DDraceNetwork - developer

DDraceNetwork

Development / developer

Development discussion. Logged to https://ddnet.org/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 2024-09-12 00:00 and 2024-09-13 00:00

Let me know if it was good then I’ll get it too

Replying to @Ryozuki got this book home from work

00:21

Yes the keepalive is not needed when the input is being sent. It never hurts. But the way it is implemented in the official client is. Send keepalive only if nothing else was sent for too long. That’s usually happening when the client gets into a weird state. But when actively playing the input is supposed to be sent in a high frequency too.

Replying to @jxsl13 what is that supposed to do, the 10hz part? on one hand you got every 2 …

00:22

@jxsl13: why? What are you working on?

can you fix my pr xD

00:22

just looking through the client code

Oh you made a pr

wanted to see of I could connect like 100 clients to 100 servers

Ehm later gotta hop under the shower quickly

and check the chat for bot client messages

00:23

no hurry, it's 2:23 am here

sus ddoser jiggsel

Replying to @jxsl13 wanted to see of I could connect like 100 clients to 100 servers

call me Google, the Data Kraken

<ChillerDragon> bra u did a huuge pr again with 100 different things @jxsl13

always ._.

<ChillerDragon> stap it

<ChillerDragon> ctx, cancel := signal.NotifyContext(context.Background(), os.Interrupt)

00:42

<ChillerDragon> this is for ctrl+c?

I reduced it from 45 files

00:42

yes it is

<ChillerDragon> i cant beleive i did not implemented that

00:42

<ChillerDragon> @jxsl13 you can do 200 files but please split it up in prs that only do one thing makes the review easier and faster

00:43

<ChillerDragon> will also help you find your bug if there is only a smaller diff

first part is the snap storage

00:44

where error was replaced with a bool, indicating whether an item was found or not

<ChillerDragon> the graceful exit is for sure its own pr. The gitignore could also be merged instantly. Than there is the typo. And than there is lots of other stuff

let me grab my laptop

<ChillerDragon> I personally would prefer 10+ prs xd but if thats too annoying for you you can also bundle all the little things that dont actually change anything into one big pr

I will do some small prs

<ChillerDragon> your gitignore pr broke the pipeline ^

I work like this: start at one location and then prow through the whole codebase and at the end everything depends on each other and it's a giant pain to split that up xD

00:53

plow*

00:54

but that's ok

<ChillerDragon> multiple ways of doing it i use all of them

00:55

<ChillerDragon> as soon as you find something that is not tied to your initial task you either write it down for later, create a new branch and pr right now, or in the end split everything by manually committing with git add -p onto the correct branch

00:56

<ChillerDragon> yes dependencies belong together but not everything in your pr depends on each other

<ChillerDragon> I don't want to annoy you too much xd. If you want me to help you split let me know.

one pr ist ready

01:42

._.

<ChillerDragon> bro wat xd

<ChillerDragon> @zwelf the whole database failsafe and backup code relies on the fact that ddnet only does inserts right? Updates to existing records such as points are not covered by that right? This and chaning map ratings is why the recompute points script has to be run right?

02:53

<ChillerDragon> I am asking because I want to use that system to update existing records in my database and that does not seem to be support the backup stuff. So I can just if(w != Write::NORMAL) return false i guess.

[ddnet-maps:master] 1 new commit

d2232a9 A Amber - ddnet-maps

2024-09-12 12:30:30 E http: https://raw.githubusercontent.com/tee-community/FlatCity-maps/main/FlatCity_2024_09_10.map failed. libcurl error (63): Maximum file size exceeded

2024 3mb is big filesize?

[ddnet/ddnet] Pull request opened: #8934 Add client setting to restore old freeze stars

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 (especially base/) or added coverage to integration test
[ ] 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-ad...

ws-client

<ChillerDragon> @zwelf the whole database failsafe and backup code relies on the fact that ddnet only does inserts right? Updates to existing records such as points are not covered by that right? This and chaning map ratings is why the recompute points script has to be run right?

no ping moment @Zwelf

Matodor

2024-09-12 12:30:30 E http: https://raw.githubusercontent.com/tee-community/FlatCity-maps/main/FlatCity_2024_09_10.map failed. libcurl error (63): Maximum file size exceeded

2024 3mb is big filesize?

