00:39 <+bridge_> [ddnet] <Chairn> @heinrich5991 how are the files supposed to be executed for #5017 ?
00:39 <+bridge_> [ddnet] <DDNet> https://github.com/ddnet/ddnet/issues/5017
00:40 <+bridge_> [ddnet] <Chairn> i miss the files 'UnicodeData.txt'
00:51 <+bridge_> [ddnet] <Chairn> okay, i managed to retrieve files from the texlive source that i had on my disk 😄
00:52 <+bridge_> [ddnet] <Chairn> so `UnicodeData.txt` and `confusables.txt`
00:52 <+bridge_> [ddnet] <Chairn> thanks `locate` 🙂
02:27 <+bridge_> [ddnet] <M Emile> @Hav
03:17 <+bridge_> [ddnet] <Nagi01 {LAN}> now i can say `i use arch btw`
04:56 <+bridge_> [ddnet] <Jeremy> oh boy
06:09 <+bridge_> [ddnet] <lynn> is it possible to change the coordinates to integers with subpixel values
06:10 <+bridge_> [ddnet] <lynn> or is this not worthwhile
06:10 <+bridge_> [ddnet] <lynn> it would break existing maps that depend on floating point stuff maybe
06:10 <+bridge_> [ddnet] <Learath2> It would break a lot, no way
06:10 <+bridge_> [ddnet] <Learath2> Also gpus don't really love integers, Jupstar might not enjoy it either
06:11 <+bridge_> [ddnet] <lynn> gotcha
06:11 <+bridge_> [ddnet] <lynn> that's very odd
06:11 <+bridge_> [ddnet] <lynn> but i guess it makes sense
06:11 <+bridge_> [ddnet] <Learath2> Which part is odd?
06:11 <+bridge_> [ddnet] <lynn> oh the GPU liking floats
06:12 <+bridge_> [ddnet] <lynn> sprry
06:12 <+bridge_> [ddnet] <lynn> sorry
06:12 <+bridge_> [ddnet] <Learath2> Ah, yeah, gpu people love their floats
06:12 <+bridge_> [ddnet] <lynn> qualitative logic or something idk
06:13 <+bridge_> [ddnet] <lynn> embracing imprecision
06:17 <+bridge_> [ddnet] <lynn> oh i had an idea
06:18 <+bridge_> [ddnet] <lynn> some play_sound command so you can even have audio cues for binds
06:18 <+bridge_> [ddnet] <lynn> or to quickly test your sounds in game maybe
08:30 <+bridge_> [ddnet] <Jupstar> why should this be a joke?
08:30 <+bridge_> [ddnet] <Jupstar> 
08:30 <+bridge_> [ddnet] <Jupstar> i find it inresponsible that we havent done it yet actually
08:30 <+bridge_> [ddnet] <Jupstar> every shitty static html site has encryption nowadays
08:31 <+bridge_> [ddnet] <Jupstar> its not that i wouldnt enjoy better math than floats, but changing everything to ints is not an easy task
08:31 <+bridge_> [ddnet] <Jupstar> and would defs require to drop gl 1.x
08:33 <+bridge_> [ddnet] <Jupstar> i mean in the end floats were invented to make real number math kinda possible. But inaccuracy ofc can suck, for graphical stuff its probs not that horrible since pixels are kinda a huge mathimatical limiation already
08:33 <+bridge_> [ddnet] <Jupstar> 😄
08:34 <+bridge_> [ddnet] <Learath2> I thought you'd have performance concerns with using ints
08:35 <+bridge_> [ddnet] <Jupstar> i can not say if ints solve a problem for us. But i guess new hardware to handle fixed size floating point types would be at least cool
08:36 <+bridge_> [ddnet] <Jupstar> real double precision on GPUs is also cool, independent of if they'd fix smth for us
09:11 <+bridge_> [ddnet] <chillerdragon> Ah Classic. But there is no place where I can check if this is the case right?
09:11 <+bridge_> [ddnet] <chillerdragon> (@deen)
09:25 <+bridge_> [ddnet] <Ryozuki> this is true
09:29 <+bridge_> [ddnet] <Ryozuki> > There is no real reason to encrypt ddnet chat
09:29 <+bridge_> [ddnet] <Ryozuki> is something someone who doesnt value privacy would say
09:40 <+bridge_> [ddnet] <Ryozuki> https://cdn.discordapp.com/attachments/293493549758939136/970228271465783316/RDT_20220501_094030754543657093547411.jpg
09:56 <+bridge_> [ddnet] <heinrich5991> the scripts/generate_unicode_confusables_data.py script tells you where to get UnicodeData.txt and confusables.txt
09:56 <+bridge_> [ddnet] <heinrich5991> encrypting ddnet chat without authenticating the encryption doesn't do much
10:00 <+bridge_> [ddnet] <Learath2> It would prevent packet injection I guess
10:03 <+bridge_> [ddnet] <heinrich5991> hm? what do you mean?
10:03 <+bridge_> [ddnet] <heinrich5991> it'd allow people to manually detect MITM if we displayed the session fingerprint somewhere
10:03 <+bridge_> [ddnet] <heinrich5991> people usually don't do that though
10:04 <+bridge_> [ddnet] <Learath2> If we do dh kex, and use that to encrypt mitm can't inject anything into the connection, can it?
