01:21 < bridge> <ChillerDragon> @zwelf2 I quoted you in my protocol docs. Let me know if thats not okay. https://chillerdragon.github.io/teeworlds-protocol/07/snap_items.html#obj_character_core
01:27 < ws-client> <ChillerDragon> This guy just forked the biggest garbage repo I had https://github.com/berkcelc So i wondered why. Now I see it. He is building a github portfolio with forks lmao. Looks like a active githubber at first glance but i am not sure if he ever published even a single commit.
01:28 < ws-client> <ChillerDragon> Oh yes 2 commits and 183 forks xxxxxxxxxD
01:29 < ws-client> <ChillerDragon> I need to learn from this guy to boost my stats
01:52 < bridge> <learath2> turkish moment
03:11 < bridge> <miguilim> can someone explain why the code checks if IsSeqInBackroom? It doesnt seems to make sense to me
03:11 < bridge> <miguilim> https://cdn.discordapp.com/attachments/293493549758939136/1211857999640006747/image.png?ex=65efb9b6&is=65dd44b6&hm=27878426c7fde2a5b7671de19ce838c532b7be1fc36fb62a4bd4e40758fc0fa4&
03:11 < bridge> <miguilim> the future ack is calculated with `% NET_MAX_SEQUENCE`
03:15 < bridge> <miguilim> so the reset of the ack number is already predicted no?
03:15 < bridge> <miguilim> so the reset of the ack number is already predicted, no?
03:57 < ws-client> <ChillerDragon> oh im so looking forward to reading your ack docs :3 
03:57 < ws-client> <ChillerDragon> might fix chat messages being printed twice in my client hehe
03:57 < bridge> <ratsouphd> is DD opensource?
03:58 < bridge> <ratsouphd> like can any joe contribute to the code?
03:59 < ws-client> <ChillerDragon> yes @ratsouphd ddnet is open source and very community driven
04:00 < bridge> <ratsouphd> cool is there a link to the github i would like to read the docs and maybe contribute
04:00 < ws-client> <ChillerDragon> https://github.com/ddnet/ddnet
04:00 < bridge> <fazi.> It is opensource, I'm about to graduate from software engineering at my respective university so I will no doubt contribute.
04:01 < bridge> <ratsouphd> thx
04:01 < ws-client> <ChillerDragon> where are two ppl coming from wanting to contribute at the same time? woo
04:02 < bridge> <ratsouphd> im new to the game and just want to give back to the community
04:02 < bridge> <fazi.> Haha well I won't be contributing until atleast 6 months down the line, I'm graduating in May then taking a very extended vacation to travel after suffering for the past 5 years
04:03 < ws-client> <ChillerDragon> cool
04:08 < bridge> <miguilim> man u will see the snaps doc, I'm rewrintg it right now. I just a snap handler on teeframe and its working 100%
04:10 < bridge> <miguilim> now I have the power of snaps :kek:
04:13 < bridge> <miguilim> man u will see the snaps doc, I'm rewriting it right now. I just implemented a snap handler on teeframe and its working 100%
06:40 < bridge> <miguilim> there it is mate:
06:40 < bridge> <miguilim> 
06:40 < bridge> <miguilim> https://protocol-docs.teeframe.com/fundamentals.html#sequence-number
06:40 < bridge> <miguilim> 
06:41 < bridge> <miguilim> https://protocol-docs.teeframe.com/packets/default-packets.html
06:41 < bridge> <miguilim> https://protocol-docs.teeframe.com/packets/chunks-concept.html
06:41 < bridge> <miguilim> https://protocol-docs.teeframe.com/packets/snaps-concept.html
07:35 < bridge> <ryozuki> https://www.whitehouse.gov/oncd/briefing-room/2024/02/26/press-release-technical-report/
07:35 < bridge> <ryozuki> @learath2 THE UNITED STATES OF AMERICA!!!!!
