IRC logs for #openttd on OFTC at 2023-09-18
⏴ go to previous day
01:45:21 *** Wormnest has quit IRC (Quit: Leaving)
02:47:25 *** debdog has quit IRC (Ping timeout: 480 seconds)
03:04:20 *** NGC3982 has quit IRC (Ping timeout: 480 seconds)
03:12:30 *** NGC3982 has joined #openttd
05:36:18 *** keikoz has quit IRC (Ping timeout: 480 seconds)
06:28:35 *** D-HUND is now known as debdog
07:26:21 <peter1139> Climate vs landscape...
08:52:47 <peter1139> I wonder, are their any sets that fiddle with cargo types without changing industries (much)?
08:54:23 <andythenorth> there must be something that fiddles with payments or something
08:54:34 <andythenorth> what does OGFX+ Industries do?
08:54:48 *** ahyangyi has joined #openttd
08:54:48 <ahyangyi> peter1139: Like adding YETI workers and Farming Supplies to the vanilla temperate economy?
08:55:01 <locosage> some novapolis event grfs do stuff with cargo payments
09:02:40 <peter1139> e.g. the lumber mill always produces in cargo slot 0. I'm not sure if the lumber mill could exist without WOOD being present.
09:02:53 <peter1139> (industry cargo slot 0 that is, not cargo type slot 0)
09:03:13 <andythenorth> should we make a test grf? 😛
09:08:46 <peter1139> Moving passengers/mail around might be interesting to test ;)
09:36:23 <peter1139> (There is at least one that moves mail.)
09:44:58 *** Flygon_ has joined #openttd
09:45:11 <_pruple> the industry grf spec has a flag that cuts trees to produce cargo, it doesn't specify that cargo has to be WOOD
09:50:21 <peter1139> Yes, that's not my point :)
09:51:56 *** Flygon has quit IRC (Ping timeout: 480 seconds)
09:52:51 <_pruple> what was the point? Maybe I should read up 😛
09:53:39 <peter1139> If, for some reason, WOOD doesn't exist but the default lumber mill does exist, and the cargo id it uses is invalid, will it still spawn on the map (I think so), and will it still produce cargo, even though it's invalid cargo (I think so)
09:54:23 <_pruple> does that question not apply to every industry? eg, can you have coal mines without COAL?
09:54:55 <ahyangyi> it does, I think. WOOD is just an example.
09:55:58 <peter1139> No. For most industry production, it loops through the produced-cargo types and that checks if the cargo type is valid before adding production.
09:56:04 <locosage> iirc I've seen some broken grfs produce invalid cargoes but didn't look much into it
09:56:19 <peter1139> But cargo production for lumber mill is handled separately based on that flag, and that does not check.
09:56:40 <peter1139> I can just add a check, so it's not a big issue. I just found it interesting (I'm weird)
09:56:41 <_pruple> I am with understanding
09:58:15 <ahyangyi> Lumber mills are weird
10:02:16 <peter1139> And I guess the question about the check, if the cargo is invalid, should it still cut down trees anyway.
10:02:41 <ahyangyi> It should only cut down invalid trees
10:02:46 <peter1139> And also, if an industry uses that flag with a different cargo type that isn't wood, should it still cut down trees...
10:03:29 <ahyangyi> I'm already envisioning some weird industry set where every primary industry is a lumber mill
10:03:38 <peter1139> There probably isn't much of a use-case for either of those two options to be honest, but it makes sense not to produce invalid cargo.
10:04:10 <peter1139> I suppose you could have a comb paper mill that uses locally sourced wood...
10:04:20 <peter1139> I was editting that...
10:05:59 *** nowthatsfunny has joined #openttd
10:05:59 <nowthatsfunny> i'm sure this has been brought up before, possibly, but would a cheat that allows any train to ride on any rail type be possible?
10:06:21 <nowthatsfunny> it would make converting a large network to a different type so, so, so much easier
10:12:12 <andythenorth> I have no idea if it works, not a thing I'd use
10:12:38 <debdog> nowthatsfunny: I am using NewGRF "Universal Rail Type" for that
10:13:33 <peter1139> I guess that works if you think of it before starting your game, and it doesn't conflict with other custom rail types.
10:13:36 *** brickblock19280 has joined #openttd
10:13:36 <brickblock19280> O don't think it does the exact thing you asked for but it accomplishes the same thing
10:15:51 <andythenorth> I remember how painful it was converting rail to monorail then maglev
10:15:55 <andythenorth> when I played vanilla game
10:16:20 <andythenorth> I think the solution is not to play the vanilla game
10:16:58 <locosage> would be nice to have some option for vanilla
10:17:50 <peter1139> I think a cheat is possible but I'm sure we've shot the idea down before for whatever...
10:19:12 <andythenorth> probably some angst about MP or something
10:19:13 <locosage> elrail option had to introduce a compatibility flag so might be some similar reason
10:22:47 <peter1139> Probably the main issue is breakage if you turn the cheat option off. So maybe disallow that.
