IRC logs for #openttd on OFTC at 2023-01-07
⏴ go to previous day
01:13:20 *** HerzogDeXtEr has joined #openttd
03:17:05 *** Wormnest has quit IRC (Quit: Leaving)
03:29:21 *** debdog has quit IRC (Ping timeout: 480 seconds)
04:14:03 *** TROILUS2 has joined #openttd
04:14:27 *** TROILUS has quit IRC (Quit: Ping timeout (120 seconds))
04:14:27 *** TROILUS2 is now known as TROILUS
04:25:48 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
04:34:57 *** APTX has quit IRC (Remote host closed the connection)
04:57:38 *** APTX has quit IRC (Remote host closed the connection)
05:48:36 *** Smedles has joined #openttd
06:52:12 *** supermop_toil_ has joined #openttd
06:58:41 *** supermop_toil has quit IRC (Ping timeout: 480 seconds)
07:20:25 *** supermop_Home has quit IRC (Ping timeout: 480 seconds)
09:57:40 <andythenorth> what are we doing?
09:57:46 <andythenorth> trying to clean repro a bug?
10:01:01 <andythenorth> ok it only happens in my savegame
10:01:08 <andythenorth> can't repro in a new game
10:07:36 <andythenorth> hmm maybe 13.0 RC1 won't crash
10:07:48 <andythenorth> the only thing that I am aware of changing is my build environment
10:07:53 <andythenorth> savegame was fine yesterday
10:08:01 <andythenorth> 13.0 RC1 .dmg won't open though
10:08:32 * andythenorth tries the nightly instead
10:09:44 <andythenorth> ok nightly crashes also
10:10:07 <andythenorth> but now I need to find out why 13.0 RC1 doesn't work
10:11:27 <andythenorth> ok downloaded the 13.0 RC1 .dmg again, it worked, seemed like Chrome was holding onto the file the first time
10:16:15 <andythenorth> so....a reproducible bug with 2 different triggers, involving train groups and autoreplace, in a savegame where grfs have been changed many times....and the mac crash log seems to be borderline useless
10:16:25 <andythenorth> fun to investigate, or gloss over and move on?
10:16:54 <andythenorth> can't repro outside of this savegame, and it's not present in older versions of the savegame
10:26:46 <TrueBrain> It sounds like you have a bug that is specific to your savegame and that you have tried multiple approaches to troubleshoot it. If you have not been able to reproduce the bug in a new game and the crash logs are not helpful, it might be challenging to determine the root cause of the issue.
10:26:46 <TrueBrain> In this situation, it might be best to gloss over the issue and move on, especially if the bug is not causing significant problems in your game. Alternatively, you could try to create a minimal reproduction of the issue in a new game, which might make it easier to identify the cause of the problem. This could involve creating a new game with only the necessary grfs and replicating the steps that lead to the bug.
10:26:52 <TrueBrain> Is what ChatGPT thinks andythenorth
10:28:04 <TrueBrain> this part surprises me: `only the necessary grfs` .. like it knows what grfs are 😛
10:32:13 <andythenorth> it hasn't done a bad job
10:32:39 <andythenorth> it hasn't detected that I tried the minimal repro already and the bug isn't manifesting
10:32:49 <andythenorth> there is nothing about this AI future that worries me 😛
10:33:16 <TrueBrain> it is funny to feed it bug-reports
10:33:25 <TrueBrain> it comes up with good responses I could just copy/paste in there, and be done with it 😛
10:34:21 <TrueBrain> this reply would have been better than mine 😛
10:35:46 <TrueBrain> When I feed it #10309:
10:35:56 <TrueBrain> okay, I will stop now 🙂
10:38:33 <TrueBrain> okay, one more, I was feeding it info about #10326, including the implementation
10:38:35 <TrueBrain> `One potential issue with this implementation is that it may not properly handle cases where the vehicle is carrying multiple cargoes, or where the vehicle has been refitted with a different cargo than its default. This could be causing the incorrect capacity to be displayed in the buy menu, as described in the bug report.`
10:41:19 <TrueBrain> not sure, but from the whole conversation I just had, I have the feeling I understand more about the PR. Just still no clue what the actual problem is 😄
10:43:04 <TrueBrain> That was a lot more detail than I expected ... peter could have added that to the PR 😛
10:44:52 <andythenorth> is there a github plug in for it?
