01:44 < bridge> <soulyvevo> chillerdragon: does your bot on multeasymap track block stats?
06:08 < bridge> <chillerdragon> Yes @Souly checkout !rank or !top3
06:08 < bridge> <chillerdragon> @soulyvevo: pinger
07:20 < bridge> <chillerdragon> Who added backdoor to Adorenarine? o0
07:22 < bridge> <ryozuki> ```
07:22 < bridge> <ryozuki> Visit https://github.com/ipfs/kubo/releases or https://dist.ipfs.tech/#kubo and update to version 0.36.0 or later.
07:22 < bridge> <ryozuki> panic: runtime error: invalid memory address or nil pointer dereference
07:22 < bridge> <ryozuki> [signal SIGSEGV: segmentation violation code=0x1 addr=0x20 pc=0x561fd1b28346]
07:22 < bridge> <ryozuki> ```
07:22 < bridge> <ryozuki> great
07:22 < bridge> <ryozuki> was trying kubo, a ipfs node
07:23 < bridge> <ryozuki> and it crashes
07:23 < bridge> <ryozuki> go ftw
07:23 < bridge> <ryozuki> golang ftw
07:27 < ws-client> **<ChillerDragon>** sigmatationfault <:gigachad:960830678465933403> 
07:28 < ws-client> **<ChillerDragon>** @ryozuki urgent reminder that there are 108 pending prs
07:32 < bridge> <ryozuki> has anyone thought of making a archive  of maps on ipfs
07:32 < bridge> <ryozuki> yeah prs, mm ill review sometime later
07:36 < ws-client> **<ChillerDragon>** <:poggers2:1008007455936094328> mergozuki
07:50 < bridge> <ryozuki> https://www.iroh.computer/
07:51 < bridge> <ryozuki> i wonder if a p2p masterserver could work
07:56 < bridge> <Ewan> how would registration work
08:03 < ws-client> **<ChillerDragon>** how does registration in any p2p work? I guess you have a public list of known nodes to talk to somewhere.
08:04 < bridge> <Ewan> what would differentiate a tracker list from a matter server itself then
08:04 < bridge> <Ewan> master*
08:11 < bridge> <ryozuki> well the tracker list would list some of the initial nodes
08:11 < bridge> <ryozuki> but from then on u can find more ppl i guess
08:12 < bridge> <ryozuki> it would be hard to ban bad servers tho
08:12 < bridge> <ryozuki> ig
08:12 < bridge> <ryozuki> also random ppl can make random tracker lists too
08:12 < ws-client> **<ChillerDragon>** @zhn bru your github display name does not match your username I pinged some random https://github.com/vlad in the release because of the changelog script ._.
08:46 < bridge> <dark_distanc> hi
08:56 < ws-client> **<ChillerDragon>** elo
09:06 < ws-client> **<ChillerDragon>** did you ever have to test something locally that required players with multiple ips? I found this neat hack to get 3 ips locally.
09:06 < ws-client> **<ChillerDragon>** https://zillyhuhn.com/cs/.1753599940.png
09:06 < ws-client> **<ChillerDragon>** local network range, loopback and ipv4/ipv6
10:10 < bridge> <Assa> gumo ^.^
10:10 < bridge> <Assa> #10527 and #10501 should be easy, they are approved and ready
10:11 < bridge> <DDNet> https://github.com/ddnet/ddnet/pull/10527
10:11 < bridge> <DDNet> https://github.com/ddnet/ddnet/pull/10501
10:11 < bridge> <Assa> all others from me need some love, but I feel like I have too many PRs open simultaniously
10:17 < bridge> <ryozuki> nice, added to queue
10:18 < bridge> <Assa> Does anyone know how the laser in the laser-shotgun works? :justatest:  like I need to find it if it is affected by LaserBounceDelay, but from testing it seems to have no effect
10:25 < bridge> <Assa> but it should have oO idk
10:29 < ws-client> **<ChillerDragon>** @avolicious is rufmord illegal on kog? i would like to report an incident 
10:29 < ws-client> **<ChillerDragon>** https://zillyhuhn.com/cs/.1753604911.png
10:58 < bridge> <Assa> It actually looks like I never destroy the buffer objects in the render layers
10:58 < bridge> <ryozuki> https://cdn.discordapp.com/attachments/293493549758939136/1398952714796404826/image.png?ex=68873bbe&is=6885ea3e&hm=b3d10af2413ddf6335c4322442cd77cc47bcb95ab470e8b7feea12f505f67856&
10:58 < bridge> <Assa> to be fair, the previous code also didn't do this!
