IRC logs for #openttd on OFTC at 2024-05-14
⏴ go to previous day
02:15:07 *** ChanServ sets mode: +v tokai
02:17:10 *** Wormnest has quit IRC (Quit: Leaving)
02:22:04 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
02:50:31 *** debdog has quit IRC (Ping timeout: 480 seconds)
04:40:26 <DorpsGek> - Update: Translations from eints (by translators)
04:58:08 *** Guest3122 has quit IRC (Write error: connection closed)
04:58:09 *** jact[m] has quit IRC (Write error: connection closed)
04:58:09 *** grag[m] has quit IRC (Write error: connection closed)
04:58:09 *** patricia[m]1 has quit IRC (Write error: connection closed)
04:58:09 *** yubvin[m] has quit IRC (Write error: connection closed)
04:58:12 *** menelaos[m] has quit IRC (Remote host closed the connection)
04:58:12 *** paulus[m] has quit IRC (Write error: connection closed)
04:58:12 *** imlostlmao[m] has quit IRC (Write error: connection closed)
04:58:12 *** igor[m] has quit IRC (Write error: connection closed)
04:58:12 *** jeeg[m] has quit IRC (Write error: connection closed)
04:58:12 *** magdalena[m] has quit IRC (Write error: connection closed)
04:58:12 *** joey[m] has quit IRC (Write error: connection closed)
04:58:12 *** pikaHeiki has quit IRC (Write error: connection closed)
04:58:12 *** tuxayo has quit IRC (Write error: connection closed)
04:58:12 *** soylent_cow[m] has quit IRC (Write error: connection closed)
04:58:12 *** andythenorth[m] has quit IRC (Write error: connection closed)
04:58:12 *** karl[m]123 has quit IRC (Write error: connection closed)
04:58:12 *** zzy2357[m] has quit IRC (Write error: connection closed)
04:58:12 *** tonyfinn has quit IRC (Write error: connection closed)
04:58:12 *** Rdolfs[m] has quit IRC (Write error: connection closed)
04:58:12 *** karoline[m] has quit IRC (Write error: connection closed)
04:58:12 *** jeremy[m]123 has quit IRC (Write error: connection closed)
04:58:12 *** FelixActually[m] has quit IRC (Write error: connection closed)
04:58:12 *** CornsMcGowan[m] has quit IRC (Write error: connection closed)
04:58:12 *** xyii[m] has quit IRC (Write error: connection closed)
04:58:12 *** hamstonkid[m] has quit IRC (Read error: Connection reset by peer)
04:58:12 *** playback2396[m] has quit IRC (Write error: connection closed)
04:58:12 *** luffy[m] has quit IRC (Write error: connection closed)
04:58:13 *** JamesRoss[m] has quit IRC (Write error: connection closed)
04:58:13 *** nolep[m] has quit IRC (Write error: connection closed)
04:58:13 *** royills[m] has quit IRC (Write error: connection closed)
04:58:13 *** elliot[m] has quit IRC (Write error: connection closed)
04:58:13 *** Gadg8eer[m] has quit IRC (Remote host closed the connection)
04:58:13 *** Heiki[m] has quit IRC (Remote host closed the connection)
04:58:13 *** shedidthedog[m] has quit IRC (Remote host closed the connection)
04:58:13 *** philip[m] has quit IRC (Write error: connection closed)
04:58:13 *** emilyd[m] has quit IRC (Write error: connection closed)
04:58:13 *** phil[m]1 has quit IRC (Write error: connection closed)
04:58:13 *** Elysianthekitsunesheher[m] has quit IRC (Write error: connection closed)
04:58:13 *** thelonelyellipsis[m] has quit IRC (Write error: connection closed)
04:58:13 *** Farrokh[m] has quit IRC (Write error: connection closed)
04:58:13 *** giords[m] has quit IRC (Write error: connection closed)
04:58:13 *** cjmonagle[m] has quit IRC (Write error: connection closed)
04:58:13 *** citronbleuv[m] has quit IRC (Write error: connection closed)
04:58:13 *** blikjeham[m] has quit IRC (Write error: connection closed)
04:58:13 *** temeo[m] has quit IRC (Write error: connection closed)
04:58:13 *** osvaldo[m] has quit IRC (Write error: connection closed)
04:58:13 *** luk3Z[m] has quit IRC (Write error: connection closed)
04:58:13 *** leward[m] has quit IRC (Write error: connection closed)
04:58:13 *** gdown has quit IRC (Write error: connection closed)
04:58:13 *** einar[m] has quit IRC (Write error: connection closed)
04:58:13 *** Bilb[m] has quit IRC (Write error: connection closed)
04:58:13 *** audunm[m] has quit IRC (Write error: connection closed)
04:58:13 *** thomas[m]1234567 has quit IRC (Write error: connection closed)
05:15:49 *** keikoz has quit IRC (Ping timeout: 480 seconds)
05:27:15 <merni> you can't make something longer than 7 by attaching more wagons, but if one unit is already 7 it still allows it
06:00:07 *** SigHunter has quit IRC (Ping timeout: 480 seconds)
07:14:13 *** SigHunter has joined #openttd
07:16:44 <peter1138> merni: Yes, but my attempts to fix it lead to more issues so far 🙂
07:17:54 <peter1138> To know that it is too long, it needs to be have been built. so it then needs to be sold again.
