01:10 < ws-client> <ChillerDragon> > str_format is also very slow, str_from_int is much faster if you just need to convert an integer to a string
01:10 < ws-client> <ChillerDragon> @robyt3 woah i did not know that
01:11 < ws-client> <ChillerDragon> Can we add a grep for str_format "%d" to the CI to auto check that?
01:13 < bridge> <heinrich5991> it's kinda sad that we can't simply optimize `str_format`…
01:28 < bridge> <sans._.> How exactly does it go from pink to blue idk, and I've spent 2 hours trying to figure it out
01:28 < bridge> <sans._.> 
01:28 < bridge> <sans._.> I need help ;-;
01:28 < bridge> <sans._.> https://cdn.discordapp.com/attachments/293493549758939136/1237546686436606024/image.png?ex=663c0aaa&is=663ab92a&hm=087a7cc660d2b7227799a400b6ae9ed182962ab938a6b8dce7f3ba50596be5a9&
01:28 < bridge> <sans._.> https://cdn.discordapp.com/attachments/293493549758939136/1237546686654451713/image.png?ex=663c0aaa&is=663ab92a&hm=0f4549885aec86dbf0ab9f3b161f46070ef3991c6de718385f54b7686f5f9168&
01:29 < bridge> <sans._.> This is after converting the color code to hsl, and translating saturation and lightness to a percentage of 255
03:01 < ws-client> <ChillerDragon> @heinrich5991 fyi i will give editor text typing another go now
03:02 < ws-client> <ChillerDragon> trying to finally clean up that old stale issue of mine https://github.com/ddnet/ddnet/issues/3570
07:23 < bridge> <fokkonaut> All that and it still broke hahaha
10:23 < bridge> <robyt3> The hue value is not correct, teeworlds/ddnet color components are in the range 0-255, but the tool you are using uses hue from 0-359. `206 / 255 * 360 = 290`
10:31 < bridge> <cellegenrih> Ah yes, diagonal colors
10:33 < bridge> <cellegenrih> Wait, so the tool just adds the remainer of the number exceeding 255 as the result then?
10:34 < bridge> <cellegenrih> Or what that calculation represents? 206 as the value you want, divided by the max value of hue, multiplied by the tool's max hue value?
10:35 < bridge> <robyt3> yes, it's how to get the correct hue value for the tool
10:35 < bridge> <robyt3> probably this one, given the similarity of the UI: https://hslpicker.com/
10:36 < bridge> <cellegenrih> pog
10:36 < bridge> <cellegenrih> makes sense then :owo:
10:37 < bridge> <cellegenrih> basically percentage convertion between 255 and 360, using the base value.
10:38 < bridge> <cellegenrih> Ah, so what the guy wanted, converting the color code to hsl... shit even I had trouble with it sometimes
10:59 < ws-client> <ChillerDragon> > All that and it still broke hahaha
10:59 < ws-client> <ChillerDragon> @fokkonaut all wat? xd
11:00 < ws-client> <ChillerDragon> There isn't even a 0.7 connecting to the server check in the CI it would have caught it.
11:00 < bridge> <fokkonaut> all that work and effort and political discussions
11:01 < ws-client> <ChillerDragon> you mean mine?
11:01 < bridge> <fokkonaut> in order to be broken by a pr which fixes something
11:01 < bridge> <fokkonaut> yes for 0.7
11:01 < bridge> <fokkonaut> Wait is it even through?
11:01 < bridge> <fokkonaut> or why did u create an issue
11:01 < ws-client> <ChillerDragon> no
11:01 < ws-client> <ChillerDragon> i spend most time client side
11:01 < ws-client> <ChillerDragon> only a bit of lobbying which got lerato and timakro to implement the server side
11:02 < bridge> <fokkonaut> Ah this is about 0.7 client not working
11:02 < ws-client> <ChillerDragon> and the server side is what broke
11:02 < bridge> <fokkonaut> ah
11:02 < bridge> <fokkonaut> misunderstood
11:02 < bridge> <sans._.> I didn't even notice the hue is capped at 255
11:02 < ws-client> <ChillerDragon> imagine my pr getting merged xd
11:02 < bridge> <sans._.> hi chiller
11:02 < bridge> <fokkonaut> Comic Sans MS
11:02 < bridge> <sans._.> fr
11:02 < ws-client> <ChillerDragon> but ye fokko i tried to lobby a bit for the ddnet server bridge to happen in the first place
11:03 < bridge> <fokkonaut> did u ?
11:03 < ws-client> <ChillerDragon> not sure if my lobbying did help to make it happen tho
11:03 < bridge> <sans._.> 2 of the greatest ddnet minds, teach me thine ways
11:03 < bridge> <fokkonaut> xD
11:07 < bridge> <robyt3> This fixed the linking errors :poggers:  now it's time for gradle errors :pepeW:
11:07 < ws-client> <ChillerDragon> android gaming powered by robsti?
11:08 < ws-client> <ChillerDragon> that one dude who has like 90k points on ddnet and does like tripple drag pseudofly without ever having touched a mouse and keyboard in his life will thank you robsti
11:08 < ws-client> <ChillerDragon> he is still stuck in ddnet 6.2.3
11:08 < bridge> <cyberFighter> no one has 90k points
11:09 < ws-client> <ChillerDragon> i may have used the stylistic device hyperbole in that sentence
11:09 < bridge> <fokkonaut> theres someone playing on android?
11:09 < bridge> <fokkonaut> like actually playing ?
11:09 < ws-client> <ChillerDragon> yes 
11:09 < bridge> <fokkonaut> tf
11:09 < ws-client> <ChillerDragon> like never tried desktop
11:09 < bridge> <fokkonaut> impossible
11:09 < ws-client> <ChillerDragon> and having more points than me ._.
11:09 < bridge> <fokkonaut> anytime you walk it automatically jumps
11:09 < bridge> <cellegenrih> lmao 💀
11:09 < ws-client> <ChillerDragon> and he mentioned once that he would really appreciate an client update
11:09 < ws-client> <ChillerDragon> so its really epic that robsti works on it
11:10 < ws-client> <ChillerDragon> robsti and jopsti androiders
11:10 < bridge> <cellegenrih> chinese players found pc to be too easy :KEKW:
11:10 < bridge> <cyberFighter> hes doing it without the left joystick so it doesnt count in my eyes
11:10 < bridge> <cyberFighter> not real mobile ddnet
11:10 < ws-client> <ChillerDragon> what does he use instead?
11:11 < bridge> <cellegenrih> btw maybe cuz chinese people have more time on phones, thats the reason they want android version
11:11 < bridge> <cyberFighter> he uses 3 buttons
11:12 < bridge> <cellegenrih> only 3??
11:12 < bridge> <cellegenrih> jump, hook, hammer?
11:12 < ws-client> <ChillerDragon> yea imagine finishing a ddnet map without 20 emote binds
11:13 < bridge> <cyberFighter> no he uses 3 buttons for movement
11:13 < bridge> <cyberFighter> the fuck are you smoking
11:13 < bridge> <cyberFighter> left jump and right
11:13 < bridge> <cyberFighter> instead of the joystick
11:14 < bridge> <fokkonaut> warning
11:14 < bridge> <fokkonaut> inappropiate use of language
11:15 < bridge> <cyberFighter> ok mr fork on these nuts
11:15 < ws-client> <ChillerDragon> forkonaut xd
11:16 < bridge> <fokkonaut> cyberForker 🍴
11:16 < bridge> <cyberFighter> btw i played ddnet on a car multimedia
11:16 < bridge> <cyberFighter> it sucked xd
11:17 < bridge> <fokkonaut> There's no vacuum cleaner emoji
11:17 < bridge> <fokkonaut> 🪠
11:18 < bridge> <Ewan> whats up fokkonaut
11:18 < bridge> <fokkonaut> do you know updog?
11:18 < bridge> <Ewan> wtf is updog???
11:18 < bridge> <cyberFighter> the sky
11:18 < bridge> <cyberFighter> duhh
11:18 < bridge> <fokkonaut> Not much, just chilling
11:19 < bridge> <Ewan> sweet
11:19 < bridge> <Ewan> eating pretzels and watching tv on my way to sleep
11:19 < bridge> <fokkonaut> Nice
11:19 < bridge> <fokkonaut> what time is it
11:19 < bridge> <Ewan> 3:19
11:19 < bridge> <fokkonaut> am ?
11:19 < bridge> <Ewan> yep
11:19 < bridge> <fokkonaut> bruh
11:19 < bridge> <fokkonaut> its 1120 here
11:20 < bridge> <fokkonaut> am
11:20 < bridge> <Ewan> yep
11:20 < bridge> <Ewan> maybe tim will respond soon
11:20 < bridge> <fokkonaut> Who?
