IRC logs for #openttd on OFTC at 2025-12-28
⏴ go to previous day
00:02:04 *** MinchinWeb[m] has joined #openttd
00:07:21 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
01:02:19 *** ChanServ sets mode: +v tokai
01:09:14 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
02:12:32 *** MinchinWeb[m] has quit IRC (Ping timeout: 480 seconds)
02:17:07 *** MinchinWeb[m] has joined #openttd
02:47:23 *** Wormnest has quit IRC (Quit: Leaving)
03:28:04 <reldred> @peter1138 nice work!
04:17:03 *** hexdog has quit IRC (Remote host closed the connection)
04:47:02 <DorpsGek> - Update: Translations from eints (by translators)
05:32:17 *** MinchinWeb[m] has quit IRC (Read error: Connection reset by peer)
05:35:43 *** MinchinWeb[m] has joined #openttd
05:42:18 <reldred> Also desert fences and desert farms when 😭
05:48:52 <reldred> Probably needs some grf wizardry around those though. Only really makes sense for pineapple landscape or other situations where desert != irradiated hellscape
06:07:31 <emperorjake> Snowy fields would be nice too
06:10:12 <reldred> also peter the children still yearn for the RGB/HSL CC 🙏
06:11:45 <reldred> Snowy fields really needs grf aware fields though, and grf aware fields could then eventually let us have varied graphics controlled by the industry
06:12:04 <reldred> which you know, would be absolutely *fantastic*
09:29:23 <_zephyris> NewGRF fields, woop
09:29:50 <_zephyris> Would really enhance how the world looks
09:30:21 <_zephyris> More generally, a few more options for random start of game object placement would be nice
09:31:17 <_zephyris> Lighthouse-like distribution, transmitter-like distribution, and a few creative new options
11:17:36 <andythenorth> like...a script? 👀
11:29:56 <sittinbythefire> Fields would be cool. I was just thinking how nice it would be if I could make farm fields look greener over time if water/supplies were delivered
11:30:16 <sittinbythefire> Would be a nice effect to show progress 🙂
12:54:52 <emperorjake> I've realised an interesting side effect - because the 15.0 title screen is toyland, it uses the toyland palette for the water pixels in the climate icons
13:40:31 *** dojando has joined #openttd
13:40:31 <dojando> I think the toyland water should be replaced with honey or chocolate 😁
13:45:05 <emperorjake> I think other climates (especially tropic) should be able to use the toyland water
14:25:09 <xarick> I'm having a hard time deciding what's the best std map set build
14:27:26 <Rubidium> the one with lowest cyclomatic complexity?
14:29:17 <xarick> the effect of value iter tracking: it's strong showing in ValuateTests
14:29:31 <xarick> but a poor show in RemoveItemTests
14:32:20 <xarick> I am almost certain the poor show in RemoveItemTests is due to swap list
14:32:57 <xarick> because all the iterators need to be refreshed
14:33:17 <xarick> and there's 2-6 million iterators
14:34:22 <Rubidium> that'd be about 2-6 million (give or take a dozen) too many iterators
14:34:27 <xarick> let me test that test without SwapList... just curious
15:04:43 <_zephyris> Idle thought, would it be possible to embed game and add-on metadata in screenshots?
15:05:55 <_zephyris> Should be possible to dump map settings and seed, newgrf list, other bananas IDs etc. in the EXIF data.
15:12:57 <xarick> apparently I triggered an infinite loop :|
15:19:30 <_zephyris> So it's already embedded?
