IRC logs for #openttd on OFTC at 2026-02-03
⏴ go to previous day
00:11:24 *** MinchinWeb[m] has quit IRC (Ping timeout: 480 seconds)
00:13:22 *** MinchinWeb[m] has joined #openttd
00:46:47 <goddess_ishtar> the diffs look identical???
00:46:50 <goddess_ishtar> why would you do that
00:47:40 <goddess_ishtar> oh, the commit name didn't match the format
00:48:14 <goddess_ishtar> (just. amend your commit and force-push the branch?)
01:27:30 <_glx_> possible reasons for the peaks are ScriptList constructors, or Valuate calls inside valuators
01:29:28 <_glx_> oh and garbage collection of huge lists
01:56:00 *** urdh has quit IRC (Ping timeout: 480 seconds)
02:43:03 *** Wormnest has quit IRC (Quit: Leaving)
03:24:01 *** WormnestAndroid has quit IRC (Remote host closed the connection)
03:24:04 *** WormnestAndroid has joined #openttd
04:33:54 *** Zathras_1 has joined #openttd
04:37:20 *** Zathras_4 has quit IRC (Ping timeout: 480 seconds)
04:54:51 *** MinchinWeb[m] has quit IRC (Read error: Connection reset by peer)
04:55:07 *** MinchinWeb[m] has joined #openttd
05:29:28 <andythenorth[d]> Are script authors supposed to manage performance themselves?
05:29:40 <andythenorth[d]> Genuine question
05:31:57 <goddess_ishtar> andythenorth[d]: I mean, no? at least, not directly? the game is supposed to be the one responsible for provisioning how many opcodes you get
05:32:13 <goddess_ishtar> I guess efficiency is everyone's problem but
06:07:37 <reldred> are grf authors supposed to manage performance themselves?
06:08:18 <reldred> coz a grf can definitely have a performance impact, you just got done fixing FIRS for that
06:16:58 <ahyangyi> I think they are different
06:18:08 <ahyangyi> a slow grf has to drag down everything else, a slow ai might but does not have to.
06:22:13 <reldred> a slow AI can absolutely tank game performance same as a grf can.
06:22:24 <reldred> we don't see it on PC much but it affects the mobile players
06:31:57 <ahyangyi> I mean, it can but there is potential to prevent it.
06:43:21 <andythenorth[d]> Generally a grf isn’t supposed to manage performance
06:43:56 <ahyangyi> And cannot manage performance.
06:44:03 <andythenorth[d]> But it’s a 90:10 situation because of certain spec affordances
06:45:55 <ahyangyi> It would help if slower effects can be turned off though
06:48:12 <reldred> likewise though should the openttd devs be responsible for every broken piece of dogshit uploaded to bananas?
06:48:51 <reldred> i mean you can try and make ai/gs/grfs idiot proof and then you just end up with superior idiots
06:53:47 <rito12_51026> Can performance expensive GRF couse desync?
06:53:51 *** Smedles has quit IRC (Read error: Connection reset by peer)
06:54:05 <reldred> Dunno, but grfs can cause desyncs in other ways
06:54:23 *** Smedles has joined #openttd
06:56:23 <ahyangyi> I'm fine with occasional performance spikes since not everything can be handled by the opcode count
07:52:28 <Rubidium> rito12_51026: they can, but there is not necessarily a causal link. It's conceptually the same as asking "Can big vehicles cause accidents" => yes they can, but the vehicle being big isn't the cause for accidents, it's being a vehicle that causes the causal link.
08:12:52 <ahyangyi> If A is not the cause of B, can we say "A can cause B" at all?
08:13:09 <ahyangyi> Oh, the big vehicle example, I see.
09:27:32 <_zephyris> Back on terraform/build/demolish, I suspect cost estimates on terraforming or demolishing areas are incorrect.
09:28:04 <_zephyris> Terraforming because changes to neightbouring tiles change the cost of subsequent changes, demolishing because of multi-tile buildings/objects/industries.
09:28:20 <_zephyris> So, allowing/disallowing an action based on estimated cost is harder
09:31:27 <peter1138> That's why those commands have the NoTest flag set which allows the estimate and actual cost to differ.
09:32:57 <peter1138> (basically the current behaviour is deliberate)
09:34:36 <peter1138> And though it's different from a user's point-of-view, placing objects and buying land are the same operation.
09:38:51 <peter1138> `CommandFlag::NoTest` could probably do with renaming :)
09:41:07 <Rubidium> from a user's POV, should placing objects and placing trees also behave the same?
09:41:16 <_zephyris> But does't make sense for buy land/place objects then, because their cost is predictable
09:41:31 <_zephyris> (and yup, matching trees would be good)
09:44:50 <peter1138> Tree behaviour is correct, right?
