<<Prev Next>> Scroll to Bottom
Stuff goes here
L1[00:48:48] <Lizzy> %tonk
L2[00:48:48] <MichiBot> Lizzy! You beat Kodos's previous record of 1 hour, 58 minutes and 6 seconds! I hope you're happy!
L3[00:48:49] <MichiBot> Lizzy's new record is 3 hours, 1 minute and 38 seconds
L4[00:50:33] <Kleadron> shit i forgot to do that
L5[00:51:19] <Lizzian> Mwahahahahahah
L6[00:51:56] <Kleadron> not like the bot would respond anyways because it doesnt half the time
L7[00:52:20] <Mimiru> Ya know, no one is forcing you to interact with her... Right?
L8[00:52:35] <Kleadron> who are you talking to
L9[00:52:45] <Mimiru> You.
L10[00:53:22] <Kleadron> yea, i know that nobody is forcing me to interact with a bot just for a command that has no real purpose
L11[00:53:32] <Kleadron> but i have the choice to care about it
L12[01:02:26] ⇨ Joins: MacDeth (MacDeth!~macdeth@205.185.152.173)
L13[01:03:20] ⇦ Quits: MacDeth (MacDeth!~macdeth@205.185.152.173) (Client Quit)
L14[01:03:40] ⇨ Joins: MacDeth (MacDeth!~macdeth@205.185.152.173)
L15[01:04:26] ⇦ Quits: MacDeth (MacDeth!~macdeth@205.185.152.173) (Client Quit)
L16[01:04:35] ⇨ Joins: ben_mkiv (ben_mkiv!~ben_mkiv@109.250.153.212)
L17[01:09:55] ⇨ Joins: CheeseCake (CheeseCake!webchat@8.200.182.58.starhub.net.sg)
L18[01:11:46] <CheeseCake> Hello?
L19[01:12:11] <Forecaster> %hello
L20[01:12:13] <MichiBot> Forecaster: Hello! Welcome to #oc! The one and only opencomputers channel! Please ask your questions directly (dont ask to ask) and provide error/code examples! (Use pastebin.com if theyre more than one line!) Dont mind the random conversation you might have walked into.
L21[01:12:55] <CheeseCake> I've found something strange(?) about the walk signal
L22[01:13:44] <CheeseCake> it only triggers when I walk from one part of a screen multiblock to another
L23[01:14:46] <CheeseCake> walking onto, off of, etc. the same screen block (or a multiblock with only one screen) fires the walk signal only once
L24[01:16:13] <CheeseCake> *once as in continuing to walk onto and off the same screen triggers no additional signals
L25[01:17:16] <CheeseCake> is this behaviour intended?
L26[01:26:52] <Forecaster> it might be meant as a detector of sorts
L27[01:27:13] <Forecaster> and thus there's no reason to repeatedly signal when stepping on and off multiple times
L28[01:27:19] <Forecaster> I'm not sure, I've never used it
L29[01:31:18] <CheeseCake> Hmm, it does make some sense for performance, was just wondering if anyone here also had this happen before because the wiki doesn't mention anything about it (then again the wiki doesn't mention quite a few things)
L30[01:32:33] <CheeseCake> It's also mildly annoying that I need a minimum of two blocks and the screen saying "please wiggle around" to reliably detect a player xD
L31[01:55:01] ⇦ Parts: CheeseCake (CheeseCake!webchat@8.200.182.58.starhub.net.sg) (webchat.esper.net))
L32[02:39:28] <Lizzian> gotta love it when you uninstall stuff and it doesn't give you an option to not reboot but then you also have windows updates to do
L33[02:42:54] <Corded> * <Lizzian> sighs
L34[02:43:06] <Lizzian> I should probably rebuild this pc but I don't want to
L35[02:44:14] <Lizzian> cause it hasn't been fully rebuilt (wipe and reinstalled) since we were first testing win10 in the latter part of 2015
L36[02:47:15] <Forecaster> ohno
L37[02:58:08] <Lizzian> k, seems to have calmed down now
L38[04:08:17] <Forecaster> %tonk
L39[04:08:19] <MichiBot> Forecaster! You beat Lizzy's previous record of 3 hours, 1 minute and 38 seconds! I hope you're happy!