07:35 < bridge> <ryozuki> 🇺🇸 🇺🇸 🇺🇸 🇺🇸
07:40 < bridge> <ryozuki> good morning btw
07:43 < bridge> <ryozuki> Government mandated rust rewrite
07:43 < bridge> <ryozuki> GMRIIR
07:43 < bridge> <ryozuki> GovRIIR
08:27 < bridge> <ryozuki> https://cdn.discordapp.com/attachments/293493549758939136/1211937527133511700/image.png?ex=65f003c7&is=65dd8ec7&hm=014e2317196bca05da21414bd11d2650f9a1b0078f7d63e33b829cbad23c591c&
08:39 < bridge> <hiyaremir_00254> can yall check #bugs
08:41 < bridge> <ryozuki> its 8:41 bro
08:48 < bridge> <Jupstar ✪> Learath always wanted to contribute to NSA 😬
08:48 < bridge> <matodor> when mathematical safe in rust 
08:48 < bridge> <matodor> :gigachad:
08:48 < bridge> <matodor> https://cdn.discordapp.com/attachments/293493549758939136/1211942956693454848/image.png?ex=65f008d5&is=65dd93d5&hm=f75d3a6e81050f77416f881a29f51bebf02192d99239169bd7ea2b35e4aca4d0&
08:48 < bridge> <ryozuki> its safe
08:50 < bridge> <Jupstar ✪> I guess you could argue a memory bug is a logic bug. But having no memory bugs then means having less logic bugs. But obviously not zero
09:12 < bridge> <milkeeycat> Fair
09:12 < bridge> <milkeeycat> https://cdn.discordapp.com/attachments/293493549758939136/1211948963196833812/IMG_20240227_101203.jpg?ex=65f00e6d&is=65dd996d&hm=66764586b5faa2c2ce6c38757a64c39414d69c5ec57d093da4a618ac53fe50c5&
09:15 < bridge> <heinrich5991> that's really weird. I'd usually say all of them print the same. since you say they don't, maybe 84, 42, 42?
09:17 < bridge> <heinrich5991> ah, not the reason I thought
11:15 < bridge> <Jupstar ✪> I'm surprised it doesn't crash in debug mode tho
11:15 < bridge> <Jupstar ✪> Are there no debug checks enabled
11:17 < bridge> <milkeeycat> 🤷‍♂️
11:37 < bridge> <zhn> @fokkonaut bw servers have more than 1 tw server hosted in different places, they share the same rules. teesports has more that only 1 modified ddnetpp server (that i would say joined teesports eventually) like fng servers (refng by k0rae). idk why you don't get the community status though but bw and teesports host more than one server
11:37 < bridge> <zhn> than*
11:37 < bridge> <fokkonaut> true, BW made a RUS server
11:39 < bridge> <Jupstar ✪> My servers don't even show in master server at all anymore
11:40 < bridge> <heinrich5991> are there error messages in the log?
11:40 < bridge> <Jupstar ✪> I wonder if the attacker switched to dosing them with spoofed IP of my server xd
11:41 < bridge> <Jupstar ✪> I haven't really looked into it yet. But I checked my iptable rules
11:55 < bridge> <meloƞ> i agree with the "needed requirements" since it's not clear to anyone as of today, but to put it into perspective:
11:55 < bridge> <meloƞ> 
11:55 < bridge> <meloƞ> having a single server in the community filter seem's... weird - maybe support a custom badge and an approved stamp for single server's ? so you can atleast get rid of the KoG tab ._.
11:55 < bridge> <meloƞ> https://cdn.discordapp.com/attachments/293493549758939136/1211989890967797791/image.png?ex=65f0348b&is=65ddbf8b&hm=4613b4738b87f6c94bf3f127faf76d08a961b09b5c5bedc7d80e10dd7f38b026&
11:55 < bridge> <meloƞ> https://cdn.discordapp.com/attachments/293493549758939136/1211989891324579870/image.png?ex=65f0348b&is=65ddbf8b&hm=e6c3c084268bcecda5ff745e1f1bf2f6960ca14b55d8ca7cece0db6bfd206baf&
11:55 < bridge> <meloƞ> i agree with the "needed requirements" since it's not clear to anyone as of today, but to put it into perspective:
11:55 < bridge> <meloƞ> 
11:55 < bridge> <meloƞ> having a single server in the community filter seem's... weird - maybe support a custom badge and an approved stamp for single server's ? so you can atleast get rid of the KoG tab ._. - but i agree with you fully there, dont get me wrong :P
11:55 < bridge> <meloƞ> https://cdn.discordapp.com/attachments/293493549758939136/1211989890967797791/image.png?ex=65f0348b&is=65ddbf8b&hm=4613b4738b87f6c94bf3f127faf76d08a961b09b5c5bedc7d80e10dd7f38b026&
11:56 < bridge> <meloƞ> https://cdn.discordapp.com/attachments/293493549758939136/1211989891324579870/image.png?ex=65f0348b&is=65ddbf8b&hm=e6c3c084268bcecda5ff745e1f1bf2f6960ca14b55d8ca7cece0db6bfd206baf&
11:56 < bridge> <fokkonaut> I know that BW and TeeSports have multiple servers, hence I said "de facto 1 server"
11:56 < bridge> <teero777> wdym get rid of kog tab? kog tab isnt there if not favorited
11:57 < bridge> <fokkonaut> Last time I checked only 1 server of each community was actually active