Github doesn't want you to use them as a static file hoster I guess

deen

Github doesn't want you to use them as a static file hoster I guess

it is no ddnet map limit size for HTTPS download?

08:33

hm, it's no ddnet limit

Image attachment

ws-client

<ChillerDragon> @zwelf the whole database failsafe and backup code relies on the fact that ddnet only does inserts right? Updates to existing records such as points are not covered by that right? This and chaning map ratings is why the recompute points script has to be run right?

yes, points aren't covered by the database failsafe code. But working on a solution for https://github.com/ddnet/ddnet/issues/8798#issuecomment-2319081388, that might change the model a bit.

@Robyt3 @Learath2 can you explain why this should be?

Image attachment

Wouldn't it be more correct for a tee moving through freeze because it would take into account the fractional part of the tick?

that's very complicated

09:34

also previous behavior did not do that

ig if u want old freeze stars you also have to use the it like the server did it, simply the cur tick pos xxd

there's no variable for "mix % where the tee entered freeze" you need to calculate the intersection

Tater

there's no variable for "mix % where the tee entered freeze" you need to calculate the intersection

Why? The actual true current pos of the tee on the screen doesn't depend on when the tee entered freeze

wdym

09:37

the stars get summoned and do not follow the tee afterwards

09:37

we only care about the tee position 1 time

09:38

which is taken from the most recent snapshot where the server told us that tee was frozen

Position = mix(vec2(Prev.m_X, Prev.m_Y), vec2(Player.m_X, Player.m_Y), IntraTick); this is where the tee will be rendered

09:40

(Technically this but cached and maybe predicted and smoothed, available in m_RenderPos)

Since the freeze stars, as tater said, aren't moving it almost doesn't matter anyway. in worst case it's 1/50 tick off

Learath2

Position = mix(vec2(Prev.m_X, Prev.m_Y), vec2(Player.m_X, Player.m_Y), IntraTick); this is where the tee will be rendered

Isn't this just going to be wrong?

09:41

it will not render from the position where the tee entered freeze

Well you asked me "why this should be". Not whether it makes much of a difference, or how much it's off

Tater

it will not render from the position where the tee entered freeze

Why should the stars be rendered from where you entered freeze? The stars start at your tee position even if you enter freeze and then leave it

I never said they should that was Learath

I didn't?

idk what you mean then

09:44

the lack of prediction is intentional

Tater

there's no variable for "mix % where the tee entered freeze" you need to calculate the intersection

You brought up the "where the tee entered freeze". After which I said that doesn't matter

Tater

the lack of prediction is intentional

Even the unpredicted position of the tee is mixed, it's not just the snapshot position

but if the server was summoning the stars it would never mix the positions?

09:45

what are we mixing them by?

09:45

I don't understand

That I won't argue with, this is probably a more accurate emulation of the old stars

tbh I don't understand why the air jump effect does the mixing either.

09:46

I think it's more wrong

09:47

you're using the prediction intratick to mix the non-predicted positions

09:47

ah wait nvm

Whenever I'm working on the teeworlds or ddnet source my first rule is that "If I'm questioning matricks' code, I'm likely just understanding it wrong"

Events should use the unpredicted position.

09:49

With unpredicted I mean non interpolatable, but anti-ping is ok

I'm not sure what IntraGameTick even means to the client

09:51

tee's cannot jump on intraticks. they can only jump on real ticks. so their position should be aligned to real ticks right?

Tater

tee's cannot jump on intraticks. they can only jump on real ticks. so their position should be aligned to real ticks right?

Yes

so why we are mixing the air jump position lol

The server sends a jump as event with a position

Tater

so why we are mixing the air jump position lol

Do they?

Image attachment

To me it's a visual effect so it should appear more closely with the rendered tee position, and not at the position where the event happened on the server

Is that prediction code?

