00:19 <+bridge_> <chillerdragon> Heinrich Hammer in big 2026 https://youtu.be/rfsKd2ZXRI4
00:30 <+bridge_> <kebscs> bumm explosion*
01:22 <+bridge_> <kebscs> @tsfreddie ^ if you could take a look, idk if thats 100% proper fix
01:42 <+bridge_> <12944qwerty> is there a reason that rotating in editor doesn't actually rotate speedups
01:49 <+bridge_> <kebscs> It should, I'm pretty sure I added it
01:51 <+bridge_> <kebscs> Can you try selecting just a single speeder, and with the outdated one too
01:56 <+bridge_> <12944qwerty> ok it's really weird
01:56 <+bridge_> <12944qwerty> s
01:57 <+bridge_> <12944qwerty> sometimes it just doesn't work ig
03:13 <+bridge_> <gorp_tw> for any max speed 1, you cant start accelerating in the direction of the arrow
03:13 <+bridge_> <gorp_tw> 1,1: can move against arrow even from 0 speed
03:13 <+bridge_> <gorp_tw> 2,1: can only move against arrow if you had some run speed before
03:13 <+bridge_> <gorp_tw> 3,1: can only move against if you had more than usual run speed from before (hook, other source, good floating point run speed) otherwise
03:13 <+bridge_> <gorp_tw> https://cdn.discordapp.com/attachments/293493549758939136/1361407779482374285/2025-04-14_11-27-00.mp4?ex=69c034cc&is=69bee34c&hm=50c51d9d6e7293c6d8581b738362d32006c098d26158092fd0cea9a54588c5ae&
10:36 <+bridge_> <Assa> huh :justatest:
10:37 <+bridge_> <Assa> @ryozuki are you okay?
10:38 <+bridge_> <ryozuki> just making hein life easier
10:38 <+bridge_> <ryozuki> i do it myself quicker
10:38 <+bridge_> <ryozuki> xD
10:39 <+bridge_> <Assa> I liked the spinner rendering 🙁 and scorebord render improvementd
10:39 <+bridge_> <ryozuki> feel free to copy the branch
10:41 <+bridge_> <ryozuki> ill just work on my projects this way i dont "waste reviewers time"
10:49 <+bridge_> <qetoast> :cammo:
11:08 <+bridge_> <heinrich5991> https://github.com/ddnet/ddnet/pull/11975 seemed like it was done by LLM only, and I don't want to have reviewers fix up LLM generated PRs, so I closed it
11:08 <+bridge_> <heinrich5991> if reviewers want to use LLMs to generate code, they can do it themselves
11:17 <+bridge_> <inv41idu53rn4m3> See I was feeling disappointed seeing this at first but sloppers deserve no mercy
11:19 <+bridge_> <heinrich5991> I think you can generate useful code with LLMs, but you should review what it outputs yourself
11:19 <+bridge_> <learath2> Drama Sundays
11:20 <+bridge_> <heinrich5991> btw, I was thinking to maybe to a git subtree for libtw2 code so we can have a copy of it in-tree
11:20 <+bridge_> <heinrich5991> @learath2 would that sound sensible to you?
11:21 <+bridge_> <heinrich5991> I also don't like having a https://github.com/heinrich5991/libtw2 reference in the ddnet repository, I agree on the "core algorithm" part
11:21 <+bridge_> <heinrich5991> but I'd also like to not maintain two copies of libtw2's code, or even its snapshot code
11:22 <+bridge_> <learath2> I guess that's acceptable, doesn't quite avoid the heinrich5991/libtw2 reference, but at least it's all in one repo
11:23 <+bridge_> <heinrich5991> I think a subtree doesn't have the heinrich5991/libtw2 reference, except if we provide scripts to update it, which we probably should
11:23 <+bridge_> <heinrich5991> not sure, haven't worked with subtrees myself yet
11:25 <+bridge_> <heinrich5991> (different from submodule)
12:07 <+bridge_> <kebscs> anyone here is packager? who is a packager of ddnet?
12:10 <+bridge_> <kebscs> <https://github.com/ddnet/ddnet/pull/11951#issuecomment-4103397443> or what the packagers do?
12:12 <+bridge_> <heinrich5991> packager are people who create packages of ddnet for linux distributions
12:14 <+bridge_> <kebscs> doesnt seem like it involves editing ddnet config 🤔
12:25 <+bridge_> <Assa> downloading cachyos
12:26 <+bridge_> <ryozuki> i did, maybe i simply confused the issue but its ok
12:30 <+bridge_> <heinrich5991> sorry for that then. the idea was that the backward compatibility thing shouldn't be recorded into demos at all. what the change did was recording it into the (server-side) demo if an old client is present. should I reopen the PR?
