IRC logs for #openttd on OFTC at 2021-11-27
⏴ go to previous day
00:11:35 <DorpsGek> [OpenTTD/OpenTTD] ldpl commented on pull request #9716: Fix: Don't try to rename OWNER_DEITY signs in-game https://git.io/JM3tD
00:53:51 *** frosch123 has quit IRC (Quit: be yourself, except: if you have the opportunity to be a unicorn, then be a unicorn)
01:37:00 *** Compu has quit IRC (Ping timeout: 480 seconds)
02:48:55 *** gelignite has quit IRC (Ping timeout: 480 seconds)
03:17:50 *** geli has quit IRC (Quit: Stay safe!)
03:24:38 *** Wormnest has quit IRC (Quit: Leaving)
03:51:18 *** roadt has quit IRC (Ping timeout: 480 seconds)
03:58:37 *** debdog has quit IRC (Ping timeout: 480 seconds)
04:02:20 *** D-HUND is now known as debdog
04:08:58 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
04:26:37 *** _aD has quit IRC (Quit: leaving)
04:34:45 *** felix has quit IRC (Ping timeout: 480 seconds)
04:48:40 *** felix has quit IRC (Ping timeout: 480 seconds)
04:55:15 *** felix_ has quit IRC (Ping timeout: 480 seconds)
05:14:06 *** while has quit IRC (Read error: Connection reset by peer)
06:35:26 *** tokai|noir has joined #openttd
06:35:26 *** ChanServ sets mode: +v tokai|noir
06:42:32 *** tokai has quit IRC (Ping timeout: 480 seconds)
07:03:18 <greeter> playing this game again after a 2 year hiatus. forgot how much fun this can be, especially when other games have you stressed out lol
07:37:19 *** WormnestAndroid has quit IRC (Remote host closed the connection)
07:47:57 *** ChanServ sets mode: +v tokai
07:54:47 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
08:28:02 *** andythenorth has joined #openttd
08:43:26 *** jottyfan has joined #openttd
08:45:47 *** andythenorth has quit IRC (Quit: andythenorth)
08:48:46 *** andythenorth has joined #openttd
09:50:05 *** supermop_Home_ has quit IRC (Ping timeout: 480 seconds)
10:15:06 <andythenorth> maybe I can install arm native photoshop
10:15:17 <andythenorth> then it will stop blocking 1 core even when idle
10:15:25 <andythenorth> such computing adventures
10:24:38 <peter1138> Charge you extra for that
10:29:30 <dP> If I replace 16 "new" shore sprites with action A does it include the 8 "old" ones as well or do I still need to replace them separately for some reason?
10:29:35 <dP> opengfx seems to replace both
10:30:30 *** HerzogDeXtEr has joined #openttd
10:31:12 <dP> hm, or maybe it was opengfx+ landscape or something...
10:45:38 <dP> yeah, it was landscape, what opengfx does I've no idea
10:45:54 <dP> why are there 2 sets of shore sprites in opengfx for everything but toyland?
10:47:19 *** while has quit IRC (Remote host closed the connection)
10:57:14 <andythenorth> many years later
10:57:20 <andythenorth> I have arm photoshop
10:57:26 <andythenorth> first upgrade creative cloud
10:57:46 <andythenorth> then it sequentially downloads and installs 4 entire versions of photoshop to get the latest binary
10:57:57 <andythenorth> many gigabytes of bandwidth wasted
10:59:54 <andythenorth> but only 20% of a core used, instead of 60%
11:00:12 <nielsm> it's funny how linux and bsd distributors solved those problems of software updates 20-25 years ago and big corps still keep reinventing things poorly
11:00:53 <andythenorth> Adobe is probably not a destination employer
11:24:11 *** frosch123 has joined #openttd
12:26:28 <frosch123> does it keep retrying when stuff fails?
12:27:11 <TrueBrain> these are fire-and-forget events
12:28:19 <frosch123> after the nightly was finished yesterday, it ran #412-415, with about 1 hour waiting inbetween
12:28:33 <TrueBrain> yeah, as I said: when new stuff gets uploaded to the CDN, it triggers a new website
12:30:01 <TrueBrain> (remember we have more than only the OpenTTD nightly ;) )
12:30:55 <frosch123> i don't remember that :p
12:31:19 <TrueBrain> OpenSFX? OpenMSX? Rings a bell? :D
12:31:25 <TrueBrain> everything eints can push an update to, basically? :P
12:32:16 <frosch123> ok, so it's all the fault of the new galician translator :)
12:51:59 *** andythenorth has quit IRC (Quit: andythenorth)
13:32:29 *** andythenorth has joined #openttd
13:32:35 *** WormnestAndroid has joined #openttd
13:52:13 *** gelignite has joined #openttd
14:34:55 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
14:50:29 *** supermop_Home has joined #openttd
14:50:39 *** WormnestAndroid has joined #openttd
14:58:01 <supermop_Home> how's it going andy?
