IRC logs for #openttd on OFTC at 2023-02-27
⏴ go to previous day
01:10:06 *** tokai|noir has joined #openttd
01:10:06 *** ChanServ sets mode: +v tokai|noir
01:16:49 *** tokai has quit IRC (Ping timeout: 480 seconds)
01:46:52 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
01:46:55 *** WormnestAndroid has joined #openttd
01:56:48 *** _aD has quit IRC (Quit: leaving)
03:26:12 *** Wormnest has quit IRC (Quit: Leaving)
03:38:54 *** debdog has quit IRC (Ping timeout: 480 seconds)
03:38:59 *** D-HUND is now known as debdog
07:06:57 *** milek7_ has quit IRC (Ping timeout: 480 seconds)
07:07:12 *** azubieta60822 has quit IRC (Ping timeout: 480 seconds)
07:07:34 *** azubieta60822 has joined #openttd
07:24:35 *** sla_ro|master has joined #openttd
07:40:12 <petern> There's always a reason
07:54:37 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
07:55:45 *** WormnestAndroid has joined #openttd
08:00:16 *** hylshols7qui[m] has quit IRC ()
08:20:17 *** supermop_toil_ has quit IRC (Read error: Connection reset by peer)
08:39:45 <petern> Hmm, it's cold again. That'll be why.
08:42:25 <Rubidium> nah, it's getting colder as we're getting further from the sun ;)
08:44:39 <Rubidium> after all, we're closest to the sun around begin of January and furthest from the sun around July, by about 5 Gm
08:59:25 <petern> Hmm, of course now that I automated the build & deployment process, uploading a special different version just for demo is now a bit more awkward. Hmm.
09:29:04 <Rubidium> hmm, does github support some search in PRs with "modifies file with map in its name"? I can't really find it, but that would be really useful to see what PRs my PR would be breaking
09:33:26 *** gelignite has joined #openttd
10:13:47 <petern> Breaking existing PRs is the best 😄
10:16:03 <petern> LordAro, I knew I wasn't imaging it, you did 100 miles on the 18th!
10:17:07 <LordAro> and then broke when i tried to do 60 miles the day after
10:18:31 <LordAro> like, things went a bit wobbly sort of broke
10:18:34 <LordAro> not done that in a while
10:18:58 <petern> That can happen, which is why I just did a short ride instead.
10:19:18 <petern> Although it was (for around here) quite a climb.
10:20:02 <petern> Thinking of doing Semaine Federale in July. Week of cycling around Metz in France, based out of a tent.
10:21:10 <LordAro> that sounds like excellent fun
10:34:22 <petern> I last went 7 years ago, christ doesn't time fly.
11:15:56 *** yur3shmukcik[m] has quit IRC ()
11:19:32 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
11:19:41 *** WormnestAndroid has joined #openttd
11:43:05 <andythenorth> started a few games
11:43:16 <andythenorth> re-generated another 100 or so maps 😛
11:46:12 <Xarick> let's see if i can convert myself into using discord full time
11:46:39 <Xarick> the interface is quite bloated
12:01:58 *** Samu has quit IRC (Quit: Leaving)
12:29:53 <Xarick> why am I getting depot failed after it was successfully built?
12:30:47 <Xarick> this is a piece of code from CoronaAI, btw
12:34:37 <glx[d]> Because the test is wrong
12:35:06 <glx[d]> It says failed if depot is present and has the correct owner
13:37:30 <Xarick> so, I installed discord app
13:38:35 <Xarick> I expected it to be different than the browser version
13:41:51 <LordAro> welcome to the modern world of UI applications
13:42:17 <LordAro> they're all actually websites
14:21:44 <Xarick> how do i stop receiving notifications from all ?
14:44:00 <Xarick> how do i hide the letfside pannel, the one that lists channels, I'm only interested in this one
14:45:29 <Xarick> don't need to have it listed
14:45:55 <LordAro> you might find you have other things to say in other channels
14:48:32 <kamnet> You could mute the other channels and hide them, but the server badge, Events and Browse Channels are always going to be available on that bar, as well as the shortcut to your personal settings, mic, and headphone muts at the bottom
14:51:23 <Xarick> it's too bloated, imop
14:59:33 <kamnet> Xarick: Aha! There's a userscript for it.
15:05:19 <Xarick> not sure I wanna install
15:05:53 <Xarick> need tampermonkey, but i'm using the discord app, not discord on browser
15:17:02 *** HerzogDeXtEr has joined #openttd
15:55:59 <petern> What kind of stupid mug breaks when you put hot water in it...
16:05:15 <Xarick> I have a dream of having 15 AIs with 5000 trains, road vehicles, ships and aircraft
16:05:32 <Xarick> still unsuccessful at that dream 😦
16:06:21 <Xarick> i think not enough stations was the problem
16:06:42 <Xarick> then after that, it's the extreme slow down the game gets
16:25:49 <dP> what was 10399 supposed to achieve?