11:58 < bridge> <meloƞ> oh im sorry i somehow overread the "de facto" - sry
11:58 < bridge> <zhn> teesports are quite popular though
11:58 < bridge> <meloƞ> i agree with the "needed requirements" since it's not clear to anyone as of today, but to put it into perspective:
11:58 < bridge> <meloƞ> 
11:58 < bridge> <meloƞ> having a single server in the community filter seem's... weird - maybe support a custom badge and an approved stamp for single server's ? so you can atleast get rid of the KoG tab ._. - but i agree with you fully there, dont get me wrong :P
11:58 < bridge> <meloƞ> https://cdn.discordapp.com/attachments/293493549758939136/1211989890967797791/image.png?ex=65f0348b&is=65ddbf8b&hm=4613b4738b87f6c94bf3f127faf76d08a961b09b5c5bedc7d80e10dd7f38b026&
11:58 < bridge> <fokkonaut> np
11:58 < bridge> <meloƞ> i agree with the "needed requirements" since it's not clear to anyone as of today, but to put it into perspective:
11:58 < bridge> <meloƞ> 
11:58 < bridge> <meloƞ> having a single server in the community filter seem's... weird - maybe support a custom badge and an approved stamp for single server's ? so you can atleast get rid of the KoG tab ._. - but i agree with you fully there, dont get me wrong :P
11:58 < bridge> <meloƞ> i agree with the "needed requirements" since it's not clear to anyone as of today
11:58 < bridge> <meloƞ> 
11:58 < bridge> <meloƞ> having a single server in the community filter seem's... weird - maybe support a custom badge and an approved stamp for single server's ? so you can atleast get rid of the KoG tab ._. - but i agree with you fully there, dont get me wrong :P
11:58 < bridge> <zhn> russian fng gets filled 20/32 and players are switching to them instead
11:59 < bridge> <meloƞ> i also agree with this:
11:59 < bridge> <meloƞ> 
11:59 < bridge> <meloƞ> ```
11:59 < bridge> <meloƞ> Other than that: I have no problem hosting multiple servers or different gamemodes, but I'm not sure if doing the same as other communities and spamming the list with useless servers is the way to go, but if that's the only possibility to gain community status I'll do so.
11:59 < bridge> <meloƞ> ```
11:59 < bridge> <meloƞ> 
11:59 < bridge> <meloƞ> forcing you to open new server's that just sit there as placeholder's shouldnt be a way to get a community tag
11:59 < bridge> <meloƞ> (bw opened them for future plans tho) <a:pepelaugh:1106634559723937853>
12:12 < bridge> <ryozuki> @reitw https://netflixtechblog.com/announcing-bpftop-streamlining-ebpf-performance-optimization-6a727c1ae2e5
12:15 < bridge> <learath2> I can only speculate since I consciously distanced myself from this set of decisions because... well you see.
12:15 < bridge> <learath2> 
12:15 < bridge> <learath2> One of the primary reasons the community feature was made was to group large networks so they don't overwhelm the list. So with the lack of any concrete rules, I'd guess it's indeed the fact that having one server in a group is just weird
12:17 < bridge> <meloƞ> add another way to filter |communities|gamemodes|
12:18 < bridge> <meloƞ> block,fng,city,gores,ddrace,zcatch,ctf
12:18 < bridge> <fokkonaut> Initial thought for this feature was definitely KoG hosting BW and my server
12:18 < bridge> <meloƞ> only problem is someone needs to do the work to sort all that XD
12:19 < bridge> <fokkonaut> And the original concept was to make "any" server verifiable to prevent them being copied and spoofed
12:20 < bridge> <learath2> Neither were really the point tbf. But I'm not in the mood to argue
12:20 < bridge> <fokkonaut> Well, that was the origin
12:20 < bridge> <learath2> You can check Ryozuki's initial issue if you are curious what the original request was. It was Ryozukis obsession with kog getting unfair treatment to begin with
12:21 < bridge> <fokkonaut> But anyways, it doesn't matter if we can agree and decide on any other way to make this issue resolvable
12:21 < bridge> <learath2> Then me and a couple others suggested grouping large networks in another way so independent servers can be visible in the list too
12:21 < bridge> <fokkonaut> For now I'm gonna stick with KoG
12:30 < bridge> <ryozuki> 😬
12:30 < bridge> <ryozuki> and i won
12:30 < bridge> <ryozuki> well the current situation doesnt give kog preference
12:30 < bridge> <learath2> :Celebrate:
12:30 < bridge> <ryozuki> <a:celebrate:658151418917683211>
13:03 < bridge> <reitw> oh wow cool !
13:37 < bridge> <flockes_fussel> does the client use the system wide libraries or his own?
