02:30 < ws-client> <ChillerDragon> @ryozuki i did not find your neovim config on github :C Do you have some lsp settings for ddnet? So it doesn't get confused with include paths
09:07 < bridge> <meloƞ> using clangd You Just define the clangd path in your Compiler Flags for the LSP to be Happy
09:37 < bridge> <milkeeycat> Or just use Mason and everything just works :gigachad:
10:02 < bridge> <Jupstar ✪> Yuzu got discontinued
10:02 < bridge> <Jupstar ✪> Rip
10:02 < bridge> <ryozuki> i want my 7900x3d already
10:02 < bridge> <ryozuki> will make llvm dev easier
10:04 < bridge> <Jupstar ✪> U still don't have it?
10:05 < bridge> <ryozuki> i orderered the whole thing without gpu and m2 and have it built
10:05 < bridge> <ryozuki> ill get it by end of month
10:05 < bridge> <ryozuki> i was lazy and have money so
10:05 < bridge> <ryozuki> xd
10:06 < bridge> <ryozuki> https://cdn.discordapp.com/attachments/293493549758939136/1214499181444726794/image.png?ex=65f95581&is=65e6e081&hm=090ea9e57cccfc65d9bf96836a4dab191f901bd77bc5828d3c3ca34a507d860f&
10:06 < bridge> <ryozuki> here i am once again building llvm
10:06 < bridge> <learath2> I never really liked how they were semi-paywalled
10:06 < bridge> <ryozuki> i dont like nintendo
10:06 < bridge> <ryozuki> or derivatives
10:07 < bridge> <learath2> Probably why they folded so quick. Making money off of emulators is a no no
10:07 < bridge> <ryozuki> the console has 0 appeal
10:07 < bridge> <ryozuki> ppl only use it cuz its a walled garden
10:07 < bridge> <learath2> The appeal is the exclusives
10:07 < bridge> <ryozuki> i laugh at 30 fps gamers
10:07 < bridge> <ryozuki> in 2024
10:07 < bridge> <learath2> If nintendo released their games on other consoles too no one would prefer a switch
10:08 < bridge> <ryozuki> and they laugh at their customers
10:08 < bridge> <ryozuki> with those graphics
10:08 < bridge> <ryozuki> did u see the pokemons
10:08 < bridge> <ryozuki> lmao
10:08 < bridge> <ryozuki> billions of $$ and that
10:08 < bridge> <scrumplex> I never quite understood it too
10:08 < bridge> <scrumplex> They should have just used Patreon as a way to fund development
10:08 < bridge> <scrumplex> and not for early access to new stuff
10:09 < bridge> <learath2> I honestly don't agree with this being an issue. For me if it looks fluid it's fine. Movies are 24fps and we don't complain about that either
10:09 < bridge> <ryozuki> @scrumplex btw im using prism, any modpack u recommend? preferably on modrinth since it seems free software
10:09 < bridge> <ryozuki> i thought of trying these
10:09 < bridge> <ryozuki> https://cdn.discordapp.com/attachments/293493549758939136/1214499960951803985/image.png?ex=65f9563b&is=65e6e13b&hm=3496dedb128b6a4e8d3c64587c5d3548d571c6382d23f63eade4ef01e09026f9&
10:09 < bridge> <scrumplex> At least people were redistributing the early access code as it's G:PLed
10:09 < bridge> <scrumplex> At least people were redistributing the early access code as it's GPLed
10:09 < bridge> <scrumplex> https://github.com/pineappleEA/pineapple-src
10:09 < bridge> <ryozuki> for me its a big issue, i need the smooth
10:09 < bridge> <scrumplex> I usually make my own xD
10:10 < bridge> <ryozuki> i dont think i will make my own xd
10:10 < bridge> <scrumplex> But movies are not interactive
10:10 < bridge> <ryozuki> i like the kitchen sink stuff
10:10 < bridge> <ryozuki> like sky factory 2
10:10 < bridge> <scrumplex> I think it's more about input latency than FPS, but those two are correlated
10:10 < bridge> <learath2> Should probably archive all these repos. Recently I've been noticing governments and companies can actually disappear stuff off the internet
10:10 < bridge> <ryozuki> ill mirror that on my forgejo instance
10:10 < bridge> <scrumplex> Nintendo could probably take this repo down
10:11 < bridge> <learath2> Do you feel movies aren't smooth? Input lag is indeed a far more annoying phenomenon
10:11 < bridge> <ryozuki> @scrumplex btw im hosting forgejo feel free to use it https://git.edgarluque.com/explore/repos
10:11 < bridge> <learath2> Careful. Don't get nintendo'd
10:11 < bridge> <scrumplex> IMO movies being 24fps is fine enough
10:11 < bridge> <ryozuki> ill make it private maybe
10:12 < bridge> <scrumplex> YouTube videos being 30 FPS or 60 FPS is fine. I think fast paced stuff is better if it's 60 FPS, but it's not like I'll complain
10:12 < bridge> <heinrich5991> doubt they can disappear something so popular
10:12 < bridge> <scrumplex> But games have to be at least like 40-50 FPS for me
10:12 < bridge> <scrumplex> But I also think stable framerates are much more important than high framerates
10:13 < bridge> <scrumplex> like 150 FPS with 1% lows at 60 is much more annoying than a stable 60 FPS
10:13 < bridge> <learath2> A couple weeks ago I played a bit of gta san andreas for a bit of nostalgia. The game is locked to 25fps. I did not even think about it until right now since we are talking about it
10:13 < bridge> <learath2> Because the game has exquisite frame pacing, it just feels good
10:13 < bridge> <scrumplex> But GTA SA isn't really a face paced game
10:13 < bridge> <ryozuki> did u play with mta
10:14 < bridge> <scrumplex> But GTA SA isn't really a fast paced game
10:14 < bridge> <learath2> I'm sure it'd feel worse the faster paced the game is, but mario isn't exactly a twitch shooter
10:14 < bridge> <scrumplex> yeah
10:14 < bridge> <learath2> Yep
10:14 < bridge> <scrumplex> Though I did always notice how choppy Animal Crossing was on the Switch. Even though it's a very slow and cozy game
10:14 < bridge> <Jupstar ✪> What, that confuses me even more. U have ur CPU already?
10:15 < bridge> <scrumplex> I think the low framerates are more annoying with games like Zelda
10:15 < bridge> <learath2> I also played through hollow knight on my switch. The 30 fps didn't really bother me that much there either. Though the input does feel less responsive than on pc, I'd be lying if I said I did not get used to it almost instantly
10:15 < bridge> <ryozuki> i dont? xD thats why  isaid i wish i had it
10:15 < bridge> <scrumplex> Again, not very fast paced, but because it has combat I think the low framerates might be annoying
10:16 < bridge> <ryozuki> @jupeyy_keks i ordered to have my pc built
10:16 < bridge> <ryozuki> i got new case and stuff
10:16 < bridge> <Jupstar ✪> I c
10:16 < bridge> <Jupstar ✪> Unepyc moment
10:16 < bridge> <learath2> Though I might have a higher tolerance to lag than the average person. I played games with 90 ping for most my life 😄
10:16 < bridge> <ryozuki> but  i ordered it without gpu or m2
10:16 < bridge> <scrumplex> what are the specs? :O
10:16 < bridge> <ryozuki> sec
10:17 < bridge> <ryozuki> this + a nvidia 3080 and my m2 ssds
10:17 < bridge> <ryozuki> https://cdn.discordapp.com/attachments/293493549758939136/1214501973643235348/image.png?ex=65f9581a&is=65e6e31a&hm=e7acce50d4fa6b6c479780371fbb72ec7d28a42ff27fd8ca583c752c0678b25d&
10:17 < bridge> <scrumplex> I'd argue it heavily depends on the game. 