10:04 <+bridge_> [ddnet] <heinrich5991> it mitms the dh kex
10:05 <+bridge_> [ddnet] <heinrich5991> you can fundamentally never be MITM free without a way to authenticate your peer out-of-band. the server in the middle can just chat with each of you and reencrypt everything
10:06 <+bridge_> [ddnet] <Learath2> I was thinking of injection after the connection got established
10:06 <+bridge_> [ddnet] <heinrich5991> ah yes, that's secure then
10:16 <+bridge_> [ddnet] <deen> chillerdragon: only my emails 😄
11:12 <+bridge_> [ddnet] <heinrich5991> it seems that there are two different concepts tangled up in the console/logging
11:12 <+bridge_> [ddnet] <heinrich5991> one is logging
11:12 <+bridge_> [ddnet] <heinrich5991> the other is "responses to the command you just entered" (for rcon, chat commands)
11:18 <+bridge_> [ddnet] <Learath2> Yep, those two have been conjoined forever
11:29 <+bridge_> [ddnet] <Jupstar> whats the reason if ping completly missing?
11:29 <+bridge_> [ddnet] <Jupstar> 
11:29 <+bridge_> [ddnet] <Jupstar> geo ip lookup failed?
11:29 <+bridge_> [ddnet] <Jupstar> https://cdn.discordapp.com/attachments/293493549758939136/970255523452702720/unknown.png
11:32 <+bridge_> [ddnet] <heinrich5991> that's a bug
11:41 <+ChillerDragon> @deen where can i signup for that mailing list? hehe
11:42 <+ChillerDragon> so your hoster always sends you a mail when a new ddos is detected and he starts filtering?
11:43 <+bridge_> [ddnet] <heinrich5991> hetzner does that, at least
11:43 <+ChillerDragon> i see
11:43 <+ChillerDragon> i am having such a hard time to find the motive behind the constant ddos over years thats why i am a bit skeptic if it is always ddos or something else we can maybe fix
11:44 <+ChillerDragon> "ddos" seems to be the goto excuse. I see it in game a lot when people just have a poor connection they immedetialy suspect beging ddosed -.-
11:46 <+bridge_> [ddnet] <deen> GER2 last week, every pike is an attack. For FRA it's harder to show since the hoster blocks them
11:46 <+bridge_> [ddnet] <deen> https://cdn.discordapp.com/attachments/293493549758939136/970259963425472532/ger2.ddnet.tw-net-7d.png
11:46 <+bridge_> [ddnet] <deen> spike*
11:47 <+ChillerDragon> that is the unfiltered traffic?
11:47 <+bridge_> [ddnet] <heinrich5991> logarithmic scale!
11:47 <+bridge_> [ddnet] <deen> yes
11:47 <+ChillerDragon> that wont be shown on the ddnet website?
11:47 <+bridge_> [ddnet] <deen> It's from https://ddnet.tw/stats/server/
11:47 <+ChillerDragon> any chance to make this graph public?
11:47 <+ChillerDragon> a
11:47 <+ChillerDragon> ok nvm then
11:47 <+bridge_> [ddnet] <heinrich5991> https://ddnet.tw/stats/server/
11:47 <+ChillerDragon> heinrich u still got wire?
11:48 <+bridge_> [ddnet] <heinrich5991> you and fokkonaut are my last contacts there. I don't open it regularly
11:48 <+ChillerDragon> fokkonaut :D 
11:48 <+ChillerDragon> i just slid in your dms i dont want to discuss this in public -.-
11:49 <+bridge_> [ddnet] <Learath2> Ooh secrets
11:50 <+ChillerDragon> nudes
11:52 <+bridge_> [ddnet] <fokkonaut> xd
12:14 <+bridge_> [ddnet] <c0d3d3v> tls encription of a html side is e2s not e2e, so if you want to encrypt chat of a game e2e we would get attention of NSA soon
12:15 <+bridge_> [ddnet] <heinrich5991> no, that's not a problem
12:15 <+bridge_> [ddnet] <heinrich5991> the NSA is not interested in a game just because it has e2e chat
12:15 <+bridge_> [ddnet] <fokkonaut> @heinrich5991 wouldnt it be cleaner to not log the chat responses anyways so instead of using Print() use SendChatTarget directly like its done in any other places. The old commands still have Print because thats probably from greyfox ddrace
12:16 <+bridge_> [ddnet] <fokkonaut> the chat responses always fill the logs while it doesnt make sense imo
12:16 <+bridge_> [ddnet] <fokkonaut> the messages from SendChat, also server messages sent to everyone get logged is right, as its shown to everyone
12:20 <+bridge_> [ddnet] <heinrich5991> `SendChatTarget` doesn't really make sense either, since it's more of a "respond to the person executing the command"
12:20 <+bridge_> [ddnet] <heinrich5991> the same problem exists for rcon commands
12:20 <+bridge_> [ddnet] <heinrich5991> anyway, with the PR, the chat responses are no longer in the logs
12:22 <+bridge_> [ddnet] <fokkonaut> isnt that what is wanted here? xd
12:23 <+bridge_> [ddnet] <heinrich5991> what is wanted is "respond to the person executing the command"
12:23 <+bridge_> [ddnet] <heinrich5991> but `SendChatTarget` doesn't fit that description
12:25 <+bridge_> [ddnet] <Jupstar> lmao, do you fear the NSA? XD
12:28 <+bridge_> [ddnet] <fokkonaut> why?
