IRC logs for #openttd on OFTC at 2024-02-06
⏴ go to previous day
00:14:16 *** royills[m] has joined #openttd
00:17:16 <xarick> and now I'm off to bed. Tomorrow I expect pitchforks as you say it 😐
01:18:29 *** throwawayaccount[m] has joined #openttd
01:45:44 <belajalilija> Autosep finally coming to vanilla ?
01:56:18 *** Eddi|zuHause2 is now known as Eddi|zuHause
02:17:25 *** freu[m] has joined #openttd
02:17:35 *** enick_974 has joined #openttd
02:31:00 *** Wormnest has quit IRC (Quit: Leaving)
02:37:45 *** VincentKadar[m]12345 has joined #openttd
03:10:02 *** Flygon has quit IRC (Read error: Connection reset by peer)
03:16:26 *** temeo[m] has joined #openttd
03:21:15 *** debdog has quit IRC (Ping timeout: 480 seconds)
03:22:36 *** D-HUND is now known as debdog
03:27:40 *** gnu_jj_ has quit IRC (Ping timeout: 480 seconds)
04:06:06 <talltyler> It’s a depot action, and all the separation delay takes place in the depot instead of out in the game world where it can cause logjams.
05:06:18 *** rain[m]1 has joined #openttd
06:02:25 *** kaomoneus has joined #openttd
06:02:25 <kaomoneus> Hi all! Did anyone experienced color issues on macos? I'm trying to rebuild 13.4 for macos and it seems that there is something wrong with pixel format.
06:24:22 <kamnet> jfs: Yes! I was trying to find the PR for that but I couldn't locate it. Help, please?
06:41:45 *** keikoz has quit IRC (Ping timeout: 480 seconds)
06:55:49 <truebrain> kaomoneus: Update to 14.0-beta2 and the blue wash will be gone
06:56:33 <kaomoneus> truebrain: , thanks! Is here any particular commit fixing that?
06:57:09 <truebrain> Just look on Github if you want to know 🙂
06:57:20 <truebrain> Or in the changelog
06:58:55 <kaomoneus> Any keywords? Smth like "blue screen"?
07:08:12 <emperorjake> kaomoneus: Did you try turning hardware acceleration on or off?
07:10:30 <kaomoneus> emperorjake: , nope, just tried it out of box. How to do it? Is it some build definition, or in-game?
07:10:57 <emperorjake> It's in the game options, under graphics
07:26:01 <kaomoneus> emperorjake: , yep, it did work. Thanks!
07:26:25 <kaomoneus> So that was a cocoa driver issue then
07:50:02 <LordAro> bit of a shame from twpol there
08:32:37 *** thelounge345 has quit IRC (Quit: Ping timeout (120 seconds))
08:32:43 *** thelounge345 has joined #openttd
08:41:00 *** m1cr0m4n has quit IRC (Quit: G'luck)
08:41:17 *** m1cr0man has joined #openttd
09:06:48 *** thelounge345 has quit IRC (Quit: Ping timeout (120 seconds))
09:07:04 *** thelounge345 has joined #openttd
09:20:03 <peter1138[d]> Dunno what the story is but a 4 year old PR is a bit strange to deal with.
09:55:07 <twpol> It was fully rebased on 17th Dec when I resumed work on it, so I don't think the age should matter
10:21:01 <truebrain> hmm .. working with GitLab CI syntax, you see that some parts of GitLab are better, but many parts of GitHub are 😛
10:21:09 <truebrain> boy, this is annoying to setup 😄
10:49:26 <xarick> procedural trees gonna solve savegame sizes?
10:50:24 <peter1138[d]> The potential savings are pretty small for a normal game.
10:50:56 <xarick> oh 😦 I was hoping for 4kx4k viability in multiplayer
10:51:09 <locosage> xarick: maybe one day but not yet
10:51:14 <peter1138[d]> 5% reduction on 4k map.
10:51:18 <locosage> savings are pretty good if growth is not allowed
10:52:26 <peter1138[d]> Get faster connections if you want to play a massive map :p
10:52:50 <locosage> 4k used to timeout even on a localhost xD
10:53:18 <peter1138[d]> You need to improve your localhost connection, clearly.
10:53:47 <locosage> connection won't help compression speed :p
10:59:29 <locosage> Btw, it ridiculous how much entropy trees add. After #11955 without spreading the only thing stored on the map is the fact of tree existance. And that still eats like 3/4 of the save size.
10:59:52 <Eddi|zuHause> trees are only significant on an empty map. once you get started with a big network you get swamped in cargopackets.
11:00:33 <xarick> limitation is lzma being slow
11:00:44 <xarick> it should be multithreaded
11:01:14 <locosage> I don't remember exact numbers but I checked big citymania games and all the player infrastructure still takes way less save size than trees
11:03:05 <truebrain> hmmm ... trees .....
11:05:20 <xarick> tree growth speed in openttd is not realistic, but I'm no tree expert. They grow too fast compared to real life
11:09:12 <Eddi|zuHause> if that's your only issue...
11:18:49 <peter1138[d]> The game must never have any randomness.
11:18:56 <peter1138[d]> That's my take-away.
11:19:58 <peter1138[d]> The only correct gameplay is to maximise profits, even though money doesn't matter.
11:28:44 <locosage> game can have randomness, but there is no good reason to waste resources on storing random numbers :p
11:29:07 <locosage> peter1138[d]: and I don't even know where you got this from
11:29:23 <merni> I think that is sarcasm
11:33:28 <merni> twpol: Yeah... jgrpp-style auto-separation (without needing depots) is a feature I would be very excited to see in vanilla. I am also not sure why talltyler says it can no longer be done thanks to the integration of autosep with depots... if some logic is developed that allows (in most cases) deadlock-free autoseparation in the game world, why can't autosep in depots be made a subset of that,
11:33:28 <merni> replacing the existing implementation?
11:35:37 <locosage> merni: nonsense doesn't make a good sarcasm :p
11:36:43 <reldred> I dunno, I've no problems with how he went about implementing unbunching, but for me it'll really start cooking once mixed with JGRPP style drive through depots, or ultimately, PR8480
11:37:31 <peter1138[d]> twpol's PR completely passed me by, so I don't know what was up with it.
11:37:34 <reldred> Or do unbunching at a station instead of a depot, since we have nice .grf's already for yards.
11:37:55 <merni> I think unbunching at station is what twpol's PR was about, right?
11:38:14 <merni> Not at a fixed station but by waiting at stations in general
11:38:24 <_jgr_> merni: They're not fundamentally incompatible, they just couldn't be both active at the same time.
11:38:35 <_jgr_> The real problem is a UI one
11:39:30 <merni> I would think solving that would mean an overall "autosep" button (as in jgrpp), removing the autosep-at-depot option, and converting all trains that have autosep at depot active in old saves to overall autosep
11:40:45 <_jgr_> The auto-separation in my branch works in such a different way that it doesn't make sense to auto-convert on load like that
11:41:06 <merni> I just meant that the button would be as in your branch, not the functionality
11:42:08 <locosage> I haven't tried auto-separating much but I see it as something that should be just on by default
11:42:42 <merni> and even otherwise, I think the main appeal of auto-separation is that you don't have to think about headways or timetables any more. It's a "set it and leave it" option for those who don't particularly care about the way it works or whether it is absolutely perfect.
11:43:02 <merni> locosage: I think it is in jgrpp or at least there is a setting to make it on by default
11:43:19 <reldred> Tying it to a depot order also has the side effect of ensuring players set a depot order, which will avoid a LOT of complaints about lost vehicle.s
11:43:41 <merni> only if people play with breakdowns/servicing on
11:43:43 <reldred> Which for a vanilla OpenTTD feature is probably A Good Thing (tm)
11:43:55 <locosage> even with breakdowns depot orders have their issues
11:44:11 <merni> without breakdown, a depot order is superfluous
11:44:35 <merni> personally the only time my trains visit a depot is when I want to change the composition of the train or sell it
11:44:41 <reldred> Most vanilla players I'd wager are running default daisies.
11:44:50 <reldred> The rest of us weirdos already know better one way or another.
11:44:54 <merni> well, let's check survey :p
11:44:56 <_jgr_> There is still autorenew and so on
11:45:35 <reldred> merni: which is only catching jgrpp and nightly players atm 😛
11:45:49 <_jgr_> The betas will be there soon enough
11:46:33 <locosage> beta it's still not the majority of players
11:46:34 <merni> well, I don't know what `1` means here
11:46:34 <xarick> no idea what auto separation does, just hope it doesn't break AIs
11:46:53 <reldred> remove it from the game
11:46:57 <reldred> last minute 14.0 feature
11:47:19 <Eddi|zuHause> we did that once already.