10:45:06 <andythenorth> also...I think it may have use on reddit 😛
10:45:33 <andythenorth> I suspect the AI is already much better equipped to navigate life than the posters on reddit
10:45:46 <TrueBrain> Bing want to integrate it
10:45:57 <TrueBrain> but it is hard to judge if it is actually telling the truth or not
10:46:26 <LordAro> TrueBrain: i like the bit about the importance of compatibility with Windows <3.1
10:48:11 <TrueBrain> " It was developed as a free and open-source alternative to Transport Tycoon Deluxe, and has been highly praised for its accessibility, moddability, and multiplayer support."
10:50:29 <TrueBrain> I keep asking it who developed OpenTTD, it spits out one or two names that are correct, and follows it with many that are not
10:50:34 <TrueBrain> like just names I have never heard about and shit 🙂
10:50:43 <TrueBrain> it keeps saying: "I apologize for the error in my previous responses"
10:51:48 <TrueBrain> `I apologize for the errors in my previous responses. I do not have access to current information about the development team of OpenTTD, and my responses were based on outdated knowledge. I apologize for any confusion or misinformation my previous responses may have caused. If you have any further questions about OpenTTD, please don't hesitate to ask.`
10:51:55 <TrueBrain> okay, I should really stop now 😛
10:54:57 <andythenorth> is a sorry AI really sorry?
10:55:08 <TrueBrain> no, as it keeps telling lies anyway
10:55:34 <TrueBrain> it is really terrible with facts like this
10:55:46 <TrueBrain> so no replacement for wikipedia 😛
10:55:49 <andythenorth> so is the internet 😛
10:56:00 <andythenorth> see: PBS is slow
10:56:24 <LordAro> you should definitely ask it if path signals are slow
10:56:31 <LordAro> and separately, if they're fast
10:56:36 <TrueBrain> ha, it can solve CodeQL issues 🙂
10:56:48 <TrueBrain> I asked it to document a function that CodeQL complained was too long without documentation .. it did just that
10:58:40 <TrueBrain> hahaha, it does exactly what you expect 😛
10:59:37 <TrueBrain> so yeah, it is very much like reddit
11:00:25 <andythenorth> that is probably the definitive answer
11:00:28 <andythenorth> set of answers *
11:00:56 <TrueBrain> ChatGPT now tells me we are a bunch of failures, and shouldn't promote PBS
11:01:16 <TrueBrain> in very kind words it tells us to fuck off
11:02:36 <TrueBrain> `For players who are new to OpenTTD and are still learning the basics of the game, block signals might be a good choice to start with. Block signals are generally easier to set up and use than path-based signals, which can make them more suitable for players who are still learning the game.`
11:02:50 <TrueBrain> pretty sure PBS signals are easier to setup 😛
11:03:31 <TrueBrain> it does understand what PBS is btw, which is kinda cool 🙂
11:03:36 <andythenorth> we are entering a world in which scaling 'wrong' will be practically free 🙂
11:03:46 <andythenorth> whereas before, you had to pay people, or rely on volunteers
11:04:33 <LordAro> dare you to put this on reddit
11:05:06 <Rubidium> does it have a working solution to speed up code scanning in github? :)
11:05:24 <andythenorth> TrueBrain: please ask it why I can't build OpenTTD on macos
11:06:05 *** CplBurdenR3294 has joined #openttd
11:06:05 <CplBurdenR3294> I am not sure making an AI feel shame, which will lead to guilt, anger, and then rage, is a good idea 😄
11:06:35 <TrueBrain> andythenorth: `I'm not sure why Andy is having trouble building OpenTTD on MacOS. `
11:07:02 <andythenorth> I was quite hopeful it would find the answer
11:07:12 <TrueBrain> it does give 3 suggestions, which ar enot even weird
