IRC logs for #openttd on OFTC at 2024-03-02
⏴ go to previous day
00:17:39 *** Wormnest has joined #openttd
00:25:52 *** wensimehrp has joined #openttd
00:25:52 <wensimehrp> _zephyris: oof I can imagine how *some* languages are going to be like
00:43:45 *** Leopold has quit IRC (Remote host closed the connection)
00:49:17 *** Leopold_ has joined #openttd
00:58:53 *** michi_cc has joined #openttd
00:58:53 <michi_cc> truebrain: What's the planed one after it, so I can think of a good teaser? 🙂
01:22:21 *** tony_pixel has joined #openttd
01:22:21 <tony_pixel> xarick: certified python moment
01:28:46 *** ChanServ sets mode: +v tokai
01:35:46 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
02:01:15 *** Wormnest has quit IRC (Quit: Leaving)
02:11:40 *** Leopold_ has quit IRC (Remote host closed the connection)
02:17:09 *** Leopold_ has joined #openttd
03:50:39 *** debdog has quit IRC (Ping timeout: 480 seconds)
03:56:26 *** gnu_jj_ has quit IRC (Ping timeout: 480 seconds)
05:51:37 *** Flygon has quit IRC (Read error: Connection reset by peer)
06:14:20 *** Leopold_ has quit IRC (Remote host closed the connection)
06:20:47 *** Leopold_ has joined #openttd
08:33:05 <_zephyris> wensimehrp: Yeah, C is ~1200, J is ~1000 and K is ~700
08:45:53 *** SigHunter has joined #openttd
09:12:29 *** SigHunter has joined #openttd
09:23:13 <truebrain> now all I need is an approval 😉
09:26:18 <truebrain> fixing a typo, while creating a typo; nice
10:04:21 <truebrain> Discord done, Steam done.
10:07:49 <xarick> alright, time for TicToc
10:18:20 *** tycekcz has joined #openttd
10:18:20 <tycekcz> I have a question about the plugins. Are there any plans, to make game data accessible via plugins? Fer example information about cities, industries, vehicles etc? Also are there any limitations considering programming language of the plugin?
10:18:56 <truebrain> the plugin API is meant to communicate with social platforms, and nothing broader than that. So no, currently no plans for the things you mention.
10:20:53 <truebrain> (this is purely because it is totally impractical to use it for anything else, given we, as OpenTTD, have to sign the plugin before it can be used)
10:20:58 <jfs> the game data are some really complex structures and will very much end up as an "offer an inch and they demand a mile"
10:21:22 <truebrain> however, we do have an interface that allows you to interact with that kind of information, which we call GameScript
10:21:32 <truebrain> maybe that already has everything you want / need 🙂
10:22:26 <jfs> hmm I wonder if we could make a way for a GS to make arbitrary data buffers available to the social API, but as a one-way communication channel, so the sociali API can't send anything (not even the most basic command) to the social API
10:23:11 <jfs> the thing is that it's really annoying to get data out of a GS, the only way is basically the admin port, and the admin port is really difficult to program against
10:24:15 <jfs> see the player on the forum who's trying to make a GS to do server admin tasks, because admin port is too difficult to build a good UI on top of, and console does not offer scripting features
10:24:21 <truebrain> jfs: maybe address the issue of the admin port, rather than creating yet-another-way? 😄
10:24:40 <tycekcz> jfs: Yes, I tried to figure out, if there is a way to export data via GS and I kinda failed. So that is, why I asked about the accessibility of the game data. I understand the decision behind that though.
10:24:53 <truebrain> GS -> admin-port is the way to go
10:25:16 <jfs> I really think we need to offer some kind of official admin port client
10:25:40 <truebrain> what is stopping you? 😄
10:25:43 <jfs> because the current stance of "here's some docs, just write your own" seems to be taken as quite hostile
10:26:24 <tycekcz> Also the admin port is available in multiplayer games only right?
10:26:35 <jfs> yes admin port is only active during multiplayer
10:26:45 <jfs> but GS also only runs on the server
10:27:31 <tycekcz> Yeah, but you can run GS in singleplayer games, which means, that the local machine is acting like a server I guess?