11:00 < bridge> <Assa> I see the same issue in other parts of the code, the TextContainer may calls RecreateBufferObject but never deletes it
11:03 < bridge> <Assa> ```
11:03 < bridge> <Assa>     // destroying all buffer objects means, that all referenced VBOs are destroyed automatically, so the user does not need to save references to them
11:03 < bridge> <Assa>     void DeleteBufferContainer(int &ContainerIndex, bool DestroyAllBO = true) override;
11:03 < bridge> <Assa> ```
11:03 < bridge> <Assa> What is a reference VBO? Do I need to care about these?
11:46 < bridge> <vahemaaa> chillerdragon crash when enter plot zone in BlmapChill, i compiled the latest version just now
11:46 < bridge> <vahemaaa> https://cdn.discordapp.com/attachments/293493549758939136/1398964771117862922/image.png?ex=688746f8&is=6885f578&hm=9b16741024b9160172492c18b4d603a2347292e317a311bdc36a6be05a2d9ee2&
11:46 < bridge> <vahemaaa> it happens everytime
12:55 < ws-client> **<ChillerDragon>** could you get a backtrace idk which vector that is @vahemaaa 
12:56 < ws-client> **<ChillerDragon>** doesnt crash for me
12:57 < bridge> <robyt3> It's probably an MSVC specific assertion
12:57 < ws-client> **<ChillerDragon>** asan should catch it i guess
12:58 < bridge> <robyt3> Yeah, or compile in RelWithDebInfo and click the Retry button in the popup, @vahemaaa
13:10 < ws-client> **<ChillerDragon>** even asan doesnt have it maybe its a setting idk. Might be some war list entry i dont have 
13:10 < ws-client> **<ChillerDragon>** @vahemaaa can u try tuning off the warlist?
13:11 < bridge> <Assa> Does the Buffer-memory always resizes itself to a max value? It seems like it doesn't get cleared, but it seems like it also doesn't get bigger if I load a smaller map
13:13 < bridge> <vahemaaa> same
13:14 < ws-client> **<ChillerDragon>** this looks interesting tho ``prediction/entities/character.cpp:1366:4: runtime error: index -1 out of bounds for type 'struct WeaponStat[6]'``
13:16 < bridge> <robyt3> Maybe depends on the graphics backend
13:17 < bridge> <vahemaaa> https://cdn.discordapp.com/attachments/293493549758939136/1398987670830579712/image.png?ex=68875c4c&is=68860acc&hm=ff8f4ec79733f6b38d8461291d25148e8b0b079e0d1e12252175342d279a14b5&
13:18 < bridge> <kebscs> Show callstack
13:18 < bridge> <robyt3> Check the call stack for the first frame in DDNet code
13:18 < bridge> <kebscs> Bottom right
13:18 < bridge> <vahemaaa> ```
13:18 < bridge> <vahemaaa> >    chillerbot-ux.exe!std::vector<SSwitchers,std::allocator<SSwitchers>>::operator[](const unsigned __int64 _Pos) Line 1930    C++
13:18 < bridge> <vahemaaa> ```
13:19 < bridge> <robyt3> Can you copy the entire call stack?
13:19 < bridge> <vahemaaa> https://cdn.discordapp.com/attachments/293493549758939136/1398988235631366197/image.png?ex=68875cd3&is=68860b53&hm=8f93321d2de046cc5d1a1558f9c0a9fa767980419033458017d3cc8b4045ec4b&
13:21 < bridge> <robyt3> Thanks. Looks more like a general issue in DDNet to me.
