L2[00:44:19] <Amanda> Night girls

L4[02:43:16] <walks​anator> the mod is coming along https://cdn.discordapp.com/attachments/125649403162656768/1153521665767129178/image.png

L5[02:55:05] <Ocawes​ome101> nice!

L6[02:55:11] <Ocawes​ome101> %tonkout

L7[02:55:12] <MichiBot> Jiminy Cricket! Ocawes​ome101! You beat your own previous record of 2 hours, 26 minutes and 7 seconds (By 1 hour, 10 minutes and 30 seconds)! I hope you're happy!

L8[02:55:13] <MichiBot> Ocawes​ome101 has tonked out! Tonk has been reset! They gained 0.003 tonk points! plus 0.004 bonus points for consecutive hours! Current score: 1.14823781, Position #3 Need 0.08495043 more points to pass Forec​aster!

L9[02:55:16] <Ocawes​ome101> %sip

L10[02:55:17] <MichiBot> You drink a fluffy aluminium potion (New!). Ocawesome101 grows whiskers until they download a car.

L11[04:43:35] <Izzy> minitel PSA: I'm getting ready to push out a change to MT-RPC that will be incompatible with old versions, so you'll want to update everything in one go. it provides better error handling and I'm just doing some testing on OpenOS and PsychOS versions to make sure they'll be okay.

L12[04:46:48] <Izzy> cc Amanda, summarily the true from pcall is no longer discarded on success

L13[04:46:58] <Izzy> and is now handled by the client library

L16[04:59:48] <Izzy> also, ported the OpenOS RPC ACL system to PsychOS

L17[05:00:42] <Izzy> now I guess I need to port the workgroup daemon to OpenOS, huh

L18[05:01:13] <Forec​aster> %tonk

L19[05:01:13] <MichiBot> Goshhawk! Forec​aster! You beat Ocawes​ome101's previous record of <0 (By 2 hours, 6 minutes and 1 second)! I hope you're happy!

L20[05:01:14] <MichiBot> Forecaster's new record is 2 hours, 6 minutes and 1 second! Forecaster also gained 0.0021 tonk points for stealing the tonk. Position #2. Need 0.04671037 more points to pass Va​ur!

L23[05:23:22] <Izzy> something I want to poke at is a k,v store accessible over RPC or sockets

L24[05:24:11] <Izzy> use a client library to wrap queries so it can transparently go via RPC, or sockets if the data is too large to fit in a packet

L25[05:29:17] <Izzy> https://media.shadowkat.net/media/408c9f7344ff359acc0dd8e78a32120435b99d70ff9430c5642613ad4915518b.png

L29[07:47:06] <ar2​000> >Izzy: use a client library to wrap queries so it can transparently go via RPC, …

L30[07:47:07] <ar2​000> IP support fragmentation, TCP must avoid it... some things are weird

L31[07:47:31] <Izzy> minitel RPC does not support fragmentation

L32[07:47:33] <Izzy> :)

L33[07:48:20] <Izzy> hence wanting support for sockets for large responses

L34[08:02:55] <Na​dja> >ar2000: IP support fragmentation, TCP must avoid it... some things are weird

L35[08:02:55] <Na​dja> TCP doesn't need to avoid fragmentation, it just lowers performance with zero benefits.

L36[08:06:41] <ar2​000> >Nadja: TCP doesn't need to avoid fragmentation, it just lowers performance with …

L37[08:06:41] <ar2​000> That's not how I interpreted that : https://datatracker.ietf.org/doc/html/rfc9293#name-maximum-segment-size-option

L39[08:14:13] <Na​dja> >ar2000: That's not how I interpreted that : https://datatracker.ietf.org/doc/htm���

L40[08:14:13] <Na​dja> Well you interpreted wrong. TCP aims to not incur IP fragmentation, but MSS does not solve that problem by itself, as the lowest MTU of any routing middlebox is the limiting factor (and finding that lower limit out is a process called PMTUD, which is very much separate from MSS)

L41[08:16:16] <ar2​000> While avoiding fragmentation is not a MUST, it is an important part of TCP (see sections under)

L42[08:16:38] <Na​dja> Only for performance reasons

L43[08:16:48] <ar2​000> TCP will work fragmented in IP, but SHOULD don it's best to avoid it

L44[08:17:41] <Na​dja> Oh yeah, it certainly SHOULD, as theres a performance downside to unnecessary fragmentation, but it doesn't *need* to. TCP works fine with IP fragmentation happening on a path in the middle.

L47[11:39:02] <Forec​aster> %tonkout

L48[11:39:02] <MichiBot> Dad-Sizzle! Forec​aster! You beat your own previous record of 2 hours, 6 minutes and 1 second (By 4 hours, 31 minutes and 47 seconds)! I hope you're happy!

L49[11:39:03] <MichiBot> Forec​aster has tonked out! Tonk has been reset! They gained 0.006 tonk points! plus 0.01 bonus points for consecutive hours! Current score: 1.25128824, Position #2 Need 0.03071037 more points to pass Va​ur!

L52[16:21:03] <Forec​aster> dammit

L53[16:21:09] <Forec​aster> media server is running out of space again

L56[16:29:49] <Ocawes​ome101> %tonk

L57[16:29:49] <MichiBot> Huzzah! Ocawes​ome101! You beat Forec​aster's previous record of <0 (By 4 hours, 50 minutes and 46 seconds)! I hope you're happy!

L58[16:29:50] <MichiBot> Ocawesome101's new record is 4 hours, 50 minutes and 46 seconds! Ocawesome101 also gained 0.00485 tonk points for stealing the tonk. Position #3. Need 0.09820043 more points to pass Forec​aster!

