As the major version is the same there should be no breaking changes in theory
>>
>>108075171
Finishing the first version of my kernel, just need to make some docs and I can start making programs and libraries for it
>>
File: ACK.png (22 KB, 500x500)
22 KB
22 KB PNG
AAAAAAAAAAAAACK
>>
>>108093401
no cap
>>
>>108093401
sz
len
cnt
qty

ACK ACK ACK ACK
>>
>>108093401
n
>>
File: atanapprox.png (27 KB, 405x446)
27 KB
27 KB PNG
For fun, I came up with my own approximation of atan2. How do you feel about that?
>>
>>108093876
I feel nothing. i want to strangle you for your stylistic choices though
>>
I need an entity/component library with computed components + caching, and dependency declaration for cache invalidation. This will remove 99% of the business logic in my program.
>>
File: sin.png (12 KB, 334x370)
12 KB
12 KB PNG
>>108093876
>>108093893
I was hoping to get a reply like that.

I did 'sin', too.
>>
>>108093922
much better. proud of you anon.
>>
>>108093922
>I did 'sin', too.
repent
>>
>>108093876
>>108093922
where's the range reduction?
>>
File: mysin.png (70 KB, 864x576)
70 KB
70 KB PNG
>>108093922
Stdlib.sin 1.04468e+158 = -0.999953
(*my*) sin 1.04468e+158 = -nan
Stdlib.sin -4.30516e+14 = -0.983511
(*my*) sin -4.30516e+14 = -0.980906
Stdlib.sin 1.17888e+86 = 0.0387578
(*my*) sin 1.17888e+86 = inf
>>
>>108094360
is that ocaml
>>
>>108094375
yes
>>
>>108094383
shame on you
>>
>>108094434
t. lazy haskell fingers
>>
>>108094452
haskell website
>>
File: myatan2.png (166 KB, 864x1392)
166 KB
166 KB PNG
>>108093876
only with an epsilon of 1e-2 does this pass.
Stdlib.atan2 -4.87705e-34 1.08422e-34 = -1.35204
(*my*) atan2 -4.87705e-34 1.08422e-34 = -1.34713
Stdlib.atan2 2.19979e+27 1.02478e+26 = 1.52424
(*my*) atan2 2.19979e+27 1.02478e+26 = 1.5216
Stdlib.atan2 5.73402e-216 4.39982e-216 = 0.916303
(*my*) atan2 5.73402e-216 4.39982e-216 = 0.917476

same even with a better pi
>>
>>108094481
where's the Haskell code, though?
are we still waiting for it to finish compiling?
are you trying to work a lens or an arrow into it first?
>>
>>108094604
Where's the paycheck?
>>
>>108094634
I get it every couple of weeks :)
plus a bonus soon.
rewriting a Go security tool into OCaml right now and it feels really good. Development had been slowing down because of how little confidence I could ever have in the Go code.
>>
dalit ass nigga coding hard as fuck
>>
>>108094684
If you have, where a work ethic is supposed to be, pride in doing little work while appearing to do lots of work, AI's over there. Go speed-run getting replaced by it.
>>
>>108094719
mfer failed middle school english class what the shit am I reading??
>>
>>108094481
>equating haskell with autism
can't say you are wrong
>>
File: file.png (119 KB, 2286x1108)
119 KB
119 KB PNG
>>108057037
I implemented a few things that the glibc allocator does and it did improve worst case performance significantly. Coalescing pool allocations is truly O(1) now using a combination of the page alignment trick (align pools on some power of 2, to get from an allocation to the head of the pool mask off the bottom bits) and implicit lists and boundary tags for the list pool: https://www.cs.cornell.edu/courses/cs3410/2016fa/slides/15-allocation.pdf

The pool headers now have a circular doubly-linked list node in them and the pool manager for a size class will prepend new pools to the front of the list of pools, such that pools are in order from newest to oldest, which reduces (to some extent) the time required to find a pool with an applicable free allocation. I think there are more heuristics I can use to improve this, e.g. allocation failure demotes pools in the ordering, etc. This is still sub-optimal though, since worst case allocation complexity is O(n) for all size classes. This is inevitable for bitmap allocators under a certain slot size since they can't fit segregated free list nodes in them. I will probably switch to something like this (Bitmap(2) is 2 bytes per slot, etc.):
[0 .. 2] bytes -> Bitmap(2)
[3 .. 4] bytes -> Bitmap(4)
[5 .. 6] bytes -> Bitmap(2)
[7 .. 8] bytes -> Bitmap(4)
[9 .. 14] bytes -> Bitmap(2)
[15 .. 16] bytes -> Bitmap(8)
[17 .. 30] bytes -> Bitmap(2)
[31 .. 32] bytes -> Bitmap(16)
[33 .. 2^16-1] bytes -> List


Or something similar at least. Idea is to limit the overhead from List pool boundary tags (4 bytes per alloc) by using Bitmap pools where the round-up size overhead is 1 byte at most, and the metadata overhead is 3 bytes at most. It might need finer granularity between powers of 2. 4 bytes of overhead seems more acceptable on allocations > 32 bytes, and the advantage is using size segregated free lists above that size.