L40[04:08:20] <MichiBot> Forecaster's new record is 3 hours, 19 minutes and 29 seconds
L41[04:08:26] <Forecaster> %tonkout
L42[04:08:27] <MichiBot> Forecaster has tonked out! Tonk has been reset! They gained 0.003 tonk points! Current score: 0.003
L43[04:08:37] <Lizzian> %tonk
L44[04:08:59] <Lizzian> hm
L45[04:09:01] <Forecaster> timeout
L46[04:11:30] <Forecaster> %tonk
L47[04:11:30] <MichiBot> You still hold the record Forecaster, for now... <0
L48[04:21:34] <Forecaster> %tonk
L49[04:21:36] <MichiBot> You still hold the record Forecaster, for now... <0
L50[04:21:42] <Forecaster> well then
L51[04:25:33] <Lizzian> %tonk
L52[04:25:34] ⇦ Quits: MichiBot (MichiBot!~MichiBot@eos.pc-logix.com) (Remote host closed the connection)
L53[04:25:40] <Lizzian> rip
L54[04:26:49] <Forecaster> sigh
L55[04:27:17] <Forecaster> why does that never happen when I'm testing
L56[04:38:04] <Bob> %tonk
L57[04:38:07] <Bob> :GWchadMEGATHINK:
L58[04:38:18] <Bob> Someone broke it
L59[05:01:26] <Kodos> https://media.discordapp.net/attachments/518929325807108116/524719089420075052/rMRj97S.jpg?width=385&height=682 Holy shit the link but this is hilarious af
L60[05:05:37] <Bob> :thonk_dye:
L61[05:06:11] ⇨ Joins: Vexatos (Vexatos!~Vexatos@p200300C107205E74FB8A6F040851060F.dip0.t-ipconnect.de)
L62[05:06:11] zsh sets mode: +v on Vexatos
L63[05:06:22] <Wattana Gaming> Heyo
L64[05:06:26] <Bob> As soon as he popped in discord ?
L65[05:07:06] ⇨ Joins: wattana (wattana!~wattana@node-9sv.pool-125-27.dynamic.totbb.net)
L66[05:07:24] <wattana> Guess I better use IRC for now, discord is going nuts
L67[05:27:34] ⇦ Quits: wattana (wattana!~wattana@node-9sv.pool-125-27.dynamic.totbb.net) (Quit: wattana)
L68[05:52:13] <Forecaster> nice
L69[05:52:30] <Forecaster> Plex finally pushed the new player interface to the self-hosted player
L70[05:53:02] <Forecaster> their online player has had it for ages now...
L71[07:43:50] <AmandaC> %8ball halucinate?
L72[07:43:55] <AmandaC> ... right
L73[07:56:19] <Forecaster> nope!
L74[07:56:47] <Forecaster> my clients roll command should work from irc (the command character is !)
L75[07:56:57] <Forecaster> I forget if I disabled that or not
L76[09:23:54] ⇦ Quits: linuxdaemon (linuxdaemon!linuxdemon@bnc.linuxdemon.xyz) (Quit: RIP.)
L77[09:29:08] <Wattana Gaming> be gone, demons!
L78[09:38:36] ⇨ Joins: linuxdaemon (linuxdaemon!linuxdemon@bnc.linuxdemon.xyz)
L79[09:51:17] <ZefTheFox> I swear he is literally 12
L80[09:55:26] ⇨ Joins: Inari (Inari!~Pinkishu@p5DEC6F19.dip0.t-ipconnect.de)
L81[09:58:11] <Lizzian> wat?
L82[09:58:38] <Forecaster> Wattana
L83[09:59:15] <Inari> Meow
L84[09:59:16] <Inari> %tonk
L85[09:59:28] <Inari> Oh, its dead
L86[10:00:02] <Forecaster> yep
L87[10:00:23] <Inari> Why is it dead?
