<<Prev Next>> Scroll to Bottom
Stuff goes here
L1[00:36:33] <CompanionCube> Ariri: someone high-up in the english courts taking the blockchain hype at face value
L2[00:36:35] <Amanda> %pet Michiyo
L3[00:36:35] <MichiBot> Amanda is patting Michiyo with amazing, multiple-use grenades. Michiyo regains 1d4 => 3 hit points! Amazing, multiple-use grenades ruptures and deflates.
L4[00:36:59] <Amanda> D:
L5[00:37:26] * Amanda quickly packs her things before her grenade scam gets any more exposed
L6[01:18:48] <dequbed> %pet Ariri
L7[01:18:48] <MichiBot> dequbed is petting Ariri with a slightly-out-of-date nixos 15.03 install cd. Ariri regains 1d6 => 1 hit points! The slightly-out-of-date nixos 15.03 install cd was taken out by the mafia.
L8[01:28:53] <sapphicf​ettucine> the mafia having an interest in nixos is hillarious to me
L9[01:29:16] <sapphicf​ettucine> ay, tony, ya ever heard about reproducible environments?
L10[01:38:13] <dequbed> I mean organized crime of today has to have proper digital infrastructure too
L11[01:42:40] <sapphicf​ettucine> the equivalent of al capone being gotten on tax fraud is gonna be a mob boss being found out because of k8s vulnerabilities
L12[02:11:29] ⇦ Quits: jackie (~jackie@banana-new.kilobyte22.de) (Ping timeout: 192 seconds)
L13[02:12:03] ⇦ Quits: dequbed (~steak@banana-new.kilobyte22.de) (Ping timeout: 192 seconds)
L14[02:12:56] <Amanda> Or someone committing an AWS key in their dotfiles repo
L15[02:15:18] ⇨ Joins: dequbed (~steak@banana-new.kilobyte22.de)
L16[02:15:21] ⇨ Joins: jackie (~jackie@banana-new.kilobyte22.de)
L17[02:15:21] zsh sets mode: +v on jackie
L18[02:18:57] ⇦ Quits: ThePiGuy24 (~ThePiGuy2@host-92-17-127-204.as13285.net) (Read error: Connection reset by peer)
L19[02:19:05] ⇨ Joins: ThePiGuy24 (~ThePiGuy2@host-92-17-127-204.as13285.net)
L20[02:28:43] * Amanda cuddles up around elfi, lays her head on her, updates Elfi's graphics drivers while she zzmews
L21[02:28:43] <Amanda> Night nerds
L22[04:17:07] <Forec​aster> %sip
L23[04:17:07] <MichiBot> You drink a dull aluminium potion (New!). Forecaster reboots for an update for 5 minutes.
L24[04:20:32] ⇨ Joins: ben_mkiv (~ben_mkiv@2001:16b8:1e79:e500:fe34:97ff:fea9:75f2)
L25[04:20:39] ⇦ Quits: ben_mkiv (~ben_mkiv@2001:16b8:1e79:e500:fe34:97ff:fea9:75f2) (Remote host closed the connection)
L26[04:27:39] <CompanionCube> sapphicfettucine: like how a misconfiguration caught the silk road guy
L27[04:51:10] ⇨ Joins: Hawk777 (~chead@2607:c000:827b:3300:b289:e3a:7fd2:391d)
L28[06:02:25] <Va​ur> %tonk
L29[06:02:25] <MichiBot> Wut! Va​ur! You beat Mic​hiyo's previous record of 4 hours, 35 minutes and 40 seconds (By 1 hour, 32 minutes and 22 seconds)! I hope you're happy!
L30[06:02:26] <MichiBot> Vaur's new record is 6 hours, 8 minutes and 3 seconds! Vaur also gained 0.0077 (0.00154 x 5) tonk points for stealing the tonk. Position #1.
L31[06:11:06] <CompanionCube> Izaya: it's snowing in march
L32[07:06:04] <f03​n1x> has anyone used the ocdevices addon for opencomputers? I made a tier 3 upgrade but I don't know how to actually apply it to the case :/
L33[07:12:19] <f03​n1x> ah found out how to use it, for anyone that wants to know you use it to upgrade the ocdevices cases
L34[07:12:33] <Ar​iri> CompanionCube: Yes but w h y
L35[07:13:04] <Ar​iri> "haha nft go brr"
L36[07:16:31] <CompanionCube> well, that's a why, if not a good ome
L37[07:23:04] ⇨ Joins: ben_mkiv (~ben_mkiv@2001:16b8:1e79:e500:fe34:97ff:fea9:75f2)
L38[07:34:06] ⇦ Quits: ben_mkiv (~ben_mkiv@2001:16b8:1e79:e500:fe34:97ff:fea9:75f2) (Quit: Leaving)
L39[07:38:26] ⇦ Quits: DBotThePony (~Thunderbi@31.220.170.28) (Remote host closed the connection)
L40[07:38:30] ⇨ Joins: DBotThePony (~Thunderbi@31.220.170.28)
L41[07:38:56] ⇦ Quits: Hawk777 (~chead@2607:c000:827b:3300:b289:e3a:7fd2:391d) (Quit: Leaving.)
L42[07:40:40] ⇨ Joins: Vexatos (~Vexatos@port-92-192-6-121.dynamic.as20676.net)
L43[07:40:40] zsh sets mode: +v on Vexatos
L44[09:00:44] <For​est> Heya, what languages does OC2 support? Only Lua or?
L45[09:01:11] <B​ob> its a RISC-V VM
L46[09:01:21] <B​ob> anything that compiles to RISC-V machine code
L47[09:01:24] <B​ob> ie anything mostly
L48[09:01:30] <For​est> Ah okay then
L49[09:01:39] <B​ob> we got Rust, C++, Zig, all the cool stuff running there
L50[09:01:40] <B​ob> and libs
L51[09:01:57] <B​ob> the Linux kernel is all C and it runs perfectly
L52[09:01:59] <For​est> Is there an API that's defined for interacting with the world of MC?
