23:03 < bridge> Reply from 104.18.11.44: bytes=32 time=5ms TTL=58 23:04 < bridge> 23:04 < bridge> Ping statistics for 104.18.11.44: 23:04 < bridge> Packets: Sent = 4, Received = 4, Lost = 0 (0% loss), 23:04 < bridge> Approximate round trip times in milli-seconds: 23:04 < bridge> Minimum = 5ms, Maximum = 5ms, Average = 5ms` 23:08 < bridge> curl -v master1.ddnet.org 23:08 < bridge> curl -v https://master1.ddnet.org/ddnet/15/register 23:09 < bridge> but im on windows 23:09 < bridge> no 23:09 < bridge> now 23:09 < bridge> bruder 23:10 < bridge> dann nerv jmd anderes 23:10 < bridge> Tja 23:10 < bridge> why 23:10 < bridge> call microsoft support 23:10 < bridge> do i look like bill gates? 23:11 < bridge> u ask me midnight how to fix windows problems 23:12 < bridge> 24/7 ddnet build support team 23:32 < bridge> @rutujek maybe disable ipv6 23:32 < bridge> sv_register ipv4 23:32 < bridge> i think 23:32 < bridge> only ipv6 seems to fail according to your logs 23:33 < bridge> i hope i finished my karma quest with that, gn8 00:01 < bridge> I will test it later 00:01 < bridge> Have a nice dream about codes without error 00:24 < bridge> Wake up, I need Windows support :kek: 01:43 < bridge> don't ask him about big endian... 01:50 < bridge> btw bit fields may completely fuck this up 01:51 < bridge> u can also use them to pad completely arbitrarily 01:52 < bridge> ```c 01:52 < bridge> struct S { 01:52 < bridge> char x : 3; // 3 bits 01:52 < bridge> int : 0; // "advance to int alignment boundary" 01:52 < bridge> char y : 3; // 3 bits 01:52 < bridge> }; 01:52 < bridge> ``` 01:52 < bridge> this is how i know it behaves in cpp but i think c is the same in this regard 01:52 < bridge> ```c 01:52 < bridge> struct S { 01:52 < bridge> char x : 3; // 3 bits 01:53 < bridge> int : 0; // "advance to int alignment boundary" 01:53 < bridge> char y : 3; // 3 bits 01:53 < bridge> }; 01:53 < bridge> ``` 01:53 < bridge> this is how i know it behaves in cpp but i think c is the same in this regard 01:54 < bridge> this shouldnt matter if u dont care about transparent C struct representation tho 04:53 < bridge> anyone knows what's happened? 04:53 < bridge> I found that status.tw has updated 04:56 < bridge> wdym 05:45 < bridge> #5781 can be closed 05:45 < bridge> https://github.com/ddnet/ddnet/issues/5781 06:23 < bridge> are u not a contributor on the repo? seems like a big oversight 06:23 < bridge> are u not a collaborator on the repo? seems like a big oversight 06:48 < bridge> No, its fine 07:03 < bridge> I've never used bit fields xd and I've seen somewhere that they're bad but idk 07:56 < bridge> They are pretty rarely used anyway. I wouldn't sweat over them 08:09 < bridge> is there a way to disable `ctrl + w` bind in firefox, every time i try to write code in godbolt, I try to delete a word and it closes a tab xd 08:26 < bridge> rip 08:32 < bridge> honestly, i barely understand what it generates from a quick look but as long as it spits out correct answer I think it works correctly 08:32 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1271717715635081216/image.png?ex=66b85aea&is=66b7096a&hm=7de0c7f5aafda08687899d8c3f8203ebaef57a3ffda2bcd15fd7a8b45d03553b& 10:29 < ws-client> lerato anti christ https://zillyhuhn.com/cs/.1723278566.png 10:30 < bridge> Dam, now I can't contribute ever again 10:30 < ws-client> exactly 10:30 < ws-client> time to change git email 10:31 < ws-client> so #10 can finally surpas your rank 10:31 < bridge> 777 10:33 < bridge> I should commit 111 white space fix commits 10:33 < ws-client> ez 10:35 < ws-client> @milkeeycat yeet this into your js console ``window.onbeforeunload = (e) => e.preventDefault()`` 10:35 < ws-client> you can also use extensions or bookmarklets to auto load js 10:37 < bridge> it asked before closing a page, is that it? xd 10:37 < ws-client> ye 10:38 < ws-client> you can write a firefox extension in 5 lines of json and then include one line of js with this and tell it in the 5 lines of json to only be active on godbolt 10:39 < bridge> https://tenor.com/view/this-is-fine-fire-house-burning-okay-gif-5263684 10:40 < bridge> ah, ye that could work 10:41 < ws-client> while at it you can also use js to let the ctrl+w do what you want it to do :D 10:42 < bridge> or install an extension which has terminal shortcuts xd 10:51 < ws-client> So C structs have padding? 