L88[10:01:09] <Forecaster> previous tonk killed her
L89[10:03:09] <Inari> :<
L90[10:03:37] <Inari> Michiyo: I've still got it stuck in my head
L91[10:48:49] * Inari hands AmandaC some maraccas
L92[11:48:29] <Inari> Oh neat
L93[11:48:35] <Inari> I found a site that'll help me find a nice artist
L94[11:53:08] <Forecaster> https://xkcd.com/2087/
L95[11:53:12] <Forecaster> seems accurate
L96[11:53:19] <Inari> https://artistsnclients.com/ that one
L97[12:38:45] <Skye> don't want to consider VicNightfall? :P
L98[12:48:21] ⇨ Joins: ben_mkiv|afk (ben_mkiv|afk!~ben_mkiv@109.250.156.54)
L99[12:49:26] ⇦ Quits: Dimtree (Dimtree!~dimtree@74-196-36-39.nbrncmtk01.res.dyn.suddenlink.net) (Quit: Peace)
L100[12:50:20] <Inari> Skye: I mean, if Vic has improved a ton lately? :p
L101[12:51:10] ⇦ Quits: ben_mkiv (ben_mkiv!~ben_mkiv@109.250.153.212) (Ping timeout: 183 seconds)
L102[13:01:30] <Inari> Panties inhernetly have a "buy one, get one free" offer
L103[13:04:33] <Skye> Inari, https://twitter.com/VicNightfall/status/1074198953409814528
L104[13:05:04] <Inari> Well, much better than a while ago, but still not quite there
L105[13:06:03] <Inari> Usually takes a lot of time for artists to get to a certain point
L106[13:07:13] <Skye> What level do you want? :p
L107[13:17:57] <Inari> I think the thing that bugs me most abotu that pic is the face, it looks sorta flat and oddly shaped. Maybe it's just the style the cahracter is regarding the shape though.
L108[13:21:24] <Skye> hem
L109[13:28:26] <Michiyo> %tonk
L110[13:28:30] <Michiyo> Oh.. I see
L111[13:28:54] <Forecaster> what happened this time?
L112[13:29:05] <Michiyo> no clue, and no time to check.
L113[13:29:22] ⇨ Joins: MichiBot (MichiBot!~MichiBot@eos.pc-logix.com)
L114[13:29:22] zsh sets mode: +v on MichiBot
L115[13:29:22] <Forecaster> >:
L116[13:29:24] <Michiyo> %tonk
L117[13:29:33] <Inari> Tonking of the dead
L118[13:29:56] <Inari> Guess I'll have to start downloading pics instead of justl iking them, stupid Twitter
L119[13:29:58] <MichiBot> Michiyo! You beat Forecaster's previous record of <0! I hope you're happy!
L120[13:29:59] <MichiBot> Michiyo's new record is 9 hours, 21 minutes and 29 seconds
L121[13:30:19] <Inari> Oh look
L122[13:30:21] <Inari> I get AmandaC's tells
L123[13:38:58] <Inari> Skye: And depends a bit on the style, I like various ones. And it can be a bit hard to categorize them into "levels". E.g. for "refined" (or whatever it woudl be called) where the lines feel strong and delimiting, and it feels maybe even a bit artifical or perfectionist, there'd be something like https://twitter.com/ordan/status/593371820549476352/photo/1
L124[13:38:58] <Inari> But I also like more natural water colour-y things at times. Or just fun and free looking stuff like kind of an oil-painting ish thing. But i can't find those two pics currently
L125[13:51:01] ⇨ Joins: Stary_ (Stary_!~Stary@2001:470:ca8f::ffff)
L126[13:52:30] ⇨ Joins: Fridtjof_ (Fridtjof_!~prassel@osiris.9net.org)
L127[13:53:02] ⇦ Quits: Fridtjof_ (Fridtjof_!~prassel@osiris.9net.org) (Remote host closed the connection)
L128[13:53:02] ⇦ Quits: Stary_ (Stary_!~Stary@2001:470:ca8f::ffff) (Client Quit)
L129[14:53:13] ⇦ Quits: Vexatos (Vexatos!~Vexatos@p200300C107205E74FB8A6F040851060F.dip0.t-ipconnect.de) (Quit: Insert quantum chemistry joke here)
L130[14:59:18] ⇦ Quits: ben_mkiv|afk (ben_mkiv|afk!~ben_mkiv@109.250.156.54) (Ping timeout: 190 seconds)
L131[15:12:58] <Kodos> %tonk
L132[15:13:00] <MichiBot> I'm sorry Kodos, you were not able to beat Michiyo's record of 9 hours, 21 minutes and 29 seconds this time.