L53[09:02:14] <For​est> >Bob: we got Rust, C++, Zig, all the cool stuff running there
L54[09:02:14] <For​est> So Python and Nim would work? 👀
L55[09:04:42] <B​ob> what are the reasons for them to not work ?
L56[09:04:52] <B​ob> besides python being too fat and slow ?
L57[09:04:58] <For​est> Idk i just find this really cool lol
L58[09:05:03] <B​ob> there's micropython in OC2 already
L59[09:05:22] <For​est> Is there an emulator for OC2? Like CraftOS-PC?
L60[09:10:14] <B​ob> you run KVM or QEmu, its RISC-V anyways
L61[09:12:45] <For​est> Ah okay then
L62[09:13:08] <For​est> >Forest: Is there an API that's defined for interacting with the world of MC?
L63[09:13:08] <For​est> hiw about this?
L64[09:13:11] <For​est> how about this? [Edited]
L65[09:13:58] <B​ob> the HLAPI, rest is Linux things and MMIO
L66[09:14:11] <For​est> Ah okay
L67[09:14:23] <B​ob> the HLAPI is just JSON RPC trough a `/dev` STTY port (`/dev/hvc0`)
L68[09:14:45] <For​est> ah okay then, thanks!
L69[09:15:12] <B​ob> we made Rust, C++ and Zig libraries already
L70[09:15:17] <For​est> Does OC2 include a C compiler by default? (last question i think)
L71[09:15:19] <B​ob> i made a heavely generic version
L72[09:15:30] <For​est> I'll need to make a Nim binding to it lol
L73[09:15:43] <B​ob> >Forest: Does OC2 include a C compiler by default? (last question i think)
L74[09:15:43] <B​ob> no, but it provides one on the github (its buildroot anyways) (i got mine from bootlin)
L75[09:15:55] <B​ob> can link my Rust template, has the toolchain link in case
L76[09:17:53] <dequbed> The emulated machine is reasonably close to what QEMU does but it has some additional hardware such as soft RTCs. Most of that isn't relevant for userspace code though.
L77[09:21:53] <For​est> thanks!
L78[09:23:20] <dequbed> And also; the VM image has no compiler, mainly because it'd be an exercise in frustration trying to compile with 2MB of RAM in this day and age. But as Bob has already explained you can get yourself a cross-compilation toolchain for linux-unknown-riscv64-musl and be good
L79[09:23:48] <dequbed> It's a bit of an odd target so expect manual fiddling.
L80[09:24:59] <dequbed> riscv64-unknown-linux-musl too, I am tired.
L81[09:26:03] <B​ob> yeah i'm tired af too
L82[09:54:27] ⇦ Quits: ThePiGuy24 (~ThePiGuy2@host-92-17-127-204.as13285.net) (Ping timeout: 192 seconds)
L83[10:25:21] <For​est> >dequbed: And also; the VM image has no compiler, mainly because it'd be an exercis…
L84[10:25:22] <For​est> Ah fair
L85[10:25:59] <For​est> Hm does anyone know if there's a project that allows WebAssembly to be ran easily? It's portable and many langs can be compiled to it
L86[10:26:18] <For​est> Like if someone has already wrapped the HLAPI in WASM
L87[10:28:02] <dequbed> Then you'd need to have a WA VM.
L88[10:28:36] <nan​odn> I believe the madman himself used tcc to compile a helloworld
L89[10:28:46] <nan​odn> let me find the twitter link
L90[10:28:54] <nan​odn> here: https://twitter.com/SangarWasTaken/status/1494706020517482496
L91[10:29:34] <For​est> >dequbed: Then you'd need to have a WA VM.
L92[10:29:35] <For​est> There's stuff like Wasmer and Wasmtime so?
L93[10:30:32] <dequbed> I'm not saying it's impossible. But OC2 is resource-constrained as it is. Adding another VM on top will not help your course. There's a reason we're not running the JVM, BEAM, et.al. but Lua and micropython.
L94[10:30:45] <PwnageP​ineapple> I'm not sure running either of those VMs is viable on OC2 computers
L95[10:31:29] <dequbed> Oh no running the BEAM is definitely viable. I'm not sure about the JVM, I know too little about it
L96[10:31:40] <For​est> Fair
L97[10:32:34] <nan​odn> java me exists, but it's not really open
L98[10:32:42] <nan​odn> sim cards use it
L99[10:33:10] <For​est> Fair
L100[10:38:06] <B​ob> who wakes up in the morning and in their right mind thinks that SIM cards need Java
L101[10:38:53] <dequbed> What else are they supposed to run?
L102[10:39:14] <For​est> A language meant specifically for sim cards?
L103[10:40:35] <dequbed> .... like Java ME for example?
L104[10:41:08] <For​est> ¯\\(ツ)\/¯
L105[10:45:59] ⇨ Joins: ben_mkiv (~ben_mkiv@2001:16b8:1ed6:3200:fe34:97ff:fea9:75f2)
L106[10:58:16] <PwnageP​ineapple> Just run Rust on your sim cards
L107[10:59:20] <dequbed> *sigh* No.
L108[11:09:36] <B​ob> ~~i fucking will~~
L109[11:10:27] <nan​odn> this exists: https://github.com/SquirrelJME/SquirrelJME
L110[11:10:50] <nan​odn> open source java me 8
L111[11:54:15] <Amanda> You've heard it here first, @Bob just volunteered to make a jvm target for rust
L112[11:55:04] <Amanda> We look forward to your future work, @Bob
L113[11:56:29] * Amanda wanders over to dequbed, meows half-sleepily about how you shouldn't use compile time evaluation to do a gender transition, the dependencies are annoying
L114[11:56:33] <B​ob> 😊 whenever i have time
L115[11:56:52] * dequbed pets Amanda
L116[11:57:44] * Amanda purrs softly
L117[12:10:55] <For​est> What is Evey-chan for :p
L118[12:12:02] <Forec​aster> %tonk
L119[12:12:02] <MichiBot> Awesome! Forec​aster! You beat Va​ur's previous record of 6 hours, 8 minutes and 3 seconds (By 1 minute and 34 seconds)! I hope you're happy!
