DDraceNetwork - developer

რილია

is there anything to be considered if im going to use unsigned char ? why they made the returned value unsigned char in the first place ?

i've always heard that unsigned is better for raw data and signed is better for text

03:10

in practice there's very little difference

<ChillerDragon> @Jupstar ✪ wtf where did you get that name from xd

Depends on the shell. Thats one of the portability hacks. Tbh I usually avoid sed -i in my scripts because its so annoying. Creating a temporary file is much more portable. So the issue is that macOS sed -i refuses to work without backup file. But the backup file is not supported by linux. So you need basically a if statement saying. If mac do backup else do nothing. And the -i'' hack does that in some cases I still think it fails on most macs but at least it should not break linux.

Replying to @heinrich5991 that should be precisely equivalent to `-i`

05:12

https://unix.stackexchange.com/a/92907/429645

How can I achieve portability with sed -i (in-place editing)?

I'm writing shell scripts for my server, which is a shared hosting running FreeBSD. I also want to be able to test them locally, on my PC running Linux. Hence, I'm trying to write them in a portabl...

05:17

i use clangd. Go send your compiler flags

Replying to @meloƞ using clangd You Just define the clangd path in your Compiler Flags for …

05:20

I use mason. But ddnet does not work out of the box. I assume because of the generated code. This has to be a solved problem by one of you i just cba to reinvent the wheel here

Replying to @MilkeeyCat Or just use Mason and everything just works

05:20

Replying to @MilkeeyCat Or just use Mason and everything just works

ddnet-maps

[ddnet-maps:master] 1 new commit

0b11235 M Cavings I - ddnet-maps

chillerdragon: https://github.com/MilkeeyCat/ddnet_demo_web_viewer/assets/40633857/6211920f-c5e8-4bee-a15e-6455d5437cd1 (edited)

06:09

all i had to do is add -DCMAKE_EXPORT_COMPILE_COMMANDS=1 to cmake command

@Jupstar ✪ webgl doesn't support geometry shaders, a gg

Switch to webgpu. But maybe they also don't support it xd

def-

[ddnet/ddnet] Issue opened: #8069 undefined reference to `curl_easy_getinfo'

http.cpp:(.text+0x4fc): undefined reference to `curl_easy_getinfo'
/usr/bin/ld: CMakeFiles/engine-shared.dir/src/engine/shared/http.cpp.o: in function `CHttpRequest::OnCompletionInternal(void*, unsigned int)':
http.cpp:(.text+0x16b8): undefined reference to `curl_easy_getinfo'

I guess this comes from https://github.com/ddnet/ddnet/pull/8063

08:20

def-

[ddnet-libs:master] 1 new commit

b387f0c Add curl_easy_getinfo to fake curl - heinrich5991

08:20

def-

[ddnet/ddnet] Issue closed: #8069 undefined reference to `curl_easy_getinfo'

heinrich5991

I find it hard to find information on this, but I'm trying

Hm, I definitely remember fraunhofer holding mp3 patents. I wonder how that happened with EU not doing software patents to begin with

MilkeeyCat

all i had to do is add -DCMAKE_EXPORT_COMPILE_COMMANDS=1 to cmake command

Then symlink the generated file to the main folder

i dont have to symlink anything :p

Hmm. I had to do it for nvim clangd

MilkeeyCat

all i had to do is add -DCMAKE_EXPORT_COMPILE_COMMANDS=1 to cmake command

This ^ or cxxcompiler Something ( im at Work and cant Check atm, i Made it an alias so i dont remember lmao )

@heinrich5991 it seems they really did hold patents to it. I guess these are not "software patents" for some reason. EP0393526A3

Ewan

i've always heard that unsigned is better for raw data and signed is better for text

note that signed char, char and unsigned char are three different types

10:07

whether char has a sign or not depends on the platform, but AFAIK independent of that, it's a type different from signed char and unsigned char