11:20 < bridge> <Ewan> my boy
11:20 < bridge> <Ewan> need help with unique db
11:20 < bridge> <fokkonaut> a
11:21 < bridge> <Ewan> usa is disconnected from the race ranks db
11:21 < bridge> <Ewan> i can technically take auth from can servers but idk if he would want that
11:21 < bridge> <Ewan> so i wait for response
11:21 < bridge> <ryozuki> i got 2kg of whey protien
11:21 < bridge> <ryozuki> :justatest:
11:21 < bridge> <Ewan> nice
11:21 < bridge> <ryozuki> buying in bulk is always more cheap right
11:21 < bridge> <Ewan> usually yea
11:22 < bridge> <Ewan> depends on where tho
11:22 < bridge> <fokkonaut> about to smoke a legal joint in Germany👍
11:22 < bridge> <Ewan> hell yeah
11:22 < bridge> <ryozuki> xd
11:22 < bridge> <ryozuki> i used to casually smoke on weekends with friends
11:22 < bridge> <Ewan> 11:20 AM  wednesday weed hits the best
11:22 < bridge> <ryozuki> but i dont rly like it i think
11:22 < bridge> <cyberFighter> thats good
11:22 < bridge> <ryozuki> it makes me feel like im losing my time
11:22 < bridge> <Ewan> yeah
11:22 < bridge> <ryozuki> cuz u literally become dumb
11:22 < bridge> <Ewan> well
11:22 < bridge> <ryozuki> cant be productive
11:22 < bridge> <fokkonaut> yes, if u overdo it
11:22 < bridge> <Ewan> i think u become dumb the next day
11:22 < bridge> <fokkonaut> + dont take so much
11:22 < bridge> <Ewan> usually im just happy but usually similarly sharp
11:23 < bridge> <Ewan> when just idly high
11:23 < bridge> <fokkonaut> dont do it everyday then its fine, like only at a weekend and just if u have time and _want_ to do it
11:23 < bridge> <fokkonaut> for relaxation etc
11:23 < bridge> <Ewan> yeah
11:24 < ws-client> <ChillerDragon> any windows cmake enjoyer wants to test if my project builds on your machine? :D
11:24 < bridge> <Ewan> sure
11:24 < bridge> <Ewan> send
11:24 < ws-client> <ChillerDragon> https://github.com/ChillerDragon/Craft
11:24 < ws-client> <ChillerDragon> ty babe
11:24 < ws-client> <ChillerDragon> make sure to clone recursive it builds dependencies from source
11:25 < ws-client> <ChillerDragon> no blobs involved
11:25 < ws-client> <ChillerDragon> whats your favorite build workflow ewan? how do you build ddnet?
11:25 < bridge> <Ewan> in ide i press play button
11:25 < ws-client> <ChillerDragon> yes i am hoping that works here as well
11:25 < bridge> <Ewan> on cmdline i do cmake -B build && cmake --build build
11:27 < bridge> <Ewan> i think u should change your submodules to work over https so that i dont need proper auth
11:27 < bridge> <Ewan> over ssh
11:28 < ws-client> <ChillerDragon> is it different than ddnet?
11:29 < ws-client> <ChillerDragon> https://github.com/ddnet/ddnet/blob/f214c1101b67ce231edabb80416c3dc021716f68/.gitmodules#L3
11:29 < ws-client> <ChillerDragon> oh yes
11:29 < ws-client> <ChillerDragon> ok sure i can do that
11:32 < bridge> <Ewan> iadded ssh key anyway
11:32 < bridge> <Ewan> idk how i didnt have one already
11:32 < bridge> <fokkonaut> Just had a call with my instructor, I might be able to go to france in September
11:32 < bridge> <Ewan> nice
11:33 < bridge> <Ewan> oh no
11:33 < bridge> <fokkonaut> wat
11:33 < bridge> <Ewan> build fail
11:33 < bridge> <fokkonaut> mine worked 🚬
11:34 < bridge> <Ewan> chiller u building glfw from source
11:34 < bridge> <Ewan> crazy
11:34 < bridge> <fokkonaut> joke
11:34 < ws-client> <ChillerDragon> ye ikr xd
11:34 < bridge> <Ewan> it wants real msvc windres :feelsbadman:
11:34 < ws-client> <ChillerDragon> wat
11:34 < bridge> <Ewan> gnu windres doesn't have arg it wants
11:34 < ws-client> <ChillerDragon> send erro
11:34 < bridge> <milkeeycat> hey, is it possible to declare variables like dis in nasm? or they all have to be in .data section 😦
11:34 < bridge> <milkeeycat> https://cdn.discordapp.com/attachments/293493549758939136/1237699250553622558/image.png?ex=663c98c0&is=663b4740&hm=ba612168adb0be5f99757e63cf791ad1cabf595a0544307197cd5fc6cebfdf77&
11:35 < ws-client> <ChillerDragon> fokko wat instructor
11:35 < bridge> <fokkonaut> apprenticeship
11:35 < ws-client> <ChillerDragon> cool
11:35 < bridge> <Ewan> very cool
11:35 < bridge> <fokkonaut> right
11:35 < ws-client> <ChillerDragon> how long?
11:35 < bridge> <fokkonaut> 2 weeks
11:35 < bridge> <Ewan> im applying for apprenticeship program rn
11:35 < bridge> <Ewan> tests are intensive
11:35 < ws-client> <ChillerDragon> ewan wats erro
11:35 < bridge> <fokkonaut> :D
11:35 < bridge> <Ewan> it's cooking
11:36 < ws-client> <ChillerDragon> i thought it failed
11:36 < bridge> <Ewan> yeah
11:36 < bridge> <Ewan> hold on
11:36 < bridge> <fokkonaut> 🧑‍🍳
11:36 < bridge> <Ewan> have to parse command line gobbletygook without highlighting
11:36 < bridge> <Ewan> for whatever reason
11:37 < ws-client> <ChillerDragon> wat xd
11:37 < bridge> <Ewan> https://cdn.discordapp.com/attachments/293493549758939136/1237699872753451039/message.txt?ex=663c9955&is=663b47d5&hm=d2c8fd8b96f899935480ee5cc791fb26f38af6a53e5697768c8e5fa88a33ef8b&
11:37 < bridge> <Ewan> this is all output
11:37 < bridge> <Ewan> there are a few things here
11:37 < ws-client> <ChillerDragon> This content is no longer available.
11:37 < bridge> <Ewan> haha
11:38 < bridge> <Ewan> hold on
11:38 < ws-client> <ChillerDragon> irc bullying
11:38 < bridge> <Ewan> http://john2143.com/f/Kt6A
11:38 < ws-client> <ChillerDragon> john com xd
11:38 < bridge> <Ewan> my boy john
11:45 < ws-client> <ChillerDragon> and i made it work with local in the github actions windows CI
11:45 < ws-client> <ChillerDragon> you seem to be failing on glfw not on curl
11:45 < bridge> <Ewan> im on vs now
11:46 < bridge> <Ewan> it succeeded on msys because it was presumably finding system curl
11:46 < ws-client> <ChillerDragon> can you run the craft.exe ?
11:46 < bridge> <Ewan> nope
11:47 < bridge> <Ewan> build fails on both
11:47 < bridge> <Ewan> but at least msys finds curl xd
11:49 < ws-client> <ChillerDragon> xd
11:49 < ws-client> <ChillerDragon> sadge
11:49 < bridge> <Ewan> so i wonder why u opt to build curl first and then link to it
11:49 < ws-client> <ChillerDragon> i was expecting that getting it to pass in the github CI was only the first step but did not expect it to be that bad
11:49 < bridge> <Ewan> instead of just building it in source
11:49 < ws-client> <ChillerDragon> wdym building it in source?
11:49 < bridge> <Ewan> include the headers from the source
11:49 < ws-client> <ChillerDragon> i need to compile the lib
11:49 < ws-client> <ChillerDragon> its not a header only thing
11:49 < bridge> <Ewan> right
11:49 < bridge> <Ewan> so include curl sources
11:49 < bridge> <Ewan> in cmake
11:49 < bridge> <Ewan> lolol not in ur cpp
11:49 < ws-client> <ChillerDragon> it has its own cmakelists.txt
11:49 < ws-client> <ChillerDragon> i dont wanna replicate their whole build settings
11:49 < ws-client> <ChillerDragon> just call it
11:49 < ws-client> <ChillerDragon> i feel like thats how it is supposed to be done
11:49 < ws-client> <ChillerDragon> anyways im getting late for beach bye
11:49 < ws-client> <ChillerDragon> thanks a lot again
11:52 < bridge> <meloƞ> Now i wanna try aswell, Break me free of my chains, let me Go Home :feelsbadman:
11:53 < bridge> <ryozuki> try what
11:55 < bridge> <meloƞ> Build chillers stuff on windows
11:58 < bridge> <Ewan> http://2143.christmas/f/P53k.png
11:58 < bridge> <Ewan> chiller rly
12:11 < bridge> <Ewan> it's building libcurl.dll
12:11 < bridge> <Ewan> and iirc curl can't be built statically through cmake
12:12 < bridge> <Ewan> ```cmake
12:12 < bridge> <Ewan> set(CMAKE_C_STANDARD 99)
12:12 < bridge> <Ewan> set(gcc_like_cxx "$<COMPILE_LANG_AND_ID:C,ARMClang,AppleClang,Clang,GNU>")
12:12 < bridge> <Ewan> set(msvc_cxx "$<COMPILE_LANG_AND_ID:C,MSVC>")
12:12 < bridge> <Ewan> add_compile_definitions(
12:12 < bridge> <Ewan>   "$<${gcc_like_cxx}:$<BUILD_INTERFACE:-O3>>"
12:12 < bridge> <Ewan>   "$<${msvc_cxx}:$<BUILD_INTERFACE:/02>>"
12:12 < bridge> <Ewan> )
12:12 < bridge> <Ewan> ```
12:12 < bridge> <Ewan> but here's the portable version of this
12:12 < bridge> <Ewan> and if ur building it yourself i would just advise foregoing find_package
12:12 < bridge> <Ewan> just link_libraries
12:12 < bridge> <Ewan> since you're already including the headers
12:13 < bridge> <Ewan> im going to bed
12:13 < bridge> <Ewan> maybe melon can crack it
12:15 < bridge> <meloƞ> I'll try lol
12:15 < bridge> <meloƞ> Gn!
