IRC logs for #openttd on OFTC at 2025-10-20
β΄ go to previous day
00:18:29 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
00:23:04 *** WormnestAndroid has joined #openttd
02:13:40 *** gnu_jj_ has joined #openttd
02:17:14 *** gnu_jj has quit IRC (Ping timeout: 480 seconds)
04:02:02 *** Zathras has joined #openttd
04:05:26 *** Zathras_11 has quit IRC (Ping timeout: 480 seconds)
04:05:36 *** WormnestAndroid has quit IRC (Remote host closed the connection)
04:05:49 *** WormnestAndroid has joined #openttd
06:26:54 <mmtunligit> I'm trying to build openttd from source with cmake but i keep getting the error
06:26:54 <mmtunligit> ```CMake Error at CMakeLists.txt:7 (project):
06:26:54 <mmtunligit> GHS_TOOLSET_ROOT directory "/usr/ghs" does not exist.```
06:26:54 <mmtunligit> which i have been unsucessful in figuring out how to solve with google. I'm running kubuntu. should i continue trying to get cmake to work or is there a different way i can build it
06:28:15 <mmtunligit> all i know is the error means its having trouble locating the compiler, is that something i need to install seperately?
06:51:03 *** SigHunter has joined #openttd
07:06:15 <DorpsGek> - Add: summary for week 42 of 2025 (by OpenTTD Survey)
07:07:44 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
07:08:30 *** WormnestAndroid has joined #openttd
08:01:28 <peter1138> Oh a Debian-based system y ou can install build-essential to get the basics required for compiling c/c++.
08:02:56 <LordAro> hmm, no mentions of that in COMPILING.md, though i guess there's a certain "already has a functioning C/C++ development environment" assumption
08:03:10 <mmtunligit> yes im pretty sure that's my issue
08:03:14 <LordAro> Greenhills is unlikely to be a compiler that you want to use though, wonder how cmake has picked that up...
08:03:14 <mmtunligit> im very out of my depth here
08:04:34 <ahyangyi> I don't know but you might want to undo the tutorial that led you to install Greenhills, which might had made the issue more complex
08:05:12 <mmtunligit> i never installed greenhills, it seems to come with cmake
08:05:49 <LordAro> it definitely doesn't
08:05:52 <LordAro> but detection for it does...
08:07:06 <peter1138> Could just be a last ditch attempt by cmake in the absense of the normal build stuff.
08:08:18 <mmtunligit> so sounds like what i actually need help with is setting up a dev enivornemnt?
08:14:29 <mmtunligit> peter1138: ok i did this and tried building with cmake using unix makefiles and got farther but now ive run into this error
08:14:29 <mmtunligit> ```CMake Error at CMakeLists.txt:186 (message):
08:14:29 <mmtunligit> SDL2 or Allegro is required for this platform```
08:15:19 <peter1138> You need a c++ compiler, which it seems you know have. You also need the dependencies, which on Debian-based systems are installed via lib...-dev packages.
08:15:59 <peter1138> The dependences are listed at the top of COMPILING.md, though you should ignore breakpad tbh.
08:17:13 <mmtunligit> i thought i already installed the required packages but maybe i have a wrong version or something? ill try installing again
08:19:15 <mmtunligit> i tried ```sudo apt install libsdl2-2.0-0``` but that doesnt seem to have helped anything
08:19:48 <LordAro> you've installed the library itself, not the headers needed to be linked against
08:21:04 <peter1138> Yup. The files to use the libraries and the files to compile with the libraries are split into separate packages, as you don't need all of the latter.
08:24:12 <mmtunligit> ok that seemed to actually get it to build, though not without a few errors
08:26:46 <mmtunligit> well it got it to build *something* but whats inside the build folder i have no clue if its correct or not
08:27:41 <reldred> yeah work your way through the errors, usually at the start of the cmake process, look for the 'abcd not found' errors, find the dev package, install it, delete the build folder and start again.
08:28:18 <reldred> like if you're missing one of the libs you can't open compressed files like scenarios and saves
08:28:24 <reldred> but the game will still run
08:28:28 <LordAro> without seeing what you're seeing, it's difficult to be sure :)
08:28:36 <LordAro> crystal ball is in the repair shop
08:29:02 <mmtunligit> reldred: i cant find the game in the build folder
08:29:20 <peter1138> Yeah, with build/compile errors, you need to start at the top instead of the looking at the last thing. Problems cascade.
08:29:31 <reldred> i didn't say that's what your *current* fuck up is
08:29:40 <LordAro> my bet is that you've only run cmake (which is the 'configure' step), not the actual build
08:29:43 <reldred> i just said its one possible fuckup
08:30:11 <peter1138> If it actually completed successfully, there'll be an `openttd` binary (alongside other stuff.)
08:30:24 <mmtunligit> the issue here is all th "could not find" are generic names for the packages, not the specific thing i need to install based on my experience with SDl2
08:30:41 <reldred> mmtunligit: yeah, welcome to linux
08:31:07 <LordAro> `apt search` (or package manager of choice)
08:31:08 <reldred> openttd is comparitively easy to compile on linux compared to some things.
