IRC logs for #openttd on OFTC at 2023-07-06
⏴ go to previous day
00:13:59 *** Wormnest has quit IRC (Ping timeout: 480 seconds)
00:28:12 *** Flygon has quit IRC (Remote host closed the connection)
00:44:53 *** Wormnest has joined #openttd
01:39:04 *** Wormnest has quit IRC (Ping timeout: 480 seconds)
02:23:27 *** Wormnest has joined #openttd
02:26:45 *** debdog has quit IRC (Ping timeout: 480 seconds)
02:35:43 *** D-HUND is now known as debdog
02:55:28 *** Wormnest has quit IRC (Quit: Leaving)
03:38:31 *** Flygon has quit IRC (Remote host closed the connection)
04:16:32 *** ChanServ sets mode: +v tokai
04:23:12 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
06:34:08 *** keikoz has quit IRC (Ping timeout: 480 seconds)
06:37:13 <pickpacket> Why is mastodon mobile gui so shit? I can’t find a way to search for a user
06:53:35 <kamnet> peter1138: Seems like a reasonable request to add.
06:55:36 <pickpacket> The question is how to add it. If it’s a setting then it’s no harder to find or change than any other setting
06:56:41 <pickpacket> no harder than entering multiplayer in the first place
06:59:59 <LordAro> settings aren't necessarily displayed ingame
07:00:34 <LordAro> that said i'm not sure this is our problem to fix
07:00:57 <LordAro> firewall, or just straight disconnecting the computer from the internet seem like more general solutions
07:01:14 <kamnet> True - but then you can't download new content or updates.
07:01:28 <truebrain> DNS blacklist `coordinator.openttd.org` and done 😛
07:01:42 <LordAro> responsible adults can do that as needed
07:04:05 <kamnet> truebrain: Ah, this does seem like the better idea. Besides, if the child has access to the game, the have access to the settings and config file and can likely figure out how to re-enable anything that's disabled from within the game.
07:04:27 <truebrain> if not now, in a year or so they would 🙂
07:07:15 <truebrain> there will be many of those PRs today 😛
07:07:41 <Rubidium_> just install Windows 98 on the computer, then you cannot use a recent OpenTTD and as such there's essentially no server to connect to. And it's quite hard to circumvent without completely reinstalling the OS :D
07:08:14 <truebrain> give it a year extra 😉
07:08:22 <peter1138> truebrain: Nice, I have suggested this.
07:08:45 <truebrain> I WAS HELPFUL TODAY \o/
07:09:32 <peter1138> Day complete, go back to bed.
07:15:37 <peter1138> > The core of the earth is an estimated 2.5 years younger than the crust, not because it arrived later, but because time runs slower there
07:15:54 <peter1138> I can live slightly longer by moving to the centre of the earth...
07:25:20 <peter1138> Is it lunch time yet?
07:25:40 <peter1138> Time would need to be going faster for that. Conflicts of interest.
07:29:15 <truebrain> time is an illusion anyway
07:29:20 <truebrain> (I played too much Path of Exile)
07:34:34 *** HerzogDeXtEr has joined #openttd
07:40:16 <truebrain> right .. minus a few repos that miss a dependabot configuration file at all, this should be it .. and if you think this is spammy, just be happy you don't see all the PRs dependabot is/was making 😛
07:48:58 <truebrain> I think that is all? 🙂
07:49:11 <truebrain> so now talk away _zephyris 😛
07:50:37 <_zephyris> I was wondering about the best way to implement a base set with optional ez or 32bpp sprites...
07:51:36 <_zephyris> Currently I preprocess to (optionally) uncomment the alternative sprites
07:53:22 <_zephyris> But making 4 versions (8bpp, 8bpp EZ, 32bpp, 32bpp EZ) isn't the most user friendly!
07:56:17 <Rubidium_> the grfcodec repository has a grfstrip which could strip out certain color depths and zoom levels from a GRF. So build once, strip 3 or 4 times
08:01:19 <_zephyris> I'd like to separately handle GUI sprites from viewport/map sprites...