10:17 < bridge> <scrumplex> Teeworlds/DDNet at 90 ms? HELL NO
10:17 < bridge> <scrumplex> Minecraft at 90 ms? Fine enough
10:17 < bridge> <ryozuki> i got the case big enough for a 4090
10:17 < bridge> <Jupstar ✪> Ez
10:17 < bridge> <scrumplex> i dont recognize the first three brands 👀
10:18 < bridge> <Jupstar ✪> If I'm honest with u, it's a weird time to buy high end
10:18 < bridge> <ryozuki> 300€ mobo xd
10:18 < bridge> <ryozuki> smth happened and i got quite big spare cash
10:18 < bridge> <learath2> I played tw with 90ms for a very long time. Until Turkey got good peerings with the EU. It is indeed annoying
10:18 < bridge> <ryozuki> lets say smth like a extra pay xd
10:18 < bridge> <learath2> And it was before antiping was a thing
10:18 < bridge> <Jupstar ✪> In q3 new gen comes
10:18 < bridge> <ryozuki> then ill buy it
10:18 < bridge> <scrumplex> I think if you have consistent 90ms with anti ping you should be fine for the most part
10:18 < bridge> <ryozuki> :gigachad:
10:19 < bridge> <ryozuki> well ill wait until more revisions come
10:19 < bridge> <Jupstar ✪> Epyc rich moment
10:19 < bridge> <ryozuki> first batch always has issues
10:19 < bridge> <Jupstar ✪> Lmao
10:19 < bridge> <learath2> The hardest part was catching falling tees before antiping. I couldn't play for a while after the internet got better since I was leading my aim so much
10:19 < bridge> <Jupstar ✪> Apple already beta tested 3nm
10:19 < bridge> <scrumplex> I don't agree with your choice of cooler. Arctic released their new Liquid Freezer 3 recently and it's the best cooler full stop. Like it's cheaper than most competitors AND performs the best regarding cooling as well as noise
10:20 < bridge> <scrumplex> I feel this. I don't use Anit Ping because I am trained for 30ms ping
10:20 < bridge> <scrumplex> I feel this. I don't use Anit Ping because I am trained for 20-30ms ping
10:20 < bridge> <ryozuki> oh lol its ok i didnt rly look into it much, as long as it cools
10:20 < bridge> <scrumplex> I feel this. I don't use Anti Ping because I am trained for 20-30ms ping
10:20 < bridge> <ryozuki> wanted to try a liquid cooling since it makes the pc look prettier
10:20 < bridge> <ryozuki> but i think its the same as air cooling
10:21 < bridge> <learath2> Yep, I still don't use it exactly for that reason
10:21 < bridge> <scrumplex> IMO it's nicer to work with, because if you ever need to reach your RAM or anything else that's close to the CPU a big air cooler is always in the way
10:21 < bridge> <ryozuki> yeah
10:21 < bridge> <scrumplex> IMO a liquid cooler nicer to work with, because if you ever need to reach your RAM or anything else that's close to the CPU a big air cooler is always in the way
10:21 < bridge> <ryozuki> and its close to the gpu sometimes
10:21 < bridge> <scrumplex> IMO a liquid cooler is nicer to work with, because if you ever need to reach your RAM or anything else that's close to the CPU a big air cooler is always in the way
10:21 < bridge> <ryozuki> but i think this case allows vertical gpu or smth
10:22 < bridge> <learath2> When I get some money and time on my hand, I'm definitely doing hard tubing liquid cooling. It looks so cool
10:22 < bridge> <scrumplex> Sadly my water cooler's pipes hug the GPU so if I ever want to take out my GPU, I need to take out my CPU cooler
10:22 < bridge> <ryozuki> damn
10:22 < bridge> <Jupstar ✪> Custom loop?
10:22 < bridge> <scrumplex> front mounted the radiator with the pipes coming out at the bottom of the case (so the air bubbles get trapped in the radiator, far away from the pipes
10:22 < bridge> <scrumplex> nah
10:22 < bridge> <ryozuki> the cool thing is, ill have a 2nd home pc!
10:22 < bridge> <ryozuki> with a ryzen 5800x kek
10:22 < bridge> <learath2> I have a noctua nh-d14. I can't do anything in my case without removing it. So annoying
10:23 < bridge> <meloƞ> Even when using Mason, clangd didnt find Std Headers for me when i used cmake :feelsbadman:
10:23 < bridge> <milkeeycat> Make a ddnet tournament, the winner gets the pc
10:23 < bridge> <ryozuki> lol no xd
10:23 < bridge> <ryozuki> i dont even play ddnet
10:23 < bridge> <milkeeycat> Too baf
10:23 < bridge> <milkeeycat> Bad
10:23 < bridge> <ryozuki> baf baf
10:24 < bridge> <ryozuki> https://tenor.com/view/cat-cats-cat-meme-meme-meme-cat-gif-14470917232397934693
10:24 < bridge> <milkeeycat> Bam mentioned
10:24 < bridge> <ryozuki> this cat looks like me looking at C and python coders
10:24 < bridge> <learath2> Wtf. I'm taking your dev role
10:24 < bridge> <ryozuki> lol no xd
10:24 < bridge> <ryozuki> i play like maybe to talk to ppl
10:24 < bridge> <ryozuki> once a week
10:24 < bridge> <learath2> Ok good
10:24 < bridge> <Jupstar ✪> I play it almost every day
10:24 < bridge> <ryozuki> also i got a pr merged a week ago!
10:24 < bridge> <ryozuki> im a active dev
10:24 < bridge> <Jupstar ✪> Best game
10:24 < bridge> <ryozuki> :owo:
10:25 < bridge> <ryozuki> @jupeyy_keks i know its the best game, but i feel i dont have like a partner to play with a lot
10:25 < bridge> <meloƞ> When hostile 3 2p ? Im on at 16:00 :gigachad:
10:25 < bridge> <Jupstar ✪> Actually one reason I play it every day is, that it starts in one second
10:25 < bridge> <meloƞ> When hostile 3 2p ? Im on at 16:00 UTC+2 :gigachad:
10:25 < bridge> <ryozuki> xd
10:25 < bridge> <Jupstar ✪> Most games are horrible loading wise
10:25 < bridge> <ryozuki> well it doesnt have much to load lets be real
10:25 < bridge> <milkeeycat> Best reason 😬
10:26 < bridge> <ryozuki> KISS
10:26 < bridge> <meloƞ> No man's Sky and destiny 2 silently hoping they dont get mentioned
10:26 < bridge> <ryozuki> gta 5 online
10:26 < bridge> <ryozuki> kek
10:26 < bridge> <Jupstar ✪> Ikr. Performance is everything
10:26 < bridge> <meloƞ> :kek:
10:26 < bridge> <Jupstar ✪> Lmao
10:26 < bridge> <Jupstar ✪> I hate this shit so much
10:26 < bridge> <learath2> I play while I'm in queue for league
10:26 < bridge> <ryozuki> ew
10:26 < bridge> <meloƞ> :ban:
10:26 < bridge> <ryozuki> does league have long queues
10:26 < bridge> <ryozuki> the situation reversed
10:26 < bridge> <milkeeycat> Oh yea, u have seen my 5fps crap 😏
10:26 < bridge> <ryozuki> i find fast in dota
10:26 < bridge> <learath2> When I play with my friends yes. Our ranks are too far apart
10:27 < bridge> <ryozuki> do u have friends?
10:27 < bridge> <ryozuki> get out
10:27 < bridge> <Jupstar ✪> U already quintupled it
10:27 < bridge> <Jupstar ✪> Nice
10:27 < bridge> <ryozuki> <a:hotcoffee:635654970275921960>
10:27 < bridge> <learath2> :sadge:
10:27 < bridge> <ryozuki> brb making coffee
10:27 < bridge> <milkeeycat> Rn i go over all tiles
10:27 < bridge> <meloƞ> Send Pics of fishies
10:27 < bridge> <meloƞ> I have to start working again in 3 Minutes
10:27 < bridge> <meloƞ> God Help me
10:28 < bridge> <milkeeycat> I think i can apply some mafs to loop only over elements in view
10:28 < bridge> <ryozuki> @learath2 do u drink coffee with milk or not
10:28 < bridge> <ryozuki> if so how much % of milk
10:28 < bridge> <scrumplex> My PC
10:28 < bridge> <scrumplex> https://cdn.discordapp.com/attachments/293493549758939136/1214504909190471680/1f0d8f04-ade6-443f-901f-d6fc1235eb36.jpg?ex=65f95ad6&is=65e6e5d6&hm=890e9ed119e1ae1b5b4cb63801b9c1979a7eb1ca94dfa6de2636c566e0e3adff&
10:28 < bridge> <scrumplex> https://cdn.discordapp.com/attachments/293493549758939136/1214504909664165888/de72b071-109f-4aee-accc-923c53d25b8f.jpg?ex=65f95ad6&is=65e6e5d6&hm=d427fca3a665d812d2bd6133e402e251a2e7b8a3a42ee902a987dd68041476ea&
10:28 < bridge> <scrumplex> https://cdn.discordapp.com/attachments/293493549758939136/1214504910092242974/66139dc4-08c2-416b-ba46-5af9a729cb58.jpg?ex=65f95ad6&is=65e6e5d6&hm=0bfa37eea8748b54f1796ecc2d3c9a20f4f696be2340a69343b420d669579d4f&
10:29 < bridge> <Jupstar ✪> Coffee -> #off-topic 
10:29 < bridge> <Jupstar ✪> Tea -> #developer
10:29 < bridge> <learath2> I'll usually have it with no milk at home. But sometimes I enjoy a latte
10:29 < bridge> <Jupstar ✪> Hot
10:29 < bridge> <scrumplex> I am honestly not quite happy with my liquid cooler mount
10:30 < bridge> <ryozuki> https://cdn.discordapp.com/attachments/293493549758939136/1214505257673953330/PXL_20240305_092939620.jpg?ex=65f95b29&is=65e6e629&hm=246cf872f6f74a98207c30c50f9e6fa9b790528f82a2c141b9d815130c645148&
10:30 < bridge> <scrumplex> The fans are blowing onto the radiator from the front of the case. Which means all the dust gets pushed into the radiators
10:30 < bridge> <scrumplex> The fans are blowing onto the radiator from the front of the case. Which means all the dust gets pushed into the radiator
10:30 < bridge> <learath2> I like my latte about 30-35% milk, but it's a subjective thing
10:30 < bridge> <ryozuki> same more coffee than milk
10:30 < bridge> <learath2> And it depends on the coffee. Darker roasts can tolerate more milk before you lose the flavors
10:31 < bridge> <ryozuki> lols non gamer
10:31 < bridge> <ryozuki> looks*
10:31 < bridge> <scrumplex> i disabled all the rgb
10:31 < bridge> <Jupstar ✪> Coffee addiction is the weirdest majority human thing
10:31 < bridge> <ryozuki> btw im a gamer and i spilled coffee on my old 200€ keyboard
10:31 < bridge> <ryozuki> but i got a new one
10:31 < bridge> <ryozuki> corsair k70
10:31 < bridge> <ryozuki> are those 2 dacs?