10:27:55 <jfs> yes but in standard singleplayer mode, the admin port is not active
10:32:17 <jfs> I'll probably go write a flatfile.social plugin at some point soon, one that just dumps the current status into a text file you can then read out with whatever else you want
10:32:47 <jfs> that should at least satisfy some of the "extracting stats" hunger
10:34:18 <frosch123> that's no plugin, that's just a patch
10:34:31 <jfs> it can be implemented as a social plugin
10:36:18 <frosch123> any custom interfaces can just use the admin port
10:36:35 <truebrain> that it is MP-only, is fixable 🙂
10:36:36 <jfs> no because the admin port only works in multiplayer
10:36:43 <frosch123> someone wanted to drop singleplayer anyway
10:36:51 <truebrain> frosch123: buy me some time 😛
10:37:02 <frosch123> i said "someone" :p
10:37:13 <truebrain> I identified that someone 😛
10:38:58 *** Leopold___ has joined #openttd
10:42:25 <xarick> rip multi destination feature
10:42:30 *** Leopold_ has quit IRC (Ping timeout: 480 seconds)
10:43:48 <andythenorth> just cut Horse grf from 32 MB to 26 MB 😛
10:43:58 <andythenorth> more duplicate realsprites deleted
10:44:01 <truebrain> now do the same for aBase!
10:44:46 <andythenorth> change all sprites to point to one realsprite, 1 px dimensions, blue
10:45:08 <truebrain> tycekcz: anyway, as you are just asking about solutions, if you can explain a bit what you are trying to solve, maybe we have some other ideas for you 🙂
10:51:00 *** Leopold___ has quit IRC (Ping timeout: 480 seconds)
10:51:46 <tycekcz> truebrain: Well some time ago I had an idea for an app just for me, that could work with data from OTTD, but the app idea is kinda silly, so I won't go into too much details. 😄 I tried approaching it with a GS, but I soon figured out, that I probably don't have the tools, that I would need. My app idea required information about towns in the game, I figured that is easy with GS, but I also needed
10:51:46 <tycekcz> an information, if the town contains a building of a specific type, which I did not found a way, how to check that.
10:52:55 <michi_cc> truebrain: Made some reddit thingy
10:54:21 <truebrain> tycekcz: If it is just for you, one way of getting data out of GS into a tool outside OpenTTD (at least, I assume that is what you are going for), is by starting the game with `-dscript=4` and capturing the output. A bit more tricky on Windows than on Linux, but both possible
10:54:48 <truebrain> That way GSLog entries in your GS end up on the console, into your tool
10:55:16 *** Leopold_ has joined #openttd
10:57:47 <michi_cc> Hmm, I was wondering. Do we have any access logs to see how many people actually visit the full dev blog posts?
10:59:28 <truebrain> Logs, no. Statistics, yes.
11:01:15 <truebrain> Any specific blog you want to know stats about?
11:01:22 <lea001> tycekcz: IIRC you could also use ServerGS and commuicate with OpenTTD over the admin port.
11:01:22 <lea001> I think there is a more robust and up to date version in the forum thread than the last version that I have released.
11:01:51 <tycekcz> truebrain: I kinda looked around the -d parameter, maybe a little bit more approacheable way, would be running it with -d script=5, that way it should output the log into stdout and capture that?
11:03:06 <lea001> ServerGS basically opens up all GS API to be accessed remotely.
11:05:30 <michi_cc> truebrain: Ah yes, I do mean statistics of course. I don't care about individual accesses 🙂 No specific blog, I was just wondering how the reception has been in generall.
11:05:42 <lea001> But it could be that also the latest one on forums is out of date with current OpenTTD and needs some love.
11:05:47 <truebrain> michi_cc: Seems about 10k unique people read the ship PF blog
11:05:49 <tycekcz> lea001: I see, but that would require running the game in a multiplayer mode, which I kinda did not want to do. But thank you for the suggestion, I can look into it.
11:06:16 <truebrain> 10k click through rate is pretty decent, yes 🙂
11:07:05 <lea001> tycekcz: Just pointing out that option. Parsing logs may be easier to pull off due to less parts involved.
11:09:25 <xarick> jeebus, NPF is so slow
11:11:46 <andythenorth> forgive us for being surprised that you're surprised?
11:12:04 <andythenorth> I appreciate the enthusiasm but "shall we remove NPF" isn't a standing joke for no reason
11:18:26 <xarick> how to make `CalcClosestStationTile` faster and smarter at the same time?
11:19:42 <xarick> and by proxy, `GetStationDockingTiles`faster as well?
11:23:11 <xarick> in case of YAPF, just GetStationDockingTiles
11:23:26 <xarick> the other isn't called
11:30:54 <xarick> gonna try removing reference, see if it changes anything
11:41:59 <xarick> well, it's faster... about ~4%
11:47:17 <xarick> I need a DistanceManhattan cache! lel
11:48:46 <xarick> how slow would this be? return std::find(m_destTiles.begin(), m_destTiles.end(), tile) != m_destTiles.end() && HasTrackdir(TrackStatusToTrackdirBits(GetTileTrackStatus(tile, TRANSPORT_WATER, 0)), trackdir);
11:53:24 <xarick> let's reinstance TrackdirBits m_destTrackdirs;
12:05:42 <merni> OK, I'm working on a PR to delete NPF (whether it will be merged is not what I'm considering right now...)
