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 2022-05-25 00:00:00Z and 2022-05-26 00:00:00Z

Chairn

@ᶰ°Konͧsti ^

?

[ddnet/ddnet] Issue closed: #3396 Memory leak in CMasterServer::SetDefault()

[ddnet/ddnet] Pull request opened: #5214 Use CMAKE_PROJECT_NAME instead of hardcoding ddnet

Output for ddnet is the same. But now it is easier and less conflicting for downstream projects to do rebranding.

holy shit new UI

08:07

i finally looked into merging the nanoseconds code and boom i get this beast of an update

08:09

any chance we can customize the ddrace HUD further? I would really like to hide the jump status

[quakenet] ChillerDragon

any chance we can customize the ddrace HUD further? I would really like to hide the jump status

but you like the rest or what? -.- add a recommendation to this https://github.com/ddnet/ddnet/issues/5159 (edited)

i very much like the weapon status. I got so use to the weapon indicator from F-DDrace that i always wonder what my current weapon is when frozen in ddnet.

08:41

Its awesome!

08:42

also like healthbar in ddrace is not too useful :D

08:42

but the dj i see on my tee and the indicator in the corner is just bloat and distracting

08:43

yea classic heinrich agianst configs there is no chance he will allow dat

08:45

Anyone having xp with hosting own github runners? Mine are always reaching limits and failling due to that. How is ddnet not running into limits? Do you get more CI runtime due to the amount of stars and contributors? Or am i actually commiting more? :D

please no config options

[quakenet] ChillerDragon

Anyone having xp with hosting own github runners? Mine are always reaching limits and failling due to that. How is ddnet not running into limits? Do you get more CI runtime due to the amount of stars and contributors? Or am i actually commiting more? :D

all free projects have the same runner limmits

09:10

https://docs.github.com/en/actions/learn-github-actions/usage-limits-billing-and-administration

Usage limits, billing, and administration - GitHub Docs

https://docs.github.com/en/actions/learn-github-actions/usage-limits-billing-and-administration#usage-limits

Usage limits, billing, and administration - GitHub Docs

09:11

I don't see any limits there ChillerDragon

@heinrich5991 could you take a quick look at this: https://github.com/ddnet/ddnet/pull/5210#issuecomment-1136988350 maybe you can think of something that could be the problem

Move ddnet character display info in ddnet character by C0D3D3V · P...

I also added the default value mechanic to the datatype.py (copied from seven protocol) and changed the rule so that we can use NetIntRange if validate_size = false if you have set a default value ...

no particular insights, likely a buffer overflow somewhere

heinrich5991

no particular insights, likely a buffer overflow somewhere

maybe if a snap comes in, it mallocs only for the size of the snap and not the size that the snap could have

09:22

and then I write the default values in non-allocated space

that sounds about right

09:23

the snap has allocated the memory for precisely the objects it contains

09:23

you can't write larger stuff into it without some changes

maybe it makes sense if validate_size is off to copy the data in a new allocated space, of the size of the true snap... I will try it

[ddnet/ddnet] Pull request closed: #4833 predicted weapons now uses non predicted player positions if player prediction is turned off

just noticed that default values are only applied if it is a NetMessage not a NetObject XD

ouch

09:54

now I remember how I added it

09:54

I wondered how that might possibly work

09:59

do you know of some standard or "standard" binary serialization formats, @Learath2?

09:59

it almost seems like protobuf2/3 is the only one widely used

no limits!?

10:02

Embedded image

10:02

so this (!) is just a buggo?

click on it, probably upload faild

Embedded image

10:02

or this

10:02

killed the compiler by signal!?

10:03

when clicking it it has no additional info https://github.com/DDNetPP/DDNetPP/runs/6587744914?check_suite_focus=true

Build · DDNetPP/DDNetPP@557a351

DDraceNetwork + city + block! Based on ddnet (www.ddnet.tw) which is based on teeworlds (www.teeworlds.com). - Build · DDNetPP/DDNetPP@557a351

probably out of memory?

so buildimg ddnet takes to much ram at once? :D

perhaps it's the IPO linking stuff

it alr fails on debug mode

10:04

that does not use the IPO right?

depends on your merge state wiht master, I think

10:05

not sure

makes no sense to me to generate network stuff by python, and in the end building the package on our own without the generated packer xD

Image attachment

im at latest master

I agree with that @c0d3d3v

[quakenet] ChillerDragon

im at latest master

just try to rerun the CI

heinrich5991

do you know of some standard or "standard" binary serialization formats, @Learath2?

Protobuf is pretty much what everyone uses, there is also msgpack which we use in the web code

[quakenet] ChillerDragon

any chance we can customize the ddrace HUD further? I would really like to hide the jump status

we shouldnt have a jump status there anyways, since its marked on the tee

fokkonaut

we shouldnt have a jump status there anyways, since its marked on the tee

I like the jump status

how is that useful tho?

10:52

if it only indicates dj or not

10:52

it shouldnt always be there imo

10:52

only if u have 0 jumps for example

fokkonaut

how is that useful tho?

