IRC logs for #openttd on OFTC at 2025-11-02
⏴ go to previous day
01:21:29 *** lobster has quit IRC (Read error: Connection reset by peer)
01:22:05 *** lobster has joined #openttd
01:47:29 *** wangli has quit IRC (Quit: User went offline on Discord a while ago)
02:33:12 *** Wormnest has quit IRC (Quit: Leaving)
03:54:15 *** gnu_jj_ has joined #openttd
03:57:49 *** gnu_jj has quit IRC (Ping timeout: 480 seconds)
04:36:55 *** Zathras has joined #openttd
04:40:29 *** Zathras_11 has quit IRC (Ping timeout: 480 seconds)
05:10:11 *** Zathras_4 has joined #openttd
07:54:27 <zanooda2000> talltyler: several-step dropdown seems not that great, at least for dropdown lists
07:54:27 <zanooda2000> but renaming, hiding, sorting and filtering on dropdown list through menu seems a great alternative
07:54:27 <zanooda2000> I'd say that those actions shouldn't be instant, but rather done by some button in menu, so you can have different lists in dropdown and menu, like bare minimum in dropdown to make it quick and full eyecandy lin menu
08:12:51 <andythenorth> If there’s a good dropdown way to do it, could we also move the station picker to it?
08:14:15 <peter1138> Hmm, time to leave IRC as well.
08:14:45 <LordAro> peter1138: don't leave me here on my own
08:21:14 <Rubidium> hmm... what news did I miss?
08:27:59 <LordAro> i mean, there's a reasonable chance he's just gone on a bike ride
08:28:31 <peter1138> Sadly not, back is playing up.
08:32:10 <alfagamma7> My back's paining like hell every single hour
08:33:46 <zanooda2000> andythenorth: Yeah, such a great comparsion of at most 30 tracktypes of which you probably will use like 7 and hundreds of station/object elements which are also should have preview on each.
08:37:28 <zanooda2000> Hmmmm... but adding multi-layer collapsable lists to the object picker can make it better
08:48:06 <kuhnovic> peter1138: All those pitchforks eventually take a toll on you
08:48:31 <andythenorth> zanooda2000: fair argument, wrong facts though
08:48:39 <andythenorth> JP+ already has 63 railtypes
08:48:55 <andythenorth> and there are spec changes that possibly make hundreds more railtypes available to browse
08:50:00 <andythenorth> if the available railtype limit is increased to 65536, then the dropdown is not going to work well
08:56:01 <reldred> andythenorth: I dare say most UI's are going to come apart pretty quickly with 65536 railtypes.
08:56:33 <peter1138> Badges do already show in the railtype selector, by the way.
08:57:08 <peter1138> (But of course there's no filtering like in the vehicle window)
08:58:33 <reldred> That said, I feel like some problems with the current UI could be solved by better icons in the purchase menu. Compare JP+ Roads alpha from the github with JP+ Tracks. I trust JP+ Tracks will get a similar treatment in the future:
08:58:34 <andythenorth> for comparison, Horse has 8202 trains currently, and people do use it with other sets
08:58:54 <andythenorth> we're in the 'thousands of entries' type range
08:59:07 <rito12_51026> talltyler: made the last 2 recently-used rail types show on the top of the dropdown
08:59:58 <reldred> andythenorth: one thing I would say to that is, I usually am working a bit slower and more thoughtfully when selecting a train, when I'm trying to select a railtype to build I'm generally trying to get about things a bit quicker.
09:00:17 <andythenorth> I didn't give any thought to my original sketch other than 'dropdown fails', but we already have a pattern for all / used / saved
09:01:18 <reldred> I earnestly believe the solution to almost all problems in OTTD are badges.
09:01:28 <reldred> they really are quite neat
09:02:03 <peter1138> Can we please address what the fuck mac OS is doing to the custom truetype fonts...
09:02:41 <reldred> throw macs in the river?
09:02:55 <reldred> probably shouldn't do that
09:03:31 <andythenorth> what is it doing to them?
09:03:41 <andythenorth> I like the new font
09:04:10 <andythenorth> I have zero nostalgia for this
09:04:32 <reldred> lol, I'm still using pixel font
09:04:47 <reldred> but currently I'm back on pixels
09:05:20 <peter1138> i.e. match the dimensions of the bitmap font.
09:05:40 <peter1138> andythenorth, try going back to 1.8 instead, I hear that's where all the development is happening.
09:05:59 <andythenorth> ok so something is screwy in line height maybe
09:06:19 <andythenorth> I know nothing about font technology
09:07:08 <peter1138> Ok, so before we go smashing up the rail/road type selectors.
09:07:17 <peter1138> How many but-ground-types do we need?
09:07:57 <peter1138> Do we end up with a huge list for that too?
09:09:35 *** eratonysiad has joined #openttd
09:11:08 <eratonysiad> I can probably fill 16
09:11:35 <reldred> ISR gravel, the three main chips ground types, urban concretey looking stuff, a couple ballast variations.
09:12:02 <eratonysiad> Beach sand, desert sand, rocky landscape, snow
09:12:16 <eratonysiad> Maybe some transitions
09:13:45 <reldred> but this for instance wouldn't need to exist
09:13:53 <peter1138> Sounds like you're talking about terrain.
09:14:11 <peter1138> BGT is for underneath rail/road tiles.
09:14:25 <reldred> with a regular narrow ballast shouldn't need to cover various sand/desert types.