08:01:26 <locosage> I don't think there is much point in doing 8bpp other than filesize but then you need separate file anyway
08:07:12 <_zephyris> It's all a user preference thing - 8bpp does look very different to 32bpp, if a player wants the retro feel.
08:11:53 *** DorpsGek has joined #openttd
08:11:53 *** ChanServ sets mode: +o DorpsGek
08:12:25 <locosage> well, there is 8bpp blitter
08:12:39 <locosage> but ultimately it's a question of baseset parameters I guess
08:12:45 <locosage> that aren't currently supported
08:13:24 <locosage> rn solution is just to make it a newgrf like BRIX
08:16:01 <locosage> also, making ez optional probably duplicates non-ez sprites...
08:16:12 <locosage> at least, I don't see how it can be done without duplication
08:21:14 <ahyangyi> Can we somehow pack the four styles in the same newGRF and rely on global settings. If the user enables 32bpp then use 32bpp, ditto for ez
08:21:45 <ahyangyi> there's the very weird corner case that the user wants 8bpp baseset + 32bpp vehicles, but I can't understand why
08:23:17 <_zephyris> ahyangyi: That's what I was wondering... Just what openttd needs though - more settings!
08:24:50 <_zephyris> locosage: Yeah, that's what I think too. Because newgrf switches happen on sprites not sprite alternates...
08:25:21 <truebrain> also take in mind diskspace; not everyone enjoys having a 300+ MiB baseset, while they only wants 8bpp graphics 🙂
08:26:00 <_zephyris> Very true. The 8bpp 1x is 10mb, 32bpp 4x is 150mb so far.
08:26:25 <truebrain> so having at least two packages seems reasonable .. 8bpp no-EZ, and the "full", or something 🙂
08:27:53 <truebrain> and if you brand that a bit nice, the user won't even notice in that sense .. something like `OpenGFX2 (Classic look)` and `OpenGFX2 (Modern look)`, or what-ever terms floots your boot 🙂
08:28:21 <truebrain> floots? Well, at least I was consistent with missing the a .. floats! ffs, spelling is hard 😛
08:28:41 <locosage> and not for basesets but for newgrfs in general ideally there would be a way to choose how much to download
08:28:51 <locosage> like if server requires yeti but you don't want 200mb
08:29:30 <ahyangyi> _zephyris: I mean, OpenTTD already has those settings
08:29:54 <ahyangyi> enforce 8bpp and force-disable EZ
08:30:12 <truebrain> you can avoid EZ sprites being used, yes
08:30:18 <truebrain> but forcing 8bpp is a bit of a different story 🙂
08:30:36 <locosage> it kinda has but should be a setting per newgrf, not global imo
08:30:50 <_zephyris> Sort of... Max zoom for map/viewport forces 1x zoom sprites, but also stops extra zoom in, right?
08:31:00 <truebrain> all or nothing baby
08:31:03 <ahyangyi> kinda, but since we don't have settings for basesets
08:31:50 <truebrain> but so yeah, personally I wouldn't do 4, but just 2. As in my mind, if you like EZ, you also kinda want 32bpp .. 8bpp EZ is a bit odd, in that sense
08:32:09 <truebrain> and if you like EZ, you most likely want 32bpp too, as you clearly like graphics 😛
08:32:13 <ahyangyi> Timberwolf is 8bpp EZ 😛
08:33:06 <truebrain> well, for me it comes from: 8bpp is basically just a color palette picked. As we render on all platforms shit as 32bpp anyway, this is more about a style we push forward. With 8bpp, the look and feel of most sets are very much the same
08:33:16 <truebrain> with 32bpp you break that open and you can just make any style you like
08:33:30 <truebrain> hence why I always consider 8bpp vs 32bpp "looks"
08:33:43 <truebrain> I might be completely off, with that statement btw, but that is how it works in my head 😄
08:34:01 <_zephyris> truebrain: Yeah. Opengfx2 32bpp is very much an 8bpp 'look'
08:34:22 <_zephyris> Because that's what floots my boot 😉
08:35:09 <truebrain> 😄 😄 I see what you did there!
