DDraceNetwork - developer

TsFreddie 10/19/2023 2:08AM

TsFreddie

Click to see attachment 🖼️

almost a perfect meme if the bottom right actually wrote SSL instead of SNL lmao

Gm

Morning

@Jupstar ✪ have any controversial or interesting topics crossed your mind today?

8:01AM

https://youtu.be/sHzdsFiBbFc?si=izInuKUPozRlt3ki

Andrew Struthers

Spiders On Drugs

8:01AM

This for example?

8:01AM

very interesting web development

GM

Teero

@Jupstar ✪ have any controversial or interesting topics crossed your mind today?

Yes. When do robots finally replace my boring useless work

Jupstar ✪

Yes. When do robots finally replace my boring useless work

Interesting. A war could help

10:39AM

In times of need people will be forced to innovate.

10:39AM

Let's start a war

No

10:39AM

This is a common miss thinking

10:40AM

The golden ages for humans were in longer peace times

Jupstar ✪

Yes. When do robots finally replace my boring useless work

what do u do at work

Ryozuki

what do u do at work

Coding xd

Welp I mean if everything is at peace and everyone is happy then it's good and people have free time.

10:41AM

to think of new things

Teero

Welp I mean if everything is at peace and everyone is happy then it's good and people have free time.

Yes. In fact most innovative ppl are not poor

the world is ending

So pressure is not a motivator

So ideas come from boredom ig

10:42AM

Mb

I think they come only from motivation

10:43AM

But having to think about, if u can manage life, does not help

10:43AM

E.g. being poor and forced to work

Jupstar ✪

But having to think about, if u can manage life, does not help

Hmm true

10:45AM

My classmate says necessities makes inventions

I'm really not sure about it. We have enough knowledge to print houses etc. But we don't do it

10:46AM

Because rich society and/or politics doesnt seem interested

It isn't necessary to print houses

10:47AM

Doesn't have enough gains

Rich Africans dont built wells

10:48AM

I'd say it's a matter of motivation to see the problems and actually wanting to solve them. And it's a societally effort

10:49AM

Some problems are simply complex and need lot of workers etc

Teero

My classmate says necessities makes inventions

But yes i could imagine that this could at least be a starting factor to talk about it more

10:50AM

Like if climate crisis hits harder. We will think about it more. But if we unlucky it will actually even demotivate ppl

10:50AM

We'll see

10:51AM

I guess there is no definite answer^^

thx for convo school is boring ^^

10:52AM

I might just invent something new

Creating something is also a matter of discipline

10:53AM

Sometimes it's hard and annoying and boring

10:53AM

And if u give up because of that.. rip

Jupstar ✪

So pressure is not a motivator

There are a couple things that were greatly accelerated by war and insane war budgets. The jet engine and radar technology both come to mind. Also the tiny vacuum tube transistors the US got made.

Jupstar ✪

I'm really not sure about it. We have enough knowledge to print houses etc. But we don't do it

Isn't it just less efficient than the precast concrete buildings we make nowadays?

Learath2

There are a couple things that were greatly accelerated by war and insane war budgets. The jet engine and radar technology both come to mind. Also the tiny vacuum tube transistors the US got made.

I bet the science behind these technologies were invented long before and probably not even by the military

11:04AM

So the missing piece was cheap labor force or whatever was expensive about these processes

Jupstar ✪

I bet the science behind these technologies were invented long before and probably not even by the military

That's true. Most was already discovered. Just needed the stupid levels of war budget

Learath2

Isn't it just less efficient than the precast concrete buildings we make nowadays?

Depends on the size of the buildings i guess. But in the end it probably also doesn't matter how high the price is.. at least in European countries u are probably low on humans rather than money xd

Jupstar ✪

Depends on the size of the buildings i guess. But in the end it probably also doesn't matter how high the price is.. at least in European countries u are probably low on humans rather than money xd

I'd actually argue it's neither money nor labour. It's just not good for the old people in charge to build more. If they build more their own homes lose value

Xd

If they needed workers they would import them from wherever. When germany needed workers turkey was happy to oblige

True

Teero

@Jupstar ✪ have any controversial or interesting topics crossed your mind today?

As @zogtib / (gotroyb) would say C99 is the best language to build upon. But it still has problems. tee_thinking

. So if the best things have problems why do invent new things which are easier to grasp at first but got more problems? Maybe it's because we want ppl to use smt that will not inovate things. (btw do not take this serious I just typed a word which would make sense after the previous with the context.) (edited)

Mr.Gh0s7

As @zogtib / (gotroyb) would say C99 is the best language to build upon. But it still has problems. tee_thinking

. So if the best things have problems why do invent new things which are easier to grasp at first but got more problems? Maybe it's because we want ppl to use smt that will not inovate things. (btw do not take this serious I just typed a word which would make sense after the previous with the context.) (edited)

I read the msg and I can't remember a single word

11:28AM

C best language tho

See this I just typed a word which would make sense after the previous with the context. worked

ghostGPT

It was I.

Most annoying thing (my opinion) with not C programming languages is that they don't actually fix the many problems C has but instead add many other and very annoying problems. So, in the end, C is still a very good option, which is sad, because something with less problems would worth it (if it actually has less problems).

wrong opinion

11:37AM

rust fixes the biggest issue with C, it prevents memory unsafety (not totally, but an improvement is better than nothing), it also helps preventing data races

11:37AM

u just dont want to accept progress

11:38AM

the only place where rust is less right now is: support for obscure platforms (which is not a problem of the language itself, but a problem of the compiler backend implementation), and compile times (atleast vs C which compiles fast iirc)

zogtib

Most annoying thing (my opinion) with not C programming languages is that they don't actually fix the many problems C has but instead add many other and very annoying problems. So, in the end, C is still a very good option, which is sad, because something with less problems would worth it (if it actually has less problems).