09:15:29 <peter1138> Though mostly road doesn't really need it.
09:15:45 <reldred> But if I just do a quick pass through my current jp+ setup it would be seven
09:16:18 <reldred> Two ballasts, urban, a couple of concretes, dirt, and the typical ISR gravel.
09:16:23 <peter1138> I did have some ideas with someone on Discord but as I'm no longer on the Discord I don't have those ideas any more :P
09:17:29 <reldred> I think in terms of the broad categories <ground><track><catenary><fence> would probably suffice. Any more layering than that is probably a bit daft. The JP+ 'universal undergrounds' would just be a custom catenary.
09:17:53 <reldred> I'm sure someone will want custom foundations but ehhhhhhhhh
09:18:33 <reldred> Maybe not even catenary tbh.
09:18:45 <reldred> catenary doesn't make sense with maglev or monorail
09:19:05 <reldred> though vac trains abuses catenary in new and exciting fashions.
09:19:07 <peter1138> Catenary is always part of the railtype.
09:20:14 <reldred> Probably for the best. I'd imagine any hypothetical new system really can't fundamentally mess with how railtypes actually work
09:32:13 <peter1138> andythenorth, have you completed the de-animation in FIRS now?
09:32:32 <peter1138> I have that version you sent but I think it was 3/4s done.
09:33:29 <peter1138> I also wonder if I should add my visualise-animated-tiles patch as a proper debug/highlight feature.
09:34:40 <andythenorth> peter1138: 65536
09:35:24 <andythenorth> I was trying to work out the spec while driving to kids rugby 😛
09:36:48 <andythenorth> I think we’d need two UIs to it
09:36:48 <andythenorth> * a ground tile painter
09:36:48 <andythenorth> * a way to select the ground for objects, stations etc, pre-composed before placing
09:37:04 <reldred> Honestly I think only a dozen at most. JP+ at the moment is only using seven distinct ground types apart from the full width and regular width ballast (which would be eight). No other track set has as many as JP+
09:37:42 <andythenorth> Even water could rapidly grow to 10s of types
09:37:43 <reldred> if, hypothetically, fences got added as a user selectable you'd see at most half a dozen.
09:38:12 <andythenorth> Also nothing gained by limiting it, we’ve seen this movie before
09:38:47 <andythenorth> There was a TTDPatch compatible HEQS with limited vehicles 😠
09:39:00 <peter1138> Why do objects and stations need ground types? Objects come with their own. Stations use the ground type of the rail or road they're placed on.
09:39:32 <reldred> I don't think thjey do tbh
09:39:44 <reldred> Stations is abit of a hjoldover from the olden times
09:39:45 <andythenorth> Because object sets duplicate the objects frequently to give different ground
09:40:01 <reldred> and non-track tile stations I suppose
09:40:11 <andythenorth> It’s not necessary to deal with objects in bgt
09:40:24 <andythenorth> It’s just weird to not
09:40:52 <reldred> Road and rails are the nails that are sticking up at the moment.
09:41:03 <andythenorth> If we have a spec for ground, then we might as well put it everywhere, like badges
09:41:06 <reldred> stations and newobjects are an 'eh' for me.
09:41:15 <andythenorth> Unless it’s a performance issue?
09:41:17 <reldred> But I do agree with the sentiment
09:42:15 <peter1138> Okay, so the only (almost) universally free bits across all tile types are 4 bits in m8.
09:42:21 <peter1138> That's only 16 types, which is not 65536 types.
09:43:13 <peter1138> But I do not see us adding 2 bytes (16 bits for 65536 types) to the map array for something that most people will never use.
09:43:21 <andythenorth> Goes it use more RAM?
09:44:26 <reldred> Doesn't leave any room for other cosmetics but tbh the fence type usually always directly matches the ground type in track sets so that's probably not a problem. Urban concrete usually gets a concrete barrier wall, gravel usually always gets a chainlink fence, yadda yadda yadda.
09:44:43 <reldred> 65536 types tbh feels silly
09:44:48 <peter1138> Given 64-out-of-lots is seen as unacceptable for rail/road types, doing 16 out of 65536 is not going to fly for either.
09:44:59 <andythenorth> FTR I’m not convinced by BGT for personal use, but there’s a lot of model railroad playing
09:45:17 <reldred> I would use the everloving shit out of BGT personally.
09:45:20 <peter1138> Fences are part of railtypes because only rails have company fences.
09:45:29 <andythenorth> I wondered about precomposing instead
09:45:53 <andythenorth> So consume IDs for railtypes/stations/objects
09:46:10 <andythenorth> But with a better UI to it
09:46:22 <peter1138> And fences are mutually exclusive with snow/desert state (uses the same bits)
09:46:24 *** super_vin has joined #openttd
09:46:24 <super_vin> Hi, I've looked everywhere but can't find the answer. I hope I'm in the right forum to ask the following question:
09:46:24 <super_vin> I found a typo in one of the French strings, but the French translation team is complete! Is there a way to report the mistake?
09:49:21 <andythenorth> Hmm station sets might have 1k tiles though?
09:49:21 <peter1138> andythenorth, FIRS s5.0.1 released?
09:49:30 <peter1138> So is FIRS 5.0.1 released?
09:49:54 <andythenorth> Yes 5.0.1 is, doesn’t have the tile de-animation though
09:50:21 <peter1138> Maybe it's not compatible, dunno.