L133[15:13:01] <MichiBot> 1 hour, 43 minutes and 3 seconds were wasted!
L134[15:18:06] <CompanionCube> it's a nice record earned because of MichiBot's brokeness.
L135[15:19:13] <Inari> %pet CompanionCube
L136[15:19:14] * MichiBot pets CompanionCube with a Shiny anti-pudding! (10%). 24 health gained!
L137[15:19:35] <Inari> Forecaster: I see
L138[15:47:14] <payonel> %lua ponk
L139[15:47:14] <MichiBot> nil
L140[15:47:38] * payonel bows to Inari, queen-immortal of the ponk
L141[15:47:43] <Izaya> do you ever like
L142[15:47:49] <Izaya> have a coffee
L143[15:47:55] <Izaya> then accidentally write a gopher server
L144[15:54:40] <Inari> payonel: It was killed/ xD
L145[15:54:43] <Inari> * ?
L146[16:00:32] <payonel> when michibot restarts the lua state is reset
L147[16:04:21] <Izaya> GNU Icecat has a cute logo https://i.4cdn.org/g/1545160740927.png
L148[16:06:04] <payonel> what is icecat?
L149[16:06:51] <Izaya> the FSF's Firefox fork
L150[16:07:15] <Izaya> It's basically Firefox LTS with assorted things removed
L151[16:08:08] <Izaya> IIRC there's no DRM'd media support, none of the proprietary service tie-ins, etc
L152[16:12:56] <Forecaster> Inari: what?
L153[16:13:14] <Inari> Forecaster: Those magic items in user invs seem to be more powerful than your standard item
L154[16:14:02] <Forecaster> They're not unless they trigger the keyword system
L155[16:14:07] <Inari> Oh :p
L156[16:14:56] <Forecaster> Which adds a +something to it depending on the keywords
L157[16:17:55] <Izaya> apparently ungoogled-chromium is pretty okay too
L158[16:18:07] <Forecaster> Which stack
L159[16:18:16] <Izaya> maybe I'll use that for the 3 annoying web pages that use 1000000% CPU
L160[16:21:50] <CompanionCube> Izaya: how did you write your gopher server
L161[16:22:46] ⇨ Joins: Vexatos (Vexatos!~Vexatos@p200300C107205E74FB8A6F040851060F.dip0.t-ipconnect.de)
L162[16:22:46] zsh sets mode: +v on Vexatos
L163[16:23:01] <Izaya> perhaps amusingly it looks like my frequest server
L164[16:23:59] <Inari> payonel: Lame
L165[16:24:00] <CompanionCube> how much like?