It is more present then the feet color

nonono, especially not if you are playing instead of watching

10:54

for me it would maybe even distract me when the icon is blinking all the time

10:54

And thats no ddrace hud thing, its just a jump that is already indicatef

I had a lot of times sitting on a teetower and tired to dj... + move ---- resulting in all dead, only because feed color was not a good indicator

yea i see

10:55

But we should have like a box like the flag selector in the browser of ddnet tab, to filter for countries (edited)

10:55

that for the ddrace hud

10:56

so we can individually select which ones to render and which not

10:56

Bcs i personally would use a few only, and definitely no freeze indicator etc

10:57

If we add stuff like this we usually need to add a bit of customization, and i think such a selection box is the best way to go without adding hundreds of configs

10:57

we can just save the config in a bit config var

Image attachment

mhh

you mean this

Image attachment

does the character info work btw? like after merging the net objects? did u add flags btw?

c0d3d3v

you mean this

ye on country side

fokkonaut

does the character info work btw? like after merging the net objects? did u add flags btw?

yes... works also without the new data

so flags are added to let the server customize it from gameinfoex snap?

we could also visualise it like for chat you could implement that (edited)

Image attachment

c0d3d3v

we could also visualise it like for chat you could implement that (edited)

could be doable

fokkonaut

so flags are added to let the server customize it from gameinfoex snap?

no

11:00

I did not say anything about that in the PR

11:00

so no

toxic i didnt check ur pr xd

11:01

monkagun

then look at it, and see that your Idea does not work as good as you have thought

11:02

extending net objects is a pain

It was just a general idea, i know that its a pain

no you forced me xD

No? xd

Someone will have to suffer it sooner or later

1

It is just a clean way, because those fit together

11:03

and 2 net obj for 1 thing bascially

11:03

is useless

11:04

i think we should just add versioning code when packing the snapshot, when creating new events, and only allocate and use fields, and also send ofc, that are valid for client version x of the client

11:05

we dont have anything else besides version checking, still no client capabilizy

just fix packer und unpacker... to work and do the defaults correct... I will try that later. No versioning needed

oke

mh but versioning would not be to bad xD

11:06

but not client version

11:06

only snap version

Ye, bcs its not possible for older clients

you could implement that

(edited)

snap Version?

11:06

wym?

so I do not have to check if all values are there, or check about default values (edited)

11:07

mh probably not a lot better

the Server cant handle any default values

11:07

the server has to send a smaller snapshot object to older clients

11:07

missing the new fields

[ddnet/ddnet] Pull request closed: #5176 Add Rus (white blue white) flag

or is the client not so restrictive when upacking snaps? bcs iirc it is

sounds wrong

11:08

yes it is not

ah

only if the snap recieved is to small

so it can just unpack the object with new data appended?

11:08

ah okay

11:08

well

11:08

where is the problem then xd

in unpacker of netobjects

due to error?

I do not know... fetch the PR and debug it yourself xD I just have to do other things

ah i see

@c0d3d3vfix scoreboard pls, in new update ddnet when i open scoreboard i have 30 fps

11:43

i downgrade scoreboard and have 500 fps

Ясно Понятно

@c0d3d3vfix scoreboard pls, in new update ddnet when i open scoreboard i have 30 fps

we are working on it... see https://github.com/ddnet/ddnet/pull/5194 and https://github.com/ddnet/ddnet/pull/5206

12:16

and https://github.com/ddnet/ddnet/pull/5193

c0d3d3v

we are working on it... see https://github.com/ddnet/ddnet/pull/5194 and https://github.com/ddnet/ddnet/pull/5206

ty

uh https://github.com/ddnet/ddnet/pull/5063 is now on ger1 server

time to test jao shooter and increase your speed

uh I finished increase your speed, seems to work

and is way less annoying then before

nightly joining old server versions:

murpi

nightly joining old server versions:

on specific map? or general

all maps, the server version is extremely old though so not sure if that's worth fixing

12:34

DDNet 13.2.2

and joining with 16.0.3 is better?

I joined with nightly on a 13.2.2 version server

12:38

this bug doesn't happen when joining any of the more recent versions

does it tell you in debug hud about a lot of net_correction_errors?

what's debug hud

12:39

I just tested it, looks really strange xD the map looks out of sync

12:42

no net_correction_errors

murpi

what's debug hud

ctrl + shift + d

12:42

if you turn debug hud on it also looks double as strangle

ah, debug mode

12:43

you're the very first person I know to call it debug hud

it is called like that in the code xD I only use the code class name

12:44

I think it is related to the teehistory change

12:44

I think we now use early inputs

12:44

not sure

12:50

it is already in 16.1 buggy so... not sure what change that could have introduced

[ddnet/ddnet] Issue closed: #5213 Inform the player about using a "secure" password

murpi

nightly joining old server versions:

seems to be input related, if you are freezed or stay against a wall it is fine (edited)

no idea :D

[ddnet-scripts:master] 1 new commit

2e08e68 IPO fails on win32 with mingw: DWARF error: could not find variable specification at offset - def-

