IRC logs for #openttd on OFTC at 2025-01-15
β΄ go to previous day
01:59:03 *** Wormnest has joined #openttd
02:04:11 *** herms has quit IRC (Quit: bye)
02:18:06 *** Wormnest has quit IRC (Quit: Leaving)
03:59:59 *** gnu_jj_ has quit IRC (Ping timeout: 480 seconds)
04:41:51 <DorpsGek> - Update: Translations from eints (by translators)
06:07:24 *** keikoz has quit IRC (Ping timeout: 480 seconds)
07:44:26 <andythenorth> I find it confusing π
08:19:12 *** reldred has joined #openttd
08:24:38 <LordAro> (β―Β°β‘Β°οΌβ―οΈ΅ pΗΙΉplΗΙΉ
08:25:42 <reldred> great now I'm stuck like this
08:26:20 <LordAro> reldred γ( γ-γγ)
08:45:04 <andythenorth> my confusion arises from
08:45:40 <andythenorth> - why fuck around with weird ways to do it, instead of introducing an explicit Consist container, which supports reversing the vehicles within it, amongst other things?
08:46:05 <andythenorth> - but introducing Consist to a game that relies heavily on assumptions about lead vehicle, is not a small thing
08:47:03 <andythenorth> - UI in depot, orders, groups; pathfinding, all the money and aging stuff, possibly sprite drawing, and the entire legacy grf stack of vars that make assumptions
09:01:28 <LordAro> i don't think they claimed to have the full solution
09:40:01 <orudge> Any OpenTTD folk heading to FOSDEM this year?
09:41:37 <peter1138> Today is "remember that deadline? well we need the users to be using it before that deadline"
09:41:46 <peter1138> So my answer is basically, er, fuck off.
09:45:52 <peter1138> Oops, but not to orudge. Just a simple no for that.
09:46:49 <LordAro> peter1138: i did wonder how that was reading on the discord side :p
10:02:07 <reldred> "i don't like this thing, so nobody should have it"
10:02:53 <pickpacket> reldred: I do sound like that there, yeah π
10:04:51 <reldred> That said I also want pr8480 merged (you know and fixed as well) so that depots arenβt magical bags of holding, but anyways.
10:05:20 *** k-man has quit IRC (Ping timeout: 480 seconds)
10:07:07 <pickpacket> reldred: fair dues :D
10:08:01 <pickpacket> As long as it's a setting and doesn't break NewGRFs I'm all for it
10:13:54 <peter1138> git newbie detected, everything is done in their master.
10:15:32 <pickpacket> peter1138: github newbie :) I know git well, but I didn't know how PRs in github worked when I made my first
10:15:41 <pickpacket> also put everything on my master
10:20:17 <pickpacket> now I want to play again
10:23:01 <peter1138> Uh... no that sounds bad.
10:23:52 <pickpacket> one could say that I have a "master game" going. A real longplay with lofty goals
10:24:01 <pickpacket> let's interpret it that way ;)
10:24:58 <peter1138> And I misread that as lefty goals.
10:25:10 <peter1138> I suppose it is. Public transport everywhere...
10:41:21 <pickpacket> peter1138: oh, no. Nothing leftist about this. Hyper capitalism and urban dystopia. Hardly any passenger transport at all
10:45:30 <pickpacket> trams in the city centres only.
10:45:57 <pickpacket> Goals: 1) serve every tea farm on the map, 2) cover the entire map in cities
10:46:53 <pickpacket> there's some 300+ cities on the map (1024x1024). I don't know exactly because there's no indicator showing totals in the town listing
10:47:15 <pickpacket> roughly half of the tea farms are serviced so far.
10:49:45 <peter1138> Service 100% of tea farms.
10:55:03 <pickpacket> all tea leaves will be shipped to four tea factories, which in turn support four tea houses
10:56:34 <pickpacket> those living near any of the four tea houses will practically inhale tea. The rest of the population will live in abject poverty
10:58:34 <pickpacket> At least that's how I choose to think of it. 99% won't have access to any commercial or public means to get around. Eventually all industries except for tea production will despawn and give way to ever more dense urban landscapes
11:08:08 *** kuhnovic has joined #openttd
11:08:08 <kuhnovic> Tea is the new coffee
11:26:02 <pickpacket> tea is the second most consumed beverage in the world. Only behind water
11:46:16 <xarick> we brought tea to the british
11:47:06 *** _zephyris has joined #openttd
11:49:05 <xarick> oh, so my history books lied to me
11:50:05 *** johnfranklin has joined #openttd
11:50:05 <johnfranklin> Dutch or Spanish colonies, which were the first?
11:50:09 <LordAro> that's purely the linguistic influences
11:50:44 <_zephyris> Yeah, just where the word came from, not necessarily where the big trade routes were.
11:50:53 <_zephyris> Hehe, just realised chai tea means "tea tea"
11:52:35 *** ahyangyi has joined #openttd
11:52:35 <ahyangyi> Yeah, I started to call that thing "Masala Chai" instead after realizing Chai *is* Tea.
13:03:59 <peter1138> Hehe, I was just wondering that.
13:10:56 <LordAro> why do the icons need to be included in the string code at all?
13:11:31 <LordAro> it's not translatable and is always on the left (or right for rtl i suppose)
13:23:01 <peter1138> Okay, we could just remove the icon, it's not really necessary.
13:24:37 <_zephyris> I like the icons π Just good to avoid another RTL bug.
13:26:19 <peter1138> None of the other options have icons so it's a bit out of place.
13:48:00 <xarick> it's not easy to see 1020 length sized rivers
13:48:21 <xarick> need some special tgp criteria
14:05:14 <merni> _zephyris: what the heck is chai tea
14:05:50 <merni> most google results talk about masala chai (for me)
14:06:22 *** florafex has joined #openttd
14:07:01 <merni> so it's just another name for masala chai
14:07:20 <andythenorth> chai is just chai
14:07:28 <andythenorth> chai tea is only a thing for people who are wrong
14:07:47 <merni> masala chai would generally just be called "chai" here except when you want to distinguish it from other kinds
14:08:01 <andythenorth> small cups or large?
14:08:29 <andythenorth> had a lot of it in shops in Mumbai
14:08:50 <merni> I hope you didn't have any 'chai wallah' π
14:09:25 <andythenorth> sorry, we had chai from the chai wallah
14:09:32 <andythenorth> lots of the big shops have one
14:10:06 <merni> usually I would think of chai wallahs as selling from street stalls
14:10:22 <merni> are you sure you don't mean a market :p
14:10:29 <andythenorth> no, like if you go to buy a suit
14:10:41 <andythenorth> the salesperson calls the chai wallah over while you sit
14:11:02 <merni> I've never had that experience... but then I've never bought a suit π
14:11:27 <merni> And the shoe shops I go to are not very fancy... and not in Mumbai
14:11:56 <merni> wow that is some expensive stuff
14:12:12 <andythenorth> yeah, it used to be relatively cheaper
14:12:28 <andythenorth> 18 years ago π
14:14:59 <pickpacket> andythenorth: wow... 18 years ago. I was 5 at the time
14:18:48 <pickpacket> Actually I was 23. I was born in '85 and have thus been 23 for almost 18 years now
14:22:07 <pickpacket> the math gets a little murky sometimes. Since I'm 23 now I should have been 5 at the time, but I forget that I've been 23 for some time now
14:34:05 <xarick> i can't make 1020 sized rivers
14:49:46 <xarick> impossible to get 1 river at all
14:50:17 *** kuka_lie has joined #openttd
15:02:46 <xarick> `uint radius = std::min<uint>(3, (current_river_length / (long_river_length / 3)) + 1);`
15:02:46 <xarick> this math is a bit off for values different than default
15:02:50 <_glx_> no node at all? I'd blame the node detection
15:04:16 <_glx_> looks like radius is most likely always 1
15:06:25 <_glx_> I'd rewrite `(current_river_length / (long_river_length / 3))` `((current_river_length * 3) / long_river_length)`
15:08:52 <_glx_> integer math works better that way
15:10:31 <xarick> closer to spring = less radius
15:12:40 <xarick> only from 8 tiles and below from source it begins to thinner
15:17:16 <xarick> i can make use of begin_end_points.back()
15:30:27 <xarick> spring removed from River_UserData struct
15:30:39 <_glx_> just to show the effect of integer math π
15:42:18 <xarick> long_river_length could be stored in River_UserData
15:42:32 <xarick> avoids computing distancemanhattan all the time
15:44:07 <xarick> github copilot is scary
15:44:16 <xarick> it totally guessed what I was going for
15:46:06 <andythenorth> no scarier than autocomplete
15:46:14 <andythenorth> it's just fancy autcomplete
15:58:22 *** Flygon has quit IRC (Read error: Connection reset by peer)
16:12:15 <xarick> AyStar::loops_per_tick exists...
16:14:19 <xarick> I mean, the feature is ready to be used
16:14:32 <xarick> feels bad to just remove it though
16:24:41 *** gelignite has joined #openttd
16:33:43 <xarick> rats, i found a bug thx to this DistanceManhattan
16:33:59 <xarick> lakes were not pushing begin and end nodes
16:35:13 <xarick> begin_end_points.push_back(lake_centre);
16:37:14 <xarick> actually end, because lake_centre is a moving target
17:30:12 *** Wormnest has joined #openttd
17:44:03 *** kuka_lie has quit IRC (Quit: leaving)
17:44:41 *** kuka_lie has joined #openttd
18:47:35 *** ChanServ sets mode: +v tokai
18:53:06 <xarick> why does AyStar finder = {}; must be initialized ?
18:53:22 <xarick> I tried AyStar finder; - bad stuff happened
18:54:33 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
19:20:16 *** michi_cc has joined #openttd
19:22:08 <truebrain> wait, is this actually happening? π
19:22:54 <michi_cc> The great Rubidium decreed it so π
19:27:56 <Rubidium> well... peter1138 has almost as many approved PRs as stashes ;) :D
19:53:27 *** gelignite has quit IRC (Quit: Stay safe!)
20:11:26 <xarick> std::make_unique<uint16_t[]> weird funny code
20:29:33 <xarick> I made AyStar finder; work
20:36:14 *** kuka_lie has quit IRC (Quit: leaving)
20:55:27 <xarick> can i cap long river size to
20:55:44 <xarick> something that the tgp can create?
20:56:42 <xarick> length = 1020 had a 0% success rate
20:57:09 <xarick> and causes the rivergen to try hard and slow down world gen
20:58:00 <_glx_> 1020 doesn't fit in normal size maps
20:58:29 <xarick> oh, I never thought of that
21:04:57 <xarick> min_river_length = Clamp(min_river_length, 2, Map::SizeX() + Map::SizeY());
21:06:21 <xarick> at the maximum value that's still quite the impossible task though
21:18:40 <johnfranklin> 18 years ago I was in kindergarten and sometimes secretly dug into computer room to play minesweeper and solitaire
21:23:08 <xarick> it's still having a hard time to generate rivers this sized
21:55:14 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
22:09:48 *** keikoz has quit IRC (Ping timeout: 480 seconds)
22:22:15 <reldred> xarick: have a look at my river patches in JGRPP, in particular theres a set of clauses for allowing a river to spawn and one of them is very very hard to achieve, removing the 2nd one allows multiple smaller rivers to form a very believable long chain.
22:24:41 <reldred> so you can keep min length low(ish) leading to less wasted CPU cycles on failed river creation attempts, but still have epic rivers that span 2000 tiles from mountains to the coast. I suggest loading up Cpt.K's Cliffy Coastline 2 in JGRPP with the river options turned on, set the map to be 2k x 4k and setting rivers to extremely many and watch what it does.
22:29:03 <xarick> π¦ where can i check the patches
22:29:44 <reldred> I'll get some screenshots
22:34:25 <xarick> i'm looking at the landscape.cpp
22:34:26 <reldred> "Rivers require their springs to generate at the top of hills" is the most important tweak to get the rivers to actually form chains across thousands of tiles and hundreds of height level change
22:35:41 <xarick> hmm why isn't this be ported in openttd π¦
22:36:21 <xarick> guess I'm going to copy paste some ideas from there
22:38:02 <reldred> 2nd option, not the first. I disabled the first one after that commit as it caused some asserts
22:38:46 <reldred> xarick: Nobody seemed particularly interested, and I wasn't particularly interested in duplicating all of my work.
22:39:20 <xarick> that code is present in FindSpring
22:39:32 <xarick> but as an "always on", no setting at all
22:40:19 <reldred> I just wrapped it in a setting to turn it off.
22:40:29 <reldred> Really really braindead simple stuff
22:40:44 <xarick> ok, let me try and see what it comes up with
22:40:49 <reldred> just the 2nd check though, not the first.
22:41:00 <reldred> /* Are we near the top of a hill? */
22:41:47 <reldred> the check at "/* Are there enough higher tiles to warrant a 'spring'? */" prevents some asserts/crashes
22:42:47 <reldred> I don't know how much effect it will have with TGP though, I made those settings for working with heightmaps better.
22:43:10 <reldred> where rivers would fizzle out despite the heightmap explicitly being generated to allow long river chains.
22:43:13 <xarick> the tile that's passed to FindSpring is a "RandomTile"
22:43:30 <xarick> can be a void tile, that's why it can screw if it's not ensured it's inside
22:49:25 <xarick> not resulting in rivers yet π¦ I'll take a better look at it
22:51:06 <reldred> It's also going to depend on the landscape in general, I see best results with a good heightmap with the maxheight not set too high (Notable exceptions being Cpt Klutz' Cliffy Coastline 1 which can achieve them at maxheight 255)
22:51:40 <reldred> teragenesis while a big step up from the original map gen... is kinda trash these days.
22:52:42 <xarick> got 288, 291, 266 distancemanhattan rivers
22:53:20 <xarick> tried to make 128 rivers though
22:53:32 <xarick> threshold was a min of 255
22:54:36 <reldred> well, the map gen needs to support rivers that are worth a damn to begin with, the software captain klutz uses, worldengine, can simulate rainfall erosion patterns
22:55:09 <reldred> like, that map you've got there xarick is just a blob, there's hardly any elevation, nowhere for a river to really 'flow'
22:56:43 <reldred> thats what the heightmap from my screenshots looks like
22:56:58 <reldred> theres somewhere for the rivers to actually flow
22:57:24 <reldred> really, we need to get the mapgen to the point it can actually generate believable terrain
22:57:33 <xarick> hmm, thinking of a way to persuade FindSpring to start on the highest heights, be more selective
22:57:44 <xarick> RandomTile just isn't enough
22:57:56 <reldred> doesn't take rivers too much tweaking to get them to generate believable rivers, but you need believable terrain first.
22:58:55 <reldred> like my only issue with rivers currently is it's slow, I don't think rivers themselves are too problematic.
22:59:22 <xarick> let me find that heihhtmap
23:01:35 <xarick> found it, what highest peak u got there?
23:01:59 <reldred> that was 2k x 4k maxheight 255
23:02:31 <reldred> I play on *very* mountaineous maps
23:03:00 <reldred> The highest whitepoint on that heightmap isn't 255 anyway, so you'll never actually achieve that max height
23:04:14 <xarick> woah, gigantic rivers indeed
23:05:35 <xarick> tgp should generate that kind of terrain
23:06:22 <xarick> dist 1186 unsigned int
23:08:14 *** nielsm has quit IRC (Ping timeout: 480 seconds)
23:08:23 <xarick> heightmap has its own settings?
23:08:36 <reldred> not that I'm aware of
23:08:37 <xarick> i wanted 8192 rivers, it picked 4096
23:08:56 <reldred> yeah well it doesn't always win when it makes its picks
23:09:09 <reldred> but, turn off rivers and look at the actual heightmap
23:09:22 <xarick> i meant the amount of rivers settings was the default
23:09:36 <reldred> each tile is a full flat tile with a full slope tile pointing at another full flat tile
23:09:46 <reldred> perfect conditions for the river to generate
23:10:24 <reldred> thats what you need TGP to generate
23:10:36 <reldred> it needs to generate realistic terrain if you want realistic rivers
23:11:14 <xarick> i think there's a bug in openttd itself
23:11:35 <xarick> heightmaps aren't using client defined settings
23:11:56 <xarick> maybe it's a me problem
23:12:54 <reldred> nah heightmaps for me in jgrpp use the settings I set
23:13:06 <reldred> so you must have busted your local copy
23:13:40 <reldred> like, I did a *LOT* of iteration to get the rivers behaving the way they do in jgrpp, and JGR helped with the tropic biome change stuff π
23:14:33 <xarick> amount of rivers: many, and i know for sure that entails 8192
23:14:45 <reldred> it also depends on map size
23:15:18 <xarick> too used to 4kx4k, forgive me openttd π
23:15:35 <reldred> I added two extra quantity amounts, and 4k x 2k at the highest setting was 16k river steps
23:15:52 <reldred> but that whole approach is pure brute force
23:16:10 <reldred> nobody actually wants eight thousand distinct rivers
23:16:48 <xarick> im still not getting the pretty results you're getting, I think 1020 dist min is just too much
23:17:11 <reldred> look at my settings screenshot, replicate those in jgrpp and watch what it does
23:17:15 <xarick> i get long rivers, but not ... exactly many branches, for each branch they would need to be 255 min
23:17:33 <reldred> I don't crank up the min river length very far at all
23:18:12 <reldred> turning off 'springs generate at the top of hills' is the key
23:18:20 <reldred> my minlegth is only 64
23:18:29 <reldred> randomness doesn't do a huge amount
23:26:38 <xarick> min river length 64 on both images
23:28:05 <xarick> gonna try improve the "RandomTile" choice
23:28:37 <xarick> get a list of the highest tiles first and try randomly chosing there
23:28:52 <xarick> instead of a random tile anywhere on the map
23:33:27 *** ian01223 has joined #openttd
23:33:27 <ian01223> lol at that - did I stumble into the local deity's workshop?
23:33:35 <ian01223> ...would that make Xarick an angel apprentice or something?
23:33:51 <reldred> eh it's just replicating what I've already put into jgrpp
23:34:02 <reldred> but he's doing other river tweaks
continue to next day β΅