09:46:33 <peter1138> (Not showing any error for objects when nothing is placed due to no money is definitely incorrect, no argument there :))
09:53:59 <peter1138> talltyler added the objects-by-area feature, presumably he knows what he intended ;)
10:08:46 <_zephyris> I'll PR the fix, pretty sure it's right, and see what he thinks
10:35:05 <andythenorth[d]> "Civilization creator Sid Meier famously had a rule for sequels: 33 percent of a new game should retain systems that were established in the prior entry, 33 percent should substantially improve prior systems, and 33 percent should be brand-new systems and mechanics."
10:36:16 <peter1138> Do IH and FIRS follow that rule?
10:36:51 <LordAro> 33% redrawn for every release?
10:39:49 <xarick> read a bit of the forums
10:39:57 <xarick> ppl want passwords back
10:40:55 <xarick> haven't tried playing a multiplayer game yet to feel the new experience
10:41:30 <jfkuayue> Where is the remaining 1 percent?
10:46:40 *** gwyd4016 has joined #openttd
10:46:40 <gwyd4016> Systems that are made substantially worse since the last edition
10:52:33 <peter1138> Microsoft are doing well then.
10:59:25 <xarick> oh, SimpleAI, the original, limits number of routes to 1000
10:59:58 <xarick> my edited version removes this limit
11:01:32 <xarick> if (routes.len() < 999999999
11:07:05 <LordAro> tbf, Civ6 & Civ5 had similar issues/criticism on release
11:07:15 <LordAro> (and i daresay the others too)
11:07:23 <LordAro> peter1138: maybe a year?
11:09:57 <LordAro> 11 Feb 2025, almost exactly
11:10:50 <_zephyris> Yeah, it's had time to settle in. I do remember the no unit stacks and hex tiles controversy of V.
11:10:56 <_zephyris> 91% (III), 92% (IV), 95% (V), 84%(VI), 51% (VII)
11:12:31 <LordAro> i did enjoy playing V
11:13:14 <LordAro> i'll never forget the diplomatic victory a friend got against me even though my giant death robots were literally half a turn from crushing his capital
11:15:04 <jfkuayue> 2025 was one year ago??
11:22:18 *** Smedles has joined #openttd
11:24:02 <xarick> oh, i'm like 95% sure if the script_vehiclelist stuff slowing down
11:25:04 <xarick> wish you could do something about it
11:27:13 <xarick> other than telling me to use default limits
11:34:01 *** firepup650 has quit IRC (Quit: Going down with the ship)
11:34:18 *** firepup650 has joined #openttd
11:37:40 <xarick> in the name of AI's spamming vehiclelists!
11:50:40 *** firepup650 has quit IRC (Quit: Going down with the ship)
11:50:57 *** firepup650 has joined #openttd
11:55:46 *** Alkel_U3 has quit IRC (Quit: maintenance)
11:56:56 *** Alkel_U3 has joined #openttd
12:01:28 <xarick> I'd like to create a PR, but... it's so very niche
12:01:38 *** lobster has quit IRC (Read error: Connection reset by peer)
12:02:14 *** lobster has joined #openttd
12:31:34 <xarick> new game started, testing my branch, let's see how stuff looks like in a few hours
12:52:34 *** Flygon has quit IRC (Remote host closed the connection)
13:00:47 <_glx_> Hmm that also means version mismatch for network games
13:01:15 <peter1138> Okay, so we need a CI task that will abort the release build if it doesn't match t
13:01:53 <peter1138> That will allow us to delete the tag and recreate it.
13:02:27 <locosage> don't remember where network gets it but 15.1 doesn't work on 15.0 servers
13:19:17 <_glx_> Ah right network check uses the tag, but bananas and newgrf use internal version
13:20:11 <peter1138> Could fix it and retag 15.1
13:20:25 <peter1138> Otherwise that bug report can never be closed.
13:22:46 <orudge> No need to retag it I would think, happy for the report to be closed - just wasn't sure if there was anything we could do to make it harder for this to happen. A change in FindVersion.cmake or similar might be worthwhile.
13:30:32 <xarick> additional conditions to be able to bribe?
13:30:57 <xarick> exclusive transport + bribe
13:31:47 <xarick> infinite money + AIs = you can't transport!
13:36:09 <peter1138> Infinite money is basically a sandbox mode. Don't use it for competitive games with AIs.
13:41:12 <xarick> NoNoCAB with that option enabled is really abusive
13:41:25 <xarick> not sure it targets every town
13:41:34 <xarick> or just the towns he transports
13:42:03 <xarick> but the log is really huge
13:42:10 <pickpacket> How come towns that I have no stations in hate me because I have rails running through their territory?
