L2[01:34:44] <Izzy> does anyone have a working example of CraftTweaker blocks with custom models?

L3[01:37:45] <Kristo​bear38> @RedstoneParkour yeah that's not the problem; the problem is that data that should arrive does not. I think this is rather a bug in the STEM bridge. I would need to host my own instance and check its debug logs

L4[01:38:08] <Kristo​bear38> Provided that I can still replicate the bug with the bridge hosted elsewhere

L5[01:53:54] <Kristo​bear38> For the record I'm connecting to the public instance at https://stem.fomalhaut.me/

L6[02:14:06] <Forec​aster> alrigh, time to check out these corvette things

L7[02:15:31] <Corded> > <Kristo​bear38> @RedstoneParkour yeah that's not the problem; the problem i…

L8[02:15:31] <Redston​eParkour> are you using normal oc or some sort of emulator?

L9[02:15:44] <Redston​eParkour> nvm its normal oc

L10[02:18:58] <Redston​eParkour> hmmmm i think i get why

L11[02:19:20] <Redston​eParkour> stem registers the handler before starting the finishConnect loop

L12[02:19:32] <Redston​eParkour> so those packets essentially get lost

L13[02:23:48] <Redston​eParkour> (for external programs, that is)

L14[02:26:03] <Kristo​bear38> Not sure if I understand

L15[02:27:37] <Kristo​bear38> Which packets get lost?

L16[02:31:01] <Redston​eParkour> Those that get sent before the finishConnect loop finishes

L17[02:31:29] <Redston​eParkour> stem still processes them, but since you can't subscribe to channels you'll miss any messages posted

L18[02:34:19] <Corded> > <Kristo​bear38> Not sure if I understand

L19[02:34:19] <Redston​eParkour> perhaps i should ask this: are you expecting to be able to receive messages *before* the finishConnect loop finishes?

L20[02:43:08] <Izzy> guys is it ~gay~ too unhinged to register 396 separate blocks with crafttweaker in order to implement 11 types of sheet metal roof times 9 roof block models times 4 rotations?

L21[02:49:24] <Forec​aster> if you do it manually, definitely

L22[03:15:06] <Kristo​bear38> @RedstoneParkour no, the problem is (I think) that I'm sending some messages to the stem bridge and the other side connected to the bridge responds, but this response isn't relayed to the OC side by the bridge

L23[03:16:09] <Kristo​bear38> And this only happens when internet_ready arrives before finishconnect loop is done

L24[03:22:30] <Amanda> Izzy: Fun fact, I'm pretty sure crafttweaker supports loops!

L25[03:22:52] <Forec​aster> I can second that pretty sure

L26[03:22:53] <Izzy> oh no

L27[03:22:55] <Izzy> it does

L28[03:39:54] <Amanda> %remindme 1h check CI

L29[03:39:55] <MichiBot> I'll tell you "check CI" in 1h at 08/28/2025 04:39:55 AM

L30[04:04:25] <Corded> > <Kristo​bear38> @RedstoneParkour no, the problem is (I think) that I'm send…

L31[04:04:25] <Redston​eParkour> how can you send messages before the finishConnect loop is done?

L32[04:04:54] <Kristo​bear38> I am not sending messages before finishconnect loop is done

L33[04:05:06] <Kristo​bear38> I am *getting* messages before it's done

L34[04:05:19] <Kristo​bear38> From the stem bridge apparently??

L35[04:07:19] <Redston​eParkour> probably ping messages then

L36[04:07:37] <Redston​eParkour> have you checked if anything is actually being sent on the channel?

L37[04:10:25] <Kristo​bear38> Yeah the other side is responding

L38[04:10:43] <Redston​eParkour> hmm

L39[04:10:45] <Kristo​bear38> So a message is being sent on the channel

L40[04:10:53] <Kristo​bear38> Just not relayed to the OC side

L41[04:10:55] <Redston​eParkour> open your browser on stem.fomalhaut.me

L42[04:11:47] <Redston​eParkour> then type in the channel id and click the button (iirc called 'join')

L43[04:11:55] <Kristo​bear38> Yeah that's literally what I did

L44[04:11:58] <Redston​eParkour> ah

L45[04:12:28] <Kristo​bear38> I see message from OC being sent, I see a response from the other side being sent, and the latter is not being relayed to OC side

L46[04:12:39] <Redston​eParkour> hmm

L47[04:13:37] <Redston​eParkour> this is confusing me as well...