L120[12:12:03] <MichiBot> Forecaster's new record is 6 hours, 9 minutes and 37 seconds! Forecaster also gained 0.00021 (0.00003 x 7) tonk points for stealing the tonk. Position #3. Need 0.01765432 more points to pass Mic​hiyo!
L121[12:12:18] <Forec​aster> >Forecaster: %tonk
L122[12:12:18] <Forec​aster> Removing spam mainly
L123[12:12:37] <Forec​aster> .. Not what I meant to reply to
L124[12:13:01] <Indrid​Cold96> Got a simple question, why did the robot model change from OC to OC2?
L125[12:13:15] <Forec​aster> Because it's not the same mod
L126[12:13:18] <dequbed> Because everything changed. Why should it *not* change?
L127[12:13:26] <Forec​aster> All the models changed
L128[12:13:43] <Indrid​Cold96> I know, just wondering, the older robot was more distinct.
L129[12:14:14] <Forec​aster> I haven't even seen the oc2 one
L130[12:14:19] <Indrid​Cold96> A minute
L131[12:15:23] <Indrid​Cold96> https://tinyurl.com/y9pkzpwd
L132[12:24:37] <Forec​aster> Oh, I've seen that picture, I didn't know that was the robot
L133[12:24:53] <Forec​aster> It does have eyes though
L134[12:25:07] <Forec​aster> It's cuter than the oc1 robot
L135[12:25:49] <Forec​aster> Elfi dunno if you've seen this picture before
L136[12:25:52] <Forec​aster> https://tinyurl.com/ydbb75t2
L137[12:27:34] <For​est> >Forecaster: Removing spam mainly
L138[12:27:34] <For​est> lol, but thanks!
L139[12:46:55] <PHO​BOSS> Hello, Is there a way to sortout multiple replies that arrive at the same time from different drones?
L140[12:46:55] <PHO​BOSS> I have a swarm of drones and I'v set all of them to reply whenever they hear a command that I broadcast. But I'm not sure how to recieve their replies that reach all at the same time.
L141[12:46:55] <PHO​BOSS> Is there a cache that I can iterate through? Are they stacked?
L142[12:50:32] <Indrid​Cold96> >Forecaster: It's cuter than the oc1 robot
L143[12:50:33] <Indrid​Cold96> Can't argue with that, but I liked the format of the old robot, gave it a distinct appearence, other than the eyes this robot looks lika driver input.
L144[12:50:38] <Kristo​pher38> You should get their network card UUID as part of the `modem_message` event
L145[12:50:54] <Kristo​pher38> Or you could include your own id as part of the reply
L146[12:51:18] <Forec​aster> Events stack and should be processed in the order received
L147[12:51:32] <Forec​aster> At least if you use an event listener
L148[12:51:34] <Kristo​pher38> Oh, the question is about sth different
L149[12:52:41] <Kristo​pher38> Yeah, replies basically get put in a queue so you should be able to read all of them by repeatedly calling event.pull, or your event listener will be called multiple times if you're using event.listen
L150[12:55:06] <PHO​BOSS> >Kristopher38: Yeah, replies basically get put in a queue so you should be able to read a…
L151[12:55:07] <PHO​BOSS> Thanks! Is there a limit tho? I've seen the tutorial video about the switches, does the receiver have its own cache limit too?
L152[12:57:40] ⇨ Joins: ThePiGuy24 (~ThePiGuy2@host-92-17-127-204.as13285.net)
L153[12:59:54] <Forec​aster> I'd expect there to be a limit
L154[13:43:07] <B​ob> in OC2 you gotta do polling 😭
L155[13:58:07] ⇦ Quits: ben_mkiv (~ben_mkiv@2001:16b8:1ed6:3200:fe34:97ff:fea9:75f2) (Ping timeout: 192 seconds)
L156[14:14:44] <Amanda> uhhhh
L157[14:14:46] <Amanda> what
L158[14:15:09] <Amanda> somehow my zig-export program generated a 7M core dump
L159[14:16:14] <Amanda> damnit, and busybox xz doesn't support compressing, only decompressing
L160[14:20:17] <dequbed> Nice!
L161[14:20:29] <dequbed> 7M *is* pretty hefty so good job ;)
L162[14:27:09] <B​ob> 😎
L163[14:27:22] <B​ob> i wanted to try zig
L164[14:28:02] <Amanda> dequbed: I think what happen is I used the wrong function, so it defaulted to using the page allocator for commandline parsing, and that allocates an entire page for each alloc call
L165[14:28:24] <dequbed> Which to be fair *is* a reasonable approach for C-esque alloc patterns
L166[14:29:11] <Amanda> most c-esq stuff then subdivides that page, though. argv of 2, so that's 8192 bytes allocated immediately
L167[14:29:50] <dequbed> Hmm isn't the Zig page allocator just libc malloc or am I remembering wrong?
L168[14:30:02] <Amanda> nope, it's the kernel alloc
L169[14:30:05] <dequbed> Aaaah
L170[14:30:07] <Amanda> there's std.heap.c_allocator
L171[14:30:16] <Amanda> wwhich uses libc
L172[14:30:31] <dequbed> Okay and I guess the page allocator doesn't then sensibly re-use partial pages for future allocs
L173[14:30:46] <Amanda> nope, you get 4k, and that's what you'll use, liek it or not
L174[14:30:52] <dequbed> Meh
L175[14:31:11] <Amanda> you can then subdivide it by wrapping that allocator in an arena or other sub-dividing allocator, but you don't get it automatically
L176[14:31:14] <dequbed> I mean implementing an allocation manager yourself is fun and all but not when I'm not getting paid for it.