12:29 <+bridge_> [ddnet] <fokkonaut> its sends a chat message output to the target being the executor of the command
12:29 <+bridge_> [ddnet] <fokkonaut> xd
12:29 <+bridge_> [ddnet] <fokkonaut> Print doesnt make sense here, since Print would be print to console (rcon, local, etc)
12:44 <+bridge_> [ddnet] <gerdoe> :troll:
12:45 <+bridge_> [ddnet] <Jupstar> what do u have trouble with? 😄
12:46 <+bridge_> [ddnet] <Nagi01 {LAN}> hey, what do you guys use to generate this chart ?
12:49 <+bridge_> [ddnet] <heinrich5991> no, because you have to care about the fact that the command sends it to chat and need to track the caller yourself
12:50 <+bridge_> [ddnet] <heinrich5991> this doesn't work generically, if you did that in an rcon command, the output would end up in the wrong place (the chat)
12:50 <+bridge_> [ddnet] <heinrich5991> we'd need some way to do this response generically, in a way that works with rcon and chat, IMO
12:56 <+bridge_> [ddnet] <gerdoe> just walking around with idea of youtube like demo database site kekw
12:57 <+bridge_> [ddnet] <gerdoe> so current problem is demo render
12:59 <+bridge_> [ddnet] <Jupstar> u can use the client in the web to render demos
13:00 <+bridge_> [ddnet] <gerdoe> :kek: nice
13:01 <+bridge_> [ddnet] <Jupstar> https://aliveclan.de/ddnettest/load_map.html
13:01 <+bridge_> [ddnet] <Jupstar> 
13:01 <+bridge_> [ddnet] <Jupstar> here example
13:01 <+bridge_> [ddnet] <Jupstar> 
13:01 <+bridge_> [ddnet] <Jupstar> maybe a client side cache for the ddnet data would be nice to add, so it doesnt always load it from the web
13:01 <+bridge_> [ddnet] <Jupstar> 
13:01 <+bridge_> [ddnet] <Jupstar> 
13:01 <+bridge_> [ddnet] <Jupstar> dunno what the limits in javascript localstorage are or the databases
13:02 <+bridge_> [ddnet] <Jupstar> the "load demo" button can be replaced by a url parsing
13:03 <+bridge_> [ddnet] <Jupstar> u can also call javascript code directly in c++ with emscripten
13:03 <+bridge_> [ddnet] <Jupstar> that makes stuff really ez to comunicate
13:51 <+ChillerDragon> i think localstorage in browser can be pretty high i remember some dude doing full blown database stuff in the client D 
13:57 <+ChillerDragon> @Jupstar when ddnet electron app? for 1000000 fps... oh wait
13:58 <+bridge_> [ddnet] <Jupstar> we need -1 fps
13:58 <+bridge_> [ddnet] <Jupstar> then we overflow eventually
13:58 <+bridge_> [ddnet] <Jupstar> or underflow xd
14:00 <+ChillerDragon> smart
14:05 <+bridge_> [ddnet] <Ryozuki> browsers have a indexdb
14:05 <+bridge_> [ddnet] <Ryozuki> or smth like that
14:05 <+bridge_> [ddnet] <Ryozuki> https://developer.mozilla.org/en-US/docs/Web/API/IndexedDB_API
14:44 <+bridge_> [ddnet] <Cipy29> Wut? Did you also run into that? xd
15:03 <+bridge_> [ddnet] <deen> RRD and some scripts https://github.com/ddnet/ddnet-scripts/tree/master/rrd as well as the raw serverstatus data from https://ddnet.tw/status/
15:03 <+bridge_> [ddnet] <deen> I wrote a bit about it here: https://hookrace.net/blog/server-statistics/
15:06 <+bridge_> [ddnet] <Nagi01 {LAN}> ty, rrd looks very clean and suckless
15:09 <+bridge_> [ddnet] <Ryozuki> :monkalaugh:
15:55 <+bridge_> [ddnet] <Voxel> woah
15:55 <+bridge_> [ddnet] <Voxel> https://github.com/teeworlds/teeworlds/blob/5c5b54c8eb00ac70c4eb1bd09b74554a0c865bdc/datasrc%2Fgame.png
15:55 <+bridge_> [ddnet] <Voxel> whats this weapon?
15:55 <+bridge_> [ddnet] <Voxel> https://cdn.discordapp.com/attachments/293493549758939136/970322677443227678/unknown.png
15:57 <+bridge_> [ddnet] <Jupstar> harpoon
15:57 <+bridge_> [ddnet] <heinrich5991> harpoon, I think
15:57 <+bridge_> [ddnet] <Voxel> interesting
16:54 <+bridge_> [ddnet] <nouaa> what it does? something like a hook?