L48[04:14:04] <Redston​eParkour> are `internet_ready` events being generated when the response is sent?

L49[04:14:13] <Kristo​bear38> No

L50[04:14:19] <Redston​eParkour> hmm

L51[04:14:34] <Kristo​bear38> But i think they are for heartbeat pings from the stem bridge!

L52[04:14:44] <Redston​eParkour> stupid question: did you actually subscribe to the channel?

L53[04:14:50] <Kristo​bear38> I see an internet_ready every half a minute or so

L54[04:15:08] <Corded> > <Kristo​bear38> I see an internet_ready every half a minute or so

L55[04:15:08] <Redston​eParkour> that's a ping yes

L56[04:15:10] <Kristo​bear38> Yes if I wasn't subscribed I would see a message from OC side

L57[04:15:23] <Redston​eParkour> wrap the subscribe call in an `assert`

L58[04:15:54] <Kristo​bear38> Besides this all works if there aren't any internet_readys before finishconnect loop finishes

L59[04:16:13] <Redston​eParkour> that's also confusing yes

L60[04:22:46] <Kristo​bear38> Anyway I think I can have an easy workaround without trying to find a proper solution: reconnect until there are no internet_readys before finishconnect loop is done

L61[04:23:25] <Redston​eParkour> worth trying

L62[04:23:52] <Kristo​bear38> I think conditions for this to happen are rare enough that this could work

L63[04:25:07] <Kristo​bear38> For the bug to happen I mean

L64[04:37:45] <Forec​aster> wellp... I'm stuck on a derelict freighter...

L65[04:37:49] <fingercomp> uh, I'm not sure I understand

L66[04:37:56] <fingercomp> the finishConnect loop doesn't have any yield points, does it?

L67[04:38:06] <fingercomp> so how can it handle events before it stops spinning?

L68[04:39:55] <MichiBot> Amanda REMINDER: check CI

L69[04:41:15] <Amanda> Oh bah, another job started

L70[04:41:22] <Amanda> %remindme 1h check CI again

L71[04:41:23] <MichiBot> I'll tell you "check CI again" in 1h at 08/28/2025 05:41:22 AM

L72[04:46:04] <Forec​aster> oh that's what the button next to the hatch is for

L73[04:46:53] <Forec​aster> it's for beaming to the freighter

L74[04:47:54] <Forec​aster> but for some reason the reverse isn't working

L75[04:49:49] <Forec​aster> this is apparently true both for regular freighters and derelict ones

L76[04:49:58] <fingercomp> @Kristobear38 so the problem is that you're unable to receive any messages (aside from pings?) in this unlikely scenario, even long after reconnect() returns, is that right? does it print any errors to /tmp/event.log?

L77[05:41:23] <MichiBot> Amanda REMINDER: check CI again

L78[06:45:19] <Amanda> %remindme 1h15m farm's upgraded

L79[06:45:20] <MichiBot> I'll tell you "farm's upgraded" in 1h15m at 08/28/2025 08:00:19 AM

L80[07:10:12] <Kristo​bear38> fingercomp: ad1. it does not handle any events since as you said, there are no yield points there. But you can see internet_ready signals being queued by OC before the loop finishes. ad2. yes, you understand correctly, I'm not able to receive any messages aside from pings after reconnect() returns. Nothing is printed to the

L81[07:10:12] <Corded> event.log

L82[07:23:35] <Kristo​bear38> Actually i need to confirm that these messages that I can still receive *are* in fact pings and not some other weird traffic but so far it seemed like so

L83[08:00:20] <MichiBot> Amanda REMINDER: farm's upgraded

L84[08:26:45] <Va​ur> %tonk

L85[08:26:47] <MichiBot> Bingo! Va​ur! You beat Forec​aster's previous record of 4 hours, 55 minutes and 54 seconds (By 6 hours, 54 minutes and 45 seconds)! I hope you're happy!