07:18:40 <peter1138> I did that in CmdBUildVehicle, and it worked great.
07:18:58 <peter1138> As long as you test + exec.
07:19:09 <peter1138> But then autoreplace comes along with only execs
07:37:33 *** SigHunter has quit IRC (Ping timeout: 480 seconds)
07:45:16 *** SigHunter has joined #openttd
07:48:34 *** XYZ_ has quit IRC (Ping timeout: 480 seconds)
07:53:26 *** SigHunter has quit IRC (Ping timeout: 480 seconds)
07:55:00 *** SigHunter has joined #openttd
07:57:19 <merni> why not make autoreplace also test?
07:59:46 <peter1138> I just didn't get that far.
08:20:28 *** SigHunter has quit IRC (Ping timeout: 480 seconds)
08:23:15 *** SigHunter has joined #openttd
08:39:53 *** amal[m] has joined #openttd
08:55:58 *** SigHunter has quit IRC (Ping timeout: 480 seconds)
08:56:06 *** tokai has quit IRC (Quit: c('~' )o)
09:30:44 *** SigHunter has joined #openttd
09:36:52 *** SigHunter_ has joined #openttd
09:39:58 *** SigHunter has quit IRC (Ping timeout: 480 seconds)
09:57:41 *** SigHunter_ has quit IRC (Ping timeout: 480 seconds)
09:58:12 *** SigHunter has joined #openttd
10:06:24 <kuhnovic> Ugh, std::prority_queue looked nice for use in a pathfinder. Too bad you can only remove the "best" element...
10:32:43 <Eddi|zuHause> that's kind of the point of a priority queue?
10:33:46 <FLHerne> Isn't that usually what you want in a pathfinder? Find the best node in the open list, evaluate it and remove it
10:41:53 <Rubidium> FLHerne: yes, but... when two paths merge you want to remove the worse of the two options. Otherwise you would have the same node twice in the priority queue
10:42:51 <Rubidium> the "removing" from the priority queue kuhnovic is talking about is likely "updating an element to give it a different priority" and std::priority_queue does not support that
10:48:16 <ahyangyi> You can do "lazy removal" though, when you want to update an element you just insert the new version of it; and when you pop an element you check whether it's a stale version.
10:49:56 <ahyangyi> Kind of annoying to work with, but might still be simpler than alternatives like a map to sort things and another map to find iterators in the first map quickly
10:54:39 <_jgr_> At that point you'd be much better off just writing a better adapter instead if std::priority_queue
10:54:54 <locosage> heap as a data structure doesn't support updates
10:55:10 <locosage> marking works but may ruin the speed
10:56:32 <Rubidium> in any case, I'd guess he either looked at the backlog where I mentioned looking into priority queue for the pathfinder last sunday, or maybe he looked at my open PR to remove one incarnation of priority queue from OpenTTD's code (and leave the other alone)
10:59:49 <locosage> hm, though I guess it's not impossible to update heap, it's just not algorithmically faster than removing and re-inserting the element
11:00:22 <ahyangyi> and the problem with C++ priority queues is that there is no API for removing either
11:00:29 <ahyangyi> besides "removing the top"
11:01:11 <ahyangyi> in principle, support for removing random elements at O(log n) speed can be done with a reverse lookup. But that's an intrusive change, not implementable by composition
11:02:57 <locosage> yeah, stl structures tend to have these issues
11:03:31 <locosage> iirc there was some non-standard api in gcc to work with rb-trees
11:03:37 <locosage> maybe there is something for heap as well
11:06:37 <locosage> I was to lazy to learn it and just wrote a treap whenever I needed stuff that stl trees didn't have xD
11:07:12 <kuhnovic> Rubidium: Indeed, sort of. An A* pathfinder sometimes needs to update the cost of an existing node if a better path to it is found. This means removing it from the priority queue and reinserting it again, but std::priority_queue doesn't support that. Too bad.
