IRC logs for #openttd on OFTC at 2022-08-23
02:05:24 *** Wormnest has quit IRC (Quit: Leaving)
02:56:58 *** D-HUND has joined #openttd
03:00:16 *** debdog has quit IRC (Ping timeout: 480 seconds)
04:49:55 *** D-HUND is now known as debdog
05:57:20 *** nielsm has joined #openttd
06:53:49 *** WormnestAndroid has quit IRC (Remote host closed the connection)
07:10:57 *** Wolf01 has joined #openttd
07:42:16 <DorpsGek> [OpenTTD/team] Bouke opened issue #340: [fy_NL] Translator access request
08:50:43 *** Flygon has joined #openttd
09:46:10 *** vasca15 has joined #openttd
09:46:29 *** vasca15 has left #openttd
09:55:21 <Wolf01> Bah, bored to play games, every day I get bored sooner
10:04:23 *** OsteHovel_ has quit IRC (Read error: Connection reset by peer)
10:04:24 <DorpsGek> [OpenTTD/team] glx22 commented on issue #340: [fy_NL] Translator access request
10:41:06 *** ely-the-kitsune has quit IRC (Ping timeout: 480 seconds)
10:41:36 *** WormnestAndroid has joined #openttd
10:42:12 *** virtualrandomnumber has joined #openttd
10:42:30 *** virtualrandomnumber has quit IRC (Remote host closed the connection)
10:49:29 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
10:49:36 *** WormnestAndroid has joined #openttd
10:59:36 <DorpsGek> [OpenTTD/OpenTTD] pon-droid updated pull request #9973: Add #7889: Company rating via Server Admin Port
11:01:13 <DorpsGek> [OpenTTD/OpenTTD] pon-droid updated pull request #9973: Add #7889: Company rating via Server Admin Port
11:05:05 <DorpsGek> [OpenTTD/OpenTTD] pon-droid commented on pull request #9973: Add #7889: Company rating via Server Admin Port
11:13:35 <DorpsGek> [OpenTTD/OpenTTD] pon-droid updated pull request #9973: Add #7889: Company rating via Server Admin Port
11:17:48 *** OsteHovel has joined #openttd
11:29:37 *** WormnestAndroid has quit IRC (Read error: Connection reset by peer)
12:11:44 <NEKOMASTER> Wolf01: I know the feeling, apathetic depression sucks
12:14:46 *** glx has joined #openttd
12:14:46 *** ChanServ sets mode: +v glx
13:06:32 *** WormnestAndroid has joined #openttd
13:52:19 <DorpsGek> [OpenTTD/OpenTTD] James103 opened issue #9974: [Bug]: `getsysdate` returns a blank value (instead of the current date and time)
13:56:39 <LordAro> bet that's the same miscompilation as the other issue
13:57:00 <JGR> The buffer is too small
13:57:59 <JGR> I get a partial result with the last part malformed on Linux
14:00:54 <LordAro> how is it too small?
14:01:00 <LordAro> oh, null terminator?
14:02:29 <DorpsGek> [OpenTTD/OpenTTD] JGRennison commented on issue #9974: [Bug]: `getsysdate` returns a blank value (instead of the current date and time)
14:17:53 <glx> buffer size is correct and includes the \0
14:20:22 <glx> but lastof points to the character before the \0, and last - buffer is then of by one
14:21:38 <andythenorth> meh, with these larger outpost layouts, FIRS Steeltown regularly fails to place the industries on mountainous maps with low water, high roughness, where the industry has rules requiring it to be near other industries
14:22:17 <andythenorth> not really viable 😐
14:24:51 <EmperorJake> I wasn't a huge fan of the outposts because they fix a problem that I don't have
14:25:31 <andythenorth> I'm on the fence currently
14:25:38 <andythenorth> they're the best solution found so far
14:26:10 <andythenorth> one option might be to delete Steeltown
14:27:21 <NEKOMASTER> I was thinking perhaps if I made my own industry set from scratch, that perhaps most industry should be scaled to the typical OpenTTD town size
14:27:43 <NEKOMASTER> For example, only some massive industries like oil refinaries, steel mills and car plants should be as large as a OpenTTD town
14:28:21 <NEKOMASTER> The rest should be small enough to better fit around towns
14:28:23 <andythenorth> your count per map will be 0
14:28:34 <peter1138> Yay, new brake pads :D
14:28:58 <peter1138> I'll probably end up contaminating them getting them in...