08:36:09 <_zephyris> Thanks for the feedback everyone! Looks like an 8bpp 1x 'classic' and 32bpp 4x 'modern' is the way to go, at least for now.
09:11:34 <Timberwolf> fwiw I do 8bpp for filesize reasons, because Trains has an utterly ludicrous number of sprites (all those intermediate turn angles, hill climbing/descending and the duplicates to avoid clipping bridges/tunnels on straight track) it would have a silly file size if I provided 32bpp versions.
09:12:39 <Timberwolf> Also the workflow starts with a voxel object in the TTD palette, and even with lighting and shadowing never deviates particularly far from that, so there's not a huge visual benefit other than some of the more restricted colour ramps being very awkward to dither with.
09:30:28 <peter1138> 2x 8bpp is a nice stylistic compromise.
09:30:56 <peter1138> More detail but not trying too hard to be photorealistic.
10:10:03 *** m3henry has joined #openttd
10:19:47 <peter1138> Urgh, C# bugging me that you can't have a class member name be the same as the class name.
10:23:20 <_glx_> Would be the constructor
10:23:20 <Rubidium_> is that case sensitive of not? As I always wonder who thought "everything should start with an uppercase character" was a good idea to quickly see what you're dealing with
10:24:38 <peter1138> Yeah, it's obvious, and probably not allowed in C++ either.
10:25:53 <m3henry> Is it possible to have an override file for english (or other) translations without recompiling?
10:32:26 <peter1138> It was fixed by #8659
10:34:02 <m3henry> I mean to actually provide *man *woman as titles. *person Just strikes me as silly.
10:35:07 <peter1138> Oh. I would say no. -person is correct.
10:37:11 <m3henry> I guess I'm really asking if I put in the work to implement it, would it be considered?
10:38:11 <m3henry> Hmm, not the response I was hoping for
10:39:32 <m3henry> Just seems wierd, like calling a policeman a policeperson.
10:41:13 <Rubidium_> or maybe a police officer?
10:42:14 <m3henry> Indeed, but still not -person
10:43:14 <truebrain> we have policepersons in OpenTTD? What are they doing there? 😄
10:43:46 <Rubidium_> though maybe there is a better synonym for the chair/board/... whatever
10:44:24 <m3henry> Atleast for chairman, an alternative could be "managing director"
10:49:02 <Rubidium_> anyway, English is not my native tongue and they describe some levels of management within a company, so... I can't really help with finding a better name/label
10:51:17 <_zephyris> On a totally different subject, how do OpenTTD and NML determine grf MD5?
10:52:45 <m3henry> For Businessman, the alternative could be merchant?
10:52:54 <_zephyris> Unless I'm doing something really dumb, I'm getting the same MD5 from `nmlc --md5` for two different `.nml` inputs. And, I get a different MD5 from `md5sum`...
10:54:11 <Rubidium_> I reckon the md5 sum ignores the extra depths and zoom levels over the base graphics
10:56:09 <Rubidium_> though OpenTTD seems to just get the MD5 checksum of the whole file
10:56:26 <_zephyris> Yeah, that makes sense. The only difference in the nmls is whether to use the extra zoom/32bpp alternates
10:57:46 <_zephyris> Rubidium_: It uses something other than a simple file md5, openttd gives corrupt file warnings if I use md5sum rather than nmlc md5 in the baseset definition.
