IRC logs for #openttd on OFTC at 2025-03-17
⏴ go to previous day
00:11:49 <peter1138> And in 2025, it's "Rust will fix everything".
00:16:27 <reldred> time to recode openttd in rust? *ducks*
00:22:26 <peter1138> Good luck with that.
00:26:10 <reldred> I don't think you'll have to worry about that from me 😛
02:10:18 *** Wormnest has quit IRC (Quit: Leaving)
04:03:42 *** geizeskrank has quit IRC (Ping timeout: 480 seconds)
04:05:51 *** D-HUND has quit IRC (Ping timeout: 480 seconds)
04:07:11 *** geizeskrank has joined #openttd
04:46:56 <DorpsGek> - Update: Translations from eints (by translators)
05:22:36 *** WormnestAndroid has quit IRC (Remote host closed the connection)
05:22:41 *** WormnestAndroid has joined #openttd
07:17:34 <DorpsGek> - Add: summary for week 11 of 2025 (by OpenTTD Survey)
07:55:31 <peter1138> wensimehrp, have you tried to use it?
07:57:30 <wensimehrp> mostly because there isn't NML support at this point
08:00:33 <andythenorth> if it's just a var, can patch locally
08:00:38 <andythenorth> or just access by var number
08:09:57 <wensimehrp> > Because this variable is likely to be expensive to evaluate, especially if testing multiple tiles, it should probably only be used rarely i.e. during construction or periodically, rather than in a graphics chain.
08:09:57 <wensimehrp> What does periodically mean here?
08:14:48 <andythenorth> presumably in specific callbacks?
08:15:00 <andythenorth> feature-appropriate?
08:15:36 <wensimehrp> but how do you define not all the time?
08:16:02 <andythenorth> it's a moving target
08:16:17 <andythenorth> but a monthly callback per industry is "not all the time"
08:16:42 <andythenorth> but "on every tile for every redraw" would be
08:17:33 <andythenorth> related: delete large maps from the game, unleash pitchforks
08:18:01 <andythenorth> "all the time for every tile" on 256x256 is very different to 4kx4k
08:18:21 <peter1138> That ship sailed when it was added.
08:25:01 <andythenorth> on a 4k map, it's probably still sailing
08:25:07 <andythenorth> racking up negative profits 🙂
08:25:43 <andythenorth> oof, bad andythenorth
08:28:10 <reldred> speaking of things that are expensive to evaluate, I've got a plumber coming tomorrow 😢
08:28:43 <andythenorth> hmm action D var 0x13?
08:29:48 <andythenorth> make all my grfs disable if the map has more than 1,048,576 tiles
08:30:30 <andythenorth> yeah, but 1/16 of the resources
08:30:35 <andythenorth> compared to current vanilla limit
08:30:55 <reldred> what else am I going to do with this 5800x3d
08:31:51 <andythenorth> dunno, generate very large maps?
08:32:13 <andythenorth> then be surprised that AI and GS are possibly slow, possibly eat RAM, and possibly hit the opcode limit consistently?
08:32:44 <andythenorth> then be confused about why FIRS doesn't have payment rates suitable for using horses and carts over 3000 tile route, with some weird daylength adjustment?
08:32:54 <andythenorth> then ask for cargo aging property to be used?
08:37:44 <reldred> andythenorth: spoiler alert I don't run AI's or GS's if I can help it 😛
08:38:37 <reldred> and I only use horse and cart for like 8-9yrs tops, I don't expect them to make money they're just to help me not upset towns quite as much 😛
08:39:04 <reldred> andythenorth: dinner more like it
08:39:10 <reldred> I have pizza in the fridge
08:43:05 <peter1138> It's likely you can't use a feature name like that there.
08:43:37 <wensimehrp> but doesn't nmlc just internally translates FEAT_STATIONS to 4?
08:48:01 <wensimehrp> can't even test. immediately crashes
08:49:35 <peter1138> That seems like testing to me.
08:53:22 <wensimehrp> peter1138: you said that checking for badges is expensive. Is checking tile class and station id also expensive?
08:53:26 <reldred> this is the do that people don't think it be like
08:57:27 <reldred> i've had that before testing new features
08:57:41 <reldred> sometimes you feed it a .grf and it just poops itself
09:01:27 <peter1138> wensimehrp, any test is expensive if you do it enough. Without any feedback, I have no idea if this is actually prohibitively expensive.