11:47:27 <xarick> I saw orders being changed, not sure if AI API was taken care of
11:47:34 <LordAro> it made people very angry
11:48:09 <reldred> actually, I do use citylife ai at least
11:48:23 <reldred> gotta have my eyecandy vehicles shittifying my road networks
11:48:45 <xarick> clicked orders list button
11:49:05 <merni> ok so 66% of people have breakdowns on `1` and 30% have it on `0`, whatever those mean. Among jgrpp players that is pretty much exactly reversed
11:49:53 <reldred> come on merni, what do you think 1 and 0 could mean
11:49:54 <Eddi|zuHause> well, odds are that people using experimental builds tend also to experimental settings
11:50:02 <merni> and some people have it on `64` in jgrpp...?
11:50:09 <merni> and whatever does `2` mean
11:50:30 <merni> reldred: I am guessing 0 = no breakdown, 1 = reduced, 2 = original
11:50:45 <reldred> improved in jgrpp probably
11:50:49 <xarick> 64 is probably breakdown every tick
11:51:02 <LordAro> there is a move to stringify the settings in the config file
11:51:10 <LordAro> you could move that on a bit further :)
11:51:15 <merni> might also be "very reduced"
11:51:28 <merni> since improved is not an option in the dropdown but a separate seting
11:51:51 <merni> LordAro: where can I find this move?
11:52:04 <LordAro> no, it's just all the settings added recently
11:52:24 <merni> I can certainly look at it when I have time
11:52:24 <LordAro> i might be making it up
11:52:53 <xarick> did you fix order flags or something for AIs?
11:53:15 <merni> does it also crash in nightly?
11:54:18 <merni> I don't see how that commit is relevant
11:54:40 <xarick> it's the version I'm on
11:54:58 <merni> that is the latest existing, so you should file an issue
11:55:52 <truebrain> LordAro: You are not making it up 🙂 enums and strings for settings when it is a dropdown is what we have been doing .. if the author/reviewer didn't forget 😛
11:56:36 <truebrain> Readability .. such a lovely concept 🙂
11:56:47 <merni> there is a bug with the survey webpage, if one version has not enough data, the next one gets an extra indentation level
11:57:19 <merni> Yes, let me dust off my HTML knowledge such as it is
11:57:56 <merni> where are the templates?
11:58:03 <merni> sorry, I never worked with this framework before
11:58:13 <_jgr_> merni: 64 is very reduced
11:58:21 <merni> or really any beyond "write html and hope" :p
11:58:35 <truebrain> Search for it; you will find it 🙂
12:04:06 <merni> also, does anyone else think it would be better if the first-level organisation on that page was by settings, instead of by versions?
12:04:44 <merni> the page is really long because each version duplicates the entire huge list of settings, and even when using Ctrl+F it is not so easy to know which version a result pertains to
12:05:38 <merni> instead, listing the settings only once and having columns on the table for different versions (only the ones which have enough data) feels like a better option to me
12:10:14 <LordAro> just as well, or we'd have needed beta4
12:10:27 <LordAro> merni: seems reasonable
12:10:32 <LordAro> but i think you know the answer :p
12:11:51 <merni> Yes, "PR it", but I don't want to PR something that would be against the intentions or otherwise likely to be rejected
12:11:59 <merni> That is different from a simple bugfix
12:12:24 <locosage> that may be a bit too many columns if each beta gets one
12:12:55 <merni> well, only as many as people are using
12:14:17 <merni> xarick: does the crash happen only with that AI, or with any AI, or even without an AI?
12:14:48 <xarick> let me test myself building a train
12:16:01 <xarick> no crash when I do it myself
12:16:15 <peter1138[d]> It's an obvious fix 🙂
12:18:10 <xarick> crashed with my AI, this time it was a bus
12:18:29 <_jgr_> It's because it's using the other company view
12:18:33 <truebrain> merni: no, that would be bad. I did consider moving every version to their own page; that might be a bit better suited. But cross-version settings have little meaning, honestly
12:19:08 <merni> truebrain: Yes, each version on its own page might be better
12:19:34 <merni> peter1138[d]: Evidently what is obvious to you isn't to me :P
12:20:10 <_jgr_> This particular failure mode has happened before
12:20:31 <xarick> so if i move to the competitors company, I wont get a crash?
12:20:57 <_zephyris> truebrain: Font blog?
12:21:05 <truebrain> So next to unittests, we need e2e tests? 😄
12:21:21 <peter1138[d]> User Acceptance Testing Tests
12:21:22 <truebrain> _zephyris: If you are up for it, go for it!
12:23:06 <xarick> confirmed, no crash if I move to competitors company before clicking show orders
12:28:08 <peter1138[d]> merni: Based on error and the line it crashed on, it was either vehicle is invalid (unlikely) or widget is invalid (likely)
12:43:52 <_glx_> We are kinda used to this type of errors
13:40:20 <xarick> ``` if (NWidgetCore *nwid = this->GetWidget<NWidgetCore>(WID_O_DEPOT_ACTION); nwid != nullptr) {
13:40:20 <xarick> nwid->tool_tip = STR_ORDER_TRAIN_DEPOT_ACTION_TOOLTIP + v->type;
13:43:15 <LordAro> it's a new C++ feature
13:43:15 <peter1138[d]> xarick: It was accepted, so I guess so 🙂
13:44:06 <peter1138[d]> It reduces the scope of nwid so that it can't accidentally be used or shadow elsewhere.
13:59:18 <talltyler> I was wondering about that
13:59:29 <talltyler> Thanks for fixing my bug ❤️
14:08:15 <_zephyris> truebrain: Website PR?
14:08:37 <truebrain> Why are you asking me?
14:08:44 <truebrain> owh, in relation to that
14:08:48 <truebrain> lol, sorry, context is everything 😄
14:08:53 <truebrain> and yes, much like any other post 🙂
14:09:28 <truebrain> haha, I forgot the big list of tickets in 10709 😄 That was funny 🙂
14:11:12 <truebrain> the gift that keeps on gicing, or something 🙂
14:20:47 <truebrain> _zephyris: what most likely is best, if the other two also do their blogs they said they would, is to have one every other weekend or so till the 14.0 release. Just because we can 🙂
14:21:58 <twpol> merni: Yes, I thought many people were excited to see auto-sep go in too. My implementation seemed decently deadlock-free (only one, unreproducible, from an earlier version) but obviously didn't get beta-levels of exposure.
14:26:15 <_zephyris> Aiming for 01/04 release presumably... So 24/02 for fonts, 09/03 for ship pathfinder (kuhnovic) & 23/09 for time changes (talltyler) ?
14:27:26 <truebrain> I would it up one week if we can, but yeah, something like that. The sooner we get the blogs the better ofc, but blog-being-finished doesn't have to mean blog-published 😄
14:28:13 <truebrain> there are some other subjects we could write about (UI, Survey, Social Plugin, new player experience (settings, help-window, ..)) .. but not sure if anyone else likes writing 🙂
14:44:36 <_zephyris> truebrain: Too much news!
14:45:09 <LordAro> it'll be like us having dev blogs again
14:48:56 <_zephyris> Compete with factorio facts friday
14:49:57 <talltyler> I assume you don’t mean for my blog to be in September 😛
14:50:21 <truebrain> only Americans .... only Americans ....
14:50:41 <talltyler> 23/09 is September, no?
14:51:10 <talltyler> American would be “what month is 23 you crazy Europeans”
14:51:17 <truebrain> hahahaha, that too 😛
14:51:29 <truebrain> I was more poking fun of the fact that common sense left the building 😛
14:51:57 <_glx_> usually it we put month before day, then we always have year before month 🙂
14:52:07 <talltyler> Anyway, I don’t want to steal the spotlight or jump the queue but I want to get the time post published reasonably soon, because people are already asking about how it works
14:52:23 <truebrain> and we figure it out after
14:52:39 <talltyler> Such a logical solution 😄
14:52:51 <truebrain> for PR reasons, doing your last is better, but I can also understand that doing it first helps with the questions / testing
14:53:06 <talltyler> “Save the big features for last”?
14:53:30 <talltyler> Hmm, is that still true with an open beta that anyone can play?