09:50:23 <andythenorth> I’m afk on my phone or I’d send it
09:50:44 <andythenorth> There are about 8 indistries left to de-animate
09:51:19 <andythenorth> It wil need a savegame bunp before release
09:52:39 <reldred> I hope my network termination device turns up tomorrow. I'd quite like some internets I think.
09:52:43 <peter1138> It's neat to watch with animation visualisation now.
10:01:10 <peter1138> Hmm, though even the current release does start and stop some animations.
10:01:27 <peter1138> Just the unanimated tiles were always left started.
10:05:55 <peter1138> It is not exciting. It is just doing as expected.
10:07:44 <peter1138> Needs to be transparent because the highlight is drawn underneath the buildings.
10:08:28 <peter1138> I have wondered if tile highlights should be drawn on top.
10:12:04 <xarick> can I clone a ScriptPriorityQueue ?
10:14:04 <peter1138> IIRC the other author who uses animation state for various things doesn't actually publish their NewGRFs.
10:16:12 <xarick> why isn't priority queue using SQInteger?
10:17:54 <xarick> i suppose int is sufficient for my needs
10:19:07 <andythenorth> I’d be surprised if all the FIRS forks haven’t just cargo culted animation 😏
10:20:22 <peter1138> Let's see, what's forked off FIRS?
10:20:46 <reldred> that's all that comes to immediate mind
10:20:55 <peter1138> AIRS, yup, all tiles are animated for no reason.
10:20:57 <reldred> xis is forked off firs3, the other two off 4
10:22:38 <peter1138> Now, did they fork it properly, or did they just take the precomposed NML...
10:23:42 <emperorjake> Both XIS and AXIS are proper forks using the python compile
10:23:53 <emperorjake> XIS is a mess though so it had to be redone
10:27:14 <peter1138> FIRS animation is being fixed, it would be good for the forks to copy it.
10:27:32 <peter1138> Anything with Auz in the name I just assume is an eyecandy object set :p
10:27:54 <andythenorth> Maybe parts of SPI
10:28:09 <andythenorth> There are 5 or 6 I think
10:30:22 <reldred> Garry's still around but he's nowhere near as active as he used to be. Fella's getting on in years.
10:31:38 <peter1138> I don't think he uses animation at all.
10:36:01 <emperorjake> Some Auz objects do have animation
10:36:29 <peter1138> But an object set is unlikely to have been forked from FIRS.
10:38:21 <peter1138> PIRS starts off with every tile animated, and then they immediately become unanimated.
10:38:48 <reldred> I thought PIRS was a totally seperate codebase? I didn't think it was a FIRS fork
10:39:08 <peter1138> It is, I'm just looking at things with my animation highlight :)
10:45:14 <andythenorth> FIRS probably has lots of needlessly animated objects also
10:50:48 <xarick> something strange happening with GS.
10:51:19 <xarick> the GS hit a breakpoint and pauses the game. But from time to time, i get a hiccup
10:51:30 <xarick> even with the game paused
10:52:34 <xarick> garbage collector or something?
10:53:31 <xarick> gonna investigate what it is
11:08:00 <xarick> how does Collect Garbage work?
11:21:33 <xarick> hmm but the game is paused, by a GS breakpoint
11:21:42 <xarick> why run the collector?
11:22:12 <xarick> also, why is the collector somehow repeating the hiccups even paused
11:25:30 <xarick> 2.554 seconds hiccups every 255 frame_counter
11:37:41 <rito12_51026> Forked from AXIS
11:46:24 <andythenorth> Shall I make an industry object planter in GS?
11:46:46 <andythenorth> And can we embed GS in a grf yet?
11:59:14 <xarick> garbage collector was screwing my benchmark results 🙂
12:11:55 <andythenorth> So if the game could precompose 100 railtypes with 100 groundtypes, that’s only 10k IDs used?
12:17:12 *** gelignite has joined #openttd
12:26:41 <xarick> question: what will happen to openttd if the garbage collector is never run?
12:28:44 <kaji_kaede> xarick: Somewhere between nothing and a memory leak.
12:28:48 <LordAro> possibly it'll use up all the memory on your system
12:30:08 <xarick> but these hiccups even when the GS hits a breakpoint and pauses
12:30:29 <xarick> any possibility to not run the collector during pause?
12:30:45 <xarick> but still run the GS normally?
12:30:56 <xarick> which does nothing cus its paused
12:32:57 <LordAro> not in a safe way, i suspect
12:38:08 <Rubidium> IMO the maintenance of a feature that stops GC when a breakpoint is hit, is not worth the negligble gain. You boast 2.5s, but you'd hit that in the game every 10 seconds making your AI/GS already unplayable.
12:47:22 <_jgr_> The GC is only taking a noticeable amount of time because the "test" is generating an absurd amount of crap
12:48:16 <_jgr_> You could disable or defer the GC in your own test build if you wanted to though
12:48:27 <andythenorth> I don't understand a method of testing which is 'generate unrealistic load, then optimise for it'
12:48:50 <andythenorth> it's meaningless
12:49:14 <andythenorth> is it just a meta-game?
12:55:32 *** ChanServ sets mode: +v tokai
13:00:56 <Borg> andythenorth: well said.. well said
13:01:01 <mmtunligit> codingstyle.md told me to commit in logical chunks and i am doing my best to adhere to that
13:01:15 <Borg> mmtunligit: I hope you later squash those?