Jupstar ✪

Is that prediction code?

void CGameClient::OnNewSnapshot()

Robyt3

To me it's a visual effect so it should appear more closely with the rendered tee position, and not at the position where the event happened on the server

Yes, but that can be predicted too

but it will never align to the rendered tee because these are unpredicted positions

09:53

unless you have 0ms ping

If you jump at 0.75 of a tick. then the next tick will generate the jump effect

09:54

In fact you cannot generate events intertick

09:54

bcs then you'd generate millions of events

09:54

every time you call the interpolation code

09:55

Wait i send you the holy grail of prediction

Tater

Click to see attachment 🖼️

For a very fast moving tee, this ends up creating the jumping effect actually under the tee, instead of at the end where the air jump really happened but the tee isn't rendered there locally yet because we haven't interpolated all the way to that tick yet

Well maybe that is too much off-topic for now. But anti ping can predict to the same tick as the server currently is on. So the effect is accurate

Atleast that's how I remember making sense of this code when I read it years ago

for local players

Learath2

For a very fast moving tee, this ends up creating the jumping effect actually under the tee, instead of at the end where the air jump really happened but the tee isn't rendered there locally yet because we haven't interpolated all the way to that tick yet

but these are unpredicted positions

For anti-ping players it's a problem anyway, because they jump around too much

Predicted or unpredicted doesn't matter, we are between ticks, the tee will be rendered in between the two positions in the snapshots whether you like it or not, that's just how it is

but it wont? lol

ok

09:58

I'm going to do something else

but it's not going to get rendered between the two snapshots? am I wrong?

Tater

Click to see attachment 🖼️

Are these snapshots the anti ping predicted snapshots. or the "real" server snapshots?

"real"

Then it defs makes no sense to interpolate

antiping does not create snapshots

Tater

antiping does not create snapshots

Ah ok

you're just mixing for lulz as far as I can tell

10:00

it's gonna be in the wrong spot either way

10:00

ah wait

10:00

this is specifically for if you don't have prediction

Wtf

10:01

but I still find it weird

10:01

I mean nobody uses that anyway

10:01

But still feels wrong

people do use it

10:01

it's off by default lol

10:01

possibly a majority use it? (edited)

Ok wait

10:02

let's not talk in predict. Do you mean Antiping vs non-antiping

huh

10:02

ok nvm it's on by default

10:03

I had a friend download the game a few days ago and for some reason his antiping was off when he installed it

10:03

but the client config says enabled

I think anti-ping is off or not

this one

Image attachment

but your local tee is still predicted using cl_predict 1 (edited)

MACRO_CONFIG_INT(ClAntiPingPlayers, cl_antiping_players, 1, 0, 1, CFGFLAG_CLIENT | CFGFLAG_SAVE, "Predict other player's movement more aggressively (only enabled if cl_antiping is set to 1)")

but cl_antiping?

that's just a global config for all of them

I assume that setting you posted is only on if the one i posted is on (edited)

Tater

that's just a global config for all of them

Sure?

10:04

Like global yes, but if not on, others are also not on, I assume

idk what you mean

10:05

oh wait

If cl_antiping is 0 then ALL antiping is off

it is off by default

MACRO_CONFIG_INT(ClAntiPing, cl_antiping, 0, 0, 1, CFGFLAG_CLIENT | CFGFLAG_SAVE, "Enable antiping, i. e. more aggressive prediction.")

10:06

"more aggressive" implies there's any prediction lol (edited)

Yeah ok, but your code is specifically for non-anti ping i guess

10:07

So i even less understand why they interpolate

it makes sense for non-anti ping a little bit

There is ofc prediction even without antiping. This game is unplayable without any prediction with anything above 10 ping

One thing however which is important is, that teeworlds snapshots also include events, and are only send every 2nd tick, which makes the whole code wrong anyway

10:08

in all cases

Learath2

There is ofc prediction even without antiping. This game is unplayable without any prediction with anything above 10 ping

sorry I meant to refer to the cl_antiping_players tip

MACRO_CONFIG_INT(ClAntiPingPlayers, cl_antiping_players, 1, 0, 1, CFGFLAG_CLIENT | CFGFLAG_SAVE, "Predict other player's movement more aggressively (only enabled if cl_antiping is set to 1)")

Those are also still "predicted" even if anti-ping is off

10:10

If you turn off cl_predict you'll get pretty close to how it really looks like

I'm aware of cl_predict

10:11

technically yes it always predicts their movement it just doesn't show it to you

Learath2

Those are also still "predicted" even if anti-ping is off

What, other players are not predicted if anti-ping is off, are they?