10:58:46 <truebrain> md5sum is nothing what it looks like, sadly .. bit of a misnomer 😛
10:58:53 <truebrain> it uses md5, but not in a classical way
10:58:53 <Rubidium_> ah, if GRF container format v2, then it's doing the MD5 checksum on the "data" section
10:58:56 <_glx_> Depends on grf container, for v2 I think only actions are used
10:59:16 <truebrain> for basesets for example it XORs the md5sums of the individual files together
10:59:20 <truebrain> silly things like that
11:02:41 <truebrain> hmmm .. it is even possible that uploading a baseset to BaNaNaS that consists of multiple files renders the wrong md5sum .. I shouldn't look at this code, it scares me 😄
11:03:13 <truebrain> ah, no, I am wrong, it takes the md5sum of the catalog .. with there in the md5sums of all individual files .. lolz
11:03:18 <truebrain> scripts are XOR'd ..
11:04:46 <peter1138> Basesets are not like NewGRFs, it doesn't matter if the MD5 hashes don't match the GRFs in another baseset.
11:04:55 <truebrain> wasn't the idea with v2 that it would be possible to optional download 32bpp, if a GRF had both 8bpp and 32bpp, or something? 🙂
11:05:24 <peter1138> For NewGRF you can have different files with the same MD5 hash, just different graphics.
11:05:32 <peter1138> But that isn't necessary for basesets.
11:07:47 <_zephyris> It is colliding somehow, can't quite work it out.
11:08:28 <_zephyris> I've got two base sets, different file names for each grf but the same md5 for each grf, and only one appears in the baseset options list.
11:09:43 <m3henry> Another option would be to use vice-president? As the next higher rank is president
11:12:21 <Eddi|zuHause> truebrain: i don't think there's a sane use case for splitting 8bpp/32bpp. maybe splitting extra zoom...
11:12:32 <truebrain> _zephyris: they both have a different shortname in the catelog file?
11:16:14 <_zephyris> truebrain: Yup. ogfx2_8 and ogfx2_32ez
11:17:27 <truebrain> shouldn't it be 4 chars? Can't remember
11:18:22 <Rubidium_> yyp, only the first four characters are taken
11:20:17 <_zephyris> Really? ogfx and ogfx2 seem to coexist happily
11:21:44 *** FLHerne has quit IRC (Quit: There's a real world out here!)
11:24:30 <Rubidium_> so... ogfx (small graphics) and oGFX (big graphics) :D
11:25:04 <Rubidium_> and lowercase is later than uppercase in the ASCII table... so ogfx > OGFX :D
11:28:04 *** m3henry_ has joined #openttd
11:29:07 *** m3henry has quit IRC (Read error: Connection reset by peer)
11:29:07 *** m3henry_ is now known as m3henry
11:34:58 *** keikoz has quit IRC (Ping timeout: 480 seconds)
11:35:04 <_glx_> f/F for full range color
11:36:05 *** FLHerne has joined #openttd
11:39:36 <_glx_> _zephyris: Nmlc has flags to do that at compile time
11:49:01 *** Kitrana2 has quit IRC (Ping timeout: 480 seconds)
12:02:59 <_zephyris> Yup... I don't use them (at least for now) because I want the option of keeping 2x/4x GUI sprites while stripping out 2x/4x map sprites. May well switch to the new nmlc feature in the future though.
12:09:23 <LordAro> dunno about "classic", it suggests old and out of date. Would probably call ogfx1 "classic"
12:20:53 <_zephyris> Fair enough. Something to think about. Is OpenGFX2 even the best name?
12:24:28 <peter1138> Sticking with OpenGFX(2) makes sense, otherwise nothing implies that the current OpenGFX is... out of date.
12:36:35 <LordAro> very much depends on the direction _zephyris wants to go, but why the 1/2 split at all? why not just "overwrite" ogfx1? (with added bonus 32bpp version)
12:37:41 <peter1138> Well, I can imagine some people might want to use the old one for some reason.
12:38:34 <LordAro> yeah, but people are wrong
12:38:55 <LordAro> and all we do is try to restrict people's play styles
12:44:51 <ahyangyi> I suddenly feel very wrong and want to change my rendering setting to support NightGFX 😮
12:44:53 <Rubidium_> since when are you required to update to the newest OpenGFX?