10:32 < bridge> <ryozuki> tbh the liquid cooling doesnt look cool there
10:32 < bridge> <learath2> Why? It's a relatively harmless stimulant that's completely legal. It'd be more weird if a majority of people didn't use it
10:32 < bridge> <ryozuki> and noctua being noctua with the color xd
10:32 < bridge> <scrumplex> dac and amp
10:32 < bridge> <ryozuki> nice
10:32 < bridge> <scrumplex> it was a leftover from my previous case xD
10:33 < bridge> <ryozuki> https://cdn.discordapp.com/attachments/293493549758939136/1214505978607833088/PXL_20240305_093257279.jpg?ex=65f95bd5&is=65e6e6d5&hm=33502c37177b78c03abc82cb6e0a8a87c8ae9eab44fa708b3a8a74fd0c81497a&
10:33 < bridge> <ryozuki> :owo:
10:33 < bridge> <scrumplex> Yeah I think I'd have to flip it for it be okay. But then I might get dripping noises over time
10:33 < bridge> <scrumplex> https://cdn.discordapp.com/attachments/293493549758939136/1214506081603166258/2024-03-05-10-32-25-278.jpg?ex=65f95bee&is=65e6e6ee&hm=8fe0b4c4bb690c2669f0d521843090a885231800e0a571e219bcfb157758369c&
10:33 < bridge> <scrumplex> unfiltered view of my desk
10:33 < bridge> <scrumplex> usually i would clean up a little before taking a picture xD
10:34 < bridge> <ryozuki> my setup
10:34 < bridge> <ryozuki> https://cdn.discordapp.com/attachments/293493549758939136/1214506250373693440/PXL_20240304_093728447.png?ex=65f95c16&is=65e6e716&hm=96e275d516b4e6c0cfbcc18f9cf48a0a2b998831a4cc4ef351666e8d870dd6c7&
10:34 < bridge> <ryozuki> i need a ferris plushie
10:34 < bridge> <learath2> Why are your desks so clean wtf
10:35 < bridge> <scrumplex> why is yours not lol
10:35 < bridge> <ryozuki> xd
10:35 < bridge> <ryozuki> gotta e a refined gamer
10:35 < bridge> <learath2> I'm unsure whether I should show you mine so you can judge me
10:35 < bridge> <ryozuki> kek
10:35 < bridge> <scrumplex> Before anyone calls me out on it: that macbook on the right is from work 🤮
10:36 < bridge> <learath2> The state of this desk is unmanageable
10:36 < bridge> <ryozuki> i also got a m1 from work
10:36 < bridge> <scrumplex> same
10:36 < bridge> <scrumplex> but I want a Linux notebook :((
10:36 < bridge> <scrumplex> I hate macOS' window management
10:36 < bridge> <ryozuki> how can u live with small keyboards
10:36 < bridge> <ryozuki> i cant
10:36 < bridge> <scrumplex> and keyboard shortcuts
10:36 < bridge> <ryozuki> i need 100%
10:36 < bridge> <ryozuki> or 110%
10:36 < bridge> <scrumplex> At least I got my framework hiding next to me in the shelf
10:37 < bridge> <scrumplex> The only problems I have with this is pressind F1-F12 and a number at once
10:37 < bridge> <scrumplex> like in Minecraft to do some debugging stuff (I can't do F3+3 for example)
10:37 < bridge> <ryozuki> xD
10:37 < bridge> <scrumplex> The only problems I have with this is pressing F1-F12 and a number at once
10:38 < bridge> <scrumplex> Also ~ is a bit annoying
10:38 < bridge> <scrumplex> I bound it to Fn+Esc
10:38 < bridge> <scrumplex> or rather Fn+Shift+Esc for ~ and Fn+Esc for `
10:38 < bridge> <scrumplex> also Home/End is a little annoying (Fn+PgUp/PgDown)
10:38 < bridge> <scrumplex> but besides that. I am quite happy with this
10:39 < bridge> <ryozuki> pains
10:39 < bridge> <scrumplex> but besides that, I am quite happy with this
10:39 < bridge> <scrumplex> do you really need a numpad though? :O
10:39 < bridge> <ryozuki> i like it
10:39 < bridge> <ryozuki> i used it in blender
10:41 < bridge> <learath2> Here judge me
10:41 < bridge> <learath2> https://cdn.discordapp.com/attachments/293493549758939136/1214507977332232202/20240305_103938.jpg?ex=65f95db2&is=65e6e8b2&hm=33f17ebc76341663e894c2468c207e05f7d655159502b74ba047d5db8b41b5ca&
10:41 < bridge> <Jupstar ✪> I dunno. It smells like shit. Coffee machines are expensive and often break. They get pretty uncomfortable without their coffee. Coffee itself only grows on very few places in the world.
10:41 < bridge> <Jupstar ✪> 
10:41 < bridge> <Jupstar ✪> It's just weird that so many ppl like it ( for me)
10:41 < bridge> <ryozuki> not bad
10:41 < bridge> <milkeeycat> nice lamp xd
10:41 < bridge> <ryozuki> same mouse? also whats that keyboard
10:42 < bridge> <Jupstar ✪> Nice Rubik's cubes
10:42 < bridge> <ryozuki> filthy anime user!
10:42 < bridge> <ryozuki> https://cdn.discordapp.com/attachments/293493549758939136/1214508258065129512/image.png?ex=65f95df5&is=65e6e8f5&hm=2ec2b8e452371be544ff5fddafad167565e55e5fb9594cf34e2388057d1c0907&
10:42 < bridge> <scrumplex> looks okay to me
10:42 < bridge> <learath2> Smells like shit is the issue. Most people would disagree with you on that one
10:42 < bridge> <ryozuki> i like coffee smell
10:42 < bridge> <scrumplex> lmao
10:42 < bridge> <scrumplex> Coffee smell goos
10:42 < bridge> <learath2> It's a leopold fc660m. So nice to have arrows in a 60%
10:42 < bridge> <scrumplex> Coffee smell good
10:42 < bridge> <Jupstar ✪> And gives bad breath
10:42 < bridge> <scrumplex> Tea yummier though
10:43 < bridge> <scrumplex> like if you want something that keeps you awake, black tea can also do the jon
10:43 < bridge> <scrumplex> like if you want something that keeps you awake, black tea can also do the job
10:43 < bridge> <Jupstar ✪> I only drink herb tea. Without coffeine
10:43 < bridge> <learath2> I wonder if I can still solve them. It's been a while since I was interested in them
10:43 < bridge> <scrumplex> The cup you can see on my desk is fennel tea
10:44 < bridge> <scrumplex> I am such a tea snob that I bought 1kg of fennel seeds that I grind myself and brew tea with
10:44 < bridge> <Jupstar ✪> But tbh I generally only drink tea in winter, if at all
10:44 < bridge> <Jupstar ✪> Water is best
10:44 < bridge> <Jupstar ✪> Water is just so amazing
10:44 < bridge> <learath2> Tea is so good too. Especially some pu'er tea :chefsKiss:
10:44 < bridge> <ryozuki> i have a coworker that cooks the coffee beans or smth i forgot word
10:44 < bridge> <Jupstar ✪> It's the best refresh
10:44 < bridge> <scrumplex> i assume roast? :o
10:44 < bridge> <ryozuki> ye
10:45 < bridge> <ryozuki> xd
10:45 < bridge> <Jupstar ✪> Drinking 500ml water at once feels amazing
10:45 < bridge> <Jupstar ✪> Like u clean your body
10:45 < bridge> <scrumplex> I use one of these things to grind my seeds:
10:45 < bridge> <scrumplex> https://cdn.discordapp.com/attachments/293493549758939136/1214509113761857577/HI2liQr.png?ex=65f95ec1&is=65e6e9c1&hm=7e37b515fe5e48cf49aac8463e3730e294048ef0eb0d552845a855f6ba2c7b76&
10:45 < bridge> <scrumplex> I feel like someone who smokes weed
10:45 < bridge> <ryozuki> sure to grind seeds bro
10:45 < bridge> <ryozuki> :biggest_abuser_and_weed_user:
10:45 < bridge> <scrumplex> I use one of these things to grind my fennel seeds:
10:45 < bridge> <scrumplex> https://cdn.discordapp.com/attachments/293493549758939136/1214509113761857577/HI2liQr.png?ex=65f95ec1&is=65e6e9c1&hm=7e37b515fe5e48cf49aac8463e3730e294048ef0eb0d552845a855f6ba2c7b76&
10:45 < bridge> <learath2> My turkish genes compel me to always drink tea. Even if it's 40degC outside, I'll have atleast my breakfast tea and afternoon tea
10:45 < bridge> <Jupstar ✪> Lmao
10:46 < bridge> <ryozuki> thats called a grinder
10:46 < bridge> <Jupstar ✪> Wow tea in summer is a skill ngl
10:46 < bridge> <Jupstar ✪> I'd probably just die
10:47 < bridge> <scrumplex> turkish (and arabic) tea hits hard ngl
10:47 < bridge> <learath2> Also, the coins I hoarded are the biggest issue. I can't get rid of them because my bank branch doesnt have a coin machine :pepeW:
10:47 < bridge> <ryozuki> i feel like i need a anime figure or smth now
10:48 < bridge> <ryozuki> i will buy a frieren one
10:48 < bridge> <ryozuki> and pray
10:48 < bridge> <ryozuki> https://tenor.com/view/burger-eating-frieren-frieren-beyond-journey%27s-end-sousou-no-frieren-gif-13425073513713719938
10:48 < bridge> <learath2> I used to avoid all anime things. Then I noticed I'm dying alone either way, might aswell enjoy my hobbies
10:48 < bridge> <ryozuki> just hide it if someone comes kek
10:48 < bridge> <ryozuki> :justatest:
10:49 < bridge> <learath2> No one comes. No one will come. So that's a non issue
10:49 < bridge> <ryozuki> not like anyone comes <a:sadSnail:785072987815804929>
10:50 < bridge> <ryozuki> llvm compiled finally
10:50 < bridge> <learath2> :Celebrate:
10:50 < bridge> <ryozuki> ```
10:50 < bridge> <ryozuki> ❯ ./bin/llvm-lit tools/mlir/test/CAPI/
10:50 < bridge> <ryozuki> -- Testing: 10 tests, 10 workers --
10:50 < bridge> <ryozuki> PASS: MLIR :: CAPI/quant.c (1 of 10)
10:50 < bridge> <ryozuki> PASS: MLIR :: CAPI/pdl.c (2 of 10)
10:50 < bridge> <ryozuki> PASS: MLIR :: CAPI/transform_interpreter.c (3 of 10)
10:50 < bridge> <ryozuki> PASS: MLIR :: CAPI/transform.c (4 of 10)
10:50 < bridge> <ryozuki> PASS: MLIR :: CAPI/sparse_tensor.c (5 of 10)
10:50 < bridge> <ryozuki> PASS: MLIR :: CAPI/llvm.c (6 of 10)
10:50 < bridge> <ryozuki> PASS: MLIR :: CAPI/translation.c (7 of 10)
10:50 < bridge> <ryozuki> PASS: MLIR :: CAPI/pass.c (8 of 10)
10:51 < bridge> <ryozuki> PASS: MLIR :: CAPI/ir.c (9 of 10)
10:51 < bridge> <ryozuki> PASS: MLIR :: CAPI/execution_engine.c (10 of 10)
10:51 < bridge> <ryozuki> 
10:51 < bridge> <ryozuki> Testing Time: 0.26s
10:51 < bridge> <ryozuki> 
10:51 < bridge> <ryozuki> Total Discovered Tests: 10
10:51 < bridge> <ryozuki>   Passed: 10 (100.00%)
10:51 < bridge> <ryozuki> ```
10:51 < bridge> <ryozuki> atleast they got nice testing tools
10:51 < bridge> <learath2> llvm-lit
10:51 < bridge> <ryozuki> its lit
11:07 < bridge> <sebaxddddd> She has such a juicy ass
11:07 < bridge> <sebaxddddd> Fuck you can jerk off here - https://discord.gg/33dcJZS2
11:25 < bridge> <heinrich5991> hey! I wanted to visit you one day
11:25 < bridge> <heinrich5991> is it you who keeps sabotaging that? :p
11:30 < bridge> <ryozuki> ```cpp
11:30 < bridge> <ryozuki>   MlirAttribute compile_unit =
11:30 < bridge> <ryozuki>       mlirLLVMDICompileUnitAttrGet(ctx, id, 0x1C, file, foo, false, 0x1);
11:30 < bridge> <ryozuki> 
11:30 < bridge> <ryozuki>   // CHECK: #llvm.di_compile_unit<id = distinct[0]<"foo">, sourceLanguage =
11:30 < bridge> <ryozuki>   // DW_LANG_Rust, file = <"foo" in "bar">, producer = "foo", isOptimized =
11:30 < bridge> <ryozuki>   // false, emissionKind = Full>
11:30 < bridge> <ryozuki>   mlirAttributeDump(compile_unit);
11:30 < bridge> <ryozuki> ```
11:30 < bridge> <ryozuki> obviously the test i add to llvm would mention rust
11:52 < bridge> <cyberFighter> :monkaS:
11:56 < bridge> <Jupstar ✪> when i ever meet cyberFighter I'll ride to him with a horse
11:56 < bridge> <Jupstar ✪> have you tried paying ppl to come to you?
11:58 < bridge> <learath2> That's the lamest thing I've ever heard
11:58 < bridge> <Jupstar ✪> still sounds more sane than getting depressions for being alone
11:59 < bridge> <learath2> Then you'll be depressed for being lame. There is no out
12:00 < bridge> <Jupstar ✪> xDD
12:10 < bridge> <milkeeycat> u can pay them to play starkiller maps with you
12:23 < bridge> <ryozuki> new llvm pr https://github.com/llvm/llvm-project/pull/83992
12:27 < bridge> <Jupstar ✪> the legend strikes again, epyc
12:28 < bridge> <milkeeycat> and it's so much less laggy, holly
12:28 < bridge> <meloƞ> Just invite us all - im down :owo:
12:44 < bridge> <Jupstar ✪> already got texture support?
12:44 < bridge> <milkeeycat> no but it feels so good
12:45 < bridge> <milkeeycat> to not render 300k tiles each time
12:45 < bridge> <Jupstar ✪> xd
12:45 < bridge> <milkeeycat> but it still laggy asf when zoom out :feelsbadman:
12:45 < bridge> <Jupstar ✪> that won't go away with your approach
12:46 < bridge> <milkeeycat> it's ok with 2-3k but higher is joe over
12:46 < bridge> <milkeeycat> how does ddnet render tiles? :owo:
12:46 < bridge> <Jupstar ✪> it buffers all tiles in GPU memory.
12:46 < bridge> <Jupstar ✪> 
12:46 < bridge> <Jupstar ✪> so they are not built on fly
12:47 < bridge> <Jupstar ✪> 
12:47 < bridge> <Jupstar ✪> additionally there is not 1 draw call for every tile, instead for a whole tile row
12:47 < bridge> <milkeeycat> i didnt even know that shaders have macros :justatest:
12:47 < bridge> <Jupstar ✪> yeah GLSL is c like
12:52 < bridge> <Jupstar ✪> @milkeeycat you should try to make a renderer using geometry shaders and then do a mix out of ddnet's approach and patiga's approach.
12:52 < bridge> <Jupstar ✪> 
12:52 < bridge> <Jupstar ✪> i really wonder how well a GPU can spawn vertices vs the CPU creating draw calls of existing vertices (but having to do it row by row)
12:52 < bridge> <Jupstar ✪> so basically patigatus21's twgpu does everything on the GPU
12:53 < bridge> <Jupstar ✪> but he does most stuff on the fragment shader
12:53 < bridge> <milkeeycat> this code is already patiga's twmap + ddnet graphics + my js glue to make it work
12:53 < bridge> <milkeeycat> :justatest:
12:53 < bridge> <Jupstar ✪> xdd
12:55 < bridge> <milkeeycat> oh and libtw2
12:55 < bridge> <heinrich5991> ❤️
12:56 < bridge> <Jupstar ✪> when libtw1 😬
12:56 < bridge> <Jupstar ✪> teeworlds 1.0 release
12:56 < bridge> <Jupstar ✪> 
12:56 < bridge> <Jupstar ✪> chiller's wet dream
13:34 < bridge> <patiga> pls go back to "patiga" 😅
13:37 < bridge> <patiga> btw I revisted my tilemap renderer, and optimized it quite aggressively. iirc, it might be 5x faster than before. the fragment shader now does the discard think, the math is a single matrix multiplication, and the memory of the gpu-tilemap is 8x smaller :owo:
13:38 < bridge> <Jupstar ✪> But it sounds like u a Knight xd
13:38 < bridge> <patiga> heh, thats actually kinda accurate for how I got that name
13:39 < bridge> <patiga> I think I lost multiple gmail addresses when I was younger, so patiga, patiga21 and all that stuff was gone
13:39 < bridge> <patiga> and in german class we just had that you can make names sound old by adding "tus" or smth, so that was my next email address :d
13:39 < bridge> <Jupstar ✪> Oh that's very impressive
13:39 < bridge> <Jupstar ✪> I should try it out
13:40 < bridge> <patiga> btw I revisted my tilemap renderer, and optimized it quite aggressively. iirc, it might be 5x faster than before. the fragment shader now does the discard thing, the math is a single matrix multiplication, and the memory footprint of the gpu-tilemap is 8x smaller :owo:
13:40 < bridge> <Jupstar ✪> But discards are usually not faster, that surprises me
13:40 < bridge> <patiga> on the `new-approach` branch
13:40 < bridge> <patiga> I think the discard did essentially nothing
13:40 < bridge> <Jupstar ✪> Maybe the compiler optimizes the shader based on it
13:40 < bridge> <patiga> but I included it for testing
13:40 < bridge> <Jupstar ✪> Ah
13:40 < bridge> <patiga> I didnt really see a different with or without
13:41 < bridge> <patiga> *difference
13:42 < bridge> <patiga> (and mipmaps are next, although my naive downscaling didn't work straight away)
13:42 < bridge> <Jupstar ✪> How did you measure the increase in performance? Did you find a nice tool?
13:43 < bridge> <patiga> no time for that xd
13:43 < bridge> <patiga> just my good ol fps counter
13:44 < bridge> <Jupstar ✪> 5x fps would be insane
13:44 < bridge> <Jupstar ✪> Nice nice
13:45 < bridge> <patiga> well if you can measure it too I'd be interested in your findings as well, I was also a bit sceptical
13:45 < bridge> <Jupstar ✪> Considering the fragment shader calculates for 3 million pixels it defs has high potential ^^
13:45 < bridge> <patiga> hm I think on my laptop, the improvement wasn't as drastic
15:24 < bridge> <heinrich5991> ChillerDragon: how does `-i''` help with macos?
15:24 < bridge> <heinrich5991> that should be precisely equivalent to `-i`
16:07 < bridge> <heinrich5991> @learath2 have you continued the message queue stuff?
16:09 < bridge> <learath2> I'm still brainstorming about it, the things I want from it make me think maybe something out of the box like qpid would be bettet than me implementing it from scratch
16:09 < bridge> <learath2> Better*
16:11 < bridge> <heinrich5991> I mean kog also has a solution, don't remember what that was
16:11 < bridge> <learath2> They have some grpc stuff
16:12 < bridge> <heinrich5991> hmm. that sounds orthogonal to the stuff the queue should do
16:12 < bridge> <learath2> I think I definitely want node discovery. Manually configuring the shape of the network down to the gameserver doesn't sound very appealing to me
16:13 < bridge> <heinrich5991> node discovery with a central server?
16:13 < bridge> <heinrich5991> or more complicated?
16:14 < bridge> <learath2> With a central server
16:14  * bridge <heinrich5991> aproves
16:14  * bridge <heinrich5991> approves
16:15 < bridge> <learath2> I think I also definitely want proxies. I don't want all gameservers directly talking to the central server
16:16 < bridge> <heinrich5991> yes, I agree
16:16 < bridge> <heinrich5991> both of that seem pretty necessary
16:17 < bridge> <learath2> One thing I'm having trouble with is how to name nodes in a unique way so messages can be addressed to them
16:17 < bridge> <heinrich5991> location+port?
16:18 < bridge> <learath2> I was thinking something similar to that. Dot seperated, something like `main.ger2.8303`
16:19 < bridge> <heinrich5991> sounds good
16:19 < bridge> <learath2> Or `s/main/central/` whatever
16:22 < bridge> <learath2> Another thing that bugs me is how to handle caching and coherency. Say you issue a global ban from `main.ger2.8303` that generates a message to `main.banmaster`. However our link to `main` is down. Now that ban could wait until it's back up, but that's not desirable. It should be in effect locally atleast
16:24 < bridge> <learath2> But if it is active locally, that message still needs to be delivered when the link is back up. Some node needs to be watching and keeping track of the link state and resynchronizing stuff when it's back up
16:24 < bridge> <learath2> Maybe I should just implement the backbone of the messaging infrastructure and think about these details later
16:27 < bridge> <heinrich5991> that seems like the thing you're using isn't aligned with how you want to model it
16:27 < bridge> <heinrich5991> but I guess that also means conflict resolution once you no longer want to require the central server to be there
16:27 < bridge> <learath2> Exactly why I'm hesitant to go ahead. Because maybe there is some protocol level way to achieve this without bolting on external synchronization
16:28 < bridge> <heinrich5991> i.e. maybe we'd need to look more for a p2p system that we coudl use centrally
16:28 < bridge> <learath2> But bans are just one thing that we can use this system with. Should we really design it around them?
16:30 < bridge> <heinrich5991> what else do we want to use it for?
16:31 < bridge> <learath2> I would ideally want to move everything onto that one link slowly one by one. Like broadcasts, executing rcon commands
16:32 < bridge> <heinrich5991> broadcasts would be the same as executing rcon commands, or do they need something special?
16:32 < bridge> <learath2> Yeah they are the same, just thinking quickly
16:32 < bridge> <heinrich5991> what properties do we want for this?
16:32 < bridge> <learath2> The hardware status python thing that we run would be nice to have over this aswell
16:33 < bridge> <ryozuki> this is about ddnet infra internally right?
16:33 < bridge> <heinrich5991> it's executed at most once
16:34 < bridge> <ryozuki> reading this again made me yet again remind me of elixir and that i should learn it more
16:34 < bridge> <heinrich5991> should it be executed later when the link comes online again?
16:34 < bridge> <learath2> Another nice to have would be log/audit streams, so you could do something like `tail-log main.ger10.8306` and get logs. I know we can bolt all of these things on with bash scripts and call it a day but I don't really like it
16:34 < bridge> <heinrich5991> or should it only be executed if the lnk is there when the command is issed?
16:35 < bridge> <learath2> Another thing that might be nice is having records go over this link too, without the weird fallback databases we sync back up every night
16:35 < bridge> <ryozuki> this maybe is off topic, but u can even not define a main server, and let a consensus algorithm define it, something like https://raft.github.io/
16:35 < bridge> <learath2> I was thinking only when connected. So the link has to be up for you to issue a rcon command
16:36 < bridge> <heinrich5991> ok, that's trivial
16:36 < bridge> <heinrich5991> any system will be able to support that
16:36 < bridge> <heinrich5991> basically a broadcast
16:36 < bridge> <learath2> Heck, I could probably bolt on bans better without any messaging infrastructure. Just run a db on every server and rely on the databases own sync mechanism
16:38 < bridge> <heinrich5991> the db thing doesn't look suited tbh
16:38 < bridge> <heinrich5991> the db thing doesn't look well-suited tbh
16:38 < bridge> <heinrich5991> cause the server needs to handle the bans itself, it'd need to subscribe to updates
16:39 < bridge> <learath2> What becomes annoying without a central server is the routing and addressing. Without a defined hierarchy you need to announce stuff around so everyone can get a network map
16:39 < bridge> <ryozuki> erlang has a db thing in OTP
16:39 < bridge> <ryozuki> mnesia
16:39 < bridge> <ryozuki> https://en.wikipedia.org/wiki/Mnesia
16:39 < bridge> <learath2> Easy enough to bolt on a query into the server so it can just query the db about it instead
16:39 < bridge> <heinrich5991> no, too slow
16:40 < bridge> <learath2> Or db update hooks to execute fifo commands, even more jank
16:41 < bridge> <learath2> Anyway, implementing a whole new messaging protocol just to stream some bans over it sounds silly to me
16:42 < bridge> <heinrich5991> syncing rcon access is also interesting
16:42 < bridge> <heinrich5991> also a list that would be synchronized
16:45 < bridge> <learath2> Both can be implemented as db insert/update/delete trigger now that I think about it
16:48 < bridge> <milkeeycat> https://news.ycombinator.com/item?id=39395020 :lol:
17:06 < bridge> <heinrich5991> @learath2 perhaps something like redis would work? https://redis.io/docs/data-types/streams/
17:06 < bridge> <heinrich5991> the replication consistency looks a bit lackluster though: https://redis.io/docs/management/replication/
17:07 < bridge> <learath2> I doubt it would work any better than just postgres
17:07 < bridge> <heinrich5991> it has a stream, which would be useful for executing commands in order
17:08 < bridge> <heinrich5991> does postgres have something like that?
17:08 < bridge> <learath2> And working with redis on this is meh since it's technically a key value store. What would be the key of a ban even? It'll just be a ban id of sort
17:08 < bridge> <learath2> Postgres has streaming replication too. Not sure if there is any order guarantee
17:08 < bridge> <heinrich5991> no, I mean it has "streams"
17:08 < bridge> <heinrich5991> a thing where you can put a message
17:09 < bridge> <heinrich5991> and other people receive it, and they can receive all messages they've missed, too
17:09 < bridge> <heinrich5991> I looked at the set datatype, which seemed appropriate for bans (if you can listen for changes, was trying to figure that otu)
17:10 < bridge> <peter44h> It's not enough, even 18 is too young
17:10 < bridge> <peter44h> You can't use C++20 until you can at least drink
17:11 < bridge> <ryozuki> @learath2 how to make a uint64_t literal
17:11 < bridge> <ryozuki> omg
17:12 < bridge> <ryozuki> i guess i cant take a reference to a literal like in rust
17:12 < bridge> <ryozuki> Cannot take the address of an rvalue of type 'int'clang(typecheck_invalid_lvalue_addrof
17:13 < bridge> <learath2> UINT64_C
17:13 < bridge> <ryozuki> ```c
17:13 < bridge> <ryozuki> uint64_t args[] = { 1 };
17:13 < bridge> <ryozuki>   MlirAttribute expression_elem =
17:13 < bridge> <ryozuki>       mlirLLVMDIExpressionElemAttrGet(ctx, 1, 1, args);
17:13 < bridge> <ryozuki> ```
17:13 < bridge> <ryozuki> i just went the long wae
17:13 < bridge> <learath2> Is this C++ or C?
17:13 < bridge> <ryozuki> pure C baby
17:13 < bridge> <ryozuki> im making the llvm C api for mlir
17:14 < bridge> <ryozuki> https://github.com/llvm/llvm-project/pull/83992
17:14 < bridge> <learath2> Okay you are about to witness my favourite C feature
17:14 < bridge> <heinrich5991> ah, another nice thing would be support for the redirect on map change stuff @learath2
17:14 < bridge> <heinrich5991> https://github.com/ddnet/ddnet/issues/6754
17:14 < bridge> <learath2> `MlirAttribute expression_elem = mlirLLVMDIExpressionElemAttrGet(ctx, 1, 1, (uint64_t){1});`
17:14 < bridge> <ryozuki> well its only test code
17:14 < bridge> <ryozuki> ohh
17:15 < bridge> <heinrich5991> should the cast go into the braces?
17:15 < bridge> <ryozuki> Incompatible integer to pointer conversion passing 'uint64_t' (aka 'unsigned long') to parameter of type 'const uint64_t *' (aka 'const unsigned long *'); take the address with & (fix available)clang(-Wint-conversion)
17:15 < bridge> <learath2> Take the address, it's fine
17:15 < bridge> <ryozuki> ```
17:15 < bridge> <ryozuki> Incompatible integer to pointer conversion passing 'uint64_t' (aka 'unsigned long') to parameter of type 'const uint64_t *' (aka 'const unsigned long *'); take the address with & (fix available)clang(-Wint-conversion)
17:15 < bridge> <ryozuki> 
17:15 < bridge> <ryozuki> ```
17:15 < bridge> <ryozuki> ```
17:15 < bridge> <ryozuki>   MlirAttribute expression_elem =
17:15 < bridge> <ryozuki>       mlirLLVMDIExpressionElemAttrGet(ctx, 1, 1, &(uint64_t){1});
17:15 < bridge> <ryozuki> ```
17:15 < bridge> <ryozuki> this worked
17:15 < bridge> <ryozuki> xd
17:15 < bridge> <learath2> Yep
17:15 < bridge> <learath2> Nope
17:15 < bridge> <ryozuki> whats the feature called
17:16 < bridge> <learath2> Compound Literals, C99:6.5.2.5
17:16 < bridge> <learath2> One day C++ people will stop discussing about it and add it to their language hopefully
17:16 < bridge> <ryozuki> @learath2 look at my pr
17:16 < bridge> <ryozuki> only the tests are pure C tho
17:17 < bridge> <peter44h> Would `1UL` not work?
17:17 < bridge> <ryozuki> i tried
17:17 < bridge> <ryozuki> and u cant take a addr to it
17:17 < bridge> <learath2> Literals are rvalues, they don't have addresses
17:18 < bridge> <ryozuki> they do in rust
17:18 < bridge> <ryozuki> 🤓
17:18 < bridge> <peter44h> Ah I didn't read enough
17:18 < bridge> <peter44h> I just read "literal" when what is required is not really a literal
17:18 < bridge> <learath2> I very much doubt Rust even has the concept of an rvalue, but I would also be surprised if they don't "materialize" the literal somehow before taking the reference
17:20 < bridge> <heinrich5991> they "materialize" it, if you want to call it that way
17:20 < bridge> <ryozuki> it does
17:20 < bridge> <learath2> btw, technically the correct literal for a `uint64_t` is `UINT64_C(1)`
17:20 < bridge> <ryozuki> because i know the rust middle IR
17:20 < bridge> <ryozuki> it mentions rvalue everywhere
17:20 < bridge> <ryozuki> https://doc.rust-lang.org/stable/nightly-rustc/rustc_middle/mir/enum.Rvalue.html
17:21 < bridge> <ryozuki> but well being pedantic, its a rvalue in MIR not in rust itself xd
17:21 < bridge> <learath2> idk any deep rust so can't say much 😛
17:21 < bridge> <ryozuki> i copied this to make edlang
17:21 < bridge> <ryozuki> kek
17:23 < bridge> <peter44h> Wdym, is that a standard thing?
17:23 < bridge> <learath2> Interesting feature of redis that I did not know, if it replicates well maybe it can be used
17:23 < bridge> <peter44h> I know the "UL" is "unsigned long" which on every platform that matters is what uint64_t is going to be typedef'd to
17:23 < bridge> <peter44h> Actually maybe not on windows
17:23 < bridge> <peter44h> So I guess ULL is the safer one
17:24 < bridge> <learath2> Yes, C99:7.18.4.1
17:24 < bridge> <peter44h> I've never seen anyone use this syntax, I'll take a look
17:25 < bridge> <learath2> It's a macro that expands to whatever is the correct literal type on that platform
17:25 < bridge> <learath2> s\/platform/implementation/
17:25 < bridge> <peter44h> In general I'm not a fan of putting the bit width in type names
17:25 < bridge> <learath2> and C++11:18.4.1p1
17:25 < bridge> <heinrich5991> I'm a fan of putting bit width in type names
17:26 < bridge> <heinrich5991> why aren't you?
17:26 < bridge> <peter44h> The amount of times I've done a typo for "23" instead of "32" is enough for me to dislike it
17:26 < bridge> <learath2> Why did you think replication looked lackluster?
17:26 < bridge> <heinrich5991> https://github.com/npm/npm/issues/2933
17:28 < bridge> <heinrich5991> > Synchronous replication of certain data can be requested by the clients using the WAIT command. However WAIT is only able to ensure there are the specified number of acknowledged copies in the other Redis instances, it does not turn a set of Redis instances into a CP system with strong consistency: acknowledged writes can still be lost during a failover, depending on the exact configuration of the Redis persistence. However with WAIT the pro
17:28 < bridge> <heinrich5991> this one
17:28 < bridge> <peter44h> Beyond bit field uses they are also pointless
17:28 < bridge> <heinrich5991> I think it's important to acknowledge which bit width you want
17:29 < bridge> <heinrich5991> especially since you usually don't want to change it across systems
17:29 < bridge> <peter44h> Sure, typedef it to "size" or something and leave it at that
17:29 < bridge> <peter44h> I think forcing it into your APIs is a bad idea though
17:29 < bridge> <heinrich5991> why?
17:31 < bridge> <peter44h> Why do it?
17:31 < bridge> <peter44h> It's a niche thing beyond bit fields
17:31 < bridge> <learath2> Across modern systems, say you need 16bits of storage, but doing math with 32bit integers is faster on one of your target architectures. You don't want to strongarm the compiler into using 16bit ones
17:31 < bridge> <ryozuki> i disagree with u on this
17:31 < bridge> <Jupstar ✪> and that from the llvm dev that works with i6
17:31 < bridge> <Jupstar ✪> :lol:
17:31 < bridge> <ryozuki> i32 to me is way better
17:31 < bridge> <heinrich5991> because you usually internalize these bit width while coding. better have them in the code, too
17:31 < bridge> <ryozuki> also llvm ir does it like i32
17:31 < bridge> <ryozuki> its not a rust invented thing btw
17:31 < bridge> <peter44h> I think the worst idea is having languages with the arbitrary bit widths though
17:32 < bridge> <learath2> `uint_fast_16t i;`
17:32 < bridge> <peter44h> The typos alone can really screw you there and there is no way to diagnose them
17:32 < bridge> <peter44h> But I guess that's a separate debate
17:32 < bridge> <learath2> `uint_fast16t i;`
17:32 < bridge> <learath2> `uint_fast16_t i;`
17:32 < bridge> <heinrich5991> I still think that's a good type
17:32 < bridge> <heinrich5991> I just dislike `short`, `int`, `long`, `long long`
17:32 < bridge> <heinrich5991> and their `unsigned` variants
17:32 < bridge> <ryozuki> why not uint16_fast_t
17:33 < bridge> <learath2> Hm, yeah those aren't amazing types ngl. Mostly because to know their limits you have to refer to the standard
17:33 < bridge> <Jupstar ✪> why not u16_fast
17:33 < bridge> <ryozuki> just u16
17:33 < bridge> <ryozuki> and always fast
17:33 < bridge> <ryozuki> <a:PadoruSpeed:585542368015417574>
17:33 < bridge> <Jupstar ✪> true
17:33 < bridge> <Jupstar ✪> no_u16
17:33 < bridge> <learath2> no, sometimes you actually want 16 bits
17:33 < bridge> <ryozuki> tbh idk what the fast does so
17:34 < bridge> <ryozuki> i was just memeing
17:34 < bridge> <learath2> fast tells the compiler to pick the fastest integer type that can contain the range of interest as in fastest to operate on in the target architecture
17:34 < bridge> <heinrich5991> what does that mean in practice, in say gcc, clang or msvc?
17:35 < bridge> <Jupstar ✪> all except u8 are 64bit or not
17:35 < bridge> <Jupstar ✪> u8fast
17:35 < bridge> <learath2> I doubt it changes anything on x86_64 in 2024
17:35 < bridge> <heinrich5991> I was wondering what their definitions are
17:35 < bridge> <ryozuki> im such a clown
17:35 < bridge> <ryozuki> https://cdn.discordapp.com/attachments/293493549758939136/1214612379795722300/image.png?ex=65f9beed&is=65e749ed&hm=4929a9b19366547af4e92f32d9d4b7552b95836f1c5b75cc4508c4e735188aa5&
17:36 < bridge> <heinrich5991> is `uint_fast16_t` `uint16_t` or `uint64_t`?