13:02 < bridge> <meloƞ> ```cmake
13:02 < bridge> <meloƞ> [build] cl : befehlszeile  warning D9002: Unbekannte Option "-std=c99" wird ignoriert. [C:\Users\Anwender\Projects\Craft\build\craft.vcxproj]
13:02 < bridge> <meloƞ> [build] cl : befehlszeile  warning D9002: Unbekannte Option "-O3" wird ignoriert. [C:\Users\Anwender\Projects\Craft\build\craft.vcxproj]
13:02 < bridge> <meloƞ> ``` :kek:
13:02 < bridge> <meloƞ> 
13:02 < bridge> <meloƞ> 
13:02 < bridge> <meloƞ> also yeah just use precompiled binaries for libcurl i guess.
13:02 < bridge> <meloƞ> 
13:02 < bridge> <meloƞ> anway CHILLER
13:02 < bridge> <meloƞ> 
13:02 < bridge> <meloƞ> it works.
13:02 < bridge> <meloƞ> https://cdn.discordapp.com/attachments/293493549758939136/1237721244074512487/image.png?ex=663cad3c&is=663b5bbc&hm=8f16c87c5125216dd078b96787963b59d2ceebd51c34fa680de97a48930b82dd&
13:02 < bridge> <meloƞ> https://cdn.discordapp.com/attachments/293493549758939136/1237721252773236736/image.png?ex=663cad3e&is=663b5bbe&hm=5011c7595154496b225d9c73c2ee1fcb2e3c5bb2b6372ded1311eef60da7094c&
13:02 < bridge> <meloƞ> ```cmake
13:02 < bridge> <meloƞ> [build] cl : befehlszeile  warning D9002: Unbekannte Option "-std=c99" wird ignoriert. [C:\Users\Anwender\Projects\Craft\build\craft.vcxproj]
13:02 < bridge> <meloƞ> [build] cl : befehlszeile  warning D9002: Unbekannte Option "-O3" wird ignoriert. [C:\Users\Anwender\Projects\Craft\build\craft.vcxproj]
13:02 < bridge> <meloƞ> ``` :kek:
13:02 < bridge> <meloƞ> 
13:02 < bridge> <meloƞ> 
13:02 < bridge> <meloƞ> also yeah just use precompiled binaries for libcurl i guess.
13:03 < bridge> <meloƞ> 
13:03 < bridge> <meloƞ> anway CHILLER
13:03 < bridge> <meloƞ> 
13:03 < bridge> <meloƞ> it works.
13:03 < bridge> <meloƞ> good game
13:03 < bridge> <meloƞ> https://cdn.discordapp.com/attachments/293493549758939136/1237721422038831154/image.png?ex=663cad66&is=663b5be6&hm=1558db5d0dd7dca0d4d0218fc4c9a2d8cbeedb69ab20ed51c133a9c68ebb7725&
13:03 < bridge> <meloƞ> good game 11/10
13:03 < bridge> <meloƞ> https://cdn.discordapp.com/attachments/293493549758939136/1237721422038831154/image.png?ex=663cad66&is=663b5be6&hm=1558db5d0dd7dca0d4d0218fc4c9a2d8cbeedb69ab20ed51c133a9c68ebb7725&
13:11 < bridge> <meloƞ> Msvc really is a Bully for Splitting Up every single build target Up into 20 subdirectories thus Not correctly linking stuff
13:55 < bridge> <ryozuki> question: if u want to go from u8 to i32, do you sign extend or zero extend
13:55 < bridge> <ryozuki> it should bezero i think right?
13:55 < bridge> <heinrich5991> if I go from 255_u8 to i32, I expect 255_i32
13:56 < bridge> <heinrich5991> so zero-extend
13:56 < bridge> <ryozuki> ok i had a bug xd
13:57 < bridge> <egyt> Minetest is better
14:08 < bridge> <meloƞ> https://cdn.discordapp.com/attachments/293493549758939136/1237737829690245250/image-2.png?ex=663cbcae&is=663b6b2e&hm=0d4cb6f12efa9e562b3c6779a934343ea285412238d53d91c883540c5bf4c2e0&
14:15 < bridge> <patiga> Isn't zero-extend the same as sign-extend on u8, as the only sign is positive, where you would zero-extend?
14:17 < bridge> <patiga> in general I'd always expect sign-extend I think
14:18 < bridge> <heinrich5991> the CPU doesn't care about u8 vs i8, generally. that means "sign-extend" will extend the highest bit from the u8, effectively treating it like an i8
14:19 < bridge> <patiga> ah, I thought we were talking in the context of a compiler
14:20 < bridge> <heinrich5991> I guessed it was about emitting instructions to the CPU, in a compiler
14:25 < bridge> <patiga> hm yea these terms seem to map directly to cpu instructions, so I'll go with your interpretation ^^
14:29 < bridge> <ryozuki> yeah
14:29 < bridge> <ryozuki> well it was about emitting llvm ir intrustions but yes
14:30 < bridge> <ryozuki> llvm also doesnt care about u8 or i9
14:30 < bridge> <ryozuki> i8
15:05 < bridge> <robyt3> Progress. Map rendering still broken though
15:05 < bridge> <robyt3> https://cdn.discordapp.com/attachments/293493549758939136/1237752285271162992/image.png?ex=663cca25&is=663b78a5&hm=22c68f7e6a1e725a7e1e7b8bbefc94eb0fcf6bec2bf3d79a45fc34e1f21dbfc9&
15:09 < bridge> <heinrich5991> amazing, wow 😮
15:09 < bridge> <heinrich5991> that seems like half the way already
15:17 < bridge> <ryozuki> ```
15:17 < bridge> <ryozuki> 
15:17 < bridge> <ryozuki> How is GEP different from ptrtoint, arithmetic, and inttoptr?
15:17 < bridge> <ryozuki> 
15:17 < bridge> <ryozuki> It’s very similar; there are only subtle differences.
15:17 < bridge> <ryozuki> 
15:17 < bridge> <ryozuki> With ptrtoint, you have to pick an integer type. One approach is to pick i64; this is safe on everything LLVM supports (LLVM internally assumes pointers are never wider than 64 bits in many places), and the optimizer will actually narrow the i64 arithmetic down to the actual pointer size on targets which don’t support 64-bit arithmetic in most cases. However, there are some cases where it doesn’t do this. With GEP you can avoid this problem.
15:17 < bridge> <ryozuki> 
15:17 < bridge> <ryozuki> Also, GEP carries additional pointer aliasing rules. It’s invalid to take a GEP from one object, address into a different separately allocated object, and dereference it. IR producers (front-ends) must follow this rule, and consumers (optimizers, specifically alias analysis) benefit from being able to rely on it. See the Rules section for more information.
15:17 < bridge> <ryozuki> 
15:17 < bridge> <ryozuki> And, GEP is more concise in common cases.
15:17 < bridge> <ryozuki> 
15:17 < bridge> <ryozuki> However, for the underlying integer computation implied, there is no difference.
15:17 < bridge> <ryozuki> 
15:17 < bridge> <ryozuki> ```
15:17 < bridge> <ryozuki> interesting
15:58 < bridge> <heinrich5991> AFAIK that's the reason for the unsafety of the `offset` method on raw pointers: https://doc.rust-lang.org/std/primitive.pointer.html#method.offset-1
17:16 < bridge> <ryozuki> https://docs.rs/regex-automata/latest/regex_automata/dfa/trait.Automaton.html
17:16 < bridge> <ryozuki> this is rly some piece of work
17:17 < bridge> <learath2> what are you making?
17:18 < bridge> <ryozuki> nothing just found this from reddit
17:18 < bridge> <ryozuki> i just admired the amount of work this probs took
17:18 < bridge> <ryozuki> look the docs detail
17:18 < bridge> <ryozuki> xd
17:18 < bridge> <learath2> I looked at this while trying to figure out whether I can generate an aho corasick automata at compile time
17:19 < bridge> <heinrich5991> permalink (press 'y' anywhere on docs.rs): https://docs.rs/regex-automata/0.4.6/regex_automata/dfa/trait.Automaton.html
17:19 < bridge> <heinrich5991> maybe I should write a bot 🤔
17:20 < bridge> <ryozuki> https://faultlore.com/blah/abi-puns/
17:21 < bridge> <ryozuki> ```
17:21 < bridge> <ryozuki> To recap, abi-cafe was created in a feverish mania as I became increasingly obsessed with the disaster that was C’s technically-nonstandard-but-also-literally-used-in-standards __int128. Like sure there’s the whole intmax_t debacle, but that’s understandable. The truly unforgivable thing was that clang and gcc couldn’t reliably pass __int128 between each other. This of course meant that if Rust’s i128 wished to interoperate with them… i
17:21 < bridge> <ryozuki> 
17:21 < bridge> <ryozuki> Actually I’m telling this story a bit backward. Initially we definitely knew that Rust and Clang disagreed in at least one case on at least platform. Discussion of this issue was stalled out because people weren’t sure of the scope and nature of the problem. Which cases were broken? Who was right or wrong and needed to be fixed? What needed to be fixed? Did __int128 FFI work right on any platforms?