L166[16:24:12] <Inari> Also N> free FPGA simulation software
L167[16:24:20] <Izaya> luasocket + luafilesystem, a coroutine is created for each client, requests are handled atomically
L168[16:24:32] <CompanionCube> ah
L169[16:24:54] <Izaya> if I replaced luasocket with minitel and luafilesystem with the fs lib it'd probably run on OC
L170[16:26:23] <Izaya> I tried a new client - cgo - and I like how it works
L171[16:26:36] <Izaya> Maybe I'll implement actual gopher over minitel
L172[16:27:38] <Inari> Guess I'll just have to buy a dev board though
L173[16:29:17] ⇨ Joins: hi2 (hi2!~hi2@177.22.81.90)
L174[16:29:37] ⇦ Parts: hi2 (hi2!~hi2@177.22.81.90) ())
L175[16:35:34] <Michiyo> @Forecaster Java frames: (J=compiled Java code, j=interpreted, Vv=VM code)
L176[16:35:35] <Michiyo> J 2648 org.sqlite.core.NativeDB.column_text_utf8(JI)[B (0 bytes) @ 0x00007f079157d158 [0x00007f079157d120+0x38]
L177[16:35:38] <Michiyo> is where it crashed today
L178[16:35:53] <Michiyo> the same thing as every other recent crash :P
L179[16:37:52] <Forecaster> Sigh
L180[16:38:21] <Michiyo> Which is why we're moving off of SQLite :P
L181[17:07:23] <Inari> Weird
L182[17:16:02] ⇦ Quits: Inari (Inari!~Pinkishu@p5DEC6F19.dip0.t-ipconnect.de) (Quit: KVIrc 5.0.0 Aria http://www.kvirc.net/)
L183[18:24:42] <Izaya> CompanionCube: https://git.shadowkat.net/izaya/tiny-gopherd
L184[19:04:43] <Wuerfel_21> TFW your keyboard driver just gives up
L185[19:04:58] <Wuerfel_21> no more meta key, no more tab
L186[19:05:13] <Izaya> shoulda used PS/2
L187[19:07:46] <Wuerfel_21> oh, it isn't the keyboard driver, it's teamviewer running in the background inexplicably mucking up the keyboard
L188[19:08:45] <Izaya> at one point I had VMWare fuck up the keyboard drivers on Windows
L189[19:09:04] <Izaya> qemu/kvm stronk
L190[19:28:08] ⇦ Quits: Vexatos (Vexatos!~Vexatos@p200300C107205E74FB8A6F040851060F.dip0.t-ipconnect.de) (Quit: Insert quantum chemistry joke here)
L191[19:34:58] <Wuerfel_21> It's really strange. It captures the special keys and just sends 'em off to the remote PC without checking that it is actually the selected window
L192[20:09:35] <S3> Thinking I can decrease the size of the trotwood actor core even more
L193[20:09:45] <S3> or if not, simplify it
L194[20:12:03] <S3> I wonder how much of the core I lost when I replaced my hard drive
L195[20:12:36] <S3> ugh. almost all of it
L196[20:12:44] <S3> \well I can get it bac
L197[20:13:10] <S3> I just need to connect the disk to some USB to sata or something
L198[20:16:10] <S3> wait, it can't be too old
L199[20:16:33] <S3> my subscribe call has a lambda argument
L200[20:19:20] <CompanionCube> did you not transfer all the things?
L201[20:20:15] <S3> well. I kinda wanted to but it took 10 minutes just to type ls sometimes
L202[20:20:34] <S3> and the disk would randomly shut off and click then start back up
L203[20:20:57] <Izaya> That sounds uncomfortably familiar
L204[20:20:58] <S3> funny thing is
L205[20:21:07] <S3> now I need to find where I put the disk XD
L206[20:21:27] <S3> so what I have left of the actor core for Trotwood is the code that contains everything but all of the yield call implementations themselves
L207[20:22:06] <S3> so it has the calls, send, recv, wake, sub, unsub, spawn, and quit.. but they just aren't written
L208[20:22:17] <Izaya> S3: thoughts on jehanne?
L209[20:22:19] <S3> the only thing that wasn't done iirc was the inter computer routing
L210[20:22:50] <S3> why does this sound familiar Izaya
L211[20:23:05] <Izaya> unsure
L212[20:23:13] <S3> what is it
L213[20:23:30] <Izaya> tl;dr it's a plan 9 fork that is going for simplicity and correctness over all else
L214[20:23:36] <S3> oh
L215[20:23:37] <Izaya> ie all calls are blocking
L216[20:23:41] <S3> I wouldn't know
L217[20:24:34] <S3> So they think nonblocking is bad
L218[20:24:43] <Izaya> not necessarily
L219[20:24:55] <Izaya> just that having both blocking and non-blocking complicates things
L220[20:25:17] <S3> well there are times to block and times you want a promise or something instead
L221[20:25:33] <S3> they have their reasonings
L222[20:25:48] <Izaya> *it does have a way to prematurely exit a blocking call, in the form of setting an alarm in advance
L223[20:26:03] <S3> neat
L224[20:26:51] <S3> Heh. Makes it seem so weird when I think of how calls work in trotwood
L225[20:27:04] <S3> where the moment you call any useful function your current process is suspended
L226[20:27:22] <S3> and it never wakes up until its done or something different comes
L227[20:28:21] <S3> If they want consistency and one or the other I would say blocking is the way to go
L228[20:29:00] <S3> the problem with non blocking is that it can actually require more work on the processing side, because in order to conserve power, you should yield your process
L229[20:29:03] <S3> or put it to sleep
L230[20:29:18] <S3> but if you're doing non blocking, you will need to do both
L231[20:29:39] <S3> and if you're doing non blocking io in a blocking like manner it's kind of a waste if it's not doing more than just that
L232[20:30:52] <S3> the weirdness in trotwood though comes with the idea that not only does every useful call yield your process and put it to sleep, but also the return of the sleep does NOT mean your result has arrived
L233[20:32:09] <S3> for example, you can call some routine somewhere else and you get suspended, then some new data comes and you handle it, but you didnt get the return of your call yet.
