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 2023-02-04 00:00:00Z and 2023-02-05 00:00:00Z

@heinrich5991 yeah, in the character core, but i dont know how to get the authed state of another player from it

[ddnet/ddnet] Issue opened: #6326 Add "add clan" button

Add clan to find clanmates, like 0.7 !likes

05:04

[ddnet/ddnet] Pull request closed: #6283 update scoreboard server info

https://github.com/orlp/glidesort

GitHub - orlp/glidesort: A Rust implementation of Glidesort, my sta...

A Rust implementation of Glidesort, my stable adaptive quicksort/mergesort hybrid sorting algorithm. - GitHub - orlp/glidesort: A Rust implementation of Glidesort, my stable adaptive quicksort/mer...

newlesstee

@heinrich5991 yeah, in the character core, but i dont know how to get the authed state of another player from it

maybe you could add something in the character

08:58

that you set automatically when a player authes

i forgot, why are 0.7 skins brighter than 0.6 ones?

i think only darker

Im trying out reply to message from irc o.O @Learath2 https://github.com/TeeworldsDB/irclogs/blob/1d3a714c6cab8b7937bbf6cf0625ebde59b8e55d/ddnet/2023-01-21.log#L24-L26

irclogs/2023-01-21.log at 1d3a714c6cab8b7937bbf6cf0625ebde59b8e55d ...

Mirror of https://ddnet.tw/irclogs/. Contribute to TeeworldsDB/irclogs development by creating an account on GitHub.

09:37

i still dont get Obj_SevenToSix

09:37

last time i asked for message and u said its binary complete so what about objects then?

09:38

oh wait i think objects are how i thought message work

09:38

it translates the id not states if they are binary complete

09:39

no never mind im confused

09:40

both sixtoseven and seventosix map 4 to 4

09:40

Embedded image

09:40

https://chillerdragon.github.io/teeworlds-protocol/06/snap_items.html#obj_pickup

09:40

https://chillerdragon.github.io/teeworlds-protocol/07/snap_items.html#obj_pickup

09:40

but in 0.7 the m_Subtype got removed so they are not binary complete?

09:41

s/binary complete/binary compatible/

Idk if if applies to objs. I said that about messages

ChillerDragon

but in 0.7 the m_Subtype got removed so they are not binary complete?

This shouldn't be an issue since it was removed from the end I guess

[ddnet/ddnet] Pull request opened: #6327 Fix memory leak of editor preview image texture and data

The texture of the previous preview image was not unloaded before loading the texture of the new image. The PNG data was never freed either. It's not necessary to keep this data loaded after loading the texture.

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 poss...

oh i see

@heinrich5991 Can you merge #6282? It's been ready for a while. I tested that the selection is kept based on the filename when files are added/removed by refreshing.

https://github.com/ddnet/ddnet/pull/6282

Add a refresh button to the editor file browser by Robyt3 · Pull Re...

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 (especially base/) or add...

[ddnet:master] 2 new commits

624f7df Fix memory leak of editor preview image texture and data - Robyt3 b6f49bd Merge #6327 - bors[bot]

11:09

[ddnet/ddnet] Pull request closed: #6327 Fix memory leak of editor preview image texture and data

@Learath2 iirc last time i asked u about how to translate snap items for the 0.7 client bridge you said as early as possible with changing as little as possible of the existing code

11:15

im not really sure what the smartest way would be ._.

11:15

changing the actual snapshot data when the items are unpacked seems impossible

11:16

preprocessing the data and let the 0.6 code handle it is also kinda tricky because it uses pointers into the actual snap data

11:16

so when the snap data is 0.7 and the client is pointing 0.6 structs with different shapes in there it wont work

11:17

so i would kinda need to store the memory in some seperate storage which also seems weird

I guess it needs to be translated during unpacking

which unpacking?

11:17

snapshot.cpp or gameclient.cpp ?

