DDraceNetwork - developer

[ddnet/ddnet] Pull request opened: #9211 Implement bans_find command.

This rcon command simplifies searching for all ban records for a specific IP if the ban list has too many records. The motivation behind comes from Blockworlds as we have around 900 ban records and searching for a ban record without having access to the bans.cfg can be quite annoying. I guess the way to search for records and print the index can be optimized but didn't find how for now. I'm open to suggestions. The corresponding issue for this pull request is #9202.

Checklist

...

<ChillerDragon> @MilkeeyCat I did not get my C lib into a usuable state yet -.- https://github.com/ChillerDragon/teeworlds/tree/chiller_sans_io_netcode

GitHub - ChillerDragon/teeworlds at chiller_sans_io_netcode

A retro multiplayer shooter. Contribute to ChillerDragon/teeworlds development by creating an account on GitHub.

ChillerDragon

03:41

Maybe I'll translate something else

@jxsl13 Can you help me?

Hey developers, I remember we talked about correcting the written Arabic in game chat will take huge space (or idk what was the problem) So if its really hard to make it written correct by making the letter connected like مرحبا and written from right to left.. At least just make it written in the chat from right to left like م ر ح ب ا Rather than ا ب ح ر م So arabic players can read easier than flipping head or brain (edited)

04:36

Its like olleh rather than hello !!!!

04:36

.

this sounds like a huge oversight tbh

05:03

all that IME work but rtl doesn't work

<ChillerDragon> @Evelyn nice

def-

[ddnet-web:master] 1 new commit

cd9c038 Add 17 € DDNet DB funding by 番茄炒蛋不 - def-

Go , I want to translate

github-merge-queue[bot]

[ddnet:master] 2 new commits

82b274c Implement bans_find command. - Rei-Tw b29723d Merge pull request #9211 from Rei-Tw/bans-find - def-

07:11

def-

[ddnet/ddnet] Pull request closed: #9211 Implement bans_find command.

07:14

Pioooooo

[ddnet/ddnet] Issue opened: #9212 Last input before spectate is dropped

I've enabled show local player key presses and if you press D and release it right after you spectate, the client thinks you released D but the server doesn't. This also applies to aim direction and possibly other stuff. This video is too big. Try again with a file size less than 10MB.

Evelyn

@jxsl13 Can you help me?

with what?

Palestinian

Hey developers, I remember we talked about correcting the written Arabic in game chat will take huge space (or idk what was the problem) So if its really hard to make it written correct by making the letter connected like مرحبا and written from right to left.. At least just make it written in the chat from right to left like م ر ح ب ا Rather than ا ب ح ر م So arabic players can read easier than flipping head or brain (edited)

@Learath2 is that possible?

jxsl13

with what?

I want to send commands in a separate discord server, and for them to be executed in rcon on the server

https://github.com/jxsl13/TeeworldsEconDiscordModerationBot

GitHub - jxsl13/TeeworldsEconDiscordModerationBot: small tool to mo...

small tool to moderate multiple zCatch servers via econ and discord. - jxsl13/TeeworldsEconDiscordModerationBot

07:35

you can use this one for that, tho it's a little bit old

07:38

and you probably don't need the chat features

github-merge-queue[bot]

[ddnet:master] 2 new commits

06fcfe0 Improve map sound validation and performance - Robyt3 fb7d8ab Merge pull request #9210 from Robyt3/Client-Mapsounds-Validation - def-

07:45

def-

[ddnet/ddnet] Pull request closed: #9210 Improve map sound validation and performance

@ChillerDragon.* We need to translate this

<ChillerDragon> yes

<ChillerDragon> @Evelyn looks a bit like you are trying to copy the TS servers -.- you even set the gametype to S-DDR thats a bit weird bro

09:39

<ChillerDragon> but yea for the translations put them here in the format i used. So I have the full english and russian text (as text) not images. And I can see if i find time to do it. https://github.com/DDNetPP/DDNetPP/issues/new

Build software better, together

GitHub is where people build software. More than 100 million people use GitHub to discover, fork, and contribute to over 420 million projects.