17:21 < bridge> <ryozuki> 
17:21 < bridge> <ryozuki> So I made abi-cafe, which let me describe a function signature in the abstract and generate code for the caller and callee, in Rust or C, and compile and link them together. Notably the generated code had both sides “report” the values they saw with callbacks, so the test harness could just run them and compare what each side reported.
17:21 < bridge> <ryozuki> ```
17:22 < bridge> <learath2> https://docs.rs/regex-automata/0.4.6/regex_automata/dfa/dense/struct.DFA.html#method.from_bytes you can serialize a dfa and save it
17:22 < bridge> <chillerdragon> lmao I watched that video
17:23 < bridge> <ryozuki> epyc
17:24 < bridge> <chillerdragon> Ah nice it just worked for you? Or did you have to change something? How did you build it?
17:24 < bridge> <ryozuki> @learath2 burntsushi is a mastermind and its epic to have him in rust tbh
17:24 < bridge> <ryozuki> https://crates.io/users/BurntSushi
17:25 < bridge> <ryozuki> totally not famous crates
17:25 < bridge> <ryozuki> https://cdn.discordapp.com/attachments/293493549758939136/1237787432376864819/image.png?ex=663ceae0&is=663b9960&hm=965887d777517851a1a015f61df89c2cd84f1756be29e72e003d1ebbe4d8c3d5&
17:25 < bridge> <ryozuki> :justatest:
17:26 < bridge> <ryozuki> https://www.sophiajt.com/search-for-easier-safe-systems-programming/
17:28 < bridge> <Jupstar ✪> epyc, how did u fix the link errors?
17:28 < bridge> <Jupstar ✪> 
17:29 < bridge> <Jupstar ✪> what backend does it use? 
17:29 < bridge> <Jupstar ✪> 
17:29 < bridge> <Jupstar ✪> i assume that's an emulator?
17:29 < bridge> <chillerdragon> Yea it seems to be the most maintained block game. But their visual design is so bad it‘s unplayable for me.
17:30 < bridge> <heinrich5991> what part? can it be modded?
17:31 < bridge> <chillerdragon> All of it is soo ugly. And sadly the project is too big to simply mod everything
17:31 < bridge> <robyt3> 1. Using https://github.com/bbqsrc/cargo-ndk, but I can't figure out how to configure it's output directory correctly so you need to copy the libraries manually and build twice
17:31 < bridge> <robyt3> 2. I tried all backends, Vulkan doesn't work, it reverts back
17:31 < bridge> <robyt3> 3. Yeah, it's an emulator. Once you go ingame you are stuck since you can't open the menu because escape doesn't work on the emulator for some reason
17:31 < bridge> <robyt3> 1. Using https://github.com/bbqsrc/cargo-ndk, but I can't figure out how to configure its output directory correctly so you need to copy the libraries manually and build twice
17:31 < bridge> <robyt3> 2. I tried all backends, Vulkan doesn't work, it reverts back
17:31 < bridge> <robyt3> 3. Yeah, it's an emulator. Once you go ingame you are stuck since you can't open the menu because escape doesn't work on the emulator for some reason
17:31 < bridge> <heinrich5991> https://www.minetest.net/#gallery doesn't look so bad
17:32 < bridge> <chillerdragon> There have been mods that tried to mimic the exact look of Minecraft. But even for those you have to go through the ugly launcher. And the mods them self look still way worse than Craft
17:32 < bridge> <Jupstar ✪> send me apk, i can test on my phone 😄
17:32 < bridge> <Jupstar ✪> i want vulkan
17:32 < bridge> <Jupstar ✪> last i tested it still worked
17:32 < bridge> <robyt3> Is it x86? I only have that apk ready
17:32 < bridge> <meloƞ> i built it just using `cmake build -G "Visual Studio 17 2022" && cmake --build .`
17:32 < bridge> <meloƞ> 
17:32 < bridge> <meloƞ> apart from having to move the glfw and libcurl headers aswell as the .exe manually, it kind of worked straight out of the box, having to move those files is a msvc issue (or im just to bad to set it up correctly)
17:32 < bridge> <chillerdragon> Every single screenshots makes me wanna puke
17:32 < bridge> <Jupstar ✪> i think the emulators have no vk support
17:32 < bridge> <Jupstar ✪> oh no it's arm
17:32 < bridge> <ryozuki> do you know the OG of minecraft is no one else but Zachtronics
17:32 < bridge> <Jupstar ✪> arm64
17:32 < bridge> <chillerdragon> It’s just not my style
17:32 < bridge> <ryozuki> maker of TIS-100 and exapunks
17:32 < bridge> <ryozuki> > Zachtronics was founded by Zach Barth in 2000, who serves as its lead designer. Some of their games include SpaceChem, Infinifactory, TIS-100, and Shenzhen I/O. Infiniminer (2009) inspired the creation of Minecraft.
17:33 < bridge> <ryozuki> https://www.zachtronics.com/infiniminer/
17:33 < bridge> <ryozuki> > Infiniminer is a first-person competitive mining game that takes place in a procedurally generated block world allowing players to mine, build, and explore. Sound familiar? That’s because Infiniminer is the game that started the “block genre” that everyone knows and loves! It’s also quite buggy and old, and is listed here purely for historical reasons.
17:33 < bridge> <ryozuki> xd
17:35 < bridge> <chillerdragon> But yes if minetest would have a design I like and physics that feel good it would be amazing. But bending everything to my needs seems to be more effort than starting from Craft which seems to have enough features to function as a good game engine.
17:37 < bridge> <learath2> Talking about aho-corasick again reminded me that I had one more string search algorithm I wanted to share with youse. Has anyone heard of Rabin-Karp algorithm?
17:39 < bridge> <learath2> It basically uses a rolling windowed hash over the haystack, quite a clever idea
17:39 < bridge> <robyt3> Untested, it's signed with the debug key
17:39 < bridge> <robyt3> https://cdn.discordapp.com/attachments/293493549758939136/1237790931411931196/DDNet-arm64.apk?ex=663cee23&is=663b9ca3&hm=97dd963293603c882ef4bfb2942a76bd114a9052b0736d7297806a9019ad4c42&
17:40 < bridge> <robyt3> Supposing you can connect keyboard and mouse it should be playable if Vulkan works
17:40 < bridge> <Jupstar ✪> i use scrcpy
17:40 < bridge> <chillerdragon> What did you have to move from where to where? To build it? Or to run it?
17:42 < bridge> <Jupstar ✪> ```
17:42 < bridge> <Jupstar ✪> 
17:42 < bridge> <Jupstar ✪> #03 pc 0000000000b4f4e0  /data/app/~~CB3gJmYnDypE7Tt0GeY6pg==/tw.ddnet-ZaQHILJQjp03xFlIOoioFQ==/lib/arm64/libDDNet.so (io_write(void*, void const*, unsigned int)+148) (BuildId: 87efde8ff66ec7895f494b893a16f33dd0813378)
17:42 < bridge> <Jupstar ✪> 05-08 17:41:27.458 22528 22528 F DEBUG   :       #04 pc 0000000000d9a590  /data/app/~~CB3gJmYnDypE7Tt0GeY6pg==/tw.ddnet-ZaQHILJQjp03xFlIOoioFQ==/lib/arm64/libDDNet.so (InitAndroid+448) (BuildId: 87efde8ff66ec7895f494b893a16f33dd0813378)
17:42 < bridge> <Jupstar ✪> 05-08 17:41:27.458 22528 22528 F DEBUG   :       #05 pc 0000000000cac3c8  /data/app/~~CB3gJmYnDypE7Tt0GeY6pg==/tw.ddnet-ZaQHILJQjp03xFlIOoioFQ==/lib/arm64/libDDNet.so (SDL_main+204) (BuildId: 87efde8ff66ec7895f494b893a16f33dd0813378)
17:42 < bridge> <Jupstar ✪> 05-08 17:41:27.458 22528 22528 F DEBUG   :       #06 pc 000000000122d314  /data/app/~~CB3gJmYnDypE7Tt0GeY6pg==/tw.ddnet-ZaQHILJQjp03xFlIOoioFQ==/lib/arm64/libDDNet.so (Java_org_libsdl_app_SDLActivity_nativeRunMain+712) (BuildId: 87efde8ff66ec7895f494b893a16f33dd0813378)
17:42 < bridge> <Jupstar ✪> 
17:43 < bridge> <Jupstar ✪> ```
17:43 < bridge> <Jupstar ✪> 
17:43 < bridge> <Jupstar ✪> crashes in InitAndroid io_write
17:45 < bridge> <Jupstar ✪> https://cdn.discordapp.com/attachments/293493549758939136/1237792444448247891/image.png?ex=663cef8b&is=663b9e0b&hm=32d96a711ec31cad2f017d8c424526d2536fea17c22ce97160dec7c8647e3f0e&
17:45 < bridge> <Jupstar ✪> i still have an apk of that version
17:45 < bridge> <Jupstar ✪> that works
17:45 < bridge> <teero777> pog i can render spritesheets xd
17:45 < bridge> <teero777> https://cdn.discordapp.com/attachments/293493549758939136/1237792523846684823/image.png?ex=663cef9e&is=663b9e1e&hm=704f0f0254c435e657aea650e8979b61e9f993f2bfc5a24c9a3579cd7291298a&
17:45 < bridge> <robyt3> Try giving it all permissions before launching. Maybe it can't ask for storage permissions.