many lang does fix lol (edited)

rust is as fast as C

just you'll loose stuff such as performance

11:39AM

(gc etc...), but rust does fix C issues without loosing perf

Rust doesn't have the simplicity and the clarity C code can have (as I understand Rust by now) and that has great value for me. I'd better debug clearly-written programms instead of a big complicated stuff that its (only?) good is that is does not do invalid memory accesses and runs fast. (edited)

the clarity of C is a double edge

11:42AM

it hurts you more than not

@zogtib be honest, did u code rust for at least 6 months? xd

but by what ur saying it shows u rly havent tried rust

11:43AM

jupstar was the same

11:43AM

tbh Rust is hard when you start

i think

but after it's pure enjoyment

ReiTW

tbh Rust is hard when you start

C is hard too if u dont know C like langs

Ryozuki

jupstar was the same

i was always pro rust. and tbf i still think some things are annoying xd

Jupstar ✪

i was always pro rust. and tbf i still think some things are annoying xd

no u werent, i remember u not knowing rust

i'm less of a fanboy than u xd

xd

11:44AM

i also acknowledge annoying things in rust

11:44AM

and whish to be fixed

@Ryozuki send good rust books

1

Ryozuki

no u werent, i remember u not knowing rust

yes but i learned it, because i accepted that i should learn it xd

yes but C is very small and "easy", like there's ,ot much stuff to learn tho. Only the way devs will write their code will make things harder

Teero

@Ryozuki send good rust books

the oficial rust book https://doc.rust-lang.org/book/ atomics https://marabos.nl/atomics/ and https://www.oreilly.com/library/view/programming-rust-2nd/9781492052586/

The Rust Programming Language - The Rust Programming Language

Rust Atomics and Locks by Mara Bos

Low-level Concurrency in Practice. This practical book helps Rust programmers of all levels gain a clear understanding of low-level concurrency. You'll learn everything about atomics and …

Programming Rust, 2nd Edition

2

C is hot asf. Best language ever conceived

Jupstar ✪

Yes. In fact most innovative ppl are not poor

wdym by that?

It's proof of intelligent creation

Jupstar ✪

but maybe godot with rust bindings could be project to test for me

@Ryozuki xddd

Learath2

It's proof of intelligent creation

rust is, its proof of improvement without a sacrificy thought to be needed (runtime speed with gained safety)

Ryozuki

the oficial rust book https://doc.rust-lang.org/book/ atomics https://marabos.nl/atomics/ and https://www.oreilly.com/library/view/programming-rust-2nd/9781492052586/

ho lol the atomics book is definetly smth I need to read lmao

ReiTW

ho lol the atomics book is definetly smth I need to read lmao

read it, its one of the best books

ty

its online btw

yeah saw

Teero

Click to see attachment 🖼️

windows is largely written in C++ afaik

heinrich5991

wdym by that?

if you have a living standard that allows you to not care too much about money.. u can concentrate more about being innovative

Ryozuki

rust is, its proof of improvement without a sacrificy thought to be needed (runtime speed with gained safety)

You meant C, right?

Learath2

C is hot asf. Best language ever conceived

scratch

4

I remember that I was learning OpenGL using C, but then learned about Rust and that felt that it could do things better and "clearer", like "match" etc, but I got annoyed by its pursue to hide complexity by introducing more complexity.

zogtib

I remember that I was learning OpenGL using C, but then learned about Rust and that felt that it could do things better and "clearer", like "match" etc, but I got annoyed by its pursue to hide complexity by introducing more complexity.

It's the pursuit of every modern language. Extreme abstraction

i think rust's enums are one of the coolest things. they are a clean way to represent variants

11:47AM

match is at least as clean as a switch

Jupstar ✪

i think rust's enums are one of the coolest things. they are a clean way to represent variants

True. Rust enums are probably the best thing about rust after the way errors are handled

I mean I'd better write if (a > 0) { ... } else if (a < 0) instead of use std::signs; match signs::of(num) signs::Positive: ... signs::Negative: ...

in the context of enums actually cleaner, bcs you can only handle the data if you match the enum not like switch a.type {...}

zogtib

Rust doesn't have the simplicity and the clarity C code can have (as I understand Rust by now) and that has great value for me. I'd better debug clearly-written programms instead of a big complicated stuff that its (only?) good is that is does not do invalid memory accesses and runs fast. (edited)

C code is not "simple" for me. e.g. any addition can invoke UB

There isn't a single thing I hate more than exceptions in programming I think

zogtib

I mean I'd better write if (a > 0) { ... } else if (a < 0) instead of use std::signs; match signs::of(num) signs::Positive: ... signs::Negative: ...

but u can do that

11:49AM

i'd also not write the sign stuff xD

C is not simple if u know its a minefield

11:50AM

im never too sure about anything

Learath2

There isn't a single thing I hate more than exceptions in programming I think

so that's why we have -wno-exceptions in ddnet xdd

whereas i rust i am

Learath2

There isn't a single thing I hate more than exceptions in programming I think

wdym by exceptions? automatically propagated errors? errors not appearing the type signature?

Jupstar ✪

i'd also not write the sign stuff xD

When use is_odd and is_even crate?

Learath2

When use is_odd and is_even crate?

xdd

heinrich5991

C code is not "simple" for me. e.g. any addition can invoke UB

C does have its problems and they are not tiny. But alternatives seem worse to me.

honestly, you can write rust code that is pretty similar to c, u just have to accept using unsafe

Also Ada has safety features etc without overcomplicating everything.

Jupstar ✪

honestly, you can write rust code that is pretty similar to c, u just have to accept using unsafe

unsafe still has strong guarantee s

Btw what happened to carbon?

Ryozuki

unsafe still has strong guarantee s

pointers alone can easily destroy most of these (if you really want to).. not saying it is good

but of u dont know what u doing it can be a stronger minefield

heinrich5991

wdym by exceptions? automatically propagated errors? errors not appearing the type signature?

I hate both aspects. If you HAVE to have an exception system in your language it shouldn't have automatic propagation and you should be forced to enumerate all possible errors from every function

zogtib

Also Ada has safety features etc without overcomplicating everything.

it seems to me that rust isn't "overcomplicating" C, to me it's what has to happen inside my brain in order to write C, but it's not chhecked there

zogtib

Also Ada has safety features etc without overcomplicating everything.

does ada have protection against dangling pointers? afaik no

Exceptions lead to "Unhandled Exception"s. Always

Jupstar ✪

pointers alone can easily destroy most of these (if you really want to).. not saying it is good

yes but in unsafe rust code u need to uphold to more stuff than in C, you need to make sure safe code cant do ub

Teero

Btw what happened to carbon?

just announced so far. creator told us to use rust if we want a usable language right now

Learath2

I hate both aspects. If you HAVE to have an exception system in your language it shouldn't have automatic propagation and you should be forced to enumerate all possible errors from every function

so java's checked exceptions are actually a step up from C++'s unchecked ones? I tend to agree

Jupstar ✪

honestly, you can write rust code that is pretty similar to c, u just have to accept using unsafe

So I can sell simplicity for longer compile times? What would that worth? (edited)

Ryozuki

yes but in unsafe rust code u need to uphold to more stuff than in C, you need to make sure safe code cant do ub

my point is really. if u just wrap everything in unsafe, u can write code similar to c.. with some differences ofc. but it feels similary low level

zogtib

C does have its problems and they are not tiny. But alternatives seem worse to me.