10:23:49 <locosage> yeah, that's what the flag is for
10:24:28 <locosage> irc it allows all electric trains drive on regular rail even after you change the option back
10:31:26 <peter1139> ttps://mentourpilot.com/update-ural-airlines-plans-to-fly-a320-out-of-field/
11:16:26 <peter1139> Well, with no production and no acceptance... the industry window is a bit wonky :D
11:18:59 <peter1139> Should industries with no production/acceptance be disabled. Probably.
11:21:16 <ahyangyi> Well, isn't production/acceptance dynamic in some sets?
11:28:34 <peter1139> Sorry, I mean no valid producing/accepting cargo types.
11:28:57 <peter1139> Although maybe even that is dynamic.
11:30:19 <ahyangyi> Hmm, the only dynamic thing I can think of is `available_year`
11:30:42 <ahyangyi> Oh, no. I don't think it affects cargo types
11:31:09 <ahyangyi> it just makes it possible that some industries exist but cannot be effectively fed, or some cargo exist but not accepted anywhere
11:31:12 <peter1139> No, there is a callback to change input and output cargo types on build, so...
11:31:32 <ahyangyi> I think SPI 1.0 used that cb to make random cargos
11:31:45 <peter1139> But it doesn't change after build. Hmm.
11:31:51 <ahyangyi> Then they decided it was a bad idea and removed that from 1.1
11:32:16 <ahyangyi> How does the in-game flowchart look like in that case
11:32:24 <peter1139> Ah, and the cargo type has to be listed in the spec anyway, it seems.
11:32:52 <ahyangyi> So, effectively it's only *subsetting*?
11:33:45 <peter1139> Yes, you specify the possible types in the spec, then the callback can be used to limit those -- once, on industry creation.
11:46:24 <peter1139> What happens if industry-tile and industry acceptance differ... Hmm.
11:48:03 <andythenorth> "who would do that?"
11:50:24 <locosage> peter1139: probably not
11:50:34 <locosage> even if it's broken it's more clear when it still appears
11:51:18 <peter1139> Thinking about it, the main reason not to disable industries would be to prevent savegame problems -- where an industry that used to appear no longer appears.
11:51:42 <peter1139> So the best thing is to just not produce invalid cargo.
11:52:02 <peter1139> That already shouldn't be possible though (except for the sawmill.)
11:57:56 <peter1139> Ah, secondary industry production doesn't check for valid cargo types.
12:13:34 *** pickpacket6 has joined #openttd
12:16:50 *** pickpacket has quit IRC (Ping timeout: 480 seconds)
12:16:50 *** pickpacket6 is now known as pickpacket
13:01:25 <peter1139> And there's another production mechanism. Fun.
14:02:06 *** berndj-blackout has joined #openttd
14:02:55 *** berndj has quit IRC (Ping timeout: 480 seconds)
14:03:08 <peter1139> Nothing new, just one code-path I'd missed.
14:08:54 *** MikZyth has joined #openttd
15:04:21 <ahyangyi> andythenorth: Vanilla? 😛
15:04:56 <ahyangyi> I vaguely remember that vanilla industries tend to accept fractions of passengers
15:27:25 *** Extrems` has joined #openttd
15:31:05 *** Extrems has quit IRC (Ping timeout: 480 seconds)
15:31:05 *** Extrems` is now known as Extrems
16:01:48 *** Flygon_ has quit IRC (Quit: A toaster's basically a soldering iron designed to toast bread)
16:27:12 *** HerzogDeXtEr has joined #openttd
16:46:44 *** gelignite has joined #openttd
17:33:36 <Eddi|zuHause> some accept fractional values per tile, some concentrate all acceptance on very few tiles, while the others get nothing
17:49:58 <andythenorth> it's weird that we still have it
17:55:35 *** belajalilija has joined #openttd
17:55:35 <belajalilija> Is it only primary industries that accept passengers?
17:55:43 <belajalilija> I think it should be kept
18:04:50 *** gelignite has quit IRC (Quit: Stay safe!)
18:11:38 <truebrain> booo, peter1139 ruining the game!
18:12:07 <truebrain> no more invalid cargo
18:12:12 <truebrain> just the worst change ever
18:46:47 <frosch123> I don't understand your "exploit" discussion
18:46:57 <frosch123> the travel is summed as vector
18:47:04 <frosch123> if you take a detour or don't makes no difference
18:48:07 <frosch123> if you travel 20 tiles west, and then 20 tiles east, you end up at 0 tiles travelled, not 40
18:48:25 <andythenorth> all stations in OpenTTD are rail stations yes?
18:48:43 <andythenorth> roadstops are roadstops, docks are docks, and airports are airports
18:48:44 <frosch123> andythenorth: especially buoys
18:49:00 <andythenorth> stations have tiles, but don't
18:49:05 <andythenorth> roadstops just are tiles?
18:49:16 <andythenorth> docks have exactly 2 tiles?
18:49:25 <andythenorth> airports have tiles, or is that work-in-progress?
18:49:41 <peter1139> No. They are all tied to a Station object.
18:49:52 <andythenorth> isn't that a BaseStation?
18:49:54 <andythenorth> (context: I have to create python classes)
18:50:12 <peter1139> Oh, are you talking about NewGRF-wise?