09:02:48 <peter1138> That this requires 3 parameters is expensive by default, imho :D
09:03:47 <peter1138> Also means adding an NML-way to do this needs to be careful, as there are orders-of-operations that minimise the amount of temp register setting that might be tricky to arrange in NML.
09:04:45 <peter1138> You can set the two temp registers once and test a few neighbours.
09:05:04 <peter1138> Or you can test tile one repeatedly for different badges, then test another tile, then another.
09:05:18 <peter1138> The latter will incur more operations.
09:05:48 <peter1138> At least badges are static ;)
09:06:30 <wensimehrp> I'm thinking this is (0, -1)
09:07:45 <peter1138> I guess `signed_tile_offset` isn't exposed as something you can use directly :(
09:08:50 <wensimehrp> is it `0xF0` or `0xE0` for (0, -1)?
09:11:09 <peter1138> 0xF0 or 0x0F I guess.
09:11:30 <peter1138> For stations the axis of the station affects things as well.
09:11:52 <wensimehrp> does it automatically flips for stations?
09:12:33 <peter1138> All the "nearby tile" type stuff does that.
09:17:54 <peter1138> Hmm, I guess it needs rebasing too, commit checker fails :)
09:22:13 <wensimehrp> peter1138: I cannot check the var0x79 in the station debug panel
09:22:55 <peter1138> As it requires parameters and temp registers, it wouldn't be particularly useful.
09:27:13 <peter1138> If you fancy trying it, add `NIV(0x79, "badge test of nearby tile"),` in `src/table/newgrf_debug_data.h` at the end of the `_niv_stations` section (around line 130)
09:34:22 <wensimehrp> btw, nmlc is not translating `"station/wins/right_yard"` to `1`.
09:53:15 <peter1138> I think you can do `badgetype("station/wins/right_yard")`, not sure.
09:53:39 <peter1138> Should be similar to how it works for cargotypes, railtypes and roadtypes.
11:23:30 <cu-kai> _jgr_: i have ended up in the rabbit-hole that is your openttd-patches and i had a question about the multi-aspect signalling (i have the GRF installed). i cannot see a way to enable this for block signals, they are, by default, either red or green. works fine for path signals though.
11:24:01 <cu-kai> realistic train braking is also enabled.
11:25:07 <mnhebi> Welcome to the wonderland...
11:25:32 <mnhebi> The hole is unending and filled with joy.
11:26:45 <peter1138> Block signals are hidden by default for a reason, they are legacy and you don't need to use them.
11:27:17 <cu-kai> peter1138: thanks, but that wasn't the question
11:27:45 <mnhebi> And yeah multi aspects are for path signals only
11:28:14 <cu-kai> that's interesting - the JGR wiki suggests otherwise
11:31:15 <cu-kai> on the topic of block signals being hidden by default, i recall back in the day there were performance reasons for using block signals over path signals :)
11:31:21 <cu-kai> old habits die hard i guess
11:31:52 <mnhebi> Ahhh I don't use blocks that much these days so I forgot they work on them
11:32:39 <emperorjake> cu-kai: Multi aspect is designed to be used with realistic braking, which is incompatible with presignals, and normal block signals behave like path signals.
11:34:26 <emperorjake> Block signals are meant to behave like simple automatic signals so they don't need to have extra aspects
11:34:45 <cu-kai> sure, but the question was in the context of the _jgr_'s mod
11:35:09 <emperorjake> Yes, that's what I meant
11:35:24 <cu-kai> i see, so the wiki is outdated?
12:00:01 <_jgr_> cu-kai: No, block signals support extra aspects in the same way.
12:00:19 <cu-kai> thx, do you know why they do not appear to work for me?
12:01:28 <_jgr_> Are you using another GRF which provides per-railtype signals?
12:04:47 <cu-kai> i even tried with _just_ the multi aspect signal GRF to see if it was some incompatibility
12:10:57 <cu-kai> (sorry, i misread your message, i was _not_ using another GRF which implemented this)
12:12:32 *** WormnestAndroid has quit IRC (Remote host closed the connection)
12:12:38 *** WormnestAndroid has joined #openttd
12:16:50 <_jgr_> I can take a closer look after work, but in the meantime it may be worth checking what GRF parameters you've set, and if you've turned on the setting to disable GRF signal graphics
12:32:57 *** WormnestAndroid has quit IRC (Remote host closed the connection)
12:33:00 *** WormnestAndroid has joined #openttd
13:43:07 *** kuka_lie has joined #openttd
13:48:38 *** kuka_lie has quit IRC (Quit: Lost terminal)
13:49:08 *** kuka_lie has joined #openttd
13:56:10 *** Flygon has quit IRC (Remote host closed the connection)
14:16:50 *** WormnestAndroid has quit IRC (Remote host closed the connection)
14:17:01 *** WormnestAndroid has joined #openttd
14:57:13 <peter1138> Tile-positioned errors might be an issue if there's no viewport currently covering the tile, though.