Have you tried zig? It's not the worst

zogtib

So I can sell simplicity for longer compile times? What would that worth? (edited)

I don't think you'll get longer compile times if you write a better-C dialect of rust

11:53AM

the long compile times come from using the non-C features

what problem of C would that solve?

I've also enjoyed D, but haven't used for any significant time

zogtib

what problem of C would that solve?

Borrow checker might help

Learath2

Have you tried zig? It's not the worst

haven't actually tried it, but its syntex felt annoying (not a too serious problem tho)

zogtib

what problem of C would that solve?

the problem that no one seems to be able to write safe, nontrivial C programs

11:55AM

inb4 "I can"

@zogtib i generally agree that compile times are a problem of rust. But on the other side. since u have a really great ecosystem it's easy to add high quality libraries to your project, it works cross platform. that alone saves lot of time often

heinrich5991

inb4 "I can"

I can

Learath2

I can

x

in my current rust project i have not a single build script.. yet everything works

heinrich5991

the problem that no one seems to be able to write safe, nontrivial C programs

I can do just a printf("Hello world\n"); & consider it safe (edited)

ReiTW

I can do just a printf("Hello world\n"); & consider it safe (edited)

Nontrivial

ah

heinrich5991

the problem that no one seems to be able to write safe, nontrivial C programs

just think harder lul

Learath2

I can

let's add network interaction so we have a threat model?

11:57AM

or do you have an example in mind?

It's actually an unfalsifiable statement. If I wrote one it would be an exception rather than the rule

why unfalsifiable?

Well, C programs don't just magically work just because I write them, I think that the effort spent making them complete is worth it instead of overcomplicating things by adding a ton of other names etc.

1

zogtib

Well, C programs don't just magically work just because I write them, I think that the effort spent making them complete is worth it instead of overcomplicating things by adding a ton of other names etc.

but C programs seem to magically have security vulnerabilities

heinrich5991

why unfalsifiable?

Well your initial statement would be disproven since you said no one. But you'd quickly update it to "not many people" and everyone will clap and I'll have wasted a couple dozen hours of my life

@Learath2 tbh my biggest problem with exceptions is not using them together with RAII. that often leads to weird bugs, even if u handle exceptions. Java imo is a good example

11:59AM

c++ at least in this case is actually better

11:59AM

(and ofc it's hard to do it right, generally)

Jupstar ✪

@Learath2 tbh my biggest problem with exceptions is not using them together with RAII. that often leads to weird bugs, even if u handle exceptions. Java imo is a good example

The way raii and exceptions interact is also quite hard for a new learner. Overall very meh feature

Learath2

The way raii and exceptions interact is also quite hard for a new learner. Overall very meh feature

you also have that in rust, "panics"

heinrich5991

but C programs seem to magically have security vulnerabilities

One guy that worked on a Rust-written GNU-coreutils implementation said that GNU-coreutils has had just a few vulerenabilities in so many years and that they are not rewritting them in Rust for security reasons.

zogtib

One guy that worked on a Rust-written GNU-coreutils implementation said that GNU-coreutils has had just a few vulerenabilities in so many years and that they are not rewritting them in Rust for security reasons.

I think even glibc recently had an exploitable bug

zogtib

One guy that worked on a Rust-written GNU-coreutils implementation said that GNU-coreutils has had just a few vulerenabilities in so many years and that they are not rewritting them in Rust for security reasons.

The argument is always that some of those vulnerabilities couldn't have happened with rust

Rust also has had many vulerenabilities itself? not necessary memory-related

zogtib

Rust also has had many vulerenabilities itself? not necessary memory-related

The argument against that one is that Rust doesn't claim to eliminate all errors. It eliminates a couple classes of errors

zogtib

Rust also has had many vulerenabilities itself? not necessary memory-related

IME the bugs that rust reports are usually things that wouldn't even be classified as bugs in C

i'd say rust tries to force you to make a better design over your program. it e.g. doesn't prevent dead locks, but with a clear hierarchy you could evade most of these problems Memory safety is given in safe rust. that eliminates many problems. If you use Arc only with the said hierarchy you also should not fall into traps like memory leaks (edited)

12:03PM

overall rust at least helps you. it doesnt enforce everything that is needed to say it's the 100% safe alternative, but that's probably also not its goal

heinrich5991

IME the bugs that rust reports are usually things that wouldn't even be classified as bugs in C

see e.g.: https://blog.rust-lang.org/2021/11/01/cve-2021-42574.html https://blog.rust-lang.org/2022/01/20/cve-2022-21658.html

Security advisory for rustc (CVE-2021-42574) | Rust Blog

Empowering everyone to build reliable and efficient software.

Security advisory for the standard library (CVE-2022-21658) | Rust ...

Empowering everyone to build reliable and efficient software.

12:03PM

two links that appear from the ddg search for "rust cve"

12:04PM

both of these would probably be considered "working just fine" in C/C++

Personally, I'm not ready to sell C's simplicity for Rust's safety. I might be if I got fed up with C unsafety. But doesn't seem very likely to me now.

zogtib

Rust also has had many vulerenabilities itself? not necessary memory-related

give examples

zogtib

Personally, I'm not ready to sell C's simplicity for Rust's safety. I might be if I got fed up with C unsafety. But doesn't seem very likely to me now.