10:52 < ws-client> meaning i can not trust the result of `sizeof` ? 10:53 < bridge> sizeof return size with padding 10:53 < bridge> sizeof returns size with padding 10:54 < ws-client> a 10:54 < ws-client> So no need to ever worry about padding right? 10:54 < ws-client> ``printf("distance=%ld size=%ld\n", (char *)&m_aNodes[1] - (char *)&m_aNodes[0], sizeof(m_aNodes[0]));`` this indeed prints 16 twice nice 10:55 < ws-client> okay then there is a bug in my debug print again... classic 10:58 < bridge> Unless you are worried about memory footprint. In that case it might matter 10:59 < ws-client> https://zillyhuhn.com/cs/.1723280377.png 10:59 < ws-client> oh no 11:00 < ws-client> can i not count or is that the padding? 11:01 < bridge> is it C repl? :justatest: 11:01 < ws-client> yes 11:01 < ws-client> written in bash 11:01 < ws-client> by me 11:01 < ws-client> totally bug free xd 11:01 < bridge> how did I not think about it 11:01 < ws-client> i use it quite often to test stuff 11:02 < ws-client> so the struct has 2 unsigned ints (4+4) one array of two unsigned shorts (4) and one unsigned char (1) which should add up to 13 bytes right? 11:02 < ws-client> but sizeof says its 16 byte 11:03 < ws-client> i will just assume it is :D 11:03 < ws-client> i would have never thought C introduces padding if I wouldn't have seen you guys talk about it 11:03 < ws-client> #developer education 11:05 < ws-client> @learath2 idc memory i just need to debug my assembly code and i use the C implementation debug prints as reference and the offets were not adding up 11:06 < ws-client> such bloated language this C who would even use it. Its close to machine code they say ... 11:06 < bridge> You can use the packed attribute to get rid of padding 11:06 < ws-client> C pro 11:06 < bridge> ``` 11:06 < bridge> +----+----+----+----+----+----+----+----+----+----+----+----+----+----+----+----+ 11:06 < bridge> | m_Bits | m_NumBits | short1 | short2 | sym| padding | 11:06 < bridge> +----+----+----+----+----+----+----+----+----+----+----+----+----+----+----+----+ 11:06 < bridge> ``` 11:07 < bridge> i guess it should be 16 because it has to also aligned by biggest value in a struct which is unsigned, so by 4 bytes 11:07 < bridge> maybe it's not true tho :justatest: 11:07 < ws-client> thanks milkeey i will take your words as facts 11:08 < ws-client> ill just add padding to my assembly struct and call it fixed 11:08 < bridge> lemme check in my language how it accesses the struct fields 😏 11:10 < bridge> ah, nvm, i dont have 4 bytes ints xd 11:10 < bridge> ? 11:10 < bridge> :D 11:11 < bridge> i have only char and short 11:11 < bridge> and ptr size 11:11 < bridge> nice, u dont need more 11:12 < bridge> i think if i would need arrays when call C functions ill just add a bunch of shorts and pass a pointer to first one 11:12 < ws-client> 4 byte ints are my favorite ints 11:14 < bridge> ints suck 11:14 < bridge> unsigned int ftw 11:14 < ws-client> https://zillyhuhn.com/cs/.1723281263.png 11:15 < bridge> - Can we have ints? 11:15 < bridge> - We have ints at home 11:15 < ws-client> @milkeeycat shouldn't that be over 1000 if it matches the biggest value? 