IRC logs for #openttd on OFTC at 2025-07-24
⏴ go to previous day
00:13:49 *** asasnat has quit IRC (Quit: User went offline on Discord a while ago)
01:50:39 *** Wormnest has quit IRC (Quit: Leaving)
02:49:17 *** gnu_jj_ has joined #openttd
02:52:56 *** gnu_jj has quit IRC (Ping timeout: 480 seconds)
04:29:13 <ahyangyi> peter1138: Yes? An action D taints a local parameter if and only if the expression reads a baseset parameter, or an already tainted local parameter
04:46:45 <DorpsGek> - Update: Translations from eints (by translators)
06:02:42 *** Smedles has joined #openttd
06:33:41 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
06:36:21 *** ChanServ sets mode: +v tokai
07:25:32 *** tokai has quit IRC (Quit: c('~' )o)
08:11:05 <isosys> Hi, any idea why stun.openttd.org is failing only for n-ice servers?
08:11:05 <isosys> [2025-07-24 09:04:43] dbg: [net:9] NetworkClientConnectGame(): connection_string=+BWyCMXB
08:11:05 <isosys> [2025-07-24 09:04:43] dbg: [net:5] Closed UDP listeners
08:11:05 <isosys> [2025-07-24 09:04:43] dbg: [net:3] Initializing UDP listeners
08:11:05 <isosys> [2025-07-24 09:04:43] dbg: [net:9] status = CONNECTING
08:11:07 <isosys> [2025-07-24 09:04:43] dbg: [net:9] Stun::Connect(): family=10
08:11:07 <isosys> [2025-07-24 09:04:43] dbg: [net:9] Stun::Connect(): family=2
08:11:09 <isosys> [2025-07-24 09:04:43] dbg: [net:6] stun.openttd.org:3975 resolved in:
08:11:09 <isosys> [2025-07-24 09:04:43] dbg: [net:6] - 3.254.230.109:3975 (IPv4)
08:11:11 <isosys> [2025-07-24 09:04:43] dbg: [net:6] - 52.213.78.185:3975 (IPv4)
08:11:11 <isosys> [2025-07-24 09:04:43] dbg: [net:6] stun.openttd.org:3975 resolved in:
08:11:13 <isosys> [2025-07-24 09:04:43] dbg: [net:6] - [2a05:d018:5fd:4605:5234::1]:3975 (IPv6)
08:11:13 <isosys> [2025-07-24 09:04:43] dbg: [net:6] - [2a05:d018:5fd:4604:4420::1]:3975 (IPv6)
08:11:15 <isosys> [2025-07-24 09:04:43] dbg: [net:5] Attempting to connect to [2a05:d018:5fd:4605:5234::1]:3975 (IPv6)
08:11:15 <isosys> [2025-07-24 09:04:43] dbg: [net:1] Could not connect to [2a05:d018:5fd:4605:5234::1]:3975 (IPv6): Network is unreachable
08:11:16 <isosys> [2025-07-24 09:04:43] dbg: [net:5] Attempting to connect to 3.254.230.109:3975 (IPv4)
08:11:16 <isosys> [2025-07-24 09:04:43] dbg: [net:5] Attempting to connect to [2a05:d018:5fd:4604:4420::1]:3975 (IPv6)
08:11:19 <isosys> [2025-07-24 09:04:43] dbg: [net:1] Could not connect to [2a05:d018:5fd:4604:4420::1]:3975 (IPv6): Network is unreachable
08:11:19 <isosys> [2025-07-24 09:04:43] dbg: [net:9] Stun::OnFailure(): family=10
08:11:21 <isosys> [2025-07-24 09:04:43] dbg: [net:3] Connected to stun.openttd.org:3975
08:11:21 <isosys> [2025-07-24 09:04:43] dbg: [net:5] - using 3.254.230.109:3975 (IPv4)
08:11:23 <isosys> [2025-07-24 09:04:43] dbg: [net:9] Stun::OnConnect(): family=2
08:11:23 <isosys> [2025-07-24 09:05:03] dbg: [net:9] Client::OnFailure(): connection_string=+BWyCMXB
08:13:33 <LordAro> stun.openttd.org does not appear to be pinging (v4 or v6), though I'm not sure whether it's supposed to be
08:14:08 <LordAro> can't nc connect to it at all either, might just be down...