12:46:10 <LordAro> it's quite hard to go downgrade
12:47:03 <LordAro> and i imagine most people don't like the outstanding "update available" box in the content download window
12:48:03 <Eddi|zuHause> the problem with the 1/2 split is now you got both versions to maintain
12:48:25 <Rubidium_> LordAro: "update available" does not seem to be a concern for many people; basically all team members at work have a pending update for their IDE when I see their IDE
12:48:59 <LordAro> Eddi|zuHause: i wouldn't say it's a huge burden
12:49:16 <LordAro> (i realise i have completely switched sides in this argument)
12:49:46 <truebrain> At least you realise 😄 most people don't..... 😛
12:50:55 <Eddi|zuHause> sorry, i forgot my role in this, i'm meant to destroy people's playstyle
12:51:44 <Eddi|zuHause> (such a silly argument, if that were true, i would have majorly attacked openttdcoop...)
12:52:46 <peter1138> We're the plague, you're the perpetrator.
12:53:17 <peter1138> Oh that isn't the word that was used. Never mind.
12:56:40 <FLHerne> Eddi|zuHause: people keep downloading aBase and no-one maintains that
12:56:59 <Eddi|zuHause> FLHerne: not sure that's a good argument for anything :p
12:57:23 <peter1138> (It's Pi Zero at its core but I imagine 400x240 screen might need work 🙂
12:57:51 <Eddi|zuHause> that's a bit tiny for a game designed for 640x480
12:58:32 <Eddi|zuHause> also, the "buy" link goes to 404
13:03:40 <m3henry> I've been testing OpenTTD on the PineTab2 recently, and I get a solid 60FPS at 4x zoom out, with 87 trains running about.
13:04:34 <peter1138> > The PineTab2 began shipping on June 2, 2023.
13:04:58 <peter1138> Who imports this thing :/
13:05:17 <m3henry> I think they import it themselves
13:06:23 <m3henry> Still, I'm impressed with the software renderer working so well without any specificoptimizations for mobile aarch64
13:07:39 <m3henry> I think I paid £170 in total
13:07:41 <Eddi|zuHause> back when TT came out, i got maybe 2fps with 80 trains
13:08:25 *** johnfranklin has joined #openttd
13:08:25 <johnfranklin> $250 for this? Better buy GPD Pocket with a little higher price.
13:09:24 <peter1138> And Eddi still hasn't upgraded their computer.
13:10:03 <johnfranklin> When TT came out, my father was in high school😛
13:10:06 <Eddi|zuHause> using underpowered computers is kind of a theme running through the ages :p
13:10:44 <m3henry> Anyway, it might be a good idea to not have the hardware rendering on by default as the performance is near 10FPS
13:10:54 <peter1138> I don't think John Franklin is old enough to be on Discord.
13:13:00 <johnfranklin> I was born in 2003
13:13:26 <johnfranklin> johnfranklin: Oh sorry, it’s not “a little”
13:14:30 <peter1138> Double the price, a smaller screen, and it's Intel instead of ARM so likely worse battery (although screens are the main draw there these days)
13:16:04 <m3henry> my record has been roughly 10 hours of battery with light usage and no suspending
13:23:42 <Eddi|zuHause> that sounds a little low for "i want to use this for a day and not worry"
13:23:54 *** virtualrandomnumber has joined #openttd
13:24:21 *** virtualrandomnumber has quit IRC (Remote host closed the connection)
13:24:24 <Eddi|zuHause> remember when phones lasted 2 weeks with one charge?
13:26:23 <m3henry> Most days that I use it, I manage to use about 25% of the total capacity, so it really doesn't bother me
13:28:07 <LordAro> i'm quite interested in the nothing phone
13:28:20 <LordAro> my oneplus5 is getting a bit tired
13:30:17 <m3henry> As is my Galaxy Core Prime xD
13:31:16 <locosage> replacing ogfx1 with ogfx2 makes no sense whatsoever, they're completely different
13:31:24 <locosage> if anything, remove basesets then
13:33:15 <locosage> ogfx2 has as many pixels as abase, same style as original and same name as opengfx
13:33:22 <locosage> perfect replacement for everything xD
13:33:31 <_zephyris> LordAro: As the person who drew huge chunks of OpenGFX, I have to disagree... The way OpenGFX1 looks wasn't entirely deliberate, it was because I wasn't as good at pixeling.