16:55 <+bridge_> [ddnet] <louis> yeah its like a hook
16:55 <+bridge_> [ddnet] <louis> it can grab hearts and other weapons too
16:56 <+bridge_> [ddnet] <nouaa> :poggers:
16:57 <+ChillerDragon> spawnblocker on 176.9.114.238:8331
16:58 <+ChillerDragon> sos
17:15 <+bridge_> [ddnet] <c0d3d3v> has someone an idea why we need id for snapshots? I have not read the complete code, but I have no clue yet why they are necessary or what they help with 😄 but they can be used to DOS the servers xD https://youtu.be/G3nVtdH0--Q
17:15 <+bridge_> [ddnet] <c0d3d3v> sell also #5054 and #4723
17:15 <+bridge_> [ddnet] <DDNet> https://github.com/ddnet/ddnet/issues/5054
17:15 <+bridge_> [ddnet] <DDNet> https://github.com/ddnet/ddnet/issues/4723
17:16 <+bridge_> [ddnet] <c0d3d3v> has someone an idea why we need IDs for snapshots? I have not read the complete code, but I have no clue yet why they are necessary or what they help with 😄 but they can be used to DOS the servers xD https://youtu.be/G3nVtdH0--Q
17:16 <+bridge_> [ddnet] <c0d3d3v> sell also #5054 and #4723
17:16 <+bridge_> [ddnet] <Jupstar> i wonder why we assert
17:16 <+bridge_> [ddnet] <Jupstar> but deen didnt really give me a real answer
17:16 <+bridge_> [ddnet] <Jupstar> all snapitems are checked for nullptr
17:18 <+bridge_> [ddnet] <c0d3d3v> has someone an idea why we need IDs for snapshots? I have not read the complete code, but I have no clue yet why they are necessary or what they help with 😄 but they can be used to DOS the servers xD https://youtu.be/G3nVtdH0--Q  (the DOS attack starts at 1:13)
17:18 <+bridge_> [ddnet] <c0d3d3v> sell also #5054 and #4723
17:19 <+bridge_> [ddnet] <c0d3d3v> has someone an idea why we need IDs for snapshots? I have not read the complete code, but I have no clue yet why they are necessary or what they help with 😄 but they can be used to DOS the servers xD https://youtu.be/G3nVtdH0--Q  (the DOS attack starts at 1:13) In the terminal I log the free snap ids
17:19 <+bridge_> [ddnet] <c0d3d3v> sell also #5054 and #4723
17:19 <+bridge_> [ddnet] <c0d3d3v> has someone an idea why we need IDs for snapshots? I have not read the complete code, but I have no clue yet why they are necessary or what they help with 😄 but they can be used to DOS the servers xD https://youtu.be/G3nVtdH0--Q  (the DOS attack starts at 1:13) In the terminal I log the number free snap ids
17:19 <+bridge_> [ddnet] <c0d3d3v> sell also #5054 and #4723
17:22 <+ChillerDragon> eeedits.....
17:23 <+bridge_> [ddnet] <Jupstar> improve quakenet and add edits xd
17:24 <+bridge_> [ddnet] <c0d3d3v> has someone an idea why we need IDs for snapshots? I have not read the complete code, but I have no clue yet why they are necessary or what they help with 😄 but they can be used to DOS the servers xD https://youtu.be/G3nVtdH0--Q  (the DOS attack starts at 1:13) In the terminal I log the number of free snap ids
17:24 <+bridge_> [ddnet] <c0d3d3v> sell also #5054 and #4723
18:32 <+bridge_> [ddnet] <Chairn> gpus loves integers
18:37 <+bridge_> [ddnet] <Chairn> any cpu/gpu prefer integers over float
18:38 <+bridge_> [ddnet] <Jupstar> tho certain calculations are slower on ints
18:38 <+bridge_> [ddnet] <Jupstar> than flots
18:38 <+bridge_> [ddnet] <Jupstar> a
18:38 <+bridge_> [ddnet] <Chairn> floats are a mess to handle at the hardware level
18:38 <+bridge_> [ddnet] <Chairn> which ones?
18:38 <+bridge_> [ddnet] <Jupstar> division for example
18:38 <+bridge_> [ddnet] <Chairn> for single precision yes
18:38 <+bridge_> [ddnet] <Chairn> you still have a 23 bits division
18:39 <+bridge_> [ddnet] <Chairn> for double, you have a 50 sth bits division, so better than 64 bits
18:39 <+bridge_> [ddnet] <Chairn> but for instance, if we had to change all floats to integers in DDNet, 32 bits would be enough
18:39 <+bridge_> [ddnet] <Chairn> with 16 bits integer part and 16 bits fractional part for fixed point
18:40 <+bridge_> [ddnet] <Jupstar> probs could work yeah
18:40 <+bridge_> [ddnet] <Chairn> well, we need roughly 5 or 6 bits for the fractional part due to a tile being 32 pixels
18:40 <+bridge_> [ddnet] <Chairn> or 16 is more than enough
18:41 <+bridge_> [ddnet] <Chairn> so 16 is more than enough
18:41 <+bridge_> [ddnet] <Jupstar> but 16 bits for integer part also means we have a strict maximum
18:41 <+bridge_> [ddnet] <Jupstar> 1 tile is 32 units i think in the simulation part
18:41 <+bridge_> [ddnet] <Jupstar> so around 2048 tiles
18:42 <+bridge_> [ddnet] <Jupstar> or put that in the fractional part
18:42 <+bridge_> [ddnet] <Chairn> right now, maximum velocity is around 10k, maximum map size is around 4k ?
