01:17 <+bridge> <soulyvevo> yes 🥹
05:24 <+bridge> <corneum> 1 second into the run. looks like more than 4 blocks from the start to me. i need something to tell me if i got a fast swing or slow swing because you cant really tell otherwise. on this map specificly, you need to hold your speed for so long that it adds up
05:24 <+bridge> <corneum> https://cdn.discordapp.com/attachments/293493549758939136/822482005148434522/screenshot_2021-03-19_15-41-05.png?ex=6a505dcf&is=6a4f0c4f&hm=933ee2a7c4e831ecfca862974139ad231efd5bc544176fa5f5ee32dfa1741e2f&
10:32 <+bridge> <Assa> I get this error currently on the steam beta:
10:32 <+bridge> <Assa> ```
10:32 <+bridge> <Assa> 2026-07-09 10:31:47 I sdl: SDL version 2.32.10 (compiled = 2.0.6)
10:32 <+bridge> <Assa> 2026-07-09 10:31:47 I gfx: Created Vulkan 1.1 context
10:32 <+bridge> <Assa> 2026-07-09 10:31:48 E gfx/vulkan: Initialization failed for unknown reason.
10:32 <+bridge> <Assa> 2026-07-09 10:31:48 W gfx/vulkan: Initialization failed for unknown reason.
10:32 <+bridge> <Assa> 2026-07-09 10:31:48 W gfx: Failed to initialize graphics. Setting GL version 1.1.0 and trying again.
10:32 <+bridge> <Assa> 2026-07-09 10:31:48 I sdl: SDL version 2.32.10 (compiled = 2.0.6)
10:32 <+bridge> <Assa> 2026-07-09 10:31:48 I gfx: Created OpenGL 3.0 context
10:32 <+bridge> <Assa> X Error of failed request:  BadValue (integer parameter out of range for operation)
10:32 <+bridge> <Assa>   Major opcode of failed request:  150 (GLX)
10:32 <+bridge> <Assa>   Minor opcode of failed request:  3 (X_GLXCreateContext)
10:32 <+bridge> <Assa>   Value in failed request:  0x0
10:32 <+bridge> <Assa>   Serial number of failed request:  326
10:32 <+bridge> <Assa>   Current serial number in output stream:  327
10:32 <+bridge> <Assa> ```
10:32 <+bridge> <Assa> never seen this before :thonk:
10:33 <+bridge> <Assa> oh might be a case of "installed updates and didn't reboot you idiot"
11:57 <+bridge> <chillerdragon> Yes it is
11:57 <+bridge> <chillerdragon> Seen it before when I forgot to reboot arch after updating
11:59 <+bridge> <chillerdragon> Maybe there is a way to detect this version mismatch at runtime and show a more useful error message
12:19 <+bridge> <heinrich5991> you could also PR it to the ddnet-translations repo
12:26 <+bridge> <Assa> is there a reason the discord button has a slightly different alpha?
12:26 <+bridge> <Assa> https://cdn.discordapp.com/attachments/293493549758939136/1524723459710783579/screenshot_2026-07-09_12-26-20.png?ex=6a50c8e0&is=6a4f7760&hm=a049f5a0daf48cf9fa05fb99bde7ce11c6f6a44edbe7d10d7feec767e9aa540c&
12:27 <+bridge> <Assa> for "news" it's obviously intentional, but this doesn't look intentional to me
12:30 <+bridge> <heinrich5991> I don't see that in the code
12:31 <+bridge> <heinrich5991> they're all `ColorRGBA(0.0f, 0.0f, 0.0f, 0.25f)`
12:33 <+bridge> <reitw> could be the background fade ?
12:34 <+bridge> <reitw> could be the background color fading ?
12:34 <+bridge> <reitw> nvm wouldn't be like this
12:35 <+bridge> <Assa> no, I am adressing an issue that looks almost unpatchable to me, since it involves a default config variable which needs change
12:36 <+bridge> <heinrich5991> tell me about it
12:36 <+bridge> <heinrich5991> maybe we can find a solution 🙂
12:39 <+bridge> <Assa> I am currently implementing https://github.com/ddnet/ddnet/issues/12155 properly, with coloring the UI instead of the background. But the default value for g_Config.m_UiColor already has already a value of `// 160 70 175 228` hasalpha which is not black and not at alpha 255
12:40 <+bridge> <Assa> I guess I can fix the last point in this, so you can make the UI more opaque
12:41 <+bridge> <heinrich5991> if you change the default value of a config option, people with the existing default will get the new default
12:42 <+bridge> <heinrich5991> that's desirable here, no?