11:09:11 <kuhnovic> I'm just generally playing with the idea of refactoring YAPF a bit. It's not a very flexible implementation, and the heavy templating makes it hard to work with / understand.
11:10:08 <kuhnovic> And if I can generalize it a bit then we should also be able to use it for river generation, and AyStar would be obsolete.
11:10:38 <peter1138> Yeah, it obfuscates things a bit, trying to work out what it's really doing and which bits are just container reimplementations is ... fun.
11:10:59 <locosage> there is a workaround with push/pop_heap functions to implement deletion if you really want to use stl
11:11:26 <locosage> but at that point it's getting more complex than implementing the heap itself xD
11:12:13 <kuhnovic> I would like to use stl, but not if it means pain and workarounds
11:12:34 <Rubidium> kuhnovic: that would be nice, but sounds like a lot of complicated work. I just simplified Aystar by using YAPF's nodelist
11:12:34 <peter1138> I think it makes sense to have one good implementation, instead of two slightly different implementations 🙂
11:13:03 <Rubidium> but going fully with YAPF was a bit too rabbit holey for me ;)
11:18:02 <locosage> there is probably some library that can be used with a fast hardware-optimized heap
11:18:17 <locosage> like b-heap or min-max heap or smth
11:33:01 <kuhnovic> I got a chance to cut my teeth on YAPF while working on the Ship PF, I might be able to do it 😉
11:34:57 <kuhnovic> locosage: I doubt that it will make a big difference. I recently did some profiling and it turns out that a lot of time is spent calculating node costs. The mechanisms of YAPF itself seemed quite fast already.
11:37:06 <kuhnovic> It differs a bit between pathfinder implementations though. Train-YAPF has by far the most complicated cost calculation, although it is somewhat mitigated by caching. RV-YAPF also does a fair bit of stuff to calculate costs, and no caching. Ship-YAPF has quite a simple cost function, and the water region's cost function is even simpler, so there the performance of the YAPF mechanics are more
11:37:16 <locosage> but if heap performance isn't a bottleneck anyway then sure
11:49:57 <andythenorth> Multi-hold ships for OpenTTD 17?
11:50:28 <peter1138> Stuck Record for OpenTTD 18?
11:51:33 <peter1138> I'm aware there's a balance between knowing what users want and doing our own thing.
11:58:13 <FLHerne> tugs+barges, plus hacks for multiple holds
11:58:49 <ahyangyi> parameters organized hierarchically, like vehicle variants
12:00:42 <kuhnovic> Hyperloops that break down all the time
12:01:16 <ahyangyi> can hyperloops overtake each other?
12:01:40 <ahyangyi> hmm, "overtaking from above, not sideways"
12:03:56 <andythenorth> peter1138: Flat Records
12:04:09 <peter1138> Hopefully. They sound weird if not.
12:04:15 <andythenorth> With escape depots, so they don’t get stuck
12:06:35 <kuhnovic> ahyangyi: There will be a cheat for that. Sorry, I mean a sandbox option 😄
12:23:03 *** gwyd4016 has joined #openttd
12:23:03 <gwyd4016> We just got rid of this for RVs
13:17:46 *** toktik has quit IRC (Quit: Ping timeout (120 seconds))
14:00:18 *** toktik has quit IRC (Quit: Ping timeout (120 seconds))
14:46:13 *** HerzogDeXtEr has joined #openttd
14:57:43 *** tony_pixel has quit IRC (Quit: User went offline on Discord a while ago)
15:53:18 *** Wormnest has joined #openttd
15:54:31 <peter1138> Urgh, when you merge something then immediately spot an issue with it...
15:54:42 <peter1138> Well, I suppose that's better than not spotting the issue.
16:14:14 *** gelignite has joined #openttd
16:23:53 *** gelignite has quit IRC (Quit: Stay safe!)
16:52:09 *** SigHunter has quit IRC (Ping timeout: 480 seconds)
17:01:28 *** masteroktagon has joined #openttd
17:01:28 <masteroktagon> Hi, I want to program a Program that extracts train data from a savegame file. Is there a documentation about the format (I suspect so) and where can I find it?