10:11

Else they'd teleport even without anti-ping

they're still in the predicted world

10:12

otherwise you would fall through them

Tater

otherwise you would fall through them

That is the client side prediction (aka cl_predict 1)

10:12

But i mean the actual other players

cl_predict 0 just shows you the server snapshots (edited)

Jupstar ✪

That is the client side prediction (aka cl_predict 1)

you can't just delete the other tees even if you use cl_antiping_players 0

Ok, other take

inline float IntraGameTick(int Conn) const { return 1.0f; } do this change, turn off cl_predict you'll see what all the mixes are doing

Without antiping other players are not predicted for future ticks. They are always very close to the server snapshots

10:13

if no new snapshot arrives maybe they are predicted for 1-2 ticks

Jupstar ✪

Without antiping other players are not predicted for future ticks. They are always very close to the server snapshots

only visually

10:14

the prediction physics keeps them in the gameworld

Yes, sure (edited)

10:14

but a very old version of them

10:14

with 300ms ping they are 300ms old

yes but their "ghosts" can still bump you in the prediction even though it doesn't show them bumping you (edited)

I can almost do Linear with 60 ping without any prediction, surprising

I think actually mixing the jump effects is correct but mixing the freeze stars is wrong

If you want to emulate the old behaviour exactly, you are right

the freeze stars should come from the position where the tee entered freeze. If you mix them they will spawn from outside the freeze

Tater

yes but their "ghosts" can still bump you in the prediction even though it doesn't show them bumping you (edited)

Yeah, but that is kinda the problem. Prediction is such a bad word. Anti-ping is also a prediction, but not the same as local player prediction which ofc has all information available of the local players directly. The non-anti ping prediction for these airjumps clearly uses snapshot prev and cur. Which are (I assume) server snapshots. So why mixing here

10:17

1. Because snaphots only arrive every 2nd tick 2. It's still wrong

10:17

The event should instead say in which tick the jump happened and there is the position where the tee was when the jump happened

10:18

or other way around. the jump effect should be an event with a position the server sends

Jupstar ✪

1. Because snaphots only arrive every 2nd tick 2. It's still wrong

In this case there is no event, that's why the hack is utilized

Learath2

In this case there is no event, that's why the hack is utilized

Can the server not send that event? Or why is it missing

sending the tick where the jump happened is wasteful for little benefit

Jupstar ✪

Can the server not send that event? Or why is it missing

It's not an event at all, the state change between the snapshot is used to trigger it. It's probably missing because matricks didn't want to waste the bandwidth, but that's a guess (edited)

no one will notice

Wtf. isn't it a core-event?

10:19

Are they not synced?

10:19

All other shit events are send too

10:19

Why not this

you have to wait 2 ticks anyway

Jupstar ✪

Are they not synced?

Why wouldn't it be synced? The snapshots provide the synchronization

10:20

The effect isn't synced if you mean that part

Yeah

10:21

I am glad i changed that

there is no NETEVENT_AIRJUMP_EFFECT

Now i also understand why twgame didn't give me an airjump event

10:21

but an air-jump sound

10:21

lmao

10:21

so the sound is completely off from visual

10:21

nice

Yes, how have you never noticed that before?