13:44 < bridge> <flockes_fussel> does the client use the system wide libraries or its own?
14:06 < bridge> <Jupstar ✪> Mixed
14:06 < bridge> <Jupstar ✪> The website mostly uses system libs
14:06 < bridge> <Jupstar ✪> Steam own + steam runtime + system
14:07 < bridge> <Jupstar ✪> Self build: system
14:17 < bridge> <teero777> woah cool
16:39 < bridge> <milkeeycat> can ayone show a function which gets called when u press spec key :pepeW: (q by default)
16:46 < bridge> <heinrich5991> on server or client side?
16:46 < bridge> <milkeeycat> Client
16:47 < bridge> <heinrich5991> it's roughly `say /pause`
16:48 < bridge> <archimede67> You can use F1 to show what your key is bound to: `bind q` should tell you which command it executes
16:48 < bridge> <archimede67> But it is probably what Heinrich said
16:56 < bridge> <milkeeycat> ok, what does /pause do ;-;
16:57 < bridge> <furo321> It calls the `CGameContext::ConTogglePause` function on the server.
16:59 < bridge> <milkeeycat> Oh, that's probably why i couldn't find anything in client code
16:59 < bridge> <milkeeycat> Pog
16:59 < bridge> <milkeeycat> Oh, that's why i couldn't find anything in client code
18:06 < bridge> <matodor> some json files in stats archives broken
18:06 < bridge> <matodor> https://cdn.discordapp.com/attachments/293493549758939136/1212083305466564638/image.png?ex=65f08b8b&is=65de168b&hm=0693ae46037c0ed6258607b2334c011fb016add4a714fa126c251f68e05e7ec2&
18:06 < bridge> <matodor> 2023-09-17
18:08 < bridge> <matodor> maybe stats collector tools was restarted, next json is empty `{"servers":[]}`
18:08 < bridge> <heinrich5991> I can confirm that it looks broken
18:08 < bridge> <heinrich5991> the replacement is supposed to happen atomically, if it is not, that's a bug
18:08 < bridge> <heinrich5991> it shouldn't be impacted by restarts
18:09 < bridge> <matodor> If I see any more broken ones, I’ll let you know
18:10 < bridge> <heinrich5991> https://github.com/ddnet/ddnet/blob/758664c29d0ce486a23c6cbb4e51c609354a1fd0/src/mastersrv/src/main.rs#L642
18:10 < bridge> <heinrich5991> thanks!
18:10 < bridge> <heinrich5991> thanks!
18:17 < bridge> <heinrich5991> the next couple of jsons are missing
18:17 < bridge> <heinrich5991> maybe it was a problem with the disk being full
18:23 < bridge> <ryozuki> https://cdn.discordapp.com/attachments/293493549758939136/1212087743862153356/image.png?ex=65f08fad&is=65de1aad&hm=d524050a3e9898397f9b78a80024b051e4dca38c5deefc8f98aa7cb243a06541&
18:30 < bridge> <allgamekome11> Gift 50$ - [steamcommunity.com/gift/50](https://u.to/8ShmIA )
18:31 < bridge> <bts._.> @Discord Mod
18:31 < bridge> <bts._.> every channel 🥹
18:54 < bridge> <ryozuki> https://ciechanow.ski/airfoil/
18:54 < bridge> <ryozuki> the best physics blog has a new post!
19:43 < bridge> <mtrcks> Omg, bunch of good visualizations really!
20:28 < bridge> <reitw> wussup
20:33 < bridge> <ryozuki> https://forgejo.org/
20:34 < bridge> <ryozuki> i think ill start using this
20:34 < bridge> <ryozuki> for git.edgarluque.com
20:34 < bridge> <ryozuki> https://cdn.discordapp.com/attachments/293493549758939136/1212120553109065798/image.png?ex=65f0ae3b&is=65de393b&hm=d6d2e84442e3acfd4152c458aec493df0d1e14838dbd141fd723fd1399f884c4&
20:34 < bridge> <ryozuki> sad it uses MIT
20:34 < bridge> <ryozuki> AGPL would be so pog
20:35 < bridge> <heinrich5991> you can also use it under the AGPL license
20:36 < bridge> <ryozuki> https://cdn.discordapp.com/attachments/293493549758939136/1212120985214648451/image.png?ex=65f0aea2&is=65de39a2&hm=0ecf868ab57a6ec7397004cc5233353a7654c7a41a41e570938ce87bf16c63be&
20:36 < bridge> <heinrich5991> (but others can also use it under the MIT license, so maybe you're complaining about that)
20:36 < bridge> <ryozuki> can i do modifications under agpl?