08:31:15 <peter1138> It's not specific package names because those vary depending on distribution and version.
08:31:21 <reldred> yee, apt search is useful
08:31:29 <mmtunligit> LordAro: ive done that but that comes up with a huge list and i dont know what i need form those lists
08:31:35 <reldred> but distros sometimes like to rename things because fuck you
08:31:40 <reldred> usually whatever ends in -dev
08:31:46 <LordAro> filtering on -dev likely
08:32:02 <LordAro> (and starts with lib)
08:32:09 <reldred> and every distro and version of every distro is different
08:32:12 <reldred> also because fuck you
08:33:11 <LordAro> and yet it's still better than Windows
08:33:37 <reldred> at least you're not having to deal with vcpkg
08:33:47 <reldred> :cryingbloodbutstillcool:
08:34:13 <peter1138> Someone sounds emotional this evening.
08:34:31 <reldred> look I think I've just had enough of computers
08:34:33 <LordAro> at work our windows dependencies are handled via chocolatey, pacman (MSYS), cpan & uv
08:34:51 <LordAro> it's a wonderful situation
08:34:51 <reldred> at least I've made peace with needing to enter a password to unlock keep ass
08:35:16 <reldred> the computers themselves aren't even my problem at my new job
08:35:28 <reldred> just all the copper and glass that goes to them
08:35:50 <LordAro> all the blinkenlights
08:36:10 <peter1138> You can complain about having to remember IPv6 addresses.
08:36:28 <reldred> and if my work laptop shits itself I call off work for the day, ride an e-scooter to the actual office, go to the den of the it support workers, and dump my laptop with them
08:36:36 <reldred> peter1138: i don't even have to do that!
08:37:12 <reldred> that said, i did have to remind a senior network engineer today that the devices won't show up in show cdp/lldp neighbors if he doesn't no shut the port first
08:37:34 <reldred> i haven't touched a console in anger in like over three months now
08:38:10 <LordAro> infrastructure-as-code and similar?
08:38:23 <reldred> no, I'm literally just overpaid overqualified hands and feet
08:38:42 <reldred> granted they expect me to be smarter than just hands and feet
08:38:42 <mmtunligit> oh my god im touching computer
08:38:49 <mmtunligit> i think its actually building
08:39:08 <mmtunligit> ive only installed the required packages but thats all i care about right now
08:40:19 <andythenorth> I guess we find out how much we depend on AWS US-east eh
08:40:40 <reldred> has aws shit itself again
08:40:54 <LordAro> how have we got this far without mentioning that yet? i've known for an hour
08:41:52 <reldred> oh I clocked out an hour ago and apart from microshart stuff we host all our own kit, and I'll get a call on my mobile if either of our two DC's gets sad.
08:43:35 <reldred> 'happy us-east-1 day for those who celebrate'
08:45:53 <reldred> anyways i get to actually go to the dc tomorrow and do some actual work instead of twiddling my thumbs. i didn't go today because they didn't approve my access request that I filed last week until like 2-3pm which was a waste of time then going, and tomorrow i have lots of copper noodle cables to run between racks.
08:46:36 <reldred> and a psu to replace for a critical server that I can't do because someone else has to go do some bureaucracy before he allows me to do it.
08:46:44 <reldred> its great because it's not my problem π
08:49:20 <peter1138> That is not US-East-1 I assume.
08:49:37 <peter1138> (As it has OVH plastered aver it)
08:50:01 <reldred> I can only hope that US-East-1 is on fire
08:50:13 <peter1138> That was OVH in 2021 or something like that.
08:51:53 <peter1138> Centralising the Internet was such a good idea.
08:52:25 <reldred> just, not whatever datacentre bcicorporate.com's VPS is hosted in.
08:52:46 <reldred> and not either of the two datacentres in brisbane that I deal with are in
08:52:53 <reldred> because that would create a lot of work for me
08:52:58 <reldred> i guess it would be exciting though
08:53:54 <alfagamma7> Damn seems like Mr Red is in a bad mood today
08:54:16 <reldred> I'm actually in a fantastic mood
08:55:39 <LordAro> yeah, i've seen nothing to suggest reldred is unhappy :p
08:55:39 <mmtunligit> if i want to reference the value of a setting in terraform_gui.cpp is there anything extra id need to add to the #include section or is everything i need there already?
08:56:12 <mmtunligit> thatw was the one bit i wasnt able to figure out last night
08:57:09 <LordAro> i'd expect it to already be there
08:57:19 <LordAro> but it'll be table/settings/something otherwise
08:58:05 <LordAro> peter1138: wasn't that the DC that turned out to be wood framed with the airflow structured perfectly to make a fire impossible to stop?
08:59:51 <peter1138> Looks like a load of shipping containers to me.
