DDraceNetwork - developer

[ddnet/ddnet] Issue opened: #4610 Exit cleanly on kill signals

Should catch TERM and INT and handle them cleanly, like shutdown in rcon console. This makes the fifo file get removed properly, remaining ranks stored in sqlite file etc.

bors[bot]

[ddnet:master] 3 new commits

746fc30 Allow multiple comands when adding vote in UI - def- 7e7fac8 Escape voting commands sent from client - def- e03e621 Merge #4604 - bors[bot]

02:20

bors[bot]

[ddnet/ddnet] Pull request closed: #4604 Allow multiple comands when adding vote in UI

plsplsplslol

ye you can use 3 dummies

ok ty

1

08:25

lmao

Ravie

dummy hammer & copy

hi Ravie

ravie

@deen: ever made a native benchmark run with the m1 btw? e.g. the phoronix benchmark maps could be interestingand maybe in entities on a map like Fly to the Legends 2 without any other players(to get max possible fps xd)

hi @Souly

def-

[ddnet/ddnet] Pull request opened: #4611 Default players to being afk on new map

As reported by MitakoKotomi in https://forum.ddnet.tw/viewtopic.php?f=118&p=70154

Checklist

[x] Tested the change ingame
[ ] Provided screenshots if it is a visual change
[ ] Tested in combination with possibly related configuration options
[ ] Written a unit test if it works standalone, system.c especially
[ ] Considered possible null pointers and out of bounds array indexing
[ ] Changed no physics that affect existing maps
[ ] Tested the change with [ASan+UB...

Jupstar ✪

@deen: ever made a native benchmark run with the m1 btw? e.g. the phoronix benchmark maps could be interestingand maybe in entities on a map like Fly to the Legends 2 without any other players(to get max possible fps xd)

not a real benchmark. but I noticed that fps are much higher than with rosetta2. i don't know if phoronix even supports macos

yeah rosetta something like an emulator so thats expected i guessthe openbenchmark ddnet site says macos is supported and last time i checked the test suite didnt really check what client version you use, I could simply replace ~/.phoronix-test-suite/installed-tests/pts/ddnet-1.1.0/build/ with any build (@deen)

deen

not a real benchmark. but I noticed that fps are much higher than with rosetta2. i don't know if phoronix even supports macos

13:08

but we have benchmarking built into the client anyway, guess you just need to run it with the demo to get the fps files, right?

@bencie: whenever someone has around 300-400fps with nVIDIA card, its 100% a laptop with nVIDIA optimus, the worst shit ever invented And the only solution is to buy a new laptop without that.... and seriously suggest exactly that, there is no other fix

13:35

if he's lucky he can disable it in the bios, but probs not

Jupstar ✪

@bencie: whenever someone has around 300-400fps with nVIDIA card, its 100% a laptop with nVIDIA optimus, the worst shit ever invented And the only solution is to buy a new laptop without that.... and seriously suggest exactly that, there is no other fix

Ah i didn't know about that, but good to know. Still, an easy check would be to see if he accidentally locked refresh rate, but it's kinda annoying to even try to troubleshoot with no response other than "i want fps"

tru :P

GitHub

Click to see attachment 🖼️

Yee

heinrich5991

@fokkonaut creating a minimal example can also help you find the error

c++
#include <iostream>
#include <vector>

using namespace std;

class CDoorTile
{
public:
    struct Info
    {
      int m_Index;
      int m_Number;
    };
    
    std::vector<Info> m_vTiles;
};
    
class CDoorTile *m_pDoor;

int main()
{
    int Height = 10;
    int Width = 20;
    
    m_pDoor = new CDoorTile[Height * Width];
    
    for (int i = 0; i < Width * Height; i++)
    {
        CDoorTile::Info Data;
        Data.m_Index = 5;
        Data.m_Number = i+1;
        m_pDoor[i].m_vTiles.push_back(Data);
    }
    
    int Found = -1;
    int Index = 7;
    for (unsigned int i = 0; i < m_pDoor[Index].m_vTiles.size(); i++)
        if (m_pDoor[Index].m_vTiles[i].m_Number == 8)
        {
            Found = i;
            break;
        }
        
    if (Found != -1)
    {
        cout << Found << "\n";
        cout << m_pDoor[Index].m_vTiles.size() << "\n";
        m_pDoor[Index].m_vTiles.erase(m_pDoor[Index].m_vTiles.begin() + Found);
    }

    return 0;
}

@heinrich5991 this is basically what i am doing. And it works in this example, as expected. I have no idea why it would not work in the tw base.

fokkonaut

when i hover over m_pDoor

As i said, this seems to be the weird thing. I dont see how that can happen though

in the original, it says your m_pDoor array is empty, if I understand it correctly?

yes

but you access its first element

no, only if found

but you do in your case

what do you mean?

I can't copy-paste code from an image

fokkonaut

Click to see attachment 🖼️

if DoorIndex != -1

15:23

c++
int CCollision::GetDoorIndex(int Index, int Number)
{
    for (unsigned int i = 0; i < m_pDoor[Index].m_vTiles.size(); i++)
        if (m_pDoor[Index].m_vTiles[i].m_Number == Number)
            return i;
    return -1;
}

DoorIndex is 0, but the m_pDoor array has size 0, if I understand it correctly

heinrich5991

I can't copy-paste code from an image

there are programs that get copyable text from images

so m_pDoor[DoorIndex] is out of bounds

heinrich5991

DoorIndex is 0, but the m_pDoor array has size 0, if I understand it correctly

you mean: m_pdoor[index].m_vTiles[DoorIndex]

15:25

well, but its not

15:25

as you can see, it returned that DoorIndex

m_pDoor[Index] is out of bounds, I mean

huh?