14:53:48 <truebrain> BaNaNaS shows only 12% activity on the betas 🙂
14:53:49 <LordAro> "reduce the amount of time people have to complain about it"
14:53:49 <talltyler> I.e., how many people wait until April 1st to actually play 14.0?
14:54:04 <talltyler> truebrain: Ah, there’s the answer 🙂
14:54:08 <truebrain> on Steam that number is even lower
14:54:16 <truebrain> anyway, I am fine with what-ever order
14:54:25 <truebrain> but spreading them out might be good, as otherwise it is much at once
14:54:52 <talltyler> Yes, that’s probably smart
14:55:15 <truebrain> so we can also play the game: whoever comes first! 😛
14:55:31 <talltyler> I wonder when our one YouTube influencer (Master Hellish) will do a video about the update 😛
15:02:37 <_glx_> hmm scripts might need to learn about unbunch flag
15:05:03 <xarick> that's what I've been trying to say
15:05:19 <peter1138[d]> No harm if they don't know about it, they just can't use it.
15:05:30 <LordAro> it's not going to break existing stuff
15:06:12 <xarick> ScriptAirport need more functions
15:06:40 <xarick> scripts also can't change service interval
15:07:21 <_glx_> timetables are hard enough for humans 🙂
15:07:23 <peter1138[d]> Could be a good PR 😉
15:07:25 <xarick> there was one recent addition to vehicle groups sharing orders
15:07:40 <xarick> i stumbled upon it but it's missing for ais
15:09:30 <xarick> and the station catchment area functions are... outdated since the bitmap iterator? or something
15:10:56 <peter1138[d]> Be a bit more specific 🙂
15:11:08 <_glx_> script order flags are full of bit magic
15:13:20 <xarick> station coverage area computations, hmm how do I say it, when spreading, it used to cover the entire region inbetween?
15:14:06 <_glx_> ha you mean sparse coverage
15:15:17 <_glx_> as long as your station doesn't mix types, and it's rectangular, it should work the same
15:16:43 <peter1138[d]> Oof, that change was years ago 🙂
15:19:27 <_glx_> I think the "problematic" one is ScriptStation::GetStationCoverageRadius(StationID station_id)
15:20:24 <_glx_> it returns the biggest radius for the station
15:21:08 <peter1138[d]> Hmm, not sure how that's used.
15:21:57 <xarick> AAAHogEx doesn't build anything with infinite money
15:22:07 <xarick> bad feature confirmed 🙂 j/k
15:22:45 <_glx_> there's no reason for it to fail, unless it does weird checks
15:23:15 <xarick> let me check other AIs
15:24:30 <_glx_> funny how ScriptStation::GetCoverageRadius(ScriptStation::StationType station_type) returns -1 for airports
15:25:09 <xarick> should redirect to airport coverage function
15:25:55 <_glx_> no because it would need an extra parameter
15:26:22 <_glx_> that's the before build radius
15:30:49 <xarick> some AIs don't build anything
15:30:50 <_glx_> I guess a TileList for Station::catchment_tiles could work
15:31:02 <xarick> with infinite money setting
15:31:17 <xarick> there's a few that still build
15:32:15 <xarick> one of the AIs is spamming exclusive transport rights, maybe that's why?
15:32:21 <peter1138[d]> I guess they check if they have enough money. So...
15:32:24 <truebrain> _zephyris: I am not spying or anything, but how do you fork a repository twice? 😄
15:32:33 <peter1138[d]> But infinite money shouldn't affect that.
15:33:15 <truebrain> ugh, I need to go to the store, but I don't want to! Can one of you go please?
15:33:47 <_zephyris> truebrain: Never underestimate the power of a noob
15:37:55 <_zephyris> (I forked it, got confused, deleted the for, reforked it, remembered what to do, pr'd)
15:38:53 <xarick> Terron doesn't build, this AI was almost flawless
15:39:19 <xarick> all he do is exclusive transport rights with infinite money
15:39:23 <xarick> but not build any route
15:40:16 <peter1138[d]> Does it behave the same if you turn off infinite money?
15:40:17 <xarick> I wonder how skewed are the AI decisions with this
15:40:38 <xarick> yes, he used to build before
15:41:10 <peter1138[d]> "before" suggests you mean before infinite money was implemented, rather than with infinite money turned off.
15:41:26 <xarick> ok let me try with it off
15:41:40 <peter1138[d]> But I suppose the issue is it does some things without checking money, which sends it into the red, and then it checks for other things, and finds it doesn't have enough cash.
15:42:55 <peter1138[d]> Solution for AIs might be to fake `GetBankBalance()` if infinite money is enabled.
15:43:17 <peter1138[d]> `Gets the bank balance. In other words, the amount of money the given company can spent.`
15:43:30 <peter1138[d]> By the latter definition, returning max seems okay 🙂
15:43:50 <peter1138[d]> That's what it does. Hmm.
15:44:27 <xarick> problem stems from... "what to do once I get x money, or after I get x cash flow every y months"... infinite money ruins these decisions entirely, they misbehave
15:45:49 <peter1138[d]> So it's doing it based on income instead of bank balance?
15:46:28 <xarick> terron builds without infinite money, just checked
15:47:02 <xarick> I wonder what mine does
15:50:44 <_glx_> I always forget I must run "regression_files" target to test my modified main.nut
15:51:04 <peter1138[d]> Never forget to test your nuts.
15:53:24 <peter1138[d]> So Terron builds for a bit, and then gets stuck in buying exclusive rights
15:58:06 <peter1138[d]> I don't actually know how you debug scripts.
15:59:32 <xarick> didn't build anything when i tried 😐
15:59:47 <LordAro> peter1138[d]: print statements
16:02:14 <_jgr_> There is a breakpoint function in the script debug window, I've yet to make it do anything useful though
16:08:36 <peter1138[d]> Not really a breakpoint is it? 😄
16:18:25 <peter1138[d]> What does Pterodactyl turn that line into?
16:18:28 <xarick> strange, "LuDiAI AfterFix" should fail to be detected in that case, too, but it's not
16:19:36 <peter1138[d]> Does it handle `Renewed Village Growth = ` if it didn't have quotes, or still change it to `Renewed = `?
16:20:48 <peter1138[d]> So the issue isn't the quotes, it's the spaces.
16:23:31 <peter1138[d]> (Not related to this, just found it amusing)
16:27:09 <truebrain> So what if we rename it to .dat, would that solve the problem? 🙂
16:30:43 <xarick> why do the font looks so good on your screenshots and so bad on me?
16:32:05 <truebrain> (I always find it a bit odd to make a generic platform and be pedantic about data formats .. always feels a bit like that is defeating the purpose .. so then I wonder, if we don't call it .ini, does that solve the problem all of a sudden?)
16:32:22 <peter1138[d]> It loos fine at 2x or above.
16:32:32 <peter1138[d]> 1x or 1.25x is a bit smudgy.
16:32:43 <peter1138[d]> Well, 1x is fine.
16:33:43 <peter1138[d]> But I don't know much about go to be able to say.
16:34:00 <truebrain> Can't even find a rfc-like spec for ini
16:34:15 <peter1138[d]> There isn't one, it's always been ad-hoc.
16:35:14 <truebrain> So they could be more truthful and say: the go module we use doesn't understand OpenTTD's ini format
16:36:35 <truebrain> Instead of blaming us not following a standard 😛 we don't follow the specs as implemented by the go module, which is a bit of a different nuance 🙂
16:37:41 <truebrain> Not that any of this is helping kat08826 in the slightest 😛
16:37:57 <LordAro> lack of standardisation is why toml became a thing
16:38:09 <truebrain> On principle I am now closing the gate and manning the walls
16:38:18 <LordAro> why do we suddenly care about this game server portal thing?
16:38:55 <truebrain> And they go: why do we now care about this "small" game someone wants to use with our generic software? 😄
16:38:58 <peter1138[d]> go-ini's test cases do have keys with spaces in them.
16:39:35 <LordAro> i don't mean it like that, i mean how did it become a topic of discussion
16:39:35 <truebrain> Sorry, I am in a sarcastic mood 🙂 seems peter is being more productive 😛
16:39:37 <peter1138[d]> But an ini file parser should be able to treat quotes as part of the key, I think.