13:03:06 <mmtunligit> yeah i will, either when asked or when review is done, at least thats how it worked with my previous PR
13:04:41 <mmtunligit> ive still got one more fix to do that i just noticed, highlight doesnt swap properly with stations
13:14:34 *** gelignite has quit IRC (Read error: Connection reset by peer)
13:16:50 <peter1138> Best to squash fixes, otherwise the logical chunks get lost.
13:18:37 <mmtunligit> peter1138: good to know, thank you
13:19:23 <mmtunligit> whats the point of the Codefix keyord then, if those changes are supposed to just get squashed
13:20:30 <_glx_> It's for when the PR has already been merged
13:21:41 <mmtunligit> thanks! ill get those squashed in now then
13:22:40 <peter1138> Hmm, so, quite often NewGRFs use animation frames as a way to update if a nearby tile is changed.
13:23:12 <rito12_51026> peter1138: Isn't it better to edit commits using interactive rebase?
13:23:39 <andythenorth> peter1138 to change sprites?
13:24:38 <peter1138> andythenorth, to do things like rotate towards a road.
13:25:03 <andythenorth> I wonder if that's because the nml wiki pages encourage it
13:26:04 <peter1138> It's better than just changing on draw, which can leave things in a partially drawn state.
13:26:56 <andythenorth> I'm trying to find the page that recommends it
13:27:23 <andythenorth> there's something about using multiple realsprites, then using animation frame to select one
13:27:35 <peter1138> I doubt there's anything in the spec that says you should rotate buildings to cater to the roads.
13:27:37 <andythenorth> and then setting the frame according to [conditions]
13:27:53 <andythenorth> I might have imagined it though
13:41:44 <mmtunligit> theres a few things that will be slightly weird as a consequence of my original "logical chunks" including fixes of the same type for multiple different commits
13:41:50 <mmtunligit> but now i know better
13:51:18 <andythenorth> dunno what the final effect of FIRS de-animation will be
13:52:41 <andythenorth> but I am guessing the fraction of tiles that needed to be animated is about 1/4
13:52:49 <alfagamma7> wdym by deanimation
13:52:59 <andythenorth> maybe 1/6 in some cases
13:53:11 <alfagamma7> so no repeating animation in say the steel mills or something like that?
13:54:08 <andythenorth> every tile here was animated
13:54:14 <andythenorth> but only the 3 smoke tiles need to be
13:54:18 <peter1138> One of them has 2 animated tiles instead of 18.
13:55:32 <andythenorth> should I test the extended cargos patch?
13:58:34 <andythenorth> should I finish Iron Horse 4?
14:01:44 <alfagamma7> andythenorth: I used to think the furnace, the smoke stacks were animated in the mill?
14:02:42 <alfagamma7> Eagerly waiting for IH4
14:02:45 <peter1138> andythenorth, smoke animation on the Soda Ash Mine is "quite fast"
14:04:27 <alfagamma7> or Central Euro Horse?
14:04:58 *** gelignite has joined #openttd
14:04:59 <peter1138> Which "mill" did you mean, alfagamma7 ?
14:05:00 <_glx_> xarick: in the doxygen all integer types are shown as int (but in the code it's SQInteger)
14:05:28 <alfagamma7> in the default Temperate economy?
14:06:47 <peter1138> Okay, you just said "mill", and FIRS has loads of different mills depending on mode.
14:07:25 <_glx_> xarick: nobody implemented ScriptPriorityQueue cloning (nor save/load)
14:07:27 <peter1138> Ah you did say steel mill earlier.
14:07:39 <peter1138> Hmm, Steel Mill in my copy is still all-animated.
14:09:36 <andythenorth> this should have most of them gone
14:10:50 <andythenorth> there are 4 that are not, because reasons
14:11:43 <andythenorth> it has some fancy tile location conditions....which aren't working
14:11:53 <andythenorth> so I need to fix those before applying them to more tiles 😐
14:16:29 <andythenorth> so many projects eh 🙂
14:16:50 <alfagamma7> imagine the scheduling problem
14:16:57 <andythenorth> hmm do cargos have badges?
14:17:12 <andythenorth> hmm do cargo subtypes have badges 😛
14:17:21 <andythenorth> I want to split up Food
14:17:53 <_glx_> anyway I don't think clone/save/load of ScriptPriorityQueue will happen, due to extra handling required for HSQOBJECT
14:18:12 <andythenorth> what if cargos could have namespaces?
14:18:22 * andythenorth appropros of nothing at all
14:18:36 <andythenorth> 64 types of food
14:18:37 <andythenorth> 64 types of steel
14:18:56 <andythenorth> 1 byte for the namespace selection
14:24:54 <peter1138> There is one namespace, the label.
14:27:12 <peter1138> 15.0 at some point.
15:02:04 <kuhnovic> 16.0 is the new 15.0
15:32:51 <andythenorth> we could have had labels for the namespaces 😐
15:33:00 <andythenorth> such wiki pages could have been created
15:33:20 <andythenorth> I guess there's no font for "I am telling lies"
15:33:29 <andythenorth> maybe there should be
16:10:34 <peter1138> Animated level crossings?
16:13:23 *** yiffgirl has joined #openttd
16:13:23 <yiffgirl> oh, irc probably doesn't bridge reactions. that was an eyes emoji.
16:20:28 <andythenorth> peter1138 I misread that as Animal Crossing
16:20:36 <andythenorth> but then Nintendo would be after us
16:34:24 <xarick> i think I found a bug in jgr's safe btree stuff
16:34:37 <xarick> need to test his own build without my interference
16:40:17 *** Wormnest has joined #openttd
16:50:53 <tony_pixel> peter1138: animal crossing?