12:31 <+bridge_> <heinrich5991> users edit ddnet configs. packagers can install vim syntax files for ddnet so that users can edit configs with syntax highlighting
12:32 <+bridge_> <ryozuki> u could have just said that..
12:33 <+bridge_> <heinrich5991> the problem is that I don't want to interact with an LLM via github comments
12:33 <+bridge_> <heinrich5991> (which has happened before)
12:33 <+bridge_> <heinrich5991> I want to talk to a human who is able to learn
12:33 <+bridge_> <ryozuki> im the one making the pr and i review what i do, i just really confused what the solution should be, which i think llm or not can happen
12:34 <+bridge_> <ryozuki> also u dont have to interact with some prs if u dont want to
12:34 <+bridge_> <ryozuki> thats fine too
12:35 <+bridge_> <heinrich5991> > the idea was that the backward compatibility thing shouldn't be recorded into demos at all. what the change did was recording it into the (server-side) demo if an old client is present
12:35 <+bridge_> <heinrich5991> this is what the solution should be ^
12:36 <+bridge_> <heinrich5991> I can't reopen the PR, the branch has been deleted
12:36 <+bridge_> <heinrich5991> it'd also be nice if you could write down how an LLM was used to generate the PR, there's a place in the PR for that
12:36 <+bridge_> <ryozuki> it doesnt matter, ill open another if i feel like it later
12:36 <+bridge_> <heinrich5991> kk
12:36 <+bridge_> <heinrich5991> it'd also be nice if you could write down how an LLM was used to generate the PR, there's a place in the PR description for that
13:31 <+bridge_> <alw5> @heinrich5991 maybe you forgot to respond on https://github.com/ddnet/ddnet/pull/11487 ? ^^
13:31 <+bridge_> <inv41idu53rn4m3> I could see packagers possibly editing storage.cfg but not much else
13:31 <+bridge_> <qetoast> Can someone tell me why https://github.com/ddnet/ddnet/actions/runs/23402221141/job/68075203548?pr=11971 failed? it's not because of the pr tho right?
13:38 <+bridge_> <robyt3> I restarted the run. It's likely unrelated unless it keeps happening.
13:47 <+bridge_> <Assa> @robyt3 tried to reproduce <https://github.com/ddnet/ddnet/issues/11708#issuecomment-4106090661> but also wasn't able to anymore ... 🤔
14:06 <+bridge_> <murpi> Why are my review comments "pending"? Why is that even a thing?
14:07 <+bridge_> <murpi> https://cdn.discordapp.com/attachments/293493549758939136/1485263577140039812/image.png?ex=69c13aff&is=69bfe97f&hm=a28ffec5629cf88a7d35260c3fc970344db8894420db4ea7a22a48ded963626b&
14:07 <+bridge_> <murpi> This is so dumb
14:12 <+bridge_> <robyt3> If you were going to submit many comments then it would be inconvenient if they would arrive as soon as you created them because some comments require understanding of other comments, so when you start a review the comments are pending until you submit the review.
14:14 <+bridge_> <robyt3> Regarding the second comment: if extracting the archive failed, it leaves and empty folder, so next time it would fail with the runtime error below `if not symbols_executable.is_file()`
14:14 <+bridge_> <robyt3> Regarding the second comment: if extracting the archive failed, it leaves an empty folder, so next time it would fail with the runtime error below `if not symbols_executable.is_file()`
14:15 <+bridge_> <robyt3> Ah, I see that's the issue to begin with. If the download failed temporarily then it should ideally try again if you run the script again.
14:16 <+bridge_> <kebscs> you have to press "Submit comments" otherwise noone can see them
14:17 <+bridge_> <kebscs> if you want to submit comments 1 by 1, then use "Comment" and not "Start a review"
14:21 <+bridge_> <murpi> Having to manually clean up is a bit annoying. You could either create a temporary folder and rename it after or remove the folder every time the script is run
14:21 <+bridge_> <murpi> I see, thanks
14:22 <+bridge_> <murpi> (I don't even see a way to submit the review rn, but I guess its because it's been merged already)
14:22 <+bridge_> <melon> the same happened to me with one of furo's PRs, and i went ingame and asked him why he'd ignore me for a week, the PR is merged and my comment wasnt looked at
14:22 <+bridge_> <melon> then i saw the [pending] tag :kekWTF:
14:28 <+bridge_> <kebscs> https://cdn.discordapp.com/attachments/293493549758939136/1485269045530857552/image.png?ex=69c14017&is=69bfee97&hm=e15fd6fba936001ffe53033ba596f90add9842ae118e903c9ea0d11403d75c0c&
15:50 <+bridge_> <kebscs> Could ddnet website host a simple emscripten build?