13:34:29 <m3henry> I'm most interested in the 2x sprites
13:36:30 <_zephyris> Eddi|zuHause: I think maintenance overhead to keep OpenGFX1 and 2 running would be very small - huge chunks of the 2 coding use the same structure as 1, and the 8bpp sprites from 2 will (generally) fit well into 1.
13:37:17 <_zephyris> But no strong feelings on exact roadmap. Main aim is to make something useful and put it into the world.
13:40:54 <m3henry> How much does 4x increase the filesize?
13:43:00 <ahyangyi> 16x if we don't take compression into consideration, about 5x (I think) in practice
13:43:31 <ahyangyi> Of course it depends on the drawing/rendering style
13:43:37 <ahyangyi> details are expensive
13:44:41 <m3henry> Is it worth having a stripped down version with only 1x/2x sprites to reduce hosting costs?
13:46:09 <johnfranklin> Eddi|zuHause: When I was very young
13:50:36 <Eddi|zuHause> m3henry: you somehow have to catch the people who just download everything without thinking about it
13:51:30 <m3henry> That's a lot of boxes to tick, does anyone go to the trouble of doing that?
13:52:13 <m3henry> So a 2x version would potentially be ~30MB?
13:53:45 <Eddi|zuHause> each step is probably the same factor, so 8bpp->32bpp(x1)->x2->x4
14:00:40 <_zephyris> locosage: Hah! I missed this message, I'll take this as a massive complement!!
14:15:36 <ahyangyi> I know that a 1x tile is 64 pixels wide and 32 pixels tall. But what is the offset in pixel for each step in altitude difference?
14:15:58 <ahyangyi> I tried to measure that on a screenshot but I got readings fluctuating from 5 to 9
14:16:15 <peter1138> (But bridges and tunnels are weird)
14:16:34 <peter1138> (Because vehicles tend to be 12-14 pixels)
14:19:26 <Eddi|zuHause> there's some perspective tricks with tunnels
15:01:14 <m3henry> _zephyris: I notice that the power station sparks background is not transparent
15:01:35 <m3henry> Otherwise it looks great
15:01:50 <peter1138> ⮡ doesn't really display well... hmm.
15:02:42 <peter1138> Saying that it works okay in a terminal with fixed-width font, but still eats into the space next to it.
15:11:15 <Eddi|zuHause> it's not adhering to fixed width here either
15:32:23 *** gelignite has joined #openttd
16:36:12 *** Smedles has joined #openttd
17:01:26 *** argoneus2 has joined #openttd
17:01:26 *** argoneus has quit IRC (Read error: Connection reset by peer)
17:01:26 *** argoneus2 is now known as argoneus
17:11:01 *** Wormnest has joined #openttd
17:16:45 <peter1138> Hmm, okay, using a spare HP DL370 for random things is probably not a great idea... 300W power consumption :p
17:26:29 <Rubidium_> if random things includes heating the room :D
17:31:17 <_glx_> so weird, my local emscripten build now fails with "/developpement/github/glx22/openttd/src/network/network_survey.cpp:35:10: fatal error: 'nlohmann/json.hpp' file not found"
17:49:20 <locosage> So I want a variant group that contains two groups with two elements each, how do I set varint_group parameter for vehicle 3?
17:49:55 *** belajalilija has joined #openttd
17:49:55 <belajalilija> so the code yesterday not worrking wasnt *entirely* my fault?
17:50:24 <locosage> and may not even be entirely mine either xD
17:54:00 <peter1138> Yay for Rubidium fixing my bugs 😄
17:55:40 <truebrain> RaaS .. Rubidium-as-a-Service
18:02:40 <locosage> I guess first element of a vehicle group must always be a single vehicle, not another group
18:02:52 <locosage> so much for nesting...