L234[20:32:42] <S3> for this reason unless you are the beholder and distributor of your own state you should never carry state in between your yields.
L235[20:33:09] <S3> and instead think of yourself as somebody who is just checking the mail every time you are woken up
L236[20:33:47] <S3> to prevent crashing and ease communication between two endpoints you can develop a protocol that resembles a finite state machine perhaps with a type system.
L237[20:34:43] <S3> It's also a crime in Trotwood to have a process handle many jobs.
L238[20:35:11] <Izaya> punishable by death?
L239[20:35:13] <Izaya> :D
L240[20:35:20] <S3> which is also why supervising processes don't do any work at all unless that work is being told to create new processes or reap them :D
L241[20:35:23] <S3> rofl
L242[20:35:37] <S3> I thought I was just blabbering on at this point heh
L243[20:36:38] <S3> Either way. a feature of Trotwood I want to work out is the ability to optionally package your programs each in their very own one single file
L244[20:36:53] <S3> and it will sort of disperse processes for you using some helper functions
L245[20:37:31] <S3> so you can choose to seperate a program in multiple files or just distribute a single file program that spawns a tree of many supervisors and worker processes
L246[20:38:28] <S3> I always liked my single file <= 1000 lines of code programs
L247[20:38:31] <Izaya> That seems interesting.
L248[20:39:06] <Izaya> So you give it a file and it figures out "this is the supervisor function, we need x many of those, and they will spawn y many worker threads"?
L249[20:39:25] <S3> Well. The other problem is deciding how to handle cluster control. It's a big one
L250[20:39:51] <S3> say you have 5 OC computers that are tagged with the same cluster ID, so you have 5 computers in your 'environment'
L251[20:40:18] <S3> there are times when you want to spawn 100 processes and you want them each to distribute across the cluster randomly
L252[20:40:31] <S3> so maybe ~ 20 procrsses on each machine by the time you are done
L253[20:40:52] <S3> however, there are times when you want to specify a node to run a process, that one is also easy.
L254[20:41:06] <S3> I have found other situations though that complicate things
L255[20:41:49] <S3> For example, I realized that in the event of say a process that handles a particular type of component, sometimes it is important to spawn one process across each node of the cluster.
L256[20:41:56] <Izaya> At a guess, wonderful for the intended purposes, but kinda ass for anything unexpected?
L257[20:42:12] <S3> ?