13:23 < bridge> <Assa> Wanted to ping ryo for review, already in merge queue :deen_star:
13:23 < bridge> <vahemaaa> aha
13:23 < bridge> <vahemaaa> when antiping is off
13:23 < bridge> <vahemaaa> it does not crash
13:23 < bridge> <vahemaaa> when i turned on antiping, it crashed, at plot zone
13:25 < bridge> <Assa> @robyt3 can you take a look on #10572 ? I believe it's finished like that
13:25 < bridge> <DDNet> https://github.com/ddnet/ddnet/pull/10572
13:26 < bridge> <Assa> @robyt3 can you take a look on #10572 ? I believe it's almost finished like that
13:27 < bridge> <robyt3> Yeah, can take a closer look later. `for(auto &&It = m_vRenderLayers.begin(); It != m_vRenderLayers.end(); ++It)` can probably be a for-each loop. And `m_aTextures[i].Invalidate();` should be unnecessary because `UnloadTexture` invalidates the handle.
13:27 < bridge> <kebscs> <https://github.com/ddnet/ddnet/pull/10455> @ryozuki
13:27 < bridge> <robyt3> Probably needs a rebase so the mingw workflow runs
13:28 < bridge> <vahemaaa> and yes, it happens on DDNet client also
13:37 < ws-client> **<ChillerDragon>** @robyt3 @vahemaaa yea none of that callstack code should be edited in chillerbot-ux
13:38 < ws-client> **<ChillerDragon>** ah already confirmed thanks @vahemaaa 
13:38 < ws-client> **<ChillerDragon>** go open ddnet issue
13:42 < ws-client> **<ChillerDragon>** @vahemaaa will you open an issue or should i?
13:43 < bridge> <vahemaaa> i was about to, but if you do it, it will be clearer to understand i guess xD
13:43 < bridge> <Assa> duplicate, can be closed
13:46 < ws-client> **<ChillerDragon>** ffs i could only reproduce once
13:46 < bridge> <robyt3> Where on the map does it crash?
13:46 < ws-client> **<ChillerDragon>** plots
13:46 < ws-client> **<ChillerDragon>** they use special switch tiles
13:47 < ws-client> **<ChillerDragon>** ``set_view 515 365`` @robyt3 45.141.57.154:8303
13:47 < ws-client> **<ChillerDragon>** xd saw u crash  @vahemaaa 
13:48 < bridge> <vahemaaa> xd
13:48 < bridge> <robyt3> I see
13:48 < bridge> <robyt3> Why is there a switch with number `-1`?
13:48 < ws-client> **<ChillerDragon>** its for plots
13:48 < bridge> <robyt3> Well, `-1` is out of bounds :monkaS:
13:49 < ws-client> **<ChillerDragon>** thats the idea xd its not a normal switch
13:49 < bridge> <Assa> :omo:
13:49 < ws-client> **<ChillerDragon>** its a fake switch used for plot tiles 
13:49 < ws-client> **<ChillerDragon>** do you also get this @robyt3 ? ``prediction/entities/character.cpp:1366:4: runtime error: index -1 out of bounds for type 'struct WeaponStat[6]'``
13:49 < bridge> <robyt3> What is the expected behavior? Am I going to break the game mode when simply ignoring those invalid values
13:50 < ws-client> **<ChillerDragon>** i dont think so
13:50 < ws-client> **<ChillerDragon>** server side F-DDrace is forked and client side it would only hide the entities. Not sure if it already requires a custom client to map that.
13:51 < bridge> <robyt3> No, I only added assertions manually, but that one looks worse. Probably a result of supporting no weapon with `-1`
13:52 < ws-client> **<ChillerDragon>** maybe https://github.com/ddnet/ddnet/pull/9230 is not supported too well in prediction code?
13:52 < bridge> <robyt3> yeah, it's missing a check at least there
13:57 < bridge> <vahemaaa> lmao i hit edge of this drawing, it also crashes with drawings
13:57 < bridge> <vahemaaa> https://cdn.discordapp.com/attachments/293493549758939136/1398997682781032538/image.png?ex=6887659f&is=6886141f&hm=d8c6d80560890add8082db8ac4600c35acd9eeea3ea322bb799a8114d92dae0d&
14:04 < bridge> <vahemaaa> im a bad issue opener
14:08 < bridge> <Assa> can you append a crashlog @vahemaaa ?