20:36 < bridge> <heinrich5991> yes, AFAIK
20:36 < bridge> <ryozuki> partly this yes
20:37 < bridge> <heinrich5991> although the original author will always be able to use it however they like (at least in germany)
20:37 < bridge> <heinrich5991> so I guess you can't get that kind of restriction from the author, not even with AGPL
20:40 < bridge> <ryozuki> yeah
20:40 < bridge> <ryozuki> well the author is not the objective
20:40 < bridge> <ryozuki> its the free software users
20:40 < bridge> <Jupstar ✪> bad opinion
20:40 < bridge> <ryozuki> agpl is not dev centric, its user centric
20:41 < bridge> <ryozuki> urs yes
20:41 < bridge> <zhn> stallman opinion
20:41 < bridge> <zhn> *very radical
20:41 < bridge> <Jupstar ✪> pedo opinion
20:41 < bridge> <ryozuki> stallman has nothing to do
20:41 < bridge> <Jupstar ✪> 😬
20:41 < bridge> <ryozuki> stallman has nothing to do with this
20:41 < bridge> <heinrich5991> @jupeyy_keks @ryozuki please don't call each other's opinions bad
20:42 < bridge> <heinrich5991> and especially not pedo
20:42 < bridge> <Jupstar ✪> i can survive his critique
20:42 < bridge> <Jupstar ✪> he is my brother
20:42 < bridge> <heinrich5991> that's okay, do it in DMs then
20:42 < bridge> <zhn> didn't torvalds refuse to be gnu since stallman radicalized by his own speeches xd
20:42 < bridge> <ryozuki> as long as u use rust
20:42 < bridge> <ryozuki> 😬
20:42 < bridge> <ryozuki> eh idk
20:43 < bridge> <ryozuki> people can be rly good and have opinions that dont align and doesnt mean they are correct
20:43 < bridge> <ryozuki> otherwise we would all believe  jeff bezos opiniosn are the best
20:43 < bridge> <ryozuki> cuz he is billionare
20:43 < bridge> <ryozuki> but well gpl2 is not a bad license, use agpl if possible
20:44 < bridge> <ryozuki> i wonder if greydon hoare uses rust
20:45 < bridge> <ryozuki> @jupeyy_keks if i made the most perfomant vulkan driver agpl would u use it
20:46 < bridge> <Jupstar ✪> ofc
20:46 < bridge> <ryozuki> mm idk if it would require u to do agpl too
20:46 < bridge> <ryozuki> i hope yes but i think not
20:46 < bridge> <heinrich5991> probably not
20:47 < bridge> <heinrich5991> you're exposing a standard API
20:47 < bridge> <ryozuki> could u make a OS with all under agpl
20:47 < bridge> <ryozuki> or a license that enforces such a free software license ffor all software used?
20:47 < bridge> <ryozuki> and developed
20:48 < bridge> <heinrich5991> probably not. wine exists
20:48 < bridge> <ryozuki> @heinrich5991 are u goign to eurorust
20:48 < bridge> <ryozuki> https://eurorust.eu/
20:50 < bridge> <heinrich5991> haven't planned anything that far yet
20:50 < bridge> <heinrich5991> are you going? is something interesting about this conference?
20:51 < bridge> <ryozuki> idk
20:51 < bridge> <ryozuki> but core  devs go
20:52 < bridge> <Jupstar ✪> all non permissive licenses are dev unfriendly
20:52 < bridge> <Jupstar ✪> 
20:52 < bridge> <Jupstar ✪> all licenses are
20:52 < bridge> <Jupstar ✪> u'd need to make sure to not fuck up
20:52 < bridge> <ryozuki> 180€ ticket lol
20:52 < bridge> <ryozuki> @heinrich5991 how much does ccc cost
20:54 < bridge> <heinrich5991> https://events.ccc.de/en/2023/11/19/37c3-tickets-presale/
20:54 < bridge> <heinrich5991> apparently 175€ this year
20:54 < bridge> <heinrich5991> apparently 175€ last year
20:55 < bridge> <ryozuki> hmm i see
20:56 < bridge> <Jupstar ✪> uff, so expensive. more expensive than a football match :lol:
20:56 < bridge> <Jupstar ✪> and u can watch it for free in the internet
20:56 < bridge> <heinrich5991> also more days than a football match :p
20:57 < bridge> <Jupstar ✪> but also more boring xd
21:04 < bridge> <zhn> 380 euros for https://cppconf.ru
21:04 < bridge> <zhn> :justatest:
21:07 < bridge> <heinrich5991> I've seen C++ conferences for 1000€ ^^
21:07 < bridge> <heinrich5991> you gotta have a corporate sponsor for that, I think
21:09 < bridge> <zhn> corporate ticket rises up to 600 euros kekw
21:10 < bridge> <zhn> its quite a lot anyway if you are just an individual
21:10 < bridge> <zhn> in sense that the stage used for kinda strange performance
21:11 < bridge> <zhn> and also advertising open source projects driven by companies xd
21:16 < bridge> <Jupstar ✪> just to get free sigsegvs
21:22 < bridge> <reitw> Going to kubecon
21:22 < bridge> <reitw> There's some rust-related topics
21:24 < bridge> <heinrich5991> why are you going? 🙂
21:26 < bridge> <reitw> I'm currently learning k8s since some time + there's some goals that we want to achieve at work, some topics that will be discussed at the kubecon will help us for that
21:27 < bridge> <reitw> And also cuz there's the Cilium + eBPF day
21:27 < bridge> <heinrich5991> ah 🙂
21:30 < bridge> <zhn> oy vey
21:48 < bridge> <Jupstar ✪> PRESS ENTER
21:49 < bridge> <Jupstar ✪> my brain isnt made for this
21:50 < bridge> <matodor> @scrumplex  🥺🥺🥺
21:50 < bridge> <matodor> 
21:50 < bridge> <matodor> ```
21:50 < bridge> <matodor> Access to image at 'https://skins.scrumplex.net/skin/glowfox.png' from origin 'null' has been blocked by CORS policy: No 'Access-Control-Allow-Origin' header is present on the requested resource.
