DDraceNetwork - developer

DDraceNetwork

Development / developer

Development discussion. Logged to https://ddnet.tw/irclogs/ Connected with DDNet's IRC channel, Matrix room and GitHub repositories — IRC: #ddnet on Quakenet | Matrix: #ddnet-developer:matrix.org GitHub: https://github.com/ddnet

Between 2020-04-25 00:00:00Z and 2020-04-26 00:00:00Z

@heinrich5991 hm, won't that cause rust to become a mess chock full of backwards compatibility code for deprecated features?

the modern way in rust to do a try! is with the ? operator right?

08:00

https://doc.rust-lang.org/book/ch09-02-recoverable-errors-with-result.html#a-shortcut-for-propagating-errors-the--operator

Recoverable Errors with Result - The Rust Programming Language

@Learath2

08:02

Why my points in command dont update?

08:03

justatest

Wat?

@Learath2 there's the concept of editions, which allows to deprecate features

08:14

but it's opt-in

08:14

you can opt into a newer edition

08:14

code is interoperable in both directions, so it's not like a python 2/3 split

how can i set up my server so that grenades dont disappear when i /kill

sv_destroy_bullets_on_death 0

i wish rust rls worked better

12:36

its rly a drawback for me

12:36

clangd works way way better and c++ syntax and even dependency and include dirs seem more complicated yet works better than rls

@Ryozuki try intellij with rust plugin

13:27

it's really nice

it was worse for me cuz for some reason intellij froze every time

hm

maybe with 2020 update it works

it worked before, dunno what's different about our setups

starting to test my map parser too see what i need to fix

19:56

tho i have much more fun right now arguing why on of the maps is broken, not my tool

19:56

the moderate map "DarkRoad" has like 20 envelopes, didnt figure out yet where they actually get used

19:57

the 20th envelope however

19:57

has a normal first point

19:57

the second is on -402456.576 secs

19:57

with the curve type 16777216 (only curve types 0-4 exist to my knowledge)

19:58

seems like some shifting of bytes happened, tho all the other envelopes seem fine

.

.

I was hoping that'd scroll my weechat up

hmm

20:14

there is more fun to be had on another map

20:15

the solo map "Extension 1.0" has a envelope with type 8

20:15

there only are types 1, 2, 4

20:16

when you open it in the ddnet editor, the value buttons look really messed up and display some random string

20:16

and hovering over them either displays some random data or crashes your client

20:16

im a tiny bit scared to click them

instant RCE

same with our favourite map "nut_hardcore_bestof"

20:22

they each share that their maps have 6 envelopes with the third being broken and 4-6 empty

20:22

mightve been a broken (maybe non-standard) map editor

maybe from a time with the ddnet/ddrace split?

[ddnet/ddnet-scripts] Pull request closed: #19 Allow non-interactive queries

20:47

[ddnet-scripts:master] 2 new commits

cf0f893 Allow non-interactive queries - Learath2 61675db Merge pull request #19 from ddnet/th-query - def-

Maybe from even earlier than that? Could be the 0.5 -> 0.6 map conversion

whats ddnet/ddrace split

ddnet used to be a competitor to the ddrace mod

21:01

due to lack of development, ddnet easily won the competition

How can I use a template type within a cpp file?

21:39

I just want to use it in one function

template<typename T>

21:42

void yourfunction(T a, T b)

but i dont want to pass a T

21:42

i want to use a T inside

i dont get it

int SendPackMsg(T *pMsg, int Flags, int ClientID)
    {
        int Result = 0;
        T Tmp;

like this Tmp, but I dont want to pass a T like *pMsg

i dont think u can do that

21:43

well

21:43

u can

21:43

with

21:43

template<typename T> void SendPackMsg<T>(..) { T tmp;

21:43

i think

21:43

iirc

21:44

test it

21:44

u have to pass a type either way

21:44

or the compiler wont know

21:44

maybe another knows more

The most funny thing about the gamebreaking physics change to the hammer, heinrich5991 and Learath2 did without listening to the arguments of the community is that the thing which was meant to be fixed still works bugged like before

void SendPackMsg<T>(..) {

dont get it?

u will have to call SendPackMsg<withyourtypehere>()

21:46

i dont get exactly what u try to achieve

oh oh, I remember that change, I didn't know the community consisted of 2 name banned cheaters and that guy who complains about everything who'd rather play on 0.5

feelsamazingman

1

@Ryozuki I want to use T inside CEventHandler::Snap, so I can create a T *pEvent there, so I dont have to copy code for a few different events

"cheaters" is a good argument since it opens a lot of cheats in a lot of maps. Still its a bad thing which should never had been done

i dont think people are getting namebanned for the type of "cheating" that konsti is referring to?

i was never banned from DDNet Servers

maybe u just want to use simple inheritance? make a event interface

21:49

sry i gotta play a dota ranked now

@Ryozuki basically, i want to make this code just one block: https://github.com/fokkonaut/F-DDrace/blob/F-DDrace/src/game/server/eventhandler.cpp#L54-L93

yeah i guess with a event interface u can do that

Whats that?

a interface is like a class

21:50

but it doesnt implement anything

21:51

brb

the solution to this would probably be to use dynamic casts with inheritence

how do they work?

but we don't do dynamic casts in teeworlds code, we instead fake them by keeping information about the type of the thing and doing an unsafe cast

21:53

you can try an interface like ryozuki says

if (m_aTypes[i] == NETEVENTTYPE_DEATH || m_aTypes[i] == NETEVENTTYPE_DAMAGE)
                {
                    CNetEvent_Death *pEvent = (CNetEvent_Death *)&m_aData[m_aOffsets[i]];
                    if (m_aTypes[i] == NETEVENTTYPE_DAMAGE)
                        (CNetEvent_Damage *)pEvent;

well, that works too :D

21:56

but i wont do it haha

21:58

Well, I could

21:59

@Learath2 is there anything bad about it? xD

just no

xD

Is release mode server on linux also faster than debug?

rls mode is always going to be slightly faster than dbg afaik, idk if its enough to be noticeable, and it doesnt matter which os

Well, it should be "way" faster. I remember when I played around with 256 tees in 0.6, in rls mode it was no problem, in debug tho it lagged very much xD

i think dbg mode turns off compiler optimization and rls mode turns it on

22:24

if that causes the code to become way faster then maybe it could be manually optimized a bit?

256 players is a number where you should probably stop doing O(n^2) algorithms

@heinrich5991 :D i am using it on 128 right now

Hi, I can't open DDNet I don't know why, can someone help me please? (edited)

Image attachment

@Shrod can you see if you find /usr/lib/libGLEW.so.1.xx in your system, where XX is some number?

wow uh what is it?@Rafael Fontenelle

23:42

i write /usr/lib/libGLEW.so.1.xx in terminale ?

23:45

or this?

Image attachment

Try running ls /usr/lib/libGLEW.so.1* in the terminal

23:51

@Shrod

ls: unable to access '/usr/lib/libGLEW.so.1*': No such file or folder @Rafael Fontenelle

What distro are you using?

linux mint 19.03

Exported 115 message(s)