14:08 < bridge> <vahemaaa> let me try
14:09 < bridge> <Assa> should be config_dir in /dumps
14:10 < bridge> <xcwqw1> Will DDNet support special color characters in the future, like in Minecraft? It would be helpful if server messages could display different colors in different parts to clearly show their function.
14:12 < bridge> <vahemaaa> i couldnt find the log there, so i added visual studio's log
14:37 < bridge> <ryozuki> 97 prs now @chillerdragon
14:59 < bridge> <Assa> what's wrong with the test here? https://github.com/ddnet/ddnet/pull/10572
15:02 < bridge> <chillerdragon> Nice @ryozuki
15:05 < bridge> <chillerdragon> @xcwqw1: probably not happening any time soon. Here my last attempt https://github.com/ddnet/ddnet/pull/8283 if you go and write a html parser there might be a chance otherwise I don’t think anyone will work on it
15:32 < bridge> <robyt3> Try running the test locally with `scripts/integration_test.py your-build-folder`
15:36 < bridge> <robyt3> Seems like we exceeded the maximum runners though, so some were removed from the queue again due to timeout
15:41 < ChillerDragon> opening console or clicking a tab in menu had a slight delay
15:41 < ChillerDragon> the first time
15:42 < bridge> <robyt3> If you have a lot of console messages that have not been rendered yet and then open the console it may delay a bit
15:42 < bridge> <robyt3> Switching between the browser tabs would probably be due to sorting and filtering of servers being performed every time
15:43 < ChillerDragon> i thought its some super smart render cache feature
15:43 < ChillerDragon> especially in console
15:44 < ChillerDragon> but today the delay increased to like seconds and i full on failed my run xd
15:44 < ChillerDragon> might also be my bot client
15:44 < ChillerDragon> im just surprised
15:44 < bridge> <robyt3> The console rendering isn't very smart at the moment
15:44 < ChillerDragon> was menu
15:44 < ChillerDragon> i clicked "call vote"
15:44 < ChillerDragon> and was stuck like 1s
15:45 < bridge> <robyt3> Weird, but if it's reproducible you could try profiling
15:46 < ChillerDragon> i clicked other menu entries they were as slow as always like just noticable but not game breaking
15:46 < ChillerDragon> lemme boot other client quickly
15:47 < ChillerDragon> cant reproduce gg
15:53 < bridge> <Assa> I believe the tests timeouted because of this, I can't rerun them
15:54 < bridge> <robyt3> Doesn't really seem related, it's only one test case that times out in both runners and other test cases pass after that
15:54 < bridge> <learath2> Just how long is the CI/CD pipeline now?
15:55 < bridge> <learath2> Actually we don't even do any CD
15:57 < bridge> <robyt3> The longest appears to be clang-tidy around 30 minutes
15:58 < bridge> <robyt3> But the queue was hung up on the shorter workflows like checking the rust format, licence etc.
15:58 < bridge> <robyt3> Because the early queue entries were waiting for those workflows to finish, some later queue entries timed out
15:59 < bridge> <robyt3> I don't recall clang-tidy being the slowest run though, was probably from switching to clang-tidy 20
15:59 < bridge> <ryozuki> xD
15:59 < bridge> <ryozuki> i added o ne again
16:00 < bridge> <learath2> clang-tidy is pretty slow, I never run it locally
17:07 < bridge> <ChnFreddie> @blaiszephyr we never talked about that one PR after you said you were gonna test it
17:08 < bridge> <ChnFreddie> did you run into anything
17:14 < bridge> <Assa> I can't even run the integration tests, do I need a specific python version?