08:22:54 <isosys> if it was stun server that was down, every server would be affected. Right? But I can connect to others servers. Only n-ice seem to be unreachable (unless if stun is bypassed)
08:22:54 <isosys> I've already tried to restart without any success.
08:29:35 <Rubidium> if port-forwarding/firewalls are configured properly (at the server side) stun isn't needed. Potentially the n-ice servers are only running IPv6 (or IPv4 doesn't work for some reason), and then I doubt stun is going to work when you connect using IPv4. Although it should then fallback to turn I'd reckon, so maybe both stun and turn are down?
08:36:21 <truebrain> LordAro: It blocks pings and you would have to sent a package with nc before you get a response. So both are expected to me
08:36:32 <isosys> I can connect directly if I do it from the command line or from the game console. Only from the Server list it gets stuck when reaching for stun server
08:36:48 <truebrain> Service is reporting to work; so might be good to start the game and sees if it works for you LordAro 😉 😄
08:37:30 <truebrain> Anyway, the log dump above says it connected to the stun server
08:39:09 <truebrain> So not sure why we are concluding things are down 😄 I am confused 🙂
08:44:42 <peter1138> I read that more as "I don't know, maybe truebrain can help" rather than a conclusion :)
08:46:35 <truebrain> Haha, I read two devs saying "it might be down". But then I read the logs it was connected. So I was confused 😉
08:47:42 <Rubidium> okay, I shouldn't have said down but "not functioning as expected"
08:47:48 <truebrain> Anyway, all service are up and running. The logs above just show happy eyeballs at work. The client clearly doesn't have IPv6, and it falls back to IPv4 🙂
08:48:11 <Rubidium> for what it's worth, it's not only the n-ice servers
08:48:49 <truebrain> So I am missing something .. what isn't working?
08:49:15 <truebrain> For context, at work on mobile, getting a ping "things might be down" 😛 So I need a bit of help 😉
08:49:43 <truebrain> The client does connect. So what part of stun fails? Can you see that?
08:50:48 <Rubidium> I've tried 5 servers with 14.1. 1 doesn't use STUN and connects, 4 use STUN stay at '(1/6) Connecting...' until an error message 'Connection to the server timed out or was refused' pops up
08:51:42 <Rubidium> the client does connect to STUN, but after that it seems to be all crickets
08:52:25 <truebrain> Do the logs tell anything useful?
08:52:49 <Rubidium> no, it's exactly what isosys has
08:53:16 <Rubidium> at least with -dnet=9
08:53:49 <truebrain> That suggests the connection hand-off fails (where the STUN connection is reused for a connection to the server)
08:54:02 <truebrain> Would be curious to see what the server logs say
08:54:22 *** Flygon has quit IRC (Remote host closed the connection)
08:57:35 <truebrain> Normally if you setup a server yourself and connect to it yourself, given you don't have one of those annoying routers, it will also use stun (and works). Might give some insights what is happening
08:57:37 *** iSoSySt has joined #openttd
08:57:57 *** aquila3129 has joined #openttd
08:57:57 <aquila3129> I've also had issues joining through STUN today while trying to test a dedicated server
08:58:01 <truebrain> Backend is also not reporting any oddities
08:58:20 <aquila3129> Rubidium: same thing as this
09:00:31 <Rubidium> Stun::Stun, Stun::StunResult and Stun::CloseConnection are local changes I added to get some more information
09:01:35 <truebrain> So they both report their stun result .. after that the GC should be telling both which connection to use and to connect to what
09:01:40 <truebrain> That isn't happening?