This is still such a map on the server, soon I will add a new author's map Copy Love box

Evelyn

@ChillerDragon.* We need to translate this

Kog train on ddnetpp, lol

Mʎɹ シ

Kog train on ddnetpp, lol

Запомни

3

09:54

А не не запоминай

Unstable, RB, kio killer kek

Mʎɹ シ

Kog train on ddnetpp, lol

? xd

Evelyn

Запомни

Do you mind please following the general rule to either: translate your message to English, especially in topic related channels such as #developer or go to #off-topic ? We enforce this rule so everyone is able to understand what someone is saying. - deepl is really good and then everyone is able to immediately help if capable then.

10:17

Oh and ofc:

1

Sry

10:26

These messages are mine without meaning

Avolicious

? xd

Map on his ss: "KoGTrain"

ChillerDragon Is it possible to put a color to your server mod, like there DDRace , S-DDR, gores have a color, but I want to my mod

3

MariaDB [teeworlds]> SELECT COUNT(*) FROM record_race;
+----------+
| COUNT(*) |
+----------+
| 44112090 |
+----------+
1 row in set (4 min 48.738 sec)

Row counts are surprisingly expensive

Learath2

MariaDB [teeworlds]> SELECT COUNT(*) FROM record_race;
+----------+
| COUNT(*) |
+----------+
| 44112090 |
+----------+
1 row in set (4 min 48.738 sec)

Row counts are surprisingly expensive

You can use EXPLAIN EXPLAIN SELECT COUNT(*) FROM record_race;

12:54

What does it tell you?

or more directly: SELECT table_rows FROM information_schema.tables WHERE table_name = 'record_race';

Learath2

or more directly: SELECT table_rows FROM information_schema.tables WHERE table_name = 'record_race';

But then its not 100% accurate, no?

yeah

12:55

Oh you were wondering what explain says to see why so slow?

12:55

It's 100% going to be a full scan

Mhmm, but there are some indexes, right?

Yeah but you still need to scan the entire index iirc

but 44 mio not "that big"

let me check, now I'm curious what it does

12:57

MariaDB [teeworlds]> EXPLAIN SELECT COUNT(*) FROM record_race;
+------+-------------+-------------+-------+---------------+-----------+---------+------+----------+-------------+
| id   | select_type | table       | type  | possible_keys | key       | key_len | ref  | rows     | Extra       |
+------+-------------+-------------+-------+---------------+-----------+---------+------+----------+-------------+
|    1 | SIMPLE      | record_race | index | NULL          | Timestamp | 4       | NULL | 27346608 | Using index |
+------+-------------+-------------+-------+---------------+-----------+---------+------+----------+-------------+

mhmmm

12:58

I am bit surprised. InnoDB?

I thought you wanted me to check the rows number there, it's from information schema

Avolicious

I am bit surprised. InnoDB?

Yes

NVMe?

Hm, we've been changing around db servers a lot lately, idk what this one is tbh

Learath2

Hm, we've been changing around db servers a lot lately, idk what this one is tbh

I see

Yeah nvme, checked

13:01

Actually, querying on the db server directly is much faster, I forgot we had a replica on the main server

13:01

MariaDB [teeworlds]> SELECT COUNT(*) FROM record_race;
+----------+
| COUNT(*) |
+----------+
| 44117349 |
+----------+
1 row in set (9.443 sec)

is sudo apt update a bad command to run in a github action?

kinda yes

13:59

u shouldnt need it

13:59

but depends

13:59

what u doing?

Ryozuki

what u doing?

I want to test my rust workflow locally so I use https://github.com/nektos/act the ffmpeg-sys-next crate needs some dependencies to build but act doesnt install them unless I run apt update first

- name: Install dependencies
  run: sudo apt update && sudo apt install libavutil-dev libavformat-dev libavfilter-dev libavdevice-dev

(edited)

ah then do

14:12

@zaer1n iirc there is a cache action for apt

14:12

https://github.com/actions/cache

GitHub - actions/cache: Cache dependencies and build outputs in Git...

Cache dependencies and build outputs in GitHub Actions - actions/cache

14:13

https://github.com/marketplace/actions/cache-apt-packages

Ryozuki

@zaer1n iirc there is a cache action for apt

But apt update wil still take time even if I cache the dependencies no?

@zaer1n im fairly sure the action runs apt update by itself

Ryozuki

https://github.com/marketplace/actions/cache-apt-packages

use this one

14:16

i think it caches that too

14:16

anyway its definitly a time save

Ryozuki

@zaer1n im fairly sure the action runs apt update by itself

it didnt

<ChillerDragon> @Evelyn the server browser colors are client side

is this a me issue?