C isn't simple. it only looks simple before you know all the UB stuff

zogtib

Personally, I'm not ready to sell C's simplicity for Rust's safety. I might be if I got fed up with C unsafety. But doesn't seem very likely to me now.

i don't want to lie. alone the fact that rust has RAII and c not. already makes so much stuff simpler

12:05PM

i dont see what u mean with simplcity

12:05PM

in c u code everything, even stuff that i'd call intuitive

destructors are actually a thing that C really lacks

12:05PM

it'd be really nice if you could just defer free(a); in C

12:05PM

like in go

12:06PM

I think there's a gcc extension for this

heinrich5991

C isn't simple. it only looks simple before you know all the UB stuff

C is absolutely trivial compared to Rust. The UB stuff is unexpected but C just doesn't have the depth to compare to Rusts complexity

12:06PM

You can't have a box pin dyn send sync

Learath2

C is absolutely trivial compared to Rust. The UB stuff is unexpected but C just doesn't have the depth to compare to Rusts complexity

it's not simple to program in, and it doesn't have a simple model for your mind

12:06PM

libtw2 also doesn't have that

12:07PM

glib, however, has self-built vtables

12:07PM

which I'd consider less simple than compiler support

heinrich5991

give examples

I prob remember some CVE's like you sent. I remember something about directory recursive deletion but maybe we'd just forget it for the current conversation, not a big deal even if rust had no CVEs

zogtib

I prob remember some CVE's like you sent. I remember something about directory recursive deletion but maybe we'd just forget it for the current conversation, not a big deal even if rust had no CVEs

the directory deletion thing would be considered "no bug" in C/C++

12:08PM

the C++ std says that concurrent access to the filesystem is UB

12:08PM

completely insane

heinrich5991

which I'd consider less simple than compiler support

I mean, is it really? It's less simple when programming but it's actually far more intuitive. Vtables are a very simple thing hidden behind lots of magic

Learath2

I mean, is it really? It's less simple when programming but it's actually far more intuitive. Vtables are a very simple thing hidden behind lots of magic

yes. just like having functions is more simple than having to invent your own calling conventions in x86 asm, e.g.

heinrich5991

it's not simple to program in, and it doesn't have a simple model for your mind

It builds generally with a simple way upon what it has. So using C you get access to what C has with a "direct way".

or stack variables are easier than manual stack management, structs are easier than only having access to primitive types

12:10PM

C's idea of strings is also very bad, and not very performant

1

for example, C gives you access to the filesystem (with libc let's say) but doesn't actually care what it does on higher level.

heinrich5991

it's not simple to program in, and it doesn't have a simple model for your mind

Actually idk how I let this one sneak by, just because you don't program like that doesn't mean no one does. You'll absolutely see a pin box dyn send sync hashmap box box pin

zogtib

for example, C gives you access to the filesystem (with libc let's say) but doesn't actually care what it does on higher level.

not sure what you mean by this. rust gives you access to the file system in the same way

Learath2

Actually idk how I let this one sneak by, just because you don't program like that doesn't mean no one does. You'll absolutely see a pin box dyn send sync hashmap box box pin

I've been able to avoid it (in non-work related settings)

‍♀️

12:12PM

hmm, maybe not

12:12PM

the mastersrv

heinrich5991

it's not simple to program in, and it doesn't have a simple model for your mind

Also this I disagree with. You could say that the abstract machine has very little to do with the modern machines but it is an absolutely trivial machine

I disagree. the rules around integer overflow, type promotion, pointer validity rules, all disagree with the simple model

zogtib

for example, C gives you access to the filesystem (with libc let's say) but doesn't actually care what it does on higher level.

I mean that it's relation to the OS/filesystem is simple and clear, but the programmer has to take care of correctness. I'm not ready to accept much complexity so that I don't fall into incorrect behaviours.

note that C's standard library actually buffers your files by default

12:15PM

so you're not even close to the OS there

I was thinking about systems calls rather than library calls.

system calls have little to do with C

12:16PM

you can do system calls from all sorts of languages, including C, rust, …

I did say libc but i was thinking about making available the open(2) function e.g.

which is then implemented as openat

All languages have those available, C's wrappers are just very thin usually

doesn't sound very transparent

12:17PM

you can absolutely make these syscalls from rust, too. I participated in a rust stdlib rewrite using the syscalls directly

12:17PM

was pretty fun

heinrich5991

doesn't sound very transparent

Using open(2)'s model is enough for many cases. Knowing all the details isn't something very practical.

heinrich5991

you can absolutely make these syscalls from rust, too. I participated in a rust stdlib rewrite using the syscalls directly

Using them from C seems more enjoyable.

zogtib

Using open(2)'s model is enough for many cases. Knowing all the details isn't something very practical.

We just seem to disagree about the level of abstraction. I like my wrappers thin. Rust people like them thick and featureful

C is a bit preferred here, not because it's a good language, but because it was chosen (maybe implicitly) to be the model for POSIX interactions

Learath2

We just seem to disagree about the level of abstraction. I like my wrappers thin. Rust people like them thick and featureful

rust's File is less of an abstraction than C's FILE *

12:20PM

but it actually works well on windows and linux ^^

I prefer C over Rust, it's not that Rust is useless compared to C (Rust solves many problems C has) but I don't like Rust's solution to those problems.

Learath2

We just seem to disagree about the level of abstraction. I like my wrappers thin. Rust people like them thick and featureful

I think open(2) existed already because of historical reasons?, so even if under the hood things changed, open(2) is still useful.

open(2) still exists

But even if it isn't, openat(2) is available.

the syscall open still exists

12:23PM

it's just not called by the glibc function open anymore

heinrich5991

rust's File is less of an abstraction than C's FILE *

I would be very very very surprised if it's not about the same level given they provide similar features

12:24PM

The only way I can imagine it being a lower level abstraction is cheating like msvcrt

Learath2

I would be very very very surprised if it's not about the same level given they provide similar features

C's FILE * has buffering

12:24PM

rust's File is just a file descriptor

If what open(2)'s manual fit's the programmer needs, then what it does in detail is not important imo.

(rust's File is literally 4 byte in size on linux)

zogtib

If what open(2)'s manual fit's the programmer needs, then what it does in detail is not important imo.

it's not just a thin wrapper over the open syscall then, though

12:25PM

it's about as thin as the File::open function in rust

heinrich5991

C's FILE * has buffering

As far as the standard is concerned what is buffering even?

idk, you're the standard person

12:26PM

but it even has functions for adjusting the buffer

12:26PM

in the standard

FILE could be whatever the implementor wants. Msvcrt uses that to implement it through a fake fileno table

fact is, FILE * is buffered in all libcs that I know

12:27PM

do you know any where it's not?