Ewan

i've always heard that unsigned is better for raw data and signed is better for text

in that manner, I think you should use unsigned char or uint8_t for bytes, and char for text

chillerdragon

Depends on the shell. Thats one of the portability hacks. Tbh I usually avoid sed -i in my scripts because its so annoying. Creating a temporary file is much more portable. So the issue is that macOS sed -i refuses to work without backup file. But the backup file is not supported by linux. So you need basically a if statement saying. If mac do backup else do nothing. And the -i'' hack does that in some cases I still think it fails on most macs but at least it should not break linux.

name me one shell where -i and -i'' isn't exactly the same

10:09

chillerdragon (edited)

Learath2

@heinrich5991 it seems they really did hold patents to it. I guess these are not "software patents" for some reason. EP0393526A3

I have no clue why this wouldn't be a software (or math) patent

10:12

there are no math patents either in europe AFAIK (edited)

heinrich5991

[ddnet-libs:master] 1 new commit

a4fbe0e Add curl_easy_getinfo to fake curl - heinrich5991

heinrich5991

[ddnet/ddnet-libs] New branch created: pr_cm

heinrich5991

I have no clue why this wouldn't be a software (or math) patent

I think(tm), the patent does not represent itself as a computer software patent, or a math patent. It's just the right amount of both to be considered neither. As in the math and the software are a means to an end, rather than the things getting patented

10:44

Article 52 is the piece of law that grants these exceptions to patentability and it is worded in such a bizarre way with that "as such" at the end that it's hard for me to interpret

has the mp3 patent been tested in court?

10:46

I've only found settlements wrt. that patent yesterday

It's for some reason extremely hard to google for this stuff. But I'm guessing not

it might be related to the fact that google tries to prioritize new results

Also this is like the core patent. There are like a dozen more around this that patent more specific parts of mp3. Maybe they litigated one of those somewhere

morning

Ryozuki

morning

Morning

MlirAttribute mlirLLVMDISubroutineTypeAttrGet(MlirContext ctx,
                                              unsigned int callingConvention,
                                              intptr_t nTypes,
                                              MlirAttribute const *types) {
  SmallVector<Attribute, 2> attrStorage;
  attrStorage.reserve(nTypes);

  return wrap(DISubroutineTypeAttr::get(
      unwrap(ctx), callingConvention,
      llvm::map_to_vector(unwrapList(nTypes, types, attrStorage),
                          [](Attribute a) { return a.cast<DITypeAttr>(); })));
}

10:59

is this unholy?

11:03

https://github.com/llvm/llvm-project/releases/tag/llvmorg-18.1.0

Release LLVM 18.1.0 · llvm/llvm-project

LLVM 18.1.0 Release A note on binaries Volunteers make binaries for the LLVM project, which will be uploaded when they have had time to test and build these binaries. They might not be available di...

11:03

llvm 18 is out

Ewan

i've always heard that unsigned is better for raw data and signed is better for text

that hurts -.-

what's wrong with it?

11:05

these are conventions I've seen in software

11:05

use unsigned char if you mean bytes and char if you mean strings

რილია

is there anything to be considered if im going to use unsigned char ? why they made the returned value unsigned char in the first place ?

the standard states that it's in general undefined if char is signed char or unsigned char

implementation-defined

11:05

but yes

11:06

wait no

11:06

it's implementation defined if char is signed or not

11:06

but it's never signed char or unsigned char AFAIK

The signedness of char depends on the compiler and the target platform: the defaults for ARM and PowerPC are typically unsigned, the defaults for x86 and x64 are typically signed.

(edited)

this does not say anything about the point of contention (edited)

erm yes, signed char, unsigned char and char are distinct types iirc

11:09

(even if char is one of the other 2 under the hood)

I just verified this using a compiler

sqlite works for different platforms and doesn't want to mess around with what char really is on the concrete platform/compiler ^^

it wouldn't matter for sqlite3

11:12

it could simply™ cast it to char just at the API boundary if it wanted to be internally oblivious of char's signedness

Just use char8_t Then u epyc c23 coder

11:14

Also it's unsigned

11:15

But this whole mess about data types is something I always disliked about c and cpp