L258[20:42:21] <Izaya> oh okay that makes sense
L259[20:42:53] <S3> yeah. so an abstracting library may exist on any node of a cluster sure,
L260[20:43:37] <S3> but the thing is, a component on a computer seperated by a relay can ONLY be handled by a process on that node or in that network of nodes in the relay
L261[20:43:43] <S3> behind the relay*
L262[20:44:06] <S3> so it gets complicated and I want it to be simple
L263[20:44:37] <S3> The only ide aI've come up with so far that makes any sense is providing a processing plan optionally when you spawn a process
L264[20:44:46] <S3> by default, it spawns the process on the current node
L265[20:45:22] <S3> but you could provide a plan such as spawn this process on all nodes, or spawn this process on any node available
L266[20:46:08] <S3> perhaps then the nodes would decide which one gets to spawn and the process' contents are sent over the network via their special storage protocol
L267[20:46:14] <S3> or something
L268[20:47:29] <S3> in the end I want the result to be a super lightweight codebase that's super easy to use and all of those advanced features are usable by anyone but half hidden away from the people who don't need them
L269[20:47:50] <S3> i.e. you shouldn't be forced to cluster compute unless you actually intend to do so
L270[20:49:09] <S3> That's another problem as well, shipping process code over the network
L271[20:50:13] <S3> It seems better to finish my distributed filesystem and let each node grab it from disk / cache
L272[20:50:38] <S3> but distributed ilesystem means distributed processes
L273[20:50:48] <S3> \so I run into a loop here :)
L274[20:51:38] <S3> I also have not anticipated disk failure
L275[20:52:13] <S3> so I thought maybe perhaps making use of the raid component to store metadata optionally if you have people destroying your computers
L276[20:52:33] <S3> this way they could be replaced and rebuild themselves if they werent properly removed
L277[20:54:36] <Izaya> that seems interesting
L278[20:55:30] <S3> yeah but it sort of beats the idea of distributed disk storage
L279[20:55:54] <S3> So I've been lately doing research on redundant or recoverable distributed hashing databases
L280[20:56:03] <S3> which is exactly the kind of disk storage system I was going for
L281[20:56:45] <S3> Oh yeah! and because of stuff like that I've also thought it'd be cool to "categorize" your cluster into subclusters
L282[20:57:12] <S3> for ecample you could have a cluster where a subsluster just does disk stuff so its entire RAM in the OC computer is just for disk caching for faster access etc
L283[20:57:22] <S3> but you could have a telecom subcluster
L284[20:57:28] <S3> with as many telecom nodes as you need
L285[20:57:32] <S3> perhaps one switch in every region or base
L286[21:00:41] <S3> Don't forget, there's also the protocol I need to work on for determining if a switch needs to become a router to hop a relay for you.
L287[21:00:59] <S3> which is dependant on your distance metric.
L288[21:01:24] <S3> I'm thinking of going an ospf like route, and when you talk to them the switches figure out how far away you are from them
L289[21:01:31] <S3> using some crazy spamming mess..
L290[21:01:42] <S3> Unless you have any ideas of course :D
L291[21:02:25] <Izaya> you know me
L292[21:02:32] <Izaya> my ideas are 100% dumb by design
L293[21:02:42] <Izaya> and I don't think dumb ideas will solve that
L294[21:05:54] <Izaya> https://s3.wasabisys.com/sergal-assets/media_attachments/files/000/756/715/original/4377f3037a88c036.jpg
L295[21:20:13] ⇨ Joins: BILLPC2684 (BILLPC2684!~billpc268@158.69.124.22)
L296[21:22:02] <BILLPC2684> hey does enyone know why OpenOS crashes with `Attempt to call field 'insert' (a nil value)`?
L297[21:22:24] <BILLPC2684> it makes everything unusable
L298[21:22:46] <FuriousRedstoner> http://tinyurl.com/yc2nl9uw
L299[22:03:03] ⇦ Quits: BILLPC2684 (BILLPC2684!~billpc268@158.69.124.22) (Remote host closed the connection)
L300[22:36:59] <AmandaC> %choose be good or be bad
L301[22:37:00] <MichiBot> AmandaC: Elementary dear Watson, be bad is the obvious choice!
L302[22:44:21] <AmandaC> %tell Inari s/paperbook/ereader/ https://i.imgur.com/SbVLph0.jpg
L303[22:44:22] <MichiBot> AmandaC: Inari will be notified of this message when next seen.
L304[23:03:54] ⇦ Quits: andreww (andreww!~xarses@24.5.179.230) (Ping timeout: 190 seconds)
L305[23:39:06] ⇨ Joins: xarses (xarses!~xarses@24.5.179.230)
L306[23:47:38] ⇦ Quits: xarses (xarses!~xarses@24.5.179.230) (Ping timeout: 190 seconds)
L307[23:47:43] ⇨ Joins: xarses (xarses!~xarses@24.5.179.230)
<<Prev Next>> Scroll to Top