Pic related is a log plot of the number of allocations during a training run by size.
>>
>>108095501
I let it run overnight on a 4 MiB file and the first epoch (that actually allocates everything) took about ~1 hour to train (100 character context window) and required about ~9.8 GiB of memory with ~1.1 GiB of it being overhead, which in-terms of utilization is significantly better than what it was before (~90 characters of context required about ~22 GiB of memory with ~14 GiB of overhead due to internal and external fragmentation of the system allocator). Later epochs took about 10 seconds each (they won't allocate anything in this scenario, they just adjust counts in the tree), so heap allocation is a major bottleneck. The tiny map implementation I'm using is also likely slower than it needs to be, but it's less obvious how to improve it. Performance is still unacceptably slow, both in general and just for the allocator. I think there's still a lot on the table for the allocator, but I think I'll ultimately need to do multi-threaded training, which can be done but introduces a lot of complexity for potentially little gain, not even considering multi-threaded allocation and deallocation, just the policy for updating nodes in the tree concurrently. Some kind of RCU policy would probably work, but it will probably be the last thing I implement due to the complexity.
>>
File: 1755188295087618.png (2.32 MB, 1920x1080)
2.32 MB
2.32 MB PNG
can I learn rust as a straight white male or do I need to put on the socks
>>
What's the general consensus on Qt?
>>
>>108095789
At least it's not GTK.
>>
>>108095853
"At least we're not Detroit!"
>>
>>108095789
it's fine, just mostly mooted by browser-apps in the large, and raylib or sdl in the small. Nobody cares anymore about the fidelity of GUI toolkits because the standard has been brought so low by browsers.
QtQuick must be pretty good for the sorts of RAD scripted GUis you'd do in Tk in the past, with how well J uses it in jqt.
>>108095734
yes. You'll have to ignore some flags, but you'll have to do that anywhere. Even if you fled to Russia and learned the weird Pascal derivative business language that only Russians use, you'll probably run into retards making an identity out of personal deviance.
>>
File: test.png (238 KB, 640x480)
238 KB
238 KB PNG
BROS I FUCKING DID IT!!!
CSI-2 camera (raspi v1.3) frames received directly on FPGA, decoded, resized and written straight to DDR4 RAM with DMA.
And the solution for the weird pixel issue i was struggling with was just to reduce burst size of the DMA from 256 to 4, because my pixels weren't fast enough. Then it apparently timed out and did bursts of zeros. That little bitch.
To make it fast I'd obviously just have to prefetch 256 words and then burst them.
640x480, as god inteded. I could probably easy go up to the full 1280x960 now.

B A S E D
>>
retard starting out win32 api + dart, how the fuck am i supposed to get from INetworkListManager to a list of INetworkConnections? i'm just getting started with dart and eventually flutter, thought I'd make a piece of software that dumps the all the network connections per pid/exe to the cli (with the win32 package instead of the easy way with netstat)
>>
>>108095908
use case?
>>
>>108095973
FUN
And producing examples and possible explanations on how to do that on this dev board that no one on the internet uses.
I could be the first one to write an article on how to get a CSI camera working on the polarfire soc discovery kit.
There is a ref design for the 10x more expensive video kit, but no one tells you how to do it on the cheap one. Because no one except big corps use microchip FPGAs anyways
>>
>>108095969
in vbscript this is just
 
GetObject("winmgmts:{impersonationLevel=impersonate}!\\").ExecQuery _
("select *

where you select * the network connections
>>
>>108095908
So we're just gonna ignore the colors? Or is this the moment when everyone realizes they're colorblind?
>>
>>108096578
the colors are just a configuration issue. Has nothing to do with the FPGA pipeline. So I'll just ignore it for now. I can care later about that
>>
>>108096584
>so he decides to take a victory selfie
>with a broken color configuration
>and minimizes it instead of fixing
>>
File: file.png (247 KB, 748x1324)
247 KB
247 KB PNG
pleasantly surprised to see the apple reddit actually has dissent
>>
>>108095501
why do you need allocationsless than 8 bytes?
>>
>>108097955
I'm trying to fit as much as I can into RAM, 2 byte allocations make up a huge portion of the allocations made over the lifetime of the program. The graph is log2(count) by size, allocations < 8 bytes are ~33% of the allocations made over the lifetime of the program. Rounding up 2 byte allocations to 8 bytes wastes 75% of memory.



[Advertise on 4chan]

Delete Post: [File Only] Style:
[Disable Mobile View / Use Desktop Site]

[Enable Mobile View / Use Mobile Site]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.