10:22

It's actually something I notice in normal gameplay with meh ping

K, but anyway

10:22

I'd send an event

10:22

That is the only logically thing to do in my head

ironically we could remove the air jump sound event and the prediction of it would be more accurate

IMHO events shouldn't be predicted in an interpolation fashion anyway

agreed

predicted in anti-ping way yes

(idk why there is an airjump sound event btw, it'd definitely be more accurate to do it along with the effect however that one gets generated)

else local players have no jump effects, that sucks too lol

lol

Learath2

(idk why there is an airjump sound event btw, it'd definitely be more accurate to do it along with the effect however that one gets generated)

tja

it would be very easy

Image attachment

Do you question matricks' decisions?

Tater

it would be very easy

That is literally why i did that

10:24

for my vanilla implementation

Jupstar ✪

Do you question matricks' decisions?

I said I'm "likely" wrong, not always wrong

Learath2

I said I'm "likely" wrong, not always wrong

lol

Learath2

I'm going to do something else

sorry for hassling you. The mixing suggestion was not incorrect.

My motivation just increased by 5%. Fixing weird design choices with backward compat breaking changes is just so epyc dev moment

Tater

sorry for hassling you. The mixing suggestion was not incorrect.

It's fine. I was more annoyed at some nasty code a coworker wrote, I just needed a second

I cba to set up a bar on this laptop yet, so I check my battery with acpi -b

Learath2

I cba to set up a bar on this laptop yet, so I check my battery with acpi -b

based

Learath2

I cba to set up a bar on this laptop yet, so I check my battery with acpi -b

on wat laptop?

<ChillerDragon> http://paste.pr0.tips/rBd checkout this epic html moment

is that... a green background?

<ChillerDragon> yes

holy

poggers

<ChillerDragon> https://www.youtube.com/shorts/kAzfu1K4_xo got it from this short xd

Coding with Lewis

The STRANGEST color on the web

‍♂️ #softwaredeveloper #programming...

MilkeeyCat

on wat laptop?

on my framework

I looove when I can just put whole words in color property and it will be like: "yep seems good, lemme just replace non hex values to 0s and pad it to next number divisible by 3"

<ChillerDragon> @MilkeeyCat most sane number parser. We should have used that for colored broadcasts in ddnet

what's the default rcon password for a server started from ingame?

jxsl13

what's the default rcon password for a server started from ingame?

randomized

rip

12:45

how imma kick my dummy

12:49

is there any way to retrieve that password :o?

github-merge-queue[bot]

[ddnet:master] 2 new commits

80e2de1 Add client side freeze stars - sjrc6 8d431f8 Merge pull request #8934 from sjrc6/pr_freeze_stars - def-

12:50

[ddnet/ddnet] Pull request closed: #8934 Add client setting to restore old freeze stars

It’s printed in the beginning of the log

Replying to @jxsl13 is there any way to retrieve that password :o?

13:10

If you are too lazy to check the log you can buy a ddnet 0day rcon exploit from me for 10 bucks (edited)

13:12

But tbh I just recommend adding a password to your config and restarting the server

13:14

Omg we have freeze stars before snowflake fade out xd

got another pr for you >:D

13:37

autoexec_server.cfg in DDNet.app, I guess

13:38

yay, it works

chillerdragon

Omg we have freeze stars before snowflake fade out xd

Snowflake fade out was hard to see tho gigachad

gigachad

GitHub

Click to see attachment 🖼️

@Aoe should be available on steam nightly in a few days gigachad

gigachad

cl_freeze_stars

<ChillerDragon> @jxsl13 yea i saw it im just a bit busy i will have a look tomorrow hopefully

k

[ddnet/ddnet] Issue opened: #8935 Floats written with commas are now considered invalid

While I think that's the correct behavior and that's also how the editor command validation validates it. There is currently three maps that need to be fixed because of it.