14:20

[ddnet/ddnet] Issue opened: #5215 Crash in swap with teehistorian

Reported by murpi: ``` (gdb) bt #0 0x0000000000000000 in ?? () #1 0x00005572dc8c911b in CTeeHistorian::Write (DataSize=, pData=0x7ffc265d0840, this=0x7ff86447bdf8) at /home/teeworlds/src/master/src/game/server/teehistorian.cpp:367 #2 CTeeHistorian::WriteTick (this=0x7ff86447bdf8) at /home/teeworlds/src/master/src/game/server/teehistorian.cpp:390 #3 0x00005572dc8ca1a5 in CTeeHistorian::EnsureTickWritten (this=0x7ff86447bdf8) at /home/teeworlds/src/master/src/game/server/teehistorian....

14:23

[ddnet/ddnet] Issue opened: #5216 Do not snap dragger if it is no longer in reach?

I have noticed on jao shooter that you can now see the dragger for one tick that drags the dummy back to its position. This is mainly caused because of the efficient way I implemented now the draggers. In the old version I think there where just to much lag to see the dragger beams, with much luck you sometimes see them. https://youtu.be/xvPzGEXlhKQ The behaviour is like this: ``` # two dragger beams created to drag the dummy (jao shooter uses most of the time two draggers to pull th...

14:24

[ddnet:master] 2 new commits

630ff22 Use CMAKE_PROJECT_NAME instead of hardcoding ddnet - ChillerDragon f484ada Merge #5214 - bors[bot]

14:24

[ddnet/ddnet] Pull request closed: #5214 Use CMAKE_PROJECT_NAME instead of hardcoding ddnet

14:25

[ddnet/ddnet] Issue closed: #2057 Logic warning from clang [-Wtautological-bitwise-compare]

14:27

[ddnet/ddnet] Issue opened: #5217 Crash in CDraggerBeam

``` Program terminated with signal SIGSEGV, Segmentation fault. #0 0x000056394101727e in CGameWorld::GameServer (this=) at /home/teeworlds/src/master/src/game/server/gameworld.h:45 45 class CGameContext *GameServer() { return m_pGameServer; } [Current thread is 1 (Thread 0x7fd3da9e7e00 (LWP 3327556))] (gdb) bt #0 0x000056394101727e in CGameWorld::GameServer (this=) at /home/teeworlds/src/master/src/game/server/gameworld.h:45 #1 CEntity::GameServer (this=) at /home/teeworlds/src/mas...

14:27

[ddnet/ddnet] Issue closed: #3623 Client crashes after computer standby

14:28

[ddnet:master] 1 new commit

049a249 Crash fix (fixes #5217) - def-

14:28

[ddnet/ddnet] Issue closed: #5217 Crash in CDraggerBeam

@c0d3d3v my weapon indicator seems to glitch while frozen

14:33

idk its ddnet++ server so might also a server issue

GitHub

Click to see attachment 🖼️

nooice

ᶰ°Konͧsti

?

you got a bug to exploit there

will it be 0seconds or what

[quakenet] ChillerDragon

@c0d3d3v my weapon indicator seems to glitch while frozen

I already have a issue for that.... and already made a PR to patch it, but heinrich wanted that I investigate if I can fix it without sending more data

ᶰ°Konͧsti

will it be 0seconds or what

maybe even negative

[quakenet] ChillerDragon

@c0d3d3v my weapon indicator seems to glitch while frozen

https://github.com/ddnet/ddnet/issues/5149

c0d3d3v

I already have a issue for that.... and already made a PR to patch it, but heinrich wanted that I investigate if I can fix it without sending more data

I'm still investigating

[ddnet/ddnet] Pull request opened: #5218 do not dragger beam snap, if the target is out of reach

fixes #5216

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
[x] Considered possible null pointers and out of bounds array indexing
[x] Changed no physics that affect existing maps
[x] Tested the change with [ASan+UBSan or valgrind's memcheck](https://github.com/ddnet/ddnet/#using-addresss...

only me get crash when try join server?

Ясно Понятно

only me get crash when try join server?

server crash or client crash?

deen

server crash or client crash?

client

what server are you joining?

Image attachment

deen

what server are you joining?

ddnet

14:49

or kog

any dump files? can you check with a debugger?

w8

deen

any dump files? can you check with a debugger?

Before that, I had crashes in moments in the game referring to std::clamp, algorithm.h. and i see new commit fix crashes, and i update and then i cant join to server , i get crash

14:54

idk its me did something or no

this crash had nothing to do with clamp

14:54

I can join servers

Image attachment

deen

I can join servers

ok

Can you get more details of that?

14:55

Like the call stack / stack trace

idk how to do that

14:56

i not pro

at the bottom right on your screen I see call stack

14:56

Is this straight ddnet or your own changes merged in? You could test without your changes

deen

Is this straight ddnet or your own changes merged in? You could test without your changes

i run a clear ddnet client, (master) without my changes

14:58

and it still crash

[ddnet/ddnet] Pull request opened: #5219 Don't try other linkers on macOS

``` ld64.lld: warning: /opt/homebrew/opt/freetype/lib/libfreetype.6.dylib has version 12.0.0, which is newer than target minimum of 11.0.0 ld64.lld: error: LC_DYLD_INFO_ONLY not found in /opt/homebrew/opt/freetype/lib/libfreetype.6.dylib ld64.lld: warning: /usr/local/lib/libGLEW.2.2.0.dylib has version 12.0.0, which is newer than target minimum of 11.0.0 ld64.lld: error: LC_DYLD_INFO_ONLY not found in /usr/local/lib/libGLEW.2.2.0.dylib ld64.lld: warning: /opt/homebrew/Cellar/wavpack/5.4....