12:27PM

musl, glibc, msvcrt all buffer AFAIK

wtf r u guys all even doin with filesystems xddd i need open file, write file maybe some helper functions like list directory

12:27PM

r u like coding the next sql database xdd

Jupstar ✪

wtf r u guys all even doin with filesystems xddd i need open file, write file maybe some helper functions like list directory

list directory is not available in C

tru

12:28PM

times before c++17 were hard

Besides the system-specific information necessary to access the device (e.g., a POSIX file descriptor), each FILE object directly or indirectly holds the following: 1. (C95) Character width: unset, narrow, or wide. 2. (C95) Parse state for conversions between multibyte and wide characters (an object of type mbstate_t) 3. Buffering state: unbuffered, line-buffered, fully buffered. 4. The buffer, which may be replaced by an external, user-provided buffer. 5. I/O mode: input, output, or update (both input and output). 6. Binary/text mode indicator. 7. End-of-file status indicator. 8. Error status indicator. 9. File position indicator, accessible as an object of type fpos_t, which, for wide streams, includes parse state. 10. (C11) Reentrant lock used to prevent data races when multiple threads read, write, position, or query the position of a stream.

12:29PM

https://en.cppreference.com/w/c/io/FILE

c/io/FILE

heinrich5991

Besides the system-specific information necessary to access the device (e.g., a POSIX file descriptor), each FILE object directly or indirectly holds the following: 1. (C95) Character width: unset, narrow, or wide. 2. (C95) Parse state for conversions between multibyte and wide characters (an object of type mbstate_t) 3. Buffering state: unbuffered, line-buffered, fully buffered. 4. The buffer, which may be replaced by an external, user-provided buffer. 5. I/O mode: input, output, or update (both input and output). 6. Binary/text mode indicator. 7. End-of-file status indicator. 8. Error status indicator. 9. File position indicator, accessible as an object of type fpos_t, which, for wide streams, includes parse state. 10. (C11) Reentrant lock used to prevent data races when multiple threads read, write, position, or query the position of a stream.

Where does rust shove these? I have a feeling we are not comparing equivalent abstractions

rust's file has none of 1-10

12:29PM

you can wrap a file in a BufReader

i could imagine that, if you code smth where low level filesystem access is important (even in rust) you might just use system related operations anyway (win32 on windows etc.)

(you probably should, if you're not just reading the complete file)

Jupstar ✪

i could imagine that, if you code smth where low level filesystem access is important (even in rust) you might just use system related operations anyway (win32 on windows etc.)

no need in rust, because it's just a thin wrapper here

heinrich5991

no need in rust, because it's just a thin wrapper here

yeah but like permission stuff and so on

12:30PM

or does all this exist as wrapper?

Ok, so these are not the same abstractions. C doesn't have a similar thin wrapper for files

Learath2

Where does rust shove these? I have a feeling we are not comparing equivalent abstractions

it doesn't do 1,2. 3,4: buffering can be done by wrapping it in a BufReader, 5,6,7,8,9 is not done in rust, 10 is guaranteed by rust's guarantees

Learath2

Ok, so these are not the same abstractions. C doesn't have a similar thin wrapper for files

correct

12:32PM

so rust has a thinner abstraction around files than C does

You mean than C libc does.

man alone not having unicode support must suck in c

12:32PM

u HAVE to use win32 api

12:32PM

xd

zogtib

You mean than C libc does.

I mean C. you might be talking about POSIX, C on linux or libraries (edited)

heinrich5991

so rust has a thinner abstraction around files than C does

It's a little disingenuous when no one would use those

no one uses FILE *?

No File you almost always want a bufreader or bufwriter

yes, but that's simpler, no?

12:34PM

you don't put all the abstractions into one big file struct

It exists but almost everyone will be using the fatter abstraction

even with BufReader, it's still thinner than C's stuff

heinrich5991

even with BufReader, it's still thinner than C's stuff

Through having a fatter abstraction for strings iirc

in what way?

12:35PM

you mean having one at all?

Learath2

Through having a fatter abstraction for strings iirc

fater abstraction = more code? or fater abstraction = less performance?

12:36PM

i kinda have the feeling this isnt clear in this conversation

I think we agreed on the fact that C and rust code is in the same league for speed

heinrich5991

in what way?

C strings are nothing but a bunch of bytes. That's why we need to keep track of the wchar stuff

Learath2

C strings are nothing but a bunch of bytes. That's why we need to keep track of the wchar stuff

no, C strings are nul-terminated char arrays

rust strings can be faster, null terminator was a bad decision

ok then fat as in more code for the user to write? or fat as in the standard is more fat?

char * can refer to a pointer to a char or to such a C string, or to a pointer to a char array, even (edited)

heinrich5991

no, C strings are nul-terminated char arrays

Since when are null terminated char arrays are not a bunch of bytes?

it just doesn't say

Learath2

Since when are null terminated char arrays are not a bunch of bytes?

because they carry an additional guarantee

Ok, done for the day

1

pack it up we done

heinrich5991

even with BufReader, it's still thinner than C's stuff

All details are not important all of the time. Also C standards have long history so they can't just appear perfect, so I don't mean "just the standarised C" as C.

Learath2

Ok, done for the day

lmao, don't be sad over programming languages. if you like c so much use it. it's your life and your time and if u like it so much do it

Thanks for sharing your opinions and knowledge with me, have nice days. Cheers

ye its just a discussion

12:39PM

end of the day each uses the lang they want

no, don't use python

3

<flamebait>but people writing critical software hopefully don't start their new projects in C</flamebait> (edited)

they use ada

Jupstar ✪

lmao, don't be sad over programming languages. if you like c so much use it. it's your life and your time and if u like it so much do it

I don't care about your opinion of C, nor do I care about @heinrich5991's opinion on C. I'm just annoyed at the disingenuous bad faith discussion techniques employed

heinrich5991

<flamebait>but people writing critical software hopefully don't start their new projects in C</flamebait> (edited)

programmer tone tags

at the end of the day, i don't want to sit in a plane that panics, nor runs on UB so just proof the software or smth

Learath2

I don't care about your opinion of C, nor do I care about @heinrich5991's opinion on C. I'm just annoyed at the disingenuous bad faith discussion techniques employed

I'm sorry. maybe we can come together some other time

Learath2

I don't care about your opinion of C, nor do I care about @heinrich5991's opinion on C. I'm just annoyed at the disingenuous bad faith discussion techniques employed

i think u call disingenuous any argument not in ur favour tho

Jupstar ✪

at the end of the day, i don't want to sit in a plane that panics, nor runs on UB so just proof the software or smth

UB as in go to your destination. Go to Heaven!

heinrich5991

<flamebait>but people writing critical software hopefully don't start their new projects in C</flamebait> (edited)

OpenBSD say they are serious about security and they use C, I'm just letting you know.