18:42 <+bridge_> [ddnet] <Jupstar> but would require a rewrite
18:42 <+bridge_> [ddnet] <Chairn> not really
18:42 <+bridge_> [ddnet] <Chairn> sed -i 's/float/Q1616/g' **
18:42 <+bridge_> [ddnet] <Jupstar> xd
18:42 <+bridge_> [ddnet] <Chairn> just define a proper class with proper operator overload
18:42 <+bridge_> [ddnet] <Chairn> i already have one
19:02 <+bridge_> [ddnet] <Jupstar> i would assume even a short would loose in division against double
19:02 <+bridge_> [ddnet] <Jupstar> 
19:02 <+bridge_> [ddnet] <Jupstar> do u have some benchmarks?
19:12 <+bridge_> [ddnet] <Ryozuki> developers coding ddnet
19:12 <+bridge_> [ddnet] <Ryozuki> https://cdn.discordapp.com/attachments/293493549758939136/970372210827092028/unknown.png
19:20 <+bridge_> [ddnet] <everestkio> yes
19:20 <+bridge_> [ddnet] <everestkio> now
19:20 <+bridge_> [ddnet] <everestkio> Graphics()->TextureSet(Graphics()->LoadTexture("imagesogay.png", IStorage::TYPE_ALL, CImageInfo::FORMAT_AUTO, 0));
19:20 <+bridge_> [ddnet] <everestkio> why this show me a multicolor nothing? xd
19:23 <+bridge_> [ddnet] <Jupstar> i think i havent sit straighter than rn now after seeing this xD
19:37 <+bridge_> [ddnet] <Chairn> not really, but look at the operation itself
19:37 <+bridge_> [ddnet] <Chairn> short division is 16 bits integer division
19:37 <+bridge_> [ddnet] <Chairn> double division is complex operation
19:38 <+bridge_> [ddnet] <Jupstar> but i only trust numbers dxd
19:38 <+bridge_> [ddnet] <Chairn> lemme look at my old courses
19:39 <+bridge_> [ddnet] <Chairn> in hardware design, its really clear
19:41 <+bridge_> [ddnet] <Chairn> this is what i have in a course
19:41 <+bridge_> [ddnet] <Chairn> 1. handle special case for denormalized numbers
19:41 <+bridge_> [ddnet] <Chairn> 2. substract exponents and divide mantissas (integer division)
19:42 <+bridge_> [ddnet] <Chairn> 3. renormalize mantissa if it's lower than 1 and update exponent
19:42 <+bridge_> [ddnet] <Chairn> if one of the number is denormalized, 3. has extra step to ensure the value of the result mantissa
19:43 <+bridge_> [ddnet] <Chairn> 4. truncate results and handle rounding
19:43 <+bridge_> [ddnet] <Chairn> 5. handle overflow or underflow
19:43 <+bridge_> [ddnet] <Chairn> in the design itself, you see that floating point division is way more complex
19:44 <+bridge_> [ddnet] <Chairn> as division is a quite complex and rare operation, hardware operators are not pipelined so they have poor throughput
19:44 <+bridge_> [ddnet] <Jupstar> but integer divisions are basically a loop
19:44 <+bridge_> [ddnet] <Jupstar> doesnt need to be faster
19:46 <+bridge_> [ddnet] <Jupstar> e.g.
19:46 <+bridge_> [ddnet] <Jupstar> https://stackoverflow.com/a/41570985
19:48 <+bridge_> [ddnet] <Chairn> they don't give units ??
19:49 <+bridge_> [ddnet] <Chairn> but i think you and i talk about 2 different thing
19:49 <+bridge_> [ddnet] <Chairn> you're interested in throughput
19:49 <+bridge_> [ddnet] <Chairn> i am interested in one operation time
19:49 <+bridge_> [ddnet] <Chairn> i believe they have pipelined floating point operator while integers one are not and takes 5 full cycles
19:52 <+bridge_> [ddnet] <Jupstar> mh ok, i'd still like to see numbers, but yes its not 100% comparable per se
19:52 <+bridge_> [ddnet] <Jupstar> the delay of the processing units might also take into account eventually
19:52 <+bridge_> [ddnet] <Jupstar> hard to say
19:52 <+bridge_> [ddnet] <Jupstar> maybe if u finished your module for teeworlds we can see
19:52 <+bridge_> [ddnet] <Chairn> im also looking for the number of execution unit
19:53 <+bridge_> [ddnet] <Chairn> integers division is often single execution unit while floating point have multiple execution unit
19:53 <+bridge_> [ddnet] <trml> are there many divisions in the code (except by constant values)?