L177[14:34:45] <Amanda> Ah, I guess not, in posix the allocator's unused, and optional: https://github.com/ziglang/zig/blob/master/lib/std/process.zig#L545-L554
L178[14:35:08] <Amanda> so I have no idea why it SIGSEGV'd and dumped 7M as a core
L179[14:35:22] <Elfi> Oooh pretty birbs c:
L180[14:35:22] <dequbed> Want me to have a look?
L181[14:36:49] <Amanda> Well, it's not happening again when I rerverted back to std.process.args()
L182[14:37:35] <Amanda> but now device iteration's SIGSEGV'ing for some reason
L183[14:40:24] <Amanda> dequbed: Havn't committed what I'm doing now, but I've not touched the hlapi code since, and that's what seems to be randomly spewing chunks: https://git.camnet.site/amandac/oc2-zig-fuckery/src/branch/main
L184[14:42:51] <Amanda> Huh. I tried compiling without linking against libc, and now std.os.tcgetattr is SIGSEGV'ing when defining the return variable?
L185[14:43:32] ⇦ Quits: ThePiGuy24 (~ThePiGuy2@host-92-17-127-204.as13285.net) (Remote host closed the connection)
L186[14:43:39] <dequbed> Amanda: Hmm, I'd look for any recursion first. That'd explain the gigantic coredump for once but also stack size in OC2's linux is either tiny or it's bigger than total available memory. Default is something like 8M IIRC
L187[14:43:48] <dequbed> s/once/one
L188[14:43:48] <MichiBot> <dequbed> Amanda: Hmm, I'd look for any recursion first. That'd explain the gigantic coredump for one but also stack size in OC2's linux is either tiny or it's bigger than total available memory. Default is something like 8M IIRC
L189[14:47:50] <dequbed> Amanda: Also the last thing doesn't suprise me. tcgetattr *does* come from the libc. I'm surprised you could even get that to produce an ELF without linking to a libc
L190[14:48:14] <Amanda> It's a kernel syscall isn't it?
L191[14:48:33] <dequbed> Yes but like almost all syscalls it's wrapped the libc to be less painful to use
L192[14:48:41] <dequbed> by the libc*
L193[14:49:12] <Cyborg​Potato> just call it directly anyways /j
L194[14:50:21] <Amanda> https://github.com/ziglang/zig/blob/master/lib/std/os/linux.zig#L1495-L1497
L195[14:50:48] <dequbed> ... interesting. I did *not* expect zig to implement that much themselves.
L196[14:56:47] <Amanda> Wait. Could the stack being tiny be why I keep getting SIGSEGV's in random fucking places constantly, that then go away just as randomly
L197[14:57:01] <dequbed> Well yes that can be an explanation
L198[14:57:29] <dequbed> I mean it's stack and you're writing Zig so if you give your program the exact same input I do expect it to crash at the exact same position every time.
L199[14:57:52] <Amanda> yeah, it's the same place until I change literally anything, then it jumps somewhere else
L200[14:58:10] <dequbed> But e.g. JSON parsing is probably done recursively to a degree in which case stack usage is very different for different inputs
L201[14:58:49] <dequbed> But it can be a very likely explanation. Zig is usually very heavy on the stack usage
L202[15:00:22] <dequbed> And it also reserves the right to do full pass-by-value which in the case of structures means the structure is put on the stack twice for that one call
L203[15:01:25] <dequbed> As in it's on the stackframe of the caller and copied to the stackframe of the callee as well for that call. And if that function then calls *more* functions… you get the gist
L204[15:02:08] ⇨ Joins: ThePiGuy24 (~ThePiGuy2@host-92-17-127-204.as13285.net)
L205[15:10:47] <dequbed> Amanda: I just checked. Zig apparently doesn't do specific stack overflow detection using e.g. canaries so a stack overflow would only be seen as an segmentation fault signal.
L206[15:14:34] <Amanda> ... goddess damned it all
L207[15:16:43] <Amanda> Damn.
L208[15:16:59] <Amanda> https://matrix.camnet.site/_matrix/media/r0/download/camnet.site/wrKQcPhNhynHjWUvJPMPiakY/screenshot-1648739798.png
L209[15:30:25] <B​ob> trolled
L210[15:42:17] <Forec​aster> The birds are super adorable
L211[15:49:41] <For​est> Anyone know if RustPython is small enough to be ran in OC2? And if it even runs lol
L212[15:50:10] <For​est> Can't test myself because currently i am walking home from college lol
L213[15:51:54] <B​ob> there's micropython ?
L214[15:54:10] <For​est> Yeah but does MicroPython have access to stuff like asyncio? And can it use other libs like aiohttp?
L215[15:55:34] <For​est> I wanna try seeing if i can use stuff like discord.py (or a fork of it) to make a stupid bot from MC lol
L216[15:56:13] <B​ob> OC2 doesnt have internet access yet
L217[15:56:20] <For​est> Oh really? Bummer
L218[15:56:29] <B​ob> a fork does but thats dequbed mad witchcraft at pkay
L219[15:56:38] <B​ob> a fork does but thats dequbed mad witchcraft at play [Edited]
L220[15:56:43] <For​est> ~~I'm interested~~
L221[15:56:46] <For​est> What fork is it lol
L222[15:57:00] <For​est> And 1.18.1 support? 👀
L223[15:57:11] <B​ob> i think it only runs on 1.18.1
L224[15:57:26] <PwnageP​ineapple> Are there any plans to bring internet access to OC2?
L225[15:57:27] <B​ob> or maybe im just way too sleep derived
L226[15:57:30] <For​est> >Bob: i think it only runs on 1.18.1
L227[15:57:31] <For​est> Nice
L228[15:57:36] <B​ob> >PwnagePineapple: Are there any plans to bring internet access to OC2?