Image attachment

[ddnet/ddnet] Issue closed: #5209 Prefix for `std::vector` variables

deen

Like the call stack / stack trace

that?

Image attachment

Ясно Понятно

that?

looks like it is related to my attached hooks list

15:03

what server do you try to connect with?

that one

Image attachment

15:03

and i try others, but still crash (edited)

@Ясно Понятно do official ddnet servers work?

heinrich5991

@Ясно Понятно do official ddnet servers work?

-

ah you crash as soon someone hooks you

nope

15:05

i crash on step connecting

15:05

w8

I just crashed, as soon as you hooked me

c0d3d3v

I just crashed, as soon as you hooked me

xD

15:06

but i crash on connect to server

15:06

w8 i give screenshot

it is enough that a hook is in your view

But it crashes me even when I don't see the game scene

[ddnet:master] 3 new commits

590eb2c Revert "Render full text when it still fits" - def- 0c8bf5c Revert "Ellipsis at end in scoreboard and hud" - def- 00c1671 Merge #5194 - bors[bot]

15:08

[ddnet/ddnet] Pull request closed: #5194 Revert Ellipsis (alternative to #5193)

15:08

[ddnet/ddnet] Issue reopened: #2057 Logic warning from clang [-Wtautological-bitwise-compare]

c0d3d3v

it is enough that a hook is in your view

on that i get crash

Image attachment

[ddnet/ddnet] Issue reopened: #3623 Client crashes after computer standby

Ясно Понятно

on that i get crash

always?

c0d3d3v

always?

yes

ah I see World.m_apCharacters is only availible on ddnet servers

c0d3d3v

ah I see World.m_apCharacters is only availible on ddnet servers

but when i connect in ddnet i get crash too

c0d3d3v

xD the code does no longer exists ... so yes should be fixed

if there are issues with code references that should be updated, e.g. bcs code was moved you can also just update the issue comment to the new code position if you want (at least in this case it defs makes sense ^^)

Jupstar ✪

if there are issues with code references that should be updated, e.g. bcs code was moved you can also just update the issue comment to the new code position if you want (at least in this case it defs makes sense ^^)

true

ah and just in this moment heinrich did that

@c0d3d3vwhy only me get crash?

try latest master

15:19

do you have show all on?

c0d3d3v

try latest master

pull?or fetch

Image attachment

15:19

i press pull, compile, and still crash

can also test with asan

c0d3d3v

do you have show all on?

idk

15:19

mb no

c0d3d3v

try latest master

can u give me ur build

you can download it from the bottom of this page https://github.com/ddnet/ddnet/actions/runs/2367748567 (edited)

15:22

updated the link

c0d3d3v

you can download it from the bottom of this page https://github.com/ddnet/ddnet/actions/runs/2367748567 (edited)

i still crash

c0d3d3v

updated the link

ok w8 try

@c0d3d3v have u used asan yet btw?

c0d3d3v

updated the link

that not crash

15:23

before u say me a link its still crash

15:23

but that no crash

Jupstar ✪

@c0d3d3v have u used asan yet btw?

for what problem? I already know the source of the crash for my attached looks list. But I always have asan on yes

yeah i am generally asking

15:24

but then fine

at least I had it on in my build setup, since I use the cmake/clang extension I'm not sure what that configures :/

[ddnet/ddnet] Issue opened: #5220 Vanilla 0.7 Scoreboard Ping display not working correctly

@c0d3d3vwhat should I do?

[ddnet/ddnet] Issue opened: #5221 Prefix/naming for `std::map` variables

Prefix m seems reasonable, but what about the types of the map? std::map<int, std::vector> *m_pTeleCheckOuts = nullptr; should be std::map<int, std::vector> *m_pmvTeleCheckOuts = nullptr; ?

@Ясно Понятно Botters & Bot developers wont get support

15:30

^ @Dev

murpi

@Ясно Понятно Botters & Bot developers wont get support

i`am not bot developer

[ddnet/ddnet] Issue closed: #5216 Do not snap dragger if it is no longer in reach?

15:32

[ddnet/ddnet] Issue closed: #5216 Do not snap dragger if it is no longer in reach?