18:04:59 <belajalilija> didnt andy do it?
18:05:22 <peter1138> Did andy mention GS again?
18:06:23 <belajalilija> ive not actually used horse since vars was added but im sure he said it was possible
18:08:32 <ahyangyi> I see `Groups can also be nested (this is experimental as of December 2022 and may change with testing)` in the wiki though
18:09:20 <peter1138> Groups have been nestable for years.
18:09:28 <locosage> afict horse always expands last element only
18:11:45 <ahyangyi> Uh, in the wiki for `variant_group`
18:12:16 <locosage> yeah, they're supposed to be nestable but looks like it's limited
18:15:05 <peter1138> Oh, you're talking about variants. They are nestable, and always have been.
18:15:39 <peter1138> Not sure why NML calls them "group" though.
18:17:10 <ahyangyi> Hey, this feature PR called it "engine variant groups" as well:
18:17:37 <peter1138> Influenced by andy I guess. Oh well.
18:17:58 <peter1138> Anyway, yeah, they are nestable, up to 10 levels.
18:19:15 <peter1138> Although no particular reason for 10.
18:20:01 <peter1138> George uses nested variants in xUSSR (or whatever it is called)
18:20:17 <jfs> someone posted a somewhat interesting idea in a comment on a youtube video of mine: screensaver mode for OTTD
18:20:42 <jfs> basically the game starting, loading and running a title screen, but with no menu or overlays, and just quits when you move the mouse/etc
18:20:46 <locosage> you have youtube channel?
18:21:00 <peter1138> Can it randomly go HONK HONK with ships on the map? 😄
18:21:02 <jfs> yes and no, just a regular personal one with a few random videos
18:22:48 <belajalilija> i think having the scene change is good
18:23:00 <belajalilija> but the movement hurts my brain for some reason
18:23:42 <jfs> many have said that the scrolling is problematic for them to watch
18:24:13 <jfs> so maybe it's best to recommend that future title games stick to only static displays instead of pans/vehicle follows
18:28:39 <locosage> locosage: so, back to original question, am I missing something or should I open a bug?
18:36:22 <locosage> hm, maybe adding an unavailable vehicle as a group header can work...
18:36:31 <locosage> won't be purchasable though at first iirc
18:36:35 <locosage> and with grey text...
18:38:10 <locosage> 1 will add it to group B
18:38:47 <brickblock19280> Add two into one?
18:40:15 <ahyangyi> 3->1 will move 3 into the group B
18:40:22 <brickblock19280> That would be silly
18:40:40 <DorpsGek> - Update: Translations from eints (by translators)
19:03:02 <locosage> yeah, with invisible vehicle it works but looks weird
19:03:24 <locosage> and it seems to ignore string color so can't fix it there
19:06:07 <belajalilija> this has got to be one of the most unintentionally cured grfs
19:08:26 <brickblock19280> wouldn't this be better fixed on game side tho
19:11:01 <locosage> yeah, but this spec is so cursed I don't even know how
19:11:29 <locosage> problem is, it uses first vehicle as a group header
19:11:36 <locosage> but here it's a head of multiple groups
19:12:25 <brickblock19280> make separete groups property of head veh
19:19:21 <locosage> how would that work?
19:19:44 <locosage> I guess a property to directly set "first bought" vehicle would work
19:19:54 <locosage> so group can be invisible but get replaced without buying
19:21:52 <locosage> or, actually, why is the game even showing disabled vehicle?
19:22:06 <locosage> can't it just choose any that is allowed?