14:29:10 <NEKOMASTER> andythenorth: I was more thinking like at most, how large the ECS Vectors car plant or largest oil refinary is
14:29:28 <JustANortherner> peter1138: No problem. Brakes aren't important...
14:29:52 <andythenorth> industries above about 5x5 are really a no
14:30:18 <NEKOMASTER> Well I was thinking no larger than that
14:30:52 <NEKOMASTER>
14:30:52 <NEKOMASTER> For example, this car plant here is HUGE compared to this town
14:31:04 <glx> ah no, lastof is the \0, but there's still an obiwan in last - buffer
14:32:19 <peter1138> JustANortherner, just stick my feet on the tyres/ground :D
14:32:42 <JustANortherner> Exactly! What could possibly go wrong?!
14:32:58 <JustANortherner> You might need a new pair of shoes every few weeks though
14:33:21 <JGR> Trying to match scales is an impossible task in OpenTTD
14:33:31 <JGR> Really, towns are too small compared to everything else
14:33:39 <andythenorth> that car plant is about 7x9
14:33:46 <andythenorth> surprised it gets built on anything but 'flat'
14:33:59 <andythenorth> try it on 'mountainous' and 'rough'
14:34:05 <peter1138> I liked Pikka's TaI as an attempt to increase town area.
14:34:46 <andythenorth> ITI does a similar thing
14:34:51 <andythenorth> large areas of low-rise
14:34:54 <andythenorth> works well
14:35:33 <peter1138> Does that reduce the population density as well?
14:36:02 <NEKOMASTER> Part of the issue with OpenTTD and scale is that to make things feel more realistic, requires more land and theres only so much space a map can give
14:36:02 <andythenorth> yup
14:36:19 <NEKOMASTER> For example, you could do a country in 4K and have everything kinda "small"
14:36:37 <NEKOMASTER> Or do a region in 4K and do it like its zoomed in
14:37:14 <NEKOMASTER> Though I think to get to realistic scale with making a map to have real population numbers you'd have REALLY large towns and little space for much
14:37:32 <NEKOMASTER> but might be fun for a very local setup depending on the city you focus on
14:43:52 <andythenorth> meh
14:44:06 <andythenorth> mountainous maps really challenge FIRS
14:44:18 <andythenorth> this is not new, but eh
14:44:54 <andythenorth> would it be helpful if industries could level land?
14:45:30 *** ely-the-kitsune has joined #openttd
14:50:10 <andythenorth>
14:50:10 <andythenorth> the game has some setting for it, but it doesn't do anything with FIRS
14:59:08 <andythenorth> hmm there's no precedent in TTD for split industries also
14:59:14 <andythenorth> so that's a problem
14:59:17 <andythenorth> in canon
14:59:27 <JGR> Does that really matter?
15:00:14 <JGR> Players who want the vanilla no-GRF experience can just do that
15:00:34 <andythenorth> irrationally bothers me
15:00:37 <andythenorth> but that's irrational
15:00:51 <DorpsGek> [OpenTTD/OpenTTD] glx22 opened pull request #9975: Fix #9974, aa5a8fe: strftime buffer maxsize is off by one
15:00:54 <andythenorth> no canon for many of the things we've added πŸ˜›
15:02:01 <DorpsGek> [OpenTTD/OpenTTD] nielsmh approved pull request #9975: Fix #9974, aa5a8fe: strftime buffer maxsize is off by one
15:02:04 <andythenorth>
15:02:04 <andythenorth> visually two more equally sized layouts might be better
15:02:10 <andythenorth> the implementation supports it fine
15:02:22 <andythenorth> but that doesn't solve the placement problem on mountainous maps
15:02:30 <andythenorth> nothing is going to solve the placement problem except smaller industries
15:02:38 <andythenorth> and the goal is larger industries
15:02:44 <andythenorth>
15:17:50 <andythenorth> hmm adding more position variants will rapidly exceed the number of allowed layouts πŸ˜›
15:26:21 *** Wormnest has joined #openttd
15:27:48 * andythenorth looks if TGP settings are exposed to grf
15:27:53 <andythenorth> 99% sure they're not
15:46:00 <andythenorth> hmm maybe I can check the height level during layout selection
15:46:21 <andythenorth>
15:46:21 <andythenorth> but then again placing an industry here ...