15:32

[ddnet:master] 2 new commits

82a0354 do not dragger beam snap, if the target is out of reach - C0D3D3V ac15943 Merge #5218 - bors[bot]

@Ясно Понятно what's your ingame name?

heinrich5991

@Ясно Понятно what's your ingame name?

nameless tee

15:33

i reinstall game

you can try this patch in gamecore.cpp

Image attachment

c0d3d3v

you can try this patch in gamecore.cpp

did not help

c0d3d3v

you can try this patch in gamecore.cpp

https://github.com/ddnet/ddnet/actions/runs/2385350722 that still crash

Merge #5218 · ddnet/ddnet@ac15943

DDraceNetwork, a cooperative racing mod of Teeworlds - Merge #5218 · ddnet/ddnet@ac15943

[ddnet/ddnet] Pull request opened: #5222 Fix crash because of hooks on non ddnet server

Reported by @Ясно Понятно on non ddnet server, this code gets run to detect strong-weak ids: https://github.com/C0D3D3V/ddnet/blob/6a5673aa04c216948ce76de1a505c419d0715b1e/src/game/client/gameclient.cpp#L2540-L2541 which causes a crash because this CCharacterCore is not bound to a gameworld the check for m_pWorld failed because m_pWorld is fill of random data https://github.com/ddnet/ddnet/blob/ac15943e85b04f29fc8e85096222c5f6a85ac7f1/src/game/gamecore.cpp#L625 So I would also like to...

Ясно Понятно

https://github.com/ddnet/ddnet/actions/runs/2385350722 that still crash

you have probably a huge default zoom

15:47

or showall

Image attachment

c0d3d3v

or showall

showall disabled

GitHub

Click to see attachment 🖼️

try this patch

c0d3d3v

try this patch

idk still crash

can you debug it again

c0d3d3v

can you debug it again

how

16:05

i just download build

16:05

not source

did you try still the same server?

yes

16:06

w8 try ddnet

16:08

its crashing

you are on windows?

+

[ddnet:master] 3 new commits

6a5673a fix crash on non ddnet server if hook is in view - C0D3D3V a5cf355 init m_pWorld with nullptr - C0D3D3V 1ca4220 Merge #5222 - bors[bot]

16:09

[ddnet/ddnet] Pull request closed: #5222 Fix crash because of hooks on non ddnet server

c0d3d3v

can you debug it again

Image attachment

ah do you have predict off?

Image attachment

first

on

16:13

cl_predict 1

c0d3d3v

first

i set cl_predict 0 , but still crash

you trace seems not to be the latest master, line number does not match the code

16:15

can you click on the second entry in the trace? and screenshot it (edited)

[ddnet/ddnet] Issue closed: #3934 Crash in client prediction on Avoid map

c0d3d3v

you trace seems not to be the latest master, line number does not match the code

i scroll code

16:15

first

Image attachment

16:15

second

Image attachment

c0d3d3v

you trace seems not to be the latest master, line number does not match the code

latest

Image attachment

[ddnet/ddnet] Issue closed: #4297 My game is crashing no matter which map I enter

16:26

[ddnet/ddnet] Issue closed: #4438

clock not centered

@Rafael Fontenelle @Ryozuki Could one of you update AUR? https://aur.archlinux.org/packages/ddnet

I just tried it in a windows VM, I see it crashes, I will try to debug it

c0d3d3v

I just tried it in a windows VM, I see it crashes, I will try to debug it

what ur OS

Arch linux

[ddnet/ddnet] Issue opened: #5223 Install reminders to better handle issues

I just stumbled on https://github.com/ddnet/ddnet/issues/4647 where heinrich5991 waits for more info, but there are other issues with similar waiting for more info. I thought it would be nice to have a reminder after 6 months for instance and i found this bot: https://github.com/probot/reminders Are you okay with installing it?

[ddnet/ddnet] Issue closed: #5223 Install reminders to better handle issues

Ясно Понятно

what ur OS

can you send me your build? it is a pain to compile in the VM (edited)

c0d3d3v

can you send me your build? it is a pain to compile in the VM (edited)

yes

deen

@Rafael Fontenelle @Ryozuki Could one of you update AUR? https://aur.archlinux.org/packages/ddnet

Will do when i get home

17:09

Hope i remember xd

does someone have experience with mingw crosscompile for windows, I always get :

CMake Error in CMakeLists.txt:
  IMPORTED_IMPLIB not set for imported target "GTest::gtest" configuration
  "RelWithDebInfo

there's an issue about it

17:21

https://github.com/ddnet/ddnet/issues/4457

Cross-build for windows fails when gtest is installed on Linux · Is...

IMPORTED_IMPLIB not set for imported target "GTest::gtest" configuration Uninstalling gtest on Arch Linux worked around it. @heinrich5991 any idea what's happening? My gue...