16:25:54 <dP> I can't see any usecase for it
16:26:21 <dP> there was category talk at the time but it can't do that afaict
16:30:25 *** gelignite has quit IRC (Quit: Stay safe!)
16:37:27 *** audigexJon has joined #openttd
16:37:27 <audigexJon> 10399 definitely improves the functionality options for variants. It's still not perfect but it's better than the current implementation - especially for real-world sets where it's most valuable
16:38:58 <audigexJon> Although it would be nice if there was a way to disable buying the "Group header" item, that would be even better, that way it could just be an actual header rather than a purchasable unit
16:39:45 <dP> if there was a way to disable buying header (and it didn't switch) 10399 would be unnecessary :p
16:40:16 <dP> you could just make a header instead of renaming all vehicles depending on the context
16:40:38 <petern> Setting no climates should do it.
16:40:49 <petern> (That doesn't stop the second part though)
16:40:54 <dP> petern: it disapearss after first buy
16:41:07 <audigexJon> It would still be necessary if you didn't want to duplicate the header in the list
16:41:07 <audigexJon> Without 10399 you'd need a separate unit as a header, and they'd show up inside the group too
16:41:50 <dP> yeah, but I'd argue having a separate unit is way better than renaming every vehicle to category name
16:42:06 <audigexJon> All I want is the ability to have
16:42:06 <audigexJon> Electrostar (non purchasable)
16:42:26 <dP> though I guess if it wasn't change you only need to rename top vehicle
16:42:53 <audigexJon> I don't care about renaming every vehicle and having a callback, it's not ideal but I'd deal with it
16:43:08 <petern> So there's loads of space for flags to do extra things like, maybe, not replace the header
16:44:21 <audigexJon> audigexJon: I'll straight up pay someone to make this possible. Couldn't give a shit how it's implemented as long as I can do it without the user being able to purchase the top item, or duplicating it in the list
16:44:34 <dP> if you make it newgrf-configurable you'll end up in a weird state where some groups work one one and some another
16:45:43 <audigexJon> tbh I think the actual solution is just make it so you can create a separate vehicle as the header, and it's not purchasable
16:45:43 <audigexJon> No idea why it wasn't done like that in the first place, I assume there's some technical reason but it would make so much more sense than duplicating a vehicle in the group IMO
16:46:49 <petern> Main reason is nobody wants to test anything until it's merged, and then it's done.
16:47:48 <petern> And then after it's done, people have "ideas" about how it could be brilliantly better, but never do anything about that apart from say how good it could be.
16:48:49 <audigexJon> I didn't even know it was happening tbh - perhaps when there are newGRF changes, a quick "@newgrfdevelopers - take a look at X we could do with your input" would be useful?
16:49:22 <dP> there wasn't even a clear goal for that change to test it :p
16:49:22 <dP> it was just, hey, we can do trees, weee
16:49:28 <dP> we'll merge and figure it out :p
16:51:45 <glx[d]> oh andy had some usage ideas
16:58:06 <dP> andythenorth: oh, yeah... 😆
17:02:04 <audigexJon> This (for the 377, assume the others have similar trees) is pretty much my dream - except without the headers being repeated in the sub-list
17:02:04 <audigexJon> I've implemented this with separate "header" vehicles, but I'd be completely okay with being able to do it via something similar to 10399 with a switch on the vehicles in the group
17:03:15 <petern> Replacing the variant happens so that you don't need to open the full tree every time to buy the same thing. It is to help declutter the list when you have a few similar items in the list.
17:04:11 *** gelignite has joined #openttd
17:04:29 <dP> but it conflicts with the concept of the group
17:04:41 <dP> because any vehicle can be the group header
17:05:19 <petern> Wrong away around if it does apply.
17:05:29 <dP> and if you rename it to show the header you don't know what you're buying anymore
17:06:04 <petern> I don't really know the purpose of 10399, but making the last-purchase variant the default on your next purchase was the intention, and the purpose of the feature.
17:07:56 <audigexJon> That makes sense for some types of set, but makes no sense for others, hence it would be good to have an ability to do one or the other depending on how the set works
17:08:49 <audigexJon> For real world sets a group is just a group
17:09:09 <audigexJon> Also surely everyone just clones existing trains anyway, when they want another one, rather than trawling the purchase list?
17:13:19 *** Wormnest has joined #openttd
17:15:58 <dP> petern: as I see it this only makes sense when there is a clear superior option that you just want to chose once and forget about others
17:16:31 <dP> like, put the whole iron horse into a group, choose the fastest and be done 😜
17:18:57 <dP> but there is already vehicle hiding feature that serves the same purpose
17:22:21 <dP> because if you're switching between options constantly it's better to have a consistently named category that stays in the same place and has the same icon then trying to remember if the current variant is in the same group as the one you want
17:24:32 *** sla_ro|master has quit IRC ()
17:28:06 <petern> Might as well throw it out
17:29:12 <Xarick> does @calc work fine on discord?
17:30:34 <dP> no, you'll have to do your math privately 😜
17:41:42 <Xarick> OrderListPool 64000 is too short
17:42:10 <JGR> Use shared orders like everyone else?