L229[15:57:36] <B​ob> yeah, already a thing
L230[15:57:38] <B​ob> just not official
L231[15:57:41] <For​est> What's the fork name?
L232[15:57:49] <For​est> Like the user
L233[15:57:53] <B​ob> look in the PRs in OC2's github
L234[15:58:12] <For​est> Aight
L235[15:58:16] <PwnageP​ineapple> >Bob: just not official
L236[15:58:17] <PwnageP​ineapple> I meant officially or via an addon. As opposed to a fork of the base mod
L237[16:02:38] <dequbed> Yes if a sensible implementation is made Snagar isn't opposed to merging it
L238[16:02:55] <For​est> What's the issue with the one that was PRed?
L239[16:03:09] <dequbed> Which one?
L240[16:03:24] <For​est> The one Bob mentioned
L241[16:03:36] <dequbed> ...
L242[16:03:42] <PwnageP​ineapple> I think this one: https://github.com/fnuecke/oc2/pull/63
L243[16:03:42] * dequbed looks at her nickname
L244[16:03:48] * dequbed looks at Bobs message
L245[16:04:16] <For​est> Yeah but what's the issue with it? :p
L246[16:04:22] <dequbed> There isn't one.
L247[16:04:44] <For​est> Oh
L248[16:04:47] <dequbed> And the issue with !63 is that it's a bad approach
L249[16:05:33] <For​est> Welp, looks like it's time to make a very, very crappy CC <--> OC2 bridge if even possible
L250[16:06:37] <PwnageP​ineapple> What makes it a bad approach?
L251[16:06:53] <Forec​aster> not enough cover on the west flank
L252[16:09:06] <dequbed> Implementing your own network stack in userspace is not a manageable approach. It takes a lot of work to do and things like NAT aren't simple.
L253[16:12:52] <Z0id​berg> fuck nat
L254[16:14:26] <dequbed> Our approach (see https://twitter.com/SangarWasTaken/status/1487903774857388033?s=20&t=TXMtjqLKbQf5fi4p0tsNCA ) was to just plug OC2 into a VXLAN terminator and deal with networking using the proper tools. It was @gruetzkopf 's and Kilobyte 's idea, I just helped with the execution. Go ask them for gory details, I'm busy
L255[16:16:27] <For​est> How do i do that? Lol
L256[16:16:46] <B​ob> linux network stack, my beloved
L257[16:16:47] <For​est> (I'm fine with all the setup)
L258[16:16:53] <For​est> Linux ftw
L259[16:16:58] <Kilobyte> Asking me? I don't know how you'd do that
L260[16:17:13] <dequbed> @Forest have you set up a VXLAN like … ever?
L261[16:17:16] <For​est> Anyone who knows how xD
L262[16:17:28] <For​est> >dequbed: <@909883978717204561> have you set up a VXLAN like … ever?
L263[16:17:29] <For​est> Nope, but I'm willing to learn because i have nothing better to do lol
L264[16:17:38] <dequbed> Then go learn that :)
L265[16:17:58] <For​est> ~~Aight wish me luck~~
L266[16:18:28] <B​ob> cursed
L267[16:18:46] <For​est> https://vincent.bernat.ch/en/blog/2017-vxlan-linux hmm gonna try follow this but on my RPI
L268[16:18:50] <dequbed> VXLAN is very much the least goddamn cursed option for this. Go away @Bob
L269[16:19:04] <B​ob> still lol
L270[16:19:12] <Z0id​berg> Your next task
L271[16:19:13] <Z0id​berg> https://wiki.freebsd.org/riscv
L272[16:19:26] <For​est> What about it?
L273[16:19:41] <dequbed> No @Z0idberg stop that
L274[16:19:49] <Z0id​berg> :>
L275[16:19:53] <dequbed> No.
L276[16:19:56] <B​ob> the whole spec huh
L277[16:20:10] <Z0id​berg> If only I had time...
L278[16:20:18] <dequbed> Don't send the noobs on a fucking goose chase. We need more skilled technicians being assholes to them just drives them awayt.
L279[16:20:37] <Z0id​berg> Oh that was for Bob
L280[16:20:45] <For​est> Code Block pastebined https://paste.pc-logix.com/cobamacesathis looks too simple ._.
L281[16:20:50] <Z0id​berg> I was telling Bob to port FreeBSD to OC2
L282[16:20:57] <B​ob> yeah, sitting in a hospital rn so
L283[16:20:59] <B​ob> no way lol
L284[16:21:03] <For​est> >dequbed: Don't send the noobs on a fucking goose chase. We need more skilled techn…
L285[16:21:03] <For​est> I live on spite
L286[16:21:49] <dequbed> Cool have fun. If you ever externalize that spite on other beings in this community then I'll personally throw you out of the door. Be nice, will you?
L287[16:21:57] <For​est> >dequbed: Our approach (see https://twitter.com/SangarWasTaken/status/1487903774857���
L288[16:21:57] <For​est> What do i do after setting up VXLAN?
L289[16:22:28] <For​est> >dequbed: Cool have fun. If you ever externalize that spite on other beings in this…
L290[16:22:28] <For​est> Oh definitely not, i live on spite but i use it to fuel my desire to do very dumb things :D
L291[16:22:33] <Z0id​berg> Oooh VXLANs are nice
L292[16:22:40] <dequbed> I guess you get to compile our cursed OC2 fork with the correct settings and then you have networking in OC2
L293[16:23:00] <For​est> ~~What fork is it tho?~~
L294[16:23:04] <Z0id​berg> @Forest Rule #1 in #oc. DON'T LISTEN TO Zoidberg OR S3.
L295[16:23:12] <For​est> XD
L296[16:23:14] <dequbed> As in networking between an OC2 computer and the VXLAN terminator.