14:58:46 <_zephyris> Yup, can only work for errors arising from clicking on a tile, and has to pop up at the cursor-selected tile, and, in factorio, there is (essentially) never a window/UI overlaying the game world.
15:05:07 *** Wormnest has joined #openttd
15:33:47 <xarick> this error is so weird
15:34:06 <xarick> and i'm not building a tunnel
15:34:13 <xarick> im building a rail track
15:48:08 <peter1138> Hmm, I have no idea how errors are translated to ScriptErrors.
15:48:55 <peter1138> RegisterErrorMap. Hmm.
15:50:44 <peter1138> > ai_tunnel.hpp.sq: ScriptError::RegisterErrorMap(STR_ERROR_CAN_T_BUILD_ON_WATER, ScriptTunnel::ERR_TUNNEL_CANNOT_BUILD_ON_WATER);
15:51:15 <peter1138> Seems a bit generic.
15:51:33 <peter1138> Are we comment-based metadataing?
15:51:42 <peter1138> ERR_TUNNEL_CANNOT_BUILD_ON_WATER, // [STR_ERROR_CAN_T_BUILD_ON_WATER]
16:01:23 *** kuka_lie has quit IRC (Read error: Connection reset by peer)
16:04:38 <xarick> oh, if this is going to be fixed... erm... I'm afraid it's going to break some AIs that expect the wrong error
16:26:38 <_glx_> clearly weird error mapping
16:29:24 <_glx_> and it's there since first NoAI version
16:39:27 *** HerzogDeXtEr has joined #openttd
16:48:31 <peter1138> Let me guess, everyone knows this issue?
16:51:27 *** simonmb6044 has joined #openttd
16:51:27 <simonmb6044> How unfeasible is it to allow multiple newObjects be placed on the same tile? From gameplay I know that it's possible to put newObjects over things like rivers, does that mean that each map-cell can contain multiple objects?
17:03:36 <peter1138> No. If you place an object over a river, it replaces the river.
17:05:20 <simonmb6044> I'm pretty sure at least some don't, because you can delete the object and the river is still underneath (or reappears?)
17:05:36 <simonmb6044> maybe that's a jgr-feature?
17:06:07 <simonmb6044> for example, one of the auz sets has fake bridges, that you can build on rivers, and when you delete them, the river is still there
17:07:19 <peter1138> When you remove the object, the original water type (and only that) can be put back. This is a special case for objects.
17:07:43 <peter1138> So it's not that the river is still there, it's that the object is no longer there, and it's now a river again.
17:29:13 <simonmb6044> ```* Look at docs/landscape.html for the exact meaning of the members.````
17:55:07 <_jgr_> Yes, 0 is the least significant bit, 7/F is the most significant bit
17:55:28 <Rubidium> yes, it's in bits. And it's using hexadecimals as when you use decimals, the columns and header would not be aligned making it even harder to work with
17:55:58 <Rubidium> though that page gives a broad overview, the non-grid version will have more details about the exact meaning of the bits
18:05:35 <simonmb6044> I understand the formatting concerns, but at first glance it gives the impression that each digit would give you 16 values instead of 2. Maybe it would be a good idea to change it to 1s and 0s wrapped in <code> tags?
18:07:31 <_jgr_> Above the table there is an explanation for what the fields are and their sizes...
18:25:46 <simonmb6044> I agree it's a little bit nitpicky, but especially because it's an informative document, if it could be more clear, maybe it should be
18:26:02 <simonmb6044> I'm not asking for some of you to pick this up, but maybe it can be my first PR
18:31:02 <_jgr_> "1s and 0s wrapped in <code> tags" does not really make sense, so I do not quite understand what it is that you think is wrong with it
18:38:31 <simonmb6044> ```bits 7654 3210 7654 3210 7654 3210 FEDC BA98 7654 3210 7654 3210 7654 3210 7654 3210 7654 3210 7654 3210 FEDC BA98 7654 3210```
18:38:31 <simonmb6044> It seems to me this row could be better represented as ```bits 10011010 10011010 etc...```
18:38:31 <simonmb6044> and because rubidium talked about the alignment of the columns and header, i thought hexadecimals were picked to improve the table layout but i see the individual spans already have some monospace font applied.