17:43:32 <Xarick> they just build road vehicles
17:44:14 <Xarick> not even trying aircraft, train, ships... my dream of 5k of everything can't be fulfilled
17:44:48 <JGR> As dreams go that doesn't sound very inspiring
17:47:23 <Xarick> let me check wentbourne
17:49:57 <LordAro> looks like calc works fine to me
17:50:02 <LordAro> do you not get the result? :D
17:53:14 <andythenorth> petern: a stupid one
17:55:41 <andythenorth> audigexJon: it was debated almost to death, but you're right, there's no clear way to ping a group
17:56:05 <andythenorth> but then again...it's not clear that more debate adds much
17:56:14 <andythenorth> implementations beat discussions
17:56:16 <LordAro> increasing the orderlist pool wouldn't be the hardest thing to do
17:56:21 <LordAro> in theory it's an completely arbitrary limit
17:57:33 <andythenorth> petern: petern goes it throw out Variants?
17:57:58 <andythenorth> I haven't included it in a Horse release anway
17:58:34 <Xarick> a typical corona ai with 5000 road vehicles
17:58:57 <Xarick> and it breaches some limits 😦
17:59:13 <frosch> LordAro: change the error message into a link about shared orders
18:00:15 <frosch> noone intends to fill the orderlist pool. it's just if people do not know about shared orders
18:01:47 <andythenorth> we haven't learnt why it was regrettable yet
18:02:04 <petern> I did through out the mug though.
18:02:11 <Xarick> if the AI has 5000 road vehicles, how come there's 5120 OrderList's in the pool?
18:02:47 <petern> Isn't it obvious? It doesn't immediately cater to everyone's different use-case so it's clearly a bad ill-thought-out feature.
18:02:48 <Xarick> disaster or effect vehicles also have orders?
18:03:09 <LordAro> they actually do, though i'd be surprised if there were 120 of them
18:03:52 <Rubidium> 5120 capacity, not used...
18:04:29 <andythenorth> petern: you mean the mug yes?
18:04:35 <Rubidium> after all, vehicle, orderlist and order are all a nice multiple of 1024
18:04:36 <andythenorth> it doesn't cater for your use case
18:04:55 <andythenorth> so is newgrf roadstops merged?
18:05:01 <andythenorth> so many grfs I didn't write yet
18:05:20 <Rubidium> also, if you increase the orderlist pool, increase max-vehicles-per-company by the same factor ;)
18:05:31 <petern> Not by me yet. I have some stumbling issues with leisure coding right now.
18:06:19 <petern> (Any coding tbh, but being paid by day job kinda means that has to happen anyway)
18:06:33 <andythenorth> I have too much day job to have any grf or GS brain space 😛
18:08:29 <audigexJon> petern: Seems like a bit of an aggressive interpretation of my words tbh - I just asked for a modification to the feature that would suit a different use case (which, in my opinion, seemed like the most common use case)
18:08:29 <audigexJon> That doesn’t make it badly thought out, it’s a great idea that can be built on
18:10:35 <audigexJon> Unless that wasn’t aimed at me in which case I’m the one interpreting badly
18:11:01 <WizzStar> is it ever possible to make an newgrf (not me cus im a noob xD) or included in a game update one day that we have diagonal roads just like rails ? i see the vehicles can drive diagonally in a corner of a road ... we can make so much better intersections with diag roads 🙂
18:11:05 *** sla_ro|master has joined #openttd
18:12:40 <audigexJon> It's theoretically possible, there have been patches for it at least a couple of times IIRC, but it's difficult to do well and it seems nobody's ever had the time to really implement it fully and iron out the bugs
18:13:44 <audigexJon> You can't do it with a newGRF alone, though - it requires some fairly substantial changes to the game itself
18:15:14 <audigexJon> It gets suggested a few times a year, so it's definitely a popular feature request, but it's a lot of work and absolutely no trivial to do properly - we might see it one day but it's not a high enough priority, considering the amount of work, for anyone to do it ahead of things they would rather see
18:18:02 <audigexJon> andythenorth: To be fair I do tend to dip in and out of the community, but even at times when I'm active it's easy to miss a conversation if chat flies away from it
18:18:02 <audigexJon> Perhaps something like development-forum could make sense, for requesting input and testing? Basically a "check here if you want to test and discuss ideas, but you're only allowed to discuss if you'll also commit to testing"
18:18:59 <JGR> There's no way to enforce that anyone tests anything
18:21:44 <JGR> Having some sort of hint to the add-on-development channel about NewGRF PR activity might be helpful though?
18:24:49 <andythenorth> I am open to open-ness
18:27:32 <audigexJon> JGR: No real hard way, but "Dude you keep suggesting stuff and not helping to test, don't be an ass" should shame most people into actually helping rather than expecting everyone else to do all the work for them while they just cherry pick features
18:31:46 <dP> it would be nice to have some kind of final warning before and impactful pr is going to be merged, just to let everyone know it's time to do a final test and voice their thought.