L59[16:29:54] <Ocawes​ome101> %sip

L60[16:29:55] <MichiBot> You drink a punctual bombastium potion (New!). Ocawesome101 feels slightly less agile.

L64[16:43:27] <stephan48> Amanda: the system is alive and kicking. it just did not want to upgrade python and thus ansible was bitching about python venvs not being whole. a manual apt dist-upgrade which did the six python packages made it work.

L65[16:43:47] <Amanda> I see

L66[16:44:11] <stephan48> sadly it removed the python mail stuff halfway through the install, so it could not send me a success mail after the first run :P

L67[17:26:30] <Amanda> there, my debian VMs for nomad are now up to bookworm

L68[17:27:09] <Forec​aster> %sip

L69[17:27:10] <MichiBot> You drink an included bavarium potion (New!). The next fork Forecaster touches tells them it's most deeply guarded secret.

L70[17:27:25] <Forec​aster> there are snacks hidden behind the lamp over there?!

L71[17:31:59] <Va​ur> %sip

L72[17:32:01] <MichiBot> You drink a smooth diamond potion (New!). Vaur grows slightly until hell has a nice fall.

L73[17:55:30] <stephan48> Amanda: congratulations

L74[17:55:47] <Amanda> and now my ci runner is, all's that's left is my matrix server

L75[17:56:23] <Amanda> I think

L76[17:58:23] <Amanda> oh, and my db vms

L77[17:58:29] <Amanda> meh, problem for tomorrow

L78[17:58:31] <Amanda> or later

L79[17:58:50] <stephan48> if you drop all databases, you save upgrade time and backupspace!

L80[18:00:19] <Amanda> I think I'll pass. :P

L81[18:14:32] <Mic​hiyo> fml speaking of databases... I seem to no longer be able to login to any of mine ._.

L82[18:16:02] <Forec​aster> that's a significant handicap

L83[18:20:06] <stephan48> Michiyo: mysql or postgres?

L84[18:20:16] <stephan48> and do you have local root?

L85[18:22:34] <Michiyo> MySQL, yes I have root, but it's a bit more complicated... *I* can access *my* databases with no issue. My Plesk users however are unable to use phpMyAdmin to access *their* databases.

L86[18:23:12] <stephan48> ouch Plesk.

L88[18:24:11] <Michiyo> I wanted the ability to host websites for folks without having to either turn over system control to them, or do everything they wanted.

L89[18:24:16] <Michiyo> Plesk was the easiest way to do that.

L91[20:34:20] <Forec​aster> %sip

L92[20:34:21] <MichiBot> You drink a wonderful currentcorn potion (New!). Forecaster: Nothing seemed to happen...

L93[20:34:34] <Forec​aster> >:

L94[20:44:49] <Amanda> Welp, my pleroma seems to have broken

L95[20:45:52] <Amanda> Husky just says "an error occurred", loading it in the web ui shows a single page then complains 500 internal server error

L96[20:49:48] <Amanda> Huh

L97[20:49:59] <Amanda> Somehow a user is in the db twice

L98[20:50:21] <Amanda> And pleroma doesn't like that

L99[20:50:40] <Amanda> %choose re-computer to try and fix it?

L100[20:50:40] <MichiBot> Ama​nda: I talked to the Swedish Chef, he said "Bork bork!" I think that means yes?

L101[21:00:13] <Amanda> Looks like I can blame Izzy

L102[21:00:35] <Amanda> The post they repeated from https://wetdry.world/users/garfield_ somehow caused a second entry for that user in the db

L104[21:09:58] <nadja> Ah yes, stephan48 who thinks LDAP is not cursed is scared of Plesk. Typical :P

L105[21:10:22] <stephan48> :P

L106[21:11:24] <stephan48> one i can control, manage and understand... the other is pure evil with so many moving parts I would consider it as complex as a medium sized k8s cluster. and yes i messed with Plesk internals before because it broke in interesting ways

L107[21:12:49] <Mic​hiyo> I just needed something that made it easy for users to control their own hosting without asking me to change their email password for the 37th time that week.

L108[21:12:59] <Mic​hiyo> r to add the 900th subdomain.

L109[21:13:10] <stephan48> yea sensible software in that range is practically non existent :/

L110[21:18:33] <Izzy> Amanda: and here I was thinking from the ping my RPC changes had broken something :p

L111[21:23:03] <Amanda> Nah, just pleroma breaking itself somehow

L112[21:26:43] <Va​ur> %tonk

L113[21:26:45] <MichiBot> Wut! Va​ur! You beat Ocawes​ome101's previous record of 4 hours, 50 minutes and 46 seconds (By 6 minutes and 8 seconds)! I hope you're happy!

L114[21:26:46] <MichiBot> Vaur's new record is 4 hours, 56 minutes and 55 seconds! Vaur also gained 0.0005 (0.0001 x 5) tonk points for stealing the tonk. Position #1.

L116[22:31:29] <Izzy> so OpenOS runs isDirectory, size, and lastModified against every file when it gets a dir listing, even when you're not asking for -l. MTFS for OpenOS implements something to make it grab and cache all that info in one call, rather than requiring 3 RPC calls per file when getting a listing. Implemented it in PsychOS, figured I'd find out the performance difference.

L117[22:31:31] <Izzy> https://shadowkat.net/tmp/1a1b.png

L118[22:49:35] <CompanionCube> %inv add sky's dns servers

L120[22:51:04] <CompanionCube> it's weird reading a supposedly 'national' outage of your isp on twitter, and then a quick test showing that it's their DNS that's down.

L121[23:39:17] <the_p​atmann> "I'm not down, you're down!"