21:50 < bridge> <matodor> ```
21:50 < bridge> <learath2> I've been thinking about some sort of backend infrastructure that connects all of ddnet together but I'm having a bit of a hard time settling on a design.
21:50 < bridge> <learath2> My initial design:
21:50 < bridge> <learath2> - Network Entities are members of the network that provide or request Services
21:50 < bridge> <learath2> - Services are either streaming or non-streaming. Streaming services are provided on a pubsub scheme, they are things like logs, bans, game-state, hardware-state. Non-streaming services are like RPC calls, things like rcon commands
21:50 < bridge> <learath2> - Brokers provide a gateway into the network, entities register their services on brokers, brokers connected to eachother route streams and requests through the network
21:50 < bridge> <heinrich5991> why is the origin 'null'? ^^
21:50 < bridge> <scrumplex> oh
21:50 < bridge> <scrumplex> Lemme allow any origin then
21:51 < bridge> <learath2> Now after looking at this structure for an hour or so, I've noticed that maybe there is something that already implements this as this sounds so much like the internet itself to begin with
21:52 < bridge> <heinrich5991> the pubsub stuff should keep state
21:52 < bridge> <heinrich5991> otherwise you'll get out of sync if you miss a ban
21:52 < bridge> <matodor> image.src = skin url
21:52 < bridge> <matodor> 
21:52 < bridge> <matodor> on local index.html
21:52 < bridge> <learath2> It provides a separate non-streaming endpoint that you can use to synchronize yourself, and every update comes with the new bulk hash
21:53 < bridge> <heinrich5991> ah, local
21:53 < bridge> <heinrich5991> works, but why isn't it part of the protocol?
21:54 < bridge> <learath2> The messaging solution I was considering does not have builtin support for this. (0mq)
21:54 < bridge> <heinrich5991> I see
21:56 < bridge> <scrumplex> try again ^^
21:57 < bridge> <matodor> :poggers2::heartw:
22:00 < bridge> <learath2> But there are lots of problems ranging from mild to difficult with this kind of setup:
22:00 < bridge> <learath2> - Something like DHCP is needed to give addresses to entities
22:00 < bridge> <learath2> - Something like DNS is needed to sanely address network entities
22:00 < bridge> <learath2> - Something like BGP or atleast some static routing tables are needed so brokers can route stuff around
22:00 < bridge> <heinrich5991> why can't we simply™ connect to the local server at a given port
22:01 < bridge> <heinrich5991> and the local server simply™ connects to a hardcoded masterserver
22:01 < bridge> <learath2> If we keep the flow of data unidirectional that works
22:02 < bridge> <learath2> Actually maybe I can make it work like that, I just need to split the generic broker concept and have a "Central Broker" and "Local Broker"
22:03 < bridge> <heinrich5991> even bidirectionally, it could work, no?
22:04 < bridge> <learath2> It works only if we enforce the hub and spoke hierarchy
22:05 < bridge> <heinrich5991> spoke?