12:43 <+bridge> <Assa> do they? Do we track if a person set the default value themselves?
12:45 <+bridge> <Assa> I guess with this we support full-contrast-mode
12:45 <+bridge> <Assa> https://cdn.discordapp.com/attachments/293493549758939136/1524728171822579712/screenshot_2026-07-09_12-45-10.png?ex=6a50cd44&is=6a4f7bc4&hm=8e2e230612e06ff5eca41e1a9c8c4f92ad98406766d26b4e04c63f61a29bbe3c&
12:45 <+bridge> <heinrich5991> no
12:46 <+bridge> <heinrich5991> yes. no
12:49 <+bridge> <jxsl13> two stacked diecord buttons 👀?
12:49 <+bridge> <jxsl13> two stacked discord buttons 👀?
12:50 <+bridge> <Assa> no just alpha 0.5 instead of 0.25
12:50 <+bridge> <jxsl13> hmhm
12:51 <+bridge> <heinrich5991> I see
12:51 <+bridge> <heinrich5991> ```c++
12:51 <+bridge> <heinrich5991>         if(GameClient()->m_Menus.DoButton_Menu(&s_DiscordButton, Localize("Discord"), 0, &Button, BUTTONFLAG_LEFT, nullptr, IGraphics::CORNER_ALL, 5.0f, 0.0f, ColorRGBA(0.0f, 0.0f, 0.0f, 0.25f)))
12:51 <+bridge> <heinrich5991>         if(GameClient()->m_Menus.DoButton_Menu(&s_LearnButton, Localize("Learn"), 0, &Button, BUTTONFLAG_LEFT, nullptr, IGraphics::CORNER_ALL, 5.0f, 0.0f, ColorRGBA(0.0f, 0.0f, 0.0f, 0.25f)))
12:51 <+bridge> <heinrich5991> ```
12:51 <+bridge> <heinrich5991> both 0.25f alpha
12:52 <+bridge> <robyt3> I don't think the alpha is supposed to apply to the offline menu. I use alpha ingame but I wouldn't want the offline menu alpha to be changed.
12:53 <+bridge> <Assa> I guess I must have made some copy pasta then ...
12:53 <+bridge> <Assa> yeah 🤦‍♂️
12:55 <+bridge> <Assa> I can address this
12:56 <+bridge> <robyt3> How? You'd probably have to separate it into three options: ingame menu, offline menu, background color of the checker-background when not using a theme.
12:58 <+bridge> <Assa> by
12:58 <+bridge> <Assa> - making the background none-theme constantly multiplied by the current value, so it's not configurable
12:58 <+bridge> <Assa> - making alpha only apply ingame
12:58 <+bridge> <Assa> btw the scoreboard is also not colored in the menu theme
13:02 <+bridge> <Assa> yes this would end up like this - or I introduce a background UI color variable as you suggest
13:02 <+bridge> <Assa> https://cdn.discordapp.com/attachments/293493549758939136/1524732530459410453/screenshot_2026-07-09_13-02-08.png?ex=6a50d153&is=6a4f7fd3&hm=48698c086789bf6d8c8b133bbfbacecafaeda508796c8ca882a65b646148745b&
13:03 <+bridge> <robyt3> Scoreboard isn't really a menu.
13:03 <+bridge> <robyt3> This doesn't look good I think.
13:03 <+bridge> <heinrich5991> hmmm, but it does belong to the UI
13:04 <+bridge> <heinrich5991> I wonder what people use the UI color variable for, and whether it makes sense in their use case to color the scoreboard
13:05 <+bridge> <robyt3> Ingame I mostly care about the alpha. But offline I want to change the background color instead.