[ddnet/ddnet] Pull request opened: #5225 Only record swap in teehistorian if active

Checklist

[ ] 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+UBSan or valgrind's memcheck](https://github.com/ddnet/ddnet/#using-addresssanitizer--u...

lol in mingw build the client does not crash -.-

17:40

what is wrong with these compilers now adays (edited)

[ddnet/ddnet] Issue closed: #5174 Register output even when sv_register 0 is set

17:43

[ddnet/ddnet] Issue closed: #5178 Set log_level?

is there a way to build ddnet on windows without visual studio xD

tada, i finished sorting issues

c0d3d3v

is there a way to build ddnet on windows without visual studio xD

i use mingw

17:48

i also use git bash as a terminal

Chairn

i use mingw

mingw does not contain the bug

17:48

nightly build crash on server join

can you try to cross compile without gtest?

done

17:49

I only did crosscompile

Chairn

tada, i finished sorting issues

colorfull

so what's the matter?

I need the build like it is build in the CI, but with sources xD and debug symbols

17:50

or someone with windows takes over the search ...

nice only 345 issues left

[ddnet:master] 2 new commits

93ece6f Only record swap in teehistorian if active - Zwelf 4ec67b8 Merge #5225 - bors[bot]

17:54

[ddnet/ddnet] Issue closed: #5215 Crash in swap with teehistorian

17:54

[ddnet/ddnet] Pull request closed: #5225 Only record swap in teehistorian if active

I've been testing my own pr in vs, what is the bug @c0d3d3v ?

344

@Chairn the last two issues you closed aren't actually fixed

Learath2

I've been testing my own pr in vs, what is the bug @c0d3d3v ?

On windows the client crash as soon you connect to a server, the stacktrace indicates, it should be in the evolve function maybe because there a temp world is generated instead of useing the real

Image attachment

[ddnet/ddnet] Issue reopened: #5178 Set log_level?

18:04

[ddnet/ddnet] Issue reopened: #5174 Register output even when sv_register 0 is set

@Chairn it's also nice if you include a reason for closing

c0d3d3v

On windows the client crash as soon you connect to a server, the stacktrace indicates, it should be in the evolve function maybe because there a temp world is generated instead of useing the real

Does it matter what type of server I join?

if I try to connect to high ping server I only get this message

Image attachment

18:05

on KOG servers (edited)

c0d3d3v

on KOG servers (edited)

ddnet too

c0d3d3v

if I try to connect to high ping server I only get this message

is that wrong?

do not know, I think if players are on the server I should be able to connect

18:07

this message comes nearly instant (edited)

hmm. instantly is a bit weird

18:08

(but also true :P, no answer from server yet)

Learath2

Does it matter what type of server I join?

no

@c0d3d3v can you report an issue if this indeed comes instantly and not after one second?

The client doesn't even build with VS lol

Is TempWorld initilized like this in C++? or would I get an exception if I access the member m_apCharacters of it

Image attachment

C:\Users\Learath2\source\repos\ddnet\out\build\x86-Debug\cl : Command line warning D9025: overriding '/MDd' with '/MTd'
C:\Program Files (x86)\Windows Kits\10\include\10.0.20348.0\um\winnt.h(2513): error C2338: static_assert failed: 'Windows headers require the default packing option. Changing this can lead to memory corruption. This diagnostic can be disabled by building with WINDOWS_IGNORE_PACKING_MISMATCH defined.'
  [176/179] C:\PROGRA~1\MICROS~4\2022\COMMUN~1\VC\Tools\MSVC\1432~1.313\bin\Hostx86\x86\cl.exe   /TP -DCONF_DEBUG -DCONF_INFORM_UPDATE -DGAME_RELEASE_VERSION=\"15.2.5\" -DGLEW_STATIC -IC:\Users\Learath2\source\repos\ddnet\ddnet-libs\curl\include -IC:\Users\Learath2\source\repos\ddnet\ddnet-libs\freetype\include -IC:\Users\Learath2\source\repos\ddnet\src\engine\external\glew -IC:\Users\Learath2\source\repos\ddnet\ddnet-libs\opus\include -IC:\Users\Learath2\source\repos\ddnet\ddnet-libs\opus\include\opus -IC:\Users\Learath2\source\repos\ddnet\src\engine\external\pnglite -IC:\Users\Learath2\source\repos\ddnet\ddnet-libs\sdl\include\windows -IC:\Users\Learath2\source\repos\ddnet\src\engine\external\wavpack -IC:\Users\Learath2\source\repos\ddnet\out\build\x86-Debug\src -IC:\Users\Learath2\source\repos\ddnet\src -IC:\Users\Learath2\source\repos\ddnet\ddnet-libs\sqlite3\include -IC:\Users\Learath2\source\repos\ddnet\src\engine\external\zlib /DWIN32 /D_WINDOWS /GR /EHsc /Zi /Ob0 /Od /RTC1 -MDd /MTd /MP /EHsc /GS /wd4996 /wd4244 /wd4267 /wd4800 /showIncludes /FoCMakeFiles\DDNet.dir\src\engine\client\http.cpp.obj /FdCMakeFiles\DDNet.dir\ /FS -c C:\Users\Learath2\source\repos\ddnet\src\engine\client\http.cpp
  Microsoft (R) C/C++ Optimizing Compiler Version 19.32.31329 for x86
  Copyright (C) Microsoft Corporation.  All rights reserved.

