IRC logs for #openttd on OFTC at 2026-04-08
⏴ go to previous day
00:17:51 *** WormnestAndroid has quit IRC (Remote host closed the connection)
00:18:04 *** WormnestAndroid has joined #openttd
01:01:09 *** MinchinWeb[m] has quit IRC (Ping timeout: 480 seconds)
01:01:34 *** MinchinWeb[m] has joined #openttd
01:06:33 *** MinchinWeb[m] has quit IRC (Read error: Connection reset by peer)
01:08:00 *** MinchinWeb[m] has joined #openttd
01:32:37 *** MinchinWeb[m] has quit IRC (Read error: Connection reset by peer)
01:32:59 *** MinchinWeb[m] has joined #openttd
01:49:14 *** MinchinWeb[m] has quit IRC (Ping timeout: 480 seconds)
01:49:34 *** MinchinWeb[m] has joined #openttd
02:03:04 *** herms2 has quit IRC (Quit: bye)
02:04:46 *** Wormnest has quit IRC (Quit: Leaving)
02:40:01 *** ChanServ sets mode: +v tokai
02:46:56 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
03:13:42 *** Philemon has joined #openttd
03:17:01 *** Phileman has quit IRC (Ping timeout: 480 seconds)
05:27:42 *** u0_a351 has left #openttd (rirc v0.1.7)
06:04:20 <hazardaj_nombroj> oh, right, en_GB is the base language
06:39:01 *** MinchinWeb[m] has quit IRC (Remote host closed the connection)
06:39:16 *** MinchinWeb[m] has joined #openttd
09:21:25 <x1vi_zi> Saw this in the afternoon
09:21:25 <x1vi_zi> It is from naukri.com a huge job site in India
09:21:55 <x1vi_zi> Just getting average job offers
09:25:37 <mmtunligit> this is Discord channel #off-topic-1 please
09:26:53 <LordAro> i'm not hugely fussed if it's quiet
09:27:13 <LordAro> though something that could actually generate conversation would be preferred
09:29:18 <peter1138> Oh no, Amazon will stop supporting devices that are 14+ years old.
09:34:59 <andythenorth> we could discuss lunch?
09:39:01 <peter1138> Let's discuss trees.
09:39:14 <peter1138> Tree height level ranges?
09:39:37 <_zephyris> Treeless high elevations
09:40:49 <peter1138> So far my work on trees has just been reorganising it to remove hardcoded values, adding things like climate mask and tropical zone mask (hence the scoped enum PR yesterday)
10:00:47 <reldred> Yes, height level ranges would be nice. Can force certain varieties to grow sea level, others above the known range of where snowline will be, though I suspect ‘put me around the snowline’ or ‘never put me above snowline’ might be more useful there
10:01:58 <_zephyris> (but I would like tree newgrf...)
10:03:07 <_zephyris> Kinda rekated to limitations of random object scattering
10:03:37 <_zephyris> Would be nice to have smarter tree distribution and object distribution algorithms, which newgrfs can use for their behaviour
10:09:53 <kuhnovic> And don't forget about my tree automaton thing, although it might be a bit radical and needs work (#15140)
10:13:41 <andythenorth> I always turn tree visbility off 😛
10:13:52 <andythenorth> so my opinion is invalid
10:15:18 <andythenorth> I did do a thing where trees are hidden around the cursor
10:17:32 <peter1138> Oh, the tree gui just used random tree sprites, ignoring the tile layouts and growth.
10:18:01 <peter1138> My new code always uses the same growth stage, and therefore it doesn't match original behaviour.
10:20:27 <mmtunligit> if were extending trees to newgrfs we should redo the tree gui to a pickerwindow
10:21:03 <mmtunligit> at least when you have tree newgrfs
10:21:10 <mmtunligit> like how stations work
10:21:27 <peter1138> Sure, step 1 is just a minor adjustment to the UI to make it use the available trees instead of hardcoded numbers./
10:22:12 <mmtunligit> yeah i spoke faster than i thought
10:22:34 <peter1138> And there's always more than 1 tree, so it always needs the picker window.
10:23:44 <mmtunligit> i mean you can still keep the historical style even without it, and just swap in the picker when you have newgrf trees, thats how stations work
10:24:31 <peter1138> Hmm, should switching picker collection (at least away from Default) automatically select the "Saved" button (at least if it's not an empty collection?
10:24:51 <peter1138> mmtunligit, yes but stations only does that because by default there is only 1 station type.
10:24:57 <peter1138> Trees always have more than 1.
10:26:00 <mmtunligit> peter1138: i dont think so, it should keep whatever settings you currently have for what types to show. if you want to add things to already created collections itd be annoying to have it switch everytime
10:27:18 <mmtunligit> and if youve made a new one thatd count as "switching off" as well which you really wouldnt want to change to saved items
10:28:38 <andythenorth> are fields also trees? 😛
10:28:51 <mmtunligit> i also want to change the pickerwindow to use HandlePlacePushButton instead of the weird bespoke thing we have now because we have the function, may as well use it, and the current behavior doesnt match any other placement button behavior in the game
10:29:18 <mmtunligit> hope i can get around to that soonish
10:29:58 <andythenorth> oh yeah they're an animation loop, not a growth stage
10:30:03 <andythenorth> silly andythenorth
10:30:11 <andythenorth> "animal, vegetable or mineral"
10:38:53 <peter1138> Not sure I'd want to change the actual behaviour, it matches things like depot placement, original road stop placement, etc.
10:39:10 <peter1138> It's just trees that are different :)
10:41:32 <mmtunligit> i gues yeah, its just felt weird to me
11:01:09 <peter1138> You can bet there'd be pitchforks ;)
11:04:14 *** Chompy has quit IRC (Ping timeout: 480 seconds)
11:04:30 <mmtunligit> i do need to actually fix the randomization button in my PR though, because it does feel wrong on account of it not being in the matrix widget
11:10:50 <Borg> jaj! openttdcoop wiki is back :)
11:46:49 *** toktik has quit IRC (Remote host closed the connection)
12:00:14 *** rito12 has quit IRC (Read error: Connection reset by peer)
12:00:16 *** u0_a351 has joined #openttd
12:01:31 *** rircA13D7 has joined #openttd
12:03:31 *** rircA13D7 has quit IRC (Read error: Connection reset by peer)
12:03:32 *** rircFA827 has joined #openttd
12:03:49 *** u0_a351 has quit IRC (Read error: Connection reset by peer)
12:06:47 *** rircFA827 is now known as rito12
12:11:33 <Borg> hah.. goto's are bad coding practice.. heh
12:11:57 <Borg> talltyler: are you also like to have if { blabla; return;} else {...} constructs too? ;)
12:13:28 *** toktik has quit IRC (Remote host closed the connection)
12:17:17 *** u0_a351 has joined #openttd
12:17:17 *** rito12 has quit IRC (Read error: Connection reset by peer)
12:17:39 *** u0_a351 has quit IRC (Read error: Connection reset by peer)
12:17:40 *** u0_a351 has joined #openttd
12:46:37 <rito12_13[d]> reldred: GitHub has too many buttons?
12:48:33 *** jfkuayue has joined #openttd
12:48:47 <Borg> comment from zaki* is interesting
12:48:52 <Borg> I actually noticed it myself too
12:49:05 <Borg> and.. maybe I will try to bite it :)
12:51:22 <Borg> inline int ReservationCost(Node &n, TileIndex tile, Trackdir trackdir, int skipped)
12:51:29 <Borg> I wonder if this is problematic function
13:15:09 <Borg> okey. Ill take a bite on it.. lets add DEBUG() there
13:26:18 <Borg> I know how to trigger this
13:28:56 <Borg> train 86 could take free platform, but it waits to platform to become free..
13:29:09 <Borg> with.. is good or bad depending on situation on tracks :D hard to bite
13:30:18 <Borg> train waits forever because the train blocking it.. is stopped
13:30:22 <Borg> so now I can use save for debuging
13:32:39 <Borg> but.. its not ReservationCost..
13:59:53 <talltyler> peter1138: I figured out the Disconnecting Train issue, I took your suggested change to afterload and missed the check of IsPrimaryVehicle(). 🙂
14:01:02 <peter1138> It was just the vehicle type to be removed :)
14:03:22 <Borg> okey.. when I reverse the train blocking it.. and move it just slighty backward so he will reserve back station tile
14:03:32 <Borg> blocked trains choose free platform
14:06:34 <peter1138> talltyler, getting rid of the gotos is a good idea, just accidentally changing the logic probably not :)
14:06:38 <talltyler> I'm just relieved it was a silly mistake on my part, and it all works again 🙂
14:07:38 <talltyler> Yeah, could probably use a function for the train reversing bit, and maybe split the fatal error to describe each case.
14:16:26 <Borg> I found a Guilty function
14:16:30 <Borg> managed to move the train
14:16:50 <Borg> class CYapfDestinationTileOrStationRailT : public CYapfDestinationRailBase {
14:17:10 <Borg> inline bool PfDetectDestination(TileIndex tile, Trackdir td)
14:17:23 <Borg> when I added checks for IsSafeWaitingPosition()
14:17:38 <Borg> train choose free platform instead of waiting
14:18:30 <Borg> so there is mismatch between PF choise and safe position
14:18:46 <Borg> this hack of course makes game broken, most trains now report as lost
14:20:14 <Borg> lets try to add sth less hacky ;D
14:39:49 <Borg> so, PathFinder stop when it hits *first* tile of station right?
14:45:58 <Borg> its even more interesting when you remove the pbs signal (green one)
14:46:40 <Borg> even when train moves futher away... train 86 still waits.. to reserve entire path (station + exit path to signal
14:46:52 <Borg> instead to just move into free platform w/ is safe
14:48:00 <jfkuayue> Borg: i guess, if you are… suomi perkele? ;)
14:48:24 <Borg> hmm perkele is bad finish word :P but I forgot what it means
14:48:43 <Borg> anyway... its interesting stuff...
14:49:36 <jfkuayue> You often use Satan as a greeting, arent you /j
15:18:20 *** Wormnest has joined #openttd
15:18:44 <Borg> but.. BUT.. I need someone to review the code.. and maybe there is better/cheaper way to achieve it
15:21:19 <Borg> I basically reused logic from Waypoint check in PfCostCalc() but for Station
15:22:27 <Borg> I just use ft.Follow() up to 20 tiles ahead to see if there is Signal.. if we found it and its still reserved.. I just add extra penalty to the path
15:22:38 <Borg> works like a charm.. train doesnt wait for path anymore
15:23:33 <Borg> but.. I dont like it.. because
15:23:49 <Borg> 1) uses CFollowTrackRail
15:24:05 <Borg> 2) its for stations only, and Im sure I saw identical behavior on normal rails
15:24:58 <Borg> when train is on signal tile.. and so. train aproaching PBS.. instead of choosing free path.. slow downs until train moves off signal tile
15:25:48 <Borg> lets see if I can confirm that
15:30:36 <Borg> okey.. its not the case for normal tracks
15:35:07 *** darksaw6536 has joined #openttd
15:35:07 <darksaw6536> Hi, I am the author of the Swarm AI, can I get the Discord tag swag? 🥹
15:37:11 <Borg> dbg: [yapf] [86]: PfCalcCost(): cur.tile=DBC5 cost=4900
15:37:23 <Borg> hmm apperently im wrong DBC5 is last station tile
15:38:40 <Borg> so actually I can optimized it a bit
15:38:52 <Borg> and just lookup single tile
16:06:32 *** gelignite has joined #openttd
16:14:29 *** MinchinWeb[m] has quit IRC (Ping timeout: 480 seconds)
16:14:36 *** MinchinWeb[m] has joined #openttd
16:30:29 <Borg> it can happen on non-station path too
16:30:46 <Borg> identifcal case.. train is just right on signal tile.. trains waits
16:31:21 <Borg> for free path.. instead taking free siding when it can really reserve path
16:44:37 <Borg> I think I have solution for it
16:44:44 <Borg> due to my previous patch..
16:44:57 <Borg> where I keep PBS signals green until train passes it
16:45:03 <Borg> so lets put it to good USE!
16:45:26 <Borg> and use look_ahead_signal to add penalty if we see green PBS signals
17:05:52 *** truebrain has joined #openttd
17:05:52 <truebrain> For those wondering why I just disabled these workflows without conversation and why glx was so kind to fix them properly 🙂
17:08:47 <Borg> okey.. so those 2 needs to be combined to get rid of that slowdown..
17:09:07 <Borg> if PBS -> Block signal transitions happens.. there is no issue because look ahead signals will penalize block signal anyway
17:18:56 <peter1138> Do trees have a cargo type?
17:19:47 <talltyler> (i.e., what do they produce when a Lumber Mill chops them down 😄 )
17:20:43 <talltyler> (this is not a serious answer, to what I assume is not a serious question)
17:23:59 <peter1138> Animated tree types?
17:24:21 <peter1138> Faster/slower growing trees?
17:27:38 <talltyler> I wouldn't mind seeing some meadow tiles, with or without flowers, that grow in patches to break up the expanse of trees. No idea if these make sense as objects, trees, or something else.
17:50:59 <squirejames> talltyler: Perhaps fake-industry thingies, like the field industry
17:54:44 <peter1138> If we're designing new features we're not doing fake-anything.
17:58:54 <talltyler> Agreed. Meadows make more sense as clear tiles, maybe, like rough ground and rocks? They get generated in patches.
18:01:49 *** Wormnest has quit IRC (Ping timeout: 480 seconds)
18:12:35 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
18:13:15 *** WormnestAndroid has joined #openttd
18:15:40 <squirejames> Well you know what I mean 🙂 generated like that, rather than piece by piece
18:16:15 <squirejames> (and regrowing, changing with season like fields)
18:20:32 *** digitalfox has joined #openttd
18:20:32 <digitalfox> truebrain: Thanks for being proactive on this matter
18:36:29 <_zephyris> Gets back to bits and pieces discussed before - more interesting placement patterns which newgrf objects can chose from, terrain tiles (as a simplified/streamlined/targeted feature instead of objects)...