19:53 <+bridge_> [ddnet] <Chairn> as the core are OOO, they benefit more from the multiple execution units when performing integer operations
19:54 <+bridge_> [ddnet] <Jupstar> i guess length math
19:54 <+bridge_> [ddnet] <Jupstar> and i dunno how exactly specific functions are implemented
19:54 <+bridge_> [ddnet] <Jupstar> sin, sqrt etf
19:54 <+bridge_> [ddnet] <Jupstar> sin, sqrt etc
19:54 <+bridge_> [ddnet] <Chairn> ```bash
19:54 <+bridge_> [ddnet] <Chairn> grep -rin '=.*/' | wc -l
19:54 <+bridge_> [ddnet] <Chairn> 1771
19:54 <+bridge_> [ddnet] <Chairn> ```
19:55 <+bridge_> [ddnet] <Jupstar> xd our grep pro
19:55 <+bridge_> [ddnet] <Chairn> floating point functions are often polynomes approximation
19:55 <+bridge_> [ddnet] <Chairn> + some table for faster initialization of first rounds
19:55 <+bridge_> [ddnet] <Jupstar> r u good in recursive regex, dunno if u call it like that
19:55 <+bridge_> [ddnet] <Jupstar> 
19:55 <+bridge_> [ddnet] <Jupstar> but i basically need to find a string(i dunno the name but pattern)
19:55 <+bridge_> [ddnet] <Jupstar> 
19:55 <+bridge_> [ddnet] <Jupstar> and match against that string
19:55 <+bridge_> [ddnet] <Chairn> looks like most division are by 32 😄
19:56 <+bridge_> [ddnet] <Chairn> you can do that with python regex
19:56 <+bridge_> [ddnet] <Chairn> you can match a string, and match only if it rematches later
19:58 <+bridge_> [ddnet] <Jupstar> i didnt find a solution for vscodium so guess i need to use an external tool ^^
19:58 <+bridge_> [ddnet] <Chairn> <https://en.wikichip.org/wiki/intel/microarchitectures/kaby_lake#Core> develop execution unit
19:59 <+bridge_> [ddnet] <Chairn> hum, doesn't say anything about floating point EU
20:01 <+bridge_> [ddnet] <trml> oh, nice. it seems most are by constants, but its probably used often for things like length/normalizing vectors yes
20:02 <+bridge_> [ddnet] <Ryozuki> eh, vscodium uses ripgrep, written in rust iirc, it by default doesnt have backreferences and look arounds enabled so it performs in linear time
20:03 <+bridge_> [ddnet] <Jupstar> yes, i also never needed it 😄
20:03 <+bridge_> [ddnet] <Jupstar> in my whole life XD
20:03 <+bridge_> [ddnet] <Jupstar> but then everything changed
20:03 <+bridge_> [ddnet] <Jupstar> and the firenation begin to attack
20:04 <+bridge_> [ddnet] <Chairn> i believe it's actually 2 EU for floating point, and one for integer division
20:04 <+bridge_> [ddnet] <Chairn> + the fact that floating point operation are compiled into SSE2 and not x87
20:05 <+bridge_> [ddnet] <Jupstar> anyway fixed floating points math is interesting, so defs finish it if u have time
20:06 <+bridge_> [ddnet] <Chairn> oh, i never tried it, i just said i have already a class to do this that i developped for Arduino
20:06 <+bridge_> [ddnet] <Chairn> because Arduino doesn't have floating point hardware, so it definitely helped in that case
20:07 <+bridge_> [ddnet] <Jupstar> then start xdd
20:07 <+bridge_> [ddnet] <Jupstar> :rplace:
20:07 <+bridge_> [ddnet] <Chairn> but it doesn"t have hardware divider either so it still relied on complex algorithm to perform the remaining division 😄
20:07 <+bridge_> [ddnet] <Chairn> i think you mean :rnow:
20:07 <+bridge_> [ddnet] <Chairn> i think you meant :rnow:
20:07 <+bridge_> [ddnet] <Jupstar> soon™️
20:08 <+bridge_> [ddnet] <Chairn> as you can see here, looks like the division is pipelined <https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=div&techs=SSE,SSE2&ig_expand=2734,2707,2689>
20:09 <+bridge_> [ddnet] <Chairn> throughput is lower than latency so it looks like they have between 2 to 3 stages divider
20:09 <+bridge_> [ddnet] <Jupstar> but if it can be pipelined its still an advantage tho
20:09 <+bridge_> [ddnet] <Chairn> it is 🙂
20:10 <+bridge_> [ddnet] <Chairn> it is for desktop, not for embedded device (higher power consumption)
20:10 <+bridge_> [ddnet] <Jupstar> yeah 😄
20:11 <+bridge_> [ddnet] <Chairn> LOL, did they use excel to fill the tables 😄
20:11 <+bridge_> [ddnet] <Chairn> https://cdn.discordapp.com/attachments/293493549758939136/970386954686451772/unknown.png
20:11 <+bridge_> [ddnet] <Jupstar> November? XD
20:11 <+bridge_> [ddnet] <Chairn> `11-14` seems to have been misinterpreted 😄
20:11 <+bridge_> [ddnet] <Jupstar> xD
20:22 <+bridge_> [ddnet] <Ryozuki> https://howtoprofessionallysay.akashrajpurohit.com/
20:22 <+bridge_> [ddnet] <Ryozuki> xd
20:27 <+bridge_> [ddnet] <Ryozuki> https://developers.redhat.com/articles/2022/04/25/new-c-features-gcc-12#c__23_features
20:30 <+bridge_> [ddnet] <Jupstar> LMAO
20:30 <+bridge_> [ddnet] <Ryozuki> WHAT?