15:19:53 <Rubidium> well, part of what you suggested is already implemented/embedded
15:21:08 <Rubidium> but yes, any unmodified PNG screenshot from the last ~15 years should have OpenTTD version, baseset + version, NewGRFs and companies/AIs
16:07:39 <xarick> just discovered a bug in my viter tracking :| I am very disappointed in myself
16:11:37 <xarick> "map/set iterators incompatible", but how
16:14:46 <xarick> so... the SwapList was actually saving my stuff from crashing
16:15:23 <xarick> now without it, there's some iterator mismatch :|
16:15:59 <xarick> i expected the crash the other way around
16:24:03 <xarick> I need to think. small sample, I start by adding 10 items values_inited is false therefore there is no viter values and no values
16:24:58 <xarick> next I cast .Begin on the list, this should fill values and initialize all viters, let's see if that happens
16:31:57 <xarick> now i'm into the _set metamethod
16:32:12 <xarick> setting new values to the items
16:35:58 <xarick> try_emplace wondering if try_emplace is a bad choice
16:36:27 <xarick> this->items.try_emplace(item, ItemRecord(value))
16:37:05 <xarick> ItemRecord is not modified I assume
16:38:51 <xarick> value is updated. viter still points to the correct iterator in this->values list
16:40:49 <xarick> ``` /* Re-insert node handle */
16:40:49 <xarick> auto value_iter = this->values.insert(std::move(node_handle)).position;
16:40:49 <xarick> /* Update cached iterator */
16:40:49 <xarick> item_iter->second.viter = value_iter;```
16:41:35 <xarick> viter has a new fresh iterator
16:41:44 <xarick> item_iter value updated
16:41:55 <xarick> so, i don't get it where this crash comes form
16:54:43 <Rubidium> welcome to the wonderful world of UB
17:03:22 <LordAro> hmm, who turned a setting into a list recently?
17:20:14 <Rubidium> can't find anything obvious
17:21:19 <Rubidium> I've got no clue how to get to the specific json that causes the issue, so can't even exclude it being something malicious
17:48:43 <xarick> i have a vague idea where the problem originates from
17:49:12 <xarick> ScriptList::AddList when the list is empty
17:50:07 <xarick> `this->items = list->items;` "this" list is empty, "list" list items have iterators pointing to list->values
17:50:49 <LordAro> Rubidium: first step may be improving the exception message :)
17:50:53 <xarick> a copy will make "this" items to have iterators pointing to "list" values
17:51:27 <LordAro> should be able to use the repo secrets to get at the cloudflare data ;)
18:00:28 *** Flygon has quit IRC (Read error: Connection reset by peer)
18:02:55 <_glx_> I don't see any changes in settings
18:04:29 <peter1138> I did changes to font settings a while back, and then reverted them.
18:05:44 <_glx_> yeah but it's not a list
18:10:41 <_glx_> unless it's the "configured"/"active" change
18:12:10 <_glx_> push_back feels like list
18:14:43 <_glx_> I guess any survey.json between #13303 and its revert should give enough data to reproduce the error
18:18:34 <_glx_> crash logs from #14898 also contain survey data
18:30:43 *** Wormnest has joined #openttd
18:33:00 *** Smedles has joined #openttd
18:54:02 *** MinchinWeb[m] has quit IRC (Ping timeout: 480 seconds)
18:54:45 *** MinchinWeb[m] has joined #openttd
19:04:28 *** gelignite has joined #openttd
19:25:45 *** Wormnest has quit IRC (Quit: Leaving)
19:55:04 *** toktik has quit IRC (Remote host closed the connection)
20:03:07 *** Borg has quit IRC (Quit: leaving)
20:31:33 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
20:31:50 *** WormnestAndroid has joined #openttd
20:43:45 *** chucky76 has joined #openttd
20:43:45 <chucky76> Just a little question: Has the intro menue been changed or is this for developing only?
20:44:19 <belajalilija> no, they just did it and released it as a joke
20:53:49 <chucky76> belajalilija: Really? Are you serious?
20:54:05 <rito12_51026> Simple rebase after one and a half month
20:54:37 <chucky76> I´m asking because I have to change some additional code too...
20:59:47 <rito12_51026> Load save and leave to the desktop?
21:08:36 <LordAro> scrollbar on the menu screen!
21:10:25 <mmtunligit> or just yknow, ctrl+Q
21:10:29 <xarick> i got distracted with youtube again :/
21:10:47 <mmtunligit> or whatever the quit program keybind is on your system
21:10:55 <mmtunligit> or use the little x in the title bar
21:16:03 <Rubidium> mmtunligit: sounds like you haven't read the issue
21:21:54 <mmtunligit> i did initially but then i came back 15 minutes later and had forgotten some of the details because its not something i bothered to commit mental capacity to
21:22:27 <peter1138> You can quit with alt-f4.