Cloudtide.cfg
tune_zone 4 hammer_strength 3,0

drg5.cfg
tune_zone 1 ground_control_speed 0,00

Dizzy.cfg
tune_zone 2 gravity -0,5
tune_zone 1 gravity -1,5

``` d452bcda8fce9ce294a92bb61b94192f585a3460 is the first bad commit commit d452bcda8fce9ce294a92bb61b94192f585a3460 (HEAD) Author: KebsCS Da...

1

1

1

1

== summon furo bot ===

GitHub

Click to see attachment 🖼️

How did you find the maps with invalid tunes?

meloƞ

How did you find the maps with invalid tunes?

config_retrieve then grep

That easy? Damn! ty

Assuming you know what you are looking for

13:53

Like commas in numbers I guess

13:54

Otherwise you'd have to load every map and check if any settings could not be found

[ddnet/ddnet] Issue closed: #6028 Show freeze stars instead of bar in DDRace servers

14:02

[ddnet/ddnet] Issue closed: #5527 Restore possibility to determine freeze time without freeze bar

ChillerDragon is your/any good ctf server running 0.6 or 0.7 (now that we can :B)?

[ddnet/ddnet] Issue closed: #8935 Floats written with commas are now considered invalid

<ChillerDragon> now that we can what @Mr.Gh0s7

14:43

<ChillerDragon> I can for the life of it not figure out what you are asking so i will give you a bunch of random answers @Mr.Gh0s7 there are 0.6 and 0.7 servers running and yes I also run 0.6 and 0.7 servers even bridge servers

14:45

<ChillerDragon> not sure what makes a ctf server "good" for you. But mine have shared stats between 0.6 and 0.7 that add a ranking and track kills since 6 years. If by good you mean have actual players on it that would be jsaurus bridge server that is currently the hottest CTF place to be with almost daily activity

1

https://github.com/the-nix-way/dev-templates

GitHub - the-nix-way/dev-templates: Dev environments for numerous l...

Dev environments for numerous languages based on Nix flakes [maintainer=@lucperkins] - the-nix-way/dev-templates

15:05

nix flake init --template "https://flakehub.com/f/the-nix-way/dev-templates/*#rust"

imagine needing templates and not writing them yourself, reading docs for 3 hours because you made a simple typo and nix is verbose enough that you feel like debugging your kernel

Mr.Gh0s7

ChillerDragon is your/any good ctf server running 0.6 or 0.7 (now that we can :B)?

"jsaurus" server, "tee.community" server, "mdx" server

ok ty :)

first twos are european ones afair, last one is moscow based

1

if you had a stick from earth to the moon and you pushed it, u wouldnt notice it being pushed from the moon until 21 hours pass

16:01

thanks for coming to my ted talk

Hi, can I ask for a second vanilla server for USA? The server has been full everyday.

Image attachment

16:03

Also, would it be possible to record a demo by the server, and somehow get that demo file?

the state of a server is recorded in teehistorian - regular player's can't access that tho i'm afraid

ws-client

<ChillerDragon> not sure what makes a ctf server "good" for you. But mine have shared stats between 0.6 and 0.7 that add a ranking and track kills since 6 years. If by good you mean have actual players on it that would be jsaurus bridge server that is currently the hottest CTF place to be with almost daily activity

Chillerdragon :O I missed the answers xD (now that we can :B) = now that we can connect to 0.7 server :) ty for the info

[ddnet/ddnet] Issue opened: #8936 Skin filter does not apply to 0.7 skin parts

It only applies to 0.7 skins but not to the skin parts lists at the moment.

There is sv_auto_demo_record and sv_player_demo_record, the first one records one big demo and the other one records one for each finish.

Replying to @Pathos Also, would it be possible to record a demo by the server, and somehow g…

https://github.com/sanctuary-js/sanctuary-def i dunno what this is but looks pretty cool

GitHub - sanctuary-js/sanctuary-def: Run-time type system for JavaS...

Run-time type system for JavaScript. Contribute to sanctuary-js/sanctuary-def development by creating an account on GitHub.

is it possible to run both client and server connected to visual studio debugger?

19:11

or at least build both at once

u can build "BUILD_ALL" or smth like that

kebs

is it possible to run both client and server connected to visual studio debugger?

CLion

i prefer vs

kebs

i prefer vs

i doubt vs can do this

kebs

i prefer vs

kek

Well then, debug one by one

1

zhn

i doubt vs can do this

found this 2009 so answer xd https://stackoverflow.com/questions/447840/debug-multiple-copies-of-a-program-from-one-visual-studio-instance

Debug multiple copies of a program from one Visual Studio instance