1

heinrich5991

<flamebait>but people writing critical software hopefully don't start their new projects in C</flamebait> (edited)

ppl use /s for sarcasm btw

zogtib

OpenBSD say they are serious about security and they use C, I'm just letting you know.

They don't know shit

it wasn't sarcasm

12:43PM

it was flamebait

sarcasm is often used for that

12:43PM

anyway

Ryozuki

ppl use /s for sarcasm btw

yeah but tonetags suck and most are infantilizing

but he was serious af about it

Ryozuki

i think u call disingenuous any argument not in ur favour tho

Yes

my take is, ill take rust complexity that maintains same speed as C but with improved safety

12:44PM

and i enjoy it

Learath2

I don't care about your opinion of C, nor do I care about @heinrich5991's opinion on C. I'm just annoyed at the disingenuous bad faith discussion techniques employed

i don't have an opinion about it, because i don't use it as a daily driver. But i know concepts i use, and that i like that c is missing.

but C simplicity is often just at face value, digging deeper it becomws complex because you need to be aware and careful of lot of stuff, ur cognitivie load comes later, when u need to fix those bugs, or when ur thinking how to do X safely

12:46PM

in rust the cognitive load comes at compile time, at ur initial write

Ryozuki

but C simplicity is often just at face value, digging deeper it becomws complex because you need to be aware and careful of lot of stuff, ur cognitivie load comes later, when u need to fix those bugs, or when ur thinking how to do X safely

You actually need to know how make things, right? (edited)

It's like talking about how calculus is a thin abstraction over zermelo-fraenkel set theory. Just because thin layers exist along the way

altho u can avoid the load initially by using simpler stuff but less fast

12:46PM

which is how u protoptype

Ryozuki

altho u can avoid the load initially by using simpler stuff but less fast

in fact sometimes u just accept it, because it would bloat software a lot

zogtib

You actually need to know how make things, right? (edited)

there is a difference between knowimg how to make things and how to make things correctly and safe

well remove the "sometimes" always

12:47PM

in the end c is high level compared to asm xD

i wonder what rust feature will benefit the ddnet code if it ever replaced the server code like, i imagine rust has some unique function that c doesnt

Voxel

i wonder what rust feature will benefit the ddnet code if it ever replaced the server code like, i imagine rust has some unique function that c doesnt

it wont make ddnet faster

12:48PM

it may avoid segfaults

12:48PM

altho a rewrite maybe improves perf by encountering optimization oportunities

Voxel

i wonder what rust feature will benefit the ddnet code if it ever replaced the server code like, i imagine rust has some unique function that c doesnt

i'd say derive macros and struct enums

Ryozuki

it wont make ddnet faster

i dont care about speed in this arguement LOL

?

Jupstar ✪

i'd say derive macros and struct enums

oo

a new lang wont bring new ingame features

Saying C has the level of complexity of Rust is like saying a casio watch calculator is comparable to a modern octocore cpu

Learath2

Saying C has the level of complexity of Rust is like saying a casio watch calculator is comparable to a modern octocore cpu

C has delayed complexity

Ryozuki

a new lang wont bring new ingame features

but it has different answers to the same problem ^^

u still need to think about what rust solves for u

and these answers might be more fitting

Jupstar ✪

but it has different answers to the same problem ^^

maybe iterators improve stuff

Ryozuki

u still need to think about what rust solves for u

I'm not saying rust doesnt solve problems. I'm saying it doesnt solve them at no complexity cost

yeah sure, but complexity is abstract

Ryozuki

maybe iterators improve stuff

the fact that i can easily serialize data types for example is great compared to c (which has no concept of typeids) and even compared to c++, where it's probably still a mess

for me who i care about writimt correct safe software its hard im C

12:51PM

because there is lot of unknowns

12:51PM

the fix is to read the bible

12:51PM

learn the standard

12:51PM

fck mobile

Jupstar ✪

the fact that i can easily serialize data types for example is great compared to c (which has no concept of typeids) and even compared to c++, where it's probably still a mess

Now open up serde docs and observe where the complexity is. Just because very smart people manage to hide it well doesnt mean rust is a very complex language

Learath2

Now open up serde docs and observe where the complexity is. Just because very smart people manage to hide it well doesnt mean rust is a very complex language

why do u need serde docs?

Ryozuki

the fix is to read the bible

i didnt read the c++ bible when i dived head first into ddnet code LOL

let's assume u use all fields you serialize (so no skip attribute or smth)

Learath2

Now open up serde docs and observe where the complexity is. Just because very smart people manage to hide it well doesnt mean rust is a very complex language

see thats a bad faith arg

12:52PM

its solving a complex problem about giving u a simple interface to seriakize stuff

for network stuff this makes sense for example

Ryozuki

its solving a complex problem about giving u a simple interface to seriakize stuff

Using extremely complex language constructs that equivalents of don't even exist in C

yeah cuz u need to make them urself in c

but @Learath2 what would u do in c?