heinrich5991

@c0d3d3v can you report an issue if this indeed comes instantly and not after one second?

will do it later... it is after 1 sec does it still try to connect if the message pops up? (edited)

@heinrich5991 seems to be related to your recent http rework, if I build 15xxx it's fine

did you have an empty build dir? @Ясно Понятно has build it (edited)

c0d3d3v

Is TempWorld initilized like this in C++? or would I get an exception if I access the member m_apCharacters of it

ah I have debuged it on linux... all elements of m_apCharacters are nullpointer, so I have to check that too

18:20

mh I already do :/

18:26

maybe on windows they are not nullpointer, because windows does not run the default constructor ?

windows runs default constructors

18:27

the C++ standard says that it does so

18:27

does CWorldCore have a default constructor?

so if you just write CWorldCore TempWorld; it runs this constructor ?: (edited)

Image attachment

18:29

Image attachment

18:30

Do I not have to use CWorldCore *TempWorld = new CWorldCore();

yes, you don't have to use it

is the check here sufficient to know if it is a null pointer or has been nullified with mem_zero?

Image attachment

heinrich5991

@Chairn the last two issues you closed aren't actually fixed

the commit says it fixed those

I'm really desperate, do not understand where the problem is xD

heinrich5991

does CWorldCore have a default constructor?

a lot of classes lacks constructor

18:44

would you prefer c++11 initialization or constructor?

c0d3d3v

I'm really desperate, do not understand where the problem is xD

especially why it only crashs on windows

Chairn

the commit says it fixed those

needs to be merged before its fixed

Chairn

the commit says it fixed those

the commit isn't in master

18:48

also please describe why you close issues

i will

lets play who closes issues faster

18:49

i wonder what would happen if we close all issues xd

https://tenor.com/XpB4.gif (edited)

it doesn't show anymore D:

fast af boi

i like that /tenor command

why does that even exists

to search and insert gif faster

18:51

/tenor git gud for instance

18:51

you can also use /giphy

18:52

¯\_(ツ)_/¯

since no one is able to build on windows... we should probably revert my changes with m_AttachedPlayers / SetHookedPlayer ? or @Learath2 do you still try to build?

I can try to build on windows if that's what you mean

if you can debug

yes pls (edited)

ok rebooting now

[ddnet/ddnet] Issue opened: #5226 Nightly build crashs on windows if you join a server

Reported by a lot of nightly build testers. Probably related to https://github.com/ddnet/ddnet/pull/5063/commits/f07d773492f60ca22fda41854b425aec352b7537

I tried to compile with msys2 and couldn't install all the dependencies.

what's the error message?

CMake Error at cmake/BuildVulkanShaders.cmake:25 (message):
  glslangValidator binary was not found.  Did you install the Vulkan SDK /
  packages ?
Call Stack (most recent call first):
  CMakeLists.txt:578 (include)

Use pacman -S mingw-w64-x86_64-vulkan-devel in msys2

I've already used it and installed everything that promts.

Also pacman -S mingw-w64-x86_64-glslang

20:03

I assume the glslangValidator is separate from the Vulkan SDK

Robyt3

Also pacman -S mingw-w64-x86_64-glslang

done that too

Anyone else dropping alot of frames with scoreboard open?

Mr.Gh0s7

done that too

If you type glslangvalidator into the windows search, does it find the executable?

murpi

Anyone else dropping alot of frames with scoreboard open?

sometimes when there are many players online, ye

Have you reopened the console windows after installing new packages?

murpi

Anyone else dropping alot of frames with scoreboard open?

not really when scoreboard is open but sometimes when someone random joins it lags hard for a second for me xd

Robyt3

If you type glslangvalidator into the windows search, does it find the executable?

no

Robyt3

Have you reopened the console windows after installing new packages?

yes but I haven't try reboot

Have you added the msys2 bin folder to the PATH environment variable?

No I haven't let me try.

it should be ...\msys64\mingw64\bin

20:09

with ... replaced with the path to your msys2 installation

murpi

Anyone else dropping alot of frames with scoreboard open?

On the nightly it is known, there are 3 fixes already proposed we just have to decide on which one is the most appropriate

if you have any other versions of mingw on the path you should also remove those

Learath2

On the nightly it is known, there are 3 fixes already proposed we just have to decide on which one is the most appropriate

Ah okey, I see

well glslangvalidator is part of https://packages.msys2.org/package/mingw-w64-x86_64-glslang so you have the package already

Robyt3

with ... replaced with the path to your msys2 installation

yeap done. Same error tho

Maybe check with pacman -F glslangValidator.exe

20:12

Though idk if the mingw version of pacman has the file database stuff enabled

you have to reopen console windows for changes to the path variable to take effect

yeap done that

ah, I think you also need the real Vulkan SDK

20:13

not from msys2

Learath2

Maybe check with pacman -F glslangValidator.exe

it's installed

Robyt3

ah, I think you also need the real Vulkan SDK

ok I'll try

If it's installed and you can't use it on the terminal there is some very weird issue there

My glslangValidator.exe is part of the Vulkan SDK

msys64/bin is part of the default path of the mingw console

Don't forget to add the Vulkan SDK to the path if it's not done automatically

Ok

20:21

I added C:\VulkanSDK\1.3.211.0\Bin to path and the error persists which is weird

why is it so easy for the CI to build for windows xD but not for us xD

ye xD

Did you reopen the console after changing the path?

yes

Does windows find the executable now if you search?