12:07:55 <merni> #10623 seems to take the second approach
12:08:52 *** Extrems has quit IRC (Ping timeout: 480 seconds)
12:16:14 <xarick> oh, nice, I guess! less work
12:16:41 <xarick> just be aware that river generation uses AyStar
12:16:53 <merni> Yes, I'm not deleting aystar
12:28:28 <xarick> looks like multiple destinations is the culprit
12:29:15 <xarick> if the estimate function works with only 1 tile, it's much faster
12:29:26 <_jgr_> merni: You can delete them from there, but you will need to properly handle the gap in settings_sl_compat.h
12:31:45 <merni> Is `SLC_NULL(18, SL_MIN_VERSION, SLV_REMOVE_NPF),` the right way to handle it? That's what seems to have been done for OPF
12:33:13 <_jgr_> "18" doesn't seem right, but otherwise that should be fine
12:33:29 <merni> What's the number supposed to be? The number of settings you have removed?
12:33:37 <merni> in which case, it is 18
12:34:18 <_jgr_> They're not all 1 byte in size
12:35:17 <merni> I guess I have to do some tedious calculations then
12:35:24 <_jgr_> Some of these settings also have different starting versions
12:35:58 <_jgr_> The penalty values are basically all 4 bytes each, so it shouldn't be too onerous to calculate
12:54:05 <merni> Hm, problem. Aystar uses one function `CheckIgnoreFirstTile` from NPF and I don't really understand the purpose of this function (nor are there any comments on it)
12:58:08 <_jgr_> This function should always return false for non-NPF as AyStarNode::user_data is not used by other AyStar users
12:58:11 <merni> The only check it does seems to be something npf-specific though... so maybe just replace it
12:58:13 <_jgr_> You could just remove it
13:05:49 <merni> I fucked up something or WSL fucked up something
13:06:13 <merni> But I didn't touch anything in graphics code so idk
13:12:38 <_glx_> hehe, happens some times, shutdown wsl and retry
13:15:02 <merni> not that there is anything I can properly test, since I haven't removed/changed the actual setting to switch pathfinders yet
13:17:26 *** ajmiles has joined #openttd
13:17:26 <ajmiles> merni: I saw exactly that when doing my blitter
13:29:49 <truebrain> merni: For compat header you don't even need a new SLV, you can just use the SLV table header; as everything newer will use the actual header written. Ofc if you already have an SLV it is fine to use that.
13:30:21 <merni> I will leave the SLV stuff to the reviewers, like last time :P
13:35:11 <truebrain> yeah, you don't use SLV for anything else, so just use the table-header slv
13:36:19 <merni> What is the table-header SLV?
13:36:26 <truebrain> SLV_TABLE something
13:37:20 <merni> `SLV_TABLE_CHUNKS, ///< 295 PR#9322 Introduction of CH_TABLE and CH_SPARSE_TABLE.`
13:37:35 <truebrain> _compat files are only loaded for SLVs before that version
13:37:45 <truebrain> so there is no point in having any SLV that is newer (also doesn't hurt, but there is no point)
13:38:21 <xarick> the same thing happened last time, when I was doing fixes for OPF, you decided to remove it, now it's NPF
13:47:21 <merni> xarick: Well, this time nobody decided anything yet
13:48:10 <xarick> you already have TrueBrain approval
13:49:11 <kuhnovic> xarick: And that's why you should always measure 😉
13:49:47 <merni> But honestly I don't get why you were working on it in the first place
13:52:59 <merni> Sure, I agree with that
13:53:08 <merni> It's just all we have atm
13:53:29 <merni> Anyway, this will not likely make it into 14.0 so enough time to wait
13:53:44 <truebrain> "not likely" -> "it will not" 🙂
13:56:12 <truebrain> how to crash the game in 2 steps. Step 1) implement a "reset to default" before loading settings. Step 2) load a game
13:58:43 <_glx_> I know other way to crash 🙂
13:59:04 <truebrain> it is just I don't care about those 😛
13:59:51 <xarick> CalcClosestStationTile was already doing a pseudo DistanceManhattan
14:03:05 <_glx_> it's nice when release workflow works on first try
14:03:09 <peter1138> Right, had to trim down the intro a bit but managed to do a fediverse post.
14:03:37 <truebrain> you are getting the hang of it 🙂
14:04:01 <peter1138> Slightly annoying, some servers have much larger character limits.
14:09:05 *** Wormnest has joined #openttd
14:14:53 <truebrain> I think this should work; at least, my testing shows it does.