09:00:00 <LordAro> i may be misremembering
09:00:17 <reldred> cloud is easy, booming industry, can't possibly lose money
09:01:11 <LordAro> >The Bas-Rhin fire service says that the SBG2 data center had no automatic fire extinguishing system and no general electrical cut-off switch. The facility also had a wooden ceiling rated to resist fire for only one hour, and a free-cooling design that created "chimneys" that increased the fire's ferocity.
09:01:41 <reldred> i especially like slagging on cloud nowadays because my livelihood now depends on cloud being utterly unsuitable for my corporate overlords particular business case.
09:01:57 <LordAro> i can't entirely blame them for the airflow design though
09:03:02 <reldred> yes i'm glad that none of the dc's i deal with have gone too far off the deep end with that crap
09:03:26 <reldred> we've got a couple racks in one of our dc's that are crammed full of gpu's but they're not for ai, they're for vdi
09:04:21 <mmtunligit> now to see if my change works, RECOMPILE
09:05:04 <LordAro> oh, thin clients everywhere?
09:05:17 <reldred> vmware/omnissa horizon, and various other similar technologies
09:05:35 <reldred> LordAro: or software vdi clients
09:05:48 <reldred> saves having to get laptops with quadros
09:06:38 <reldred> or having to let the terabytes of proprietary data that the gpu is crunching through have to leave your controlled network
09:07:34 <LordAro> with our latest round of workstations Dell has decreed that to get the fastest CPU (Ultra 9 whatever) you need the "T2" upgraded chassis... which will only take the 1500W PSU
09:07:53 <LordAro> it does feel excessive, given we have no GPU requirements at all, just onboard
09:08:05 <reldred> 3d visual effects production companies like it, they'll use vdi implementations that can automatically watermark the screen data so if people leak the new marvel movie before it releases they can throw the correct person into the pits of hell
09:08:46 <reldred> its great for the corporate security panopticon as well
09:09:09 <peter1138> Cool cool, I wonder how many computers with PS/2 ports meet the requirements for Windows 11.
09:09:31 <reldred> something something a third of all code is now written using genai something something?
09:09:31 <ahyangyi> PSU efficiency peaks at about 50% load so having some headroom helps, though 1500W is probably still an overkill
09:11:04 <reldred> do i want to order takeout for dinner and stay in or do i want to go ride an escooter around for an hour and then get a cheeseburger from maccas on the way home
09:11:19 <reldred> it sounds like another thunderstorm is rolling in again
09:11:27 <LordAro> ahyangyi: given it can max out at 250W...
09:11:39 <reldred> which sounds like fun to get stuck in
09:14:18 <ahyangyi> LordAro: "Built for future expansions"
09:14:42 <reldred> yeap, time to go adventure
09:16:57 <peter1138> I'm hearing rumours that the issue is...
09:18:38 <kuhnovic> Is there a reason why towns have to be at least 12 tiles away from the map edge? Other than "towns can't grow"
09:19:15 <peter1138> I imagine it's to let them have some room in all directions.
09:19:40 <kuhnovic> I'm considering lowering it just to generate my Southern Norway Map, but I don't want to end up with a broken map
09:20:55 <xarick> have you ever come across a lake without a river pointing to it?
09:21:06 <mmtunligit> my change seems to work, ill try figuring out how to submit a PR now
09:22:31 <peter1138> In the standard github workflow, you fork OpenTTD, push a branch to your fork, and create a PR from that.
09:24:44 <mmtunligit> peter1138: first i need to figure out if i have github installed(?)
09:25:36 <peter1138> You don't need github installed.
09:25:48 <peter1138> Github is a website.
09:26:12 <xarick> interesting bug: no lake, only rivers
09:26:20 <peter1138> If you cloned with git, then you have git installed.
09:26:51 <mmtunligit> i did not clone with git
09:27:11 <peter1138> If you didn't then you need to install git, and then start again with a git clone, because you don't currently have a version-tracked tree.
09:28:53 <xarick> wish I could enumerate these rivers
09:29:11 <xarick> river 1, river 2, river 3, so i could track which river is which
09:29:24 <xarick> the order they were generated
09:31:06 <andythenorth> river name generator?
09:31:17 <andythenorth> hmm.....that might be nice flavour π
09:31:22 <andythenorth> stem them from a nearby town
09:32:58 <xarick> it will be buggy, i wanted to prove some of the rivers are frankenstein rivers
09:35:52 <mmtunligit> at what point in this process do i actually edit the file to make the change?
09:36:18 <mmtunligit> wait im dumb thats step 4 right
09:42:55 <mmtunligit> have i formatted step 3 correctly? ```git checkout upstream/master -b <Change:><Adjust placement of terraform toolbar if linking enabled>
09:43:34 <xarick> std::to_string is safeguarded, use fmt::format... well how?
09:44:20 <LordAro> mmtunligit: branch names aren't formatted strings
09:44:36 <LordAro> adjust-terraform-toolbar would be fine
09:45:11 <xarick> `std::string river_number = fmt::format("{}", _river_number);`
09:45:11 <xarick> does this look correct?