18:32:24 <dP> because right now there is very little indication whether pr is going to be merged tomorrow or rejected.
18:32:58 * andythenorth remembers when things just appeared in svn trunk
18:35:27 <dP> dP: for controversial PRs it can even be accompanied by a voting system of some kind
18:36:29 <andythenorth> not sure that conforms with the desire to get as many PRs through as possible
18:37:07 <JGR> dP: This is not really practical
18:37:19 <JGR> The project is not a democracy, after all
18:37:44 <JGR> People would vote yes for everything, no matter how daft
18:37:45 <dP> yeah, but for developers that don't play the game it's hard to tell what's even useful :p
18:37:45 <andythenorth> it's more or less an anarchy, of sorts
18:37:47 <glx[d]> often the only way to get real testing is to merge
18:38:45 <dP> but some changes can't be easily fixed or reverted after merge
18:40:43 <dP> actually most changes are harder to fix once merged because there is no motivation for anyone to do that
18:41:04 <dP> unless it breaks something horribly
18:41:32 <JGR> The only things that sure really be problematic to unmerge are stuff like savegame bumps and NewGRF spec additions
18:42:05 <JGR> As for the latter, I'd argue that this is in large part a defect in the way the NewGRF spec works
18:43:12 <glx[d]> in many case nobody uses new newgrf features unless they are already merged
18:43:24 <TallTyler> andythenorth: You just don't have the ability to ping roles. There's two resolutions to deal with this: either people with the OpenTTD Developer roles gain the ability to ping `all`, `all`, and all roles, or we just make `@NewGRF Developer` pingable by anybody. I'll see what the other moderators think.
18:43:28 <glx[d]> even when a PR build is available
18:44:14 <andythenorth> well it looks like newgrf dev is somewhat moving to JGRPP anyway
18:44:24 <andythenorth> which might solve it
18:44:48 <JGR> There's a fairly small subset of NewGRF that I'm interested in
18:45:29 <JGR> (Though I do have opinions about low-level NewGRF stuff in general)
18:46:28 <andythenorth> I had a list of grf things that I thought were JGRPP-only
18:46:40 <andythenorth> - advanced path signals
18:46:48 <andythenorth> - bridges over stations
18:46:59 <andythenorth> - more terrain tiles (rocks stuff)
18:47:14 <glx[d]> roadstops is in vanilla now (a little different implementation IIRC)
18:47:46 <JGR> One of my opinions is that it's too hard to add stuff to NewGRF because in vanilla you then have to commit to supporting it until t = infinity
18:47:59 <andythenorth> only because the community get very cross
18:48:07 <andythenorth> and that's massively inhibiting
18:48:10 <TallTyler> I don't know why bridges over stations hasn't made it to vanilla besides "nobody cares enough to take the time"
18:48:14 <TallTyler> It seems pretty obvious
18:48:28 <andythenorth> well the reason is in the second half of your question 🙂
18:48:47 <TallTyler> Maybe I'll care enough someday but I have enough upstreaming tasks on my plate
18:49:18 <JGR> One of the reasons for the roundabout way in which the GRF stuff is structured in my branch is so that I can remove/replace stuff in future if need be
18:49:44 <JGR> And GRFs will still do something sensible, instead of just falling over with a meaningless error message
18:50:15 <TallTyler> I wonder if that sensible failing could be brought to vanilla
18:50:21 <TallTyler> Not knowing exactly what you mean
18:50:38 <glx[d]> special newgrf feature check
18:51:10 <glx[d]> newgrf must test for availability of the feature before using it
18:52:11 <JGR> The only version info you can look at in vanilla is variable 0x21
18:52:12 <glx[d]> but it's good for experimental stuff, not sure it would fit in vanilla
18:53:32 <glx[d]> there used to be svn version checks, but that's dead since the move to git
18:56:18 <LordAro> andythenorth: roadstops were merged yesterday :)
19:00:26 <andythenorth> I will find a suitable gif 🙂
19:05:19 *** Flygon has quit IRC (Quit: A toaster's basically a soldering iron designed to toast bread)
19:10:01 <LordAro> petern: do i want to cycle to (near) manchester before a 200k, or shall i get the train?
19:11:30 <TallTyler> lol, it built fine before I rebased it and forgot to rebuild before pushing 😛
19:11:47 <TallTyler> Good thing actions now cancel previous actions on the same PR
19:16:26 *** gelignite has quit IRC (Quit: Stay safe!)
19:18:35 <LordAro> TallTyler: also a load of warnings ;)
19:19:05 *** _aD has quit IRC (Quit: leaving)
19:19:09 <LordAro> TallTyler: CommandCost param should probably be a const ref too
19:29:10 <TallTyler> Oof, looks like actually just warning repeated a million times in three places 🙄
19:29:29 <TallTyler> But should be fixed now 🙂
19:42:48 <audigexJon> What's this roadstop merge I keep seeing mentioned?
19:45:51 <audigexJon> Thanks, looks interesting
19:50:26 <Arastais> Are there plans to make a 14.0 milestone soon on github?