Right after delta unpacking you'd have to go through the snapshot and figure out what needs changing

11:18

I honestly don't know

11:18

I didn't care enough back then. I care even less now

So you suggest that as soon as the type id is known like here https://github.com/ddnet/ddnet/blob/b6f49bd670c3baacc93081b00e85fe1e5eb9f84a/src/engine/shared/snapshot.cpp#L408

ddnet/snapshot.cpp at b6f49bd670c3baacc93081b00e85fe1e5eb9f84a · dd...

DDraceNetwork, a free cooperative platformer game. Contribute to ddnet/ddnet development by creating an account on GitHub.

11:19

one starts to look at all the fields and translate them?

11:20

if the snap item changes size where would that data be stored?

I don't see any other way to keep the 0.6 code working

would it be that bad to break the 0.6 code with a bunch of sixup if statements?

But you can't really do it there

11:20

You'd have to do it in client.cpp

can't you just™ hand the snapshot to some code that translates it to a 0.6-style snapshot? (edited)

so the whole snap data buffer thingy?

input: 0.7-style snapshot, output: 0.6-style snapshot

after all data is received?

after a snapshto is completely received

and then possibly resize it

and reconstructed

i see sounds smooth in theory

create a whole new one, just throw away the 0.7 one

ChillerDragon

would it be that bad to break the 0.6 code with a bunch of sixup if statements?

That's the last resort option. I don't want branches everywhere for a feature that will if we are being honest be used by 5 people at most

(keep it around for deltas)

that seems a bit hard to code imo tho

11:23

it will take me forever and grow the git diff even more

11:23

what are the chances of the current state being merged :)

[ddnet/ddnet] Pull request opened: #6328 Remove `bytes_be_to_int` and `int_to_bytes_be`, static assert size of `int` and `unsigned`, refactoring

Supersedes #6263.

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/dd...

yo @heinrich5991 u mean here right? https://github.com/ddnet/ddnet/blob/b6f49bd670c3baacc93081b00e85fe1e5eb9f84a/src/engine/client/client.cpp#L2200

ddnet/client.cpp at b6f49bd670c3baacc93081b00e85fe1e5eb9f84a · ddne...

DDraceNetwork, a free cooperative platformer game. Contribute to ddnet/ddnet development by creating an account on GitHub.

11:29

at the end of else if(Msg == NETMSG_SNAP || Msg == NETMSG_SNAPSINGLE || Msg == NETMSG_SNAPEMPTY)

11:29

going over all received data and modifying it

yea, basically

ok i get the idea sadly i dont even understand half of the snap unpacking code

don't touch the snap unpacking

11:33

you have a complete snap

11:33

you know how to read it (see the client code)

yea yea i know

you know how to write a new one (see server code)

11:33

anyway

11:34

it'd probably be better to talk about whether it's getting merged

but to write i kinda need to understand

11:34

ok yea

11:34

lets talk

I'm not opposed, but I don't like that it adds special cases and new parameters

11:34

and I'm gone for a couple of seconds

well it should only be a temporary solution until 0.6 is dropped

[ddnet:master] 2 new commits

aa23873 Add a refresh button to the editor file browser - Robyt3 911bd0e Merge #6282 - bors[bot]

11:36

[ddnet/ddnet] Pull request closed: #6282 Add a refresh button to the editor file browser

Im having a hard time understanding where snap data is actually stored i assume its here but it doesnt make much sense to me https://github.com/ddnet/ddnet/blob/911bd0e69a8fee9b02f0cea26d50ecc2f000bc3c/src/engine/shared/snapshot.h#L145

ddnet/snapshot.h at 911bd0e69a8fee9b02f0cea26d50ecc2f000bc3c · ddne...

DDraceNetwork, a free cooperative platformer game. Contribute to ddnet/ddnet development by creating an account on GitHub.

11:58

in snapshot.cpp it seems to store data into pItem->Data() which basically boils down to this + offset in CSnapshot

11:58

how can you store data in this ? if it this is referring to an instance of a class without any instance variables?