19:22:15 <_glx_> someone ask for the feature IIRC
19:23:32 <locosage> you mean disabled vehicles? iirc they were supposed to be a solution for group names
19:23:43 <locosage> but it never worked so name cb was added
19:36:51 <locosage> I guess some kind of "vehicle group behavior" property would be the best solution
19:36:51 <locosage> >0 - vehicle group header (automatically set to NO_CLIMATE)
19:38:33 <brickblock19280> two and three should imo be chosen by the user not the grf
19:42:49 <locosage> yeah, that's debatable
19:43:47 <locosage> but imo that ship has already sailed with name callback
19:44:42 <locosage> would be even more weird if grf sets up name callback for last bought but user configures something else
19:45:48 <locosage> grf can make it configurable in its own parameters xD
20:12:44 <brickblock19280> it should be fine if grfs cant know then they shouldn't do it like that
20:13:05 <brickblock19280> and if they do then that is their
20:19:20 <_glx_> truebrain: it's strange, savegames and downloaded files are saved, but not the config it seems
20:20:18 <truebrain> is the mainloop properly shut down? As saving the config happens REALLY late in OpenTTD
20:20:24 <truebrain> like, after the video mainloop closes late
20:21:18 <_glx_> ah emscripten finishes with the video loop
20:21:25 <_glx_> nothing happens after it
20:23:35 <_glx_> ok if I change settings in the settings window they are saved immediately, but with game option or the survey window it's not the case
20:24:00 <_glx_> well saved immediately multiple times it seems
20:25:45 *** gelignite has quit IRC (Quit: Stay safe!)
20:30:54 <_glx_> yeah config saving happens after emscripten exit
20:41:47 <_glx_> I won't try to fix that 🙂
20:47:51 <truebrain> yeah, that code has always been a bit weird, honestly
20:48:00 <truebrain> but you can also wonder if the teardown of emscripten is in the right place
20:48:14 <truebrain> but easiest way would be to put those lines in a function, and call that when closing down emscripten 🙂
20:48:44 <truebrain> and yes, this assumed all setting changed caused a config save
20:48:50 <truebrain> so maybe the bug is just that game options don't
20:53:55 <_glx_> hmm CI failure is annoying, it build fine on my machine
21:00:50 <Rubidium_> i.e. rerun the fail jobs
21:11:01 <truebrain> that is the OOM paying a visit _glx_
21:11:05 <truebrain> seems emscripten is pushing it
21:11:45 <truebrain> had it once before this week .. a retrigger fixed it
21:12:38 <truebrain> seems we found a limit ... now we need to check what we can do to avoid hitting it constantly 😛
21:14:29 <_glx_> btw survey should have received some wasm32 reports 🙂
21:18:53 *** keikoz has quit IRC (Ping timeout: 480 seconds)
21:18:59 <_glx_> hmm and it failed again, that will be annoying
21:21:29 <truebrain> emscripten can't make HTTP connections, so no, the survey did not receive any, sorry 🙂
21:23:12 <truebrain> guess `IsSurveyPossible` should be extended to make EMSCRIPTEN return false too ..
21:29:21 *** m3henry has quit IRC (Quit: m3henry)
21:37:52 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
21:39:12 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
21:48:39 <_glx_> it's so weird to not be able to HTTP from a browser
22:04:51 <truebrain> _glx_: you can, but you can't do it streaming
22:05:08 <truebrain> but when I am done with the last pieces of the infra migration, and written a blog-post, I want to fiddle with pthread support in emscripten
22:05:17 <truebrain> as if that works, it becomes easier to have a non-streaming variant
22:05:43 <_glx_> yeah I looked at commits since the version we use, I saw some pthread ones
22:06:12 <truebrain> and they are also moving towards WASI, which is a good thing
22:06:26 <_glx_> and many WASMFS commits too
22:06:41 <truebrain> yeah, them building out on what WASI offers, which is just a much better way of doing WASM
22:07:18 <truebrain> one downside is with SharedBuffers .. it requires a ton of headers on the host hosting it
22:07:25 <truebrain> but as we are those people, we should be able to manage 🙂
22:17:37 *** nielsm has quit IRC (Ping timeout: 480 seconds)
23:33:42 *** k-man has quit IRC (Remote host closed the connection)
continue to next day ⏵