20:30 <+bridge_> [ddnet] <Jupstar> this
20:30 <+bridge_> [ddnet] <Jupstar> litteraly mentioned in the article
20:30 <+bridge_> [ddnet] <Ryozuki> https://cdn.discordapp.com/attachments/293493549758939136/970391873426448484/unknown.png
20:30 <+bridge_> [ddnet] <Ryozuki> lol
20:31 <+bridge_> [ddnet] <Jupstar> tho my context was templates
20:31 <+bridge_> [ddnet] <Jupstar> but bet that works too xd
20:31 <+bridge_> [ddnet] <Ryozuki> c++ adding rust syntax XDD
20:31 <+bridge_> [ddnet] <Ryozuki> ```
20:31 <+bridge_> [ddnet] <Ryozuki> if consteval {
20:31 <+bridge_> [ddnet] <Ryozuki> 
20:31 <+bridge_> [ddnet] <Ryozuki> }```
20:31 <+bridge_> [ddnet] <Ryozuki> (rust doesnt have () on ifs)
20:32 <+bridge_> [ddnet] <Jupstar> ah but this is special case
20:32 <+bridge_> [ddnet] <Ryozuki> ye
20:32 <+bridge_> [ddnet] <Jupstar> but yeah 😄
20:32 <+bridge_> [ddnet] <Ryozuki> its a good feature
20:32 <+bridge_> [ddnet] <Ryozuki> c++ is actually quite advanced in const eval stuff
20:32 <+bridge_> [ddnet] <Ryozuki> rust needs to catch up
20:33 <+bridge_> [ddnet] <Ryozuki> https://cdn.discordapp.com/attachments/293493549758939136/970392586734936134/unknown.png
20:33 <+bridge_> [ddnet] <Ryozuki> lol u can now do this
20:33 <+bridge_> [ddnet] <Jupstar> but i'd assume a lot of cases are handled by higher optimization flags anyway arent they
20:34 <+bridge_> [ddnet] <Jupstar> tho function calls cannot be removed maybe
20:34 <+bridge_> [ddnet] <Jupstar> but ofast and lto doesnt care xd
20:34 <+bridge_> [ddnet] <Ryozuki> https://doc.rust-lang.org/reference/const_eval.html
20:36 <+bridge_> [ddnet] <Ryozuki> when enable c++23
20:36 <+bridge_> [ddnet] <Jupstar> xd
20:36 <+bridge_> [ddnet] <Jupstar> deen even upgraded to debian 10
20:37 <+bridge_> [ddnet] <Jupstar> thats already against the plan xD
20:40 <+bridge_> [ddnet] <Nagi01 {LAN}> lel interesting 😄
20:40 <+bridge_> [ddnet] <Ryozuki> xd
20:40 <+bridge_> [ddnet] <Nagi01 {LAN}> is there a palce where i can find every divisors for the teeworlds assets ?
20:41 <+bridge_> [ddnet] <Jupstar> conent.py or smth
20:41 <+bridge_> [ddnet] <Jupstar> content.py or smth
20:41 <+bridge_> [ddnet] <Chairn> woohoo, multidimensional [] !!
20:42 <+bridge_> [ddnet] <Jupstar> yes
20:43 <+bridge_> [ddnet] <Jupstar> kinda cool
20:44 <+bridge_> [ddnet] <Nagi01 {LAN}> ty, i check
20:45 <+bridge_> [ddnet] <Ryozuki> le this be reminded https://cor3ntin.github.io/posts/abi/
20:46 <+bridge_> [ddnet] <Ryozuki> google kept supporting clang till the ABI breaking vote failed (aka we dont break the abi)
20:47 <+bridge_> [ddnet] <Ryozuki> https://cdn.discordapp.com/attachments/293493549758939136/970395937891573810/unknown.png
20:47 <+bridge_> [ddnet] <Jupstar> google still uses it for android tho
20:47 <+bridge_> [ddnet] <Ryozuki> dont they use https://abseil.io/
20:47 <+bridge_> [ddnet] <Ryozuki> instead of the stdlib
20:48 <+bridge_> [ddnet] <Jupstar> internally?