17:04:36 <_glx_> was about to tell the same
17:15:21 *** SigHunter has joined #openttd
17:43:18 <masteroktagon> getting the map/train data in real-time is probably not easily possible, is it?
17:52:48 <locosage> masteroktagon: only by patching game itself
17:53:51 <masteroktagon> for this I would need more understanding of the code... but then I can also do a patch that could be PRed for OpenTTD
17:54:57 <LordAro> anything like that would want to be GS/admin port based
18:06:40 <talltyler> #12678 could use a re-review too 🙂
19:56:58 <silent_tempest> Something has increased the game options menu width
19:57:19 <peter1138> The "Get Content" buttom
19:58:14 <silent_tempest> In the Base Graphics set frame?
19:58:27 <peter1138> Graphics, Sounds and Music.
19:59:02 <_glx_> but effect is bigger in graphics since there was also the settings button
19:59:03 <peter1138> Whenever the width changes I have to update my muscle memory.
19:59:42 <peter1138> But the width varies with language anyway, so it's not a sacred width or anything 🙂
19:59:58 <silent_tempest> I've spent the last two days trying to figure what I did make the width sudddenly increase until I eventually figured out I hadn't...
20:00:34 <peter1138> I might actually have to implement something else to make that work better.
20:01:07 <silent_tempest> Well that's one less thing I have to debug. Lol
20:04:40 <peter1138> Only a problem if your widgets up end not filling the window somehow...
20:24:27 <talltyler> What needs to be determined in order to get the house picker available in normal gameplay mode? Obviously there will be a setting to enable/disable the feature, but apart from that…the cost of a house? Specs have a removal cost multiplier that could be used for the construction price 🙂
20:29:13 <peter1138> I was thinking purely sandbox mode
20:29:20 <peter1138> In which case cost does not matter.
20:29:54 <peter1138> Consider the consider of merely funding construction of new buildings in the town authority window.
20:30:13 <peter1138> Which has no gaurantee of actually increasing the popultion.
20:30:35 <peter1138> Placing houses will always increase it.
20:30:52 <peter1138> So it would need to be out-of-proportionally expensive.
20:31:11 <peter1138> So... I think it's best left as sandbox, so that that does not matter.
20:31:49 <andythenorth> is sandbox a mode or a setting? 😛
20:31:52 <peter1138> And for that, we need to decide how the Sandbox Settings window should behave.
20:32:01 * andythenorth considering other scenario editor tools
20:32:19 <andythenorth> oh yes, you took the words
20:32:35 <peter1138> So when do we give rivers flow?
20:32:53 <peter1138> So that if you disrupt them they flow into new paths and destroy the thing you were trying to serve...
20:32:58 <andythenorth> peter1138: openttd 21
20:33:16 <peter1138> Okay, my font search interface has FontConfig and Win32 implementations.
20:33:28 <talltyler> I wouldn’t make Sandbox Settings a third category of things to change about the game, just merge Cheats into regular Settings
20:33:36 <peter1138> "interface" as in code, not UI, that is.
20:33:40 <talltyler> Settings and Options seems like enough 🙂
20:33:54 <peter1138> Yeah, but we haven't done that yet 🙂
20:34:28 <peter1138> But seems odd that we moved it from Cheats to Sandbox, and then we either remove it, or leave it there, but put other sandbox settings somewhere else...
20:34:39 <talltyler> Right, but I mean that’s the route I would take instead of making a new Sandbox Settings category for all the things we deem to be “not serious gameplay”
20:35:43 <talltyler> I see the Cheats menu(now Sandbox Settings, I guess) disappearing entirely by 15.0
20:36:04 <talltyler> Once someone figures out what to do with the weird ones like date cheat
20:36:20 <peter1138> Which route, sorry?
20:36:39 <peter1138> Oh, you mean "Sandbox Settings" "Cheat" entry.
20:36:43 <talltyler> Delete Cheats, move its contents to Settings
20:36:48 <peter1138> Not "Sandbox Settings" settings entry.
20:36:50 <talltyler> Nothing says Sandbox anymore
20:37:07 <peter1138> I didn't want to put it in to Cheats 🙂
20:37:25 <peter1138> And currently there's no obvious place in settings for sandbox stuff, so I didn't do that either.
20:37:42 <peter1138> Infinite Money is under Difficulty, which doesn't quite fit.