I have a pre-alpha GUI program that I'm dogfooding and want to run under the debugger (for when things go wrong ;), but I don't want to have to launch a new copy of Visual Studio for each instance ...

i think it has smth to do with cmake since its debug targets

kebs

i think it has smth to do with cmake since its debug targets

nop, cmake is just making vs projects for u to build

zhn

nop, cmake is just making vs projects for u to build

how does it know what can it build

in CLion you can run any number of configurations at the same time

Image attachment

Image attachment

Matodor

in CLion you can run any number of configurations at the same time

* in any jetbrains products, like phpstorm, rider (edited)

kebs

how does it know what can it build

In software development, CMake is cross-platform free and open-source software for build automation, testing, packaging and installation of software by using a compiler-independent method. CMake is not a build system itself; it generates another system's build files.

19:22

so its just generator of build files, what you would do manually with make for example, or fancy .sh/.bat script

19:23

afair bam builds without any generator justatest

justatest

zhn

so its just generator of build files, what you would do manually with make for example, or fancy .sh/.bat script

yes thats true it makes the build targets i can choose

19:23

so i guess it could make client+server at once to run

kebs

so i guess it could make client+server at once to run

cmake has nothing to do with visual studio environment i mean

zhn

cmake has nothing to do with visual studio environment i mean

doesnt it make those?

Image attachment

it makes targets, vs let u to run em

19:25

lets*

yes thats what i mean

19:25

so a target that is client+server

zhn

found this 2009 so answer xd https://stackoverflow.com/questions/447840/debug-multiple-copies-of-a-program-from-one-visual-studio-instance

from what i read from there you can run another instance by pressing button in some context menu

idk how to do that

19:27

in cmake project

cant find anything in pure ddnet folder opened in vs too

19:37

i remember using cmake-gui though

19:38

it literally did everything as vs project so you could choose many to run (afair)

Ui makes me sick

i think its not possible in vs from what i saw on stack overflow

19:55

maybe new cmake target but idk how to do it

lol stars don't get destroyed

kebs

so a target that is client+server

Why would you need that :o

20:19

cmake --build . -j -t game-client -t Game-Server (or whatever they are called in the cmake file I forgot)

meloƞ

cmake --build . -j -t game-client -t Game-Server (or whatever they are called in the cmake file I forgot)

he wants to attach debugger

On both simultaneously?

yup

Hmm, yeah that sounds like a workaround is needed

20:24

Isn't the server attached to the client if launched through it? Or is it just another executable running alongside it

20:27

https://developercommunity.visualstudio.com/t/CMake-Project-Multi-target-Debugging/916069 Known and wanted suggestion tho it seems

Developer Community

Developer Community

Matodor

lol stars don't get destroyed

in latest master ?

20:30

then report it to github and tag tater

20:30

even tho i am not sure why his change should be related with it xd

GitHub

Click to see attachment 🖼️

@Tater @Learath2 I think the server should have a gameinfo flag like "allow hook coll" or "allow x skins", so it can kinda control whether or not to show the stars

It's tied fo freeze bar info

furo

There is sv_auto_demo_record and sv_player_demo_record, the first one records one big demo and the other one records one for each finish.

Right, but I can't get the demo from the vanilla server, right?

Pathos

Hi, can I ask for a second vanilla server for USA? The server has been full everyday.

Actually, who should I contact about this?

meloƞ

Isn't the server attached to the client if launched through it? Or is it just another executable running alongside it

runs separately from client

Like as a normal player? No, you would need access to the server.

Replying to @Pathos Right, but I can't get the demo from the vanilla server, right?

kebs

is it possible to run both client and server connected to visual studio debugger?

is opening 2 instances of VS not functional enough for you?

Tater

is opening 2 instances of VS not functional enough for you?

doing it rn but thought there might be smth better (edited)

Exported 340 message(s)

Timezone: UTC+0