L297[16:23:19] <dequbed> The rest is up to you
L298[16:23:25] <Izaya> or me
L299[16:23:53] <Z0id​berg> Zoidberg/S3 does really terrifying things with technology that you do not want to try, and then works too much so they can never finish anything.
L300[16:23:54] <Kilobyte> If you are doing point-to-point only vxlan isn't too complex.
L301[16:24:01] <dequbed> If you don't know who to get to the internet from that terminator I suggest learning a few things about linux networking and more importantly routing first. It's not too compicated though
L302[16:24:14] <Kilobyte> If you have more than two nodes it can get very complicated
L303[16:24:21] <Izaya> S3: this week I wrote a distributed ARK server load balancer in bash
L304[16:24:29] <Z0id​berg> 😄
L305[16:24:36] <Izaya> real manmade horrors beyond my comprehension hours
L306[16:24:46] <Z0id​berg> @Enthalpy You got our VXLAN + Wireguard Mikrotik setup plan yet?
L307[16:24:57] <PwnageP​ineapple> >Izaya: S3: this week I wrote a distributed ARK server load balancer in bash
L308[16:24:57] <PwnageP​ineapple> Why tho?
L309[16:25:06] <Forec​aster> to balance loads
L310[16:25:13] <PwnageP​ineapple> But bash
L311[16:25:15] <B​ob> bash yikes
L312[16:25:18] <B​ob> my condolences
L313[16:25:18] <Izaya> because I don't have enough memory to run a server for every map
L314[16:25:33] <Z0id​berg> Hey. It could have been written in piet.
L315[16:25:40] <PwnageP​ineapple> Malbolge
L316[16:25:43] <Izaya> people talk shit about bash but for better or worse you can get a lot done very quickly in bash
L317[16:26:07] <dequbed> ^
L318[16:26:09] <dequbed> bash is fine
L319[16:26:16] <PwnageP​ineapple> For anything that needs a bash script, I'd rather just write it in Python
L320[16:26:20] <Z0id​berg> until you end up with that one syntatic POS of a variable that doesn't seem to work until you realized you accdiently typed #!/bin/sh instead of #!/bin/bash
L321[16:26:26] <Z0id​berg> so it's using some other shell instead
L322[16:26:41] <Izaya> also played with glusterfs
L323[16:26:43] <Izaya> it's pretty cool
L324[16:26:43] <Kilobyte> There is a bug in Linux with vxlan, that makes this a bit more annoying. Basically you can only bind to ::/0.0.0.0
L325[16:26:54] <Z0id​berg> lol wut
L326[16:27:04] <dequbed> right I wanted to fix that -.-
L327[16:27:19] <Kilobyte> This means you can't terminate the vxlan on the same kernel as the Minecraft server is running on
L328[16:27:35] <Kilobyte> Well - I guess a network namespace works too
L329[16:27:38] <Izaya> https://shadowkat.net/tmp/4b36.png
L330[16:27:56] <Kilobyte> Bash is really nice for many simple things
L331[16:28:06] <Kilobyte> At some point it gets painfull
L332[16:28:20] <Z0id​berg> Kilobyte: I guess in our case with a Mikrotik cloud hosted router VM we could terminate it there and bridge it to the Minecraft server interface
L333[16:28:31] <Z0id​berg> kinda cheating
L334[16:28:38] <dequbed> Nah we terminate VXLAN in hardware too
L335[16:28:40] <Kilobyte> Bash is the hot glue gun of the sysadmin
L336[16:28:50] <Z0id​berg> CHR is fucking fantastic but
L337[16:28:55] <Kilobyte> Cisco L3 Switch :P
L338[16:28:57] <Z0id​berg> it's terrifying using a VM for your router
L339[16:29:07] <Z0id​berg> it's amazing how performant it really is
L340[16:29:29] <Kilobyte> I've used many router OSes in VMs
L341[16:29:33] <Z0id​berg> we can force all of our VPSes to route through one Mikrotik VM on the hypervisor and have no IO issues. for some strange unexplainable reason
L342[16:29:50] <Kilobyte> RouterOS, VyOS, OpenWRT at least
L343[16:29:55] <Z0id​berg> yeah
L344[16:30:08] <Amanda> almost like virtualiasition has become mainstream enough for common I/O bottlenecks to have accleration options
L345[16:30:09] <Kilobyte> All of them with near native performance, thanks to virtio
L346[16:31:12] <Kilobyte> Though I am generally not so much a fan of mikrotik, probably mainly because I'm traumatised by CAPsMAN
L347[16:31:29] <Kilobyte> Their switches are great for hobby use though
L348[16:31:56] <Z0id​berg> I'm actually looking into setting up a new OVH hypervisor with FreeBSD + Xen, I always used Slackware + Xen. I guess we'll see
L349[16:32:13] <Z0id​berg> I just have other expenses atm
L350[16:32:31] <Kilobyte> FreeBSD as dom0?
L351[16:34:14] <Z0id​berg> Yep!
L352[16:34:19] <Z0id​berg> been in kernel since.. 11.x?
L353[16:34:34] <Z0id​berg> BHyve is cool too but
L354[16:35:00] <Kilobyte> Nice, I'm not sure much a BSD guy, but I like it when there are options
L355[16:36:24] <Z0id​berg> I use it on all of my virtual servers except for like, one of them. it can be nice for a small footprint server that uses little to no RAM at all
L356[16:37:01] <Z0id​berg> It's pretty much no different in terms of experience to Linux for most things really.
L357[16:38:33] <Z0id​berg> I think my mail server running freebsd 13 is using like.. < 40 MB of ram
L358[16:38:37] <For​est> >dequbed: I guess you get to compile our cursed OC2 fork with the correct settings …
L359[16:38:37] <For​est> I still don't know what fork I'm supposed to use :p
L360[16:40:32] <Z0id​berg> I think maybe they are referring to the one grue and all were working with?