11:15 < bridge> Ints at home: 11:15 < bridge> ``` 11:15 < bridge> ``` 11:15 < bridge> - Can we have ints? 11:15 < bridge> - We have ints at home 11:15 < bridge> Ints at home: 11:15 < bridge> ``` 11:15 < bridge> struct u32 { 11:15 < bridge> u16 low; 11:15 < bridge> u16 high; 11:15 < bridge> } 11:15 < bridge> ``` 11:15 < ws-client> lgtm 11:15 < bridge> iks de 11:16 < bridge> https://www.youtube.com/watch?v=iMwepyyaj8I :O (me seeing open hardware and daydreaming about making my own stuff ^^, just daydreaming tho) 11:16 < bridge> @milkeeycat if u add 512bit ints, then you have more possibilities than there are atoms in the universe 11:16 < bridge> do you not think that should be the default? 11:17 < bridge> all i need to be happy is u8 11:17 < ws-client> u8 smol 11:17 < bridge> all those fancy shmancy 128s and 512s for nobos 11:18 < bridge> ok, it's time to add 32/64 bit ints 11:18 < bridge> u512 11:19 < bridge> what would you use it for? xd 11:19 < bridge> femto second precision 11:19 < bridge> count all atoms in the universe 11:20 < bridge> count all atoms in the universe twice 11:20 < bridge> 😏 11:21 < bridge> if u make it a f512 11:21 < bridge> ddnet would finally work above 10000 units 11:21 < bridge> no more precision problems 11:22 < bridge> i just looked how gcc handlers regular `float` and I think I don't need floats in my language 11:22 < bridge> so fixed point numbers? 11:22 < bridge> but PLSSS make them ez to use 11:22 < bridge> that is so important 11:22 < bridge> not even rust is good in that aspect 11:23 < bridge> also adding sqrt support etc is not ez 11:23 < bridge> i saw something about fixed point numbers when I was checking Patiga's twmap crate or smth 11:23 < bridge> so for such stuff i always have to convert between them to floats 11:23 < bridge> suggs 11:23 < bridge> possible he uses it in some places 11:24 < bridge> @patiga what is your experience using them 11:24 < bridge> Native fixed point numbers lets goooooo 11:24 < bridge> honest 11:25 < bridge> big integer numbers are practical for cryptography but idk what else 11:25 < bridge> Biggest issue with fixed precision numbers is accidentally losing precision. You need to bw careful 11:25 < bridge> if u have like a 512 bit prime u can have a field with it and do stuff 11:25 < bridge> to store someone's mom's weight 11:26 < bridge> What if you need to compute TREE(3)? 11:26 < bridge> what is tree 11:27 < bridge> xd 11:27 < bridge> Very big number. Idk details 11:27 < bridge> in twmap, the importance of the fixed point numbers comes from representing the map values properly in types. simply because they are fixed point numbers everywhere in tw (also in snaps) 11:27 < bridge> Something about coloring a tree 11:27 < bridge> you mean the / 1024? 11:28 < bridge> yea, but also often / 32 11:28 < bridge> sometimes / 2**15 11:28 < bridge> / 100 at one point, tho that can't be represented well 11:28 < bridge> oke yeah makes sense 11:29 < bridge> i also use them in my map format, but i use them properly, like colors are normalized fixed point numbers 11:29 < bridge> and honestly i didnt really like working with it 11:29 < bridge> i remember we talked about physics using them 11:29 < bridge> not trivial enough i guess 11:29 < bridge> also GPUs ofc expect f32 11:30 < bridge> well you can use normalized fixed points there 11:30 < bridge> but not custom formats i think 11:30 < bridge> I mean they have a couple of upsides over f32, such as no precision pitfalls and no NaN 11:30 < bridge> tru 11:31 < bridge> thats why i liked them for a format 11:31 < bridge> on the other hand they give you potential overflows 11:47 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1271766753982283776/image.png?ex=66b88896&is=66b73716&hm=1fcfb4d5a15b9d1296d6b4053a0bfad4c0b8f99294e133bb8acc44549d0f69d7& 11:47 < bridge> [OHNOOOO](https://cdn.discordapp.com/emojis/1183682565392039936.webp?size=48&quality=lossless&name=OHNOOOO) 11:48 < bridge> tja 11:49 < bridge> ban 11:52 < bridge> youtube thumbnails be like 12:19 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1271774921227239484/image.png?ex=66b89031&is=66b73eb1&hm=1c92d56be374f561a898411c99cdffdc3cee7fac8dd5295b7cf484f25081b329& 12:20 < bridge> get a vps/root server 12:22 < bridge> or the not recommended variant: open a port in your router and forward the port to your windows pc's port that the gameserver is listening on. 12:22 < bridge> or the not recommended variant: open a port in your router and forward the port to your ~~windows~~ pc's port that the gameserver is listening on. 12:22 < bridge> or the not recommended variant: open a port in your router and forward the port to your ~~windows~~ linux pc's port that the gameserver is listening on. 12:29 < bridge> didn't u get vps already huh 12:41 < bridge> he switched to windows 12:41 < bridge> 🤦‍♂️ 12:48 < bridge> is it? was also my first thought 13:08 < bridge> its vps 13:08 < bridge> with windows 13:08 < bridge> the worst decision of yor life 13:08 < bridge> i dont know how to make in debian 13:08 < bridge> your 13:09 < bridge> i tried with readme 13:09 < bridge> but there was error 13:09 < bridge> tried google 13:09 < bridge> and ai 13:09 < bridge> didnt worked 13:09 < bridge> should continue to google 13:09 < bridge> ai is shit if you don't know what you are doing 13:10 < bridge> ChillerDragon: where i can dm you? 13:22 < bridge> windows vps for teeworlds sounds fun 13:27 < bridge> it's not really that bad tbh just unnecessary 13:50 < bridge> TRACTOR 13:55 < bridge> @zillyhuhn on wire. Or chillerdragon@gmail.com on mail. Usually also on irc and in game but not too much this year. 13:57 < bridge> Oh or on matrix idk my nick xd chillerdragon 13:57 < bridge> chillerdragon:matrix.org or something like that 14:11 < bridge> ChillerDragon: chat zillyhuhn has private channels? 14:12 < bridge> Not yet. Sometimes I think about it but not sure if it’s worth the effort 14:13 < bridge> or maybe channel that not bridge'd with discord? 14:14 < bridge> chat zillyhuhn has voice channels? 14:14 < bridge> @matodor: yes \#off-topic is not bridged messages there will probably be red by less than 2 people 14:14 < bridge> Jopsti and milkeey 14:14 < bridge> give me register token 😄 14:15 < bridge> Forgot it 14:15 < bridge> It’s in the discord history of this channel 14:17 < bridge> There is two secrets. One is the token which allows you to sign up. And the other is the universal password that lets you login in to any username without account xd 14:17 < bridge> both in chat history? 😄 14:18 < bridge> I think so 14:18 < bridge> That’s the universal pw I think 14:19 < bridge> I’ll get my laptop sec 14:20 < ws-client> ah lmao 14:20 < ws-client> the sign up token is `nimrocks` 14:20 < ws-client> <:gigachad:960830678465933403> 14:21 < ws-client> elo 14:22 < bridge> elo 14:28 < bridge> chiller the person of trust 14:38 < bridge> <0xdeen> @fokkonaut Are you looking into the rest for https://github.com/ddnet/ddnet/pull/8542 ? 14:42 < ws-client> @0xdeen Are you looking into the rest for https://github.com/ddnet/ddnet/pull/5949 14:43 < bridge> :KEKW: 14:45 < bridge> <0xdeen> no, why would I? 14:46 < ws-client> :c if not you who else? 14:47 < bridge> <0xdeen> I probably missed the part of why we are doing 0.7 support. Is anyone using TW 0.7? 14:47 < ws-client> Not a lot because ddnet does not support it 14:47 < ws-client> this pr fixes that 14:47 < bridge> <0xdeen> Probably too late to bring that up after you made a 6000 line PR 14:47 < bridge> <0xdeen> Or because Vanilla Teeworlds gets boring after a while? 14:48 < ws-client> the pr did not start with 6k lines 14:48 < ws-client> lots of it is json files but yes it took me a while 14:48 < bridge> <0xdeen> Hm, I thought the direction we wanted to go was to split further from Teeworlds, not go back to it 14:48 < ws-client> So you are against supporting 0.7? 