09:02:50 <truebrain> Receive_GC_STUN_CONNECT
09:04:04 <truebrain> (I thought we made net=9 print every packet it was sending/receiving, but clearly I am wrong 😄 )
09:08:18 <Rubidium> we did for server/client, not so much for stun/turn/gc (at least in 14)
09:11:04 <Rubidium> can't help any further though, as I'm AFK for a few days ;(
09:13:24 <truebrain> No worries, tnx for being the remote hands 😄
09:15:37 <truebrain> I will just restart redis; that will restart the rest of the backend in a panic. Who knows that restores it. Although I rather find out what is actually not working.. but that will be a few days before I have that time 😛
09:16:13 <truebrain> Will take a few minutes for the system to recover and find all servers again
09:17:10 <LordAro> peter1138: burn everything
09:17:16 <reldred> Absolutely disgusting
09:17:27 <reldred> Glad I don’t have to deal with teams at my new work
09:18:14 <reldred> Hell I even asked whether I should install outlook and Webex on my personal phone (don’t have a work phone) and got told absolutely not 🤣
09:23:07 <truebrain> Can any of you check whether this fixed connecting over stun?
09:23:24 <truebrain> By any chance, that is 🙂
09:34:26 <isosys> Seems to be working again. Thanks! 🎉
09:39:37 <peter1138> Perfect, now we just have to wait another 5 years for it to happen again so we can figure out the problem :)
09:40:39 <truebrain> Basically ... how incredibly odd ...
09:41:13 <truebrain> This gives the impression redis was not relaying between the stun backend and coordinator backend .. but redis was build for that
09:41:28 <truebrain> Also very annoying there were no logs about that, and all reported healthy
09:41:41 <truebrain> But indeed: till next time this issue happens 😛
10:17:44 *** dh1 has quit IRC (Quit: My Mac has gone to sleep. ZZZzzz…)
10:27:12 <andythenorth> there's another recent one also
13:43:37 <LordAro> "All AI's were tested on a 1028x1028 map, with no GRF's"
13:46:58 <peter1138> Yeah, I guess a full test might want to include variations of NewGRF configurations too.
13:47:42 <peter1138> s/full/more complete/
14:37:54 *** Wormnest has joined #openttd
15:41:07 *** jond911 has joined #openttd
15:41:07 <jond911> Are there other bindings for AI development other than squirrel? Python, C? Rust?
15:42:16 <peter1138> No. Squirrel is the 'VM' embedded in the game.
15:42:46 <jond911> So no external library APIs? Any plans of adding ones?
15:46:10 <andythenorth> NoAI is what there is 🙂
15:47:47 <andythenorth> possibly the admin port can be used to drive an AI, I'm unclear about that
15:48:08 <jond911> Heh I see thanks guys I'll inspect how squirrel is integrated in and see if I can hack something in
15:49:24 <peter1138> No. Every so often someone thinks about that wasm thing but nothing further.
15:49:57 <peter1138> LordAro, how is it that some people can't figure out how to get a route from say Strava onto their cycle GPS?
15:50:01 <jond911> Wasm is a good idea clang had backend for it
15:52:12 <LordAro> peter1138: some units are clunky
15:52:23 <LordAro> i have to actually download the gpx to get it onto my lezyne
15:52:40 <LordAro> but garmin? yeah, just a lack of reading
15:55:10 <peter1138> Ah, they use a Hammerhead Karoo.
15:55:45 <peter1138> Still, there is integration for that too.
15:56:04 <LordAro> jond911: the trouble is linking against the entirety of clang isn't especially desirable
16:07:38 <peter1138> Why would you need to link to clang?
16:11:25 <LordAro> maybe i don't understand wasm, but isn't that the only way of getting a reasonable wasm VM?
16:23:08 <jond911> Ye I'm not sure either how it all connects that's why I was hoping there is some C api which can be easily integrated into
16:35:28 *** Smedles has joined #openttd
16:36:24 <peter1138> Nah, it needs to be sandboxed, no something you can do with native C.
16:36:49 <peter1138> (And it wouldn't be platform-independent)
16:38:36 <jond911> I see so perhaps I'll look into embedded python it can be easily sandboxed
16:39:15 <peter1138> Or just learn Squirrel ;-)
16:40:55 <rito12_51026> embedding in python seems not hard but time consuming
16:41:11 <jond911> The issue isn't me was hoping it will open options for other developers because of python popularity
16:59:50 <_glx_> squirrel integration is very deep inside openttd
17:02:27 <peter1138> truebrain, please change jobs or retire ;-)
17:12:49 *** Smedles has joined #openttd
17:21:17 *** Flygon has quit IRC (Remote host closed the connection)
17:21:38 <andythenorth> we didn't crowdfund TB yet? 😮
17:22:01 <peter1138> I think we can fund him for about 1 week.