18:38:58 <simonmb6044> so disregard that 'wrapped in <code> tags' bit
18:40:40 <_jgr_> What is the reader supposed to do with `10011010 10011010`?
18:41:40 *** ChanServ sets mode: +v tokai
18:42:44 <_jgr_> If your objection is just that it's hexadecimal, you could use RFC-style bit column headers, but that would make your table twice as wide
18:48:36 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
18:49:45 *** gelignite has joined #openttd
18:50:34 <xarick> it's been years since i touched the road pathfinder, and now I got an issue
18:55:59 <xarick> trade performance for a fix? hmm
19:22:02 <peter1138> Ring ring ring ring ring ring ring ring bananaphone, do be do be do.
19:28:43 <_zephyris> Ting tong ting, tonanaphone!
19:32:35 <xarick> oh, there was a minimap bug I reported yesterday or so
19:37:35 <peter1138> If it's not on github, you didn't report it.
19:57:55 <xarick> I wonder if I could improve this check
19:58:55 <xarick> scan_path finding a tile
19:59:02 <xarick> and immediately discards
19:59:16 <xarick> as a collision, it's not totally right
20:00:33 <frosch123> Can we delete case/gender/plural support for newgrf?
20:00:33 <frosch123> I keep on encountering corner cases which do not work
20:08:57 <peter1138[d]> frosch123: Did it ever work out did my recent changes break it....
20:11:28 <frosch123> Some things may have worked in 2010 and got broken in 2012 or so
20:13:10 <frosch123> I have like 4 bug fixes for ottd. And 2 for nml. And 1 I do not even know what the correct behaviour would be
20:21:54 <xarick> this is what I think is bugged: toggling Show height on and off for Cargo flow panel
20:24:37 <xarick> maybe it's not bugged... I can't tell
20:29:19 *** kuka_lie has joined #openttd
20:36:01 <peter1138> Hmm, St Patrick's Day, am I meant to be at the pub?
20:37:54 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
21:14:59 *** gelignite has quit IRC (Quit: Stay safe!)
21:29:25 *** nielsm has quit IRC (Ping timeout: 480 seconds)
21:47:56 *** WormnestAndroid has quit IRC (Remote host closed the connection)
21:47:59 *** WormnestAndroid has joined #openttd
21:50:23 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
22:02:16 *** kuka_lie has quit IRC (Quit: Lost terminal)
22:32:16 <Rubidium> oh joy... coverity has an issue when you have zero outstanding defects. It now shows the previous state of the top defects, defects per component and defects per category. That's unfair :(
22:33:19 <xarick> citizens are still celebrating
22:37:56 *** keikoz has quit IRC (Ping timeout: 480 seconds)
22:53:22 *** akimoto has joined #openttd
23:13:48 <truebrain> haha; nice Rubidium 😄
23:49:54 <truebrain> `It looks like you are writing in regarding usage of SteamCMD to login. I am sorry but we do not provide support for usage of SteamCMD. We are not aware of any issues with the type of use your doing, or SteamCMD commands/functionality. For help/troubleshooting with SteamCMD you will want to refer to the official documentation and other Communities in Steam and outside of Steam for help. I am sorry
23:49:54 <truebrain> that we are not able to be of more assistance with this. `
23:50:02 <truebrain> so basically I waited 48h to get a reply from ChatGPT
23:51:07 <truebrain> feels like if they paid a bit more for the AI subscription, they could answer people faster 😛
23:51:37 <digitalfox> Wow and I thought my ISP answers were bad 🥸
23:52:01 <truebrain> Steam support is not known for their service 😛
23:52:20 <truebrain> It is always hoping the right person gets the right email at the right time, to get some form of useful reply
23:53:10 <digitalfox> Well same with my ISP
23:53:10 <digitalfox> I call multiple times until I get that awesome person who actually cares 🙃
23:53:38 <truebrain> Weirdly enough, often it is enough if we make some public post of our issues with one provider or the other
23:53:44 <truebrain> really weird how that somehow does solve issues
23:54:51 <digitalfox> So… Wait and see if magic happens? 😇
continue to next day ⏵