No I probably did something wrong? Let me reboot.

Also confirm that you have the executable in the folder

20:27

maybe you need to use the Vulkan SDK maintenancetool to install the validator separately

now it finds it let me try to run cmake .. again

btw @Robyt3 if you know the process so well xD why can you not debug it? Do you also no longer use windows?

so far no error is thrown. edit same error xD (edited)

c0d3d3v

btw @Robyt3 if you know the process so well xD why can you not debug it? Do you also no longer use windows?

I haven't looked into what you are trying to debug, but I do use windows

it is a windows only bug

20:32

https://github.com/ddnet/ddnet/issues/5226

Nightly build crashs on windows if you join a server · Issue #5226 ...

Reported by a lot of nightly build testers. Probably related to f07d773

20:33

maybe it is already fixed with https://github.com/ddnet/ddnet/pull/5222... do not know (edited)

c0d3d3v

https://github.com/ddnet/ddnet/issues/5226

I can take a closer look tomorrow

2

Mr.Gh0s7

so far no error is thrown. edit same error xD (edited)

Does C:\VulkanSDK\1.3.211.0\Bin\glslangValidator.exe exist?

yes

Somehow I got msys2 to work on my machine, but it only compiles ddnet. On compiling teeworlds I get an internal compiler error..

c0d3d3v

maybe it is already fixed with https://github.com/ddnet/ddnet/pull/5222... do not know (edited)

I can't reproduce it. Not with current master and not without 5222

can you try the CI build @Robyt3 (edited)

20:47

https://github.com/ddnet/ddnet/actions/runs/2386124910

Only record swap in teehistorian if active · ddnet/ddnet@93ece6f

DDraceNetwork, a cooperative racing mod of Teeworlds - Only record swap in teehistorian if active · ddnet/ddnet@93ece6f

ok, this crashes the LAN server when connecting to it, and it crashes the client when connecting to any internet server

xD not good

20:49

it is build using visual studio 2022 (edited)

20:49

17 enterprice

20:50

https://github.com/ddnet/ddnet/actions/runs/2372936542 this is the last without my changes

Fix pos1 only scrolling up one page in console · ddnet/ddnet@0c518ba

DDraceNetwork, a cooperative racing mod of Teeworlds - Fix pos1 only scrolling up one page in console · ddnet/ddnet@0c518ba

c0d3d3v

https://github.com/ddnet/ddnet/actions/runs/2372936542 this is the last without my changes

that one works

c0d3d3v

it is build using visual studio 2022 (edited)

I could try building with VS22 tomorrow

could it be that we have IPO active on windows builds?

20:58

mh no

Knuski

not really when scoreboard is open but sometimes when someone random joins it lags hard for a second for me xd

the client is loading the font character. I suppose it only happens when someone use special characters

Chairn

the client is loading the font character. I suppose it only happens when someone use special characters

ahh good2know

[ddnet/ddnet] Issue closed: #5135 Warnings when building + long link time

VS22 won't build master at all

I am building on windows with VS19 and I don't see any crashes, is it only on some maps?

[ddnet/ddnet] Issue opened: #5227 Initial map not found -> Server can't be killed

DDNet-Server "sv_map foasdsada"

Tater

I am building on windows with VS19 and I don't see any crashes, is it only on some maps?

no on all servers

Tater

I am building on windows with VS19 and I don't see any crashes, is it only on some maps?

can you try the builds from the CI too, that I linked up there (edited)

c0d3d3v

https://github.com/ddnet/ddnet/actions/runs/2386124910

here

ok

23:04

yeah it crashes

23:04

but only on servers where my tee gets spawned

ok thanks... if you want you can try to debug it

but we also can do that tomorrow

oh ok I can replicate the crash now with my own build

23:23

idk why it didn't work before

[ddnet:master] 5 new commits

f25a7d9 Don't try other linkers on macOS - def- 64e6ba5 Increase macOS target version to 10.13 - def- 3a231b0 Fix new Obj-C deprecation warnings - def- 604142d Try other fix for tautological - def- f7434be Merge #5219 - bors[bot]

23:44

[ddnet/ddnet] Pull request closed: #5219 macOS fixes

I can confirm reverting https://github.com/ddnet/ddnet/commit/f07d773492f60ca22fda41854b425aec352b7537 fixes it

Add a set of Players that are attached to a Tee · ddnet/ddnet@f07d773

it looks like initializing any type of std::set in gamecore.h will cause a crash when it gets deleted? (edited)

23:56

adding just this line and nothing else will cause a crash

Image attachment

can you check if the destructor is called only once?

Exported 488 message(s)