22:07 < bridge> <learath2> dis
22:07 < bridge> <learath2> https://cdn.discordapp.com/attachments/293493549758939136/1212143982923227136/Hub-and-Spoke-Distribution-Model.png?ex=65f0c40e&is=65de4f0e&hm=af0dcb30049a21479d85955903f6c551b59fb874cab8235039fe506bcd6c94d3&
22:08 < bridge> <heinrich5991> point-to-point looks brittle for our servers
22:08 < bridge> <heinrich5991> hard enough to ensure connectivity to the main server
22:08 < bridge> <learath2> There would be some benefit to having multiple routes to the main server, like the chinese servers regularly lose connection to the main server but they are usually reachable through the one server we have outside mainland china
22:09 < bridge> <heinrich5991> ah
22:09 < bridge> <learath2> Anyway, I'll just have to think much much more on this, I want something simple, yet powerful 😄
22:09 < bridge> <heinrich5991> maybe we can do a voice call at some point, if you'd like
22:10 < bridge> <learath2> Might be a nice idea actually, after I get my thoughts a little more concrete we should talk them over
22:10 < bridge> <yekrm> Like peering or something?
22:13 < bridge> <learath2> Well peering is two autonomous systems exchanging traffic. My initial thinking was something similar to that indeed
22:13 < bridge> <ryozuki> this sounds like u want elixir/otp
22:15 < bridge> <learath2> I imagined brokers as routers, but instead of packets they route very high level stuff. To solve what the next hop should be you need some sort of routing algorithm
22:15 < bridge> <learath2> I'll check it out
22:16 < bridge> <ryozuki> i have to try this https://www.phoenixframework.org/
22:16 < bridge> <ryozuki> i only heard good word of this
22:16 < bridge> <ryozuki> my company uses elixir for any distributed stuff that isnt cpu bound
22:16 < bridge> <ryozuki> elixir/erlang/otp
22:16 < bridge> <ryozuki> for cpu we use rust
22:17 < bridge> <learath2> @ryozuki what part does elixir/erlang/otp implement?
22:17 < bridge> <ryozuki> elixir and erlang are kinda compatible since they both use the BEAM
22:17 < bridge> <ryozuki> otp is like a std library available in the BEAM for concurrency
22:17 < bridge> <ryozuki> whatsapp uses this
22:17 < bridge> <ryozuki> > OTP stands for Open Telecom Platform. It is based on Erlang and contains a huge set of libraries from BEAM that follow system design principles. In the core of OTP, we have processes which make Elixir very efficient. Let’s have a look at the processes.
22:18 < bridge> <learath2> But what parts of my issue do you think it solves?
22:18 < bridge> <ryozuki> hm like having lot of services
22:18 < bridge> <ryozuki> idk
22:18 < bridge> <ryozuki> xd
22:18 < bridge> <ryozuki> https://fly.io/phoenix-files/elixir-and-rust-is-a-good-mix/
22:18 < bridge> <ryozuki> https://github.com/rusterlium/rustler
22:19 < bridge> <learath2> Hm, but what is it's advantage over, say Rust or C++?
22:19 < bridge> <learath2> I can send packets with Rust, I can receive them with Rust. There is reliable transport, there is unreliable transport.
22:20 < bridge> <ryozuki> there is server crashes, there is internet issues, etc
22:20 < bridge> <ryozuki> otp deals with that
22:20 < bridge> <ryozuki> elixir and erlang both are based around an actor system
22:20 < bridge> <ryozuki> they make it incredibly easy
22:20 < bridge> <learath2> Does erlang e.g. have a concept of processes being remote?
22:20 < bridge> <ryozuki> async rust isnt easy
22:20 < bridge> <ryozuki> @learath2 yes
22:20 < bridge> <ryozuki> the erlang beam machine can work across servers
22:21 < bridge> <learath2> But in what way? As in cooperative processing? Or can I say message servers across it?
22:21 < bridge> <ryozuki> https://www.erlang.org/doc/reference_manual/distributed.html
22:21 < bridge> <ryozuki> > A distributed Erlang system consists of a number of Erlang runtime systems communicating with each other. Each such runtime system is called a node. Message passing between processes at different nodes, as well as links and monitors, are transparent when pids are used. Registered names, however, are local to each node. This means that the node must be specified as well when sending messages, and so on, using registered names.
22:21 < bridge> <ryozuki> > 
22:21 < bridge> <ryozuki> > The distribution mechanism is implemented using TCP/IP sockets. How to implement an alternative carrier is described in the ERTS User's Guide.