17:36 < bridge> <heinrich5991> on x86 etc.
17:37 < bridge> <Jupstar ✪> @peter44h was your problem standard typdefs for uint32_t  etc.?
17:37 < bridge> <Jupstar ✪> or was it about custom types
17:37 < bridge> <heinrich5991> thought it was even about rust's integer types
17:38 < bridge> <peter44h> Actually it makes me question, with a 64 bit CPU, is there a big difference between a 32 vs 64 bits multiply?
17:38 < bridge> <Jupstar ✪> depends on the value
17:38 < bridge> <peter44h> It's not the case of just "this takes n cycles"?
17:39 < bridge> <peter44h> There's no cost table Intel gives or something?
17:39 < bridge> <learath2> I'd guess 64 on x86-64, no reason to use half a register
17:39 < bridge> <peter44h> My expectation is the 32 bits is still faster but idk
17:40 < bridge> <Jupstar ✪> i guess u could find it somewhere
17:40 < bridge> <Jupstar ✪> but in the real world it's not as ez as that
17:40 < bridge> <Jupstar ✪> as the compiler can reorder some stuff etc.
17:41 < bridge> <heinrich5991> https://www.agner.org/optimize/instruction_tables.pdf
17:41 < bridge> <heinrich5991> apparently the larger the bit width, the faster the `mul`
17:41 < bridge> <heinrich5991> for ice lake
17:42 < bridge> <learath2> lol, I'm looking at the exact same table
17:42 < bridge> <heinrich5991> but 32bit and 64bit imul are the same
17:42 < bridge> <learath2> for haswell it seems imul for r32,r32 is the same as r64,r64
17:42 < bridge> <heinrich5991> (register-register-multiplication)
17:43 < bridge> <heinrich5991> but `mul` (register-constant-multiplication) is faster for 64 bit
17:44 < bridge> <heinrich5991> `imul` for both 32 and 64 bit has a latency of 3 cycles and it takes 1 cycle
17:44 < bridge> <learath2> not only the compiler btw, the cpu will also reorder stuff, it's impossible to reason about cycle level performance anymore unless you are one of the few people that seem to be aliens
17:45 < bridge> <heinrich5991> the cpu has less UB to exploit though, so the observable effect (modulo speed, … ahhh, we're talking about speed
17:47 < bridge> <heinrich5991> ah, but `idiv` is faster for 32-bit than for 64-bit
17:47 < bridge> <heinrich5991> takes 6 vs 10 cycles (latency: 12, 15)
17:50 < bridge> <peter44h> Wdym by this?
17:50 < bridge> <peter44h> Surely if it's constant the compiler would just be using shifts
17:51 < bridge> <peter44h> (if it's pot or whatever other optimization rules exist)
17:51 < bridge> <learath2> Ignore the compiler for a second
17:51 < bridge> <heinrich5991> I'm saying the 64-bit `mul` instruction is faster for 64 bit than for 32 bit
17:51 < bridge> <heinrich5991> whether it's emitted by the compiler is a different qusetion
17:51 < bridge> <learath2> This is at the level of the cpu, if you emit a mul with an immediate operand
17:52 < bridge> <heinrich5991> I just verified, `x * 1234` does not compile down to shifts, but to a multiplication
17:54 < bridge> <peter44h> Also I think div is not pipelined but mul can be
17:54 < bridge> <heinrich5991> doesn't seem to be the case
17:54 < bridge> <heinrich5991> in the table, I can see that the latency differs from the speed
17:54 < bridge> <heinrich5991> for both division and multiplication
17:54 < bridge> <heinrich5991> so something must be pipelined according to my understanding
17:55 < bridge> <learath2> @heinrich5991 what does `x / 113` compile to btw?
17:55 < bridge> <heinrich5991> I'd guess a multiplication
17:55 < bridge> <heinrich5991> yes, multiplication plus shifts
17:56 < bridge> <learath2> Yeah, I'd guess it's very hard to get a compiler to emit a division
17:56 < bridge> <peter44h> It's easy
17:56 < bridge> <peter44h> Use -Os
17:56 < bridge> <learath2> Is there even an immediate division in x86?
17:56 < bridge> <ryozuki> https://uops.info/table.html
17:56 < bridge> <peter44h> I think gcc is way too aggressive on this point
17:56 < bridge> <peter44h> It's obviously not worth the bytes it saves
17:57 < bridge> <ryozuki> zen3 :poggers2:
17:57 < bridge> <ryozuki> https://cdn.discordapp.com/attachments/293493549758939136/1214617740711825408/image.png?ex=65f9c3eb&is=65e74eeb&hm=472857ec75936db2949341c9e58ce4a7b0b554839cdb88f0ad4ee710c61f2be6&
17:58 < bridge> <ryozuki> https://cdn.discordapp.com/attachments/293493549758939136/1214617949709541437/image.png?ex=65f9c41d&is=65e74f1d&hm=601ce379479a352a17b6ba1c695b51e702e384a3e5bb316fcdae227c3d0f9087&
17:58 < bridge> <peter44h> That's behavior I'd expect from "-Oz" instead
17:58 < bridge> <peter44h> But I think that option got copied from clang and does basically nothing
18:06 < bridge> <learath2> I managed to not do anything productive today :/
18:07 < bridge> <Jupstar ✪> is your day now over or what
18:11 < bridge> <ryozuki> i did a 500 line pr to llvm
18:11 < bridge> <ryozuki> and worked
18:33 < bridge> <meloƞ> gamer
18:34 < bridge> <meloƞ> i gave a "lgtm" on gitlab today and also worked
18:37 < bridge> <learath2> I don't operate well in the evenings anymore
18:38 < bridge> <Jupstar ✪> then you are simply not motivated for your project
18:39 < bridge> <Jupstar ✪> except it's really only today
18:39 < bridge> <Jupstar ✪> everyone has bad days
18:39 < bridge> <learath2> I'm not motivated to be awake anyway
18:39 < bridge> <learath2> You forget that I'm an old man
18:41 < bridge> <Jupstar ✪> 12 is not the oldest age
18:41 < bridge> <Jupstar ✪> compared to most animals... yes
18:42 < bridge> <learath2> That's only one of my ages, I also identify as an 85 year old
18:42 < bridge> <Jupstar ✪> wise as you are, you know how to trick your brain to also work in the evening
18:43 < bridge> <Jupstar ✪> 😏
19:12 < bridge> <heinrich5991> @learath2 thinking about it. do we even need anything fancy?
19:12 < bridge> <heinrich5991> since none of the solutions fit exactly, a simple http endpoint for current state plus a tcp connection for state updates should work, no?
19:30 < bridge> <learath2> HEH, I was also pondering that. If we want to do just a solution for bans, I think that's the way
19:30 < bridge> <blitzzoo_> hi, I found a server for you with a nice girl with perfect nudes. join now ❤️
19:30 < bridge> <blitzzoo_> https://discord.gg/sexyhot
19:30 < bridge> <blitzzoo_> @everyone
19:30 < bridge> <learath2> huh
19:32 < bridge> <learath2> It would have been nice to have all our infrastructure nicely connected, but shouldn't let perfect be the enemy of works better than the last band aid
19:35 < bridge> <heinrich5991> that'd connect our infrastructure, right?
19:36 < bridge> <learath2> It just expands the patchwork of solutions
19:36 < bridge> <heinrich5991> I guess you could replace the tcp connection with a simple message queue
19:36 < bridge> <learath2> noo, I wouldn't even bother messing around with it just for bans
19:37 < bridge> <learath2> Not much of an advantage, you just get message queueing, our operation is fast enough that we don't even need to think about that
19:37 < bridge> <heinrich5991> we also get a segmentation protocol like that
19:37 < bridge> <heinrich5991> plus subscriptions, I guess
19:39 < bridge> <learath2> Eh, if we want one zeromqs pub/sub sockets aren't bad
19:39 < bridge> <learath2> and easy enough to proxy
19:41 < bridge> <heinrich5991> segmentation! ^^
19:43 < bridge> <learath2> Why would a ban update ever be segmented anyway? Are we thinking of different kinds of segmentation?
19:44 < bridge> <heinrich5991> message segmentation
19:45 < bridge> <learath2> Our messages in this case are tiny though, why would they get segmented?
19:45 < bridge> <heinrich5991> tcp doesn't give us a way do distinguish between messages
19:45 < bridge> <heinrich5991> it only gives us a byte stream
19:45 < bridge> <learath2> Ah, that's what you meant
19:45 < bridge> <heinrich5991> a message queue would give us messages, I'd assum
19:45 < bridge> <heinrich5991> e
19:46 < bridge> <heinrich5991> what kind of segmentation did you think about?
19:46 < bridge> <learath2> I was considering just sending newlines 😄
19:46 < bridge> <heinrich5991> classic
19:46 < bridge> <learath2> I was thinking packet segmentation
19:46 < bridge> <heinrich5991> if it's good enough for IRC…
19:46 < bridge> <heinrich5991> but that'd mean no binary data
19:47 < bridge> <heinrich5991> which kinda sucks
19:47 < bridge> <learath2> Well for bans it's not too bad, but we could do length prefixed messages too I guess
19:49 < bridge> <heinrich5991> parsing text is harder than parsing binary data in C++, I think
19:50 < bridge> <learath2> why?