20:37:53 <talltyler> I think the game is just customizable enough that most settings could be used for sandbox purposes
20:38:06 <LordAro> could hide date cheat behind one of the developer modes
20:38:21 <talltyler> No need to have its own category for Sandbox things
20:38:51 <talltyler> I put a lot of things into Environment subcategories recently
20:39:33 <peter1138> I notice JGRPP has a few more features in its House Picker.
20:39:44 <talltyler> I like the subcategories I made, although I’m not sure the Environment parent makes that much sense. I wonder if it could just be eliminated to promote its children…
20:39:48 <peter1138> So maybe they can be "backported" somehow.
20:39:59 <talltyler> Someday Someone will reorganize Settings 😛
20:40:10 <peter1138> I think it has the ability to choose town. I just went with nearest town.
20:40:34 <talltyler> Ctrl+Click to choose town, like with station joining?
20:42:39 <peter1138> Maybe. Needs a UI 🙂
20:43:06 <peter1138> I did not consider it important enough for now.
20:43:16 <talltyler> That would get old fast, if you had to choose for each house
20:43:18 <peter1138> Plus it was already big enough, despite the unified picker code.
20:43:41 <peter1138> Oh sure, I mean needs a UI with ctrl 🙂
20:43:41 <talltyler> Oh agreed, the scope of the PR was admirably focused 🙂
20:44:19 <peter1138> Well, actually the fake-classes was a bit of a feature creep which isn't entirely necessary, but quite nice.
20:44:53 <talltyler> I am sure players/GRF authors will want custom classes someday, so it’s nice to have UI for that already
20:44:59 <peter1138> Of course, having to implement all the GUI-side drawing on town houses didn't help. For everything else that always existed.
20:45:03 <talltyler> (And fallback for sets which don’t have custom classes)
20:45:15 <peter1138> Well, that UI bit is already handled by the picker code.
20:45:38 <peter1138> Houses DO have classes, but they're NewGRF internal, don't have names, and don't work anything like class labels.
20:45:47 <talltyler> Right, just numbers
20:46:05 <peter1138> Of course you know all this having made house sets.
20:46:29 <talltyler> I’ve used them and they don’t really correspond to categories that players would want to choose. I used them for placement rules to cluster some buildings and anti-cluster others
20:46:36 <peter1138> I do also have a branch that implements a GRF File filter.
20:46:51 <peter1138> Which would bring the ability to pick by NewGRF to ALL the picker windows.
20:47:50 <peter1138> I have not finished the GRF File picker yet though. Of course it also makes the UI larger and more complex.
20:48:46 <peter1138> And also the separate windows feature which was my first inspiration for unifying it.
20:48:52 <talltyler> That could be nice, although I’m not sure how much I’d use it versus the text filter. If I want to see only Auz objects, that’s in the class name, whereas the object might be in one of hundreds of files that Garry has released
20:48:59 <peter1138> That was actually non-unified, hence the rewrite.
20:49:43 <peter1138> People have asked for a grf filter though, so who knows.
20:52:03 <talltyler> Yes, worth asking around Discord channel #add-on-development perhaps, there are some eye-candy builders who hang out in there 🙂
20:56:58 <peter1138> Have you ever asked in there? lol
21:07:24 *** keikoz has quit IRC (Ping timeout: 480 seconds)
21:17:00 <talltyler> You’ll get plenty of opinions, I’m not saying they’re useful 😛
21:30:20 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
21:36:59 <peter1138> Right, let's see if this mac OS code compiles.
21:38:28 <peter1138> Living dangerously :p
21:39:30 <peter1138> My master branch was not up to date, it looked like a huge PR 😄
21:40:14 <peter1138> Writing mac OS code with no way to test. Yum.
21:51:47 <_glx_> you should be able to find testers here
22:08:33 <peter1138> Heh, CI page is crashing firefox :S
22:13:23 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
22:23:00 <_glx_> too many errors in the log ?
22:26:14 <peter1138> Well, at least for that I can self-compile.
22:38:03 <peter1138> Of course, it compiles for me.
22:38:38 <peter1138> So I guess a random include before stdafx 😄
22:42:30 <peter1138> ```Unused local variable
22:42:30 <peter1138> Variable _ is not used.```
22:45:47 <_glx_> usually random std includes are detected when disabling precompiled headers
22:46:15 <peter1138> No, random normal includes that clangd likes to add.
23:36:14 *** Flygon has quit IRC (Read error: Connection reset by peer)
continue to next day ⏵