18:50:39 <andythenorth> kinda, but also that's tended to change recently (more features)
18:50:47 <andythenorth> industries are easy industry -> layout -> tiles -> spritelayout
18:50:51 <andythenorth> no confusion at all
18:51:05 <frosch123> stations have sections
18:51:17 <frosch123> sections are a bit like articulated vehicles
18:51:33 <frosch123> a tile can ask its position along the whole platform, or its position along its section
18:51:49 <andythenorth> and the section is defined at tile placement time?
18:51:54 <andythenorth> or some other means?
18:52:22 <frosch123> sections are the tiles you build together at once, rectangular at the start, but can be split/made non-rectangular by removing tiles or overbuilding tiles
18:53:00 <andythenorth> ok so station -> section -> spritelayout?
18:53:04 <andythenorth> I see no tile entity
18:53:28 <frosch123> it's like articulated vehicles when you use the same id for all parts 🙂
18:54:05 <frosch123> you can build a station-in-game using different newgrf stations types
18:54:35 <andythenorth> so I need to figure out if sections are container structures, or just some other reference
18:54:45 <frosch123> the newgrf draws the tiles of its type, and can query positon within station, and position within tiles-built-at-same-time
18:54:54 <andythenorth> ok so it's more like cb36?
18:55:09 <andythenorth> or industry-tile-position-in-industry-layout var?
18:55:16 <truebrain> frosch123: Who's discussion? Comment is not all that clear what you talk about 🙂
18:55:26 <andythenorth> I don't have to rigidly define section -> tile IDs?
18:55:32 <andythenorth> [there appear to be no tile IDs]
18:58:59 <frosch123> truebrain: you/peter in the PR. is your "exploit" only about the previous PR #11281, or does the current PR #11283 add to it?
18:59:02 <andythenorth> hmm in FIRS I had to define `class GRFObject` due to `object` being rather reserved in python 😛
18:59:26 <andythenorth> are all rail stations always rail stations?
18:59:38 <andythenorth> trying to figure out if my class name is `Station` or `RailStation`
19:00:12 <truebrain> frosch123: Ah, you meant the comment. It is more 11281 than 11283. As loading now happens in the tile the wagon is on, you can place that favourable to you
19:01:11 <truebrain> That is why we all went through the motions of your offtopic, that the distance between producer and consumer should also play a factor
19:01:13 <peter1139> I thought there's some exploit to do with station walking, which this helps prevent, because moving the sign no long changes things. But I dunno any more.
19:01:42 <truebrain> peter1139: That is all fixed with 11281
19:01:57 <truebrain> As cargo is now loaded in the tile the wagon is on
19:02:10 <truebrain> 11283 fixed the free labour you get inside stations
19:02:24 <andythenorth> hmm someone hand me the tinfoil pls
19:02:24 <peter1139> Oh right, it's all moved on. Sorry.
19:02:27 <andythenorth> need to make a hat
19:02:49 <andythenorth> "Stations being named Stations just proves that OpenTTD is a train game and unbalanced"
19:03:35 <truebrain> peter1139: No need to be sorry 😄 there are a total of 3 exploits overlapping, making it difficult to keep things clear 😄
19:04:27 <andythenorth> peter1139: yes grf wiki 🙂 I am just being a tool though
19:04:48 <frosch123> andythenorth: i guess you need to distinguish tiles and variables
19:05:13 <frosch123> the "Station" is only the "station sign", it has variables and cargos, but no graphics
19:05:36 <frosch123> railstations and roadstations are tiles, they have graphics and a parent station
19:05:40 <peter1139> Truck stops are the weird one, they are not particularly a thing IRL :D
19:06:01 <andythenorth> peter1138: they are, but they belong to factories and crap
19:06:06 <frosch123> railstation tiles know their position within the whole station and within their section (tiles built at once)
19:07:22 <_pruple> it's not much different to industrial rail loading and unloading
19:07:37 <peter1139> There's no post office, and yet you pickup/deliver mail to a truck loading bay. Hmm.
19:08:26 <_pruple> maybe there's a post office?
19:09:36 <_pruple> I've worked in a few parcel smashing facilities, they're not readily distinguishable from other industrial or commercial buildings from the outside
19:09:54 <andythenorth> I made Irn Bru, just looked like a metal shed
19:10:22 <andythenorth> I supposed renaming that is out of scope after 15+ years?
19:10:35 <truebrain> frosch123: : regarding 11283, a check, you noticed it was a vector right, not a distance?
19:10:48 <truebrain> Otherwise I need to make it even more clear in the code 😛
19:12:11 <frosch123> yes, the code was clear, but the discussion was not :p i read the code as "vector" and the discussion as "distance", and thought I misunderstood the code :p
19:12:37 <truebrain> Distance is one of the two pieces of info in a vector, ofc 😛
19:14:34 <frosch123> truebrain: about the int16_t. if you use the exploit, the travel-in-station is actually the bigger vector than the travel-in-vehicle :p
19:15:03 <peter1139> 32767 is way larger than our maps.
19:15:40 <truebrain> Without adding 4 more bytes, there is actually no good solution here
19:15:47 <truebrain> One could overflow both methods
19:15:54 <truebrain> It just takes a lot of effort
19:16:18 <truebrain> As for JGRPP, JGR knows and will have to see what he does .. I would guess, make them int32s 🙂
19:16:18 <peter1139> How do you overflow int16_t with our map sizes?