13:44:07 <xarick> probably not all towns, he caps number of towns at the start
13:48:46 <peter1138> Because you built rails running through their terrority, right?
13:49:37 *** Wormnest has joined #openttd
13:58:09 <LordAro> peter1138: it's definitely been forgotten before
14:04:51 <pickpacket> peter1138: I did, but they *never* forget it
14:05:41 <pickpacket> And even if I bribe them (or plant trees) to get a better rating that rating falls again even if I don't build or destroy anything more in their territory
14:06:30 <_glx_> yes you are present in the town but not serving it
14:21:53 <peter1138> Which bit is out of date?
14:32:02 <pickpacket> peter1138: it doesn't say anything about rating being affected by infrastructure other than stations
14:37:12 <peter1138> But if you're clearing houses or roads, or trees anywhere in the town's local authority zone, the rating will be affected.
14:37:57 <pickpacket> yeah, but in time the rating should go back up. In my experience it doesn't
14:38:19 <peter1138> Only up to mediocre.
14:42:32 <_glx_> there's the counter balance of +5 every month if <200 and -15 every 50 days if bad transport in station
14:43:43 <_glx_> so if the town has a one or more bad serviced stations the auto up won't have any effect
15:46:25 *** ialokin has quit IRC (Ping timeout: 480 seconds)
16:22:38 <_jgr_> I strongly suspect that this is not a good approach
16:23:42 <xarick> doesn't have any sorting apparently
16:28:56 <xarick> weird way to set the const_iterator
16:28:56 <xarick> is this really how it's indended to be used
16:30:53 <xarick> anyway, I think it's gonna be slow during find_if
16:46:19 *** MinchinWeb[m] has quit IRC (Ping timeout: 480 seconds)
16:46:27 *** MinchinWeb[m] has joined #openttd
17:10:56 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
17:10:57 *** WormnestAndroid has joined #openttd
17:22:35 <xarick> wow NoNoCAB v9 is out!!
17:23:09 <xarick> 2026 is gonna be the year of AIs
17:24:05 <_jgr_> (strong doubts about this...)
17:25:26 <xarick> already 4 AIs in 2026, but one is fake
17:26:18 <Zathras_1> I have joined a multiplayer server and created a company. did this on my desktop PC. now I want to continue playing that game from my laptop.
17:26:18 <Zathras_1> this is on 15.0 right now (since the server has no 15.1 yet). is there some kind of identifier I can use insted of a password?
17:26:46 <Wormnest> Yeah, besides the bugfixes for NoNoCab and WormAi, I doubt I will have time to do any other development on them.
17:28:06 <xarick> WormAI also being updated?
17:28:47 <Wormnest> xarick: yes, the money crash you found a few years ago, which was in SuperLib.
17:32:40 <xarick> let me check something quick
17:33:39 <xarick> ah nevermind, I made no changes to wormAI
17:34:38 <xarick> 2026 is the year AroAI gets updated!
17:35:40 <_jgr_> Zathras_1: You'll need to copy the secrets.cfg file from the openttd config folder on your desktop to your laptop
17:37:54 <xarick> is WormAI more active now? It tends to sleep a lot
17:38:04 <xarick> between building new routes
17:39:30 *** MinchinWeb[m] has quit IRC (Read error: Connection reset by peer)
17:39:50 <xarick> How fast this AI will think?
17:39:55 *** MinchinWeb[m] has joined #openttd
17:40:40 <xarick> max distance of 1000 is kinda... massive
17:41:04 <Wormnest> By default it isn´t very fast. Nothing has changed, except the bug fix.
17:53:16 <xarick> SuperSimpleAI also updated this year
17:53:25 <xarick> too bad he doesn't upload to bananas
17:56:20 <xarick> oops, i was running NoNoCAB 8
17:56:49 <xarick> there is a bug when updating AIs
17:57:14 <xarick> if I have NoNoCAB chosen, but i update to a new version, it will still start the previous version
17:57:28 <xarick> i have to restart openttd
17:58:42 <LordAro> xarick: i don't recall seeing an issue for that
18:01:10 <Wormnest> xarick: do you have a local copy under /ai/
18:02:57 <Wormnest> I guess the MacroEcon ai doesn´t work, it doesn´t show up in the list to select from.
18:03:39 <xarick> I tried to make it boot
18:03:54 <xarick> it has wrong syntax for require
18:04:23 <xarick> it has mixing classes, like Math
18:04:51 <xarick> uses AI.AbstractList, HasNext but it's set to use API 14 or whatever
18:06:03 <xarick> after these hindrances, what the AI was able to do was exclusive transport rights and building statues
18:06:55 <xarick> it looks to connect industries that produce what it accepts
18:07:07 <xarick> doesn't end up placing any piece of infrastructure
18:09:11 <Wormnest> Ah I see, that´s too bad.