17:46 < bridge> <robyt3> In the app settings
17:46 < bridge> <Jupstar ✪> opengl works in that old version btw
17:46 < bridge> <Jupstar ✪> https://cdn.discordapp.com/attachments/293493549758939136/1237792751479820369/image.png?ex=663cefd5&is=663b9e55&hm=8c5ee54258092cbf27c85c39596348271b351173271775e70ec8759d6844e498&
17:46 < bridge> <Jupstar ✪> screenshot here is with opengl
17:47 < bridge> <Jupstar ✪> lol now it works
17:48 < bridge> <Jupstar ✪> but cant open f1 console in that version
17:48 < bridge> <Jupstar ✪> ah nice, ingame it works
17:49 < bridge> <Jupstar ✪> https://cdn.discordapp.com/attachments/293493549758939136/1237793420861243472/image.png?ex=663cf074&is=663b9ef4&hm=60dbfecb8a0842f72103335c112c5f12f43064692904fbe1ce2192c1d473033f&
17:49 < bridge> <Jupstar ✪> with vulkan
17:49 < bridge> <Jupstar ✪> epyc
17:49 < bridge> <heinrich5991> theoretically, it shouldn't need storage permissions, right?
17:50 < bridge> <Jupstar ✪> 600 FPS
17:50 < bridge> <Jupstar ✪> https://cdn.discordapp.com/attachments/293493549758939136/1237793663275503736/image.png?ex=663cf0ae&is=663b9f2e&hm=03f1f4bd892d367feaf6d6e1a4bc41b9f230fa3adff147cf8812a3cc69e31560&
17:50 < bridge> <Jupstar ✪> epyc
17:50 < bridge> <Jupstar ✪> xd
17:50 < bridge> <Jupstar ✪> better than most intel laptops
17:50 < bridge> <Jupstar ✪> :lol:
17:50 < bridge> <heinrich5991> I see @flockes_fussel
17:50 < bridge> <cofil75k> my too xd
17:50 < bridge> <heinrich5991> nice 🙂
17:51 < bridge> <meloƞ> :poggers2: very cool
17:51 < bridge> <robyt3> If an android app only accesses its own files it shouldn't need it, but I don't know how this storage system interactive with native code, maybe the permissions are  also required because of SDL
17:52 < bridge> <robyt3> If an android app only accesses its own files it shouldn't need it, but I don't know how this storage system interacts with native code, maybe the permissions are  also required because of SDL
17:52 < bridge> <Jupstar ✪> i wonder if it works now bsc i started the old version
17:52 < bridge> <heinrich5991> ah, didn't think of SDL
17:52 < bridge> <robyt3> It currently uses
17:52 < bridge> <robyt3> ```
17:52 < bridge> <robyt3> <uses-permission android:name="android.permission.ACCESS_WIFI_STATE"/>
17:52 < bridge> <robyt3> <uses-permission android:name="android.permission.CHANGE_WIFI_MULTICAST_STATE"/>
17:52 < bridge> <robyt3> <uses-permission android:name="android.permission.ACCESS_NETWORK_STATE"/>
17:52 < bridge> <robyt3> <uses-permission android:name="android.permission.READ_EXTERNAL_STORAGE"/>
17:52 < bridge> <robyt3> <uses-permission android:name="android.permission.WRITE_EXTERNAL_STORAGE"/>
17:52 < bridge> <robyt3> <uses-permission android:name="android.permission.INTERNET" />
17:52 < bridge> <robyt3> ```
17:52 < bridge> <Jupstar ✪> GLES 3 also works without problems
17:52 < bridge> <Jupstar ✪> https://cdn.discordapp.com/attachments/293493549758939136/1237794388831113277/image.png?ex=663cf15b&is=663b9fdb&hm=f0532e99fafa1723193cca30c4a47c912e542bc3dec32a6e66c3fc801d34424a&
17:53 < bridge> <Jupstar ✪> from a quick look, everything renders without bugs
17:53 < bridge> <robyt3> nice, so only the emulators that's pretty broken, it also had ~5 FPS only :pepeW:
17:53 < bridge> <Jupstar ✪> weird
17:53 < bridge> <Jupstar ✪> u can enable angle on the emulator
17:54 < bridge> <Jupstar ✪> i think i tested it and it gave me performance equal to desktop pc
17:54 < bridge> <heinrich5991> you're a wizard, @robyt3 ^^
17:55 < bridge> <Jupstar ✪> @robyt3 do you use WSL?
17:56 < bridge> <Jupstar ✪> can u send me x86_64 version?
17:56 < bridge> <robyt3> I used my Ubuntu VM, and I should have expanded the storage beforehand 😄
17:56 < bridge> <Jupstar ✪> then i can test on android emulator
17:57 < bridge> <robyt3> I can send you x86, I don't have x86_64 prepared
17:57 < bridge> <Jupstar ✪> i can try if that works
17:57 < bridge> <learath2> but 32 is less than 64, I think you should wait
17:57 < bridge> <meloƞ> nah CI is a damn troll
17:58 < bridge> <meloƞ> 
17:58 < bridge> <meloƞ> macOS wants a long long int to be  %lld 
17:58 < bridge> <meloƞ> while ubuntu wants his long long int to be %ld
17:58 < bridge> <meloƞ> 
17:58 < bridge> <meloƞ> mac:
17:58 < bridge> <meloƞ> ```m
17:58 < bridge> <meloƞ> /Users/runner/work/ddnet/ddnet/src/engine/shared/netban.h:236:64: error: format specifies type 'long' but the argument has type 'int64_t' (aka 'long long') [-Werror,-Wformat]
17:58 < bridge> <meloƞ>                         str_format(pBuf, BuffSize, "%s for %ld minutes (%s)", aBuf, Mins, pBan->m_Info.m_aReason);
17:58 < bridge> <meloƞ>                                                            ~~~                      ^~~~
17:58 < bridge> <meloƞ>                                                         %lld
17:58 < bridge> <meloƞ> ```
17:58 < bridge> <meloƞ> 
17:58 < bridge> <meloƞ> ubuntu:
17:58 < bridge> <meloƞ> ```
17:58 < bridge> <meloƞ> /home/runner/work/ddnet/ddnet/src/engine/shared/netban.h:236:63: error: format ‘%lld’ expects argument of type ‘long long int’, but argument 5 has type ‘int64_t’ {aka ‘long int’} [-Werror=format=]
17:58 < bridge> <meloƞ>   236 |                         str_format(pBuf, BuffSize, "%s for %lld minutes (%s)", aBuf, Mins, pBan->m_Info.m_aReason);
17:58 < bridge> <meloƞ>       |                                                            ~~~^                      ~~~~
17:58 < bridge> <meloƞ>       |                                                               |                      |
17:58 < bridge> <meloƞ>       |                                                               long long int          int64_t {aka long int}
17:58 < bridge> <meloƞ>       |                                                            %ld
17:58 < bridge> <meloƞ> ```
17:58 < bridge> <meloƞ> 
17:58 < bridge> <meloƞ> :angy:
17:58 < bridge> <Jupstar ✪> i'm entering a dangerous path here
17:58 < bridge> <robyt3> https://cdn.discordapp.com/attachments/293493549758939136/1237795773387771997/DDNet-x86.7z.001?ex=663cf2a5&is=663ba125&hm=7a548be1f2034423f80f587915ca957bbee36edab88543e4a1a32cc9fca91efb&
17:58 < bridge> <robyt3> https://cdn.discordapp.com/attachments/293493549758939136/1237795773761196123/DDNet-x86.7z.002?ex=663cf2a5&is=663ba125&hm=7f44722ea7c1d4fb0392a782c0e6deb68b86a5e2595e67b87ed97dbd3ea36d52&
17:58 < bridge> <Jupstar ✪> lmao
17:58 < bridge> <Jupstar ✪> we have some limit?
17:58 < bridge> <robyt3> Yeah, discord didn't want the whole thing
17:59 < bridge> <learath2> Neither are the correct way to print an `int64_t` that's the issue there
17:59 < bridge> <meloƞ> roby casually spreading malware :greenthing:
17:59 < bridge> <meloƞ> :MonkaS:
17:59 < bridge> <robyt3> I guess I didn't need to split it, 7z did pretty well
17:59 < bridge> <robyt3> Got down to the size of the arm64
17:59 < bridge> <learath2> You are looking for `PRId64`
18:00 < bridge> <learath2> `"%s for " PRId64 " minutes (%s)"`
18:00 < bridge> <robyt3> You need another percent, it's not included in the macro `"%s for %" PRId64 " minutes (%s)"`
18:00 < bridge> <Jupstar ✪> i only have andriod 8 with x86, let's see
18:00 < bridge> <learath2> Ah missed it
18:01 < bridge> <meloƞ> sometimes i wonder if i even understand anything
18:01 < bridge> <meloƞ> thanks
18:01 < bridge> <learath2> Oh since this is C++ you need to define something iirc
18:02 < bridge> <learath2> Oh, nvm I checked and that was removed after C99
18:02 < bridge> <learath2> So you can just use it and it should exist in any compliant implementation just fine
18:04 < bridge> <Jupstar ✪> @robyt3 tile maps are not rendering at all, and for some reason my emulator crashes when i want to access the device emulation settings xD
18:04 < bridge> <Jupstar ✪> i try to update android studio
18:04 < bridge> <Jupstar ✪> but i'd not say i have few fps
18:05 < bridge> <Jupstar ✪> felt quite fast in the ui
18:06 < bridge> <robyt3> The emulator was never very fast for me, maybe a Windows limitation
18:06 < bridge> <robyt3> But I didn't update it or Android Studio in a while either
18:07 < bridge> <Jupstar ✪> so the emulator itself was native on windows?
