IRC logs for #openttd on OFTC at 2023-09-20
β΄ go to previous day
00:04:11 *** muffindrake has joined #openttd
00:04:28 *** muffindrake has left #openttd
01:45:51 *** Extrems has joined #openttd
02:16:06 *** Wormnest has quit IRC (Quit: Leaving)
02:45:00 *** debdog has quit IRC (Ping timeout: 480 seconds)
05:10:51 *** blindvt has quit IRC (Quit: Leaving.)
05:38:55 *** nielsm has quit IRC (Ping timeout: 480 seconds)
05:46:23 *** keikoz has quit IRC (Ping timeout: 480 seconds)
06:13:29 *** blindvt has joined #openttd
06:13:59 *** blindvt is now known as Guest714
06:21:38 *** Guest714 has quit IRC (Ping timeout: 480 seconds)
06:27:15 <emperorjake> Since when do locomotives no longer preserve their name if they're not the first vehicle in the train? This is very unfortunate
06:32:17 *** blindvt1 has joined #openttd
06:39:00 *** blindvt1 is now known as blindvt
07:32:47 <ahyangyi> andythenorth: Perhaps we need nested classes π
07:47:02 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
08:01:09 *** Smedles has joined #openttd
09:12:59 <peter1139> Heh, my linter in VS Code now complains that "all parameters should be named"...
10:49:28 <LordAro> truebrain: i feel like i'm missing something regarding dns
10:53:39 <peter1139> Docker does magic to hide that away, but I suspect that only works on a single host.
10:55:55 <LordAro> i kinda want something that registers foo.domain.com to the (external) dns server, and updates in the event that the IP changes
10:56:09 <LordAro> google-fu hasn't revealed anything particularly promising thus far
10:57:07 <peter1139> Docker also works on the princple of hiding everything behind NAT, so you'd just use the IP of the host box.
10:57:11 <peter1139> This, imho, is shitty.
10:57:16 <peter1139> I don't know how podman does things though.
10:58:43 <LordAro> yeah, i want foo.domain.com to point to the (in this case) traefik instance, which can do all the actual routing within the cluster
11:15:47 <peter1139> Hmm, an array of 240 bools. std::bitset?
11:22:21 <peter1139> Ah, we already use bitset elsewhere, so that's fine.
13:22:11 <peter1139> Heh, it seems the depths of unused variables does not end :-)
13:44:15 *** D-HUND is now known as debdog
14:20:19 <peter1139> Hmm, calculators that support hexadecimal... with decimal points.
14:47:57 *** HerzogDeXtEr has joined #openttd
14:50:11 *** blindvt has quit IRC (Quit: Leaving.)
15:02:42 *** AngelLofey[m] has joined #openttd
16:16:09 *** Flygon has quit IRC (Quit: A toaster's basically a soldering iron designed to toast bread)
16:27:14 *** gelignite has joined #openttd
17:14:37 <pickpacket> Who maintains NML?
17:14:41 <pickpacket> "DEPRECATION: nml was installed using the legacy 'setup.py install' method, because a wheel could not be built for it. pip 23.1 will enforce this behaviour change."
17:18:33 <pickpacket> I don't know anything about wheel builds, but I know how to convert from setup.py to setup.cfg. I have no idea if that would help
17:20:09 <pickpacket> i don't get deprecation warnings when I install other things with setup.cfg at least π€·
17:20:23 <pickpacket> I'll put this on my to-do list
17:29:01 <alfagamma7> See I reported this a week before
17:29:18 <alfagamma7> Easy install is deprecated
17:38:32 <Eddi|zuHause> pickpacket: congratulations, you're now the resident expert on wheel conversion
17:54:38 <pickpacket> Eddi|zuHause: ... thanks...
18:34:48 *** gelignite has quit IRC (Quit: Stay safe!)
18:34:57 <pickpacket> huh... annoying...
18:38:57 *** nielsm has quit IRC (Remote host closed the connection)
18:38:58 <DorpsGek> - Update: Translations from eints (by translators)
18:45:15 <pickpacket> alright, here's where I'm at: I can translate the setup.py to a setup.cfg and make it installable that way, however the "entry_point" can't be the executable nmlc file because it needs to point to a function in a package. But nothing works if I point it to the nml.main:run function because all the imports are relative to the root directory. AND
18:45:15 <pickpacket> even if I can make that work the Makefile will be completely broken because it runs setup.py
18:46:32 <andythenorth> but we can rewrap that?
19:01:23 <Eddi|zuHause> changing the makefile... how hard could it be :p
19:09:42 <_glx_> I always use pip to install nml (`pip install -e .`)
19:12:41 *** Wormnest has joined #openttd
19:45:19 <andythenorth> hmm how do we do ground for stations? π
19:45:28 <andythenorth> there are two ground sprites for rail tiles?
19:47:37 <_glx_> there are 2 things for ground, one is using the automatic railtype offset, the other is preventing it
19:47:54 <frosch123> for track tiles you use the default normal rail sprites with/without snow. ottd will replace it with the railtype sprites
19:48:52 <andythenorth> ok so do I just leave `ground` blank in spritelayout?
19:49:08 <andythenorth> and then do the platforms in the `building` list?
19:49:13 <_glx_> cow pen example prevents the use of the offset
19:50:16 <_glx_> but for a rail tile you need to take care of the offset
19:50:45 <andythenorth> I tried reading the nfo, but it's all GRM etc
19:52:58 <andythenorth> ok so sprite: 2022 + a; // prevent railtype offset
19:53:03 <andythenorth> gives me a mud tile
19:53:06 <andythenorth> the classic tile
19:53:45 <_glx_> yeah it internally translates to `1 0002 Add offset to sprite, disable default usage of construction stage or railtype-offset `
19:53:57 <andythenorth> do I need to add this to docs, or did I look in the wrong place?
