DDraceNetwork - developer

[ddnet/ddnet] Issue opened: #2945 Variable naming

The meta journey continues. Proposed business rules about variable naming: Classes, structs --- 1. Name starts with C or I. 2. Second letter must exist and be a capital letter. Enum constants --- Must be all uppercase letters, digits and underscores Parameter/variable/member names --- 1. static const, member or not member may follow the Enum constant naming scheme. 2. Single letter lower-case variable and member variables are allowed. This is used for loop indices,...

Jupeyy

[ddnet/ddnet] Pull request opened: #2946 Add directory buttons

fixes #2944 windows: !screenshot_2020-09-29_03-16-32 skin: !screenshot_2020-09-29_03-20-32 but no chance for demo: !screenshot_2020-09-29_03-23-13 naming suggestions a...

With those ^ & soon assets changer, maybe config_directory.bat could be renamed to assets_directory.bat / assets_and_config_directory.bat or possibly there could be two with such names, pointing to exactly same place - just to reduce the confusion

thats true

04:10

would make sense

1

def-

[ddnet/ddnet] Issue closed: #1211 Ingame entities/gameskin changer

05:46

def-

[ddnet/ddnet] Pull request closed: #2928 Add changeable assets

05:46

def-

[ddnet:master] 2 new commits

0e71a47 Add customization - Jupeyy b88b33c Merge pull request #2928 from Jupeyy/pr_multiple_pngs - def-

Feels like downloadable game tiles are on the way too

05:50

I would love to add some custom assets in the future as I use katana pickup as a taxi sign and it confuses players, have to explain every time what it is.

def-

[ddnet/ddnet] Pull request closed: #2940 Fix empty lines in chat (fixes #1835)

05:51

def-

[ddnet/ddnet] Issue closed: #1835 Empty lines in chat

05:51

def-

[ddnet:master] 2 new commits

18ad5b5 Fix empty lines in chat (fixes #1835) - def- 16f965a Merge pull request #2940 from def-/pr-empty-lines - def-

def-

[ddnet/ddnet] Pull request opened: #2947 Dynamic Camera button: Use g_Config.m_ClDyncam as pID

bors[bot]

[ddnet:master] 2 new commits

e4023cb Dynamic Camera button: Use g_Config.m_ClDyncam as pID - def- 113ec25 Merge #2947 - bors[bot]

06:51

bors[bot]

[ddnet/ddnet] Pull request closed: #2947 Dynamic Camera button: Use g_Config.m_ClDyncam as pID

@Learath2

@Learath2 here starts the meta war

07:35

Nah variable naming I think we all mostly would agree on

yeah

07:36

i meant PascalCase

07:36

tbh

07:36

i didnt remember the name

Let's add a default editor to the ddnet project. Now that would be a war

vim obviously

all lower case, pointers end with _
every english word is seperated by _

(edited)

07:36

ah

oh

pointers end with every english word?

my vim supports :Format and uses clang format

07:37

i forgot

07:37

dont even need the script

i just hate discord xd

english words seperated by _ is snake case

07:37

snake_case

thats how i started learning cpp xd

then there is the kebab-case, which is very extremenouis

but honestly i hate it

a point that points to for example the middle of an array wouldnt be prefixed with a a right? i guess it is just a pointer

Nah, that's just a pointer

decayed pointer

07:38

or smth like that

Eh, it's actually not possible to know the difference between two pointers one of which pointing to an int and another pointing to an array of ints

i wonder who made backend_sdl.cpp and decided to ignore all the naming rules

07:39

Not at compile time anyway

i wonder who named it backend_sdl if 99% of the code there is opengl xd

xd

I wonder what other backends they were planning

directx 6

07:41

directdraw

ohgod

07:42

Honestly, the way the gfx code is laid out never really made much sense to me

tbh the whole codebase makes no sense to me

same

07:42

but u live with it

The whole base has some pattern it follows atleast

07:43

Graphics is like 5 different classes

is graphics trully threaded?

Btw this thing we do of swapping on the worker threads is a completely unsupported setup

igraphics, ienginegraphics, cgraphics, cgraphics_threade, cpgrahics_threaded_lockfree, cgraphics_threaded_lockfree_vulkan

SDL says it's undefined behaviour to swap on any thread except the main thread

sometimes having isomething and then csomething becomes tiresome specially to add things

07:45

xd

07:45

but i guess its useful somewhere..

@Learath2 interesting, normally its undefined, if it doesnt hold the opengl context

@Learath2 wasnt it ok to swap where the opengl context is?

@Ryozuki it's nice to have a defined API between two components. I kinda like interfaces honestly

but usually those interfaces are only implemented by 1 class

in our small projects yes

yeah im not talking about firefox here

07:47

OpenGL is fine with it, SDL introduces some OS specific complexities that arent safe on other threads though. Like NSWindow:update on macOS

07:49

@Ryozuki yeah only one class implements them but it gives a nice clean interface. I guess that could also be achieved by formatting the class in such a way that it's declaration can be used as documentation

07:50

This way we have a kernel which I also like actually. idk I have a soft spot for singleton patterns, mostly because pompous ivy league "engineers" don't like it

learath is based

07:51

Help me with my crusade against the academics or as I like to call them the backseat engineers

oh i also hate some academics

07:52

some teacher refused to use return inside forloops

07:52

and stuff like that

07:52

i wonder if that guy actually programmed

@Ryozuki makes control flow much easier

if you only have one place to return, easier to reason with

He refused because technically it breaks the rules of structured programming. ^^

07:53

Though even knuth himself said a single exit point is not too important. When the guy who made the paradigm says it's pointless, I think it's time to quit your fanboyism

Ah I misremembered, it was Djikstra who came up with structured programming

07:56

That man was the prime example of a toxic academic

07:58

Knuth is really the only academic that seems to understand the actual complexities of the field

07:59

Funnily enough, one of the first people who seems to think the early exit is not an issue seem to be the C++ people. (RAII helps a lot with the issues that early exits can cause)

@deen "Then the info text would also fit in 5:4 again" what info text isnt fitting?, i cant find anything

our code is not academic justatest

ah its probs only with video recorder

08:01

nvm

@Ryozuki we shouldn't stop until we can prove that ddnet runs, we should move to haskell so we can actually do that

i rly want to learn haskell

08:02

but i remember having problems with the rls i think

08:02

i rly need a lsp (edited)

08:03

idk whats happening to me

08:03

rls -> lsp

@Deleted User yes, with video recorder

I wonder why the RLS doesn't implement the LSP

u mean rust one?

08:04

RLS is going to be abandoned probs

08:04

rust-analyzer is taking over

08:04

and its better

08:04

u should use it

I don't do much rust to care anyway

based

08:04

I wish I had more to use haskell on, it's actually not too bad to code on and I'd love a reason to learn more

isnt hedgewars made with haskell

08:05

so u can prove hedgewars

08:05

runs

08:06

i wonder where the hedgewars guy that was in the irc channel is

The only good candidate for a ddnet/teeworlds rewrite seems to be rust nowadays

Sadly, I wish we had a little more choice

well competing against c++ is hard

D always looked quite promising to me, but I don't like it's runtime much

and i guess modern ppl prefer languages to be higher lvl

Why do people want their languages to be glorified english? pepehands

@Learath2 elixir or haskell

08:09

is there a difference

The difference is that I never used elixir