L361[16:41:59] <For​est> Hm okay
L362[16:44:56] <Z0id​berg> Waiiiiit a minute. looks like FreeBSD-RISC-V works with Qemu risc64..
L363[16:45:05] <Z0id​berg> therefore it shouldn't be too hard to run on OC2 I would hope
L364[16:45:12] <Z0id​berg> (Maybe I will try to run it)
L365[16:46:37] <Z0id​berg> Looks like it works with gcc or llvm but defaults to LLVM for the world build
L366[16:47:12] <Z0id​berg> I thought somebody told me though that llvm was problematic with OC2 as a riscv target
L367[16:56:04] <Kilobyte> Problem with *bsd is the lack of many tool I've gotten used to, simply because of lack of compatibility
L368[16:56:13] <Kilobyte> Mainly iproute and systemd
L369[16:56:46] <Kilobyte> Though for small single purpose systems, systemd may not be the best option
L370[16:57:16] <Kilobyte> Currently having fun getting a buildroot to work with runit as init system
L371[16:58:47] <Z0id​berg> haha. the FreeBSD community and Slackware community have both basically pledged "never to adopt systemd". Though, Slackware does use elogind from that project and that only.
L372[16:59:03] <Izaya> systemd is pretty cool
L373[16:59:03] <Kilobyte> @Forest https://github.com/paranoidlabs/oc2, probly develop branch. Careful, there are hardcoded settings
L374[16:59:08] <dequbed> @Forest and I told you to ask @gruetzkopf
L375[16:59:12] <Izaya> it does dumb shit but it's really nice sometimes
L376[17:00:39] <Kilobyte> I love that it supervises everything by default. OpenRC can supervise, but it's generally not used by default from what I've seen
L377[17:00:53] <Kilobyte> runit also supervises everything:P
L378[17:01:05] <Izaya> timer and socket units I enjoy
L379[17:01:40] <Kilobyte> That too. And networkd is also pretty nice
L380[17:02:12] <Kilobyte> Then there are things like resolvd which are a PoS, but those are all optional
L381[17:02:36] <sapphicf​ettucine> my like of systemd is not a "it's good" kind of like it's a "i have to do less effort to get shit working" like
L382[17:02:39] <Izaya> I haven't played much with networkd
L383[17:03:01] <Izaya> probably should some time soon though
L384[17:03:22] <Kilobyte> Another thing I dislike: 90 second unit timeouts
L385[17:03:33] <Kilobyte> Or even infinity on shutdown
L386[17:03:59] <Izaya> Waiting for <whatever> [12378917]
L387[17:04:50] <Kilobyte> I mean on boot that's okay, if the service is required for the system
L388[17:05:06] <Kilobyte> But how am I supposed to reboot a broken system
L389[17:05:26] <Izaya> three finger salute x 7
L390[17:07:13] <For​est> >dequbed: <@909883978717204561> and I told you to ask <@451012380709879808>
L391[17:07:14] <For​est> Where?
L392[17:07:53] <For​est> >dequbed: Our approach (see https://twitter.com/SangarWasTaken/status/1487903774857���
L393[17:07:53] <For​est> I can't see any message like that unless you mean this lesson lol
L394[17:08:09] <For​est> >Kilobyte: <@909883978717204561> https://github.com/paranoidlabs/oc2, probly develop…
L395[17:08:09] <For​est> Ah fair
L396[17:16:39] ⇦ Quits: ThePiGuy24 (~ThePiGuy2@host-92-17-127-204.as13285.net) (Ping timeout: 186 seconds)
L397[17:17:01] <Z0id​berg> Don't forget though you also have Poettering
L398[17:17:03] <Z0id​berg> with systemd
L399[17:17:18] <Z0id​berg> and Poettering is awful.
L400[17:17:45] ⇨ Joins: ThePiGuy24 (~ThePiGuy2@host-92-17-127-204.as13285.net)
L401[17:26:41] <Kilobyte> I haven't dealt with that person yet, though I'm told otherwise by people who have interacted with him in person.
L402[17:28:49] <Forec​aster> %tonk
L403[17:28:49] <MichiBot> I'm sorry Forecaster, you were not able to beat Forecaster's record of 6 hours, 9 minutes and 37 seconds this time. 5 hours, 16 minutes and 47 seconds were wasted! Missed by 52 minutes and 50 seconds!
L404[17:28:58] <Forec​aster> oh, shoot
L405[17:29:48] <Z0id​berg> I don't necessarily think that tPoettering is a bad person
L406[17:56:53] <Enth​alpy> >Z0idberg: <@!123971392906461184> You got our VXLAN + Wireguard Mikrotik setup plan ye…
L407[17:56:54] <Enth​alpy> Kinda waiting on a long-term build for ROSv7 before updating the CHR.
L408[17:57:05] <Z0id​berg> heh
L409[17:57:19] <Z0id​berg> I have 3 mikrotiks in the house
L410[17:57:38] <Z0id​berg> particularly, the one that connects my DOS machine to wifi is a good one to use as a guinea pig
L411[17:57:57] <Z0id​berg> or the spare 100 Mbit one
L412[18:07:20] <Forec​aster> %sip
L413[18:07:20] <MichiBot> You drink an oxidised tomato potion (New!). Forecaster has a feeling that their face just appeared on a random vegetable somewhere.
L414[18:12:16] ⇨ Joins: ben_mkiv (~ben_mkiv@2001:16b8:1ed6:3200:fe34:97ff:fea9:75f2)
L415[18:15:00] <Va​ur> %sip
L416[18:15:00] <MichiBot> You drink a bubbly honey potion (New!). Vaur feels chill.
L417[18:58:54] ⇨ Joins: Overloader (webchat@111-136-067-156.ip-addr.inexio.net)
L418[18:59:17] <Overloader> so whatsn the difference between openos and plan9k?