18:37:28 *** Flygon has quit IRC (Remote host closed the connection)
18:39:54 <talltyler> I am still meaning to get back to my lighthouse placement PR 🙂
18:42:27 <peter1138> talltyler, "I have a patch for that" is MY quote.
18:47:12 <kuhnovic> It's not just a patch if it's a PR 😉
18:48:10 <andythenorth> stash is most l33t
18:50:11 <peter1138> So currently there are tree sprites and tree layouts. A tree layout defines which tree sprites appear on a tile. There are 41 tree layouts, and they all have 4 tile-based random variants.
18:51:47 <Borg> okey.. changed from look_ahead_cost to firstred_penalty for PBS.. more effective
18:52:08 <Borg> but only when num_signals_passed == 0
18:52:26 <Borg> works nicely.. of course I need to playtest it more for any anomalies.. none so far
19:03:54 <talltyler> You have a patch, I have a PR that I made a draft after the first round of feedback 🙂
19:20:40 <_zephyris> I have a random text document wannabe design doc saved on my desktop for that.
19:30:47 * talltyler hides my nine open tabs in Notepad++
19:33:37 <peter1138> Put it on the wiki in your scratch space for safe keeping. Sooner or later someone will see it and treat it as gospel ;)
19:39:41 <andythenorth> frosch's page says hi
19:50:39 <_zephyris> Hmm, `newgrf_farm_fields.txt`...
19:50:40 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
19:50:44 *** WormnestAndroid has joined #openttd
19:51:04 <peter1138> 10 files changed, 750 insertions(+), 333 deletions(-)
19:51:07 <peter1138> I don't like this :(
19:51:54 *** Wormnest has joined #openttd
19:53:46 <peter1138> Fields of bulldozers
19:54:30 <andythenorth> there's probably a .diff somewhere for this 😛
20:15:29 <peter1138> I suppose I better add docs.
20:23:15 <peter1138> I suppose that's needed.
20:41:21 <peter1138> Docs checker already failed \o/
20:44:56 *** Borg has quit IRC (Quit: leaving)
20:59:52 <peter1138> Still wondering if it's possible to just have trees instead of trees and tree tiles.
21:00:22 *** MinchinWeb[m] has quit IRC (Read error: Connection reset by peer)
21:00:36 <andythenorth> can any tile have trees?
21:00:38 *** MinchinWeb[m] has joined #openttd
21:01:09 <andythenorth> can we do something like a waveform that's mapped onto tile indexes
21:02:02 <peter1138> Sorry by tree tile I mean something similar to a sprite layout for industries/objects.
21:02:53 <peter1138> There are 53 different tree tile layouts, but only 41 tree sprites.
21:03:14 <peter1138> Whoops, I got that the wrong way around.
21:03:33 <peter1138> There are 41 different tree tile layouts, but 53 different tree sprites (excluding snowy variants)
21:04:54 <andythenorth> I only have emojis to offer
21:05:01 <andythenorth> not sure that's helpful
21:05:17 <andythenorth> I could arrange them 41 different sequences though
21:05:39 <peter1138> There are 4 variants per tree tile layout.
21:05:58 <peter1138> "Randomised" by X/Y location.
21:06:44 <peter1138> So I guess 164 different tree tile layouts.
21:07:35 <mmtunligit> thatd be a great way to do it
21:07:51 <mmtunligit> just having trees instead of tree layouts
21:08:14 <mmtunligit> seems like something that was only doen due to the limitations of the time
21:08:34 <peter1138> The tree layouts are what provide the variety, but also consistency, when there are lots of trees.
21:09:54 <peter1138> Tree layouts also provide palette remaps in the variations.
21:10:24 <peter1138> That's only used for toyland though.
21:14:07 <peter1138> So yeah, I can't really see how to remove that but keep the visuals the same.
21:15:52 <peter1138> It's definitely a design choice rather than something done due to limitations.
21:19:36 <peter1138> Well I misrepresented the original code, as actually it does not have "tree" and "tree layout". The tree layouts list sprite numbers directly. "tree" is my indirection, to allow sprite numbers or act3/2/1 resolved sprites.
21:26:40 *** Wormnest has quit IRC (Ping timeout: 480 seconds)
21:38:38 *** Wormnest has joined #openttd
21:45:53 <peter1138> Okay, potentially...
21:46:13 <peter1138> Store 4x 8 bit values for each tree tile.
21:54:47 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
22:55:51 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
22:56:21 *** WormnestAndroid has joined #openttd
continue to next day ⏵