16:39:53 *** Wormnest has joined #openttd
16:39:57 <LordAro> it's not been linked to us from anywhere, it was just mentioned by peter1138[d] out of the blue, as far as i can tell
16:39:58 <truebrain> LordAro: kat08826 asked about it 🙂
16:40:04 <jfs> truebrain: INI files are whatever you got when you used the API in 16 bit Windows to write them
16:40:14 <xarick> value with quotes = "some value"
16:40:25 <peter1138[d]> Probably too long for IRC 🙂
16:40:50 <LordAro> ah, that message has not appeared at all on this side
16:41:06 <truebrain> Funny, kat08826 is not relayed at all
16:41:43 <peter1138[d]> Actually go-ini's unit tests do include keys with quotes.
16:42:00 <peter1138[d]> Or at least the testdata does.
16:42:55 <truebrain> So you are now debugging their software? 😄 curious where this leads
16:43:56 <truebrain> Why do we store gamescripts by name? That is a bit odd to me .. why not by their uniqueid?
16:44:15 <peter1138[d]> What's a unique ID?
16:44:32 <truebrain> 4 letters, like with all our content 🙂
16:44:57 <peter1138[d]> Where does that come from?
16:45:08 <peter1138[d]> function GetShortName() { return "TERR"; }
16:45:25 <xarick> "1 + 1 = 2" = does not compute
16:45:37 <xarick> no spaces between quotes for the key?
16:46:02 <truebrain> Names don't have to be unique on BaNaNaS, but those IDs do
16:46:08 <peter1138[d]> xarick: are you somehow testing go-ini?
16:46:10 <truebrain> So bit odd that our config doesn't use it
16:46:27 <truebrain> As it could resolve to more than one GS
16:47:37 <truebrain> Okay, maybe just not open that can of works 😄
16:50:01 <truebrain> So BaNaNaS does no validation on name, and sees the name as a display-only, but the client does all kinds of things with the name 😄 hihi, seems there is a disconnect between those two worlds 🙂
16:51:45 <truebrain> Pfff, my own handy work .. I should have known better 😛
16:56:43 <_glx_> oh nice regression fails locally for me (my guess is the duplicate script errors removal)
17:02:59 <locosage> "backup floppy disks" doesn't sound very reassuring
17:03:47 <peter1138[d]> It's fairly common for old software development to be stuck rotting on old floppy disks.
17:07:01 <peter1138[d]> I think Terron would have an issue with large starting money too.
17:08:46 <merni> locosage: Well, what would you expect from someone who corresponds by letter
17:17:20 <xarick> my ai doesn't build with "infinite money"
17:17:41 <xarick> but if I give him money via GS, it builds
17:17:53 <xarick> something else's going on
17:19:13 <merni> kat08826: well... PRs are welcome but... there is no real spec for INI so not sure what we should be fixing it to
17:19:38 <truebrain> kat08826: Our INI reader is rather strict. So the only solution that comes to mind, next to that platform supporting OpenTTD's ini, is manually renaming the GS name
17:22:02 <truebrain> Yeah, but #effort on a system that works for 10+ years already. So realistic ... 🙂
17:22:42 <merni> xarick: could you perhaps share the AI (if it's not on banananananas) / make a bug report?
17:23:15 <merni> because the simple AI that you gave me earlier that tried to buy 500 buses or whatever worked
17:23:19 <truebrain> How is the survey-web PR going @merni? 😄
17:23:29 <xarick> there's a older version on bananas without train support
17:23:33 <merni> Eh will do it the weekend
17:23:38 <xarick> should behave the same
17:26:12 <truebrain> I removed that empty line with the regex
17:26:32 <truebrain> Everything updated etc?
17:26:53 <truebrain> As it also passes CI
17:27:58 <_glx_> well I think it does on windows runners
17:28:14 <truebrain> But that should influence all lines I guess 🙂
17:29:14 <truebrain> But yeah, really odd. Will test later tonight what my MSVC does 🙂
17:32:20 <xarick> is something wrong here?
17:32:48 <peter1138[d]> It's probably overflowing.
17:33:31 <xarick> gonna add some debug messages here to see what money it comes up with
17:34:37 <_glx_> no need to use debug message if `bank_balance` is INT64_MAX it will overflow
17:36:09 <_glx_> it should check `bank_balance` only before checking with loan
17:36:53 <merni> I wonder though if we could provide a function that checks "do I have/can I get this amount of money" directly via script api
17:37:17 <_jgr_> It may be more pragmatic to just return a more reasonable big number
17:37:22 <peter1138[d]> Terron does somethning similar, it gets the current balance and then adds on its predicted (or something) income.
17:38:24 <_glx_> INT32_MAX should be more than enough
17:38:29 <xarick> that's why it doesn't build 🙂
17:38:45 <truebrain> No `Money` for scripts? 😄
17:38:56 <xarick> I have negative 9 whatever illion money
17:39:21 <peter1138[d]> "Not enough funds to start action..."
17:40:41 <merni> Hm, the docs in `script_company.hpp` should also probaby be updated...
17:40:45 <truebrain> _glx_: regression passes for me on MSVC
17:41:01 <merni> Ok, this PR I can do quickly :)
17:41:05 <peter1138[d]> Heh this AI just spend £63m on something.
17:42:36 <peter1138[d]> Terron spends more freely with a lower fake money.
17:42:53 <peter1138[d]> Think it still gets itself unstuck.
17:42:57 <merni> well, if I change to INT32_MAX ie 2^31 - 1, then AI won't be able to spend more than 2 * 10^9 moneys
17:45:49 <merni> outch this will mean a weird special case in the script api
17:46:18 <truebrain> change SQInteger into OverflowSafeInt 😛
17:46:29 <LordAro> OverflowSafeSQInteger
17:46:37 <truebrain> can you imagine the slowdown? 😄
17:47:06 <LordAro> can't be any worse than the terrible scripting languages we have at work
17:47:14 <merni> just slightly unsatisfying code
17:47:23 <truebrain> Javascript can be scary quick, I found out
17:47:51 <LordAro> we don't use V8 though
17:48:25 <LordAro> javascript itself isn't particularly slow, it's just terrible
17:48:34 <LordAro> it's the custom languages we have that are both slow and terrible
17:49:06 <_jgr_> In the big scheme of things, JavaScript is not that bad
17:50:13 <LordAro> let me add an additional qualifier: our javascript
17:50:31 <truebrain> yeah, now you did it! 😛
17:50:49 <truebrain> but, think about it like this ... 15 years ago you could also have picked Squirrel because it looked "good enough" 😛
17:51:24 <LordAro> no, instead we have javascript, python, lua, *and* #custom scripting languages
17:51:33 <truebrain> for a while we also worked on our own language, called bolts .. it parsed .nail files ... Not Another Interpreted Language 😛
17:51:35 <LordAro> all more or less able to interop with each other
17:51:45 <truebrain> or called NAILs and parsed .bolts files?
17:51:56 <truebrain> it was actually functional etc etc
17:52:10 <truebrain> but in the end we went with Squirrel, because .. "not written by us, must be better maintained"
17:53:08 <LordAro> writing some sort of basic squirrel parser in lua seems like the best "solution", imo
17:53:27 <LordAro> not that i am volunteering for such a thing
17:53:34 <_jgr_> That sounds like turning one problem into two problems?