19:16:29 <truebrain> 8x 4k map, let's gooooooo
19:17:03 <truebrain> As I said, lot of effort 🙂
19:17:33 <truebrain> Also the reason I didn't add code to prevent overflow
19:17:34 <andythenorth> ok I'm going to have a RailStation entity in python, even though grf has no such thing
19:17:45 <peter1139> It's not sub-tile position (*16) is it?
19:17:46 <truebrain> When someone goed through that much effort.. they can have it
19:18:09 <andythenorth> it's too weird to rely on `Station` being not `RoadStop` or `Dock` or `GRFObject` to know that it's a train stop
19:18:17 <andythenorth> but that means it won't match the spec
19:18:21 <truebrain> So frosch123 that means no changes in the PR are needed? 🙂
19:18:29 <peter1139> Also if it overflows, you just get paid nothing because it took you 1000 years to travel "0" tiles.
19:18:44 <truebrain> peter1139: True true
19:18:56 *** blindvt has joined #openttd
19:18:56 <truebrain> It first overflows ina negative way
19:19:24 <truebrain> Even with JGRPP .. you need a very very long track
19:25:26 <frosch123> truebrain: in "GetDistance" you wrote some essay, which I do not understand
19:25:37 <frosch123> wasn't the point of vectors, that you do not end up longer on detours?
19:26:45 <frosch123> truebrain: did you test with aircraft? i am not sure `Vehicle::tile` is valid for aircraft
19:27:16 <truebrain> Actually, I did not. Only RVs and trains. Will test
19:27:56 <truebrain> frosch123: You can still end up with a longer route. Especially when you use higher station spread and multiple transfer ststions
19:28:18 <truebrain> Say you unload at the top of every stations and load at the bottom
19:28:30 <truebrain> And the cargo needs to go up
19:28:41 <truebrain> You make more tiles than the actual distance
19:29:34 <truebrain> It is to not punish people that are more creative
19:29:58 <truebrain> I will make a better example in the comment 😄
19:31:10 <truebrain> frosch123: Do you also agree? 😛
19:32:05 <frosch123> longer distance is no "punishement", it's an "exploit"
19:32:57 <frosch123> you could have industries 20 tiles apart, and do 20 transfers inbetween which elonate the distance to 400 by transfering it further away, right?
19:33:12 <truebrain> Yes. But does that make more money?
19:33:21 <frosch123> together with the fast-train-reverse-travel you can make more money
19:33:37 <truebrain> So yeah, then we should keep that code 😄
19:34:15 <truebrain> I will improve the code (and test on aircrafts) tomorrow or so
19:34:54 <truebrain> Typing in mobile is shit 😄
19:35:47 <andythenorth> `RailStation` or `TrainStation`?
19:35:53 <andythenorth> has GPT learnt grf yet?
19:38:31 <ketsuban[d]> If GRFs weren't in the training data then it has no information for you. There's no mechanism for adding new information—they'd have to train a new model on an updated dataset for it to "learn" anything new.
19:39:40 *** nielsm has quit IRC (Ping timeout: 480 seconds)
19:43:53 <peter1139> Oops, getting side-tracked again :/
19:45:38 <andythenorth> yeah ok, so GPT says "it's appropriate to use the term "stations" to refer to rail stations."
19:45:53 <andythenorth> but I had to craft the question to get it to do that
19:46:14 <andythenorth> generally it keeps saying that it should be RailStation or TrainStation if it's rail (or train) related
19:46:20 <andythenorth> but it won't tell me which is correct
19:46:29 <andythenorth> it says approximately "there is no correct"
19:46:37 <andythenorth> why do I pay £20/month for GPT?
19:46:43 <andythenorth> I can be confused for free
19:47:20 <truebrain> peter1139: So just a normal day, I guess 😛
19:55:01 <truebrain> Inverting 0,0 .. silly frosch123 , that is still 0,0! 😄
19:55:28 <ahyangyi> andythenorth: Can GPT spell `recolour_sprite` correctly?
19:57:13 <andythenorth> "when working with OpenTTD NewGRF or similar projects, it's advisable to use "recolor_sprite" to ensure consistency and clarity in your code and documentation."
19:57:23 <andythenorth> why isn't GPT just in this discord?
19:57:34 <andythenorth> it would solve a lot of our problems
19:58:21 <truebrain> You are already here?
19:59:32 <frosch123> there is a fine line between humans using software, and software using humans
19:59:42 <locosage> truebrain: 11281 didn't fix station walking, it just now requires a vehicle to set loading tile first.
20:00:18 <ahyangyi> andythenorth: I doubt that "recolor" without the "u" could work 😛
20:01:54 <truebrain> Hihi, I forgot "eigenvector" is an actual word used in the English language
20:02:05 <truebrain> Still amazes me how borrowed words work in languages
20:02:26 <truebrain> Not many German words conquering the world like that 😛
20:04:24 <ahyangyi> kindergarten and zeitgeist
20:04:43 <ahyangyi> I guess they are better than direct translations "kids-garden" and "time-ghost"
20:05:17 <truebrain> And ofc spelling-nazi! (Too far?)