L86[08:26:48] <MichiBot> Vaur's new record is 11 hours, 50 minutes and 39 seconds! Vaur also gained 0.03455 (0.00691 x 5) tonk points for stealing the tonk. Position #2. Need 0.01368 more points to pass Compan​ionCube!

L87[08:27:03] <Va​ur> holy tonkpoints

L88[08:42:13] <Spider ​EveryOS> Are you for real?

L89[08:42:16] <Spider ​EveryOS> %sip

L90[08:42:19] <MichiBot> You drink a seeping water potion (New!). Spider EveryOS feels slightly more agile.

L91[08:44:00] <Va​ur> %sip

L92[08:44:04] <MichiBot> You drink a proprietary green potion (New!). As Vaur drinks the potion they become the target of a wad of llama spit! They fail to evade it with a 9 vs DC 12 and takes 1d​4 => 3 damage.

L93[08:50:20] <Corded> > <Spider ​EveryOS> Are you for real?

L94[08:50:20] <Va​ur> where you waiting for 12h mark or something ? 😄

L95[08:58:14] <Spider ​EveryOS> No, I've been preoccupied with other stuff so just forgot to check tonk

L96[08:58:42] <Spider ​EveryOS> I can't believe it went that high though

L97[09:00:53] <fingercomp> @Kristobear38 that would be helpful, yes

L98[09:01:22] <fingercomp> if the server sends pings but nor regular messages, either the client is not registered after all, or the server misbehaves

L99[09:02:31] <Brisingr​ Aerowing> Izzy: the ContentCreator mod supports blockstates IIRC. It can be used alongside ContentTweaker.

L100[09:28:30] <Amanda> https://matrix.camnet.site/_heisenbridge/media/camnet.site/mIpBKsIoTkPTtHvAZsVmShde/GVuvpNIoWLw/screenshot.jpeg

L101[09:28:34] <Amanda> Ffs

L102[09:29:03] <Amanda> Literally <500 away from turning a profit

L103[09:35:57] <Forec​aster> gotta work them harder!

L104[10:46:30] <Redston​eParkour> %sip

L105[10:46:32] <MichiBot> You drink a tacky rock potion (New!). A tiny cloud appears with a ridiculous smile on it. It follows RedstoneParkour until they have a nap.

L108[11:36:17] <Brisingr​ Aerowing> %splash @Forecaster

L109[11:36:18] <MichiBot> You fling a viscous iron potion (New!) that splashes onto @Forecaster. Dramatic music briefly plays in the distance.

L110[12:10:42] <Va​ur> %sip

L111[12:10:44] <MichiBot> You drink a simulated octiron potion (New!). Vaur gains knowledge about a random useless subject.

L112[12:57:21] <Va​ur> %sip

L113[12:57:23] <MichiBot> You drink a knifey nectar potion (New!). Vaur now has a mullet until they sneeze.

L114[13:13:59] <Forec​aster> Strategic %tonk

L115[13:14:00] <MichiBot> I'm sorry Forecaster, you were not able to beat Vaur's record of 11 hours, 50 minutes and 39 seconds this time. 4 hours, 47 minutes and 13 seconds were wasted! Missed by 7 hours, 3 minutes and 26 seconds!

L116[13:23:22] <Redston​eParkour> %sip

L117[13:23:24] <MichiBot> You drink a shaky green potion (New!). Tonk moved back 2 hours. (Rem. uses: 0)

L118[13:33:08] <Forec​aster> D:<

L119[13:53:16] <Amanda> %splash @Forecaster with mutable iron potion

L120[13:53:18] <MichiBot> You fling a mutable iron potion (New!) that splashes onto @Forecaster. @Forecaster turns into a sapphire platypus boy until hell celebrates midsummer.

L123[14:00:22] <Va​ur> %sip

L124[14:00:24] <MichiBot> You drink a crunchy diamond potion (New!). Vaur falls into a shaft and drop 3 floors!