14:16:13 *** Extrems has joined #openttd
14:16:52 <andythenorth> pff laptop crashed
14:17:06 <truebrain> MacOS .. can't be trusted
14:17:21 <truebrain> merni: lol, AyStar got used by other code? Haha .. that is a bit silly 😄
14:17:36 <truebrain> well, silly .. unexpected 🙂
14:18:52 <truebrain> guess it should be moved away from `pathfinder/npf/aystar.*` 😛
14:19:32 <merni> I mentioned that already :)
14:21:19 <xarick> a ~1.7% worse performance
14:24:12 <peter1138> Ah, infinite money.
14:33:11 *** magdalena[m] has quit IRC (Quit: Client limit exceeded: 20000)
14:43:53 <xarick> could TileIndex be smarter?
14:45:52 <xarick> I wanted to get the x and y of a tile
14:46:45 <xarick> TileIndex tile -> uint x = tile.x, uint y = tile.y
14:48:46 <truebrain> michi_cc: an image with that 10k stat I mentioned earlier 🙂
14:49:45 <truebrain> 25% of that is via Steam. 3% via reddit
14:50:36 <truebrain> needless to say, Steam is the best social platform for us to announce things on 😛
14:52:01 *** Wormnest has quit IRC (Ping timeout: 480 seconds)
14:54:19 *** Wormnest has joined #openttd
14:55:07 <peter1138> And 72% is via me hitting reload?
15:08:59 <andythenorth> we want filesystem io in GS? 😮
15:09:08 <andythenorth> /me can't think how that could be abused
15:09:53 <peter1138> @openttd@fosstodon.org now has 303 followers, nice.
15:11:02 <merni> andythenorth: Will it enable more bad ideas for FIRS-GS?
15:13:40 <peter1138> Well, I was considering a trundle around the shops, but it's raining... again 😦
15:14:26 <xarick> `std::abs` vs `abs`, what is more performant?
15:21:12 <merni> So, should I remove the `_settings_game.pf.pathfinder_for_*` settings altogether or retain them (maybe without showing in GUI) with just one option?
15:23:34 <truebrain> Do we expect a new PF any time soon? 🙂
15:24:26 <merni> Who knows. But the question is, if that happens, is there any point in keeping the settings values from before?
15:24:56 <merni> rather than recreating the setting at that time as required
15:26:56 <merni> There are some weird old settings regarding PF selection too like `new_pathfinding_all` and `*_use_yapf`
15:27:34 <truebrain> What is a bit annoying, settings are not removed from the ini file automatically.
15:29:09 <merni> truebrain: Does that matter? If there are nonexistent settings I guess they will just be ignored?
15:29:31 <merni> Isn't that what allows patchpacks to use the same cfg?
15:29:36 <truebrain> Just when you reintroduce the same setting there might be unexpected values
15:29:45 <peter1138> Given there are some settings that aren't available in the UI, it's a bit annoying as manually editing the config file isn't completely silly.
15:30:20 <merni> In that case it might be better to have it as a setting but not in GUI
15:30:41 <peter1138> I think it's best to remove it instead of leaving something around "just in case"
15:31:25 <truebrain> I tempt to agree, despite my issues with the ini not cleaning up 😛
15:31:49 <merni> I guess it is something to think about if or when NYAPF comes around :P
15:32:10 <truebrain> I did remove some settings explicit, just as I didn't want to give the user the idea it was still functional
15:32:37 <truebrain> Not saying we should in this case
15:34:09 <merni> All right. I assume it will be OK to just delete those settings (and make the changes in `settings_sl_compat.h`) ? Or does any conversion stuff need to be done?
15:35:14 *** as1 has quit IRC (Remote host closed the connection)
15:37:45 <_glx_> conversion is usually needed when setting value options are updated, but if you remove the setting you can't convert anything 🙂
15:46:32 *** Wormnest has quit IRC (Ping timeout: 480 seconds)
15:51:21 <xarick> trying to speed up the estimate calculations
15:55:51 <merni> Is `SDT_BOOL` = 1 byte for the purpose of `SLC_NULL`?
15:58:26 <truebrain> Funny enough, it is stored as an i8 .. always found that funny. A signed boolean.
15:58:27 <_zephyris> As multiple versions of openttd can use the same cfg file, removing entries unused by one version would lead to potentially odd behaviours in another...
15:58:40 <_zephyris> Maybe too weird a corner case
15:58:56 <truebrain> _zephyris: It is why we normally don't remove entries
15:59:30 <truebrain> Still annoying to the end user, seeing a setting that no matter how they change it, doesn't do anything
15:59:50 <Rubidium> truebrain: for real fun, use MSVC which has a signed bit as bool type... so true = -1
16:01:48 <_zephyris> Depreciated in version X comment after obselete settings?
16:02:27 <merni> Rubidium: I think that was true in (at least some implementations of) BASIC too
16:05:36 <merni> yay, lang file change -> full recompile
16:13:56 <merni> Hm, `InvalidateShipPathCache` isn't used by anything except the callback when changing the setting
16:14:53 <merni> Why are only ships' paths cleared but not those of trains/road vehicles?