11:07:19 <TrueBrain> it even knows we depend on zlib, liblzma and liblzo2
11:07:28 <TrueBrain> funny guy, this ChatGPT 😛
11:07:39 <LordAro> it's unlikely to have pulled liblzo2 out of its hat
11:07:53 <andythenorth> ask it about math.h header failures on mac builds, using brew
11:08:06 <andythenorth> I don't know the exact search terms to give it 😛
11:08:15 <andythenorth> "chat bot commands as a service"
11:08:35 <TrueBrain> you don't "search". You just talk to it like it is a human
11:08:57 <TrueBrain> `Without more information, it is difficult to say exactly what the problem might be.`
11:23:14 <glx[d]> Switch to vcpkg Andy, it just works
11:24:11 <andythenorth> oh there's a mac build of it? 😮
11:24:57 <andythenorth> "install via brew"
11:25:07 <andythenorth> that seems like it will just compound the problem 😛
11:25:59 <andythenorth> I don't understand C++ module search
11:26:13 <andythenorth> vcpkg recommends installing as a submodule in the project
11:26:34 <andythenorth> I don't want to just remove brew from my OS because other things depend on it
11:27:05 <andythenorth> can I install vcpkg and then have openttd cmake use the path to it?
11:29:59 <TallTyler> Is Samu happy, or are road vehicles incorrectly getting the weight multiplier?
11:30:23 <JGR> That will be fixed shortly
11:30:26 <Samu> the freight multiplier is a train only thing
11:40:00 <andythenorth> ok I'm completely guessing here
11:40:11 <andythenorth> I installed zlib with vcpkg
11:40:17 <andythenorth> then ran `cmake -B . -S .. "-DCMAKE_TOOLCHAIN_FILE=../../vcpkg/scripts/buildsystems/vcpkg.cmake"`
11:40:32 <andythenorth> and now zlib is reported by cmake as the vcpkg version, not the brew version
11:40:37 <andythenorth> repeat for all deps?
11:44:29 <andythenorth> hmm cmake is from brew though
11:50:04 <andythenorth> ok removing some of brew just broke lots of things
11:54:00 <Samu> with the capacity fix, my AI is now smarter, it's chosing better vehicle models :)
11:54:22 <andythenorth> does it know about variants?
11:54:57 <Samu> if they get their own engine ID, i guess it does
13:06:08 *** WormnestAndroid has joined #openttd
13:14:50 <reldred> Hey folks, running into an issue compiling someone elses NML, this is a released .grf so I'm not sure why I'd be running into issues? Error message from nmlc is "Temporary registers from 128 to 255 are reserved for NML's internal calculations."
13:16:20 <reldred> Shouldn't be too old. It's V5 of Fridaemons object set.
13:16:37 <glx[d]> Because the 128-255 reserve is not that recent
13:17:31 <reldred> End of the day it's not that big a deal, I ran into it with my own NML but the section I cut out doesn't need the particular switch blocks that are causing that error
13:17:40 <reldred> I just thought it was interesting is all
13:18:25 <JGR> You can change the registers used to be < 128 to make the issue go away
13:18:50 *** D-HUND is now known as debdog
13:19:11 <JGR> So long as the IDs you pick aren't also used for something else at the same time
13:19:42 <petern> Yay bought more vinyl.
13:20:51 <petern> May need a sticker to avoid the cover deadnaming the artist though.
13:22:05 <glx[d]> The check is more recent than I remembered
13:22:11 <reldred> Yeah it's all good, I gotta write brand new switch blocks for this anyway
13:24:54 <Samu> just noticed there's no load/unload delay when a road vehicle doesn't fit the loading bay platform
13:25:53 <JGR> Road stops aren't rail platforms and the over-length penalty doesn't apply
13:31:43 *** Flygon_ has joined #openttd
13:32:28 *** Flygon has quit IRC (Ping timeout: 480 seconds)
14:17:48 *** sla_ro|master has joined #openttd
14:29:01 <TrueBrain> peeve .. that is a new word to me 😄 TIL!