17:53:48 <truebrain> I did look into writing a Squirrel parser in WASM
17:53:57 <truebrain> if WASM would be a tiny bit more mature, that isn't impossible
17:54:05 <LordAro> well the complaint is always that "squirrel(2) is a dead language that no one knows"
17:54:18 <LordAro> and lua is well known for being a very fast very embeddable language
17:55:02 <FLHerne> Even with the current state of affairs, WASM looks much better than Lua to me
17:55:06 <truebrain> just lua has no concept of OO, which makes it challenging for us 🙂
17:55:33 <xarick> support both languages
17:56:04 <LordAro> WASM seems like a massive dependency to add, imo
17:56:15 <xarick> with a big notice: "we prefer you use the new language, we are depreciating squirrel in x years"
17:56:32 <truebrain> it is not that bad; depending on which you pick 🙂
17:56:52 *** HerzogDeXtEr has joined #openttd
17:57:05 <FLHerne> and the nice thing about it being a proper standard is that "which you pick" isn't binding
17:57:26 <truebrain> but currently it is lacking good support for other languages
17:57:31 <truebrain> only Rust and Go actually build to WASM properly
17:57:38 <truebrain> it needs a bit more to be practical
17:57:40 <FLHerne> very unlikely to end up in a Squirrel (or Lua) situation where you're stuck supporting some ancient dead version to preserve compatibility
17:57:41 <_glx_> or we could redo gpmi 😉
17:58:03 <truebrain> don't get me wrong, GPMI was nice, but a security nightmare 😄
17:58:40 <_jgr_> Squirrel (and lua) are simple enough that you could support it indefinitely if you wanted to
17:59:24 <FLHerne> Lua has already had exactly the same type of flag-day compat breaks as Squirrel
18:00:39 <FLHerne> supporting it doesn't (or shouldn't) just mean keeping the old interpreter: it includes documentation and libraries and so on
18:01:00 <FLHerne> a problem solved for OTTD Squirrel by ignoring it :p
18:01:36 <truebrain> too lazy to test #177 locally, so let's see how it works out on preview 🙂
18:01:41 <_jgr_> If the language doesn't change, the documentation and libraries don't need to either
18:01:52 <FLHerne> IMO switching to Lua would just be repeating the identical mistake -- in 5-10 years you're stuck with what's functionally a custom language because no-one else is using that ancient version
18:01:58 <_jgr_> In some sense something that is dead and frozen forever is easier to deal with
18:02:25 <andythenorth> javascript [cough]
18:02:38 <andythenorth> wait, is this the game where everyone names favourites
18:02:41 <andythenorth> then we start trolling
18:02:48 <andythenorth> then Truebrain just picks something anyway
18:03:11 <truebrain> mostly what makes me interested in WASM is not scripts; but things like terrain generation
18:03:24 <truebrain> I have some early drafts that actually work, but ... yeah .. frontend language support is annoying
18:03:27 <andythenorth> if it's WASM, I just write in whatever language I want and code generate or transpile?
18:06:32 <truebrain> if with "whatever" you mean Rust or Go, yes
18:07:39 <xarick> Is GetBankBalance the only tainted function?
18:08:31 <xarick> I need to find all instances of GetBankBalance to deal with infinite money properly
18:10:34 <truebrain> it is just not all what it seems
18:10:52 <truebrain> for example the Python one, they all are like 30MB WASM files
18:10:55 <truebrain> as it contains Python executor
18:11:19 <truebrain> which is not actually useful 😛
18:11:31 <truebrain> for TypeScript there is AssemblyScript, which is unmaintained
18:11:47 <_jgr_> Is that going to stop script authors from doing it anyway, if they like Python?
18:11:59 <truebrain> it requires WASI support
18:12:04 <truebrain> which basically means "access to everything"
18:12:11 <truebrain> so such scripts would fail to load
18:12:26 <truebrain> WASI is the gateway between WASM and the rest, like file support, network support, etc
18:12:32 <truebrain> WASM normally is fully isolated
18:12:41 <_glx_> yeah keep the scripts inside a sandbox
18:12:56 <truebrain> and that is a bit the issue with the current state of WASM for projects like ours
18:13:13 <truebrain> there is a lot of focus on WASI (rightfully, btw), so you can make applications in any language, and distribute them via WASM (run them in your browser, etc)
18:13:59 <_glx_> well I'm not sure it's a good idea to allow stuff running in the browse to access everything
18:14:34 <truebrain> _glx_: the browsers facilite that again to be safe
18:14:38 <truebrain> like the FS is sandboxed again
18:14:42 <truebrain> (like with OpenTTD via emscripten)
18:18:22 <truebrain> `Luwa's end goal is to JIT to WASM. Right now it's a bit of a learning environment for me as I've never written a language implementation that required real parsing` .. why is a project like that in that list, lol
18:19:23 <_glx_> xarick: but that will only fix one script
18:20:46 <merni> well, I compiled locally and tested with that AI, and it seemed to
18:21:55 <xarick> maybe return the highest value between INT32_MAX and what the company actually has
18:22:32 <merni> why? not like that matters in infinite money mode
18:22:51 <xarick> company could be richer than INT32_MAX
18:23:20 <merni> technically speaking, if the company *actually* had something close to INT64_MAX, it could overflow even without infinite money :p
18:23:38 <_glx_> I can't see any expense being above INT32_MAX
18:23:55 <_jgr_> merni: 2^64 is a *lot* of money
18:24:13 <merni> hence "technically speaking" lol
18:25:10 <xarick> but inflation exists! it inflates everything
18:26:02 <LordAro> even with zimbabwean inflation no one ever got close to exceeding 2**64
18:26:36 <_jgr_> All scripts have to work with if SQInteger, which doesn't offer anything useful at all to deal with integer overflow
18:26:43 <truebrain> LordAro: someone, not naming names, James, did make a savegame that did get to 2**63-1 by FF 🙂
18:26:52 <truebrain> but exceeding 2**64 is indeed impossible (Given it is signed 😛 )
18:27:04 <merni> > Therefore, the fourth dollar (ZWL) is equivalent to 10,000,000,000,000,000,000,000,000, or 1×1025 or 10 septillion first dollars (ZWD) (or 1 trillion third dollars).
18:27:23 <merni> Only revaluation prevented the zimbabwean dollar from exceeding INT64_MAX
18:27:40 <truebrain> nature also has a problem with exceeding 2**63
18:27:44 <truebrain> as we are a simulation after all
18:28:35 <merni> ok, but now that I have changed it to 2^31 for the script API, is that reasonable?
18:28:55 <truebrain> it is wrong no matter how you slice it 🙂
18:29:10 <truebrain> `2**31`, `2**56`, ... it is all wrong! So it is fine 🙂
18:29:21 <truebrain> if you can't make it right, just pick any wrong one that feels right 🙂
18:29:26 <_jgr_> Wrigns are not all equal
18:29:41 <andythenorth> /me considering templating rust
18:29:51 <_glx_> it's a reasonable high enough value when checking available money
18:29:53 <truebrain> _jgr_: I liked the misspelled variant more 🙂
18:30:06 <truebrain> just remember the price of long long tunnels!
18:30:11 <_jgr_> I'd like to say that that was wit, but I just can't type very well
18:30:58 <_glx_> ah yes tunnels across the map, or bombing extremely wide water areas
18:32:03 <bigyihsuan> andythenorth: i thought rust templating was generics and macros? 😛
18:33:42 <xarick> This looks very weird... No loan interest
18:34:28 <LordAro> you also have no loan
18:34:41 <xarick> yes, for some reason...
18:34:47 <merni> No, but in non-infinite mode a negativev money balance is considered a loan
18:35:18 <merni> the AI probably also takes the opportunity to pay back the loan since it has lots of cash
18:35:32 <_glx_> ah yeah like IRL, you don't have money so you need to pay more money
18:35:38 <DorpsGek> - Update: Translations from eints (by translators)
18:36:01 <merni> openttd is too realistic sometimes
18:36:37 <truebrain> merni: except for bus-height
18:38:51 *** gelignite has joined #openttd
18:44:28 *** kale91 has quit IRC (Quit: User went offline on Discord a while ago)
18:45:18 <truebrain> ppfff .. we are too active to not constantly run again the GitHub runner limit 😛
18:45:26 <xarick> ugh... these defaults post difficulty removed... are so off
18:46:41 <xarick> for best performance, is friendly should be off
18:46:57 <xarick> number of days in transit [60, 70]
18:47:06 <xarick> Fastest pathfinder profile
18:47:21 <xarick> found towns on, fund yes, advertise, yes 😦
18:48:08 <michi_cc[d]> So, do I need to re-date the beta3 news post? 😜
18:48:46 <LordAro> have we fixed all the bugs yet?
18:49:00 <xarick> capacity mode is also meh, oh well, RIP LuDiAI AfterFix
18:49:24 <truebrain> frosch123: I am testing via preview btw .. couldn't be bothered to get a setup working local 😛 So I need to test this a bit more 😄
18:50:03 <frosch123> ah, development at site 🙂
18:50:20 <truebrain> but CI has a backlog, so this will take a bit of time 😛
18:50:57 <_zephyris> Thanks truebrain, I wondered why the website CI was still not happy with me!
18:51:06 <truebrain> it is REALLY picky 🙂
18:51:50 <truebrain> Jekyll is nice, but .. has its quirks 😛
18:56:16 <frosch123> hmm, i remember we force-recruited eddi as translator. but apparently xaroth is an inactive translator as well
18:56:21 <xarick> currently testing AAAHogEx with inflation
18:56:43 <truebrain> wait, xarothbrook hasn't been useful as translator? I am absolutely shocked by this revelation 😛
18:58:48 <truebrain> frosch123: I can't find Xaroth having access to translator?