16:17:56 <merni> Version: 20240302-removenpf-ma86bd28329
16:17:56 <merni> Hash: a86bd28329ba7d3ca11b163319a7f278ff836361
16:17:59 <merni> Reason: Segmentation fault
16:18:01 <merni> ./openttd(_ZNK12CrashLogUnix16SurveyStacktraceERN8nlohmann16json_abi_v3_11_310basic_jsonISt3mapSt6vectorNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEEblmdSaNS1_14adl_serializerES4_IhSaIhEEvEE+0x32) [0x55e713077a50]
16:18:01 <merni> ./openttd(+0x133200b) [0x55e71329500b]
16:18:03 <merni> ./openttd(+0x1339e36) [0x55e71329ce36]
16:18:03 <merni> ./openttd(+0x13381f6) [0x55e71329b1f6]
16:18:05 <merni> ./openttd(+0x1336b20) [0x55e713299b20]
16:18:05 <merni> ./openttd(_ZNKSt8functionIFbvEEclEv+0x32) [0x55e7130781da]
16:18:07 <merni> ./openttd(_ZN12CrashLogUnix10TryExecuteESt17basic_string_viewIcSt11char_traitsIcEEOSt8functionIFbvEE+0x1ca) [0x55e713077d0e]
16:18:07 <merni> ./openttd(_ZN8CrashLog12FillCrashLogEv+0x2e0) [0x55e7132955d8]
16:18:09 <merni> ./openttd(_ZN8CrashLog12MakeCrashLogEv+0x120) [0x55e713297bc2]
16:18:09 <merni> ./openttd(+0x1114628) [0x55e713077628]
16:18:11 <merni> /lib/x86_64-linux-gnu/libc.so.6(+0x3bfd0) [0x7f1c5b2d0fd0]
16:18:11 <merni> ./openttd(_ZNK11SettingDesc12AsIntSettingEv+0x10) [0x55e71357faac]
16:18:13 <merni> ./openttd(_ZN12SettingEntry4InitEh+0x55) [0x55e7135896b3]
16:18:13 <merni> ./openttd(_ZN17SettingsContainer4InitEh+0x6e) [0x55e713589e84]
16:18:15 <merni> ./openttd(_ZN12SettingsPage4InitEh+0x3f) [0x55e71358a5e1]
16:18:15 <merni> ./openttd(_ZN17SettingsContainer4InitEh+0x6e) [0x55e713589e84]
16:18:17 <merni> ./openttd(_ZN12SettingsPage4InitEh+0x3f) [0x55e71358a5e1]
16:18:17 <merni> ./openttd(_ZN17SettingsContainer4InitEh+0x6e) [0x55e713589e84]
16:18:19 <merni> ./openttd(+0x162af10) [0x55e71358df10]
16:18:19 <merni> then bunch of window/video stuff
16:18:21 <merni> These binary offsets are not very useful...
16:18:39 <_jgr_> This generally means that you've removed a setting, but it's still referenced from the settings window
16:19:15 <merni> Right, because the hierarchy in the window is different and built manually...
16:19:24 <_glx_> yeah settings in the window is all manual
16:24:41 *** Wormnest has joined #openttd
16:35:19 *** blathijs has quit IRC (Ping timeout: 480 seconds)
17:02:30 *** teifion has joined #openttd
17:02:30 <teifion> Hi, I develop middleware servers for fun and was wondering what OpenTTD uses for servers. I looked at the docs and they appear focused around setting up game host instances.
17:02:30 <teifion> For those unsure, a middleware server would handle the interim between opening the app and being in a game, that is stuff like auth, lobbies, messaging, moderation, ratings etc etc.
17:03:03 <merni> > lobbies, messaging, moderation, ratings
17:03:03 <merni> None of those exist in openttd
17:03:25 <merni> and "auth" is only as far as entering a password to join a particular server, and then a company in that server
17:03:40 <_glx_> any client can be a server
17:04:16 <teifion> How do you handle discoverability of servers?
17:05:21 <teifion> Skimming the readme this looks to me like what I'd call a middleware server elsewhere, thanks!
17:05:45 <_glx_> master server is outdated, only used for pre-12
17:05:49 <merni> teifion: Perhaps but all it offers is a simple listing
17:06:08 <teifion> Yeah, I was seeing that too. Is there any desire to have something more featureful?
17:06:43 <teifion> for context, I wrote the middleware server for Beyond all Reason and am interested in writing more
17:07:34 <merni> keep in mind, at present there is no concept of "accounts"/"identities"
17:07:43 <_glx_> and local discovery uses plain UDP
17:08:26 <_glx_> it's not even needed to use the listing to join a server if you know its address/port
17:08:37 <teifion> The big reason I've found for middleware servers is for lobby discoverability but there are community advantages too. That said, I'm new here and it might well be the game and community would not want nor benefit from such a thing.