13:07 <+bridge> <Assa> I am thinking about blending these colors
13:12 <+bridge> <Assa> Blending them works (even tho I don't do this correct according to color experts 🤷‍♂️ )
13:12 <+bridge> <Assa> https://cdn.discordapp.com/attachments/293493549758939136/1524734933220986930/screenshot_2026-07-09_13-11-36.png?ex=6a50d390&is=6a4f8210&hm=b70b0abbe91c74c947ec73f61dcdb351397d2beddf688715759c30db1a81c2d6&
13:12 <+bridge> <Assa> yes this extreme case looks ugly - but if you want to have the client look like pink unicorns, I won't stop you
13:16 <+bridge> <robyt3> Having the background and UI with the same color generally seems less readable.
13:17 <+bridge> <Assa> do we have a theme as default?
13:18 <+bridge> <robyt3> Yes, the seasons-theme is the default
13:21 <+bridge> <Assa> I don't know what to do with this issue now - either we close it and let it be like it is, make this color the UI properly or remove this "feature" - right now it looks only broken to me, if I have a theme activated it does nothing in the offline menu and only colors the menu ingame, but ingame everything is colored then
13:22 <+bridge> <robyt3> The minimal change to me would be separating the menu background color and moving this setting next to the theme selection. Then rename the existing UI color to ingame UI color and also move it (it doesn't belong in the graphics settings I think).
13:34 <+bridge> <Assa> @heinrich5991 interesting that we have a hook state that can be reached for 1 tick which just does nothing <.< I wonder what the implication is
13:39 <+bridge> <Assa> I guess this is just an extra tick where the hook retracts
13:47 <+bridge> <patiga> https://discord.com/channels/252358080522747904/1524743635374051430/1524743635374051430 built a neat thing you might find fun :)
14:40 <+bridge> <chillerdragon> wat the thing do?
14:40 <+bridge> <heinrich5991> identify maps based on tile collision patterns
14:40 <+bridge> <heinrich5991> quite neat
14:41 <+bridge> <cellegenrih> so the reason is that an extra frame needed to identify whether the hook is in a block or not
14:41 <+bridge> <cellegenrih> that sounds cheesy
14:41 <+bridge> <cellegenrih> cuz, you could just count to 2 frames from the previous state instead and achieve the same effect
14:43 <+bridge> <cellegenrih> But either way, since it's just a small impact to the whole implementation, I wouldn't mess with it imo
14:48 <+bridge> <chillerdragon> Like a map finder based on a partial match?
14:48 <+bridge> <chillerdragon> What is it for? To cheat in geo guesser?
15:13 <+bridge> <Assa> you could go implement Tee guesser - have a screenshot from a map and guess which it is
15:13 <+bridge> <teero777> doesnt it already exist?
15:14 <+bridge> <cellegenrih> oooor, have a map segment as your guess and you have to go through each map to find it. The farther your chosen map's index from the actual guess, the worse points you get
15:14 <+bridge> <cellegenrih> And the closer you get your guess to the map segment, the better bonus points you get
15:15 <+bridge> <cellegenrih> maybe indexing should go in map creation date order
15:18 <+bridge> <patiga> chillerdragon: to recognize maps on screenshots based on tile patterns. I already have the neural image recognition model, which can't deal with entities properly.
16:03 <+bridge> <sedonya> I'm a bit confused, why roll back the version and why isn't this commit on a branch??
18:21 <+bridge> <fokkonaut> rip chillerdragon
18:22 <+bridge> <fokkonaut> write it in rust and its merged
18:32 <+bridge> <jxsl13> chillers adventures in the ddnet codebase feat. github PRs.
18:33 <+bridge> <jxsl13> chillers adventures in the ddnet codebase feat. github ~~PRs~~issues.
18:33 <+bridge> <jxsl13> chillers adventures in the ddnet codebase feat. github ~~PRs~~ issues.
18:45 <+bridge> <robyt3> Our releases are generally tags and not branches. Looks like the release tag was rebased on a later commit from 20.0 again, so for the release the version needed to be reverted.