09:46:27 <mmtunligit> ```git checkout upstream/master -b <adjust-terrafrom-toolbar>
09:46:27 <mmtunligit> bash: syntax error near unexpected token `newline'```
09:46:27 <mmtunligit> im doing something worng
09:46:45 <LordAro> you're not supposed to put in the `<>` :)
09:46:50 <LordAro> they have special meanings in bash
09:48:27 <mmtunligit> theyre red to indicate you *shouldnt* include them i see
09:49:11 <mmtunligit> ```git checkout upstream/master -b adjust-terrafrom-toolbar
09:49:11 <mmtunligit> fatal: this operation must be run in a work tree```
09:50:05 <mmtunligit> ok i figured it out
09:50:13 <LordAro> welcome to git error messages :)
09:50:19 <LordAro> they're often somewhat obtuse
09:50:20 <mmtunligit> was in the wrong directory
09:50:35 <locosage> kuhnovic: I don't remember there being anything in the code that would break, so should be fine to have towns on edges
09:50:42 <locosage> until someone decides to change something :/
09:51:50 <kuhnovic> I think I'll take the chance π
09:53:36 <mmtunligit> so now i just edit the file i need to change like normal right? or do i have to do that in the command line too, im kinda confused by what step 4 is asking of me
09:54:15 <LordAro> well, you'll need to learn how to commit the changes, which isn't directly specified
09:54:54 <mmtunligit> yeah ok so there is another bit assuming i know how to do something that i dont
09:57:27 <ahyangyi> The git workflow is kind of strange at first (because we didn't know what to expect), but once you get it, it makes lots of sense.
09:57:43 <ahyangyi> But the command format and error messages are still a bit obtuse π
09:58:29 <ahyangyi> LordAro: I also had this in mind π
09:58:39 <mmtunligit> so i open the file in /openttd/src add my change, save and then what
09:58:46 <mmtunligit> if im understanding correctly
09:58:47 <peter1138> Oh heaven forbid we make changes to the game.
10:00:05 <dwfreed> peter1138: changes?! openttd *must* be bug-compatible with TTD!!!1!!1111!
10:01:18 <LordAro> but read the other one too
10:01:25 <mmtunligit> can you tell im new to this?
10:02:14 <LordAro> couldn't possibly say
10:02:29 <LordAro> there is no issues being new
10:02:29 <yiffgirl> we were all new once
10:03:26 <LordAro> when i was super-brand-new, i constructed patch files by hand
10:03:51 <LordAro> i mean, it taught me about the patch file format, but not a lot else
10:04:36 <LordAro> peter1138: hey, making local commits in svn era was difficult!
10:04:56 <peter1138> Impossible, in fact.
10:05:04 <peter1138> But `svn diff` was not.
10:05:21 <peter1138> I feel sorry for the poor unfortunates who invested time in mercurial.
10:05:32 <xarick> 1, 2, 3, 4, 5, 6 are, in our eyes, the same river, but in openttd code, that's 6 rivers
10:06:17 <peter1138> That's seems to be because you are counting each segment individually.
10:06:25 <peter1138> Instead of each river.
10:12:38 <xarick> ok gonna start the counter before FlowRiver function
10:13:45 <mmtunligit> LordAro: this part of that link is what i do in step 6 of contributing.md correct?
10:14:36 <mmtunligit> er, realizing irc users might not be able to view images
10:14:40 <mmtunligit> ```# push changes to github
10:14:40 <mmtunligit> git push --set-upstream origin my-branch```
10:15:16 <peter1138> It's not mandatory to do that.
10:15:24 <peter1138> You can just always specify the remote and branch.
10:15:42 <peter1138> That behaviour is not the default now because some people had issues with it.
10:27:16 <mmtunligit> ok now im having trouble pushing, i set up 2FA but when i run the push command i get a popup on my computer asking for my username and then my password, but regardless of what i put in the password field, either my password or the one time code from the authenticator, i get ```remote: Invalid username or token. Password authentication is not supported for Git operations.
10:27:59 <LordAro> mm, you can't use https auth with 2FA
10:28:59 <mmtunligit> ok then what am i supposed to do
10:29:51 <LordAro> ssh auth will be best for you, as you're on linux
10:30:07 <xarick> I'm honestly impressed, out of 8193 rivers only one was malformed
10:31:51 <xarick> or in other words: river 26 bfs part didn't find a place to flow down to. so it tried to make a lake, it made the lake, then yapf tried to connect the lake to the start of the valley, but it was too far away... max_search_nodes limit was reached, and no connection was made
10:32:29 <xarick> lake is alone, and river ends abruptly
10:34:18 <xarick> the good news is that... there's some resilience in this river algorithm
10:34:35 <peter1138> LordAro, weirdly, I can? Hmm.
10:34:57 <xarick> other rivers eventually ended finishing in that lake
10:35:31 <xarick> and the river 26 ending abruptly was merged with other rivers
10:35:33 <peter1138> Presumably a have an access token somewhere though.
10:38:02 <xarick> the other rivers are un numbered because I ran out of signs
10:40:14 <mmtunligit> should i set up the SSH key as an authenticator or a signing key?