15:50 <+bridge_> <kebscs> Maybe one built from nightly
15:50 <+bridge_> <kebscs> Would help test emscripten changes
15:58 <+bridge_> <robyt3> Yes, my plan is to add something like web.ddnet.org to host the current nightly version of the Emscripten client. It requires integrating the Emscripten build into the scripts in ddnet-libs. Then explain necessary setup to deen. Technically we could also just pull the latest artifact from GitHub because those should be build reproducibly.
16:00 <+bridge_> <kebscs> amazing
16:00 <+bridge_> <kebscs> could do it now or is there smth blocking it?
16:00 <+bridge_> <robyt3> Testing them locally is also pretty easy though, because you can run the artifacts from the CI. You only need a minimal web server like `other/emscripten/server.py`.
16:01 <+bridge_> <robyt3> It depends on #11920 being merged to make it a full index page
16:01 <+bridge_> <DDNet> https://github.com/ddnet/ddnet/pull/11920
16:02 <+bridge_> <kebscs> ^ @essigautomat merge pls 😄
16:07 <+bridge_> <kebscs> and this <https://github.com/ddnet/ddnet/pull/11949> how many approves does it need for merge :p
16:08 <+bridge_> <qetoast> 6 - 7
16:08 <+bridge_> <kebscs> go approve
16:08 <+bridge_> <kebscs> we are at 4
16:12 <+bridge_> <robyt3> Review #11800 to progress on #11776 :justatest:
16:12 <+bridge_> <DDNet> https://github.com/ddnet/ddnet/pull/11800
16:12 <+bridge_> <DDNet> https://github.com/ddnet/ddnet/pull/11776
16:14 <+bridge_> <kebscs> 4k lines :tear:
16:15 <+bridge_> <robyt3> Maybe I could put the raw move of the definitions to `envelope_editor.cpp` into a separate commit first, but I don't know if that makes it easier to review overall
16:56 <+bridge_> <Assa> busy with cachyos rn
17:02 <+bridge_> <Assa> CachyOs gives me 1000+ more fps than windows 🤔
17:02 <+bridge_> <Assa> https://cdn.discordapp.com/attachments/293493549758939136/1485307713125875904/screenshot_2026-03-22_17-01-10.png?ex=69c1641a&is=69c0129a&hm=ae3edd68cfa91a72a599c98c62c9577e13c6f2107f7d62964017bab1e8e10269&
17:03 <+bridge_> <Assa> I have never ever reached 7K fps before
17:07 <+bridge_> <kebscs> human eye cant see 30 fps
17:08 <+bridge_> <Assa> that's a lie
17:49 <+bridge_> <Assa> haha I forgot to backup my private keys 🥳
18:06 <+bridge_> <teero777> ragebait
18:09 <+bridge_> <Assa> <https://youtu.be/Sb_7uN7sfTw?si=At-7VEE-wP1zpe8z>
18:24 <+bridge_> <Assa> vs code man, you can install cmake extensions and nothing tells you cmake is not installed :kek:
18:27 <+bridge_> <Assa> I have never compiled this fast, all of ddnet in 10 seconds
18:28 <+bridge_> <Assa> wtf is cachyos doing, like how wtf?
18:30 <+bridge_> <Assa> anybody an idea why I get this when I compile ddnet myself? I don't get this on nightly
18:30 <+bridge_> <Assa> https://cdn.discordapp.com/attachments/293493549758939136/1485329787856621731/screenshot_2026-03-22_18-29-31.png?ex=69c178a9&is=69c02729&hm=c3f469204e8c95e506c380777568dbd601358b0795555f8d41c5fa267a5d8b6f&
18:30 <+bridge_> <Assa> oh maybe I am still missing the right nvidia drivers :kek:
18:31 <+bridge_> <Assa> nah that's not it
18:33 <+bridge_> <Assa> lol this happens if you start ddnet from the root directory instead of the build directory :thonk:
18:33 <+bridge_> <Assa> on windows I always had to start from root
18:34 <+bridge_> <Assa> I guess the shaders in the data directory are actually not build which makes sense
18:44 <+bridge_> <chillerdragon> Do you finally switched to Linux?
18:44 <+bridge_> <chillerdragon> So you finally switched to Linux?
18:44 <+bridge_> <chillerdragon> No more windows?