17:09:29 <merni> What exactly do you mean by lobby here?
17:09:47 <merni> Keep in mind openttd is not like a FPS game where a game lasts 15 min and then you start another game
17:09:49 <teifion> I'd call a lobby a place where players gather to setup a game before launching into the game itself
17:10:11 <_glx_> oh it doesn't work like that
17:10:15 <merni> yeah... that's a server admin's job
17:10:36 <merni> because most multiplayer games last several days or longer
17:10:50 <teifion> okay, very different situation then 😄
17:10:56 <truebrain> Maybe start the game up and check it out yourself 🙂
17:11:05 <merni> think of it more like Minecraft
17:11:18 <peter1138> There's two types of lobby, simple one is a chat lobby to find other players wanting to play, the other style is a match-making lobby. Chat lobby is a can-of-worms with things like moderation, and I don't think match-making fits with the complexity.
17:12:06 <peter1138> We already have requests to disable in-game chat so that children can play.
17:12:17 <teifion> Yeah, I think given the duration of games something like discord is completely adequate for finding games to play and a list of lobbies as you'd see in other games wouldn't be a good fit.
17:12:17 <teifion> Chat moderation is indeed a can of worms.
17:12:18 <peter1138> (Very rarely mind you)
17:12:19 <merni> merni-nsviaGitHub: Hm, does Regression test for NPF somehow and that's why it's failing?
17:12:32 <_glx_> children should never play on public servers anyway
17:12:44 <merni> "Broken savegame - Invalid chunk size in RIFF in PATS - expected 387, got 461"
17:13:05 <_glx_> you failed in _sl stuff
17:13:35 <_glx_> means the table data doesn't match what the savegame has
17:13:52 <peter1138> Probably wrong lengths for removed stuff.
17:14:06 <merni> Let me look at that again
17:15:00 <_glx_> guess you were lucky the intro game loaded fine 🙂
17:16:05 <peter1138> Should we add a savegame of every version just for testing? 😉
17:16:06 <merni> OK, so when a setting has `to = SLV_87`, do I use that as the last version for the compat stuff?
17:16:06 *** as1 has quit IRC (Remote host closed the connection)
17:16:56 <merni> Back into the world of `git rebase --interactive`...
17:17:45 <_glx_> `from =` is important too
17:17:59 <merni> yeah, I got the from part
17:18:33 <merni> but I just used the `SLV_TABLE_CHUNKS` for all the `to`s
17:19:18 <truebrain> Only when SLV is after table chunk 🙂
17:20:11 *** Wormnest has quit IRC (Ping timeout: 480 seconds)
17:21:43 <merni> nice, another full recompile
17:30:46 <_glx_> check `SDT_NULL = SDT_NULL( $length, $from, $to),` in old_gameopt_settings.ini
17:31:24 <_glx_> the settings ini should still have the data info, even if it's no longer used
17:32:09 <_glx_> and of course it's not even used here
17:32:40 <merni> _glx_: what do you mean?
17:35:43 <_glx_> the ini are used to generate _XXX_settings_table, which are then used to determine what to read/write from/to the savegame
17:36:46 <_glx_> hmm but it doesn't matter since table header
17:37:22 <merni> so... do I need to do something different than what I have done, or not? :P
17:37:48 <_jgr_> _glx_: old_gameopt_settings is for the OPTS chunk, not PATS
17:38:35 <merni> well, regression passed this time
17:38:42 <_glx_> yes I know, but it defines SDT_NULL without using it, so now I'm digging into history to find why 🙂
17:42:19 <_glx_> seems it was added during the setting split but never used
17:54:46 *** Wormnest has joined #openttd
18:05:25 <peter1138> Didn't you already make a list of docking tiles per station?
18:06:42 *** as1 has quit IRC (Remote host closed the connection)
18:06:58 *** belajalilija has joined #openttd
18:06:58 <belajalilija> suggestion #1: the search feature should also read the grf file name
18:07:20 <belajalilija> suggestion #2: the picker tool should state the grf that something is from
18:08:06 <merni> I think 2 is already in jgrpp at least
18:08:43 <xarick> peter1138: I did, but I thought it was risky because it was acting as a cache
18:09:53 *** brickblock19280 has joined #openttd
18:09:53 <brickblock19280> picker tool isn't in vanilla but landscape info is so you might be mixing those two up
18:10:51 <brickblock19280> why would this happen when I install from the website?
18:11:04 <brickblock19280> is it the fact I am using jgrpp?