19:43 <+bridge> <crhxstophxr> https://cdn.discordapp.com/attachments/293493549758939136/1524833405089353758/1783100302490.jpeg?ex=6a512f45&is=6a4fddc5&hm=c5f71d14c0c5d28fe0cd95491a00b909585416bd09c14a60b134dca01e677c71&
19:43 <+bridge> <crhxstophxr> https://cdn.discordapp.com/attachments/293493549758939136/1524833405395534096/1781977017646.jpeg?ex=6a512f45&is=6a4fddc5&hm=357a3f17888b5f485e7116b824a2d2698f21cabcd29eabb6a5785a4d5ce8ab04&
19:52 <+bridge> <milkeeycat> https://andrewkelley.me/post/my-thoughts-bun-rust-rewrite.html
19:53 <+bridge> <aegisub> he was pulling no punches
19:54 <+bridge> <aegisub> > Jarred was already writing slop well before he had access to LLMs.
20:23 <+bridge> <louis> sour blog post lmfao
21:33 <+bridge> <cellegenrih> I need small help with my custom Variant type in https://github.com/VertaireStudio/RatLab/blob/main/Core/Types/variant.hpp
21:33 <+bridge> <cellegenrih> How should I return a function in a way, that based on a switch case, I return the type itself and not just a pre-made type to typeof() from?
21:34 <+bridge> <cellegenrih> I would save a ton of space if I get a function like that working
21:56 <+bridge> <melon> could try a visitor template that simply calls func(val) per member
21:56 <+bridge> <melon> 
21:56 <+bridge> <melon> i.e
21:56 <+bridge> <melon> 
21:56 <+bridge> <melon> ```cpp
21:56 <+bridge> <melon> template<typename T>
21:56 <+bridge> <melon> void visitor(T&& func) {
21:56 <+bridge> <melon>     switch (current_type) {
21:56 <+bridge> <melon>         case TYPES::BOOL: func(value_bool); return;
21:56 <+bridge> <melon>         case TYPES::U8:   func(value_u8);   return;
21:56 <+bridge> <melon>         ...
21:56 <+bridge> <melon>     }
21:56 <+bridge> <melon> }
21:56 <+bridge> <melon> ```
21:56 <+bridge> <melon> 
21:56 <+bridge> <melon> you'd then just do
21:56 <+bridge> <melon> 
21:56 <+bridge> <melon> ```cpp
21:56 <+bridge> <melon> visitor([&](auto &...) {...}
21:56 <+bridge> <melon> ```
21:56 <+bridge> <melon> within your setter
21:59 <+bridge> <melon> all of your
21:59 <+bridge> <melon> 
21:59 <+bridge> <melon> ```cpp
21:59 <+bridge> <melon> func void set(X p_value) { current_type == p_value.get_type() ? value_bool = p_value : get_value<typeof(p_value)>() = (typeof(p_value))p_value; }`
21:59 <+bridge> <melon> ```
21:59 <+bridge> <melon> 
21:59 <+bridge> <melon> would collapse to a single
21:59 <+bridge> <melon> ```cpp
21:59 <+bridge> <melon> template<typename T>
21:59 <+bridge> <melon> void set(T p_value) {
21:59 <+bridge> <melon>     visitor([&](auto& stored) {
21:59 <+bridge> <melon>         stored = static_cast<std::decay_t<decltype(stored)>>(p_value);
21:59 <+bridge> <melon>     });
21:59 <+bridge> <melon> }
21:59 <+bridge> <melon> ```
21:59 <+bridge> <melon> 
21:59 <+bridge> <melon> tho this drops the type-match check and always coerces into whatevers active, so it will fail instead of silently working now :D
22:00 <+bridge> <melon> all of your
22:00 <+bridge> <melon> ```cpp
22:00 <+bridge> <melon> func void set(X p_value) { current_type == p_value.get_type() ? value_bool = p_value : get_value<typeof(p_value)>() = (typeof(p_value))p_value; }`
22:00 <+bridge> <melon> ```
22:00 <+bridge> <melon> would collapse to a single
22:00 <+bridge> <melon> ```cpp
22:00 <+bridge> <melon> template<typename T>
22:00 <+bridge> <melon> void set(T p_value) {
22:00 <+bridge> <melon>     visitor([&](auto& stored) {
22:00 <+bridge> <melon>         stored = static_cast<std::decay_t<decltype(stored)>>(p_value);
22:00 <+bridge> <melon>     });
22:00 <+bridge> <melon> }
22:00 <+bridge> <melon> ```
22:00 <+bridge> <melon> tho this drops the type-match check and always coerces into whatevers active, so it will fail instead of silently working now :D
22:00 <+bridge> <melon> could try a visitor template that simply calls func(val) per member
22:00 <+bridge> <melon> 
22:00 <+bridge> <melon> i.e
22:00 <+bridge> <melon> ```cpp
22:00 <+bridge> <melon> template<typename T>
22:00 <+bridge> <melon> void visitor(T&& func) {
22:00 <+bridge> <melon>     switch (current_type) {
22:00 <+bridge> <melon>         case TYPES::BOOL: func(value_bool); return;
22:00 <+bridge> <melon>         case TYPES::U8:   func(value_u8);   return;
22:00 <+bridge> <melon>         ...