19:54:49 <frosch123> the stations specific spritelayout stuff is still undocumented
19:58:15 <pickpacket> andythenorth: too tired to figure out a working wrap right now, but I can do it. The makefile thoughβ¦ I think Iβll rewrite that to install with pip
20:04:45 <andythenorth> graph is nice π
20:05:02 <frosch123> better than the refitting one? :p
20:08:21 <_glx_> maybe I should add a second example with conversion of CHIPS magic tile (but using a more simple conversion than the one in my gist)
20:09:21 <truebrain> LordAro: ah, yes, sorry, DNS resolvement I had to solve a different way as I wasn't using consul π
20:09:26 <truebrain> that was the main downside
20:12:16 <andythenorth> _glx_: I am trying to understand the docs draft, but my brain is slow π
20:12:39 <andythenorth> can't be hard, it's just a spritenumber for ground?
20:13:25 <_glx_> just use GROUNDSPRITE_RAIL_X or GROUNDSPRITE_RAIL_Y
20:13:50 <frosch123> i think you always use the X sprites
20:13:59 <frosch123> the orientation is done by ottd
20:14:34 <_glx_> no you always need at least 2 spritelayouts
20:15:04 <peter1139> I remember the original implementation of stations had this so-called "ground relocation" thing, because of the weirdness of station tile layouts.
20:15:06 <_glx_> the automatic orientation is for the result of select_layout
20:15:18 <peter1139> It was totally broken, of course, because it was one global value.
20:22:19 <andythenorth> confirmed, GROUNDSPRITE_RAIL_X and GROUNDSPRITE_RAIL_Y work
20:28:37 <andythenorth> so otherwise I can just use any sprite refernce?
20:28:45 <andythenorth> for ground of my choice (on non-track tiles)?
20:29:11 <_glx_> using the `sprite + a` trick yes
20:29:30 <andythenorth> why is the offset there? Is it just how the syntax is?
20:31:04 <_glx_> it's to set the `Add offset to sprite, disable default usage of construction stage or railtype-offset` flag
20:31:45 <_glx_> maybe I should look into a special `NON_TRACK(sprite_id)` syntax to do that in a less hacky way
20:32:11 <frosch123> andythenorth: build your station with default monorail or maglev, and look how broken it is
20:32:29 <frosch123> for every sprite that looks broken, add the "+" thingie
20:33:43 <_glx_> and it needs to be a spritelayout parameter because nml is too smart π
20:33:54 <frosch123> the default station is a single station type, and shows different sprites for each of the 3 default railtypes
20:34:07 <frosch123> newgrf station spec assumes all stations provide different graphics for railtypes
20:34:14 *** Extrems has joined #openttd
20:34:22 <frosch123> since noone does that, you need to disable it
20:35:27 <_glx_> hmm yeah I'll probably do the `NON_TRACK(id)` thing
20:40:05 <frosch123> maybe it should be the default, except when using GROUND_RAIL_X :p
20:56:35 <LordAro> truebrain: pretty sure i should just be setting up the trifecta of nomad/consul/vault on all 3 nodes at this point
20:57:16 <LordAro> i don't think it's strictly necessary, but probably easier
20:58:07 <LordAro> reducing cost to an absolute minimum is only a small concern of mine
20:59:20 <peter1139> Curse these loops that loop over an array but also use the index for something else...
21:10:43 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
21:18:37 <andythenorth> first CHIPS 3 station tile works π
21:19:45 <peter1139> Don't forget about the different layouts...
21:20:03 <peter1139> Stations are quite flexible ;)
21:21:50 <andythenorth> one step at a time π
21:21:54 <andythenorth> wax on, wax off etc
21:22:13 <andythenorth> such efficient code reuse there
21:22:43 <andythenorth> so....new roadstop layouts? π
21:23:00 <peter1139> You've got 6 fixed layouts.
21:24:14 * andythenorth looking at lego baseplates
21:25:59 <andythenorth> one way roadstops, and vehicles drive in the centre of the tile? π
21:25:59 <peter1139> I had ideas about that but we fucked it all up.
21:26:32 <andythenorth> was it machine states?
21:28:02 <andythenorth> structure graphics will need some work eh
21:28:05 <andythenorth> to avoid clipping vehicles
21:29:45 *** keikoz has quit IRC (Ping timeout: 480 seconds)
21:29:54 <peter1139> You mean, getting the bounding boxes right?
21:37:47 <andythenorth> I mean the silo legs jut into the road area
21:38:01 <andythenorth> they're drawn to fit trains π
21:46:27 <andythenorth> wonder if roadstops has an equivalent magic tile
21:46:52 <andythenorth> trying to get it to draw road
21:47:09 <andythenorth> found a draw_mode flag bitmask, but can't get it to do anything
21:49:35 <andythenorth> sprite 1333 does something
21:49:46 <andythenorth> might be not terrain sensitive though
21:53:30 *** Wormnest has quit IRC (Quit: Leaving)
22:02:23 <andythenorth> is everything grf-ed now? π
22:02:58 <peter1139> There's a PR for bridges.
22:03:06 <peter1139> There's a ... something for docks.
22:11:08 <andythenorth> newgrf disasters? π
22:12:50 *** ckb has quit IRC (Ping timeout: 480 seconds)
22:13:09 <andythenorth> hmmmm.....might be doing this wrong π
22:13:21 <andythenorth> I guess there's an orientation var to check
22:18:27 *** ChanServ sets mode: +v tokai
22:25:03 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
23:06:58 *** NGC3982_ has quit IRC (Ping timeout: 480 seconds)
23:10:08 *** NGC3982 has joined #openttd
continue to next day β΅