As with most other things, C suffers from still supporting ancient shit

like ddnet

11:19

i get it

cyberFighter

like ddnet

Xdddd

11:19

Based

No need to support ebcdic, no need for CHAR_BIT to be anything but 8

there were architectures with less than 8bit/byte :D

Looks at the External json that cant be merged because we still support.. what was it, Ubuntu 12(?) and some other really old fkn distros and they dont Support latest upstream cammostripes

Fussel

there were architectures with less than 8bit/byte :D

The keyword being "were"

meloƞ

Looks at the External json that cant be merged because we still support.. what was it, Ubuntu 12(?) and some other really old fkn distros and they dont Support latest upstream cammostripes

link?

hey, they for sure run some military weather station

#7703 Heinrich

https://github.com/ddnet/ddnet/pull/7703

change cast of pointer to uintptr_t by BlaiZephyr · Pull Request #7...

1 liner Warning: cast from pointer to integer of different size fixed by pulling latest upstream of json-parser : https://github.com/json-parser/json-parser Checklist Tested the change ingame Pr...

There are also some exotic embedded stuff that uses chars bigger than 8 but honestly they can just compile with an older compiler. C23 has no reason to be able to target a DSP or whatever designed in the 80s

meloƞ

#7703 Heinrich

that doesn't look like "ubuntu 12", but even now, "the currently newest fedora version"

11:24

https://repology.org/project/json-parser/versions

json-parser package versions - Repology

List of package versions for project json-parser in all repositories

11:26

perhaps we could use rust to parse json ^^

I must be getting scammed on my internet bill. 93€ for 2 months is insane

heinrich5991

perhaps we could use rust to parse json ^^

Learath2

I must be getting scammed on my internet bill. 93€ for 2 months is insane

Wtf

heinrich5991

perhaps we could use rust to parse json ^^

Mh, not an amazing fit I imagine. You'd either have to fill a complex C++ structure from rust or have C++ call into rust to retrieve fields

11:28

I'd do the latter fwiw, if you want to give it a go

The size of the byte has historically been hardware-dependent and no definitive standards existed that mandated the size. Sizes from 1 to 48 bits have been used.

11:29

:D

11:29

try to write portable code with this

Fussel

try to write portable code with this

Well C truly makes this sort of portability easy. Unless you are doing something with the bit representation of a char, you'll be fine

Learath2

Mh, not an amazing fit I imagine. You'd either have to fill a complex C++ structure from rust or have C++ call into rust to retrieve fields

I was thinking the first, the structs we parse into aren't that complicated

Learath2

Well C truly makes this sort of portability easy. Unless you are doing something with the bit representation of a char, you'll be fine

I wouldn't say "easy". maybe "possible"

heinrich5991

I was thinking the first, the structs we parse into aren't that complicated

But then you need to generate a struct from each of our json schemas which we don't even have because we don't really enforce

I don't think C is compatible with a 1-bit char btw

I ask myself what the difference between signed char and unsigned char is on a 1-bit-byte architecture ^^

heinrich5991

I don't think C is compatible with a 1-bit char btw

Can't be less than 8b iirc

because a char must be able to store [A-Za-z0-9] IIRC

Learath2

Mh, not an amazing fit I imagine. You'd either have to fill a complex C++ structure from rust or have C++ call into rust to retrieve fields

Im developing a restAPI in rust that Kind of does that, and communicate to IT using ddnets http, curl stuff, If i understood it correctly thats what you meant ?

meloƞ

Im developing a restAPI in rust that Kind of does that, and communicate to IT using ddnets http, curl stuff, If i understood it correctly thats what you meant ?

link?

Private still :P

meloƞ

Im developing a restAPI in rust that Kind of does that, and communicate to IT using ddnets http, curl stuff, If i understood it correctly thats what you meant ?

Not really. We were talking about using ffi

Can Send parts to you in DM later when im Home

Just directly calling into native code compiled from rust

Learath2

Not really. We were talking about using ffi

I c

Http client would have been perfect fit for rust

Yep, not performance critical. I considered it but then I didn't feel like rewriting the entire thing