16:57:24 <xarick> nope, no crash on jgr :|, so it's me
16:59:59 <andythenorth> did we finish the game yet?
17:00:22 <xarick> somewhere about `_nexti` and `_get` is corrupted on my adaptation
17:02:10 <xarick> my initial adaptation which was very much a verbatim copy paste from jgr also does not crash
17:02:18 <xarick> so it's my recent changes :/
17:17:01 <xarick> ah, found the problem, 😛 it was me
17:17:38 <xarick> my own SetIterValue using removed iterators
17:20:08 <peter1138> (When JGRPP does have bugs, they get fixed pretty quickly. So it's unlikely you'd be finding them.)
17:20:17 <andythenorth> surprised it wasn't caused by FIRS
17:22:30 <_jgr_> I fixed an 11 month old bug earlier today. It would be nice if everything did get fixed quickly :P.
17:27:39 <peter1138> That's enough that everyone just assumes it's meant to be that way.
17:29:28 <xarick> oops, it wasn't SetIterValue
17:29:57 <zanooda2000> andythenorth: can be fixed by having colors as refit. Will be somewhat less convinient, but way better than having 99 different colors as different vehicles. Even better with sub-groups
17:29:57 <zanooda2000> And so you can recolor your 100 years old steam engine.
17:30:06 *** gelignite has joined #openttd
17:30:43 <xarick> is there a difference between passing `{}` and `this->values.end()`
17:31:56 <xarick> `this->sorter->PostErase(new_item_iter, {});`
17:33:11 <andythenorth> yes, Variants were a mistake, everyone should switch to the new subtypes system
17:33:22 <andythenorth> livery refit callback
17:33:48 <andythenorth> I'd much rather build my vehicle, then hunt through a complicated refit menu
17:34:06 <jfkuayue> Do I have rights to use refits without getting laughed?
17:34:08 <talltyler> Is this where you remind us that “time is a flat circle”?
17:34:31 <andythenorth> what's old is new again?
17:34:50 <alfagamma7> I used to think time is like a good old tape cassette
17:34:53 <talltyler> Those who don’t understand development history are doomed to repeat it
17:35:08 <alfagamma7> talltyler: Amen to that
17:37:12 <andythenorth> tapes can be cut to continuous loop
17:37:27 <andythenorth> interesting medium
17:37:56 <jfkuayue> As far as I know, quite a few players love those BAD FEATURES like cargo aging, reliability decay, and refit
17:38:39 <jfkuayue> Would Andy say “something is lost in translation”?
17:39:07 <andythenorth> dunno, grf isn't a contest with one winner
17:39:12 <andythenorth> other grfs exist
17:40:08 <jfkuayue> But Andy would joke on us…
17:41:00 <alfagamma7> Grfs are just fruits of loving labour
17:41:12 <alfagamma7> Kinda like this game itself
17:41:44 <alfagamma7> If I am being too pedantic , I really should pinch myself
17:43:07 <talltyler> I think that trying to achieve perfection or standardization across GRFs is a quick way to suck the fun out of the process. My favorite GRFs are consistent with themselves but unashamedly do their own thing compared to other GRFs.
17:43:25 *** belajalilija has joined #openttd
17:43:58 <belajalilija> i will not conform to standards unless they are my own
17:44:47 <jfkuayue> what about Standardised (insert anything) Scheme?
17:44:56 <belajalilija> i will however make a grf so vast that i will force a standard within my niche
17:45:01 <andythenorth> would docks need groundtypes? 👀
17:50:50 <andythenorth> ok so badges are static action 0
17:50:58 <andythenorth> but what if GS could attach badges to tiles?
17:51:11 <andythenorth> what if players could attach badges to tiles
17:51:16 <talltyler> I don’t think we have the map bits for that
17:51:18 <jfkuayue> I will do more China Set stuff if my academic stuff is not urgent and if I don’t want to play openttd nor airlinesim and if I don’t want to go out to enjoy real life and if I am not surfing discord and…
17:52:18 <andythenorth> wonder what clock speed GS could have on redstone tiles
17:52:23 <kaji_kaede> jfkuayue: ...How is cargo aging a bad feature?
17:52:31 <andythenorth> it's a terrible pointless feature
17:52:39 <andythenorth> see logs or something
17:52:41 <andythenorth> debated to death
17:53:05 * peter1138 tries to work out if he needs to take care of iterator invalidation.
17:53:32 <kaji_kaede> ...Apparently it hasn't been debated to death enough.
17:54:03 <andythenorth> would my redstone GS be faster if I copied the map into RAM?
17:54:10 <andythenorth> I am moderately serious btw
17:54:22 <andythenorth> process redstone actions in the event loop
17:54:25 <jfkuayue> Different speed of cargo aging are thought by many people as “useless”
17:54:29 <peter1138> I would assume that the map is already in RAM.
17:55:16 <andythenorth> jfkuayue: when you can demonstrate an empirical effect that is (1) reliable and (2) explainable to players, please write a gist or forum post?
17:55:42 <andythenorth> I should do more with the story book
17:56:52 <kaji_kaede> jfkuayue: It seems to make a lot of sense to me as a design question for a network.