15:42:50 *** jottyfan has quit IRC (Quit: jottyfan)
16:40:48 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
16:54:02 <andythenorth> it's going cold :P
16:58:30 *** Flygon has quit IRC (Quit: A toaster's basically a soldering iron designed to toast bread)
17:17:43 *** Wormnest has joined #openttd
17:18:52 *** WormnestAndroid has joined #openttd
18:06:52 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler updated pull request #9716: Fix: Don't try to rename OWNER_DEITY signs in-game https://git.io/JMeZx
18:43:18 *** Wolf01 is now known as Guest6961
18:46:01 <DorpsGek> [OpenTTD/OpenTTD] 2TallTyler opened issue #9719: [Bug]: Transfer credits calculated wrong with negative cargo values https://git.io/JMnl7
18:47:42 <DorpsGek> - Update: Translations from eints (by translators)
18:49:26 *** Guest6961 has quit IRC (Ping timeout: 480 seconds)
18:59:35 *** felix has quit IRC (Ping timeout: 480 seconds)
19:12:35 *** felix has quit IRC (Ping timeout: 480 seconds)
19:24:55 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
19:25:03 *** WormnestAndroid has joined #openttd
19:32:10 *** felix has quit IRC (Ping timeout: 480 seconds)
20:05:48 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
20:05:58 *** WormnestAndroid has joined #openttd
20:10:10 *** felix_ has quit IRC (Ping timeout: 480 seconds)
20:36:17 <DorpsGek> [OpenTTD/OpenTTD] glx22 commented on issue #9720: [Crash]: Can't start game with CityGrowthLimiter GS https://git.io/JMn1r
20:41:04 <glx> github was dying when I commented, now it's down for me :)
20:42:17 <glx> but I don't see what could crash here
20:42:39 <frosch123> i tried running the script on a dedicated server, nothing special
20:43:00 <frosch123> maybe it's something silly like oom
20:43:18 <frosch123> no, the script is trivial
20:43:25 <frosch123> it does not allocate much
20:43:38 <nielsm> github is still down for me
20:44:05 <nielsm> the savegame appears to have modified GRF config, but I doubt it'd be related for this
20:44:26 <frosch123> the savegame was 0 bytes for me
20:44:42 <nielsm> the crashlog has the game log
20:44:59 <frosch123> well, the game was just started
20:45:28 <frosch123> GSController.GetSetting is pretty much the first thing the GS calls
20:46:12 <frosch123> ok, let's actually try it via a scenario
20:46:17 <frosch123> it's definitely unusual :p
20:47:21 <nielsm> ask for two things: have the reporter try with a fresh game (new game) using the same GS and GRF configs instead of using the scenario, and have the reporter post the scenario file
20:49:36 <glx> so the scenedit step is important
20:51:14 <frosch123> gdb takes ages to start again :(
20:52:17 <frosch123> ScriptInfo "this" is nullptr
20:52:52 <frosch123> scriptconfig is valid, but "info" is nullptr
21:00:07 <glx> only thing I can imagine is ScriptConfig::Change() not being called
21:02:14 <frosch123> it's still valid in StringToSettings
21:04:26 <frosch123> hmm... where did i lose it
21:05:49 <glx> only way for info to become nullptr is Change(nullptr), but then the script should be dead/non running
21:09:43 <frosch123> there is some chicken-egg thing
21:10:13 <frosch123> somehow the script already executes in script_instance.cpp:84
21:10:26 <frosch123> but the real instance is only created in line 92
21:11:20 <nielsm> ah right, I saw this when doing the title screen contest, but that was with AIs failing to load and erroring in a similar way with nullptr configs, because a different AI was being loaded when I didn't have the reqested AI installed
21:12:48 <frosch123> "_settings_game.game_config->info" is valid during the crash
21:13:00 <frosch123> but because the instance is not really started, it accesses a different object
21:13:33 *** supermop_Home has quit IRC (Ping timeout: 480 seconds)
21:13:33 <glx> line 84 loads and compiles the script
21:14:12 <frosch123> when it crashes it actually tries to access _settings_newgame.game_config
21:14:27 <frosch123> and newgame-settings have no info=nullptr
21:15:02 <frosch123> glx: that's what i though, but apparently it already excutes there :/
21:17:15 <frosch123> town.nut calls GSController.GetSetting to initialise static variables
21:17:24 <frosch123> but GSController is not ready yet
21:18:53 <nielsm> so the fix is to make sure GSController.GetSetting crashes the script if it's called before the script has really started
21:19:05 <frosch123> _game_mode = GM_MENU <- lol
21:19:11 <frosch123> that's what fails :)
21:19:30 <frosch123> GameConfig::GetConfig then picks the wrong config
21:19:36 <glx> but that means savegame failed to load
21:19:49 <frosch123> no, it's during loading
21:19:53 <frosch123> loading is not finished yet
21:20:13 <frosch123> all this happens inside GSDTChunkHandler::Load
21:20:41 <frosch123> actualy... not even AfterLoad ran yet.. and the GS is already in busyness :p
21:20:49 <frosch123> who designed this? :p
21:21:00 <glx> so it's the Load() inside the GS ?