Thread 83 "DDNet" received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0x7ffe7ffff6c0 (LWP 15478)]
0x00007ffff3d87f46 in __memmove_avx512_unaligned_erms ()
   from /nix/store/sl141d1g77wvhr050ah87lcyz2czdxa3-glibc-2.40-36/lib/libc.so.6

happens on ddnet master and steam beta, didnt happen before

❯ ldd --version
ldd (GNU libc) 2.40

15:16

no DDNet crash log aswell - weird

ChillerDragon Is online thanks to your divine coding

https://x.com/its_bvisness/status/1853962132091449656?s=12&t=V2tQHBvGf6372bITe1EJTg

Ben Visness (@its_bvisness) on X

I’m pleased to report that as of today, the Memory64 proposal for WebAssembly is now standard (phase 4)! Finally we will all be able to use 64-bit pointers in WASM apps. This has been a very long time coming (too long!) but I’m very glad to get here. https://t\.co/N3qbj9CxiY

16:19

@Jupstar ✪

@Learath2 sadly u didn't answer and the post goes up

Palestinian

@Learath2 sadly u didn't answer and the post goes up

What didn't I answer? I may have missed a ping

Palestinian

Hey developers, I remember we talked about correcting the written Arabic in game chat will take huge space (or idk what was the problem) So if its really hard to make it written correct by making the letter connected like مرحبا and written from right to left.. At least just make it written in the chat from right to left like م ر ح ب ا Rather than ا ب ح ر م So arabic players can read easier than flipping head or brain (edited)

Ah this one, hm

1

16:31

hmm, this is not very easy still, we need to add support for ALM/RLM or bidi in general because we need something to mark the region in which we reverse the text

16:33

@Jupstar ✪ && @Robyt3 you two know the text rendering better, do you think this would be possible? The client would need to start inserting the correct markers in stuff like chat messages (idk what this would be based on, maybe unicode prescribes when to emit these), and the text rendering code would (I presume) need a bool on text containers to keep track of which direction to render them in

16:36

The bidi algorithm in general is annoying enough to implement that we probably should use fribidi for it

$demos

This command was not found.

Learath2

@Jupstar ✪ && @Robyt3 you two know the text rendering better, do you think this would be possible? The client would need to start inserting the correct markers in stuff like chat messages (idk what this would be based on, maybe unicode prescribes when to emit these), and the text rendering code would (I presume) need a bool on text containers to keep track of which direction to render them in

It's not impossible, but it seems rather challenging. I guess using a combination of harfbuzz+fribidi would make sense, then we don't need to manually insert markers as far as I understand it. Implementing this without it having a major performance impact on all text rendering would especially be my concern. I found this example https://github.com/UnickSoft/ex-sdl-freetype-harfbuzz-fribidi/blob/24e7bdeb08ca50100a03864db6829c193d40fa90/ex-sdl-freetype-harfbuzz.cpp and it's full of allocations that would happen every frame.

Robyt3

It's not impossible, but it seems rather challenging. I guess using a combination of harfbuzz+fribidi would make sense, then we don't need to manually insert markers as far as I understand it. Implementing this without it having a major performance impact on all text rendering would especially be my concern. I found this example https://github.com/UnickSoft/ex-sdl-freetype-harfbuzz-fribidi/blob/24e7bdeb08ca50100a03864db6829c193d40fa90/ex-sdl-freetype-harfbuzz.cpp and it's full of allocations that would happen every frame.

If we use harfbuzz, then we'd get full proper shaping anyway. I was hoping we just use fribidi to get us a properly reordered string array and just render that

I manage to put full screen images in console

-StormAx

I manage to put full screen images in console

cool! - how much does it affect perfoemance greenthing

same as creepy bubble texture by default ddnet CoolSkull

meloƞ

cool! - how much does it affect perfoemance greenthing

it's one textured quad my dude

risu

it's one textured quad my dude

you havent seen what i have

meloƞ

you havent seen what i have

go for it

no i rather forget

tell me at dm at least

I think it's more that they're trying to avoid reliving the trauma

xD

@meloƞ btw i become a light side dev and sharing source to every1

meloƞ

Click to see attachment 🖼️

ARL FK OF

1

-StormAx

@meloƞ btw i become a light side dev and sharing source to every1

https://discord.com/channels/1010144732921475112/1301880609840300102/1303796330048782519