18:45 <+bridge_> <teero777> rich mf
19:18 <+bridge_> <cloud9000.> Pic of a Map I made in my own DDnet Client made in python with pyglet (probably will remake it in pygame because 130 fps is not enough!!!!!!)
19:18 <+bridge_> <cloud9000.> https://cdn.discordapp.com/attachments/293493549758939136/1485341984292868340/Map.png?ex=69c18405&is=69c03285&hm=9019196e92ad62906bb54ed7ad157165af367db539b6e13409df0d134f10eef6&
19:29 <+bridge_> <qetoast> Show us gameplay
19:30 <+bridge_> <cloud9000.> okay but first I gotta record it so wait a few minutes
19:47 <+bridge_> <cloud9000.> Im so sorry but I cant make the video fps higher because recording and playing on a 7 year old laptop drops the framerate a lot ):
19:47 <+bridge_> <cloud9000.> https://cdn.discordapp.com/attachments/293493549758939136/1485349282604253396/2026-03-22_19-44-17.mp4?ex=69c18ad1&is=69c03951&hm=168a8c70dff8492c9ea0a5afc64ba5b58f80c4fce0824cfc852baaf579dc7643&
19:48 <+bridge_> <cloud9000.> also the reason why im jumping so high is because I messed up the delta-time on the jump part oops
19:49 <+bridge_> <qetoast> :kek:
19:49 <+bridge_> <qetoast> nice
19:54 <+bridge_> <cloud9000.> Does somone want the code and texture for it cuz I can send it
20:13 <+bridge_> <Nouaa> upload it on github ig
20:14 <+bridge_> <cloud9000.> Yay
20:14 <+bridge_> <cloud9000.> ok
20:14 <+bridge_> <cloud9000.> But what about the textures
20:19 <+bridge_> <cloud9000.> https://github.com/Cloud190/Im-very-proud-of-this okay here it is
20:36 <+bridge_> <gorp> honestly? this might be a great reference for me
20:37 <+bridge_> <gorp> also, can we just assume that wrong aspects are just extra X/Y canvas space on the right/boytom? instead of always showing an error at launch?
20:38 <+bridge_> <gorp> sometimes i download assets and they have like a pixel extra or something on the right or bottom edge and I think its not appropriate to always warn on launch, maybe once is fine but its not like the assets don't work properly because of this, and most people editing this wouldn't notice problems anyway
20:40 <+bridge_> <gorp> is cachyos seriously optimized somehow
20:40 <+bridge_> <teero777> @ryozuki idk about the anthropic models but gemini so fucking retarded
20:41 <+bridge_> <teero777> <80 IQ
20:41 <+bridge_> <chillerdragon> @heinrich5991: could you do another phone kyc for me? :)
20:42 <+bridge_> <teero777> and yes im using their newest "groundbreaking" model
20:42 <+bridge_> <teero777> it has 0 reading comprehension
20:43 <+bridge_> <teero777> it executes a program, the program segfaults, and it says "I have completed the implementation and confirmed everything works as expected"
20:44 <+bridge_> <teero777> i write 1 page prompt explaining everything it should do in complete detail with nothing missing and it implements about 1 sentence of it and then calls it a day
20:45 <+bridge_> <teero777> i know from tater that insulting an llm doesn't help. but i'm so mad right now. especially since their cli tool hangs without giving a single thing to know why, on some prompts it just hangs for 30min, restarting doesnt help either. complete timewaste
20:45 <+bridge_> <teero777> by know i would have finished this already
20:45 <+bridge_> <teero777> but no
20:46 <+bridge_> <qetoast> lmao reasonable crashout
20:49 <+bridge_> <Assa> okay, I should have deleted windows apparently years ago
20:49 <+bridge_> <Assa> this is soo much better
20:50 <+bridge_> <gorp> yes
20:50 <+bridge_> <gorp> i feel like Linux has been really good now for like 5+ years
21:01 <+bridge_> <chillerdragon> Decade of the Linux desktop
21:17 <+bridge_> <ryozuki> ye its bad
22:03 <+bridge_> <Assa> xD
22:03 <+bridge_> <Assa> https://cdn.discordapp.com/attachments/293493549758939136/1485383531520725173/Bildschirmfoto_20260322_220302.png?ex=69c1aab7&is=69c05937&hm=b9bedc2a2bfa4e803dc18d7549c4cac5368b035f287b08660b89516b07130c4a&
22:24 <+bridge_> <qetoast> 6 approvals and we got it merged :kek:
23:08 <+bridge_> <melon> :OMEGALUL:
23:09 <+bridge_> <melon> :poggers2:
23:10 <+bridge_> <kebscs> but win32 api..
23:10 <+bridge_> <kebscs> :tear: