Erlang

>>101249866
>Power the telecommunication infrastructure
>meme

>>101249911
Is this like those companies that had one guy write some 100 line D utility program 10 years ago, and then people going on about how the company is "built on D"?

File: 1716032014322780.jpg (46 KB, 1024x1024)

46 KB JPG

I only develop for windows, mac and linux aren't attractive

>>101250039
this but ios

>>101250039
>Frogposter
>Says something retarded
A tale as old as time.

>>101249828
H!

File: 1700734254159508.png (1.01 MB, 1467x1005)

1.01 MB PNG

So I'm looking at this crazy-ass program that compiles itself. It's from IOCCC 1989 by Jay Vosburgh

[intentional blank line]
#include <stdio.h>
#define QQ      1
#define TT         1
#define cc main(c,v) int c; char **v;{char tt[12],qq[7]; int q=0,o=1,l=1,m=1;struct {int c;} f;
#define incest qq[6]='\0';tt[11]='\0';if(QQ==atoi(v[1])+1){(void)fprintf(stderr,"%s factorial = %d\n",v[1], TT);exit(1);}o=c+f
#define x ;while(EOF!=(o=getchar())){if(l && q=='Q' && o=='Q'){l=0;(void)getchar();(void)fread(qq,6,1,stdin);(void)printf("Q %6d",atoi(qq)+1);}else
if(m && q=='T' && o=='T'){m=0;(void)fread(tt,11,1,stdin);(void)printf("T %9d\n",atoi(tt)*QQ);}else {q=o;(void)putchar(o);}}exit(0);}
cc incest.c -o x 
#define zxc ;{/*
cat incest.c | x $1 >! x1
if ($status != 0) then 
exit
endif
mv x1 incest.c
chmod +x incest.c
exec incest.c $1
exit
*/

see pic rel for an explanation
and https://ioccc-src.github.io/temp-test-ioccc/1989/fubar/index.html for submission

However, being a noob, I'm confused about one part in the "de-entangled" output:

o=c+f -o;

If I try to compile this bit (which you're not supposed to, you're supposed to always start with the original script and fubar.c), it complains about
>error: invalid operands to binary + (have ‘int’ and ‘struct <anonymous>’)

So, my question is:
how do you sum a struct and an int like this? Isn't this syntax wrong?

>>101249990
Nah it's widely used but there are few erlang jobs

File: ‪ 6890965321.gif (1.15 MB, 197x192)

1.15 MB GIF

I'm experimenting with WinUI 3 publishing and it turned out that i can create a portable exe, some people on /g/ lied to me and said i can't do that, it's an okish experience but i think i'm forced to use msix instead of traditional installer
>>101250052
Based swift dev

>>101249911
It's meme outside telecoms

File: firefox_9YMahzwFDX.png (4 KB, 295x124)

4 KB PNG

I actually want to make a memelang given how I am dissatisfied with practically every language. Even if I end up like the guy who made D and essentially make the language I think is the most good but nobody uses, it would be fun.
But how do I make a memelang?

File: .png (6 KB, 590x41)

6 KB PNG

>>101250201
that's a typo
"ouroboros" is a define, "x" is a define
so the line "cc ouroboros.c -o x" gets split into

(ouroboros)(.c -o )(x)

and the expansion results in

(...}o=c+f)(.c -o )(;while...)

so it's actually

o=c+f.c -o ;

you can do

gcc -E fubar.c result.out

and ctrl+f "c+f"
see picrel

>>101250424
invent the syntax, attempt writing bnf just to get clarity. learn to write recursive descent parsers in c. create an enum with OP_FUNCTION1... then try to parse syntax into your bytecode. make a big switch statement in a loop that implements all of these builtin functions.

>>101250424
>Even if I end up like the guy who made D
OH RIGHT right right, worst case is you end up as an author of multiple C++ books while having a pHD in computer science with accepted proposals to the C++ standard and also worked on your 'meme' language 15+ years. WORST CASE.

>>101250704
I don't mean like he's in a bad place, I mean like he made the language he thinks is the ideal language and practically noone uses it

>>101250704
lole hi walter

File: p.jpg (415 KB, 684x909)

415 KB JPG

!

>>101250704
he means walter, not andrei lmao

>>101250039
I only develop for windows but I do all my development in linux

>>101250443
Ah! I see, thanks for pointing out the typo, and also for mentioning the -E flag with gcc. Wasn't aware that was a thing for preprocessor. It makes sense now

[blank line]
#include <stdio.h>
#include <stdlib.h>
#define QQ      4
#define TT         6

int main(int c, char **v)
{
    char tt[12];
    char qq[7];
    int q = 0;
    int o = 1;
    int l = 1;
    int m = 1;
    struct {
        int c;
    } f;
    qq[6] = '\0';
    tt[11] = '\0';
    if (QQ == atoi(v[1]) + 1) {
        (void)fprintf(stderr,"%s factorial = %d\n",v[1], TT);
        exit(1);
    }
    o = c + f.c -o ;
    while (EOF != (o = getchar())) {
        if (l && q == 'Q' && o == 'Q') {
            l = 0;
            (void)getchar();
            (void)fread(qq, 6, 1, stdin);
            (void)printf("Q %6d", atoi(qq) + 1);
        } else if (m && q == 'T' && o == 'T') {
            m = 0;
            (void)fread(tt, 11, 1, stdin);
            (void)printf("T %9d\n", atoi(tt) * QQ);
        } else {
            q = o;
            (void)putchar(o);
        }
    }
    exit(0);
}