20:48 <+bridge_> [ddnet] <Jupstar> maybe i dunno
21:10 <+bridge_> [ddnet] <Ryozuki> https://github.com/openmainframeproject/cobol-programming-course
21:10 <+bridge_> [ddnet] <Ryozuki> im gonna learn this and work in a boring bank
21:10 <+bridge_> [ddnet] <Ryozuki> with good pay maybe
21:10 <+bridge_> [ddnet] <Ryozuki> :monkalaugh:
21:43 <+bridge_> [ddnet] <Chairn> at the expense of your mental health :monkalaugh:
21:44 <+bridge_> [ddnet] <Ryozuki> ill have rust therapy sessions at home while at it
21:44 <+bridge_> [ddnet] <Ryozuki> :greenthing:
21:45 <+bridge_> [ddnet] <Chairn> or maybe you'll turn into a secessionist terrorrist
21:45 <+bridge_> [ddnet] <Chairn> holding back all wiring to spanish people and sending them to catalonia
21:45 <+bridge_> [ddnet] <Ryozuki> :monkaS:
21:45 <+bridge_> [ddnet] <Ryozuki> good idea
21:46 <+bridge_> [ddnet] <Ryozuki> they already steal us
21:46 <+bridge_> [ddnet] <Ryozuki> so i would be robin hood
21:46 <+bridge_> [ddnet] <Chairn> don't forget to get 1 cent off everytime 😄
21:46 <+bridge_> [ddnet] <Ryozuki> i would only take what was ours
21:46 <+bridge_> [ddnet] <Ryozuki> :BASED:
21:47 <+bridge_> [ddnet] <Ryozuki> i hope the rust job i applied goes through, the guy told me they always answer (whether u get chosen or not)
21:47 <+bridge_> [ddnet] <Ryozuki> i have yet to get a reply
21:48 <+bridge_> [ddnet] <Ryozuki> :monkaS:
21:48 <+bridge_> [ddnet] <Chairn> like when you create a class (dont know cobol), overload operator + and minus like that:
21:48 <+bridge_> [ddnet] <Chairn> ```c++
21:48 <+bridge_> [ddnet] <Chairn> decimal operator+(const decimal& lhs, const decimal& rhs)
21:48 <+bridge_> [ddnet] <Chairn> {
21:48 <+bridge_> [ddnet] <Chairn>     decimal res(lhs.i+rhs.i-0.01);
21:48 <+bridge_> [ddnet] <Chairn>     myaccount.wire(0.01);
21:48 <+bridge_> [ddnet] <Chairn>     return res;
21:48 <+bridge_> [ddnet] <Chairn> }
21:48 <+bridge_> [ddnet] <Chairn> ```
21:48 <+bridge_> [ddnet] <Chairn> that sounds like a physicist joke
21:49 <+bridge_> [ddnet] <Chairn> i always reply, just in a non finite time
21:49 <+bridge_> [ddnet] <Ryozuki> xd
21:49 <+bridge_> [ddnet] <Ryozuki> nah but he said it seriously, when we finished the interview
21:50 <+bridge_> [ddnet] <Ryozuki> cobol is like screaming always
21:50 <+bridge_> [ddnet] <Ryozuki> functions dont have arguments
21:50 <+bridge_> [ddnet] <Ryozuki> rather they modify outside variables
21:50 <+bridge_> [ddnet] <Ryozuki> xd
21:50 <+bridge_> [ddnet] <Ryozuki> i have yet to start learning it tho
21:50 <+bridge_> [ddnet] <Chairn> `tr 'a-z' 'A-Z'` there you go
21:52 <+bridge_> [ddnet] <Ryozuki> https://cdn.discordapp.com/attachments/293493549758939136/970412367588507678/unknown.png
21:52 <+bridge_> [ddnet] <Chairn> i like this wikipedia page with SCREAMING_SNAKE_CASE: <https://en.wikipedia.org/wiki/Snake_case>
21:53 <+bridge_> [ddnet] <Ryozuki> > Add the following to your awareness of COBOL. It is an absolute myth that you must be at least 50 years
21:53 <+bridge_> [ddnet] <Ryozuki> > old to be good with COBOL. COBOL is incredibly easy to learn and understand. One of the many reasons
21:53 <+bridge_> [ddnet] <Ryozuki> > financial institutions like COBOL is the fact that it is not necessary to be a programmer to read and
21:53 <+bridge_> [ddnet] <Ryozuki> > understand the logic. This is important because critical business logic code is subject to audit. Auditors
21:53 <+bridge_> [ddnet] <Ryozuki> > are not programmers. However, auditors are responsible for ensuring the business financial statements are
21:53 <+bridge_> [ddnet] <Ryozuki> > presented fairly. It is COBOL processing that frequently results in the business ledger updates and subsequent
21:53 <+bridge_> [ddnet] <Ryozuki> > financial statements.
21:53 <+bridge_> [ddnet] <Ryozuki> interesting
21:53 <+bridge_> [ddnet] <Chairn> so it's more like descriptive language
21:53 <+bridge_> [ddnet] <Ryozuki> the "guetting started pdf" is 158 pages long
21:53 <+bridge_> [ddnet] <Ryozuki> and looks well done
21:53 <+bridge_> [ddnet] <Chairn> kinda like original TeX
21:53 <+bridge_> [ddnet] <Ryozuki> time to truly learn cobol
23:14 <+bridge_> [ddnet] <Ryozuki> https://cdn.discordapp.com/attachments/293493549758939136/970432965911912458/unknown.png
23:14 <+bridge_> [ddnet] <Ryozuki> i++++
23:14 <+bridge_> [ddnet] <Ryozuki> --i++++
23:15 <+bridge_> [ddnet] <Ryozuki> https://cdn.discordapp.com/attachments/293493549758939136/970433361304748053/unknown.png
23:17 <+bridge_> [ddnet] <plsplsplslol> i-=-i
23:46 <+bridge_> [ddnet] <Chairn> lol, does that legit comile?
23:46 <+bridge_> [ddnet] <Chairn> lol, does that legit compile?
23:46 <+bridge_> [ddnet] <Ryozuki> i havent tried