11:58

who allocated the memory?

11:59

it has the CSnapshotBuilder friend class but can one simply write into the instance variables of a friend class by memcopying into this + offset ?!

ChillerDragon

well it should only be a temporary solution until 0.6 is dropped

Problem is no one except you wants to drop 0.6 so in practice we'll have those branches forever

wants as in needs 0.6 or wants as in wants to go through the effort of removing the code?

12:18

i do not mind keeping 0.6 nor adding 0.5 support but if 0.6 and 0.7 in the code together make it too complex to maintain it makes sense to remove 0.6 right?

12:19

the end user should ideally not notice much of a change

12:20

oh i think i read the memcopy wrong its not writing into this but more from this yea pretty sure i still dont get it :D

Honestly, if 0.6 and 0.7 in the code together is complex to maintain, it's 0.7 that should go. Why would we even add it in the first place if it makes the code annoying to maintain?

to have a smooth transition with a time of backwards compatibility

12:33

if 0.7 is not there we are not compatible with upstream :c

Why transition? 0.7 is dead again, future is whatever we come up with for the protocol update, whether it be DTLS or QUIC

12:35

The only reason to have 0.7 compatibility would be if it wasn't a maintainence hassle

yea it has a small player base but i can see it being bigger if ddnet clients would support it

Honestly, it's small playerbase should just drop it and play 0.6 with us. Are there even mods unique to 0.7 getting any players? It's all bridge servers

its super annoying to switch clients imo

chiller just use ddnet

i do

the only good thing 0.7 has is skins

and as result i havent played ctf in 3 years

we just need to find a way to adapt to our client

12:36

without fucking up the tee

12:36

so it doesnt look ugly

12:37

pvp is jus not famous

ChillerDragon

and as result i havent played ctf in 3 years

Not like there were many playing ctf in 0.6 either. Last I tried I couldn't find a populated server

12:37

0.7*

yea because most users are too lazy to switch clients

12:37

or didnt even notice that 0.7 released

12:38

or think ddnet is 0.7

ChillerDragon

yea because most users are too lazy to switch clients

If there was a genuine desire to play ctf, they would play ctf on 0.6 no?

maybe instagib is the last good pvp

12:38

and fng

The truth is that pvp is dead except for fng. I honestly think your efforts would be better spent trying to get it popular again

getting ctf popular in 0.6?

12:39

that sounds wrong doesnt it

But anyway, if it's not creating too big a mess, I'm okay with having 0.7 compatibility aswell

yea im trying heinrichs approach now

ChillerDragon

that sounds wrong doesnt it

Not really. At least not to me

so you are saying ddnet client should promote ctf servers instead of kobra novice for new players?

12:41

the noob ctf5 sceene fully moved to 0.7

12:41

and even some of the pro scene is vanilla maximalists i think

12:41

or they quit ctf due to not wanting to switch clients

We could get some vanilla servers again, create a pvp tab for pvp servers, there are possible things

12:42

0.7 support in ddnet wont really start promoting ctf servers either

the servers are there but they are usually only played by users with vanilla clients

12:42

oh tru your probably right

So I have a feeling you'll get pretty disappointed when you get this 0.7 support in but there are no extra players in pvp

no i wont

12:43

only thing i want is one client to play on all servers for my self

Well if that's the case then go ahead

i love how heinrich said lets talk if it gets merged and then rage quitted the convo xd

12:44

yo lerato send brain

12:44

Embedded image

12:44

how can the highlighted line crash on a nullptr?