22:00 <+bridge> <melon>     }
22:00 <+bridge> <melon> }
22:00 <+bridge> <melon> ```
22:01 <+bridge> <melon> you'd then just do
22:01 <+bridge> <melon> ```cpp
22:01 <+bridge> <melon> visitor([&](auto &...) {...}
22:01 <+bridge> <melon> ```
22:01 <+bridge> <melon> within your setter
22:01 <+bridge> <melon> ah wait no it wouldnt... gimme a second i'll test it out
22:07 <+bridge> <cellegenrih> it's kinda an issue when I cannot identify what is and isn't supported
22:07 <+bridge> <cellegenrih> Variant should only allow specific types to interact with it, and likewise - Variant interacting with only the types it supports
22:08 <+bridge> <melon> could wrap it in a variant concept
22:08 <+bridge> <melon> 
22:08 <+bridge> <melon> ```cpp
22:08 <+bridge> <melon> template<typenam T>
22:08 <+bridge> <melon> concept Variant = std::same_as<T, bool> || std::same_as<T, uint8_t> || ...
22:08 <+bridge> <melon> ```
22:08 <+bridge> <melon> could wrap it in a variant concept
22:08 <+bridge> <melon> 
22:08 <+bridge> <melon> ```cpp
22:08 <+bridge> <melon> template<typename T>
22:08 <+bridge> <melon> concept VariantA = std::same_as<T, bool> || std::same_as<T, uint8_t> || ...
22:08 <+bridge> <melon> ```
22:08 <+bridge> <cellegenrih> std::same_as() huh
22:09 <+bridge> <melon> https://en.cppreference.com/cpp/concepts/same_as
22:09 <+bridge> <melon> 
22:09 <+bridge> <melon> its awesome ^^
22:12 <+bridge> <cellegenrih> thing is, I don't want to condense it down to a boolean, but rather just returning the type itself, so that I don't need to perform checks just to pass 'typeof(foo)'
22:13 <+bridge> <cellegenrih> Cuz that way, I can make a better 'get_value()' without worrying about the T template so much
22:15 <+bridge> <cellegenrih> as in, when having a Variant with a boolean type of 'true' as current:
22:15 <+bridge> <cellegenrih> 
22:15 <+bridge> <cellegenrih> get_current_type() should return 'Bool', as the type itself
22:15 <+bridge> <cellegenrih> get_current_value() should return 'true' as the type's value
22:15 <+bridge> <cellegenrih> I don't wanna AI it, cuz I know it will be even more unreadable
22:44 <+bridge> <jxsl13> where is matricks currently working at? only reading bad news for idTech 👀
22:44 <+bridge> <jxsl13> where is matricks currently working at? only reading bad news about idTech 👀
23:18 <+bridge> <cellegenrih> his best next place to work would be alongside valve with the source engine
23:36 <+bridge> <cellegenrih> @blaiszephyr I was thinking, can't I just merge all the data of Variant into an array which acts as a way to index into whatever variables you want, based on the indexing of 'TYPES'?
23:36 <+bridge> <cellegenrih> would it be stupid or genius
23:43 <+bridge> <melon> depends on how much you control you'd like
23:43 <+bridge> <melon> you're thinking union but rawdogging style
23:43 <+bridge> <melon> which comes with more boilerplate but more control
23:45 <+bridge> <cellegenrih> well, I could private the data anytime, nothing's stopping me from it
23:46 <+bridge> <cellegenrih> but then pointing to those data individually will be tricky
23:47 <+bridge> <cellegenrih> I think having as much control for such a versatile struct as possible is the idea