14:34:10 <TrueBrain> ChatGPT made the worst sin .. it claims OpenTTD stands for Open Transport Tycoon Deluxe 😦
14:34:41 <TrueBrain> it claims OpenDUNE and OpenTTD have no overlap in developers ..
14:51:23 <glx[d]> Though in current state it's true
14:54:50 <frosch> imagine it answered "xaroth" 🙂
15:33:06 *** Smedles has joined #openttd
15:34:22 *** supermop_Home has joined #openttd
15:39:52 *** Fairy has quit IRC (Quit: User went offline on Discord a while ago)
15:43:45 <Samu> I'm fairly sure VEHICLE_PROFIT_MIN_AGE is not 800, let me find
15:47:59 <Samu> I asked on the forum about what he intended to do with AIGroup.GetProfitLastYear, still waiting for the answer
15:49:20 <Samu> the exlusion part is confusing
15:52:46 <Samu> feels like RailwAI wants to find the group with the worst average profit per vehicle, but exclude groups which are younger than 800 days. But since groups have no age property, he does it by checking the age of the first vehicle in a group.
15:54:09 <Samu> so, in a way, seems like he wanted to only include "old enough" groups on his search for the worst average profits
16:17:55 *** TROILUS6 has joined #openttd
16:20:21 *** Flygon__ has joined #openttd
16:23:01 *** TROILUS has quit IRC (Remote host closed the connection)
16:23:01 *** TROILUS6 is now known as TROILUS
16:27:58 *** Flygon_ has quit IRC (Ping timeout: 480 seconds)
16:40:50 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
16:53:33 <TallTyler> I edited my comment to reflect VEHICLE_PROFIT_MIN_AGE = 730
17:00:25 *** Flygon__ has quit IRC (Read error: Connection reset by peer)
17:21:27 <TallTyler> A perfect candidate to try auto-merge 🙂
17:21:54 <TallTyler> New contributor, so I had to manually start CI which are now running
17:22:04 <TallTyler> And a total no-brainer for approval and merge
17:31:46 <Samu> what is SetFill (x, y) on a NWidget?
17:41:07 <JGR> It's just a layout change, there is no functional change
17:41:51 <Samu> rebasing my stuff and i stumbled upon there, with conflicts :)
17:43:53 <Samu> there were also string renames
17:55:40 *** supermop_Home has quit IRC (Ping timeout: 480 seconds)
18:11:06 *** supermop_Home_ has joined #openttd
18:43:41 <DorpsGek> - Update: Translations from eints (by translators)
18:53:58 *** gelignite has joined #openttd
18:59:57 <Samu> since you're making rewords to the AI stuff
19:41:23 <Samu> truebrain comment scares me
19:42:12 <Samu> so, I'm gonna try dissolve it into several commits
19:43:00 <TallTyler> Small PRs are less scary for us
19:43:25 <TallTyler> Also, you don't have to remove strings from other languages, eints will take care of that
19:45:15 <TallTyler> Also, it's important to use the PR template to explain what you're doing, and why. Screenshots are also helpful. A PR that doesn't explain why is confusing, and we can only guess at your intentions. It's much better for you and us if you explain why you think a change should be made.
19:52:43 <petern> Small PRs like Daylength 🙂
19:53:24 <TallTyler> I expect NoDayLength to be a scary PR 🙂
19:53:53 <TallTyler> Maybe NotDL to make pronounciation even harder
19:56:33 <Samu> the point of the PR is to eliminate the need to control when "star_date" parameter is pushed into an AI config, then control it when it's to be randomized or deviated, also to make it more predictable when loading from a savegame, what it initializes to, make the savegame thus more reproducible
19:57:53 <Samu> by making it a game setting, it is removed from the AI config, no more need to babysit it
19:58:16 <TallTyler> Makes sense, make sure that goes in the PR though because nobody's going to scrolling through IRC/Discord to find this message when reviewing 😉
20:11:19 <glx[d]> start_date being an AI setting was wrong from the start
20:12:16 <glx[d]> and it's kind of a hack
20:18:48 <Samu> how to insert a commit between commits, oh boy
20:19:50 <andythenorth> rebase -i, reorder
20:19:53 <andythenorth> won't always work
20:20:02 <andythenorth> only if the commits are orthogonal
20:20:16 <Samu> it has 1 commit, i want to finish it with 2 commits :)
20:20:23 <Samu> or more, apparently more for this case
20:20:29 <andythenorth> rebase -i is a good tool to learn
20:20:34 <andythenorth> but is easy to get in a mess with
20:21:37 *** Wormnest has joined #openttd
20:22:16 *** WormnestAndroid has joined #openttd
20:31:42 <glx[d]> to split a commit I often use rebase -i, then duplicate the commit and set to edit the first instance
20:33:07 <glx[d]> then I remove the unwanted stuff, and continue rebase which should reapply the removed stuff but in another commit
20:53:01 <andythenorth> goes it bring back train view?