19:50:38 <Arastais> (and assign the major changes to it)
19:50:50 <TallTyler> We still need to release 13.1 😛
19:51:20 <TallTyler> Bugfix builds like this don't generally contain new features, but it might be a bit early to think about 14.0
19:51:33 <TallTyler> Why? Do you have something you want to add to the milestone?
19:51:34 <Arastais> TallTyler: true lol, but i was just wondering about creating the milestones ahead of time, would make it easier to track prs
19:51:57 <Arastais> TallTyler: eventually yes, but not right now. Would take me maybe up to a month to finish.
19:53:03 <TallTyler> I'd imagine 14.0 would be well more than a month away. I've mostly seen the milestones used to spur PR completion and review before a release, not drive new PRs to be created
19:54:57 *** gelignite has joined #openttd
19:55:08 <Arastais> ah ok, i was thinking that you would use milestones on every open pr thats a candidate, in order to convey what prs will be part of what release.
19:55:25 <TallTyler> No, if a PR is merged before the release, it is always included
19:56:42 <andythenorth> I can't see us releasing major version more than quarterly
19:56:51 <andythenorth> more likely 2-3 times a year
19:56:54 <TrueBrain> In general things get merged when they are done; as we have learnt that waiting on "just that one more PR" means we delay and delay and delay 😛
19:59:09 <TrueBrain> you can image every PR being scheduled for the next milestone
19:59:50 <TrueBrain> we always have the intention to merge them all .. given enough time and energy for both any developers as the author 🙂
20:02:19 <andythenorth> we have neither the capability nor the inclination to do advanced release management 🙂
20:02:43 <Arastais> TrueBrain: yea; unfortunately it looks like many prs in openttd have been abandoned by the pr authors themselves, and I haven't really seen that with larger-scale projects like cmake.
20:02:58 <Arastais> andythenorth: yea, it would only work if prs for a milestone are taking too long and are thus just dropped
20:03:06 <andythenorth> I didn't clean any PRs or issues up recently 🙂
20:03:24 <TrueBrain> Arastais: different audience, I am sure. But yes, that happens. And not enough developer-time to pick the PRs up, and actually make them worth of merging 🙂
20:03:42 <TallTyler> I've been doing my best to keep issues and PR lists updated, but haven't closed anything for staleness
20:03:45 <frosch> i have seen plenty of major projects with many abandoned PRs
20:03:54 <frosch> cmake probably has a close-bot
20:04:02 <andythenorth> we had a close bot
20:04:05 <TrueBrain> I once installed stale-bot, but it was removed 😢
20:04:05 <andythenorth> but that caused angst
20:04:22 <TrueBrain> many projects run a stale-bot 🙂
20:04:26 <TrueBrain> makes things look tidy 🙂
20:04:58 <frosch> however, i ignoring stale PRs. there is a way too high percentage of drama when PRs are rejected
20:05:01 <LordAro> hides what actually needs doing
20:05:09 <TallTyler> Arastais: If there's a PR you want but it's unfinished with just little things to fix (we have a lot in this category), feel free to adopt it and make your own PR to supersede the original
20:05:19 <LordAro> most people report issues then don't worry about it further
20:05:30 <Arastais> TallTyler: wouldnt i need the original pr authors permission?
20:05:33 <LordAro> but you do need to have people to keep on top of things, or you end up like github, with 47000 open issues
20:05:50 <andythenorth> 10063 - I have exactly that problem trying to make a new map. Too many industries crammed together on 'normal'. Not enough on 'low'
20:05:51 <TrueBrain> frosch: that is what stale-bot is for 😛 Removes the human 😄
20:06:17 <LordAro> maybe ask first and give them a day or two to respond, but it generally doesn't matter
20:06:23 <TrueBrain> Arastais: code your PR is under GPL, so .. it is like making a GPL contribution to their PR, after which you merge their PR under GPL 🙂
20:06:28 <TallTyler> Especially if it's been abandoned for a long time
20:07:54 <andythenorth> oh 10063 needs rebased
20:07:59 <andythenorth> non-chunky bevels 😮
20:08:07 <andythenorth> I tried the rebase, but failed 😛
20:10:36 <LordAro> andythenorth: conflict seems simple enough
20:10:45 <andythenorth> I am even simpler 😛
20:10:57 <LordAro> just wrap the new Map::ScaleBySize call in ScaleByWaterLevel
20:11:02 <LordAro> in the conflicted areas
20:11:27 <LordAro> can't guarantee there's anything else wrong, mind
20:11:29 <andythenorth> this is like when I say 'just draw some pixels' 🙂
20:12:01 <TrueBrain> owh, so you meant to say: someone should rebase 10063?
20:12:06 <TrueBrain> wording is important 😄
20:12:32 <andythenorth> someone should make a bot that randomly chooses who 'someone' is today
20:12:35 <andythenorth> then aliases their nick
20:13:01 <TrueBrain> I had that at several jobs I worked, where "someone" was assigned
20:13:11 <LordAro> not suitable to be merged, but might compile now
20:16:16 <andythenorth> let's see if it compiles....