L419[18:59:50] <Forec​aster> plan9k is third-party
L420[19:00:11] <Overloader> yeah, but advantages/disadvantages does it have over open os?
L421[19:00:28] <Forec​aster> I've never used it so no idea
L422[19:00:58] <PwnageP​ineapple> I don't know if plan9k is still maintained
L423[19:01:12] <Forec​aster> it's not I'm pretty sure
L424[19:01:24] <Forec​aster> but then neither is OpenOS 😛
L425[19:01:47] <PwnageP​ineapple> Fair
L426[19:02:08] <PwnageP​ineapple> I think plan9k has been unmaintained for longer tho
L427[19:02:12] <Forec​aster> yes
L428[19:02:24] <Kristo​pher38> its broken iirc
L429[19:05:33] <Forec​aster> that's probably an important difference
L430[19:08:07] <Ocawes​ome101> Plan9k is not a good experience these days.
L431[19:08:31] <Ocawes​ome101> Heck, I’d almost recommend ULOS 2 over it, and it’s not even remotely finished
L432[19:10:00] <Overloader> alright so I got openos now. at first glance it looks pretty similar to a linux distribution. can i treat it as one (concerning the file tree and basic comands) or is there an important difference?
L433[19:10:53] <Kristo​pher38> its only like this at first glance
L434[19:11:12] <Kristo​pher38> there's no multiuser support, file permissions, etc
L435[19:11:30] <Kristo​pher38> but most basic commands you can come up with are implemented
L436[19:28:30] <Ocawes​ome101> ^^ ULOS 2 is intended have these things and more. It’s still far from user ready yet though—I anticipate another couple months of work.
L437[19:39:19] ⇨ Joins: TPG24 (~ThePiGuy2@host-92-17-127-204.as13285.net)
L438[19:40:56] ⇦ Quits: ThePiGuy24 (~ThePiGuy2@host-92-17-127-204.as13285.net) (Ping timeout: 189 seconds)
L439[19:43:08] <Amanda> dequbed: why am I not getting a core dump. It also doesn't help that the address given just shows up as "zig-export:?" with add2line:
L440[19:43:19] <Amanda> https://matrix.camnet.site/_matrix/media/r0/download/camnet.site/SeTIvGCEQZbZvCXpgzsrNqSZ/screenshot-1648755761.png
L441[19:59:48] <dequbed> Amanda: sorry dear I had a terrible day and now the mill at my lab has broken in a very expensive way. I don't have the headspace for that right now, ask me Saturday. Or sometime next week.
L442[20:00:00] * Amanda nods
L443[20:00:03] <Amanda> fair enough
L444[20:00:09] * Amanda offers a dead mouse she "found"
L445[20:23:01] <Overloader> Ok, so, maybe im just too stupid to find it in the manual, but how do I get access to connected components. and how do I see what functions I can call on them?
L446[20:23:39] <Forec​aster> go into the lua repl (aka lua prompt) (aka the `lua` program)
L447[20:23:58] <Overloader> yeah, im already there
L448[20:24:00] <Forec​aster> type `component.<tab>` to cycle through connected components
L449[20:24:27] <Forec​aster> then when you have the component you want `component.modem.<tab>` cycles through that components methods and properties
L450[20:27:03] <Overloader> is there also a documentation of the methods that I can see? Like with pythons doc strings?
L451[20:50:01] <dequbed> Amanda: thanks I guess. I take a dead mouse over being soaked in three different flavours of oil and having to pick steel shavings out of my bra. <.<
L452[20:50:43] * Amanda beams all the steel shavings into a pile
L453[20:51:02] <Kristo​pher38> There is
L454[20:51:10] <Amanda> %give MichiBot a surprisingly large pile of steel shavings
L455[20:51:10] * MichiBot accepts the surprisingly large pile of steel shavings and adds it to her inventory
L456[20:51:22] <Kristo​pher38> If you just type method's name instead of calling it
L457[20:51:37] <Kristo​pher38> i.e. `component.whatever.foo`
L458[20:51:46] <Kristo​pher38> Also there's wiki
L459[20:51:50] <Kristo​pher38> %wiki
L460[20:51:51] <dequbed> Amanda: There's also a good amount of aluminium, stainless and a bit of plastic in my clothes if you want those too
L461[20:52:17] <Kristo​pher38> Well
L462[20:52:20] <Kristo​pher38> https://ocdoc.cil.li/
L463[20:52:38] * Amanda beams them into @Inari 's pantsu drawer
L464[21:39:43] <Overloader> I assume there isnt a vim like text editor?
L465[21:52:18] <Forec​aster> Not built in
L466[21:52:30] <Forec​aster> Edit is the only included editor
L467[22:09:32] <Amanda> Izaya made a modal editor I think
L468[22:09:45] <Amanda> or was it sam?
L469[22:23:57] * Amanda curls up around Elfi, read to her about magic and other things that are banned in this universe
L470[22:26:18] <gruet​zkopf> >Enthalpy: Kinda waiting on a long-term build for ROSv7 before updating the CHR.
L471[22:26:18] <gruet​zkopf> 7.1.5 has been fine for me
L472[22:26:54] <gruet​zkopf> i haven't yet terminated the vxlan tunnel on a mikrotik device
L473[22:27:48] <gruet​zkopf> (though i did get around to doing it on a celestica dx010 - on broadcom tomahawk1 100G switch ic)
L474[22:41:46] <Enth​alpy> I have one that I am testing with at work, and it seems to be fine. Just don't want anything that I would consider production running ROSv7 yet.
L475[22:43:11] <gruetzkopf> i'm running _all_ my prod on ROSv7
L476[23:25:58] ⇦ Quits: Vexatos (~Vexatos@port-92-192-6-121.dynamic.as20676.net) (Quit: Insert quantum chemistry joke here)
<<Prev Next>> Scroll to Top