10:42:45 <LordAro> mmtunligit: authenticator
10:45:51 <mmtunligit> ok i got that set up i think, now how do i actually use it when pushing
10:46:01 <mmtunligit> the docs dont seem very clear on that to me
10:46:30 <LordAro> indeed, you'll need to switch the remote url
10:46:51 <LordAro> it'll be something like `git remote set-url origin git@github.com:myusername/openttd.git`
10:49:17 <rito12_51026> I have added all data for NewGRFs - rail-infra set from survey results and got only 39.4 %, what does the other 60.6 % represents?
10:51:17 <peter1138> What do you mean by "added all data"?
10:52:07 <peter1138> The percentages shown in the survey results are described in the text at the top of the page.
10:54:00 <rito12_51026> but it describes the usage of setting not grfs
10:56:52 <rito12_51026> peter1138: like that but for all rail-infra sets
10:57:41 <peter1138> > The following statistics are based on "seconds played". That is to say, if a setting was "true" for 100 seconds of play-time and "false" for 200 seconds of play-time, it would be 33% true and 66% false. This is to avoid biasing the results towards shorter games.
10:57:56 <peter1138> What part of that do you think doesn't apply?
10:58:31 <mmtunligit> phew that took a lot out of me
10:59:44 <rito12_51026> i'm not sure how value of setting translates into usage of a newgrf
11:00:14 <peter1138> rito12_51026, it's "how many seconds of playing uses this NewGRF"
11:00:41 <peter1138> Adding up the values is therefore nonsense.
11:01:40 <peter1138> "Metro Track Set" at 5.0% means that 5% of playing time used that NewGRF, and 95% of playing time did NOT use that NewGRF.
11:02:11 <rito12_51026> now i understand
11:02:12 <reldred> and if someones using two track sets your numbers end up even more screwy trying to add them together
11:04:44 <mmtunligit> thank you all for your patience while i figured out how to do all this
11:04:53 <mmtunligit> i would nto have been able to do it without you
11:05:40 <LordAro> nice to actually be able to help someone for a change...
11:05:47 * LordAro looks at no one in particular
11:05:52 <reldred> thanks for not being offended at me slagging off the state of linux package management π
11:07:17 <mmtunligit> reldred: lmao thats fine, ive never really installed packages via CL before so i have nothing to base my opinion of it on
11:07:46 <reldred> yeah once you've been at it for a while you'll find things that tick you off π
11:08:36 <xarick> dear kuhnovic I need your yapf expertise. I need all the 10000 items from the closed nodes when the pathfinder reaches the max_search_nodes limit. a vector of 10000 TileIndexes, for lake guessing purposes. How do I do this?
11:09:23 <mmtunligit> reldred: this may well be the only time i ever actually code something, i dont touch computer because i want to, id much rather touch grass
11:10:15 <peter1138> Dear exception, why are you occuring in a line that no longer exists...
11:10:40 <LordAro> usually it's because you forgot to rebuild :p
11:11:16 <LordAro> though i did have one the other day with lambdas that seemed to refer to the previous lambda in the file rather than the one that actually had the exception...
11:11:23 <reldred> mmtunligit: computers may actually be a mistake
11:11:40 <reldred> but i depend on them for my livelihood so
11:11:45 <mmtunligit> but you can get cat pics whenever you want so whos to say
11:12:13 *** notluke2578 has joined #openttd
11:12:13 <notluke2578> mmtunligit: This is exactly how we have been corrupted
11:12:16 <notluke2578> But we canβt go back now
11:12:33 <peter1138> LordAro, I have a suspicion it's running an old .net 6 build instead the current .net build.
11:23:25 <kuhnovic> xarick: The Node List is just a public member called `nodes`, should be accessible
11:24:02 <rito12_51026> So with the corrected calculations have estimated that in the 2nd quarter of 2025 only 2.23% OpenTTD 14 players have used rail-infra sets.
11:24:29 <kuhnovic> NodeList doesn't currently have a way to iterate the items, you'd have to make that yourself
11:24:59 <reldred> rito12_51026: yeah the majority of players don't run grf's at all
11:25:10 <reldred> niche problems for a niche set of players
11:25:30 <reldred> compare the numbers with the 15.0 betas as well, gives a different perspective.
11:25:58 <rito12_51026> reldred: only 51.4% so it is more like a half
11:27:04 <reldred> ah it's higher than i remember
11:30:55 <locosage> how did you get that 51%
11:32:17 <rito12_51026> for comparison FIRS(from v1 to v5) estimates to 0.68%
11:33:26 <peter1138> Someone there with 222 NewGRFs...
11:33:34 <locosage> hm, wonder what does that count...
11:36:49 <peter1138> Looking at the code it seems to be counting available NewGRFs, not in use?
11:37:27 <reldred> hmm, still, 51.4% of play time is with installs that don't even have a single newgrf installed let alone active
11:37:31 <peter1138> last_newgrf_count = 1235
11:38:37 <ahyangyi> peter1138: I wonder if it's me π
11:39:22 <peter1138> Okay, I'm not sure that newgrf_count is related to the newgrf_count in game.