17:56:52 <belajalilija> if there's ever proof that cargo aging has a real and noticeable impact i will include it in my set
17:57:31 <andythenorth> old font in that video
17:57:44 <kaji_kaede> It might be worth having higher-speed freight if it means my food isn't moldy by the time it gets to customers. Gives more to think about in how to handle those kinds of trains - if using a GRF that encourages it.
17:59:27 <andythenorth> hmm so redstone objects
17:59:37 <andythenorth> TBH I could do redstone with objects anyway
17:59:39 <kaji_kaede> Well. If it's 1885 that advice doesn't help much.
18:00:37 <jfkuayue> Then how could I do long distance on-land transport in a practical way…
18:00:52 <andythenorth> just don't play ridiculous maps
18:01:13 <talltyler> I’d like to see more quests given by storybook, could be fun to flesh out Busy Bee or Silicon Valley, or whatever Pikka remixed that we played in multiplayer:)
18:01:38 <talltyler> Express freight trains used to be a thing 🙂
18:01:50 <jfkuayue> We want to play ridiculous maps, and it has been proven that we can achieve a feasible way using cargo age period, though
18:01:56 <peter1138> "But you can only have one GameScript!!111!!!"
18:02:14 <andythenorth> it's not even technically impossible to have n GS
18:02:20 <andythenorth> libs are supported
18:03:13 <michi_cc> jfkuayue: You don't in 1885 is the simple answer. Choose an appropriate game year and you can have non-sucky land transport.
18:04:04 <kaji_kaede> Or fund industries so the only bakery in the country isn't half a continent away.
18:04:09 <andythenorth> I remain open to making refrigerated vehicles age slower, but no player or author has been able to explain what the effect is
18:04:52 <kaji_kaede> Intuitiveness is definitely a major issue right now.
18:05:09 <jfkuayue> The British railway speed in 1885 would be the Japanese speed in 1955, and Chinese speed in 1980
18:07:38 <andythenorth> the effect of train prop 0x2B
18:07:38 <andythenorth> needs to be related to cargo props 0x10, 0x11, 0x12
18:07:39 <andythenorth> in a way that is capable of being explained to players in one line of text
18:07:51 <andythenorth> I didn't manage it
18:08:52 <jfkuayue> My original stance was, we have rights to enjoy custom cargo aging speed without being laughed at
18:09:49 <andythenorth> guilty as stated
18:11:00 <jfkuayue> Or we are happy with it
18:11:13 <belajalilija> andythenorth: same type of thing really, currently my set uses loading speed quite heavily and gives metro type trains high speed loading and also capacity for standing to encourage use of those vehicles, suburban coaches have a similar thing going on too but more a half way point between normal and metro type trains
18:11:33 <peter1138> Why do you care with andythenorth laughs are you?
18:13:11 <peter1138> Speaking of which, have you managed 90mph now, andythenorth?
18:13:35 <andythenorth> my ebike does not go that fast
18:14:11 <andythenorth> to be clear, I don't understand cargo aging either
18:15:10 <andythenorth> but dP has charted a lot of it for citymania, and afaik, the problem tends to be a collission between 'cargo aging is useful on large maps' and the fact that there is a cutoff to aging, where only distance is relevant
18:15:30 <jfkuayue> andythenorth: Applied to the old cargo aging graph, not new ones
18:15:37 <andythenorth> there are a narrow band of circumstances where it will behave usefully
18:15:45 <talltyler> The cutoff is gone now 🙂
18:16:22 <talltyler> That cutoff is why your previous tests showed it didn’t do anything useful.
18:18:10 <jfkuayue> The typical way of operation of 1950s-1970s JNR (and China Railway later) may fall into the narrow band, though, with long distance slow trains
18:18:22 <andythenorth> oh I ignored 10596 because it has extensive Zorg
18:18:34 <andythenorth> Zorg is the primary reason I deleted cargo aging
18:18:44 <andythenorth> the secondary reason is I can't find text to explain it neatly
18:19:12 <andythenorth> Zorg wrote very long texts and kept bollockings on with savegames that didn't show what was claimed
18:19:28 <andythenorth> I spent a weekend testing cargo aging on FFWD for different route types
18:19:33 <jfkuayue> Can I say it is like Chinese herbal medicine…
18:19:35 <andythenorth> deleted everything to do with it
18:20:25 <andythenorth> if anyone can explain it reliably in one line of text for purchase menu (like a badge) I'll put it back on refrigerated vehicles
18:20:53 <andythenorth> but the text has to account for transfer routes
18:21:02 <andythenorth> and slow legs that have already aged the cargo
18:21:23 <jfkuayue> Practically, it is useful, so we use it, although not theoretically proven
18:21:31 <peter1138> Legs to take me to the pub.
18:22:06 <andythenorth> don't forget to wear shoes
18:22:09 <andythenorth> it's not summer any more
18:22:25 <andythenorth> slower legs on the way home?
18:23:16 <belajalilija> legs should also slow down as the vehicle ages
18:23:38 *** Flygon has quit IRC (Read error: Connection reset by peer)
18:23:43 <belajalilija> add leg pains (cargo aging) too
18:28:14 <xarick> why is this a conflict?
18:36:14 <rito12_51026> Why `ROADTYPE_END` is equal to 63 and not 64
18:36:14 <rito12_51026> and why `INVALID_ROADTYPE` has to be equal to `ROADTYPE_END` instead of `0xFF`?
18:36:14 <rito12_51026> Why is it needed for NRT to work properly?