18:08 < bridge> <Jupstar ✪> (not in VM)
18:08 < bridge> <robyt3> yeah, I compiled everything in the Ubuntu VM but I also read the emulator on my Windows host
18:08 < bridge> <robyt3> yeah, I compiled everything in the Ubuntu VM but I already had the emulator on my Windows host
18:08 < bridge> <Jupstar ✪> is iommu activated in your UEFI?
18:09 < bridge> <Jupstar ✪> 😬
18:09 < bridge> <Jupstar ✪> https://cdn.discordapp.com/attachments/293493549758939136/1237798476381945917/image.png?ex=663cf529&is=663ba3a9&hm=b779dbca680f20e771dbd228bb1420a9f182fffadc6ad0f36589594f12bcfc50&
18:09 < bridge> <Jupstar ✪> does it want me to login or what
18:09 < bridge> <Jupstar ✪> i hate java
18:13 < bridge> <Jupstar ✪> well guess it's broken now :lol:
18:13 < bridge> <robyt3> intel virtualization should be enabled, yeah
18:13 < bridge> <Jupstar ✪> ah nice i just deleted everything related to android studio now it works
18:17 < bridge> <Jupstar ✪> epyc gamer moment
18:17 < bridge> <Jupstar ✪> @robyt3
18:17 < bridge> <Jupstar ✪> https://cdn.discordapp.com/attachments/293493549758939136/1237800642739699804/image.png?ex=663cf72e&is=663ba5ae&hm=c6737d0ea56ad65fa01ea5a3856a26cd5b030da606fc3d27eb7b0579a50c867c&
18:17 < bridge> <Jupstar ✪> these are the settings i meant
18:18 < bridge> <Jupstar ✪> there u can enable native opengl etc
18:18 < bridge> <ryozuki> ```rust
18:18 < bridge> <ryozuki> #[doc(hidden)]
18:18 < bridge> <ryozuki> pub mod ඞ {
18:18 < bridge> <ryozuki>     pub use ::regex_automata;
18:18 < bridge> <ryozuki>     pub use ::regex_syntax;
18:18 < bridge> <ryozuki> }
18:18 < bridge> <ryozuki> ```
18:18 < bridge> <ryozuki> https://docs.rs/ctreg/latest/src/ctreg/lib.rs.html#46
18:20 < bridge> <ryozuki> https://github.com/Gankra/abi-cafe
18:20 < bridge> <Jupstar ✪> well dunno doesn't work
18:20 < bridge> <robyt3> Looks better already. Is it possible to disable the keyboard shortcuts of the emulator?
18:20 < bridge> <Jupstar ✪> but @robyt3 x86_64 would be nice.. i don't trust my 32bit drivers
18:21 < bridge> <Jupstar ✪> i am relatively sure i never had trouble with x64
18:21 < bridge> <Jupstar ✪> i dunno
18:22 < bridge> <robyt3> Seems like it catches Escape, so you can't open the menu ingame
18:22 < bridge> <Jupstar ✪> true
18:22 < bridge> <robyt3> I'll build it later
18:22 < bridge> <Jupstar ✪> arrows keys also dont work
18:23 < bridge> <Jupstar ✪> console is also bit broken lmao
18:23 < bridge> <Jupstar ✪> https://cdn.discordapp.com/attachments/293493549758939136/1237802083638313052/image.png?ex=663cf885&is=663ba705&hm=742cfa9d3e7290c7780d1f688934f8b54761905ec80f9cfc23d4366d0cfc384c&
18:23 < bridge> <ryozuki> @learath2 do u know about https://github.com/libuv/libuv
18:24 < bridge> <learath2> Yes. Never used it, though I considered it for ddnet once upon a time
18:26 < bridge> <learath2> Some others you might want to look at are libev and libevent
18:32 < bridge> <AssassinTee> some tools, that you can make with the ddnet repository, aren't in the default download, like map_create_pixelart for example. Can I download it somewhere?
18:32 < bridge> <Jupstar ✪> isnt it part of the editor
18:33 < bridge> <AssassinTee> is that tool behind "Add tileart"? lol
18:33 < bridge> <Jupstar ✪> i dunno
18:33 < bridge> <Jupstar ✪> sounds a bit like that
18:34 < bridge> <AssassinTee> oh wow, thank you
18:34 < bridge> <AssassinTee> it doesn't convert it into a row of quads as I expected
18:37 < bridge> <AssassinTee> @jupeyy_keks there are apperently 2 pixelart tools, one that adds a pixelart as tileset, group and tilelayer, and one that adds it as quads: https://github.com/ddnet/ddnet/pull/5766
18:37 < bridge> <AssassinTee> the secondary isn't available in any compiled form
18:37 < bridge> <Jupstar ✪> u have to ping deen then, i think he uploads tools manually
18:37 < bridge> <ryozuki> https://twitter.com/tsoding/status/1788032651531637247
18:38 < bridge> <AssassinTee> @0xdeen manually 🤷‍♂️ the `map_create_pixelart` tool is missing from downloads
18:39 < bridge> <learath2> Lol, true
18:39 < bridge> <AssassinTee> hmm, goto where
18:40 < bridge> <AssassinTee> hmm, goto somewhere
18:44 < bridge> <Jupstar ✪> why don't u simply builds yourself xd
18:44 < bridge> <AssassinTee> working on it, but building on windows is always ...
18:44 < bridge> <Jupstar ✪> oh yeah
18:44 < bridge> <Jupstar ✪> i bet patiga has the same tool
18:44 < bridge> <Jupstar ✪> xdd
18:45 < bridge> <AssassinTee> I am using all of patigas tools, never had it so easy to build and install anything, guess rust made it easier
18:45 < bridge> <patiga> :justatest:
18:45 < bridge> <patiga> mine works differently
18:46 < bridge> <patiga> https://patiga.eu/blog/ddnet-pixelart
18:47 < bridge> <AssassinTee> omg, next time I ask you directly if you have YET ANOTHER tool for me
18:47 < bridge> <AssassinTee> holy shit, that's 3 in a row patiga
18:47 < bridge> <patiga> xd
18:48 < bridge> <patiga> its a different approach with other upsides/downsides
18:49 < bridge> <0xdeen> add where?
18:49 < bridge> <Jupstar ✪> there was a tool download once
18:49 < bridge> <Jupstar ✪> but it's gone xd
18:50 < bridge> <AssassinTee> @0xdeen anywhere, there are no releases on github and there is no download for it in the ddnet download section. Also it's not part of the default or nightly download
18:50 < bridge> <0xdeen> yes
18:51 < bridge> <patiga> most significantly, it always does the pixelart with 3 tiles layers. the upside is that it can create nearly every single color value. however, if you have a image with very few colors (< 255 * 3), the other tool's approach is objectively better
18:53 < bridge> <patiga> the blog explains the math behind it, it was a nice puzzle :)
18:53 < bridge> <AssassinTee> @patiga the tool I linked uses quads, while the other tool that jupestar meant, and your python script, use tiles, is that right? Because I need quads, because I need smaller pixelarts
18:53 < bridge> <patiga> yup, I use tiles
18:53 < bridge> <Jupstar ✪> what's wrong with tiles?
18:54 < bridge> <Jupstar ✪> do u need to scale it?
18:54 < bridge> <AssassinTee> my very very specific usecase @jupeyy_keks xD
18:54 < bridge> <AssassinTee> https://cdn.discordapp.com/attachments/293493549758939136/1237809881894092820/screenshot_2024-05-01_11-11-55.png?ex=663cffc9&is=663bae49&hm=6df8f433eb1651643d3e204af3506d7de106b5a6b9d21e17b238f1e3e85829c6&
18:54 < bridge> <AssassinTee> https://cdn.discordapp.com/attachments/293493549758939136/1237809882472910978/screenshot_2024-05-07_20-14-46.png?ex=663cffc9&is=663bae49&hm=56aa57f72cef16686593bac5f605dbaeb895d6e49a39c8e550ebe7fab9e46bab&
18:54 < bridge> <AssassinTee> my pixelsize is basically 1/8th of a tile
18:57 < bridge> <patiga> *gough* would be a relatively easy script with *cough*
18:57 < bridge> <Jupstar ✪> what even is the use case of using it with quads instead of just embedding the image itself?