14:48 < bridge> <0xdeen> Yes 14:49 < ws-client> oof 14:49 < ws-client> thats unexpected now since you did not mention that and even said you will merge it at some point 14:49 < bridge> <0xdeen> This is so much effort to support playing with the 20 people playing TW 0.7 :/ 14:49 < ws-client> But its me who did the effort no? 14:51 < bridge> <0xdeen> Yes, but now this PR can break all the existing 0.6 support in subtle ways 14:51 < bridge> <0xdeen> and we have to maintain it in the future 14:51 < ws-client> I am not going anywhere 14:51 < ws-client> In january i will be doing less development but i will stay here and play the game. If you need 0.7 maintenance you can ping me 14:52 < ws-client> holy sh*t 14:54 < ws-client> wish all your love for 07 could be used to develop APIs for mods <:heartw:395753947396046850> 14:54 < ws-client> I would have fixed the conflict my self if id knew you merge it woah 14:57 < bridge> CHILLER 14:57 < bridge> YOU DID IT 14:57 < bridge> the worst day in my life 14:57 < bridge> :KEKW: 14:58 < ws-client> 0.7 support by deen -.- 14:58 < ws-client> https://zillyhuhn.com/cs/.1723294706.png 14:58 < ws-client> sori jopsti 14:59 < bridge> time to rq 14:59 < ws-client> nono 14:59 < ws-client> dw 14:59 < ws-client> when heinrich returns from afk he will send revert pr 15:02 < bridge> the effort to support 0.7 is already done. Just merge it :0 15:02 < bridge> press the button 15:02 < bridge> :justatest: 15:02 < ws-client> he did @jxsl13 15:02 < bridge> he did D:? 15:02 < bridge> what 15:02 < ws-client> ikr 15:03 < bridge> the hell's going on 15:03 < bridge> sick move 15:03 < bridge> party time, I guess 15:03 < bridge> 🥳 15:06 < bridge> holy shit 15:06 < bridge> best salesman ever 15:09 < bridge> chiller breaking shit in real time 15:10 < ws-client> yea i can't believe i aborted the merge 15:10 < ws-client> my ego got in the way i had to merge the conflict my self 15:10 < bridge> TIL u cant push after it's been added to the merge queue 15:10 < ws-client> it stops the merge queue 15:10 < bridge> or ig u can 15:10 < bridge> just not if u want it merged 15:10 < bridge> xddd 15:17 < bridge> Pepew. One more piece of code we have to maintain forever for the deadforkcompat 15:17 < bridge> maybe chiller will stop contributing to ddnet one day and the you will be able to drop 0.7 xd 15:21 < ws-client> heinrich is working on some translation layer thingy that should make it easier to maintain and also easier to drop 15:21 < bridge> <0xdeen> Oh? But then we probably shouldn't merge that already, should we? 15:22 < ws-client> well he did not say we should wait or that he will finish or that it conflicts 15:22 < ws-client> he told me to ask another maintainer if i want it merged 15:22 < ws-client> so i assume he is okay with it but we can also wait for him to return @0xdeen i just think if we don't merge now nothing will happen 15:23 < ws-client> by now i don't mean "now" but in the next few weeks 15:28 < bridge> Mergus maximus 15:28 < bridge> 🍷 :gg: 15:29 < ws-client> <:brownbear:346683497701834762> 15:30 < bridge> <0xdeen> @chillerdragon Is it intentional that you can't change the Tee 0.7 settings without joining a 0.7 server? 