:o

Robyt3

It's not impossible, but it seems rather challenging. I guess using a combination of harfbuzz+fribidi would make sense, then we don't need to manually insert markers as far as I understand it. Implementing this without it having a major performance impact on all text rendering would especially be my concern. I found this example https://github.com/UnickSoft/ex-sdl-freetype-harfbuzz-fribidi/blob/24e7bdeb08ca50100a03864db6829c193d40fa90/ex-sdl-freetype-harfbuzz.cpp and it's full of allocations that would happen every frame.

it sounds like a job for pango tbh

fribidi is not that hard to use

19:32

harfbuzz is more tricky

Ewan

it sounds like a job for pango tbh

Pango depends on 2.x series of the GLib library

We probably don't want to add a dependency on GLib for this.

meloƞ

is this a me issue?

Thread 83 "DDNet" received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0x7ffe7ffff6c0 (LWP 15478)]
0x00007ffff3d87f46 in __memmove_avx512_unaligned_erms ()
   from /nix/store/sl141d1g77wvhr050ah87lcyz2czdxa3-glibc-2.40-36/lib/libc.so.6

happens on ddnet master and steam beta, didnt happen before

❯ ldd --version
ldd (GNU libc) 2.40

update system, happened to me once i updated kernel and didn't reload keke

zhn

update system, happened to me once i updated kernel and didn't reload keke

okey :c

Evelyn

ChillerDragon Is it possible to put a color to your server mod, like there DDRace , S-DDR, gores have a color, but I want to my mod

its time for you to start learning programming at this point

Robyt3

Pango depends on 2.x series of the GLib library

We probably don't want to add a dependency on GLib for this.

i can see why not

-StormAx

@meloƞ btw i become a light side dev and sharing source to every1

Ehrenmann

Robyt3

[ddnet/ddnet] Pull request opened: #9213 Improve map image validation

Add check for map image data being loaded from map to avoid null pointer access.
Add check for size of map image data to avoid out-of-bounds reads.
Add check for map layer item not being found to avoid null pointer access.
Fix inconsistent handling of layers with invalid images. In some cases only -1 was considered non-textured whereas all invalid values should be.
Ensure correct texture is always used for game layers instead of considering the image property of game layers wh...

Hey, just shooting my shot in the dark here. Would someone, somehow, still have the datas for the Basket gamemod ? (edited)

def-

[ddnet-web:master] 1 new commit

4b7ef9a Remove old CHN servers - def-

@deen tested with a full banlist on a small vps, doesn't seem to cause any lags for now and the command output is pretty fast (bans_find). Was able to match 1800 records out of 3000? (idk the max size of the ban list)

bug?

22:38

when local player enters a tune zone it sets same settings for everyone (edited)

22:38

when anyone else is in it it works fine

22:39

im testing using m_pClient->m_aClients[ClientId].m_PrevPredicted.m_Tuning.m_LaserUnDeep

22:39

weapon lights up when it has set tune value, same thing with other tunes (edited)

github-merge-queue[bot]

[ddnet:master] 2 new commits

8f37729 Improve map image validation - Robyt3 91ec32b Merge pull request #9213 from Robyt3/Client-Mapimages-Validation - def-

22:54

def-

[ddnet/ddnet] Pull request closed: #9213 Improve map image validation

kebs

bug?

ok yea i found whats wrong

22:56

will push a fix

Evelyn

I want to send commands in a separate discord server, and for them to be executed in rcon on the server

https://github.com/jxsl13/tw-discord-econ-fifo

GitHub - jxsl13/tw-discord-econ-fifo

Contribute to jxsl13/tw-discord-econ-fifo development by creating an account on GitHub.

23:00

it's doing only that, nothing more

23:01

good luck

jxsl13

https://github.com/jxsl13/tw-discord-econ-fifo

you do not print cmd output to discord right ?

no

23:12

you get a response, tho "executed command xyz

23:12

"

23:14

which is ephemeral an only visible for the uset

23:14

user

KebsCS

[ddnet/ddnet] Pull request opened: #9214 Fix invalid prediction tuning in tune zone 0

Fixed a bug where entering a different tune zone would incorrectly apply the current zone's settings to tune zone 0, causing prediction issues for players in other zones (red weapon color in the video) https://github.com/user-attachments/assets/84f8c87f-f2f5-4951-a5a2-eb4b1afbe230

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 (especi...