11:39:49 <peter1138> The code has no mention of "ai_count".
11:39:59 <peter1138> So I guess it's calculated by the survey system instead.
11:40:04 <locosage> no, it's not the last count
11:40:13 <locosage> it's just the amount of active items in grfs section
11:40:31 <reldred> wtf an update got posted for 2ccTrains?
11:42:10 <peter1138> Is it breakfast time?
11:44:11 <reldred> i could do with a holiday
11:44:58 <reldred> alas the closest I'll get to it any time soon is taking a week off to drive my landcruiser across the country, I guess I'll probably stay in a few bed and breakfasts along the way?
11:51:26 <peter1138> Wonder if it counts static NewGRFs.
11:52:58 <rito12_51026> reldred: got 2.30% for 15.0 betas
11:54:37 <rito12_51026> will try JGRPP also
11:58:11 <reldred> probably similar to 15.0 betas
11:58:35 <rito12_51026> I don't think so
12:02:30 <peter1138> Nah, for JGRPP, no NewGRFs is at 1.5%
12:02:57 <reldred> oh there'll definitely be a different nuance to the .grf's used but the percentage using .grf's will still be quite dominant
12:03:09 <peter1138> What can I say, JGRPP players like shovelling mismatching graphics into the games.
12:03:57 <reldred> I'll have you know my grf loadouts are quite tasteful thankyou very much
12:04:33 <reldred> alas i haven't been playing much lately
12:05:54 <peter1138> (I'm joking, JGRPP players generally know what they are doing.)
12:06:52 <peter1138> That guy who keeps calling C++ C+, though...
12:09:23 <alfagamma7> Nvm I keep thinking about calculus
12:10:10 <ahyangyi> Is that a joke name for Java?
12:11:38 <rito12_51026> reldred: 2.73 for JGRPP
12:13:28 <ahyangyi> As a side note, is it a welcome change if it were possible to change the baseset during a game?
12:13:38 <ahyangyi> At least for single player games
12:17:44 <locosage> can't you already change it by restarting the game?
12:18:41 <peter1138> ahyangyi, I wouldn't object to it, if it's not massively intrusive.
12:19:05 <peter1138> The main issue is if changing it could cause a crash or something, losing the game.
12:23:12 <brickblock19280> Loading an invalid extra newgrf probably could
12:28:09 <xarick> closed_nodes are in a hashtable π
12:28:55 <xarick> and closed_nodes are under "protected"
12:29:48 <xarick> and the slots of the hashtable are also "protected"
12:30:36 <peter1138> We do allow changing music and sound basesets now, would be nice to allow changing graphics as well.
12:31:22 <peter1138> Reapplying NewGRFs basically reloads it anyway.
12:32:06 <peter1138> > Louvre Thieves Given Immunity After Confirming Jewels Stolen For Purpose Of Training AI Software
12:34:38 <xarick> too much abstraction for a TileIndex value
12:35:15 <ahyangyi> Let's add a layer of abstraction to solve that "too much abstraction" problem
12:35:29 <rito12_51026> peter1138: Training in what? In stealing?
12:38:10 <peter1138> I'm not sure the joke needs explaining.
12:42:15 <ahyangyi> You can ask an AI to explain the joke, and they will be able to, because someone already looted Louvre for joke explanations.
12:52:19 <xarick> according to the debugger, the HashTableSlot has unreachable code
12:54:07 <xarick> `inline const TItem *Find(const Key &key) const` - unreachable
12:54:15 <xarick> `inline TItem *Find(const Key &key)` - reachable
12:55:13 <peter1138> They are definitely reachable.
12:56:08 <peter1138> The const version is definitely reachable. It's just not used, so it's not included in the build.
12:56:43 <xarick> that's what I meant, probably
13:01:14 <xarick> these are not even hashes of TileIndexs... these are hashes of whatever a Node is made of which includes a TileIndex
13:01:32 <xarick> how am I gonna get a list of TileIndexes ...
13:02:38 *** Flygon has quit IRC (Read error: Connection reset by peer)
13:22:51 <rito12_51026> How to attatch heavy file to a bug report
13:34:04 *** Rikests has joined #openttd
13:34:12 <mmtunligit> on talking with a friend and finally figuring out how the window actually gets placed where it does in the first place ive realized i couldve done this in a cleaner fashion, testing for the (lack of the?) setting in the ```OnInitialPostion``` function instead
13:35:00 <mmtunligit> but it works as is and thats the important part
13:35:13 <peter1138> Oh yes, I enabled preview to test it.
13:36:15 <peter1138> But yes, it would be better to fix the OnInitialPosition instead of what you did.
13:36:47 <mmtunligit> i can make that change if youd like
13:37:26 <peter1138> I also think the behaviour shouldn't depend on linking.
13:37:29 <rito12_51026> peter1138: I'l rescale it as it does not need to be full hd
13:38:26 <mmtunligit> oh? im curious as to why
13:39:02 <peter1138> Basically I think you are looking for "place landscape toolbar alongside construction toolbar", and that whether its linked or not is not exactly relevant?
13:39:42 <peter1138> Like, opening the landscaping toolbar on its own leaves a gap.
13:39:48 *** johnfranklin has joined #openttd
13:39:55 <mmtunligit> peter1138: i think if its not linked its fine to put it below is the thing
13:40:17 <mmtunligit> since it would get covered up once you open up a construction toolbar, and then get put on top if ou refocused
13:40:44 <mmtunligit> peter1138: yeah that gap is my issue
13:41:42 <johnfranklin> Nearly f*cked up my cloud server by mount (some new disk) (some existing folder)
13:42:45 <mmtunligit> if it wasnt linked and it was up against the main toolbar and you opened up a construction menu id say it should probably get shunted down to where it is currently, but that sort of dynamic behavior is above my paygrade
13:43:33 <emperorjake> Rito13viaGitHub: This is why you always use non-stop orders
13:44:14 <emperorjake> I've said it before, but I think non-non-stop orders are useless and should be removed
13:44:54 <peter1138> emperorjake, it will never be removed, old savegames, and old-school players, may rely on the bevaiour.
13:44:59 <Rikests> Hey. I just finished playing a game with xUSSR trainset. Is there some other trainset that has similar detail, like cargo ageing multipliers, speed limits, catenary Voltage segregation etc as this xUSSR set? I usually increase the running costs by 8x, cargodist for pax on (would prefer cargodest, but I guess its not possible now), wagon speed limits on. Basically I like a challenging game where I really have watch the running costs and i
13:45:20 <peter1138> I thought there was a config setting that made it use non-stop by default, though.
13:45:26 <Rikests> For example 2cc seems a bit basic compared to xUSSR
13:46:52 <johnfranklin> Rikests: JP+, or China Set
13:47:03 <emperorjake> I get why non-non-stop orders can't be removed, but they could at least be hidden like presignals were
13:47:42 <Rikests> Alright, thanks johnfranklin.
13:48:03 <rito12_51026> Wow I didn't know that vehicles can stop on multiple stations
13:48:43 <LordAro> emperorjake: that's called 'changing the default'
13:48:46 <LordAro> which i'm pretty sure we did
14:07:28 *** Rikests has quit IRC (Remote host closed the connection)
14:09:54 <jfkuayue> Why is it emperor nasi goreng?
14:31:32 <kuhnovic> Nothing wrong with a good Nasi Goreng!
14:33:18 <locosage> hm, I've found save that crashes beta3 on assertion
14:35:03 <xarick> I think I managed to make lakes 100% yapf now
14:40:19 <xarick> there may be a problem...
14:40:27 <xarick> the binary heap thingie
14:40:42 <xarick> i might need to sort the vector?
14:41:08 <locosage> is there no link for older versions on openttd website?
14:41:19 <locosage> it's getting harder to google for them every time I need one
15:05:48 *** kuka_lie has joined #openttd
15:06:28 <peter1138> Setting widget state in OnPaint is bad form, tbh.
15:12:36 <peter1138> Well, it doesn't really matter when the state is forced in OnPaint, I guess.
15:13:28 <peter1138> But generally if it's forced in OnPaint it's because someone couldn't make it stick when it was set in the correct place. Because it's the wrong type :)
15:17:01 <rito12_51026> Window is not marked as dirty when the train's force_proceed changes
15:18:03 <rito12_51026> *contaning invalid data
15:18:53 <peter1138> The correct thing is to invalidate the window when the state changes, and in the window's OnInvalidateData set the state.
15:19:06 <rito12_51026> peter1138: can it be in OnDraw instead?
15:20:24 <peter1138> There isn't an OnDraw.
15:20:52 <peter1138> What you've done is fine tbh.
15:21:05 <peter1138> It's already wrong, so you haven't made it worse :)
15:21:54 <rito12_51026> it can't be because DrawWidget is const function
15:25:01 <peter1138> DrawWidget is just as wrong as OnPaint.
16:01:59 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
16:02:58 *** WormnestAndroid has joined #openttd
16:53:33 <audigex> emperorjake: I don't mind them - if you have a very straightforward branch line or something (no alternate routes, no way to bypass stations) then they can be a super-fast way to handle the orders for it
16:54:18 <audigex> I wouldn't mind if it was flipped so that the default order was "go to X" and the alternate order was "go to X and stop at any intermediate stations you pass through" but that's more confusing for the user
16:55:59 <peter1138> You mean by enabling the option "New orders are 'non'stop' by default"?
16:56:10 <peter1138> (Which defaults to On)
16:56:41 <vondpc> i wish tehre was a way to just mirror the orders for return journeys on non-loop lines
17:05:45 <mmtunligit> if youre on JGRPP there is
17:15:27 <xarick> will std::tie clean up the vector before acquiring the new contents?
17:16:46 <xarick> `auto [found, river_tiles] = YapfFlowRiver(begin, height_begin);`
17:16:46 <xarick> `std::tie(found, river_tiles) = YapfFlowRiver(begin, height_begin);`
17:20:29 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
17:35:51 *** WormnestAndroid has joined #openttd
17:37:31 <davidxn> xarick: That looks OK to me! Is the returned value [all tiles that were considered] if max_search_nodes was reached, and empty if no path is found for a different reason?
17:40:11 <xarick> yes, exactly. The different reason is "all nodes were expanded and found no place to flow down to without even reaching the max_search_nodes"
18:15:57 <rito12_51026> Eh, videos don't work in my opera
18:20:24 *** tokai has quit IRC (Ping timeout: 480 seconds)
18:21:17 *** ChanServ sets mode: +v tokai
18:31:16 *** tokai has quit IRC (Ping timeout: 480 seconds)
18:34:47 *** ChanServ sets mode: +v tokai
18:35:33 <peter1138> People still use Opera?
18:37:10 *** yiffgirl has quit IRC (Server closed connection)
18:37:10 *** yiffgirl has joined #openttd
18:37:23 <rito12_51026> I use Opera in order not to use Chrome
18:49:04 *** tokai has quit IRC (Ping timeout: 480 seconds)
18:49:48 *** ChanServ sets mode: +v tokai
18:55:17 *** tokai|noir has joined #openttd
18:55:17 *** ChanServ sets mode: +v tokai|noir
19:00:09 *** tokai has quit IRC (Ping timeout: 480 seconds)
19:03:19 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
19:09:07 <xarick> what's different from std::flat_set and openttd's FlatSet
19:10:33 <peter1138> One is part of the std, one isn't.
19:11:41 *** ChanServ sets mode: +v tokai
19:14:56 *** WormnestAndroid has quit IRC (Remote host closed the connection)
19:14:57 *** WormnestAndroid has joined #openttd
19:19:00 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
19:23:42 <xarick> would std:flat_set be faster than openttd's FlatSet
19:23:56 *** WormnestAndroid has joined #openttd
19:27:33 <_glx_> one is usable, the other one is c++23
19:29:26 <_glx_> (as a remainder, openttd uses c++20)
19:32:19 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
19:32:52 *** WormnestAndroid has joined #openttd
19:43:40 <andythenorth> well railtypes UI will rumble on eh π
19:44:12 <andythenorth> "dropdown probably fine"
19:44:25 <andythenorth> no badges, no filtering, no hiding, but eh
19:58:22 <andythenorth> can't even do 'type ahead' in the dropdown
19:58:26 <andythenorth> have to just scroll
19:58:36 <andythenorth> but I guess users are familiar with where to scroll to
19:58:40 <andythenorth> and have muscle memory
19:59:42 <rito12_51026> I like the Brickblock's idea with the large and small version
20:06:26 <andythenorth> I am curious why stations aren't a dropdown
20:06:32 <andythenorth> original only has 1 type
20:17:32 <brickblock19280> is it possible to search badges?
20:18:15 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
20:20:52 <rito12_51026> yes there is an badge string text filter
20:38:59 *** tokai has quit IRC (Ping timeout: 480 seconds)
20:43:10 <andythenorth> brickblock19280: search badges where? π
20:43:15 *** ChanServ sets mode: +v tokai
20:55:29 <brickblock19280> In the search bar
20:56:13 <brickblock19280> seems like an easy way to quickly filter on multiple but I haven't used badges so idk if that's possible rn or not
21:00:31 <andythenorth> works for vehicles
21:19:04 <brickblock19280> cool and you can add any number of other badge filters above too?
21:20:42 <andythenorth> in a world of oversized grfs like Horse....it's outstandingly handy
21:21:46 <brickblock19280> very very nice
21:38:01 <xarick> the random cost is ruining my yapf lake implementation now
21:39:48 <xarick> stoopid lakes... so hard
21:41:04 *** keikoz has quit IRC (Ping timeout: 480 seconds)
22:01:36 <xarick> fmt::format is crashing me
22:01:37 *** Wormnest has joined #openttd
22:16:48 <rito12_51026> Doesn't the "Fix #14723" state "why" ?
22:30:57 <xarick> 10000 closed nodes. The red arrow points to the beginning of the search
22:31:28 <xarick> it wants to find a lake which it didn't find
22:31:51 <xarick> because the lake was placed
22:33:42 <xarick> well... there is a lake there
22:34:00 <xarick> maybe i should point out the lakes
22:40:12 <xarick> new visual studio version is out, with a win11sdk, whatever that is
22:49:14 <xarick> okay, I understand the problem now
22:50:13 <xarick> I understand what needs to be done to fix the problem, but got no time today
22:52:51 <xarick> need to pass a dest_tile to yapf and pathfind to it. Hopefully, this will work around the random node costs and avoid reaching the max_search_nodes limit even when it previously didn't in the last run prior.
22:56:02 *** kuka_lie has quit IRC (Quit: Lost terminal)
continue to next day β΅