18:11:34 <_jgr_> belajalilija: It does already in vanilla
18:11:45 <belajalilija> just not in jgr yet?
18:11:51 <belajalilija> or not mine i guess
18:13:10 <_jgr_> I tested it in 0.56.1 and it works fine there
18:14:04 <peter1138> Does it need developer tools on?
18:14:39 <belajalilija> thats what i see
18:15:07 <_jgr_> That's rail track, not a GRF object/station/etc
18:15:59 <belajalilija> it comes from a grf, but yeah, tested on a station, that works
18:19:05 <belajalilija> it would be cool for it to find both from what it does and the grf file name
18:19:38 <xarick> well, it's either hella fast, or it's misreporting
18:19:53 <xarick> the values for CalcClosestStationTile
18:21:13 <xarick> and yes, NPF calls this function an insane amount of times
18:27:37 <xarick> if it's not misrepresented, I'm going to assume it's fast enough
18:28:18 <peter1138> belajalilija: Damn, such overly wide icons 😮
18:29:03 <belajalilija> peter1138: DACH train set
18:29:08 <belajalilija> i think you mean
18:29:37 <peter1138> Yeah, preview sprites.
18:34:51 <DorpsGek> - Update: Translations from eints (by translators)
19:36:55 <truebrain> `nothing could have been more unnecessary` .. you got to love some replies on Steam blog posts, lol
19:56:27 <Rubidium> that's about supporting MacOS, right? :D
20:29:16 <andythenorth> not quite sure what I've done here 😛
20:29:33 <andythenorth> one generation of box vans has parented to another generation
20:29:43 <andythenorth> which is parented to narrow gauge box vans which aren't buildable in this depot
20:34:55 <andythenorth> oof the nml is giving me no clues 😛
20:38:45 <andythenorth> this is also a bit exciting
20:39:42 <_glx_> I guess something is wrong in your generator
20:40:31 <_jgr_> Having that many almost identical wagons seems overkill regardless of the tree structure?
20:41:06 <_glx_> different colour scheme
20:42:26 <andythenorth> different generations, different tracks
20:42:33 <andythenorth> but yes, overkill is a feature
20:46:55 <locosage> truebrain: don't worry, you can still prove them wrong in 15.0 😜
20:48:01 <xarick> it's fast now, and has multiple destinations!
20:48:10 <xarick> need to update the PR description
20:49:21 <andythenorth> ach, these variant group bugs are bizarre
20:49:49 <andythenorth> did anything change in variants recently?
20:50:29 <locosage> there were some changes after 13.4
20:52:54 <locosage> andythenorth: what are you even trying to achieve?
20:53:02 <locosage> keep in mind that not every tree structure is possible
20:55:35 <andythenorth> should look like
20:55:51 <andythenorth> but vehicles are being grouped that should not be, recursively
20:56:35 <andythenorth> there will be an incorrect group ID somewhere, but I can't find it in the nml
20:56:48 <locosage> well, at least that should be possible...
20:57:02 <andythenorth> it also doesn't appear consistently
20:58:02 <andythenorth> the 'available vehicles' window now looks like
20:58:20 <andythenorth> whereas before, these appeared after the last Brake Van
20:58:27 <andythenorth> not sure how to debug that in nml
20:58:56 <andythenorth> nothing changed in the grf, I just opened and closed windows and variant groups
20:59:08 <andythenorth> there's no callback for variant groups?
21:00:00 <locosage> it this point I usually start to add debug output into openttd code...
21:02:28 <andythenorth> it's reproducible in Horse 3.7.1 from bananas
21:02:42 <andythenorth> but I am 99% certain it's just a wrong variant group ID somewhere in the nml
21:06:52 <locosage> oh, yeah, can confirm
21:11:31 <andythenorth> I changed a bunch of stuff relating to trying to preserve varact 2 IDs
21:11:50 <andythenorth> /me picking through nml 😛
21:16:12 <locosage> 'climates_available': grf.NO_CLIMATE,
21:16:14 <locosage> 'variant_group': 26860,
21:16:47 <_glx_> feels wrong to point to itself
21:17:02 <locosage> yeah, even if it's disabled
21:17:36 <andythenorth> I was just reading the spec for those
21:17:46 <andythenorth> not sure how to unset a variant group
21:17:56 <_glx_> not sure how OpenTTD handles this case
21:18:44 <andythenorth> can't remember why it's there
21:19:11 <_glx_> I think unset would be some kind of invalid id
21:20:12 <_glx_> yup INVALID_ENGINE (0xFFFF)
21:21:51 <_glx_> /* Engine looped back on itself, so clear the variant. */
21:21:51 <_glx_> e->info.variant_id = INVALID_ENGINE;
21:23:25 <_glx_> -d grf=1 should be noisy for you 🙂
21:24:46 <andythenorth> `Unhandled vehicle variable 0x61, type 0x0`
21:26:13 <andythenorth> lot of var 61 going on eh
21:26:47 <peter1138> Is it cheese o'clock?