21:24:23 <mmtunligit> small, unobstrusize edge case unless using zbase and the far end of the interface scale, not a crash, easily worked around
21:24:38 <rito12_51026> Allow to close main menu window, and perform division by zero if no window exist?
21:26:58 <xarick> how to refresh 2 million iterators fast
21:39:48 <Rubidium> have fun optimising the wrong thing
21:40:16 <_zephyris> The real issue is window size vs interface size, surely?
21:40:35 <andythenorth> why not 20 million iterators? 🙂
21:40:38 <_zephyris> So a "proper" fix would be to clamp max interface scale based on window size?
21:42:51 <xarick> but i can't make it shine
21:44:13 <andythenorth> why are `bitmask(TRIGGER_VEHICLE_UNLOAD_ALL, TRIGGER_VEHICLE_ANY_LOAD, TRIGGER_ALL_NEEDED)` needed to re-randomise bits when a vehicle is loaded?
21:44:27 <andythenorth> why not just `TRIGGER_VEHICLE_ANY_LOAD`xor `TRIGGER_VEHICLE_UNLOAD_ALL`?
21:46:00 <andythenorth> also bits 2 and 3 of vehicle random trigger are for the consist (parent scope)
21:46:12 <andythenorth> but NML docs have them as vehicle (self scope)
21:51:45 <_glx_> xarick: very easy, don't use 2M iterators
21:59:33 <xarick> i can get RemoveItemTests down to 334 seconds with value iter tracking
21:59:56 <xarick> this goes up to 369 seconds
22:01:31 <xarick> master is at 400 seconds
22:02:28 <peter1138> andythenorth, the text says "when the consist has been emptied and then receives new cargo"
22:02:36 <peter1138> That means it needs to have unloaded, and then loaded again.
22:02:53 <xarick> master without list swap is probably still 400 seconds
22:22:07 <andythenorth> triggers break my brain 🙂
22:23:00 <andythenorth> Horse 4 has a bug with bits being re-randomised when 100% load is reached on a vehicle
22:23:17 <andythenorth> I will test tomorrow with Horse 3 to see if it's a new bug
22:29:42 <andythenorth> hmm bug exists in Horse 3 also
22:29:48 <andythenorth> so either it was always there unnoticed
22:29:50 <andythenorth> or NML is broken
22:30:09 <andythenorth> or OpenTTD random triggers are broken (probably recently ish)
22:30:14 <andythenorth> let's go with Horse bug 😛
22:46:23 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
23:02:18 <xarick> i have two lists, i swap, what should happen to each list
23:03:16 <xarick> unconditionally swap values
23:03:38 <xarick> now whether to rebuild items or swap items
23:04:45 <xarick> rebuilditems will relink iterators
23:05:08 <xarick> iterators are toast if that happens
23:06:54 <xarick> if the values_inited is false in one of the lists, what must I do
23:08:34 <_glx_> not using iterators should help
23:09:06 <_glx_> it's always a pain to keep valid iterators
23:09:09 <xarick> if the items are not rebuilt and values_inited is false, that means i must forcefully create list->items, but, is swaping correct at that point
23:10:42 <xarick> i cannot swap if i just rebuilt for the other list
23:12:49 <xarick> i think the swap test should be first before the others
23:14:08 <xarick> if neither values are inited, swap is the only logical possible choice. okay, so I turned the problem around.
23:14:29 <xarick> but linking need to happen
23:14:38 <xarick> if any of the lists has list_values inited
23:17:04 <xarick> my goal was to avoid iterating ->items twice
23:17:47 <xarick> but apparently, i have no good brains to think this through
23:31:30 *** MinchinWeb[m] has quit IRC (Read error: Connection reset by peer)
23:31:47 *** MinchinWeb[m] has joined #openttd
23:45:38 <xarick> I can't do better than 369
23:46:24 <xarick> If i exclude the SwapList tests in RemoveItemTests, it goes to 334 seconds
continue to next day ⏵