20:06:19 <andythenorth> eigenvector...been a while since I saw that
20:06:25 <andythenorth> philosophy degree 😛
20:07:06 <frosch123> it also depends on the target language. almost all technical terms in russian have a german origin
20:07:12 <andythenorth> ok so it's RoadStop, so I guess RailStation
20:07:15 <andythenorth> why not RailStop?
20:07:47 *** kuhnovic has joined #openttd
20:07:47 <kuhnovic> We use "fingerspitzengefühl" in Dutch, not sure how that managed to sneak in but I sure like it
20:08:25 <truebrain> kuhnovic: In what part of the country do they use that?!
20:08:37 <truebrain> On the German border? 😛
20:09:14 <kuhnovic> In the Randstad region as well
20:09:18 <frosch123> we include "frikandel" into words to make them sound dutch
20:09:23 <truebrain> Never ever heard it 😛
20:09:39 <truebrain> frosch123: You are such a frikandel
20:10:15 <kuhnovic> Funny how that's both an insult and a compliment
20:11:09 <truebrain> Funny how that works 😄
20:13:50 <ahyangyi> What's the difference between a port and a wharf?
20:15:36 <kuhnovic> Not gonna lie... it did make me chuckle 😛
20:17:42 <frosch123> ahyangyi: maybe wharf is a mistranslation of shipyard?
20:20:48 <andythenorth> ahyangyi: a port can have wharves, wharves do not have ports
20:21:30 <_glx_> I like how vasistas came in french
20:22:01 <frosch123> lol, with that spelling?
20:22:47 <frosch123> oh dear... i read it as "was ist das?" and thought you borrowed that :p
20:23:32 <_glx_> germans saying it before opening the door
20:23:38 <ahyangyi> andythenorth: Yeah, but we have both in Arctic Basic
20:25:02 <andythenorth> trying to figure out if Tom Scott went to my school
20:25:40 <blindvt> in dropdown_type.h there is a Height(uint width) function. I'm curious if this is a typo (s/width/height/) or just a leftover? I may be blind, but AFAICS neither DropDownListItem::Height() nor DropDownListIconItem::Height() actually need (resp. use) any parameter.
20:28:44 <andythenorth> andythenorth: no, went to the school my dad was a teacher at
20:31:49 <frosch123> blindvt: i assumed it's for linebreaks. assuming width "width", return the resulting "height"
20:32:43 <andythenorth> ahyangyi: English ran out of suitable words
20:33:11 <andythenorth> although Harbour, Haven, Anchorage, Quay
20:33:15 <andythenorth> could all have been candidates
20:35:48 <peter1139> Yes, it returns the height of a drop down item, given the provided width.
20:36:27 <peter1139> Probably `GetHeight()` would be a clearer name.
20:36:43 <andythenorth> ok RailStation it is then
20:37:26 <andythenorth> GPT wasn't able to make it's mind up
20:37:35 <andythenorth> so I've had to decide
20:37:54 <peter1139> It is possible that width is never used though, I'm not sure :D
20:38:17 <truebrain> andythenorth: But is it the right decision?
20:38:40 <andythenorth> can I find-and-replace later if it's wrong?
20:39:00 <blindvt> peter1139: i've remove both widths from height, let's see what parts of the universe collapse
20:39:29 <peter1139> There's two more in other files.
20:39:35 <truebrain> Okay, enough teasing andythenorth for no other reason than it being fun 😛
20:39:39 <peter1139> None of them actually use width.
20:40:08 <andythenorth> truebrain: moar emojis
20:41:28 <frosch123> i set "width" to unused in various places in #11306, i did not notice it was unused for all classes 🙂
20:41:40 <blindvt> peter1139: right. I've removed some, let's see if anything needing this is left
20:53:35 <andythenorth> ach RailStation isn't a python subclass of BaseStation
20:54:01 <andythenorth> maybe BaseStation can just be Station
20:54:27 <andythenorth> BaseStation is a grf thing though
21:01:13 *** keikoz has quit IRC (Ping timeout: 480 seconds)
21:03:48 <blindvt> one thing that nags me is that SDL2 seems to turn off the screensaver. Is this considered a feature or would it be ok to turn that off or have it configurable?
21:05:10 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
21:05:49 <blindvt> i had the game paused and left, but the screen kept on glowing over night which was unexpected..
21:08:51 <peter1139> Hmm, never really considered it, I would say we can just call SDL_EnableScreenSaver and not worry.
21:15:34 <andythenorth> oof it's not RailStation, it's RailStationTile maybe
21:18:16 <andythenorth> nah it's RailStation
21:18:52 <andythenorth> ok what can I rename BaseStation to?
21:19:04 <andythenorth> in python BaseFoo is for subclassing more Foo from
21:19:05 <blindvt> i think SDL per default kicks the screensaver out in its X event plumber loop, but i did not look closely if a plain SDL_EnableScreenSaver would make everything comfty dark after the usual time configured.
21:19:22 <blindvt> peter1139: i'll try that, thanks!