Ryozuki

yeah cuz u need to make them urself in c

You can't even make them in C, that sort of access to AST through macros don't exist

u'd need to manually verify every single data chunk is correct size (size > my struct) then proof all fields individually etc.

well too bad

12:54PM

i prefer that vs textual macros

that is complex IMHO

also headers are complex

Ryozuki

i prefer that vs textual macros

The discussion is about language complexity, not your preferences. I also prefer just automagically having my stuff serialized

From my perspective: Rust's solutions (and Rust itself) are way more complex than a solution should be, and I find that as a big and primary problem.

zogtib

From my perspective: Rust's solutions (and Rust itself) are way more complex than a solution should be, and I find that as a big and primary problem.

i dont agree

Learath2

The discussion is about language complexity, not your preferences. I also prefer just automagically having my stuff serialized

If that is the measure of complexity then the best language is js

the problem rust tries to solve is rooy complex

12:55PM

if u know a simple solution u would get a Nobel prize

Learath2

The discussion is about language complexity, not your preferences. I also prefer just automagically having my stuff serialized

yes in that sense you are right. it's more complex behind the scnenes.. but therefore less complex in the scene itself

zogtib

From my perspective: Rust's solutions (and Rust itself) are way more complex than a solution should be, and I find that as a big and primary problem.

ive seen some solutions that are beautiful

less code = less potential for bugs

12:56PM

if serde is used widely, it might™️ be more tested

wdym guys, rust is not complex at all

than anthing u ever write by yourself

rust is complex but not complex

I guess I look at complexity of a language very differently than you seem to do. It's almost a linear function of the amount of features for me

Voxel

rust is complex but not complex

It moves complexity

12:57PM

Imo intuitively is more important

12:57PM

And serde is intuitive

Jupstar ✪

It moves complexity

What is the complexity it uses to move that complexity though?

In it's prime form

zogtib

What is the complexity it uses to move that complexity though?

Derive macros

when i inevitably make a file generator ill use rust so my own code doesnt trash my own pc

Basically a mini compiler

12:59PM

U know the attributes of the structure etc

Jupstar ✪

Basically a mini compiler

Even more powerful. Procmacros have access to the compiler itself, not a separate precompiler

Learath2

Even more powerful. Procmacros have access to the compiler itself, not a separate precompiler

Yes it's a fully working program

12:59PM

U can do anything u want

It's almost like monkeypatching the compiler itself

But let's be fair. In this case u wouldn't need to do this

Learath2

It's almost like monkeypatching the compiler itself

Yep

Jupstar ✪

but @Learath2 what would u do in c?

Nothing really except for some hobby projects. I see coding in C as mostly an artistic endeavour. Not much practical value in it

we all agree in one thing tho

1:13PM

linux is better than windows

1:13PM

And voxel took that personally

Ryozuki

linux is better than windows

better in what way? ^^

1:14PM

I like it better to work

<ChillerDragon> gnome looks nicer than windows ui

Next explosive opened xd

being open source is already a killer

<ChillerDragon> yes that too

1:24PM

<ChillerDragon> oh and free

1:24PM

<ChillerDragon> lightweight

Ryozuki

linux is better than windows

Can't even use directx12. Weak os

It can xd

Jupstar ✪

It can xd

Needs translator, not even native. Weak

Bloat

1:50PM

The abstraction could do smth similar to dxvk and allow to compile ur program directly to Vulkan using dx headers

1:52PM

Time to invent royalty free Direct 4D 12 To convince the fan boys xd

vulkan better

Microsoft best ❤️

1:54PM

@Jupstar ✪ when ddnet on dx?

Learath2

@Jupstar ✪ when ddnet on dx?

U can always use it

2:06PM

Microsoft contributed a Vulkan to dx12 translator to Mesa xd

tfw you stumble upon your own thread about a problem feelsbadman

Mr.Gh0s7

tfw you stumble upon your own thread about a problem feelsbadman

Tfw it only has one reply, it's wrong ohno

4

https://github.com/darthdeus/comfy

GitHub - darthdeus/comfy: Comfy is a fun 2D game engine built in Ru...

Comfy is a fun 2D game engine built in Rust. It's designed to be opinionated, productive, and easy to use. - GitHub - darthdeus/comfy: Comfy is a fun 2D game engine built in Rust. It&#3...

@Jupstar ✪ didn't think I'd spark a 800 message convo xdd

ah true, all your fault

3:14PM

xd

@Ryozuki https://www.phoronix.com/review/amd-ryzen-threadripper-7000

AMD Launches The Ryzen Threadripper 7000 Series: Up To 96 Cores, DD...

AMD today announced the Ryzen Threadripper PRO 7000 WX Series as the company's long-awaited Threadripper update that brings up to 96 Zen 4 cores, RDIMM memory is now required for Threadripper platforms moving forward, and catering to both HEDT enthusiasts and professionals.

4:08PM

96 cores

4:08PM

can u buy me one

4:09PM

AMD Ryzen Threadripper PRO 7995WX: $9,999

4:09PM

wow, really cheap... not xd

xd

Jupstar ✪

@Ryozuki https://www.phoronix.com/review/amd-ryzen-threadripper-7000

mee tooo

4:25PM

cries in Intel Xeon E5420 (8) @ 2.499GHz

Learath2

Tfw it only has one reply, it's wrong ohno

@Voxel

tf

are you red or blue ? https://www.reddit.com/r/ProgrammerHumor/comments/17bhp01/whichsideistheoneyougotusedto/

6

1

depends on the function

6:24PM

if u have early returns without any logic, u might want to not even call that function at all in first place

how to make a freeze bar was only seen by allies on ddnet-skeleton server? (edited)

7:03PM

or replace freeze bar to stars I'm trying to make a fng mod (edited)

we have timeout ppl can we fast get code? or he will disconnect swap 2 hours run 'viny ᵐᶦᵒ' would have timed out, but can use timeout protection now 45.141.57.22:8332

45.141.57.22:8332 is an official DDNet server. <steam://run/412220//45.141.57.22:8332>/

triki

we have timeout ppl can we fast get code? or he will disconnect swap 2 hours run 'viny ᵐᶦᵒ' would have timed out, but can use timeout protection now 45.141.57.22:8332

nobody has any code. if u have him in discord ask there

Jupstar ✪

nobody has any code. if u have him in discord ask there

he timeout

7:11PM

its mean we cant contact with him

the only thing mods could potentially do is increase the timeout time

Jupstar ✪

the only thing mods could potentially do is increase the timeout time

nobody have taht timeout code?

7:12PM

admins cant see that?

7:12PM

server know this code

i am not aware of that at least

7:12PM

server knows it yes

sad

7:13PM

he dropped

7:13PM

near finish

unlucky

trikiman

[ddnet/ddnet] Issue opened: #7357 public timeoutcode (swap issue)

if someone timeout and dont reconnect in t0 give acess to swap with near tees(ip) its mean alive can join dummy and get swap with timeout. it is save some race where u race 2 hours and someone just timeouted.

what do i need to do with the binary of clang on my debian? I wanna install clang 10 and this seems to be the only way to install it. do I need to unzip it or something? https://github.com/muttleyxd/clang-tools-static-binaries/releases

Releases · muttleyxd/clang-tools-static-binaries

static-linked versions of clang tools for Linux, Mac OS X and Windows, available in AUR as clang-format-static-bin - muttleyxd/clang-tools-static-binaries

why do u want to use clang 10?