17:14 < bridge> <Assa> ```
17:14 < bridge> <Assa> FileNotFoundError: [Errno 2] No such file or directory: './integration_start_client_pipkd33r/client0.fifo'
17:14 < bridge> <Assa> ```
17:18 < bridge> <robyt3> Any recent Python 3 version should work
17:19 < bridge> <robyt3> But apparently something on Ubuntu doesn't work, see #9937
17:19 < bridge> <DDNet> https://github.com/ddnet/ddnet/pull/9937
17:23 < bridge> <Assa> it should be `name_arg = os.open(name, flags=os.O_WRONLY|os.O_CREAT)` , it's missing the create flag
17:23 < bridge> <Assa> now every test is timing out on ubuntu :/ for me
17:25 < bridge> <Assa> there is a timeout multiplier?
17:26 < bridge> <robyt3> yeah, there should even be a command line argument for that
17:26 < bridge> <robyt3> but if every test times out it's probably not the timeout that's broken
17:29 < bridge> <Assa> I only get timeout errors, I can't integration test at all
17:29 < bridge> <Assa> set the timeout multiplier now to 10
17:41 < bridge> <Assa> test timeouts on wait_for_exit oO
17:43 < bridge> <Assa> does it state somewhere, that you need sudoers rights for the tests?
17:44 < bridge> <Assa> Also I can not reproduce the issue the pipeline had, it works for me on ubuntu24
17:44 < bridge> <Assa> ```
17:44 < bridge> <Assa> sudo python scripts/integration_test.py build
17:44 < bridge> <Assa> running 11 tests
17:44 < bridge> <Assa> meta_timeout ... ok
17:44 < bridge> <Assa> meta_test_timeout ... ok
17:44 < bridge> <Assa> start_server ... ok
17:44 < bridge> <Assa> start_client ... ok
17:44 < bridge> <Assa> client_can_connect ... ok
17:44 < bridge> <Assa> open_editor ... ok
17:44 < bridge> <Assa> smoke_test ... ok
17:44 < bridge> <Assa> start_mastersrv ... skipped
17:44 < bridge> <Assa> server_can_register ... skipped
17:44 < bridge> <Assa> server_can_register_tw_0_6 ... skipped
17:44 < bridge> <Assa> server_can_register_tw_0_7 ... skipped
17:44 < bridge> <Assa> 
17:44 < bridge> <Assa> test result: ok. 7 passed; 4 skipped, 0 failed; finished in 52.85s
17:44 < bridge> <Assa> ```
17:45 < bridge> <robyt3> You shouldn't need sudo for the integration tests, did it only work with sudo for you?
17:57 < bridge> <Assa> it only worked with sudo for me, and with the fix for directory creation
17:58 < bridge> <Assa> I know why the test failed, the runner was probably busy, and the smoke test on ubuntu 24 with valgrind enabled takes ages
17:58 < bridge> <Assa> and it probably cancelled itself after 10 minutes even tho it was still running fine
17:59 < bridge> <robyt3> Can you comment that on #9937?
17:59 < bridge> <DDNet> https://github.com/ddnet/ddnet/pull/9937
18:03 < bridge> <Assa> @robyt3 can you try to rerun [this action](https://github.com/ddnet/ddnet/actions/runs/16550924173/job/46805622942?pr=10572) once?
18:03 < bridge> <Assa> merci
18:25 < bridge> <Assa> tested ogl3 backend it it doesn't have a buffer-memory (it's always 0), I think it means that vulkan is double buffered. But the textures also appear to be cleared in my PR
18:25 < bridge> <Assa> tested ogl3 backend it seems it doesn't have a buffer-memory (it's always 0), I think it means that vulkan is double buffered. But the textures also appear to be cleared in my PR
18:30 < bridge> <Assa> they failed again :/ idk
18:35 < bridge> <Jupstar ✪> If someone needs an easy issue to work on:
18:35 < bridge> <Jupstar ✪> Add allocation size of all vram to all assert messages coming from the gfx backend
18:36 < bridge> <Jupstar ✪> Would give us potentially some better clue on what's going on
18:36 < bridge> <Jupstar ✪> Even better would probably current ram usage and vram usage fetched from the os too
18:36 < bridge> <Assa> ```
18:36 < bridge> <Assa>         if self.valgrind_memcheck:
18:36 < bridge> <Assa>             self.timeout_multiplier *= 10
18:37 < bridge> <Assa> ```
18:37 < bridge> <Assa> I should try to up the timeout multiplier even more, I just seem to cause more overhead or something
19:21 < bridge> <ryozuki> anyone tried keycloak
19:21 < bridge> <ryozuki> seems like a nice  thing to not have to make auth on every app
19:23 < bridge> <Assa> There are different solutions depending on your type of federation, I guess you're fine with keycloak
19:24 < bridge> <Assa> at my work we are using unity-idm, which would be probably overkill for you, or do you need to server users in the 5 digits or more?