21:19:24 <andythenorth> but RailStation isn't a subclass of BaseStation
21:20:26 <locosage> what are BaseStation subclasses then
21:21:36 <andythenorth> are there subclasses?
21:21:51 <andythenorth> the way I am reading it, there are instances of e.g. RoadStop, and they are passed an instance of BaseStation
21:21:55 <andythenorth> but I might be reading wrong
21:22:51 <locosage> what are you even reading?
21:23:51 <andythenorth> newgrf_roadstop.cpp
21:24:14 <andythenorth> and base_station_base.h
21:24:47 <andythenorth> looks like BaseStation is an instance of a station, and then it has lists of tiles for train, road etc
21:25:04 <andythenorth> quite happy to be wrong here
21:25:27 <andythenorth> hmm it is `train_station` though
21:25:38 <andythenorth> I named mine `rail_station` because it's also `road_stop`
21:25:44 <andythenorth> might need to change that
21:26:05 <locosage> you classes for grf generation don't need to replicate whatever openttd is doing
21:26:11 <locosage> they just need to make sense
21:26:34 <andythenorth> but they have to follow the spec
21:26:50 <andythenorth> Horse is a mess because it has "consists" for articulated "consists"
21:27:03 <andythenorth> but a consist in OpenTTD is the full train with a powered head
21:27:45 <andythenorth> and a unit in Horse is a single vehicle, but IRL a unit is a multiply-articulated vehicle
21:27:54 <locosage> spec just defines them as separate features
21:28:05 <andythenorth> a unit has multiple units, and each part is a unit
21:28:06 <locosage> I quite like horse consist concept
21:28:11 <locosage> though maybe it's not the right word
21:28:22 <andythenorth> it's confusing with OpenTTD and IRL definition of consist
21:28:40 <andythenorth> and Horse uses vehicle and unit interchangeably in source
21:28:55 <andythenorth> and it uses consist where OpenTTD means vehicle
21:29:17 <locosage> openttd uses "vehicle" for parts too
21:29:38 <locosage> mu is not really a thing in a code iirc
21:29:48 <andythenorth> well multi-part vehicles are made up of vehicles
21:29:57 <andythenorth> so I guess grf has to use vehicles to make vehicles
21:30:06 <andythenorth> and then the articulated vehicle is a vehicle
21:30:23 <andythenorth> but an articulated vehicle can't be made of articulated vehicles, even though articulated vehicles are vehicles
21:30:46 <andythenorth> I worked with people who just name everything 'a', and work upwards
21:31:00 <andythenorth> `Class a(self, b, c, e, z)`
21:31:11 <andythenorth> it's much easier and faster TBH
21:31:30 <_jgr_> There is a middle ground between useless one letter names and endless agonising over what to call everything 😛
21:32:07 <locosage> I think I solved that at some point by calling parts "Unit"' and multi-part just "Train"
21:32:17 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
21:32:21 <andythenorth> I have learnt to trust that not knowing what to call something is a good sign for "I don't understand the spec" 🙂
21:32:33 <andythenorth> people arguing semantics is something different 🙂
21:32:41 <Eddi|zuHause> from the game's perspective, articulated vehicles are not one single vehicle, but multiple in a linked list
21:32:48 <andythenorth> if the name is long and unique, we can always change it with find-and-replace 😛
21:33:18 <andythenorth> ok so in OpenTTD, stations of different types can be joined, so BaseStation appears to be how we handle that?
21:33:30 <andythenorth> and we expose common varact2 on BaseStation?
21:34:21 <Eddi|zuHause> what is the exact problem we're trying to solve?
21:36:17 <andythenorth> I need to name python classes for a grf compile
21:37:41 <_glx_> in grf world there's not really base station, just common variables
21:38:01 <andythenorth> yeah, I named my basic entity BaseStation, but I think it's wrong
21:38:11 <Eddi|zuHause> the internal class structure tends to be much newer than the exposed newgrf interface
21:38:39 <andythenorth> I need a class to hold, e.g. ID, class etc, from which will be generated any or all of: rail station tile, road stop, object
21:38:45 <andythenorth> maybe in future, dock
21:39:02 <andythenorth> my class needs to hold the class 😛
21:39:21 <locosage> as I understand it what newgrf spec exposes is a mix of stations, base stations, and specs in openttd code
21:39:23 <andythenorth> the class will be an attribute on the class, but not a class attribute
21:40:23 <andythenorth> ok maybe I just call the class CHIPSStation
21:40:30 <andythenorth> but all caps in Python class names 😦
21:40:36 <andythenorth> not really PEP 8
21:41:21 <blindvt> i'd better not ask, i figure, but what's that thing with python in this context?
21:41:43 <blindvt> andythenorth, oh. ah. hm. eh!
21:42:18 <locosage> it's a general theme in newgrf actually, action0 works with specs but action2 with constructed entities
21:42:22 <locosage> not a great design tbh
21:42:23 <Eddi|zuHause> $someone wasn't content with writing nml code, so wants 5 layers of code generators on top
21:42:38 <andythenorth> StationHolder? StationContainer?