7:42PM

ddnet compiles just fine with newer ones

7:42PM

ah u mean clang-format

7:43PM

well depends:

do you use some kind of IDE?
do you just want to fix style?

7:44PM

the easiest is probably to copy it to /usr/local/bin/

7:44PM

that should be in your path i guess

well yeah I just wanna fix the style

how to make a freeze bar was only seen by allies on ddnet-skeleton server? or replace freeze bar to stars I'm trying to make a fng mod

Jupstar ✪

the easiest is probably to copy it to /usr/local/bin/

will try that, I guess I will have to rename it?

He11o, wor1d

how to make a freeze bar was only seen by allies on ddnet-skeleton server? or replace freeze bar to stars I'm trying to make a fng mod

don't send them to the users that should not see them

toneey

will try that, I guess I will have to rename it?

i'd call it clang-format-10

alright thank you.

Jupstar ✪

don't send them to the users that should not see them

i am dont know how this make (sorry for my english)

toneey

alright thank you.

fix_style.py also tries to find version 10

He11o, wor1d

i am dont know how this make (sorry for my english)

mh, maybe start with something easier then

Jupstar ✪

mh, maybe start with something easier then

why?

He11o, wor1d

why?

to get a feeling for how the source code works

Jupstar ✪

mh, maybe start with something easier then

yeah that's what I noticed, I don't have it installed and it seems that there is no actual package for v10 (edited)

Jupstar ✪

mh, maybe start with something easier then

where do we start?

yeah sadly :/

He11o, wor1d

where do we start?

i dunno, try edit some source code you think you understand

7:48PM

e.g. server messages, or changing player name/skin whatever

7:48PM

try to place stars into the map

Jupstar ✪

try to place stars into the map

can you give me a hint on how to do this?)

He11o, wor1d

can you give me a hint on how to do this?)

do you know c++?

on elementary level

7:50PM

or better yet, a hint on how to send the freeze bar only to allies)

i'd try to read through https://github.com/ddnet/ddnet/blob/dad2c14abb3b89385fc2b8d1d1becd43206bad00/src/game/server/gamecontext.cpp

ddnet/src/game/server/gamecontext.cpp at dad2c14abb3b89385fc2b8d1d1...

DDraceNetwork, a free cooperative platformer game. Contribute to ddnet/ddnet development by creating an account on GitHub.

7:50PM

that is the most important file

7:51PM

i dunno how that file is called in ddnet-skeleton

7:51PM

i never used that

это подсказка для стоп-бара?)

7:51PM

sry

7:51PM

is this a tip for a freeze bar?

7:52PM

or place start into the map?

CreateDamageInd

7:52PM

creates stars

Jupstar ✪

creates stars

ok, but how to send a freeze bar only to allies?

how should i answer this question without implementing it myself? xD

7:55PM

i'd try to see what network package contains information about the current freeze time

7:55PM

and only sent it to allies

Jupstar ✪

how should i answer this question without implementing it myself? xD

Well, I would at least like the name of the function that sends the freeze bar to players)

https://github.com/ddnet/ddnet/blob/dad2c14abb3b89385fc2b8d1d1becd43206bad00/src/game/server/entities/character.cpp#L1228

ddnet/src/game/server/entities/character.cpp at dad2c14abb3b89385fc...

DDraceNetwork, a free cooperative platformer game. Contribute to ddnet/ddnet development by creating an account on GitHub.

8:00PM

this seems freeze related

8:00PM

for the other team u could simply set the skin to ninja skin or smth

8:00PM

i dunno if u can cleanly do what u want for FNG

8:00PM

the codebase favours ddrace

The freeze bars are not sent to the players directly. The server sends m_FreezeStart and m_FreezeEnd for each tee (character)

8:01PM

If you don't want the freezebars, only send 0 or -1 for both values I guess

8:02PM

Then send the damage indicator stars manually

8:03PM

Also with regards to ddnet-skeleton: https://discord.com/channels/252358080522747904/293493549758939136/1159155737738621051

nuh uh dont remove stars from fng

why is TeamMask() the same for both teams? how can this be fixed?

Robyt3

[ddnet/ddnet] Pull request opened: #7358 Validate ticks when reading demo chunk headers, check for all file errors in demo player, show demo error popup

See commit messages. The demo linked in #7349 is now considered invalid and fails to load. Demos which were simply truncated due to the client crashing during the recording can usually still be opened but the playback will be stopped entirely when an error is encountered reading the demo until the end.

Checklist

[X] Tested the change ingame
[ ] Provided screenshots if it is a visual change
[ ] Tested in combination with possibly related configuration options
[ ] Writ...

3 - this TeamMask() player 1 0 - this Team() player 1 3 - this TeamMask() player 2 0 - this Team() palyer 2 player1 - Red Team player2 - Blue Team

He11o, wor1d

why is TeamMask() the same for both teams? how can this be fixed?

I assume because ddnet-skeleton adds back the original red and blue teams without integrating DDNet team system with this properly

how this fix?

Don't know, maybe use separate TeamMask for both teams

8:26PM

Change TeamMask() to TeamMask(int Team) so you can get the mask for each team separately (and adjust the rest of the code accordingly)

a how to change int Team for blue team?

There are constants in the code for the teams

8:28PM

I'd recommend starting with something easier then trying to fix ddnet-skeleton

1

I too tried to make a mod with ddnet-skeleton but quickly forgot about it :/

github-merge-queue[bot]

[ddnet:master] 5 new commits

60807b5 Fix io_skip return type and documentation - Robyt3 6951795 Add title to engine warnings and make auto-hiding optional - Robyt3 92e2e17 Validate ticks when reading demo chunk headers - Robyt3 0e4f174 Check for all file errors in demo player, show demo error popup - Robyt3 5987d8d Merge pull request #7358 from Robyt3/Engine-Demo-Validation - def-

10:09PM

def-

[ddnet/ddnet] Pull request closed: #7358 Validate ticks when reading demo chunk headers, check for all file errors in demo player, show demo error popup