18:57 < bridge> <patiga> *cough* would be a relatively easy script with *cough*
18:57 < bridge> <patiga> *twmap-py*
18:58 < bridge> <patiga> I'd whip up vim rn, but I'm ingame ^^
19:02 < bridge> <AssassinTee> If I use a true size pixel art image, scaling it up will create artifacts, otherwise the filesize will explode with uninterpolated upscaling, because every pixel takes 8x8 pixels
19:05 < bridge> <Jupstar ✪> ah right we have no texture wrap modes for quads
19:05 < bridge> <Jupstar ✪> that sucks
19:05 < bridge> <Jupstar ✪> sampler modes
19:07 < bridge> <AssassinTee> here is an example, guess where the pixelart is
19:07 < bridge> <AssassinTee> https://cdn.discordapp.com/attachments/293493549758939136/1237813034303623348/screenshot_2024-05-08_19-06-42.png?ex=663d02b8&is=663bb138&hm=a7bf40673a8235a441574304741d70a1756dd78bfb186a3ea59a500762df1338&
19:08 < bridge> <AssassinTee> https://cdn.discordapp.com/attachments/293493549758939136/1237813310091563009/hello_pxl.png?ex=663d02fa&is=663bb17a&hm=e77020371e08b740f534a23c6a2b0d6079151ce7b204674345d21407d856bcf9&
19:44 < bridge> <robyt3> https://cdn.discordapp.com/attachments/293493549758939136/1237822566069174342/DDNet-x86_64.apk?ex=663d0b99&is=663bba19&hm=149748ac8cb68cfc3b120ab0bba0db790c0d7b6fc81b542ebf2251827301d610&
19:46 < bridge> <Jupstar ✪> works without problems
19:48 < bridge> <Jupstar ✪> lmao have to use scrcpy to the emulator to use my keyboard
19:48 < bridge> <Jupstar ✪> so stupid
19:48 < bridge> <Jupstar ✪> https://cdn.discordapp.com/attachments/293493549758939136/1237823510114471947/image.png?ex=663d0c7a&is=663bbafa&hm=5d51ce5b4b353f03103def16db7f011479cdcdec36061d346a481b996ad9f86f&
19:49 < bridge> <Jupstar ✪> renders fine.. fps are also okish
19:49 < bridge> <Jupstar ✪> https://cdn.discordapp.com/attachments/293493549758939136/1237823629350404169/image.png?ex=663d0c96&is=663bbb16&hm=7c3a7ebde8f2f39039a954c74f98f40fff436677f26fa3764ae78774dc7e97e3&
19:49 < bridge> <Jupstar ✪> 300 fps for emulator, quite ok i guess
19:55 < bridge> <robyt3> Looks like you could use the Android Accelerometer gamecontroller for left and right controls :kek:
19:55 < bridge> <Jupstar ✪> left and right works but arrow keys, esc etc not
20:00 < bridge> <learath2> I need to find some C++ portfolio thing to do. I want to do some library work, anyone have any libraries they wish they had in C++?
20:01 < bridge> <learath2> Currently I'm thinking maybe a state machine generalization, maybe with coroutines
20:02 < bridge> <learath2> Maybe something with more algo work would be nice too. I need lots of algorithms practice too but I couldn't come up with anything
20:03 < bridge> <meloƞ> task scheduler, like these AI advertisements you see everywhere
20:03 < bridge> <learath2> Or maybe I should do some golang portfolio projects. Japanese developers love go and a job in japan would be cool
20:04 < bridge> <learath2> Task as in asynchronous programming?
20:05 < bridge> <learath2> I do have a coroutine scheduler that I was messing around with, maybe I can clean that up and publish it
20:05 < bridge> <meloƞ> like a coroutine-based scheduler - like task scheduling etc
20:06 < bridge> <learath2> AI coroutine scheduler. Uses reinforcement learning to schedule tasks :thinkW:
20:06 < bridge> <meloƞ> or write yabl (yet another benchmark library) - this should require quite a lot of algorithms
20:06 < bridge> <meloƞ> or write yabl (yet another benchmark library) - this should require quite a lot of algorithms ig
20:10 < bridge> <learath2> I doubt it beats the algorithms humans developed though 😄
20:11 < bridge> <ryozuki> the future is weeb ais
20:12 < bridge> <meloƞ> i read web ai's and thought you meant like online picture generators, until i realized what you meant... :weebsout:
20:12 < bridge> <learath2> Mh training ai is just legwork. This is why I never dabbled in it
20:12 < bridge> <meloƞ> i read web ai's and thought you meant like online picture generators, until i reread what you wrote... :weebsout:
20:12 < bridge> <ryozuki> yeah
20:13 < bridge> <learath2> You just feed data into a library and it trains models. The interesting work in that niche is in developing algorithms and libraries. I'm not smart enough for either
20:14 < bridge> <patiga> something neat simd accelerated maybe ^^
20:14 < bridge> <ryozuki> the interesting work is also MLIR
20:14 < bridge> <ryozuki> compilers for AI stuff
20:14 < bridge> <learath2> Simd accelerated bans handling in teeworlds 😄
20:14 < bridge> <ryozuki> icould make a teeworlds mlir dialect
20:14 < bridge> <ryozuki> ok that makes 0 sense but its funny
20:14 < bridge> <patiga> teeworlds variable int parsing with simd is something I already put some thought into but couldn't find a satisfying solution
20:14 < bridge> <meloƞ> give it a single: ban all racist slurs
20:14 < bridge> <meloƞ> 
20:14 < bridge> <meloƞ> and dude rips away 50% of the playerbase in an instance
20:15 < bridge> <learath2> I actually have a simd accelerated ban check for ddnet. It didn't improve anything so I never pr'd it 😄
20:15 < bridge> <ryozuki> ban is a rare thing xd
20:15 < bridge> <patiga> though some bigger use-case is probably more interesting anyways ^^
20:15 < bridge> <ryozuki> wouldnt it be better SIMD packet parsing or smth
20:15 < bridge> <ryozuki> or decoding*
20:15 < bridge> <meloƞ> do.. you mind.. sharing.. if you have it? this is something blockworlds could use :kek:
20:15 < bridge> <learath2> The ban table is large, looking up ips in it is getting slower and slower since it's linear
20:15 < bridge> <ryozuki> i have avx512 but idk what to do
20:15 < bridge> <patiga> I'm just brainstorming here :p
20:15 < bridge> <ryozuki> i never did simd stuff
20:16 < bridge> <patiga> I did some simd stuff for a c project in university
20:16 < bridge> <patiga> where there was a competition with speed
20:16 < bridge> <patiga> but in the end my simd parser was slower in practice
20:16 < bridge> <learath2> Just put the ips in a btree instead. I got better performance out of that and simd lookup in the btree added no value
20:16 < bridge> <ryozuki> @patiga read this https://en.algorithmica.org/hpc/
20:16 < bridge> <patiga> why?
20:16 < bridge> <meloƞ> fair
20:17 < bridge> <meloƞ> it explains why algorithms are great af
20:17 < bridge> <meloƞ> `100x faster matrix multiplication (compared to “for-for-for”)`
20:17 < bridge> <learath2> It's really hard to get this right yeah. Simd instructions are slow, so you need to fully utilize them for it to be worth it
20:18 < bridge> <patiga> + you need to worry about error handling :c
20:18 < bridge> <meloƞ> always expected Simd was developed for the sole purpose of enhancing speed :D
20:18 < bridge> <learath2> PSHUFB is an amazing instruction for lookups
20:20 < bridge> <learath2> (Not for longer than a byte though)
20:21 < bridge> <patiga> hm yea that is neat :)
20:33 < bridge> <Jupstar ✪> i have many that u could port from c to rust
20:33 < bridge> <Jupstar ✪> if that is also an option
20:35 < bridge> <Jupstar ✪> a better data structure would already do lot
20:35 < bridge> <Jupstar ✪> the data structure sucks so hard
20:35 < bridge> <Jupstar ✪> what even is it, a normal linked list lmao
20:36 < bridge> <learath2> It wasn't much of an issue with our older policies, but nowadays mods go hard with longer and longer bans
20:37 < bridge> <Jupstar ✪> i use https://crates.io/crates/iprange currently
20:37 < bridge> <Jupstar ✪> works quite good i think
20:37 < bridge> <Jupstar ✪> only did smaller testing
20:39 < bridge> <Jupstar ✪> uses a radix tree.. makes sense i guess
20:41 < bridge> <learath2> Just putting the banned ips in an actual contiguous array actually helps a lot already. The simple linear search gets vectorized easily by the dumbest of compilers
20:42 < bridge> <learath2> And it plays very well with prefetch
20:42 < bridge> <learath2> However, our n kinda has surpassed the limits of linear searches, so a tree like structure is nice
20:46 < bridge> <ryozuki> https://m.benui.ca/@ben/112396505994216742
20:56 < bridge> <learath2> I have a friend that treats me like I'm insane because I don't like AI. Whenever I think about it more carefully, I notice the people who support this kind of thing never contribute to the common wealth of knowledge. They only know how to take, so for them this is a positive. It makes copying code from stackoverflow easier
20:57 < bridge> <Jupstar ✪> uff that is a bit naive tbh.