20:16:54 <andythenorth> feel like it should be taken round the back and disposed of humanely
20:17:10 <andythenorth> `src/genworld.cpp:90:28: error: use of undeclared identifier 'MapSize'
20:17:10 <andythenorth> for (TileIndex t = 0; t < MapSize(); t++) {`
20:17:23 <andythenorth> and similar errors with 10063
20:19:00 <andythenorth> which would also be usable to fix the river slopes that TrueBrain reported as a bug
20:19:33 <TallTyler> You might be the right someone with art skills for water cycle stuff
20:20:15 <andythenorth> I did try extending the cycle in src
20:20:21 <andythenorth> didn't get very far
20:20:45 <dP> are there any unused colours to extend the cycle?
20:20:50 <andythenorth> we have 12 spare colours in the DOS palette
20:21:04 <andythenorth> a non-highlights water cycle could be done with 4-5 colours
20:21:22 <andythenorth> I didn't pursue it because I also wanted 3CC 😛
20:22:03 <dP> it may be worth to figure out some more general way to do custom palette animations
20:22:27 <andythenorth> well we could just switch to 32bpp
20:22:33 <Arastais> frosch: i think a good stale bot would just mark stale PRs with a "stale" label instead of closing it. That way they're easier to spot for maintainers yet don't have to be closed.
20:22:33 <andythenorth> and retcon the 8bpp
20:22:43 <dP> 32bpp can't palette animations :p
20:22:57 <andythenorth> dP: well....then we can just cycle the sprites
20:23:06 <andythenorth> I feel like we talked about this before 🙂
20:23:19 <andythenorth> I think it turned to lolz instead of commits 😛
20:23:29 <dP> cycling sprites seems to be a waste
20:23:43 <dP> custom shaders though ... 😅
20:24:02 <andythenorth> anyway 3CC is no longer needed, due to Variants
20:24:19 <andythenorth> can just choose the exact company re-colour you want
20:24:39 <dP> lol, now I know one usecase for variants :p
20:24:55 <andythenorth> the pinks are free
20:24:56 <dP> don't even need to rename the group :p
20:25:28 <andythenorth> I am going to treat your comments about Variants like my comments on large maps
20:28:19 <dP> hm... add a flag that treats recolour sprite a permutation for palette cycle...
20:29:15 <dP> though recolour sprites are in layouts, not useful for landscape
20:29:31 <dP> just add permutation option to every sprite?
20:30:36 <dP> actually, for 7924 just using 32bpp may be enough
20:30:56 <dP> as you can use mask layer and darken the palette color with rgb
20:31:55 <dP> and 8bpp blitter is kinda pointless anyway
20:32:20 <andythenorth> can you scope creep it to add ice in winter?
20:32:49 <dP> and 8bpp can just have all depths the same colour, not the worst compared to it other quirks
20:33:02 <TallTyler> Ice should be objects that change sprite based on date, then you can have icebreakers
20:33:19 <TallTyler> Also I think I just invented ship lanes/tracks all over again
20:33:27 <TallTyler> Let’s not get into ship pathfinding now
20:33:49 <Xarick> speaking of ship pathfinding...
20:34:45 <Xarick> wow, discord is so flashy
20:35:15 <andythenorth> we mostly delete the previews 🙂
20:35:16 <LordAro> andythenorth: didn't work :(
20:38:37 <Arastais> TallTyler: lmao speaking of...everytime i give a ship a route thats got stations really far apart, it pops up saying its lost every time it leaves a station, then after like 4 seconds fixes itself.
20:39:18 <TallTyler> That’s an age-old problem with ship pathfinding
20:39:26 <TallTyler> There’s at least one issue open about it
20:39:44 <TallTyler> Although apparently it’s been improved several times over the years and used to be much worse
20:39:45 <Arastais> yea ik, it jsut reminded me of it. So maybe reinventing the ship pathfinder is not such a bad idea
20:40:58 <andythenorth> there is a really nice looking ship pathfinder PR
20:41:00 <andythenorth> I didn't test it
20:41:49 <andythenorth> oh maybe it's not a PR
20:41:49 <andythenorth> maybe it was a gist or something
20:44:25 <TallTyler> Not even a gist, just a promise of a PR which I hope is fulfilled eventually
20:46:29 <LordAro> aw, still doesn't work
20:49:01 <LordAro> github ui is a terrible way to fix merge issues
20:49:50 *** sla_ro|master has quit IRC ()
20:50:49 <Xarick> coronaai is now the fastest to 5000 road vehicles
20:51:13 <Xarick> given you give him a 30Million loan
20:59:24 <Xarick> AIVehicleList().Count() I wonder how heavy this function is
20:59:40 <frosch> lordaro: do you need the preview build on 10063?