19:51 < bridge> <learath2> I feel it's about the same
19:51 < bridge> <heinrich5991> you can just read data straight into a POD struct
19:51 < bridge> <heinrich5991> harder to accurately parse text IMO
20:28 < bridge> <meloƞ> is there any way to rotate a rect atm?
20:28 < bridge> <meloƞ> is there any way to rotate a UIrect atm?
20:32 < bridge> <milkeeycat> as a clueless ddnet graphics enjoyer i can say.. no =]
20:35 < bridge> <meloƞ> :brownbear: go implement it for me then pls ty
20:36 < bridge> <milkeeycat> take coords from rect and slap em in DrawRectExt
20:36 < bridge> <meloƞ> that seems way to complicated for me to try
20:36 < bridge> <milkeeycat> wat -.-
20:37 < bridge> <milkeeycat> ure pro
20:37 < bridge> <milkeeycat> ure pro!
20:39 < bridge> <milkeeycat> the more I look at `CMapLayers` the less i believe it's possible to write it :pepeW:
20:39 < bridge> <milkeeycat> the more I look at `CMapLayers` the less i believe it's possible to write :pepeW:
20:40 < bridge> <meloƞ> i just want a precoded function that i can use, idk something user friendly like rotateRect(CUIRect rect, angleInDegrees)
20:40 < bridge> <meloƞ> >:(
20:40 < bridge> <meloƞ> i pay you 1€
20:40 < bridge> <meloƞ> paypal
20:41 < bridge> <milkeeycat> xD
20:41 < bridge> <meloƞ> wait would the angle be a float or a double?
20:42 < bridge> <milkeeycat> in `graphics_threaded` it's float
20:42 < bridge> <meloƞ> ic
20:53 < bridge> <Jupstar ✪> rotate by degrees?
20:53 < bridge> <meloƞ> ye
20:54 < bridge> <Jupstar ✪> or rotate by 180° cf
20:54 < bridge> <Jupstar ✪> or rotate by 180° xd
20:54 < bridge> <meloƞ> pi / 180
20:54 < bridge> <meloƞ> i'd like to rotate a UIRect by a defined amount (45° or smth idk)
20:55 < bridge> <Jupstar ✪> then no
20:55 < bridge> <milkeeycat> did u make helper function?
20:55 < bridge> <meloƞ> i didnt even attempt it yet, im just brainstorming the possibility loudly rn XD
21:10 < bridge> <chairn> @jupeyy_keks any take on this ?
21:10 < bridge> <chairn> https://www.reddit.com/r/pcmasterrace/comments/1b76htj/cmon_eu_do_your_magic_sht/
21:12 < bridge> <heinrich5991> the thing was apparently not news
21:12 < bridge> <heinrich5991> i.e. the present tense of "bans" is incorrect
21:13 < bridge> <heinrich5991> it's been in some EULA, but likely unenforcable anyway because RE for interoperability is legal
21:15 < bridge> <chairn> RE ?
21:17 < bridge> <Jupstar ✪> what should i say xd
21:17 < bridge> <Jupstar ✪> 
21:17 < bridge> <Jupstar ✪> my ideology goes against what nvidia is trying.
21:17 < bridge> <Jupstar ✪> economy is build in a very weird way
21:18 < bridge> <Jupstar ✪> EU probably does that bcs of competition reasons, i'd probably argue a bit different
21:18 < bridge> <heinrich5991> reverse engineering
21:42 < bridge> <Jupstar ✪> ah and also i think they only doing this bcs these are american companies.
21:42 < bridge> <Jupstar ✪> 
21:43 < bridge> <Jupstar ✪> if they'd be european their lobbyism would probably be completely different
21:43 < bridge> <Jupstar ✪> xd
21:43 < bridge> <Jupstar ✪> that's my take
21:57 < bridge> <chairn> usually american companies are pretty good at lobbyism
21:57 < bridge> <peter44h> This is how all US politics function
21:58 < bridge> <peter44h> Close enough anyway
21:59 < bridge> <peter44h> Buckley v Vallejo effectively made bribery legal
21:59 < bridge> <peter44h> Buckley v Valejo effectively made bribery legal
22:02 < bridge> <Jupstar ✪> but i guess the nvidia workplaces for CUDA are all in murica
22:02 < bridge> <Jupstar ✪> so EU doesn't care as much
22:03 < bridge> <Jupstar ✪> better make the market nice for all competitors
22:03 < bridge> <Jupstar ✪> and maybe EU also doesnt like to be too dependent on NVIDIA
22:03 < bridge> <Jupstar ✪> since almost all the industry uses CUDA 😄
22:04 < bridge> <Jupstar ✪> i dunno, is their reasoning public?
22:04 < bridge> <Jupstar ✪> 
22:04 < bridge> <Jupstar ✪> was it even the EU or the european court
22:04 < bridge> <peter44h> There's probably software patent BS involved too
22:04 < bridge> <Jupstar ✪> i dont think that exists in europe
22:04 < bridge> <Jupstar ✪> at least not as in murica
22:05 < bridge> <peter44h> It definitely does
22:05 < bridge> <peter44h> I think it was "fraunhofer" that had patents on mp3
22:05 < bridge> <heinrich5991> source?
22:06 < bridge> <peter44h> That there are software parents in Europe?
22:06 < bridge> <heinrich5991> yes
22:07 < bridge> <peter44h> https://en.wikipedia.org/wiki/Software_patents_under_the_European_Patent_Convention
22:07 < bridge> <peter44h> This page ig
22:07 < bridge> <heinrich5991> > Under the EPC, and in particular its Article 52,[1] "programs for computers" are not regarded as inventions for the purpose of granting European patents,[2] but this exclusion from patentability only applies to the extent to which a European patent application or European patent relates to a computer program as such.[3]
22:07 < bridge> <peter44h> Definitely there were patents on mp3 that got enforced
22:07 < bridge> <heinrich5991> from your article
22:08 < bridge> <peter44h> I don't think that means what you think it does
22:08 < bridge> <heinrich5991> can you give a source for that?
22:09 < bridge> <heinrich5991> what does it mean?
22:09 < bridge> <Jupstar ✪> 🍿
22:09 < bridge> <Jupstar ✪> 😬
22:09 < bridge> <Jupstar ✪> @peter44h you know these stuff are very complicated
22:09 < bridge> <Jupstar ✪> laws are bloatware
22:09 < bridge> <Jupstar ✪> 😬
22:10 < bridge> <peter44h> absolutely, agree
22:11 < bridge> <peter44h> Lawyer is definitely in top 5 most dishonest professions
22:11 < bridge> <Jupstar ✪> BloodWork131: do you agree? xddd
22:11 < bridge> <Jupstar ✪> chillerdragon: do you have that name as ping?
22:12 < bridge> <peter44h> I've heard of lawsuits related to this happening before
22:13 < bridge> <Jupstar ✪> he means a real source
22:13 < bridge> <peter44h> But yeah it does seem the EU is less lenient with what counts as being "patentable" and such
22:14 < bridge> <peter44h> I am searching now and I can't find an EU example
22:15 < bridge> <peter44h> https://en.wikipedia.org/wiki/MP3
22:16 < bridge> <peter44h> 
22:16 < bridge> <peter44h> This page states:
22:16 < bridge> <peter44h> > The basic MP3 decoding and encoding technology is patent-free in the European Union, all patents having expired there by 2012 at the latest. 
22:16 < bridge> <peter44h> If there were no patents there would be nothing to expire, so I guess it can be inferred from this they exist
22:18 < bridge> <peter44h> https://fsfe.org/activities/swpat/swpat.en.html
22:18 < bridge> <peter44h> This page claims:
22:18 < bridge> <peter44h> > The European Patent Convention states that software is not patentable. But laws are always interpreted by courts, and in this case interpretations of the law differ. So the European Patents Office (EPO) grants software patents by declaring them as "computer implemented inventions".
22:18 < bridge> <peter44h> So I have no idea by which technicality something which could be described as a "software patent" could exist
22:18 < bridge> <peter44h> Looks like there is far less of a craze regarding that
22:18 < bridge> <peter44h> Patent trolling as an industry is a more USA thing from what I can tell
22:20 < bridge> <peter44h> I am searching now and I can't find an EU example of an actual lawsuit involving mp3
22:28 < bridge> <mamad_melanin> hi all
22:29 < bridge> <heinrich5991> I find it hard to find information on this, but I'm trying 🙂
22:29 < bridge> <mamad_melanin> do i need to change this returned  unsigned char into signed ?
22:29 < bridge> <mamad_melanin> ```cpp
22:29 < bridge> <mamad_melanin> const unsigned char *sqlite3_column_text(sqlite3_stmt*, int iCol);
22:29 < bridge> <mamad_melanin> ```
22:29 < bridge> <mamad_melanin> do i need to consider anything if im going to use unsigned char ?
22:30 < bridge> <mamad_melanin> and if i wanted is this the right way to do so ?
22:30 < bridge> <mamad_melanin> ```cpp
22:30 < bridge> <mamad_melanin> const char *username = reinterpret_cast<const char*>(sqlite3_column_text(pstmt, 0));
22:30 < bridge> <mamad_melanin> ```