20:57 < bridge> <Jupstar ✪> 
20:57 < bridge> <Jupstar ✪> Our goal is to automate stuff, and AI will be a required tool
20:57 < bridge> <ryozuki> i like AI tech wise but moral wise ppl find the wrong thing to do always xd
20:57 < bridge> <Jupstar ✪> if u dont like OpenAI fine
20:58 < bridge> <learath2> Our goal is to automate stuff for ourselves. Not for "Open"AI Inc. to automate it for a huge profit off the backs of our free labour
21:00 < bridge> <learath2> By all means train your ai on your own proprietary codebases, hire people to write code for you to train your ai on. You don't just get to train it on millions of stackoverflow users contributions and sell it back to them
21:01 < bridge> <Jupstar ✪> fine, it sounded more like u generally like the idea of what AIs will do in future
21:02 < bridge> <learath2> No that part is unavoidable. I just don't like the transfer of knowledge from the open to closed
21:03 < bridge> <learath2> If it was an open model, training on open data for all of us, I wouldn't care that much
21:04 < bridge> <learath2> Same reason I never really had an issue with what stable diffusion did
22:29 < bridge> <Ewan> http://2143.christmas/f/Z1nV.png
22:29 < bridge> <Ewan> i love this game
22:32 < bridge> <Ewan> as far as AI goes
22:33 < bridge> <Ewan> i think it's something that we should rely on our governments to regulate
22:33 < bridge> <Ewan> but in the interest of money it will never happen
22:34 < bridge> <heinrich5991> anyone wants to read some proper FUD? https://t.me/durov/274
22:37 < bridge> <heinrich5991> anyone wants to read some proper FUD? https://t.me/durov/274
22:37 < bridge> <Ewan> something keeps changing but i can't tell what
22:37 < bridge> <Ewan> who's deleting shit xd
22:37 < bridge> <heinrich5991> hmmm. it seems my message is getting autodeleted
22:37 < bridge> <Ewan> weird
22:38 < bridge> <heinrich5991> base64: YW55b25lIHdhbnRzIHRvIHJlYWQgc29tZSBwcm9wZXIgRlVEPyBodHRwczovL3QubWUvZHVyb3YvMjc0
22:41 < bridge> <Ewan> i bet t.me is filtered automatically
22:41 < bridge> <Ewan> because of such high scam likelihood
22:42 < bridge> <peter0x44> Nothing with such a high level of significance ever gets regulated properly
22:42 < bridge> <peter0x44> The lobbying forces are strong
22:42 < bridge> <peter0x44> Democracy isn't real
22:42 < bridge> <heinrich5991> are lobbying forces something that brought us the GDPR?
22:42 < bridge> <heinrich5991> I would not think so
22:43 < bridge> <heinrich5991> t.me
22:43 < bridge> <heinrich5991> t.me/durov/274
22:43 < bridge> <Ewan> http://2143.christmas/f/dNCt.png
22:43 < bridge> <Ewan> goodbye
22:43 < bridge> <heinrich5991> https://t.me/durov/274
22:44 < bridge> <heinrich5991> <https://t.me/durov/274>
22:44 < bridge> <Ewan> haha yep
22:45 < bridge> <heinrich5991> anyone wants to read some proper FUD? hxxps://t.me/durov/274
22:45 < bridge> <heinrich5991> perfect. now I only need to find out why t.me links are blocked
22:46 < bridge> <Jupstar ✪> why not?
22:46 < bridge> <Jupstar ✪> 
22:46 < bridge> <Jupstar ✪> US corps are clearly the world leaders, so why should EU corps not have an interest to give them less power?
22:47 < bridge> <heinrich5991> is that related to US vs EU in some way?
22:47 < bridge> <Jupstar ✪> i think yes
22:47 < bridge> <heinrich5991> both EU and US companies have to adhere to the GDPR when offering services to EU citizens
22:47 < bridge> <Jupstar ✪> if EU would have google and co, GDPR would probs not exist
22:47 < bridge> <heinrich5991> california passed laws similar to the GDPR IIRC
22:50 < bridge> <heinrich5991> or there is https://de.wikipedia.org/wiki/Lieferkettensorgfaltspflichtengesetz, which doesn't seem to be motivated by the kind of lobbying @peter0x44 referred to, either
22:54 < bridge> <louis> i mean putting together huge models without charging somehow would be p difficult
22:55 < bridge> <louis> unlesss there are open/free projects out there to pool compute for open models
22:55 < bridge> <Jupstar ✪> mh yeah. hard topic, u can always read it from two sides. if there is no lobby against smth then there might not be a high economical power for such stuff in first place or the law is does only benefit bigger corps anyway (e.g. Lieferkettengesetz might be harder for small corps).
22:55 < bridge> <Jupstar ✪> 
22:55 < bridge> <Jupstar ✪> For example could you explain why the EU decided to shift the combustion engine to a later date then planned?
22:55 < bridge> <Jupstar ✪> the end of*
22:55 < bridge> <heinrich5991> Lieferkettengesetz does not apply to small companies, only 1000 employees and larger
22:55 < bridge> <heinrich5991> because of the german car industry
22:56 < bridge> <heinrich5991> the ICE thing is because of the german car industry
22:56 < bridge> <heinrich5991> and because of the fdp IIRC
23:03 < bridge> <peter0x44> Well I don't know what's involved with that but there's the usual carbon credits and such like that (which are obviously problematic)
23:03 < bridge> <peter0x44> 
23:03 < bridge> <peter0x44> In practice a fine is just a cost of doing business
23:04 < bridge> <heinrich5991> not related to carbon credits
23:05 < bridge> <heinrich5991> here's the english version: https://en.wikipedia.org/wiki/Supply_Chain_Act
23:06 < bridge> <peter0x44> A fine of 2% is nothing
23:06 < bridge> <peter0x44> That's not enforcement
23:06 < bridge> <peter0x44> If it grows revenues by more, then it's worth it
23:07 < bridge> <peter0x44> A threat of forced nationalization, or something that actually has teeth...
23:08 < bridge> <heinrich5991> 2% of revenue
23:08 < bridge> <heinrich5991> that's a lot
23:08 < bridge> <peter0x44> If violating it can cause growth more than that, it's worth it
23:08 < bridge> <peter0x44> It really isn't
23:09 < bridge> <heinrich5991> do you believe that this would cause companies today to do losses rather than wins, if a 2% fine was applied?
23:09 < bridge> <heinrich5991> do you believe that this would cause companies today to have losses rather than wins, if a 2% fine was applied?
23:10 < bridge> <peter0x44> I don't know
23:10 < bridge> <heinrich5991> then you're oddly sure about that this isn't "having teeth"
23:10 < bridge> <Jupstar ✪> i guess u'd need to ask nestles secret stats for that :lol: 
23:11 < bridge> <Jupstar ✪> 
23:11 < bridge> <Jupstar ✪> is it worth hiding their shady activities in order to stay market leader
23:11 < bridge> <peter0x44> In practice fines aren't enough when things go wrong
23:11 < bridge> <heinrich5991> [citation needed]
23:11 < bridge> <peter0x44> If BP was "punished" properly for deepwater horizon it wouldn't exist today
23:11 < bridge> <Jupstar ✪> EU is defs an interesting experiment
23:11 < bridge> <Jupstar ✪> the only thing that could kill us is our aging society
23:11 < bridge> <heinrich5991> completely unrelated to the point
23:12 < bridge> <peter0x44> 3M putting pfas in water
23:12 < bridge> <heinrich5991> completely unrelated
23:12 < bridge> <peter0x44> It isn't
23:12 < bridge> <heinrich5991> you're talking about incidents that are bad, as an argument that completely unrelated laws are useless
23:12 < bridge> <peter0x44> I didn't call them useless, they just aren't enough
23:14 < bridge> <Jupstar ✪> @peter0x44 i guess you are more refering to "big companies seem to be punished less when something unfair happens"
23:14 < bridge> <Jupstar ✪> 
23:14 < bridge> <Jupstar ✪> which could always be true, independent of the type of law
23:33 < bridge> <robyt3> Here's the debug build for all architectures
23:33 < bridge> <robyt3> https://cdn.discordapp.com/attachments/293493549758939136/1237880016126218320/DDNet.7z.001?ex=663d411a&is=663bef9a&hm=821231c7dc6e173ca289ddb1586c9b9498f367efd1180c8aa97966ab82255ce1&
23:33 < bridge> <robyt3> https://cdn.discordapp.com/attachments/293493549758939136/1237880016562294954/DDNet.7z.002?ex=663d411a&is=663bef9a&hm=9109d857ae9b18f0fd8b6ef2154448ce7a6a35083bb86ba49985b3133a978ec4&
23:33 < bridge> <robyt3> https://cdn.discordapp.com/attachments/293493549758939136/1237880016956690432/DDNet.7z.003?ex=663d411a&is=663bef9a&hm=7b4df9e7c1e701cd165ec25be8c214e3420f3e4c5c315629dd9ac62daa352648&
23:33 < bridge> <robyt3> https://cdn.discordapp.com/attachments/293493549758939136/1237880017350823987/DDNet.7z.004?ex=663d411a&is=663bef9a&hm=b8d0f70e6184d2b2be53e2b624052849087531ea41afd3e5d337c4c3f50d4a86&
23:46 < bridge> <robyt3> Also works on my phone :poggers:
23:48 < bridge> <meloƞ> just imagine the hype, that after 9 versions android download is back
23:48 < bridge> <robyt3> It still needs UI touch input and ingame input in general though
23:49 < bridge> <robyt3> The weird relative mouse input hack we have for android is terrible to use
23:57 < bridge> <Jupstar ✪> yeah it's really about this xD
23:57 < bridge> <Jupstar ✪> we already had android builds a few versions ago
23:57 < bridge> <Jupstar ✪> but nobody cared. if that doesn't change, then this won't help in this case either
23:58 < bridge> <Jupstar ✪> tbh relative mouse mode sucks for UI anyway
23:58 < bridge> <robyt3> SDL supports touch events so the UI should be doable
23:59 < bridge> <Jupstar ✪> https://github.com/Jupeyy/ddnet/commit/3b1f72eb83beffff3c35287ab6f222b85187b984
23:59 < bridge> <Jupstar ✪> change my mind
23:59 < bridge> <robyt3> For ingame controls we need to think about something
23:59 < bridge> <Jupstar ✪> it uses relative for ingame
23:59 < bridge> <Jupstar ✪> and absolute for ui
23:59 < bridge> <Jupstar ✪> i improved on that and consequently use that for dd-pg
23:59 < bridge> <Jupstar ✪> sadly with SDL it made problems on i3 and other half finished WMs