18:14:23 <xarick> oh yeah, KrakenAI also mentioned something about updating it
18:14:47 <xarick> that's the 8th AI getting an update :)
18:14:53 <xarick> yeah, definitely the year of AIs
18:27:30 <xarick> sounds like a Valuate function that does a lot
18:32:16 *** MinchinWeb[m] has quit IRC (Read error: Connection reset by peer)
18:32:32 *** MinchinWeb[m] has joined #openttd
18:38:56 *** ialokin has joined #openttd
18:46:20 <Zathras_1> _jgr_: sorry, I am on it. want to sort out a .openttd vs .config/openttd confusion first.
18:47:56 <_glx_> xarick: well, IndustryId being 20019, that's already more than 100k ops (not counting the actual valuator cost)
18:48:40 <_glx_> assuming the list contains all industries
18:49:33 <_glx_> in master it's suspended
18:50:35 *** MinchinWeb[m] has quit IRC (Ping timeout: 480 seconds)
18:51:24 *** MinchinWeb[m] has joined #openttd
19:00:16 <xarick> actually it's taking him a long time to build :|
19:44:56 <xarick> okay, gonna investigate
19:50:19 <andythenorth[d]> weren't we removing 4k maps in OpenTTD 16?
19:50:27 <andythenorth[d]> I'm sure I saw that PR go by
19:53:00 *** SigHunter has quit IRC (Ping timeout: 480 seconds)
19:53:00 *** SigHunter has joined #openttd
20:02:57 <xarick> building openttd is taking a bit too much time
20:05:43 <xarick> there was a sea, then suddenly the sea disappears
20:05:55 <xarick> replaced by a land mass
20:23:20 *** MinchinWeb[m] has quit IRC (Read error: Connection reset by peer)
20:23:37 *** MinchinWeb[m] has joined #openttd
20:36:54 <xarick> 2k map, KrakenAI can build, still took 4 months at 250k ops
20:44:17 *** gelignite has joined #openttd
21:05:08 <xarick> i dunno what superlib does, but something there isn't fast enough for 4k maps
21:07:49 <andythenorth[d]> don't play 4k maps
21:11:07 <_jgr_> xarick: Why not do some profiling and find out what it is?
21:12:11 <xarick> I'm trying to decipher what KrakenAI is doing
21:13:00 <xarick> I suspect the Water body checks might be a source of slowdowns
21:28:54 <xarick> I made it print the evaluations, yeah this is definitely slow in a 4k map
21:29:37 <xarick> the counter must reach 0 before it starts building anything
21:40:02 <xarick> holy crap, it repeats the operation after a while
21:51:21 <xarick> AIIndustryList_CargoAccepting
21:51:49 <xarick> i mean... for every industry, he also gets every other industry that accepts the cargo
21:52:31 <xarick> let me disable the water body check
21:54:25 <xarick> without water body check... faster, but still feels slow
22:29:11 <peter1138> Right, time to reread that network change PR.
22:55:44 <xarick> i made a little something, reduced 5 years to 2 months
22:58:09 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
23:28:05 <xarick> reduced evaluation period by 4 days with minor optimization
23:31:05 *** eilonlipton5365 has joined #openttd
23:31:05 <eilonlipton5365> EilonviaGitHub: Hi, it's me from GitHub. I submitted a PR last week (https://github.com/OpenTTD/OpenTTD/pull/15158), and after iterating on it and with some great community feedback, I think it's in a good state. I'm wondering what I need to do as far as next steps go to see it part of the next release of OpenTTD. Do the project owners review community PRs every once in a while? Or I can buy
23:31:05 <eilonlipton5365> someone a coffee to review it? :openttd: I'm happy to help however possible
23:42:00 <Rubidium> those with approve/merge powers review whenever and whatever they fancy, so sadly it needs some luck to get things to progress. Sometimes things go really quickly, other times it takes weeks before someone takes a look at a PR
23:43:09 <xarick> I am waiting 3 years for a merge :P
23:44:18 <eilonlipton5365> Alright well I'll go pick up my kids from school and hope that in the next hours things speed up! 🙂 No worries, I'm just excited for the potential!
23:45:11 <eilonlipton5365> Rubidium: And thank you for your comments on my PR too!
23:51:29 <xarick> it does this for every route it builds
23:55:48 *** MinchinWeb[m] has quit IRC (Remote host closed the connection)
23:57:06 *** MinchinWeb[m] has joined #openttd
continue to next day ⏵