21:42:52 <andythenorth> think I got the idea from CETS 😛
21:43:04 <andythenorth> think how much less nml has to be written this way
21:43:08 <Eddi|zuHause> yes. but things... escalated... :p
21:43:22 <_glx_> locosage: then you have stations with action2 stuff in action0 (the spritelayouts)
21:43:25 <andythenorth> I mean...dP has now eliminated the nml step in grf-py
21:43:52 <locosage> _glx_: yeah, and that...
21:44:08 <andythenorth> what are station classes?
21:44:14 <locosage> at least for vehicles it makes some sense in nfo, it's nml that mixes up classes and objects
21:44:22 <andythenorth> I am wondering if I can just do `Class StationClass(object)`
21:44:29 <locosage> but stations are screwed from the get-go xD
21:44:48 <andythenorth> is a class a single tile, or a family of tiles?
21:45:15 <peter1139> A station class is just a label applied to a station type.
21:45:23 <andythenorth> lol, FIRS is easy: Industry, Cargo, Tile, GRFObject
21:45:24 <Eddi|zuHause> station spec is one of the oldest, and have considerably different handwriting than later specs like houses, industrytiles, objects, ...
21:45:49 <andythenorth> ok so classes are labels 🙂
21:45:57 <andythenorth> cargos have labels, but they're not classes
21:46:05 <_glx_> yeah later additions used a more sane spec for spritelayouts
21:46:28 <blindvt> Eddi|zuHause: :) Not sure if this is conceived fair, but that's far beyond the point in this context anyway i guess ;)
21:46:34 <_glx_> but at least station and all the other share a similar syntax for the layout
21:46:54 <andythenorth> from what I've seen so far, with nml hiding away station weirdness, the grf spec is all good
21:47:23 <andythenorth> the specific problem I have is because I'm going to generate 3 types of buildable grf object from one thing
21:47:39 <andythenorth> station, road stop, and object
21:47:48 <_glx_> yeah nml hides the weirdness and simplifies the use of multiple spritesets
21:48:02 <blindvt> andythenorth, s/GenericStation/CargoGIGO/
21:48:19 <_glx_> but it's still harder than industry tiles or objects
21:48:27 <andythenorth> it's just a bit...weird 🙂
21:49:04 <andythenorth> FoundationStation nearly rhymes 😛
21:49:14 <andythenorth> but conflates with station foundations
21:49:32 <Eddi|zuHause> blindvt: but that was one of the major blockers for NML implementation of stations
21:49:42 <andythenorth> RootStation might work
21:50:06 <blindvt> andythenorth, is this a train specific GIGO point, if it's generic then it's generic. Subclasses would filter by incoming vehicle type i guess and certainly a gazillion other constrains
21:50:56 <andythenorth> blindvt GIGO applies here how? 🙂
21:51:43 <blindvt> andythenorth, it accepts goods/passengers/stuff and might provide them
21:51:56 <andythenorth> think that's a different game 🙂
21:52:01 <andythenorth> RootStation, then RoadStop etc is a base class, which will be subclassed to make specific RoadStop variants
21:52:11 <andythenorth> variants not Variants
21:52:23 <_glx_> I think the first step should be a working version with different classes for different types, then you can try to extract the common stuff and subclass
21:52:38 <locosage> imo, just make RailStation, RoadStop, Airport. If there is something to put in a base class do Station or BaseStation, if not don't.
21:52:41 <blindvt> andythenorth, most likely. I should keep my mouth shut since i have zilch idea about openttd
21:53:04 <locosage> openttd is structured differently than newgrfs
21:53:26 <blindvt> locosage: ... and newgrf
21:56:49 <blindvt> s/constrains/constraints/; # my crap cheap thrash keyboard sucks. I want my Cherry back
21:57:39 <andythenorth> locosage: this would work, but then there would need to be foreign keys or string identifiers or something
21:58:03 <andythenorth> to know whether e.g. "bulk loader" is available as rail, road stop etc
21:58:13 <andythenorth> there would have to be some common key across all the implementations
21:59:12 <andythenorth> it's personal preference, but I'd rather define "bulk loader" and then "implements xyz" as flags
21:59:37 <andythenorth> it's also cleaner because there's a 1:1 mapping of spritesheet to entity
21:59:56 <_jgr_> I'm struggling to see why you'd need DB keys
22:00:38 <locosage> do you mean you want to define station graphics that can be both a rail station and a road stop?
22:00:38 <andythenorth> because there'd be a table of 'road stops' and a table of 'rail stations'
22:00:47 <andythenorth> locosage: yes, I am doing this
22:01:12 <_glx_> but the spritelayouts will be different anyway
22:01:25 <andythenorth> but that's just templating
22:01:36 <andythenorth> they'll always be different in the same way
22:02:12 <_jgr_> At some point you just have to start writing
22:02:22 <_jgr_> Otherwise you can go in circles forever without getting anywhere
22:02:27 <locosage> well, grf-py way of doing that would probably be making a generator class, say, "DualStation" that yields RailStation and RoadStop filling them with same properties.