17:24:39 <andythenorth> if we'd embedded the coin miner much earlier
17:24:47 <andythenorth> we'd be able to sell our BTC and fund TB
17:26:50 <kuhnovic> Time to sell our bag of TrueBit
17:30:52 *** tony_pixel has joined #openttd
17:30:52 <tony_pixel> andythenorth: me when I read this and look at my 99% cpu usage to run a 64x64 PAX map:
17:33:55 <tony_pixel> talltyler: hi Tyler, wanted to ask: are `return` in switches useless?
17:46:07 <truebrain> LordAro: You seem to understand WASM just fine 😉 There are two modes you can execute WASM in: "interpreter" mode, or "native" mode. The first is, ofc, slow. Although it is optimized, it executes pretty much like languages as Python.
17:46:07 <truebrain> The second uses a backend like clang (but there are other smaller ones) to compile the WASM module into native byte-code, which is ofc MUCH faster.
17:46:07 <truebrain> The first runs on any machine, the second only runs on the machine it is compiled for. So often you ship the compiler backend, to do it in a sort of Just-in-Time way.
17:46:12 <truebrain> Did that make sense? I think it did 😛
17:47:58 <tony_pixel> .NET "native" AOT compilation style
17:48:08 <tony_pixel> ship a compiler with the app
17:48:39 <truebrain> And for the visualization: WASM byte-code is about twice as slow as native compiled C++. So it is really fast.
17:48:56 <truebrain> the first mode of WASM was ... euh ..... much slower 😛
17:49:18 <truebrain> (comparison made with the TGP of OpenTTD)
17:50:46 <peter1138> Ah, so compile to WASM is fine, running WASM is "fine" but to improve performance means an extra back to make it native.
17:58:14 <truebrain> Yup! And this is what makes WASM so powerful and versatile
17:58:26 <truebrain> It can run "anywhere", and for selected platforms it can run "very fast"
18:55:15 <peter1138> Cannot build that station underneat the bridge because the pillars "block" the station.
18:57:46 <peter1138> Hmm, maybe that's another issue with this version.
19:01:26 <peter1138> Maybe the pillar test needs to take account of station continuation, but I'm not sure.
19:03:45 <rito12_51026> Can't for example bridge change its appearance depending on existence of station beneath it?
19:04:44 <cu-kai> interesting, can the pillars be present if the station design is simply a platform, and no canopy?
19:05:58 <peter1138> Not really my work, it's just ported from JGRPP, started by yiffgirl.
19:08:06 <peter1138> yiffgirl, happy for my to push to your PR, or shall I open a new one?
19:15:55 *** dh1 has quit IRC (Quit: My Mac has gone to sleep. ZZZzzz…)
19:20:03 <rito12_51026> Why is it impossible to make from first the second?
19:21:48 <yiffgirl> peter1138: should be set up to allow you to push to it directly, yeah!
19:31:01 <peter1138> Apparently not, ah well, no matter :)
19:59:40 <peter1138> Windows strikes again :p
20:06:06 <peter1138> I mean, there is a changelog PR, but LordAro took that over.
20:21:09 <talltyler> I still have PRs that I want to get into beta3 🙂
20:22:05 <peter1138> Why does this bitset not want to play...
20:29:02 <peter1138> Hmm, I guess all the other tables with enumbitsets are not constexpr.
20:29:18 <peter1138> MSVC being silly (or pedantic) somewhere.
20:29:52 <andythenorth> do we call code freeze at some point? 🙂
20:31:30 <talltyler> It's usually when we release RC1, right?
21:42:43 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
22:44:21 <yiffgirl> peter1138: oh no, I'm terribly sorry - I was visiting friends, if I'd been home I would have double checked
22:44:21 <yiffgirl> could have sworn I ticked the option though
22:52:04 <dh1> hard to reconcile "visiting friends" with being a person in this channel
23:05:50 <peter1138> yiffgirl, it's fine, improves integrity as I can't self-approve now ;)
23:06:15 <peter1138> (Not that I would.)
continue to next day ⏵