21:27:16 <_jgr_> andythenorth: That doesn't seem like soemthing that would make sense in a purchase list
21:27:28 <andythenorth> think it's only on constructing a vehicle
21:27:31 <andythenorth> unless I missed something
21:27:42 <andythenorth> (unrelated to the variants issue)
21:40:13 <_jgr_> It's part of the graphics chain so also reached in the purchase list, but it doesn't really matter that it fails
21:41:21 <peter1138> Sort of, if it's unsupported it stops processing the chain.
21:41:43 <peter1138> Oh, there's a fallback "error group" now.
21:42:19 <peter1138> `/* Dynamically allocated, this is the sole owner */`
21:42:27 <peter1138> Funny, because it's completely wrong 😄
21:45:47 <andythenorth> maybe I should write a python script to show the tree 😛
21:45:59 *** SigHunter_ has joined #openttd
21:48:07 <peter1138> `free(): invalid next size (fast)`
21:48:42 <truebrain> peter1138: I had cheese. It was good.
21:49:04 *** SigHunter has quit IRC (Ping timeout: 480 seconds)
21:49:14 <peter1138> NewGRF scan with valgrind 😦
21:50:51 <truebrain> Someone keeps saying he is going to cache it in a JSON file .. he never did 😛
21:51:36 <truebrain> Awh, he isn't even online .. 😛
21:52:44 <truebrain> Wasn't talking about you, sorry 🙂
21:53:09 <truebrain> You just have patches for random stuff 😄
21:53:23 <andythenorth> still can't see which variant_group ids are wrong 😛
21:56:04 *** nielsm has quit IRC (Ping timeout: 480 seconds)
21:56:09 <andythenorth> disabling the incorrect narrow gauge 'parent' just moves the problem up the tree
22:22:22 <truebrain> I bisected some food. It was good.
22:24:05 <andythenorth> which is mostly just changing IDs
22:24:47 <frosch123> Json? Ottd uses ini/toml
22:25:06 <truebrain> Owh, wait, you were the sqlite one
22:25:30 <truebrain> Did you know our ini is not compliant? 😄
22:26:22 <truebrain> Nope. But some kind of software doesn't understand our ini file, so they blamed us for not following "the standard"
22:26:40 <truebrain> With which they meant whatever Go implemented as "the standard"
22:27:38 <frosch123> I bet it was something from opendesktop
22:27:49 <truebrain> Some kind of server management tool
22:27:55 <truebrain> Forgot the name etc
22:28:31 <truebrain> But you say toml ... we don't do toml 😛
22:29:19 <frosch123> I use them synonymous
22:29:32 <truebrain> One is actually a standard, sadly 😛
22:30:10 <truebrain> We could be toml compliant, but does require a few changes 🙂
22:31:17 <Rubidium> ini => "is not ini" ;)
22:31:45 <frosch123> rather get rid of singleplayer
22:32:09 <truebrain> Lol, how do these two compare? 😄
22:32:15 <peter1138> Just run the game twice.
22:32:43 <frosch123> They seem to compete for truetime
22:33:11 <truebrain> Haha, I ain't touching the ini writer, no worries
22:33:53 <frosch123> Don't touch toml writers either
22:35:16 <truebrain> Well, personally more of a yaml guy, but okay
22:36:09 *** keeeek7218 has joined #openttd
22:36:35 <truebrain> Apple can suck my ....
22:37:40 <truebrain> Question of the week: what is open about OpenAI?
22:37:57 <truebrain> Answer: it opens your wallet
22:38:20 <keeeek7218> just 20$/month for gpt plus!
22:42:54 <andythenorth> it's a tax I have to pay to know what GPT is
22:43:00 <peter1138> What is AI about OpenAI?
22:52:37 <truebrain> How is the title competition going talltyler ?
22:58:34 <talltyler> I have finished my entry. Looks like one other entry so far — but in past years we’ve had a flurry right before the deadline
23:06:08 <andythenorth> dunno what to do about all the failing 0x61 but eh
23:06:14 <andythenorth> fixed the variants
23:20:01 *** keikoz has quit IRC (Ping timeout: 480 seconds)
23:24:29 *** Leopold_ has quit IRC (Ping timeout: 480 seconds)
23:24:30 *** Leopold has joined #openttd
23:49:13 <xarick> One other thing I'm experimenting with. FindNearestDepot on top of 12206
23:49:40 <xarick> each company has ~500 ship depots
23:50:47 <xarick> I'm impressed it didn't completely tank performance
23:57:18 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
continue to next day ⏵