22:02:30 <_jgr_> You can always refactor later
22:05:34 <andythenorth> locosage: or I could have made StationFactory 😉
22:06:05 <andythenorth> and little StationManifest or StationInstructions to hold config 😛
22:06:11 <andythenorth> StationOrders? 😛
22:07:04 <andythenorth> I have StationManager
22:07:11 <andythenorth> which is sort of a factory
22:08:51 <locosage> well, grf-py doesn't need managers for this, you just add stuff to newgrf
22:09:11 <locosage> it's just that "stuff" can be a generator class that yields more stuff xD
22:09:55 <locosage> not python generator I mean but a special class
22:11:19 <andythenorth> I thought you meant the iterable generator thing
22:11:22 <_glx_> yeah generating NFO is often simpler than generating NML
22:11:28 <andythenorth> I avoided learning how to use iterable generator thing 😛
22:12:21 <locosage> I probably can make it use python generators
22:12:39 <locosage> but so far it's just simply an object with `get_sprites` method
22:13:11 <locosage> and sprites totally not being the right word here xD
22:13:20 <andythenorth> it gets nfo sprites?
22:13:45 <locosage> it can return nfo sprites/actions or other generators
22:14:01 <_glx_> actions are sprites in NFO 🙂
22:14:04 <andythenorth> and then eventually the generator(s) yield sprites?
22:14:13 <locosage> _glx_: yeah, it's confusing af
22:14:21 <andythenorth> sprites are sprites
22:14:27 <andythenorth> sprites are realsprites
22:14:57 <locosage> sprites are realsprites with FD pseudo sprite
22:15:14 <andythenorth> and that offsets into another structure?
22:15:32 <_jgr_> It's usually easier to understand what the GRF gets parsed into inside OpenTTD, than the bizarre container format itself
22:16:04 <andythenorth> I don't find the sprites vs. realsprites thing confusing, once it's explained
22:16:24 <andythenorth> other game frameworks have sprites that contain code
22:16:39 <_glx_> the naming is logical when you know the history
22:17:11 <andythenorth> in flash and director, a sprite exists on a canvas or timeline, and may contain any of code, bitmap, vector, or more timelines
22:17:37 <locosage> in grf-py I'm thinking of using "action" for everything that goes into grf
22:17:42 <locosage> and sprite for graphics
22:24:00 <locosage> andythenorth: missed this, yes. sprites/actions, whatever can into binary
22:24:46 <andythenorth> really it's just `Station` but that sucks for search / replace
22:25:21 <andythenorth> I seriously considered that
22:25:29 <andythenorth> eh but FIRS just has Industry, Cargo, Tile, it's fine
22:25:56 <andythenorth> `Station` will do
22:26:05 <andythenorth> I will ignore the confusion with grf docs
22:27:47 <blindvt> TransferPoint is ambiguous WRT other logical parts where stuff is passed through i guess. StationBase, a base-class for Station'ish stations? At least it's greppable?
22:28:23 <andythenorth> yeah it's not a base class for subclasses though
22:28:26 <andythenorth> this is how all this started 🙂
22:28:47 <andythenorth> RailStation isn't a base class of BaseStation
22:29:02 <andythenorth> subclass / base class /s
22:29:26 <blindvt> meh so that's not it. I said i should just shut up when i have no idea about the context
22:31:04 <blindvt> so what is it supposed to be?
22:31:38 <peter1139> > Only a top club rider can sustain 20+mph on a flat surface for more than a few seconds, not 'many half decent cyclists'.
22:31:43 <peter1139> Uh, I mean... that's not really true :p
22:32:26 <andythenorth> unless the computers were wrong, even I used to be able to sustain 23-24mph on flat if no headwind
22:32:33 <andythenorth> and I was neither strong nor fit
22:33:39 <_glx_> the hardest part is to reach the speed
22:34:15 <locosage> that's almost 40km/h, kinda lot
22:34:17 <andythenorth> peter1139 is this on a website called "today's errors"
22:34:25 <locosage> iirc my best average was 27
22:34:45 <peter1139> andythenorth, Twitter, yes.
22:34:46 <locosage> though I probably never counted even as half-decent xD
22:35:30 <andythenorth> my trip computer says top speed I've done was about 56 mph
22:35:41 <andythenorth> which was off a reasonable hill
22:35:55 <andythenorth> I think it's overstated, and it was more like 48 mph
22:36:34 <andythenorth> ok, failure to adult again, bedtime was 1 hr ago
22:37:15 <locosage> my max was like 60-smth kmh
22:37:35 *** MikZyth has quit IRC (Quit: Konversation terminated!)
23:05:50 <blindvt> frosch123: i'm sorry but i didn't see your patch (the joy of not having a ML ;). Earlier today I've tweaked my tool to generate patches for these unused params.. i'll let it run over night, curious what it comes up with
23:15:53 <blindvt> just a thought. There's glibc version check fiddle required to support that, but i'm sure somebody did a cmake hunk for that already..
23:21:51 <blindvt> frosch123: like http://paste.debian.net/1292414/ The real challenge is to bring that further by looking if we can prune the parms altogether or if just using unnamed params. I see that you did that already, i guess for about everything.
23:23:14 <blindvt> frosch123: well, at least i had a bit of fun tweaking the tool, so let's see what's left after your cleanups are merged -- cheers! :-)
continue to next day ⏵