15:30 < ws-client> yes its a feature i suggested and heinrich and jopsti appreciated 15:30 < ws-client> originally it showed both menus at all times 15:31 < ws-client> the 0.7 is hidden away to avoid confusion why 0.7 dont show 15:31 < ws-client> 0.7 skins don't show* 15:31 < bridge> <0xdeen> yes, makes sense, ok for me 15:33 < ws-client> amazing thanks for the merge @0xdeen this was the most emotional merge for me it feels unreal ngl 15:33 < bridge> unrelated but is there any reason u can fathom why my build would fail here? 15:33 < bridge> ``` 15:33 < bridge> [ 54%] Building CXX object CMakeFiles/game-client.dir/src/game/generated/checksum.cpp.o 15:33 < bridge> /home/ewan/ddnet/ddnet/build/src/game/generated/checksum.cpp:1:10: fatal error: engine/client/checksum.h: No such file or directory 15:33 < bridge> 1 | #include 15:33 < bridge> | ^~~~~~~~~~~~~~~~~~~~~~~~~~ 15:33 < bridge> ``` 15:33 < bridge> as long as src is added in target_/include_directories this should be impossible 15:33 < ws-client> unedited ddnet code base? 15:33 < bridge> ya i think it has been a long time in the making 15:34 < bridge> very excited for u 15:34 < bridge> essentially 15:34 < bridge> o can u see replies on irc 15:34 < ws-client> i got better at guessing 15:34 < bridge> lol 15:34 < ws-client> what does essentially mean? does it work for unedited ddnet? 15:34 < ws-client> can you try to delete the build cache? 15:35 < ws-client> the checksum depends on some pyson scripts 15:36 < bridge> it's failing in a new and exciting way 15:36 < bridge> <0xdeen> Thanks for the feature, it seems to work great from what I tried 15:36 < ws-client> yes thanks to all the people who tested it and reported bugs 15:38 < bridge> idk 15:39 < bridge> Odd question But did you try a Clean rebuild and it still fails there ? 15:39 < bridge> :pepeW: 15:39 < bridge> :kek: its mostly the easy Things 15:40 < bridge> the real q here is what kind of shell are using when deving ddnet 15:40 < bridge> mine busted 15:45 < bridge> ```nix 15:45 < bridge> { pkgs ? import { } }: 15:45 < bridge> 15:45 < bridge> let 15:45 < bridge> name = "ddnet"; 15:45 < bridge> in ( 15:45 < bridge> pkgs.mkShell { 15:45 < bridge> name = name; 15:45 < bridge> 15:45 < bridge> nativeBuildInputs = with pkgs;[ 15:45 < bridge> gcc 15:45 < bridge> mold 15:45 < bridge> buildPackages.cmake 15:45 < bridge> ninja 15:45 < bridge> pkg-config 15:45 < bridge> 15:45 < bridge> rustc 15:45 < bridge> cargo 15:46 < bridge> rustPlatform.cargoSetupHook 15:46 < bridge> ]; 15:46 < bridge> 15:46 < bridge> nativeCheckInputs = with pkgs; [ 15:46 < bridge> gtest 15:46 < bridge> ]; 15:46 < bridge> 15:46 < bridge> buildInputs = with pkgs; [ 15:46 < bridge> fish 15:46 < bridge> 15:46 < bridge> SDL2.dev 15:46 < bridge> zlib.dev 15:46 < bridge> 15:46 < bridge> HE USES NIX 15:46 < bridge> and 15:46 < bridge> `cmake -B build -S . -DCMAKE_CXX_FLAGS="-Wno-format-security" -DCMAKE_EXE_LINKER_FLAGS="-fuse-ld=mold" -DCMAKE_SHARED_LINKER_FLAGS="-fuse-ld=mold" && cmake --build build --parallel` 15:46 < bridge> Letsgoooooo 15:46 < bridge> ya 15:47 < bridge> Use inputsFrom = [ ddnet ] 15:47 < bridge> Easier than to define all depa 15:47 < bridge> Easier than to define all deps, espacially when the actual Thing is packaged 15:48 < bridge> i wrote it based off the default.nix on nixpkgs but many of those pkgs were not found when i was using buildFHSEnv 15:48 < bridge> i had it compiling earlier but it couldnt load freetype .so 15:49 < bridge> and others 15:49 < bridge> dont remember what it was when it worked 15:49 < bridge> dont remember what it was when it "worked" 15:49 < bridge> Also is there a reason to specify Fish there ? 15:49 < bridge> 15:49 < bridge> Yeah you can do inputsFrom to get the stuff from the actual package, for development and stuff you can then Just include them seperately with buildInputs 15:49 < bridge> im using fish 15:49 < bridge> Try zsh :gigachad: 15:50 < bridge> lol 15:50 < ws-client> bash best 15:50 < bridge> No Idea why it fails, lgtm 15:51 < bridge> i can use right side prompt w fish 15:51 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1271828268558848062/image.png?ex=66b8c1e0&is=66b77060&hm=46050d65f14faa74d061a5ffbf7902eec8bd050d4c5ed98dd617d5ecc6a65f47& 15:51 < bridge> ``` 15:51 < bridge> outputs = { 15:51 < bridge> nixpkgs, 15:51 < bridge>