20:59:44 <Xarick> and there's tons of vehicles
20:59:56 <frosch> every build failure sends a mail to info@ and makes a popup on my desktop
21:00:11 <LordAro> i didn't put it there
21:00:35 <frosch> you did :p last october
21:00:38 <TrueBrain> guess we should make previews never fail or something, to avoid that spam 😛
21:01:13 <LordAro> well, i'll stop trying now anyway
21:06:37 *** nielsm has quit IRC (Ping timeout: 480 seconds)
21:08:48 <Xarick> oh, the Count() itself is fine, but the creation of the list all the time, however...
21:09:18 <glx[d]> but why recreate the list all the time ?
21:09:31 <Xarick> to ensure we didn't reach the max_roadveh limit
21:09:50 <Xarick> otherwise coronaai will just keep spamming stations endleslly when it can't even build vehicles
21:11:09 <glx[d]> init a variable on start with the size of the list, update the variable when adding/removing vehicle
21:11:31 <Xarick> i had an idea with group
21:11:32 <glx[d]> no need to recreate a vehicle list everytime
21:11:57 <Xarick> but then i have to send all new vehicles to the group, then it's just AIGroup.GetNumVehicles
21:13:18 <Xarick> unless... i can just count the DEFAULT_GROUP
21:13:22 *** WormnestAndroid has quit IRC (Ping timeout: 480 seconds)
21:13:27 <Xarick> need to check openttd code
21:15:06 <frosch> well, i can't find any gh api for discussion categories :p
21:15:33 <frosch> i guess it's a case for logging the requests of the browser
21:20:18 *** WormnestAndroid has joined #openttd
21:21:15 <frosch> i wonder which poor soul had to implement the emoji api
21:24:26 <TrueBrain> frosch: seems there is only an API to interact with discussions indeed
21:25:19 <frosch> sorry, i got lost in the emoji api
21:25:44 <frosch> looks like all emoji exist in a "man" and "woman" variant, which differ exactly in "short" and "long" hair
21:25:53 <frosch> including bearded-man and bearded-woman
21:26:15 <andythenorth> are we going to have a gender debate? 🙂
21:26:25 * andythenorth might have to watch YouTube for a bit
21:26:35 <andythenorth> we had pepe and femboys already in last 3 days
21:26:39 <TrueBrain> frosch: poor you .. talk to you in 2 days 😛
21:26:51 <glx[d]> Xarick: yeah use AIGroup.GetNumVehicles(GROUP_ALL, AIVehicle::VT_ROAD)
21:27:16 <glx[d]> group counts are cached IIRC
21:31:08 <glx[d]> AIGroup.GetNumEngines() use cached values too
21:32:59 <glx[d]> anyway recreating lists continuously is never a good idea
21:34:42 <glx[d]> almost 5k RVs in 1,5 year ?
21:35:15 <Xarick> it builds a bus in every town
21:35:31 <Xarick> doesn't require pathfinding, the roads already exist
21:35:41 <glx[d]> I never had that many (all types included)
21:36:43 <FLHerne> I have in a real game
21:36:50 <FLHerne> had to patch out the limit once
21:37:31 <Xarick> so, for every increase in vehicle
21:37:44 <FLHerne> on a large map if you play for a long time and build really thorough bus and mail networks in every town, 5000 is quite reasonable to hit
21:37:45 <Xarick> it completed an entire loop
21:39:26 <glx[d]> yeah building vehicle list and counting the elements is the first idea, but clearly not the best 🙂
21:46:55 <frosch> ok, timeout, it will be faster to add the categories manually 🙂
21:47:32 <TrueBrain> automation is not always the solution 😛
21:47:43 <Xarick> the ultimate test now! 14 Corona AIs
21:51:36 <andythenorth> frosch: shall we spend money on mechanical turk?
21:51:49 <andythenorth> get 5 people to classify, take weighted average of result 😛
22:09:47 *** keikoz has quit IRC (Ping timeout: 480 seconds)
22:15:20 <andythenorth> oof I rejoined midjourney discord
22:15:25 <andythenorth> won't be sleeping tonight then
22:19:28 <Xarick> reached station limit 😦
22:19:45 <Xarick> i bet it's the station limit this time, and not the orderlist limit
22:44:29 <andythenorth> ok I have to hide midjourney discord somehow
22:45:51 <andythenorth> risk of genuine addiction
22:47:40 <TallTyler> Huh, commit author's name can't have non-ASCII characters or commit checker fails
22:47:59 <TallTyler> How English-centric
22:49:08 <petern> I guess fingers hurting from being cold is better than not feeling them...
22:49:21 <TallTyler> Fixing actions is way above my pay grade 😛
22:49:26 <TallTyler> (in terms of skill)
22:49:35 <TrueBrain> it is not an action 🙂 It is a script!
22:53:35 <TallTyler> No idea how I would test any changes 🤔
22:53:43 <TrueBrain> its in the readme 🙂
22:54:49 <TallTyler> `A collection of both client-side and server-side hooks used around the OpenTTD git repository.`?
22:55:22 <TrueBrain> owh, not in that readme .. where were the instruction then? Someone wrote stories about how to add that locally
22:55:39 <TallTyler> The main question is probably, which non-ASCII characters are okay and which should legitimately be blocked?