/* this was called with ./fubar 3 */
/* 3! = 6 */
and now running ./ouroboros with 3 also yields the right output

>>101251407
I accidentally'd a codeblock inside a codeblock, that's interesting

so docker containers are just VMs basically?

>>101251542
Not quite. Most containers are more lightweight, and will share the same kernel as the host, but just be isolated from everything else.
But sometimes containers are true VMs.

The code will be closed source. It will not be malicious. Trannies will not fork my project and run it into the ground. You will accept it.

File: 414ZNWVCQCL.jpg (19 KB, 353x500)

19 KB JPG

Why are c "programmers" like this

>>101251719
I assure you there are no "chicks" programming in c.

>Boss: built our own Excel from scratch
>Me: that'll take hundreds of hours
>Boss: do it in 50
real conversation

File: 1574667018395.jpg (564 KB, 990x1200)

564 KB JPG

I'm writing some networking stuff, and the maximum size a message for this protocol is is 128MiB, but the typical size will be MUCH smaller.
Having to write a streaming parser seems kind of annoying, but having 128MiB lying around unused, or having to do a resizeable array seems annoying too.

So I was thinking about this:

size_t len = 0x8000000;
void *buf = mmap(NULL, len, PROT_READ | PROT_WRITE, MAP_ANONYMOUS, -1, 0);

while (1) {
    // Done with proper polling and likely multiple calls of this, which offsets/lengths properly tracked.
    // This is just to show the point.
    ssize_t nread = recv(sock, buf, len, 0);

    parse_message(buf, nread);

    madvise(buf, len, MADV_DONTNEED);
}

mmap won't actually allocate any memory until I actually use the pages, so while I'm fucking around with small messages, I'll only be using a small handful of memory.
And madvise(MADV_DONTNEED) will actually free the pages, so if I get a big message, I won't have a bunch of unused memory sitting around for the rest of the program.

Is my understanding of this correct, and is there anything wrong with doing it this way, outside of portability?

>>101252041
>lines and colums
>can go through them with the arrows
>save and load custom format
that is a really fucking simple task you dimwit.

>>101252046
it is fine

>>101252163
t. never used excel

>>101252175
no it's not
>>101252046
>Having to write a streaming parser seems kind of annoying
why write anything at all then?

>>101252242
>why write anything at all then?
I mean a parser that is able to stop and restart in the middle of reading, meaning you need to properly store/serialize all of your state.
It's obviously possible, but it's easier to just not.

>no it's not
Why not?

>>101252276
>why not?
nta but other people have solved the problem. just steal their implementation? it's either blocks or streaming. if you don't have consistent block sizes (or don't want to)... then stream it.

File: int skeletonBoners.png (108 KB, 656x436)

108 KB PNG

Do compilers/runtimes eliminate automatic variables as soon as they leave their scope or do they hang around for some arbitrary amount of time?
Also can they further restrict the scope of a variable if it falls out of reference before the end of a scope?
Like for example

var cond bool
if cond {
    // ...
}
// cond is no longer used; reaped by the runtime here
// more code is here before scope terminates

Normally I tighten scope manually, with braces or whatever language feature.

File: thecrane.jpg (116 KB, 411x700)

116 KB JPG

Newbie question:
You must flip a C89 style boolean (using int, there's no bool). Using a conditional because we don't want to set the variable to the same thing it's already set to (though this may already get optimized away by the compiler, not sure)

Which one do you prefer? In terms of readability

A.

#define THING_EMPTY 0
#define THING_POPULATED 1
...
if (thing_state == THING_EMPTY) {
    thing_state = THING_POPULATED;
}

B.

#define NO 0
#define YES 1
...
if (thing_populated == NO) {
    thing_populated = YES;
}

C.

#define NO 0
#define YES 1
...
if !(thing_populated) {
    thing_populated = YES;
}

In B and C, we're using the "positive" case as a variable name because I don't want to get into silly shit with double negatives (ie, !(empty) )

>>101252725

thing_state = !!thing_state;

Holy shit I love programming. I wish I could code 16 hours a day and then dream of it.

>>101252725
(Pretend "THING_" isn't there, it's easier to just say EMPTY/POPULATED for a symbolic constant that can get reused in any appropriate context)

>>101252777
what's stopping you?

>>101253096
Sometimes I have to eat or poo. But mostly stupid stuff like meetings. This week I have been almost only programming though and it's great.

>>101252751
Reading my own post again an hour later: just the one !.
Two !s are for forcing an arbitrary int value into 0/1.