15:26

But how can it be out of bounds, thats also not possible because m_pDoor = new [height * width], as in tw code

15:27

i dont get how you think that can be it

bors[bot]

[ddnet:master] 2 new commits

c033513 Default players to being afk on new map - def- cfe5032 Merge #4611 - bors[bot]

15:27

bors[bot]

[ddnet/ddnet] Pull request closed: #4611 Default players to being afk on new map

I don't get your mindset. you claim that everything is obviously okay, even though you obviously don't understand something: there is a bug

fair

in your debugger, it said that the array is length 0. did I misunderstand that?

yes and no

fokkonaut

when i hover over m_vTiles

.

15:28

it shows it as correct, including correct values if i hover over m_vTiles

15:28

if i hover over m_pDoor though, it shows m_vTiles size = 0

can you output m_Index?

Yes

Index*

wait

15:30

(Index cant be out of bounds though, because it gets set using GetPureMapIndex, but let me try real quick)

15:31

thats my index: 102882

15:31

a valid tile position

okay. I think your debugger is showing m_pDoors[0] when you hover it

would make sense

so that's a different element than m_pDoors[Index] aka m_pDoors[102882]

yea, so the size IS actually correct and i should be able to erase

I'm going to eat something, see you later

fokkonaut

[2022-01-16 19:29:09][hi]: returned: 0 [2022-01-16 19:29:09][hi]: removing: 0/1

as i can also print the values

heinrich5991

I'm going to eat something, see you later

alright

I can not even loop over the vector:

for (std::vector<CDoorTile::SInfo>::iterator i = m_pDoor[Index].m_vTiles.begin(); i != m_pDoor[Index].m_vTiles.end(); i++)
    {
        
    }

15:58

That also asserts. Very weird.

wrong iterator type? is m_vTiles initialized?

how do you initialize a std::vector?

16:01

with .clear()?

if you called normal constructor should be fine

how?

but if m_pDoor[Index] has not been initialized then you might read out garbage

16:02

default constructor should do it implicitly

m_pDoor[Index] definitely is

deen

default constructor should do it implicitly

so i dont have to call anything manually?

fokkonaut

so i dont have to call anything manually?

no, std::vector<int> my_vector; creates an empty vector

yea

I don't see the assertion with the code you sent

16:03

you were supposed to give a piece of code that causes the problem, not one that works

how, if thats exactly what im doing? :D

16:03

now that i think about it there is one thing that is different in fact

16:04

that is that m_pDoor gets mem_zero-ed after creating, could that be it?

definitely

then that has to be it

memzeroing a vector will destroy it

16:04

that's why I said it should be initialized properly

:D oh man

16:04

thinking u have everything correct and then someone elses year-old code comes in :D

16:04

i'll check one sec

where do you store your crystal balls, deen?

16:05

I also need one

heinrich5991

where do you store your crystal balls, deen?

you sure its called "balls"?

16:06

Thats sounds soo weird to me

https://en.wikipedia.org/wiki/Crystal_ball

Crystal ball

A crystal ball, also known as an orbuculum or crystal sphere, is a crystal or glass ball and common fortune-telling object. It is generally associated with the performance of clairvoyance and scrying in particular. In more recent times, the crystal ball has been used for creative photography with the term lensball commonly used to describe a cry...

fokkonaut

i'll check one sec

nice, it actually was it

heinrich5991

https://en.wikipedia.org/wiki/Crystal_ball

This article needs additional citations for verification

you think this is about the name?

Yeah, I assume

lol, definittely not

16:07

but I can give you another refernece

But I am not sure, maybe my mindset is blown

https://en.wiktionary.org/wiki/crystal_ball https://www.oxfordlearnersdictionaries.com/definition/english/crystal-ball?q=crystal+ball https://www.merriam-webster.com/dictionary/crystal%20ball

crystal ball

crystal-ball noun - Definition, pictures, pronunciation and usage n...

Definition of crystal-ball noun in Oxford Advanced Learner's Dictionary. Meaning, pronunciation, picture, example sentences, grammar, usage notes, synonyms and more.

Definition of CRYSTAL BALL

a sphere especially of quartz crystal traditionally used by fortune-tellers; a means or method of predicting future events… See the full definition

Okay okay, you right

But imho it still sounds weird

16:09

Gonna use crystal sphere instead

@fokkonaut: just save Width * Height in a member var and check if the index is valid at runtime

a error that rust would have prevented poggers

102k doors sounds alot xd

16:12

he can also use a vector and use at()

Jupstar ✪

102k doors sounds alot xd

his mod has crazy stuff i think

16:12

at does bounds checking?

yes

nice

but in debug mode i think even [] does it

16:13

for vectors

Jupstar ✪

@fokkonaut: just save Width * Height in a member var and check if the index is valid at runtime

all fine, it was because of mem_zeroing

Jupstar ✪

102k doors sounds alot xd

thats the door tiles only.

Jupstar ✪

but in debug mode i think even [] does it

well u probs need a flag for it

Jupstar ✪

102k doors sounds alot xd

Every single tile has a possible doortile :D

16:14

and the only thing i did is now to make each tile holding a vector, so multiple doors with different numbers can be put on one single tile

i see, then ur map is just huge xd

16:17

since c++11 zeroing any kind of memory allocated with new is stupid anyway

Jupstar ✪

i see, then ur map is just huge xd

Nah, that spot is actually at 400/80 or so

ok there is one exception, you can still allocate non initialized memory by leaving out the parenthesis and having no user defined constructor