22:56:09 <TrueBrain> owh, there is a Makefile with a `make test`
22:57:01 <TrueBrain> I do wonder where the instructions are, as someone has been pushing we should all install that locally for a while
22:58:10 <TrueBrain> so now, as it turns out, the newest developer never actually read the CONTRIBUTING.md file 😛
22:58:20 <TrueBrain> you are caught with your pants down! 😛 😄 🙂
22:58:41 <TallTyler> I did read it, though?
22:58:52 <TrueBrain> but ignored its content 😛
22:58:57 <TrueBrain> I am just trolling you, to be clear 🙂
22:59:13 <TrueBrain> I think nobody actually reads that file, tbfh
22:59:33 <TallTyler> I see the part about installing hooks, but I don't know where they run locally
22:59:44 <TrueBrain> if you follow that, and you make a commit locally
22:59:51 <TrueBrain> it tells you the same thing as the CI tells you
22:59:55 <TrueBrain> at least, that is the idea behind it
23:00:01 <TrueBrain> I just absolutely hate it
23:00:10 <TallTyler> Tells me where? In the git command line?
23:00:16 <TrueBrain> easiest I think is to just add a test-case in the test of the repo
23:00:28 <TrueBrain> yeah, after you close your commit message
23:00:31 <TrueBrain> it runs those scripts
23:00:36 <TrueBrain> and tells you in the CLI what is broken
23:01:55 <TallTyler> Ah, `ln` doesn't work on Windows
23:02:11 <TallTyler> So yeah, I guess I never followed that part of the instructions correctly
23:02:41 <TallTyler> Whenever I set up my build environment on this computer (2020?) I must have given up on that step
23:03:16 <TrueBrain> `git cat-file commit HEAD | sed '1,/^$/d' > tmp-commit-message`
23:03:25 <TrueBrain> `python check-message.py tmp-commit-message server`
23:03:28 <TrueBrain> is what is actually executed
23:04:08 <TrueBrain> not the easiest thing to fix, as it intentionally checks for non-ASCII
23:05:12 <TrueBrain> I have been wondering if we shouldn't just switch to "squash-only" for PRs; you can these days force that the PR title is used as commit message. That removes all of these commit-message checks basically
23:05:43 <TrueBrain> been doing that for all my other projects .. so much easier .. as you can just make small commits in the PR, fixing things as you go .. no longer the PR branch has to be pretty
23:05:49 <TallTyler> Actually, I think we just need certain parts of the extended ASCII characters, specifically 192-214, 216-246, and 248-255
23:06:30 <TrueBrain> I think the intention of that script is to prevent non-ASCII in the commit message itself
23:06:38 <TrueBrain> I think the author name was not within the intention
23:06:52 <TallTyler> The excluded characters between those ranges are multiplication and division symbols and won't be part of anybody's name, unless Elon Musk's children are better at coding than he is 😛
23:07:11 <TrueBrain> well, you also have other names in the world, which are non-ASCII 🙂
23:07:26 <TrueBrain> feels a bit weird that we check your name
23:07:34 <TallTyler> It may just be if you put "Co-authored-by" in the commit message, other commits by that author seem to pass since name isn't encoded in the commit message
23:07:38 <TrueBrain> "no no, we don't allow you to write your own name like that!!!" 😛
23:08:03 <TrueBrain> yeah, the script only checks the message
23:08:36 <TrueBrain> it basically disallows anyone from adding a rocket unicode icon in the commit message
23:08:44 <TrueBrain> which is just a bit sad that we are not allowed to do so 😛
23:09:11 <LordAro> check for valid unicode should be fine
23:09:45 <TrueBrain> issue is, it is a normal line in the commit message
23:09:54 <TrueBrain> if you only want it for author, it means we need to actually understand the message
23:10:19 <glx[d]> forcing squash might not be the greatest idea for complex PR
23:10:43 <TrueBrain> glx[d]: means no complex PRs! Which really works out for many projects 🙂
23:10:55 <glx[d]> imagine cmake PR in one commit 🙂
23:10:57 <TrueBrain> LordAro: check the link above; the current code is really simple 😛
23:11:15 <TrueBrain> glx[d]: I honestly don't see the issue 😉
23:11:28 <TallTyler> Maybe the trick is just don't use `co-authored-by`, just author seems to work
23:12:08 <TrueBrain> `if l.lower().startswith("co-authored-by"): continue`
23:12:10 <glx[d]> co-authored-by is a github thing
23:12:19 <TrueBrain> might be enough, that we just completely ignore those lines
23:13:08 <TrueBrain> worst case someone adds rocket unicode to their name ... 😛
23:46:30 <FruitSalad> Rubidium: It was the UI interface, it had reverted back to 1x where as previously it was set to 2x. Setting back to 2x restored it to how it was before. I'm not sure why the update reset the resolution, I don't remember that happening before with previous updates.
23:47:17 <FruitSalad> However I have noticed other resolution changes, such as news messages. They appear to be higher resolution.
continue to next day ⏵