18:59:03 <michi_cc[d]> truebrain: Run your changelog script just to get something moving? ❤️
19:00:55 <truebrain> I just linked that! "but I don't read backlog" .. IT IS STILL ON YOUR SCREEN! 😛
19:01:34 <kuhnovic> Nice article _zephyris . I'm still working on mine. It's currently getting "the wife test".
19:01:36 <truebrain> michi_cc[d]: when I am done fighting eints 🙂
19:02:45 <truebrain> okay, my eints change is far from perfect, as that requires access to the raw english.txt, but it at least is an improvement
19:03:00 <truebrain> sometimes it does things I don't like, but solving that requires a `while True` loop, and I am not willing to do that 🙂
19:03:14 <truebrain> but at least, `_1`, `_2`, etc are always done in order
19:03:35 <truebrain> Rubidium: you can test if you like on translator-preview, but I think I resolved the biggest obstacle 🙂
19:07:54 <truebrain> Rubidium: you can btw test it by pressing the `Get another string` button, which will always be at the start of a "block" (by some measure of block). But never at `_3`, always at `_1` 🙂
19:08:44 <truebrain> and some of our strings are horrible .. some just write out the english sentence .. which isn't very useful by any measure 😄
19:13:02 <frosch123> ok, i am going to kick ~140 translators from the org: they signed up in 2022 or earlier, and never translated anything
19:13:18 <truebrain> how mant translators do we have?
19:16:24 <frosch123> 918 are mentioned in commits, so translated something in WT2, WT3 or eints. some of them are duplicates, like glx and glx22.
19:18:01 <jfs> (where did glx1 to glx21 go?)
19:19:51 <frosch123> 400 requested access on github at some point, 65 failed the invitation, ~140 never did anything and signed up before 2023, ~60 never did anything but signed up 2023 or late, ~200 are active
19:20:06 <frosch123> not sure whether my numbers add up
19:21:00 <frosch123> ah yes, i include the failed invitaitons in the never-did-anything
19:22:35 *** xarothbrook has joined #openttd
19:22:35 <xarothbrook> truebrain: YOU ASKED ME TO!
19:23:06 <xarothbrook> "Hey Xaroth, can you help test this integration"
19:23:06 <xarothbrook> "Who the heck added Xaroth to this list"
19:29:53 <Rubidium> ugh... these strgen pragmas that end up in the language pack header are nasty (and very C-style). So for the customer number formats I either check and make the lookup tables in strgen and waste a huge amount of space that's going to be unused except for a few cases, or we just don't bother at all and it behaves just like the strings do in the PR. Or the whole language file format header writing/reading
19:29:59 <Rubidium> has to be changed...
19:33:24 <Rubidium> I've got no idea how many characters there would be needed per location, but lets say two that's already 9 bytes times 20 for each of the potential spots. Times three different formats, not to forget the abbreviations. That quickly nears 1 kB and that for each and every language that OpenTTD loads.
19:37:56 <truebrain> kuhnovic: any chance you can add some visualisation to the post? 😄
19:38:24 <truebrain> one (or more) of your debug screenshots with the regions in white boxes or something?
19:39:29 <truebrain> btw, absolutely awesome, the post 🙂
19:39:33 <kuhnovic> Sure! Let's see if I have some of those screenshot stored somehwere
19:40:03 <truebrain> poor _zephyris .. 😄
19:40:13 <_zephyris> One more broken link
19:40:42 <michi_cc[d]> _zephyris: You uploaded a gif but linked a png 🙂
19:43:38 <_zephyris> truebrain: Do you get a ping every time I do something weird on Github? 😜
19:43:56 <truebrain> only when you do something weird, yes
19:44:10 <truebrain> no, I am actually interested in the blog, so I clicked it, and saw some redness 🙂
19:44:42 <truebrain> We have some more work to do 😄
19:44:58 <truebrain> did I ever tell you our HTML/CSS sucks balls? 😄
19:45:54 <michi_cc[d]> Not sure if it is just my browser, but...
19:45:58 <_zephyris> So markdown syntax compatible with Jeckyll for setting image width? Or is that impossible?
19:46:11 *** berndj has quit IRC (Read error: Connection reset by peer)
19:46:11 <truebrain> I really do not know
19:46:27 <truebrain> I think I always scaled the images myself 🙂
19:46:47 <_zephyris> 756px looks like the magic target number...
19:46:53 <michi_cc[d]> I think you're allowed to just add some html as well.
19:50:48 <_zephyris> `{:width="25%"}`?
19:51:05 <_zephyris> Well, one obvious way to find out I suppose.
19:51:14 <truebrain> if you scale them yourself, you know how they will look. As we are fixed-width, that works fine 🙂 Also saves a few bytes of network traffic 😄
19:54:11 <locosage> but larger image looks better on retina
19:59:26 <michi_cc[d]> Do we plan on publishing the TTF blog post first? If yes, I'll add a teaster to the beta3 news.
19:59:51 <michi_cc[d]> Also, anbody fancying reviewing/approving things for beta3?
20:00:16 <kuhnovic> Balls. I don't have any of my pathfinder development screenshots anymore. I somehow have to revive that debug code hehe.
20:00:33 <truebrain> I remember some in there?
20:00:52 <kuhnovic> That only has the one that I murdered with paint
20:00:57 <truebrain> michi_cc[d]: I am fine with that; talltyler mentioned he rather would do the daylength one first, but I don't know what timeline we are looking at (no preasure, to be clear; it is done when it is done)
20:01:54 <michi_cc[d]> Well, there's still time for the announcement anyway. No approval, no news 🙂
20:02:10 <truebrain> talking about the devil 😄
20:03:19 <kuhnovic> Haha everyone's writing blogs all of a sudden
20:03:27 <truebrain> not "all of a sudden" 😛
20:03:48 <kuhnovic> I might be able to use those screenshots, but let's see if I can cook up something a bit nicer.
20:04:03 <truebrain> okay, I think we will have 5 in total, unless someone else surprises me. As for the order .. dice-roll works for me 😛
20:04:27 <truebrain> I do wonder if we should add a link to all posts, something like: Want to donate to OpenTTD? Check <here>
20:04:34 <truebrain> making this one big fundraiser campaign 🙂
20:04:34 <talltyler> That’s not the blog I promised, but I’d already written most of it in the PR description!
20:04:47 <talltyler> Still have to write the time post…
20:06:28 <truebrain> maybe a bit more generic: "Interested in helping out with the game? Join us on Discord / GitHub! Want to donate? See <here>"
20:08:34 <michi_cc[d]> talltyler: I added a tease for your unbunching post. And you really need to step up the game after Zephyris came in with pictures LOL
20:08:40 <truebrain> I see michi_cc[d] picked who is next; we can do that for the other posts too .. popularity contest! 😄
20:08:50 <_zephyris> Fingers crossed that's right now!
20:10:15 <_zephyris> michi_cc[d]: Well... It is about graphical design, it'd be a bit rubbish without pictures!
20:13:57 <michi_cc[d]> Hmm, we've all been really blind in the changelog.... 🤣
20:16:57 <talltyler> When I was a train conductor I once had a seating conflict on the annual Santa train: two groups of passengers had the same seats. I checked the dates and they were fine. Then I checked the dates more closely and realized one family had printed off their e-tickets for the previous year! Luckily their new tickets were for the same day and time, so they still got to ride, just in a different car.
20:17:39 <talltyler> I did not ruin Christmas 😛
20:18:12 <michi_cc[d]> Cross-check on website #294, too, please. Just in case :
20:18:15 <truebrain> I btw suggest to do the unbunching this Saturday, and every Saturday after that .. we will run out of blogs, but ... 🙂
20:18:46 <michi_cc[d]> We can always intersperse with beta 4, 5, 6, RC1, 2, 3, and 4
20:18:59 <talltyler> Or find more to write about 😄
20:19:00 <truebrain> basically, there are 8 weekends to fill 🙂
20:19:10 <talltyler> Infinite money mode maybe?
20:19:28 <truebrain> we now have 3, 1 more planned by talltyler , I will do one too about social plugins. So that makes 5
20:19:30 <talltyler> Maybe combined with rebranding cheats as sandbox options
20:19:36 <truebrain> that is a good idea
20:19:41 <truebrain> so 2 more slots, we will figure it out 😛
20:20:41 <michi_cc[d]> A good topic that is just missing anybody that can reasonably write it would be a general AI/GS script round-up. The whole creator scene there is quite anemic and could really use some new people.