21:22:00 <frosch123> the GS immediately reads the setting, but because loading is still in progress it actually gets the newgame settings,and crashes
21:22:23 <frosch123> but imho, it's completely unsafe to execute any squirrel code while the savegame is partially loaded :p
21:23:40 <frosch123> so, problem is GSDTChunkHandler
21:24:28 <frosch123> Game::StartNew + Game::Load must be moved to way later
21:25:00 <glx> same probably true for AIs too
21:26:22 <frosch123> gh still down for me, can't comment :)
21:26:49 <nielsm> github is going further down still
21:26:57 <glx> but if it happens only with -D it's maybe only something wrong in dedicated mode
21:26:58 <nielsm> pages and webhooks are also failing now
21:27:02 <frosch123> still amazing how such a fundamental bug only shows up after 12 years :p
21:27:34 <frosch123> glx: i guess newgame-config just happens to be valid in non-dedicated mode
21:27:42 <frosch123> so it does not crash, but reads incorrect values
21:28:00 *** Wormnest has quit IRC (Ping timeout: 480 seconds)
21:28:44 <glx> and yes AI use the exact same sequence
21:29:22 <glx> StringToSettings(), start new and load
21:30:07 <frosch123> no idea what GS/AI do in their Load function, but if they actually use any API methods there, instead of just loading data, a lot will break
21:32:20 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
21:32:45 *** WormnestAndroid has joined #openttd
21:33:54 *** supermop_Home_ has joined #openttd
21:39:31 <glx> oh but it's not really loading at that time
21:39:41 <glx> dedicated does weird things
21:39:55 <glx> if (SaveOrLoad(_file_to_saveload.name, _file_to_saveload.file_op, _file_to_saveload.detail_ftype, BASE_DIR) == SL_ERROR) {
21:39:55 <glx> /* Loading failed, pop out.. */
21:39:55 <glx> Debug(net, 0, "Loading requested map failed; closing server.");
21:39:56 <glx> /* We can load this game, so go ahead */
21:39:56 <glx> _switch_mode = SM_LOAD_GAME;
21:40:10 <glx> it crashes when testing it can load
21:52:39 <andythenorth> how many IDs are valid for articulated parts? First 16k?
21:52:48 * andythenorth is worryingly high in Horse
21:53:07 <frosch123> callback result is 14bit
21:53:30 <frosch123> if we remove the ctrl+click thingie to reverse vehicles, we can increase it to 15bit, so you had 32k
21:54:08 <andythenorth> that is Sophie's Choice
21:55:41 <andythenorth> maybe reversing should be done with the livery callback?
21:56:09 <andythenorth> so instead of 'liveries * refittable cargos' it would be 'liveries * refittable cargos * reverse states'
21:56:45 <andythenorth> let's say 4 liveries, FIRS, reversable
21:57:14 <andythenorth> @calc 4 * 48 * 2
21:57:32 <andythenorth> and then there might be mixed vehicles in the consist, which causes the strange union of options I never understand
21:57:37 <andythenorth> so maybe 3 types
21:57:53 <andythenorth> perfectly sensible refit menu
21:58:53 <glx> oh progress on github, now everything is red, except webhooks and pages (they are orange)
21:59:18 <frosch123> yeah, complete outage on a weekend
21:59:39 <frosch123> someone is having a nice weekend
22:00:02 <frosch123> maybe they are lucky and management is golfing, so they don't interfer
22:01:00 <andythenorth> it's a sign of responsible management to call the engineers every 5 minutes
22:01:07 <glx> management should just provide coffee
22:01:11 <andythenorth> otherwise how else do you know they are trying to fix it?