20:53:23 <andythenorth> 'add vehicle', 'remove vehicle'
21:00:17 <imlegos> ...Is that pre-release TTD?
21:02:58 <Rubidium> very pre-release then, as it seems to even predate TT (original)
21:03:44 <petern> With the A-Train-esque graphics.
21:08:02 <petern> Practically open source 🙂
21:09:23 <petern> Interesting the small font there is "antialiased"
21:34:12 <TallTyler> Yeah, that's significantly slower than the current 2.2 seconds per day
21:52:01 <Samu> okay, i think i have the PR ready
21:52:07 <Samu> maybe still hard to read
21:52:12 <Samu> but i split into 3 commits
21:52:29 <petern> Are you doing more than one thing again?
21:52:58 <Samu> i dont know how to split more :(
21:55:10 <petern> Is this economy days or calendar days? 😄
21:56:54 <glx[d]> hmm I wanted to add a flag to search scripts only in binary dir for regression, but that looks redundant with -X
21:57:14 <glx[d]> but -X also skips searching baseset in global dirs
22:03:12 <Samu> btw, I completely ignored the saved start_date from old savegames
22:03:22 <andythenorth> petern: goes it substring thing for variant names?
22:03:51 <Samu> by "works" i mean, doesn't crash openttd, nor is it required for anything
22:04:44 <Samu> AIs still load the other settings properly
22:06:05 <Samu> openttd.cfg becomes cleaner
22:06:45 <Samu> no more start_date in each AI configuration
22:08:07 <Samu> under [ai_players] section
22:15:11 <Samu> wondering if I'm missing anything
22:20:53 *** keikoz has quit IRC (Ping timeout: 480 seconds)
22:21:18 <petern> andythenorth: Not sure, there's nothing like that anywhere else.
22:21:50 <andythenorth> name string callback, but with context in the callback info var?
22:22:21 <petern> No, running a callback for every string in the list isn't a good idea.
22:39:41 <glx[d]> string handling is already slow enough 🙂
22:41:01 <petern> andythenorth: what variants (hah) of engine name are required?
22:44:31 <andythenorth> when it's group parent or not
22:44:41 <andythenorth> and when it's displayed in vehicle info window
22:44:46 <andythenorth> and then as child in buy menu
22:44:53 <andythenorth> it's not essential, might just be nicer
22:45:22 <andythenorth> - "Company colour"
22:46:07 <petern> What if it's a child and a parent?
22:46:23 <andythenorth> I was really hoping you wouldn't ask that
22:46:38 <andythenorth> grf could check both and use AND or something 😛
22:58:24 *** Samu has quit IRC (Read error: Connection reset by peer)
23:06:24 <Samu> I'll try to be more descriptive tomorrow
23:06:34 <Samu> Running out of time today, cyas goonight
23:41:37 <andythenorth> wonder if grfcodec can compile native to ARM
23:44:29 <andythenorth> hmm doesn't built at all currently for me 🙂
23:47:08 <petern> ... managed to fill in the template incorrectly
23:54:02 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
23:54:33 *** nielsm has quit IRC (Ping timeout: 480 seconds)
23:56:48 *** Tirili has quit IRC (Quit: Leaving)
continue to next day ⏵