18:39:22 <_jgr_> 0xFF will not fit in 6 bits, which is how wide the field is
18:42:47 <rito12_51026> But railtype is also 6 bits and `INVALID_RAILTYPE == 0xFF`
18:44:08 <peter1138> Road tiles have to store road type and tram type. If one of them is not present, then INVALID_ROADTYPE is stored.
18:44:21 <peter1138> Rail tiles don't need to do this.
18:45:27 <belajalilija> question that is probably stupid
18:45:46 <belajalilija> i had this idea yesterday but was just reminded of it
18:46:42 <belajalilija> currently we have road and tram types, and certain tram track types are just OHLE, for trolley busses
18:47:17 <belajalilija> how hard would it be to have a similar thing going on for rail? so there could be rail types, that are just plain tracks
18:47:26 <belajalilija> and power types, that add OHLE, metro or 3rd rail
18:47:37 <andythenorth> how do trolley buses use tram tracks?
18:47:52 <andythenorth> _is curious if there's a misunderstanding_ 🙂
18:48:23 <belajalilija> i mean, trolley bus ways i think are counted as a type of tram way are they not?
18:49:03 <kaji_kaede> I think it's specific "tram" road types that don't have actual rails.
18:49:22 <belajalilija> i know tram tracks can exist without roads, but the fact that they can coexist on the same tile is the idea im trying to get at
18:49:58 <talltyler> Your base understanding is wrong. Trolleybuses in OpenTTD run on electrified roads, not tram tracks.
18:50:15 <talltyler> (Although there was an early pre-NRT implementation that used invisible tram tracks
18:50:26 <andythenorth> when NRT was done, there was a bit of discussion about 'should it be power types'
18:50:52 <belajalilija> talltyler: that will be what i am remembering
18:51:01 <andythenorth> it's not something that can be changed course on at this point 🙂
18:51:11 <andythenorth> railtypes are railtypes
18:51:43 <belajalilija> just thought it would be a solution of sorts, im sure you can all see where i was coming from))
18:52:05 <andythenorth> 'would it have been better to have a tool to add power types to rails'
18:52:10 <andythenorth> is not a new idea
18:52:22 <belajalilija> yeah the same idea exists within locomotion
18:52:37 <belajalilija> which is itself 20 years old
18:54:07 <peter1138> Alright, so can FlowEdgeIterator be doing things at the same time as flows being modified?
19:47:53 *** gelignite has joined #openttd
19:53:25 *** Borg has quit IRC (Quit: leaving)
20:02:41 <xarick> **left**: me; **middle**: master; **right**: JGR'ish
20:03:04 <xarick> with garbage collector disabled
20:04:49 *** gelignite has quit IRC (Read error: Connection reset by peer)
20:08:55 *** gelignite has joined #openttd
20:09:30 <andythenorth> I've only skimmed it though
20:09:35 <peter1138> Yours is slower than JGRPP, so what's the point?
20:10:18 <peter1138> There's now two ways to select rail/road type. What.
20:10:30 <andythenorth> classic compromise
20:11:21 <peter1138> Two ways is a big fat no from me.
20:11:57 <andythenorth> well the pitchfork contingent are insistent that we can't change the dropdown 🙂
20:12:04 <andythenorth> so we have to have two ways
20:14:44 <xarick> mine uses std::map std::set
20:15:40 <peter1138> Okay, well. I still hate it.
20:15:52 <peter1138> The big green window is bad.
20:16:31 <peter1138> The drop down works because it's temporary and doesn't waste screen space once the type is selected.
20:17:37 <andythenorth> it's not a ditch I want to die in 🙂
20:17:45 <andythenorth> my game looks like
20:18:26 <andythenorth> no badges though
20:18:36 <andythenorth> [I didn't draw a sprite]
20:18:43 <peter1138> Add some badges to your railtypes, then it will, christ.
20:19:24 <andythenorth> what badge for 'high speed rail' 👀
20:19:30 <andythenorth> small orange TGV?
20:19:45 <andythenorth> I guess I have no idea how to fix this then 🤷♂️
20:20:25 <peter1138> Easy, let the eyecandy players have their own long list and don't worry about everyone else.
20:22:12 <peter1138> It's just so bloated.
20:22:37 <andythenorth> it's so colourful though
20:22:41 <xarick> do you still have your vector based scriptlist, peter?
20:22:54 <xarick> did u update it meanwhile, gonna check it out
20:28:05 <rito12_51026> peter1138: The dropdown is for choosing railtypes and the window is now for modifying what and how it is shown in the dropdown (ability to select railtype from it is just an feature I haven't removed as removing code also takes time 😄 )
20:32:33 <andythenorth> if we increase the cargo limit, what are the main food cargos I could add to FIRS?
20:35:50 <xarick> okay, gonna test scriptlist vector by peter
20:37:53 <xarick> is it a single vector? woah
20:42:27 <xarick> ah, there's a vector of indexes hidden
20:43:01 <peter1138> Yup, it's going to suck for 10,000,000 items.
20:43:37 <peter1138> It also behaves differently when iterating if the list is modified.
21:00:48 <xarick> for each i in add / list.AddItem, it's struggling a lot here
21:00:48 *** tokai|noir has joined #openttd
21:00:48 *** ChanServ sets mode: +v tokai|noir
21:01:28 <xarick> "add" is an array of [key, val] items
21:01:59 <peter1138> Yes, it sorts as it adds, which is fine for normal loads but will suck with 10,000,000 items./
21:03:01 <xarick> ok, let me test something moderate
21:05:35 <xarick> meh, too quick for 10 000
21:05:49 <mmtunligit> peter1138: i have had thoughts along this line, if the people who are actually running into the issue of the list being too long *and* actually bother to use a good chunk of the list are all eyecandy jgrpp players, maybe we should just solve it over there
21:06:40 <peter1138> How many eyecandy players actually have a problem with, e.g. the JP+ tracks list?