12:44

Embedded image

UB?

looks like hm

12:45

asan it iss i guess

12:45

or more like ubsan

12:49

omg asan soo good :)

12:49

it was the very first line where i indexed the nullptr

1

Learath2

Honestly, it's small playerbase should just drop it and play 0.6 with us. Are there even mods unique to 0.7 getting any players? It's all bridge servers

they could be non-bridge but ddnet's still protocol exclusive client

sorry ChillerDragon

13:54

I'm at fosdem ^^

ChillerDragon

Im having a hard time understanding where snap data is actually stored i assume its here but it doesnt make much sense to me https://github.com/ddnet/ddnet/blob/911bd0e69a8fee9b02f0cea26d50ecc2f000bc3c/src/engine/shared/snapshot.h#L145

it's stored behind the snapshot class

13:58

check how the server builds a snapshot I guess ^^

when is ddnet session at fosdem?

wat

I didn't get a stand at Fosdem, so we don't have an official session. But let's do a small meetup if you're also here

o nice heinrich

18:04

watever behind a class means xd

18:05

who is allocating that? the stack?

18:05

infinite stack space behind all classes? xd

it is possible to add delay(ping)? on rus i have 8 on ger i have 48 on pol 22 i wanna have on all server 48

Hm, don't think we have that, that'd actually be a nice feature

20:08

Have you tried antiping? That should feel the same on all those pings

ComebackPlay

it is possible to add delay(ping)? on rus i have 8 on ger i have 48 on pol 22 i wanna have on all server 48

people want to reduce ping and you want to increase it. what the f?

He wants it consistent across all servers with ping lower than 48, which is reasonable

20:31

but yeah, antiping is the way to go

murpi

but yeah, antiping is the way to go

antiping shit

Learath2

Have you tried antiping? That should feel the same on all those pings

ye i did. i dont like it

ComebackPlay

antiping shit

anti-antiping

ComebackPlay

it is possible to add delay(ping)? on rus i have 8 on ger i have 48 on pol 22 i wanna have on all server 48

I have requested this idea a few times but no one commented about it https://discord.com/channels/252358080522747904/293493549758939136/1061925747310219284

Iza

I think I requested this already a long time ago, but is there a way to add an artificial delay to anti ping? Since alot of people who don't use anti ping get used to compensating the ping delay on their main server, so that when they turn on anti ping it messes with their mental ping compensation and it feels too responsive in a way. You can especially feel it when you need to catch a falling tee (edited)

my main reason is hammer hit through wall. i always miss hammer when switch servers. when i play on rus i always try add brain delay, is so annoying always thinks about it. (edited)

Iza

I have requested this idea a few times but no one commented about it https://discord.com/channels/252358080522747904/293493549758939136/1061925747310219284

I think I already answered that. I did that with a custom written proxy that ran on my VPs. It worked quite well for cable internet ping jitter. But even with a delay of 200ms it still sometimes gave me lags, BCS cable internet just sucks so hard

22:43

So it cannot compensate real lags

22:43

And these are even harder then

22:43

BCS more delay

Jupstar ✪

I think I already answered that. I did that with a custom written proxy that ran on my VPs. It worked quite well for cable internet ping jitter. But even with a delay of 200ms it still sometimes gave me lags, BCS cable internet just sucks so hard

i'm not quite getting your point right now, i was talking about adding an artificial delay

Iza

i'm not quite getting your point right now, i was talking about adding an artificial delay

Well delay= execute it later

22:44

If u just want to have higher ping that's ez af

but can't there be a build-in feature for antiping where you have a delay slider? normally antiping tries to give you 0ms delay, but people like me can only play with 25-50ms ping

22:46

under 25ms is either too fast or over 50ms is too slow

22:46

im sure it is possible but i don't know how to code

Isn't there a prediction margin or smth

hmm i have never tried that

22:48

can it work in combination with anti ping?

22:49

antiping = 0ms + your given prediction margin = your desired delay

I assume so

22:50

I think it's goal is to add delay to smooth out prediction, but it will maybe also do the goal u want

alright i will try that

did something change with the new version regarding ghosts?

Jupstar ✪

I think it's goal is to add delay to smooth out prediction, but it will maybe also do the goal u want

it doesn't work like that, prediction margin adds delay to your already existing ping delay

23:47

and anti ping will try to compensate that

23:47

instead of the other way around, that it adds delay to anti ping itself like i wanted to

Exported 189 message(s)