22:01:20 <frosch123> glx: in homeoffice?
22:01:34 <frosch123> you want your manager to visit you to bring mcdonalds coffee?
22:01:37 <andythenorth> I used to order pizza and draw pixels when there was an outage
22:01:44 <andythenorth> and check the whiteboard pens didn't run out
22:01:58 <andythenorth> half of Horse probably got done in outages
22:02:22 <frosch123> ah, you actually work at gh
22:02:39 <frosch123> that's why you started with 1152 livery options
22:05:35 <andythenorth> that is the intended way, yes?
22:05:43 <andythenorth> discord keeps telling me I am wrong not to use it
22:05:49 <andythenorth> and that it's definitely definitely not broken
22:05:55 <andythenorth> even though I have made 2 videos showing it is
22:06:07 <frosch123> you only have 255 subtype options though
22:06:29 <frosch123> you have to combine them with the 64 actual cargos to get more liveries
22:06:49 <glx> and discard auto replace IIRC
22:07:05 <frosch123> no, autoreplace/renew works well
22:07:22 <frosch123> it will pick the same livery, if available
22:07:45 <andythenorth> based on the string?
22:08:05 <andythenorth> hmm, what if the cargos actually encode the liveries?
22:09:03 *** Wormnest has joined #openttd
22:09:14 <frosch123> subtypes got their bad reputation, when people used it for non-visual stuff, like vehicle-length and capacity, because people could not autoreplace all vehicles to a different subtype/capacity
22:12:18 <andythenorth> who did that? :o
22:12:32 <frosch123> it probably started with regearing
22:12:39 <frosch123> no idea who did it first
22:12:45 <glx> oh like the nml example ?
22:13:05 <andythenorth> or Unsinkable Sam
22:13:59 <glx> nml example is a MU with 3 or 4 vehicles (depending on the livery)
22:14:43 <andythenorth> such cool features
22:14:50 <frosch123> yeah, if the player wants to make the vehicle longer, they have to send all vehicles to depot, and manually refit :)
22:14:53 <andythenorth> I did invent a livery spec
22:15:00 <andythenorth> would probably be terrible
22:15:47 <frosch123> but well, i guess template-based autoreplace would make subtype-replacements work as well
22:16:34 <glx> main issue will be the GUI for it (as often)
22:16:38 <frosch123> refitting trams from 3 to 4 parts is not much different to attaching a wagon to a train
22:19:12 <andythenorth> template based replace
22:19:18 * andythenorth dreams of such things :)
22:20:30 <frosch123> i learned, if you play quick NCG games, that last less than 10 years, you never need autoreplace
22:24:24 <andythenorth> or daylength x 4
22:24:39 * andythenorth never tried daylength
22:24:45 <andythenorth> but date cheat works :P
22:27:40 *** EER has quit IRC (Remote host closed the connection)
22:44:58 <nielsm> 2 hours of downtime for github now
22:45:12 <dP> I was kinda thinking about template replace recently, is it even possible with newgrfs?
22:45:22 <dP> iirc it's impossible to preview a train without building it
22:46:08 <nielsm> I think that's also basically what the templates do, as implemented... build the trains in an off-map depot for imaginary money, then clone them into real depots when replacing
22:47:13 <frosch123> yes, you need some kind of train-building sandbox for templates
22:47:28 <frosch123> something that appears like real trains to newgrf
22:47:47 <andythenorth> I broke the patch included in JGRPP, but eh
22:47:51 <andythenorth> I break everything
22:48:03 <frosch123> i recall it broke you :p
22:48:20 <andythenorth> it's funny how it removes any other trains in the depot when template replacing
22:48:25 <andythenorth> kind of a lol feature
23:00:57 *** gelignite has quit IRC (Quit: Stay safe!)
23:10:48 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
23:10:58 *** WormnestAndroid has joined #openttd
23:13:36 *** _aD has quit IRC (Quit: Lost terminal)
23:18:43 *** andythenorth has quit IRC (Quit: andythenorth)
23:23:42 <glx> I hope they will blog about it
23:54:04 <DorpsGek> [OpenTTD/OpenTTD] frosch123 commented on issue #9720: [Crash]: Can't start game with CityGrowthLimiter GS https://git.io/JMn1r
23:58:21 <frosch123> yay, dorpsgek also works
continue to next day ⏵