21:07:08 <peter1138> Isn't it what they expect? A lot of choice.
21:07:50 *** tokai has quit IRC (Ping timeout: 480 seconds)
21:07:53 <mmtunligit> being able to sort throught he choice easier would be nice, especially if the limit gets increased, the current dropdown is fine for what we have but if it expands anymore it wont be
21:08:41 <peter1138> The sort order in JP+ is chosen entirely by the NewGRF authors.
21:09:00 <peter1138> Presumably they have a reason to have it ordered the way they have ordered it.
21:09:09 <mmtunligit> and i think im not alone in salivating over the idea of being able to pick through every possible permutation of gauge, ballast/groundsprite, electrification, catenary style and fence
21:09:19 <mmtunligit> its not the order thats the issue
21:09:30 <mmtunligit> if there are more types its the finding stuff
21:09:49 <mmtunligit> single scroll lists will become cumbersome to use, no matter the sorting
21:10:02 <peter1138> "There's not enough stuff" "Adds stuff" "There's too much stuff"
21:10:51 <mmtunligit> we arent complaing that theres too much stuff, we're complaining that its difficult to sort through. if object selection was just one giant dropdown single scroll list thatd be atrocious
21:11:30 <mmtunligit> or, will be difficult to sort through if the limit increases
21:21:28 <xarick> apples to apples, 1 million items, yeah it's kinda ruggish
21:23:23 <andythenorth> the main reason to increase the limit is to deconflict grfs IMHO
21:23:56 <andythenorth> like...currently I might have to try and detect JP+ from Horse, because Horse breaks JP+
21:24:03 <andythenorth> [depending on grf load order]
21:24:36 <andythenorth> whether that really matters 🤷
21:29:28 <peter1138> As usual, the surprising thign is that a vehicle NewGRF contains railtypes.
21:35:29 <andythenorth> which is probably why we shouldn't increase the railtype limit
21:37:35 <andythenorth> or do a new railtype spec
21:39:56 <xarick> peter's favourite container
21:41:18 <xarick> any alternative for emplace?
21:52:44 <peter1138> There is a reason I didn't do anything like make a PR.
21:54:37 <reldred> BGT I think would put more than 64 railtypes to bed for quite some time.
21:55:45 <reldred> peter1138: horse really is the edge case here, for almost any other person wanting to mix rail grf’s outside of specific combinations that were made for each other we’d generally tell them to just ‘don’t do that’
21:56:50 <reldred> That said, BGT would take JP+ tracks and reduce it down to like sixteen tracks, twenty at most.
21:57:57 <reldred> BGT personally I feel solves a lot more problems
21:58:13 <andythenorth> we don't have the map bits though
21:58:15 <reldred> Even if it’s only four bits
21:58:49 <reldred> How were we going to get the map bits for more railtypes?
22:01:18 <reldred> Even if say hypothetically we grab another byte for implementing BGT with say, four or five bits, maybe leave an extra couple bits so fences can be split off of desert/snow (for instance), I think that dodges more problems with the UI in OTTD long term than throwing another 8 bits straight to railtypes themselves.
22:01:50 <andythenorth> we could....reduce the number of railtypes? 👀
22:01:54 <andythenorth> what's the fences thing?
22:02:29 <reldred> Oh Peter knows what I’m on about, I’m just spitballing as an example of what the extra bits could be used for.
22:03:15 <reldred> andythenorth: At most we could nick one from railtypes but then we still need to find a few more to implement BGT anyway.
22:05:26 <reldred> 64RT+16BGT I think still gives some extra slippage (though the UI goes to shit) if someone does something dumb and loads multiple rail sets, but BGT also allows each rail set to remove three quarters of their rails. You get a lot of deduplication out of BGT and the UI doesn’t need to be terribly complex for it. Whereas unfucking the dropdown list for 64 or more railtypes is gonna be messy.
22:06:08 <reldred> Anyways, just some morning shower thoughts 🙂
22:07:28 <andythenorth> I think BGT would cut JP+ down to about 53 types
22:08:00 <reldred> God no, way more than that
22:08:29 <andythenorth> most of it is sleeper types
22:08:53 <andythenorth> that could be done as a BGT hack
22:08:57 <andythenorth> but would need to combine with ground combinatorially
22:09:07 <andythenorth> so it would be 'snow + wooden sleepers' etc
22:14:07 <andythenorth> snow is a terrible example 😄
22:41:29 *** keikoz has quit IRC (Ping timeout: 480 seconds)
23:18:04 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
23:32:53 <xarick> i finally understand the retarget iterator stuff
23:33:45 <xarick> my std::map std::set implementation doesn't handle that correctly :((
23:35:11 <xarick> triggered by ScriptList::SwapList(list) after a ScriptList::Begin() call
23:36:36 <_glx_> I'm quite sure `ScriptList::CopyList()` after `ScriptList::Begin()` might have troubles in current code
23:40:48 <_glx_> ah no, it's fine, it's only used when cloning and applied on the clone
23:57:05 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
23:57:12 *** WormnestAndroid has joined #openttd
continue to next day ⏵