19:25 < bridge> <Assa> at my work we are using unity-idm, which would be probably overkill for you, or do you need to serve users in the 5 digits or more?
19:25 < bridge> <Assa> if you only need SSO, I think the keycloak would already be overkill 😄
19:38 < bridge> <ryozuki> lol no
19:38 < bridge> <ryozuki> im just new into this stuff
19:38 < bridge> <ryozuki> i might do some random apps and thought why not try this fancy stuff
19:38 < bridge> <ryozuki> i just want to have auth fast and easy
19:38 < bridge> <ryozuki> xd
19:43 < bridge> <Assa> ```
19:43 < bridge> <Assa>     for _ in range(2):
19:43 < bridge> <Assa>         server.wait_for_log_prefix("server: player has entered the game", timeout=20)
19:43 < bridge> <Assa> ```
19:43 < bridge> <Assa> 
19:43 < bridge> <Assa> integration tests fail here 🤔  I already upped the timeout from 10 to 20, I guess the log is not there then?
19:43 < bridge> <Assa> idk if it would be a "fix" just to remove that check
19:44 < bridge> <robyt3> The fix is figuring out why the client doesn't enter the game
19:46 < bridge> <Assa> it's right after the sv_map command 😦
19:47 < bridge> <Assa> if the client would have crashed, the test would have noticed, wouldn't it?
19:49 < bridge> <ryozuki> i thought about making a mini social app for ddnet
19:49 < bridge> <ryozuki> but how would u link ur profile to ur player name
19:49 < bridge> <ryozuki> thats impossible i guess
19:51 < bridge> <Assa> in what sense do you speak about linking?
19:52 < bridge> <ryozuki> imagine i register as Ryozuki
19:52 < bridge> <ryozuki> and i want to verify im Ryozuki in ddnet
19:52 < bridge> <ryozuki> u cant
19:52 < bridge> <ryozuki> unless maybe u use discord but it still isnt a surefire way
19:52 < bridge> <Assa> you get a code, that you need to send as ryozuki in a chat command
19:52 < bridge> <ryozuki> hmm
19:52 < bridge> <ryozuki> that requires a ingame bot
19:52 < bridge> <ryozuki> maybe as a clanname
19:52 < bridge> <Assa> but nobody could ever stop you from faking to be sb else
19:52 < bridge> <ryozuki> so i can use master
19:53 < bridge> <ryozuki> true
19:53 < bridge> <Assa> it may require a econ bot
19:53 < bridge> <ryozuki> cuz i can join as a name ingame i want to fake
19:53 < bridge> <ryozuki> without accounts its not possible
19:56 < bridge> <Assa> if you mod a server, you could only let people play after they /login code but this is bound to cause other problems and would basically be a local account
20:03 < bridge> <louis> finish accounts pr 👀
20:11 < bridge> <Assa> client gets stuck in this screen: something like "Asking the server to join"
20:11 < bridge> <Assa> https://cdn.discordapp.com/attachments/293493549758939136/1399091788307759205/screenshot_2025-07-27_20-10-19.png?ex=6887bd44&is=68866bc4&hm=496f58a8468f7abf958897c543ca1efed71f708354c28f4b6bae89a070970d1c&
20:11 < bridge> <Assa> cancelling and retrying works
20:19 < bridge> <Assa> idk, seems like a protocol/timing issue for me, client takes a lot longer in trying to download and setup the map
20:19 < bridge> <Assa> > Requesting to join the game
20:34 < bridge> <Assa> I can't reproduce it
22:10 < bridge> <melon> huuuge!
22:51 < bridge> <kollpotato> i read the pr name as ban linear