20:21:05 <truebrain> the GUI work would also be nice, but harder to write
20:21:34 <michi_cc[d]> And, no, I've never written any AI/GS, so I don't think I am qualified to write anything usefull.
20:24:47 <truebrain> and if all fails, we just ask frosch123 to write about hovercrafts
20:25:23 <truebrain> I also think a post about new player experience could be nice btw; also a generic topic. And one you don't need to know AI/GS for michi_cc[d] 😛
20:25:28 <frosch123> ok, in the end i kicked 69 inactive translators. the rest were already no members (9 left themself, the rest were failed-invitations or rejected-applications, which my script included in the count)
20:25:38 <andythenorth> did we add hovercraft? 😮
20:25:49 <truebrain> frosch123: you did that on purpose, right, that number?
20:26:03 <truebrain> anyway, nice, tnx for cleaning up 🙂
20:26:13 <frosch123> i considered rounding to 70 to obfuscate it
20:26:17 <andythenorth> michi_cc[d]: I wrote a GS
20:26:31 <andythenorth> I am not sure that makes me a good person to write about it 😛
20:27:04 <frosch123> members can be counted in uint8 again
20:27:20 <truebrain> we also need to poke the translators to prepare for 14.0
20:27:28 <truebrain> that always really helps, I have seen in last few versions
20:28:01 <andythenorth> /me wonders how big water regions are 😛
20:28:11 <truebrain> if only there was a blog telling you
20:28:46 <peter1138[d]> They're about the same size as a Minecraft chunk.
20:28:56 <truebrain> I will give translators a poke
20:29:39 <andythenorth> so could industries check if a water tile is in a minecraft chunk?
20:30:48 <xarick> 2,147,483,647 surpassed
20:31:18 <peter1138[d]> Now spend it all in one go 🙂
20:31:47 <xarick> im waiting for 170 years inflation
20:33:15 <xarick> you underestimate inflation in openttd
20:33:43 <michi_cc[d]> You underestimate 63 bits.
20:34:00 <xarick> 15 years yields these results, 170 years.... boom catastrophic
20:34:37 <xarick> INT32_MAX is gonna be insufficient
20:35:14 <truebrain> Translators poke done
20:36:22 <truebrain> I always like it when one by one start replying: NNN done! Let's see if that happens this year too 🙂
20:37:45 <michi_cc[d]> xarick: I don't think the total funds are of any importance. AIs check if they have enough money for whatever they want to do/buy. Even with inflation buying a train or a bus will not exceed INT32.
20:38:15 <peter1138[d]> They might want to build a canal though 😉
20:38:35 <xarick> aqueducts are mighty expensive
20:39:50 <truebrain> frosch123: to confirm, the discussion in that repository does trigger an email in your mailbox (translators)
20:40:53 <truebrain> this discussion place is new for me 🙂
20:41:11 <truebrain> I like how we only spam once a year 🙂
20:41:24 <michi_cc[d]> Didn't GitHub want to shut down teams discussions at some point? Or was that idea killed off again.
20:41:31 <truebrain> no, they did, sadly
20:41:36 <truebrain> I still hate that choice, but whatever
20:41:45 <truebrain> so now we have a "team" repository, here we use it 🙂
20:43:29 <michi_cc[d]> Hmm, turing test in #490 🙂
20:44:12 <frosch123> ah, I need _glx_ annotation magic
20:44:27 *** gelignite has quit IRC (Quit: Stay safe!)
20:49:07 <_glx_> oh node16 strikes again
20:49:16 <_glx_> (we have too many repos)
20:49:33 <truebrain> but most of it is already condensed
20:49:54 <truebrain> someone is not on the ignore list 😄
20:50:14 <truebrain> found some faulty permissions in some repos too .. but for another time
20:51:39 <truebrain> finally visibility on BaNaNaS if 14.0 is going to fix the "huge" bandwidth bill
20:51:50 <truebrain> now to wait for a few hours for data to be useful to analyze 🙂
20:52:30 <truebrain> but it really looks like if a download took more than 60 seconds over HTTP, that it would fail, fall back to TCP
20:52:36 <truebrain> and _jgr_ fixed that bug \o/
20:52:58 <truebrain> so I have good hope the bill might actually go down (a lot) with 14.0 🙂 But .... first numbers 😄
20:53:23 <truebrain> frosch123: my name isn't _glx_ , but you can merge your PR now 🙂
20:54:09 <michi_cc[d]> Gentle reminder: workflows/publish-docs has a node16 warning and some "npm WARN deprecated " stuff.
21:06:27 <frosch123> hmm, why did it now show that to me
21:06:37 <truebrain> they requested access to the Translators team
21:06:37 <xarick> > support more number formats that just the 'western' one
21:06:39 <truebrain> which is a team higher
21:07:25 <frosch123> yes, but it looked like someone external to me, it did not show the org membership when i checked their profile. now it does
21:07:28 <frosch123> i guess i get old 🙂
21:07:40 <frosch123> reading backlog, looking at screen, etc
21:07:41 <truebrain> I dunno; I did not check that 🙂
21:07:52 <wensimehrp> Rubidium: I asked a few players in the Chinese OpenTTD Community and some other places. Many said that they are used to the 000,000 format. Sorry to say that, but would it be better if you could change this to a configurable setting?
21:07:52 <wensimehrp> I am very sorry if this would cause any extra work. Sorry.
21:08:33 <michi_cc[d]> Okay, GitHub must have both slow and fast macOS runners. Beta 2 was done on macOS after 18 minutes, while today for beta 3 we're not even at the half-way point at that time.
21:08:48 <truebrain> it is really random on these machines
21:08:58 <truebrain> but soon we can upgrade to M2
21:09:05 <truebrain> which should make everything much faster in general (for MacOS)
21:10:06 <michi_cc[d]> Boo for the evil MS overload to just give M2 instances away for free for OSS projects /s
21:10:29 <truebrain> their MacOS support is surprisingly good
21:10:47 <truebrain> just that they removed vcpkg from the images without any notification I could find, that was a bit more annoying .. but okay
21:11:33 <truebrain> after branching, we will fix that up 🙂 Didn't want to switch to the M2 before branching, as you never know what shit comes out of that 😛
21:12:29 <_glx_> I think they never added them to arm images
21:12:47 <truebrain> I asked; they no longer want the maintenance burden of having vcpkg in the image
21:12:55 <truebrain> so I assume it means it will also not come in other new images
21:13:05 <truebrain> but I couldn't find an announcement or decision about that before asking them
21:13:16 <_glx_> I think they'll announce if they remove
21:13:18 <truebrain> I can understand them not having vcpkg in there, but pfffff
21:13:27 <truebrain> I think they will never remove it; just not add it in new images
21:13:39 <truebrain> which will be fun when macos default rolls over
21:14:31 <truebrain> owh, sorry, the new runners are M1, not M2
21:17:22 <xarick> btw the removal of difficulty levels are making some AIs "underperform" on purpose due to the conversion.
21:17:59 <xarick> seems it's not just mine
21:18:17 <Rubidium> wensimehrp: well, the whole point is a bit that it's in a sense super configurable but with that comes a very complicated actual configuration. To the point that many wondered whether translators would know what to do with it. The same will likely hold with the number format. So it'll likely become something that will often get messed up with all problems associated with that. And let me guess, they
21:18:23 <Rubidium> actually want to mix-and-match, i.e. the Chinese power of 10000 for the short names but power of 1000 the digit separators
21:19:21 <truebrain> please don't make a "fix everything" PR 😛
21:19:24 <truebrain> saveload bumps are cheap 🙂
21:19:25 <Rubidium> or do they want some literal translation of million, billion, trillion, etc as suffix?