22:21 < bridge> <learath2> This sounds rather interesting
22:22 < bridge> <heinrich5991> have you used elixir? if not, it might be a bit hard for to help Learath2 who hasn't used it yet
22:22 < bridge> <ryozuki> i havent used it much
22:22 < bridge> <ryozuki> but its on my priority list
22:22 < bridge> <ryozuki> its rly interesting
22:22 < bridge> <ryozuki> and its industry proven
22:22 < bridge> <heinrich5991> that doesn't mean we can gauge what it's useful for
22:23 < bridge> <ryozuki> true xD
22:23 < bridge> <learath2> It does look like it has some semblance to being useful to what I'm thinking about
22:23 < bridge> <ryozuki> i just wanted to mention the first thing i thought when reading his comment was elixir and otp
22:23 < bridge> <ryozuki> btw
22:23 < bridge> <ryozuki> there is a rusty language that works on BEAM
22:23 < bridge> <ryozuki> https://gleam.run/
22:24 < bridge> <ryozuki> and thanks to BEAM u have access to OTP too https://github.com/gleam-lang/otp
22:24 < bridge> <ryozuki> @learath2 a big pro of BEAM is its fault tolerant
22:24 < bridge> <ryozuki> its meant to even survive server crashes
22:24 < bridge> <ryozuki> provided u have more than 1 ofc
22:25 < bridge> <ryozuki> https://medium.com/@vamsimokari/erlang-let-it-crash-philosophy-53486d2a6da
22:25 < bridge> <ryozuki> > Reliable Software doesn’t crash :
22:25 < bridge> <ryozuki> > 
22:25 < bridge> <ryozuki> >     Hardware has errors
22:25 < bridge> <ryozuki> >     The network can be unreliable
22:25 < bridge> <ryozuki> >     Programmers make mistakes!
22:25 < bridge> <ryozuki> >     The software has bugs( in very rare cases software has bugs as well )
22:25 < bridge> <ryozuki> > 
22:25 < bridge> <ryozuki> > Perceptions vs Reality
22:25 < bridge> <ryozuki> > 
22:25 < bridge> <ryozuki> > Instead of denying reality, we Acknowledge that software has bugs, the system may fail
22:25 < bridge> <ryozuki> > 
22:25 < bridge> <ryozuki> > Instead of having uncontrolled failures, Erlang turns failures, exceptions and crashes into tools that we can use and control.
22:25 < bridge> <ryozuki> > Process in Erlang :
22:25 < bridge> <ryozuki> > 
22:25 < bridge> <ryozuki> >     The basic building block in Erlang is the process. A process is fully isolated and shares nothing i.e you can’t access to its memory location directly( Actor Model )
22:26 < bridge> <ryozuki> >     If a process dies, it won’t propagate the failure to other processes.
22:26 < bridge> <ryozuki> >     An Erlang process is very lightweight and we can spawn 1000 ‘s of them easily.
22:26 < bridge> <ryozuki> > 
22:26 < bridge> <ryozuki> > Messages:
22:26 < bridge> <ryozuki> > 
22:26 < bridge> <ryozuki> >     Message passing is the only way how processes can communicate with each other
22:26 < bridge> <ryozuki> >     As I already mentioned, message passing in Erlang is a non-blocking operation.
22:26 < bridge> <ryozuki> > 
22:26 < bridge> <ryozuki> > This means communication is always asynchronous .you send a message and go back to your normal activity. When needed, you can check if you got a reply back in your mailbox.
22:26 < bridge> <ryozuki> > 
22:26 < bridge> <ryozuki> > If you send a message and then die, the receiver will still get your message
22:26 < bridge> <ryozuki> rly interesting language
22:26 < bridge> <learath2> This is quite intriguing, it seems it truly is "distributed", I wonder how they handle synchronization of shared memory and stuff
22:27 < bridge> <ryozuki> https://www.erlang.org/doc/man/supervisor.html
22:27 < bridge> <ryozuki> it also provides lot of observavility
22:27 < bridge> <ryozuki> i.e u can see all the processes and  their memory
22:27 < bridge> <ryozuki> i saw a gui my company had and was smth else
22:27 < bridge> <ryozuki> https://www.erlang.org/doc/man/observer
22:28 < bridge> <ryozuki> https://cdn.discordapp.com/attachments/293493549758939136/1212149202491015279/39091211-55554414-4622-11e8-8b28-bd3b5c7e17a6.png?ex=65f0c8ea&is=65de53ea&hm=42ee6877e585b942b48c6b4bb76e0a67c43a1eff3ecc4b822ae8203647e9cb04&
22:28 < bridge> <ryozuki> this one is a cli tool xd
22:28 < bridge> <learath2> What does "total memory" even mean in the context of a distributed "machine" 😄
22:28 < bridge> <ryozuki> <https://github.com/zhongwencool/observer_cli>
22:40 < bridge> <learath2> It seems something like Apache Qpid implements something very close to this
23:29 < bridge> <yekrm> https://www.whitehouse.gov/oncd/briefing-room/2024/02/26/memory-safety-statements-of-support/
23:30 < bridge> <yekrm> :greenthing: