IRC logs for #openttd on OFTC at 2025-12-12
⏴ go to previous day
01:01:04 *** ChanServ sets mode: +v tokai
01:08:05 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
02:49:17 *** Wormnest has quit IRC (Quit: Leaving)
04:18:04 *** Zathras_7 has joined #openttd
04:18:24 *** Zathras_1 has joined #openttd
04:21:36 *** Zathras_11 has quit IRC (Ping timeout: 480 seconds)
04:21:46 *** Zathras_4 has quit IRC (Ping timeout: 480 seconds)
04:44:36 <DorpsGek> - Update: Translations from eints (by translators)
06:22:12 *** MinchinWeb[m] has quit IRC (Ping timeout: 480 seconds)
06:25:09 *** MinchinWeb[m] has joined #openttd
07:44:57 <locosage> why the heck is discord keeps breaking in this channel
07:47:22 *** reldred has joined #openttd
07:47:22 <reldred> discord is a shitshow everywhere, not just this channel.
07:52:07 *** gelignite has joined #openttd
08:06:58 <ahyangyi> This channel contains useful stuff, so not as easily avoidable as everywhere else.
09:10:07 <locosage> there is a lot more room for the year if you don't try to jam it in a single column
09:16:23 <locosage> have enough open prs already :p
09:26:36 <LordAro> there you go, there's an open slot now
09:33:32 <locosage> that's a dev slot :p
09:34:23 <locosage> doesn't work for peasants
10:10:00 <xarick> "To insert the data or key in B-tree is more complicated than a binary tree."
10:10:11 <xarick> did some robot write this article?
10:11:32 <peter1138> And that's before slop was widespread. Guaranteed slop now.
10:49:33 <ahyangyi> I think I ranted each time I clicked on it because it ranked so high on Google
10:49:54 <ahyangyi> But the new "Google AI Overview" has taken its place now
10:50:58 <ahyangyi> both the place on Google, and the metaphorical place to place my rants and blames
11:18:07 <peter1138> Hm, does Visual Studio allow for interactive rebases?
11:18:20 <peter1138> Or am I going to have to give my junior dev a lesson on git cli?
11:26:59 <peter1138> Hmm, maybe cherry-pick instead.
11:38:15 <xarick> ctor means constructor?
12:12:51 <locosage> could've been merged 3 years ago 🤷♂️
12:17:06 <locosage> idk, ask whoever put that label there
12:18:01 <locosage> 2TT would probably not agree xD
12:18:26 <ahyangyi> oh, I meant the `work: needs rebase` one 😛
12:18:27 <jfkuayue> Yesterday-Today: 0430-1130
12:18:51 <ahyangyi> I imagine the 2TT label was obsolete by the time it was reopened
12:19:26 <ahyangyi> Anyways, the label stays there for some reason :S
12:20:58 <locosage> I never saw it being removed so it's just a pointless label to me
12:21:26 <locosage> rebase I can do if anyone is interested in merging it
12:21:55 <ahyangyi> also rebasing will send a message here
12:22:10 <ahyangyi> I often feel not so sure about that
12:22:27 <ahyangyi> Technically I should always keep my PRs up to date
12:22:42 <ahyangyi> But I'm not sure how people read into it
12:22:47 <locosage> yeah, like, I could spam rebase my 7 or smth prs every day but what's the point
12:24:07 <locosage> well, except some ui around it got rewritten but that's solvable
12:24:24 <locosage> instead it ended up in cmclient 😜
12:34:32 *** Zathras_1 has quit IRC (Quit: Connection reset by beer!)
15:15:15 *** moll has quit IRC (Ping timeout: 480 seconds)
15:40:03 *** MinchinWeb[m] has quit IRC (Read error: Connection reset by peer)
15:40:50 *** MinchinWeb[m] has joined #openttd
15:43:55 *** Zathras_4 has joined #openttd
15:47:19 <locosage> wth, did RC1 get released with a known desync (#14677)?
15:49:21 <locosage> kinda makes testing multiplayer pointless
16:25:27 *** MinchinWeb[m] has quit IRC (Ping timeout: 480 seconds)
16:26:42 *** MinchinWeb[m] has joined #openttd
16:30:48 *** MinchinWeb[m] has quit IRC (Read error: Connection reset by peer)
16:30:57 *** MinchinWeb[m] has joined #openttd
16:47:37 *** MinchinWeb[m] has quit IRC (Ping timeout: 480 seconds)
16:47:42 *** MinchinWeb[m] has joined #openttd
16:55:24 <LordAro> locosage: you could try to help instead of whining
16:55:31 <LordAro> depending on your preference
16:56:24 <LordAro> no one else has reported anything, so it seems pretty rare
16:57:10 <locosage> wasn't me who rushed RC1 :p
16:57:24 <locosage> don't have time to be helping with desyncs atm, sorry
16:58:17 <locosage> I did hope to run it for a while on citymania since it has some desync debugging tools
16:58:32 <locosage> but doubt it's usefulness now that there is a known one anyway
16:58:37 *** michi_cc[d] has joined #openttd
16:58:37 <michi_cc[d]> I wouldn't call 3 months rushed though, unless you advocate for an indefinite delay to releases.
16:59:15 <locosage> I called abandoning milestone targets and going for rc instead of beta rushed :p
16:59:17 <_jgr_> Having a release makes it more likely that it'll get tested in multiplayer, if the issue is still there it'll likely get eventually uncovered by users
16:59:37 <_jgr_> Nobody is using nightlies for multiplayer testing
17:00:02 <locosage> problem is, desyncs are very unspecific, all that users can uncover is that there is one
17:00:09 <locosage> and that's already known
17:01:22 <locosage> as I wouldn't bet on it accidentally getting fixed since beta3
17:02:46 <locosage> also, users tend to not report desyncs as all considering them "normal"
17:02:50 <LordAro> i'm strugging to see your reasoning about why running rc1 on citymania wouldn't be useful regardless
17:03:01 <LordAro> worst case is that you find the same problem
17:03:15 <LordAro> which gives valuable debugging information
17:03:31 <locosage> I won't know if it's the same unless I spend hours investigating it
17:03:34 <LordAro> best case is nothing ever happens because it's super rare
17:03:43 <LordAro> or that known newgrf issue
17:04:02 <LordAro> spending hours investigating it is our problem, not yours
17:04:08 <LordAro> not necessarily yours*
17:04:29 <locosage> without investigating all I can tell you is whether there is desync or not
17:04:50 <_glx_> we only received one report
17:04:52 <LordAro> "since it has some desync debugging tools"
17:05:11 <LordAro> does this not mean you have slightly more details than that?
17:05:14 <locosage> those debugging tools allow me to reproduce most desyncs
17:05:19 <locosage> but using them takes a lot of time
17:05:30 <LordAro> (also, PRs, links, branches, whatever, please)
17:19:04 *** toktik is now known as Guest33487
17:47:01 <locosage> links for what, debug tools? it's essentially the same stuff that openttd already has except it can be run in production
17:47:11 <locosage> doubt you need a pr that logs client connection frames to postgres
17:48:02 <locosage> make a save at start, log connection and desync frames, record and replay commands
17:48:19 <locosage> with enough effort that's enough to restore client and server states at the moment of desync
17:51:47 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
18:01:53 *** WormnestAndroid has joined #openttd
18:09:58 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
18:10:38 *** WormnestAndroid has joined #openttd
18:20:07 <xarick> something wrong with openttd gs config
18:20:47 <xarick> i had it set up to run "Performance Tests" GS
18:21:08 <xarick> upon restarting openttd, it's no longer there, it's set to none
18:22:12 <ahyangyi> I kinda wish that AI get better at "vibe testing" especially for things like desyncs
18:27:20 <_glx_> xarick: it works fine for me
18:28:03 <locosage> iirc it's not saved if game crashes
18:29:04 <xarick> seems to happen after I instaled RC1
18:34:23 <_glx_> nothing changed in code for that for 8 months
18:34:48 <xarick> probably windows and folder protections ...
18:38:41 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
18:39:22 *** WormnestAndroid has joined #openttd
18:52:20 <jfkuayue> *christmas break starts*
18:58:18 *** Wormnest has joined #openttd
19:05:33 <peter1138> Okay, so I need a development environment on Windows.
19:05:54 <peter1138> Is VS 2022 okay to use?
19:06:24 <peter1138> Hmm, they're pushing 2026.
19:25:47 *** Wormnest has quit IRC (Ping timeout: 480 seconds)
19:25:49 <audigex> peter1138: Minimal reason to use anything other than the most recent, unless you have a license for an older version
19:32:35 *** MinchinWeb[m] has quit IRC (Read error: Connection reset by peer)
19:32:52 *** MinchinWeb[m] has joined #openttd
19:38:12 <peter1138> Ah, good old VS. After cloning and opening OpenTTD... I have to close it all and open it again, so that it seems it is CMake.
19:39:43 <peter1138> Now, was I imagining that VS has its own copy of vcpkg already?
19:41:13 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
19:42:41 *** WormnestAndroid has joined #openttd
19:44:57 <_zephyris> I do enjoy the wetland
19:44:57 *** WormnestAndroid has quit IRC (Remote host closed the connection)
19:45:29 *** WormnestAndroid has joined #openttd
19:45:40 <xarick> I wonder how the current AIs deal with that
19:45:47 <peter1138> Nice. vcpkg, installed with VS, running inside VS, fails to "install" because it can't see VS.
19:46:22 <xarick> my AI treats water as untouchable
19:47:08 <xarick> so... I expect a poorer performance
19:49:35 <jfkuayue> my condolences to peter
19:54:14 <andythenorth> _zephyris: we need alligators
20:00:30 *** Flygon has quit IRC (Remote host closed the connection)
20:00:52 *** MinchinWeb[m] has quit IRC (Ping timeout: 480 seconds)
20:01:43 *** MinchinWeb[m] has joined #openttd
20:08:13 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
20:08:49 *** WormnestAndroid has joined #openttd
20:14:38 <_glx_> peter1138: I think you still need to clone vcpkg repo somewhere, even in VS uses its internal vcpkg (and also don't forget to set CMAKE_TOOLCHAIN_FILE and VCPKG_TARGET_TRIPLET)
20:17:15 <xarick> static cast means bad design, right?
20:19:43 <xarick> i just went through template hell with mr copilot and we ended up needing to use static_cast almost every time we need to go into Leaf vs Internal
20:21:33 <andythenorth> can mr copilot produce Iron Horse grf yet?
20:22:59 <xarick> let me show you my code
20:24:26 <_jgr_> I hesitate to throw another spanner in the works, but what I was hinting at when I pointed out that the values and children are mutually exclsuive, was that you could make a variant or equivalent of those two fields
20:24:39 <_jgr_> I didn't mean make two nodes where almost all the fields are the same
20:25:37 <_jgr_> Or you could make the common fields a base class, and inherit for the more specific node types, etc, etc
20:26:20 <tolis> quick question, does anyone know how to change the time periods? I get 3 financial reports on 3 periods, I just want one annual report
20:27:26 <_jgr_> xarick: OK, you're already using a base class, that should significantly reduce the number of casts you need
20:28:31 <jfkuayue> tolis: I don't think an "annual report" is possible in wallclock timekeeping mode
20:29:24 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
20:30:09 *** WormnestAndroid has joined #openttd
20:36:27 <xarick> let me see what happens if I remove static_cast
20:37:37 *** tolis has quit IRC (Quit: Page closed)
20:38:14 <xarick> here's an example function with static_cast
20:42:49 *** Zathras_4 has quit IRC (Quit: Connection reset by beer!)
20:44:27 <_glx_> cast is usually not needed if accessed members are in base
20:45:24 <xarick> I see that. Looks like some clean up is in order
20:45:32 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
20:47:30 *** WormnestAndroid has joined #openttd
21:10:23 <xarick> erm apparently I'm too constrained. must keep the static casts most of the time
21:10:31 <xarick> doesn't let me compile
21:18:45 <xarick> There's Node, Leaf and Internal. if node->role == leaf, then it has a Leaf struct, otherwise it has a Internal struct. but accessing them requires casting away from Node to Leaf or Internal...
21:18:45 <xarick> isn't there a more elegant way to do this?
21:21:07 <xarick> children is still a bit messy
21:24:38 <xarick> std::array<std::unique_ptr<BPlusNodeMap<Tkey, Tvalue, B>>, B + 1> children;
21:24:38 <xarick> hmm yeah, they're still Node, but actually they're Leaf or Internal
21:25:02 <xarick> but Node is the base so...
21:25:52 <xarick> not sure what I can do
21:26:33 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
21:26:56 *** WormnestAndroid has joined #openttd
21:38:18 <_glx_> you're kinda force to cast with inheritance and passing parent type
21:41:53 <peter1138> _glx_, it looks like it clones vcpkg itself
21:43:16 <_glx_> and it auto sets CMAKE_TOOLCHAIN_FILE ?
21:43:19 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
21:43:26 *** WormnestAndroid has joined #openttd
21:49:28 <_glx_> hmm if I remove CMAKE_TOOLCHAIN_FILE, clear cache and reconfigure it doesn't install packages
21:49:36 *** Wormnest has joined #openttd
21:56:19 <_glx_> oh ok I still need to set CMAKE_TOOLCHAIN_FILE in presets, but it should point to" $env{VCPKG_ROOT}/scripts/buildsystems/vcpkg.cmake"
22:02:51 <_glx_> but I forced the toolchain
22:04:48 <_glx_> else it doesn't detect it
22:42:15 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
23:02:21 <_glx_> ah when using presets there's no autodetect
23:06:18 <Rubidium> locosage: yes, I 'rushed' RC1 as I wanted to break the cycle of not doing anything appreciable towards a release for three months and then basically promoting some nightly to beta. I also hope it shifts some more of the dev capacity to fixing bugs instead of introducing them.
23:07:29 <peter1138> My dev capacity is trying to get VS 2026 going :o
23:11:33 <_glx_> is cmake able to configure if you disable vcpkg.json (renaming is enough)
23:12:52 <locosage> I'm not against releasing quicker, I just don't like how unpredictable it was.
23:13:16 <locosage> Also, desyncs should be a blocker for any release imo, at least known ones
23:14:08 <_glx_> I hope we get more reports with RC1
23:15:34 <peter1138> _glx_, yes, but of course then I don't have liblzma, etc.
23:15:45 <peter1138> Let's see if it at least builds.
23:21:00 <_glx_> oh based on some searching you might need to install `C++ CMake tools for Windows` (via visual studio installer)
23:28:47 <peter1138> Of course, now it build it clang and that won't run due to the manifest thing, or something.
23:33:48 <xarick> i have a fix for the manifest
23:38:21 <peter1138> Anyway, hopefully there'll be a enough built that I can test these fixes.
23:42:24 <peter1138> Okay. It loaded. No intro :D
23:43:38 <_glx_> and I have the same component installed
continue to next day ⏵