L125[14:00:34] <Va​ur> ouch

L126[14:11:12] <Kristo​bear38> fingercomp: I've confirmed that the only messages that are being received are stem pings; Also I'm unable to reproduce the problem with the stem bridge instance hosted locally on my PC so I'll probably be unable to debug it

L127[14:15:02] <fingercomp> that means the server doesn't send messages (well, aside from pings) at all

L128[14:15:13] <fingercomp> most likely it didn't process the client's subscriptions

L129[14:16:09] <fingercomp> does `:subscribe()` return `true` when you try to join channels?

L130[14:18:42] <Izzy> Brisingr Aerowing: that's basically impossible to look up, got a link by chance? >.<

L131[14:18:55] <Izzy> oop I found it

L132[14:19:35] <Kristo​bear38> from the looks of it I am checking that subscribe returned true but I'll double-check

L133[14:22:33] <Kristo​bear38> it does indeed return true

L134[14:22:41] <Kristo​bear38> after joining a channel

L135[14:23:00] <Kristo​bear38> the channel join of course happens after finishconnect loop is done

L136[14:49:34] <fingercomp> hmm, it might be worth looking at network traffic to/from TCP port 5733 in tcpdump/wireshark, since stem work over raw TCP

L137[14:50:21] <fingercomp> in particular, to see whether it sends subscription requests or not

L138[14:57:30] <Kristo​bear38> the thing is i'm not in control over either the server that the bridge is working on (that's you) nor the server that the minecraft computer is running on (that's Izzy)

L139[15:02:44] <fingercomp> ah, I see

L140[15:12:35] <fingercomp> how often can you reproduce the problem? if, say, I run tcpdump on fomalhaut, do you think you could get it to manifest in under an hour?

L141[15:12:51] <fingercomp> since it looks like you're the only user right now

L142[15:17:03] <Kristo​bear38> happens often enough that I can probably reproduce it in a few minutes

L143[15:18:53] <fingercomp> all right, then I can start dumping in two minutes, and you can tell me when to stop

L144[15:19:10] <Kristo​bear38> ok let me boot up mc instance then

L145[15:21:16] <Kristo​bear38> did you start?

L146[15:21:28] <fingercomp> yeah, just now

L147[15:23:47] <Kristo​bear38> ok done

L148[15:24:22] <fingercomp> all right, let's see...

L149[15:24:54] <Kristo​bear38> last 70s or so should have the issue there, as I also wanted to catch the stem ping being sent

L150[15:25:20] <Kristo​bear38> also fyi I layered another protocol on top of stem which is cbor-encoded and that also has ping and pong

L151[15:32:03] <fingercomp> so I see a SYN/SYN+ACK/ACK handshake, all at 22:22:49Z

L152[15:32:36] <fingercomp> but the next packet fomalhaut receives is a regular stem message, at 22:22:51Z

L153[15:34:07] <Kristo​bear38> as in, stem message to a channel, where no subscription has happened?

L154[15:34:26] <fingercomp> yes, exactly

L155[15:35:36] <fingercomp> there's a flurry of encapsulated ping messages sent by the client, then fomalhaut sends a stem ping, and the client responds with a stem pong

L156[15:36:02] <Kristo​bear38> yeah you're looking at the right spot

L157[15:36:21] <fingercomp> since the client is able to hold a TCP conversation, it means it couldn't have sent a subscription message

L158[15:37:46] <fingercomp> it looks like it blindly writes to the socket here without checking what the `write` call returns https://gitlab.com/UnicornFreedom/stem/-/blob/master/stem.lua?ref_type=heads#L69

L159[15:38:23] <Kristo​bear38> that might be it

L160[15:42:31] <Kristo​bear38> it does indeed return 0 when this problem happens and 20 otherwise

L161[15:43:30] <Kristo​bear38> so this bit of code does need a send loop

L162[15:44:16] <Kristo​bear38> I think I can handle it from here, thanks a lot for your help

L163[15:45:02] <fingercomp> no problem, I'll tell Totoro to fix that

L164[15:53:53] <Kristo​bear38> looping until socket.write actually writes the data solved the issue

L167[17:23:39] <Izzy> how does one force garbage collection again?

L168[17:37:14] <Re​nno> In OC?

L169[17:37:21] <Re​nno> Os.sleep 10 times