>>101253319
or != 0

150 days to aoc

>>101253666
>150
>1+5+0= 6
>101253
>1+0+1+2+5+3 = 12 = 6+6
>666
what's the most satanic problem we can have?

>>101251736
That's a baby duck

>>101253726
they're called ducklings, and it isn't

>>101253740
A childish word used only by children
You have baby duck syndrome

>>101253818
>each animal species should have only one noun to refer to it
average american, touch grass

I still don't understand "const" after years of coding...
Can anyone explain in retard words pls

>>101253895
https://stackoverflow.com/questions/10906595/is-const-a-lie-since-const-can-be-cast-away

>>101253895
It's simple, it's just the opposite of what happens when const is not there.

>>101252390
you are retarded

>>101250412
So better than Rust then, Rust is a meme everywhere

>>101250412
It is not even used in telecoms anymore

>>101250226
Not anymore, most if not all large projects have switched to other languages.

File: FruitWarpAnimation.webm (99 KB, 384x384)

99 KB WEBM

>>101249828
This video is one continuous walk through the training input space of a diffusion network trained on some fruit images.

>>101249828
Penislang

>>101254691
the pen is langier than the sword

>>101252481
>Do compilers/runtimes eliminate automatic variables as soon as they leave their scope
In c/c++ you could probably guarantee that but depends on compiler. In a garbage collected language there is not a guarantee.

>Also can they further restrict the scope of a variable if it falls out of reference before the end of a scope?
Yeah, that is a thing in optimizers to conserve registers.

You would have to ask about a specific compiler to get a specific answer.

>>101249990
pretty much

>>101252276
>it's easier to just not
again, why bother at all if it's gonna be crappy
>Why not?
madvise is placebo
it won't free shit

>>101249828
The only use of broadcasted interrupts is to distract you and waste your time

File: dontneed.png (169 KB, 672x760)

169 KB PNG

>>101255210
>if it's gonna be crappy
Do you have any actual reason why it's supposedly crappy?
>madvise is placebo
https://man7.org/linux/man-pages/man2/madvise.2.html
It doesn't sound like a placebo to me.

>>101255269
>Do you have any actual reason why it's supposedly crappy?
it'll waste 128mb while all it needs is about 4kb
>It doesn't sound like a placebo to me.
in my experience it is

>>101255286
... what you just said shows you know nothing about modern memory management and should just shut the fuck up.
>in my experience
you speak in a weirdly niggerish way, do you happen to be that one nigger who was schizoing out about swapping?

>>101255286
>in my experience it is
Were you using MADV_FREE? That seems like it does it lazily, so it would be leaving the 128MiB still sitting around most of the time.
Google around, I do see discussions in things like chromium talking about their use of MADV_FREE.

every day seems bleak than the prior

>>101252481
Local variables are traditionally allocated on the stack at the beginning of the function's stack frame, which is discarded upon return. The compiler might not allocate it and just use a register, then it can just use it for anything else if it's not needed. But compilers do all kinds of optimizations, so there's no generic answer. I wouldn't worry about this.

>>101252481
This is the type of thing a compiler would optimize out for certain

>>101254950
>>101255440
>>101255471
Thanks Anonymous. Informative.
I figured it would depend on specific implementations but wondered if this was a generally/potentially solved problem.
Seems like it's not worth thinking about unless it came up in some performance measurement, but was curious about it.

just dont forget that the default optimization for gcc is O0 and it doesnt optimize that out, not even kidding.
Matter of fact the point of "global variables bad" is that compilers are advanced enough to optimize into arguments through registers (when they dont break and do call +0 for 20 instructions) and using a global forces the memory access

>>101255494
It is a solved problem, local variable usage gets optimized even on O1 on any compiler

>>101252481
>>101255494
Once you learn a little bit of the fundamental of how compilers work, you'll learn that compilers have internal representations such as a "register transfer language".
I forget the actual details of it, but maybe something like

a = b * c + d;

would get transformed into

$1 = b
$2 = c
$3 = d
$4 = $1 + $2
$5 = $3 + $4

You never "reuse" registers. and they're not representations of actual registers.

Now none of that shit is actually important and you absolutely shouldn't focus on it, but my point is that your code gets processed and "mangled" so far from your original source code into its own abstract representation, things like variable scopes really don't have any direct correlation to what happens in the outputted assembly. All it does is describe how this abstract representation is built.

File: APL.jpg (236 KB, 1624x626)

236 KB JPG

>>101255593
>your code gets processed and "mangled" so far from your original source code into its own abstract representation, things like variable scopes really don't have any direct correlation to what happens in the outputted assembly
lel. Frightening.

>>101254645
Always nice to see some visualisations.

File: smudge smudge.webm (1.11 MB, 480x640)

1.11 MB WEBM

>>101255630
Here's some progressing training output from a different input set, the top is the training inputs, the bottom is random preselected inputs. Also I made the previous one into a music video.
https://youtu.be/Nl73LXgqRJA