21:20:11 <_glx_> yes that can be 2 more PRs
21:20:42 <_glx_> but at least my ideas are written somewhere
21:21:59 <truebrain> pfff, MacOS is especially slow today, pfffff
21:22:07 <truebrain> I need to press buttons, hurry up, I want sleeeepppppp 😄
21:22:07 <_glx_> I already sneaked a GUI QoL change in the PR
21:22:35 <_glx_> forced closing of settings window was annoyin
21:23:07 <_glx_> especially when it stays open when it starts
21:23:38 <_glx_> and debug window "configure" button was not properly disabled
21:26:20 <truebrain> it is a really slow node 😛
21:27:05 <truebrain> now waiting for Apple to approve
21:27:50 <truebrain> finally, time to upload \o/
21:28:58 <michi_cc[d]> Watching along, too? Passed the random notarization failure 🙂
21:29:42 <truebrain> published news on Steam 🙂
21:30:33 <truebrain> and pressing the GOG buttons
21:31:30 <xarick> i wish it was possible to "insert" a GS on a running game if it has none yet
21:33:44 <wensimehrp> Rubidium: Well, you are partly right
21:34:43 <wensimehrp> Does discord replies show up on the IRC channel?
21:35:10 <wensimehrp> > actually want to mix-and-match, i.e. the Chinese power of 10000 for the short names but power of 1000 the digit separators
21:35:19 <michi_cc[d]> No, IRC has no reply context.
21:38:01 <talltyler> Discord announcement made
21:38:31 <michi_cc[d]> Xitter and Reddit done, too. Mastodon, forums and whatever is left for someone(TM) to do 🙂
21:38:39 <talltyler> We seem to have forgotten beta2 on TT-Forums 🙂
21:38:47 <talltyler> I'll do beta3 though
21:39:23 <frosch123> isn't it in the same topic?
21:40:02 <talltyler> Oh, Kamnet posted it, but in a quote so I ignored it
21:40:10 <talltyler> Reading comprehension ⭐
21:40:24 <frosch123> from reddit to forums 🙂
21:40:52 <talltyler> Who has access to Mastodon?
21:41:54 <Rubidium> @topic set 1 13.4, 14.0-beta3
21:41:54 *** DorpsGek changes topic to "13.4, 14.0-beta3 | Website: *.openttd.org (source: github, translator: translator, server list: servers, wiki: wiki) | Don't ask to ask, just ask | 'Latest' is not a valid version, 'Most recent' neither | English only"
21:43:58 <frosch123> lol, the mastodon post has 10x as many likes and reposts than we ever had on xitter
21:45:59 <truebrain> Peter is excellent at PR 🙂
21:48:45 <frosch123> rubidium: do you know of a language with differing number_format and currency_fomat?
21:51:04 <Rubidium> frosch123: Pig Latin maybe?
21:55:10 <kuhnovic> Man reading all these blog posts really gets me excited for OpenTTD 14.0. There's so much cool stuff in it, and I'm not just talking about the stuff in the blogs.
22:06:59 <andythenorth> hmm so industry grf var for water tiles, number of connected regions?
22:07:56 <_glx_> hey at least it doesn't crash
22:11:24 <_zephyris> talltyler: Double bug! 🐛🦗 Just little UI things.
22:13:35 <truebrain> Pff, you should have found those during testing
22:13:41 <truebrain> Window is closed now 😛
22:13:58 <michi_cc[d]> truebrain: It's not written yet, but maybe I can offer you a blog with the title "How the sausage is made, or: How OpenTTD is developed"
22:14:18 <talltyler> #12020 had an error to prevent it, but I removed it because I thought the case was impossible 😛
22:14:41 <truebrain> michi_cc[d]: By pitchforks and the plague? 😄 without joking, go for it! Sounds interesting 🙂
22:14:56 *** keikoz has quit IRC (Ping timeout: 480 seconds)
22:15:40 <michi_cc[d]> My plan would be to stay relatively light on the technical details and focus more on the process and how decisions are made.
22:16:32 <andythenorth> "by those who turn up"
22:16:42 <andythenorth> and also [name redacted]
22:16:44 <michi_cc[d]> Thankfully, we already have some high-quality posts, so I still have a few weeks to actually write it 🙂
22:17:24 <truebrain> Don't delay too long 😛
22:17:51 <michi_cc[d]> Post-release engagement 😛
22:23:33 <peter1138[d]> Only Hackykid's implementation.
22:23:39 <locosage> there was even worse pbs bug somewhere
22:24:09 <andythenorth> how do I get 'ignore signals' button to work?
22:24:16 <andythenorth> I have tried click, ctrl-click and shift-click
22:24:33 <peter1138[d]> double click sometimes.
22:24:52 <andythenorth> I have a lot of 0 power trains in my savegame
22:24:56 <andythenorth> having to crash things into them
22:25:56 <locosage> andythenorth: usually spam-click works xD
22:26:33 <andythenorth> I appreciate whatever accidental-click protection we added 😛
22:26:57 <locosage> it's completely accidental xD
22:29:11 <frosch123> iirc it depends on path/block signals: 1 click to ignore signal, a second click to ignore path reservations, or something
22:38:51 <_glx_> beta are just promoted nightlies
22:39:44 <reldred> It's also the hype cycle, gotta stoke interest, show rapid progress
22:39:48 <reldred> feed the media engine
22:40:04 <reldred> get more useful testing and metrics caught by a wider audience
22:46:31 <xarick> i'm starting to get dangerous numbers
22:50:48 <xarick> heh, it's cheaper to build a Farm
22:50:55 <xarick> than an aqueduct of this length
22:53:05 *** nielsm has quit IRC (Ping timeout: 480 seconds)
22:53:50 <xarick> town generation lets me fund towns of size "random" for the price of a "medium"
22:55:20 <reldred> you could get a small for the price of a medium
22:56:31 *** m1cr0man has quit IRC (Quit: G'luck)
22:56:47 *** m1cr0man has joined #openttd
23:02:17 <xarick> rather just build medium
23:05:14 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
23:12:59 <peter1138[d]> Somehow this terrain feels a bit minecrafty
23:16:05 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
23:22:46 <_glx_> oh that's why the setting is disabled ingame 🙂
23:23:37 <xarick> i also suffer from this bug, it is missing for quite some time
23:24:10 <talltyler> It's supposed to be disabled in Scenario Editor too, but apparently is not
23:24:55 <_glx_> I can understand scenario creators wanting to change it
23:25:45 <_glx_> imagine spending a lot of time placing towns and industry, and then find out the setting was wrong
23:27:05 <xarick> oh, nvm, it's showing now, must have been fixed very recently
23:30:19 <xarick> left control key vs right control key
23:37:26 <locosage> but now servicing is more clicks...
23:37:58 <xarick> I can already see breakdowns removed
23:38:50 <locosage> though service if needed doesn't really work anyway
23:39:09 <_glx_> hmm seems it's `xor`, but could have been just `or`, as I don't see anything presetting service
23:39:16 <xarick> it works, but there's a distance limit
23:40:02 <_glx_> it works, but not if vehicle is already in a reserved path when the search happens
23:40:34 <locosage> doesn't it just skip the order if service not needed?
23:40:34 <_glx_> hmm no that's for auto service without service order
23:40:47 <_glx_> yeah with the order it should be fine
23:41:01 <locosage> and then just drive past depot because order is skipped
23:41:16 <_glx_> it's the search for closest depot which is "broken" with reserved paths
23:41:52 <xarick> place a sign right before depot
23:42:07 <xarick> and dont use service at nearest depot order
23:42:07 <locosage> it has exact depot, why would in search for closest
23:42:24 <_glx_> yeah with the service order it works
23:42:30 <locosage> xarick: yeah, that's the only way to make it work afaict
23:42:49 <_glx_> because it's an actual order
23:43:39 <xarick> isn't the xor to toggle between service modes when clicking?
23:43:43 *** tokai|noir has joined #openttd
23:43:43 *** ChanServ sets mode: +v tokai|noir
23:44:07 <_glx_> not here, it's creating a new order anyway
23:44:58 <_glx_> but the xor could have been for "invert the default", but default doesn't seem to be configurable
23:45:36 <_glx_> doesn't matter, xor or or are fine in this case
23:47:09 <xarick> tunnels under industries, houses, etc... should be forbidden 🙂
23:48:26 <xarick> which tells how much height is required to overcome the size of the structure
23:48:50 <xarick> same for bridges perhaps
23:50:10 <xarick> while at it, a property for ships about how much height is required to pass under a bridge
23:50:30 *** tokai has quit IRC (Ping timeout: 480 seconds)
23:53:28 *** Flygon has quit IRC (Read error: Connection reset by peer)
23:57:20 <xarick> bridges could have a similar height property, then someone could make bridge over bridge a possibility with less graphic issues (maybe)
23:57:51 <peter1138[d]> Well that's already a thing in JGRPP.
23:58:15 <peter1138[d]> (And an old patch of mine from about 18 years ago...)
continue to next day ⏵