15:46:25 <andythenorth> will be hard
15:48:14 <JGR> Not having an industry spawn there sounds like a feature not a bug
15:48:48 * andythenorth searching for mountainous countries with major steel industries
15:49:00 <andythenorth> it's not as though players have reported this
15:49:20 <JGR> Even mountainous countries have flat areas like valleys
15:51:01 <andythenorth> the specific combination I'm trying is mountainous, variety distribution none, roughness very rough, sea level medium
15:51:14 <frosch> you can check the altitude. the maximum-height from map-generation is somewhat exposed
15:51:18 <andythenorth> there are flat areas at sea level, but they get filled up with industries
15:51:36 <frosch> though there are many players who play boring flat maps, so mountain industries fail for them
15:51:51 <andythenorth> I might ignore this one as a boring pathological case
15:51:57 <andythenorth> into every life, a little rain must fall
15:52:38 <andythenorth>
15:52:41 <andythenorth>
15:53:10 <andythenorth>
15:53:20 <andythenorth> can we call that 'doing what the player requested' ? πŸ˜›
15:53:56 <andythenorth> TBH there is space for the missing industry, I was able to manually place it trivially
15:54:02 <andythenorth> but the game didn't find it
15:54:03 <DorpsGek> [OpenTTD/OpenTTD] glx22 merged pull request #9975: Fix #9974, aa5a8fe: strftime buffer maxsize is off by one
15:54:06 <DorpsGek> [OpenTTD/OpenTTD] glx22 closed issue #9974: [Bug]: `getsysdate` returns a blank value (instead of the current date and time)
15:54:20 <andythenorth> how about a hint that it should not consider sloped tiles as a starting point?
16:20:17 *** HerzogDeXtEr has joined #openttd
16:34:31 * peter1138 wonders if mp3 will sound shit, or more likely, no difference.
16:47:19 <andythenorth> are you the kind of person who can hear the difference between two different power leads?
16:49:46 * peter1138 rewrites a complete git history due to using the wrong email address...
16:50:04 <peter1138> andythenorth, I can hear bad compression sometimes.
16:51:17 <andythenorth> you might need this then
16:51:30 <andythenorth> presumably those people also run their own grid and generating stations
16:54:49 <LordAro> ...that site hard crashes chrome for me
16:56:23 <andythenorth> oh
16:56:26 <andythenorth> I am in Chrome
16:56:34 <andythenorth> probably you need a better power cable
16:56:44 <LordAro> possible
17:07:40 <andythenorth> hmm
17:07:50 <andythenorth> FIRS stuff then πŸ˜›
17:14:36 <andythenorth>
17:14:36 <andythenorth> are we preferring
17:14:54 <andythenorth>
17:14:54 <andythenorth> or a larger outpost
17:14:56 <andythenorth> ??
17:15:13 <TallTyler> Larger, strong preference
17:15:17 <Wolf01> Why not both?
17:15:26 <andythenorth> will run out of layouts
17:15:36 <Wolf01> Larger then
17:15:45 <andythenorth> also not worth the candle to do lots of types πŸ™‚
17:15:56 <andythenorth> the large outpost implies smaller main layout
17:16:05 <andythenorth> 'probably fine
17:16:12 <TallTyler> I also like the split wharf layout you posted yesterday
17:16:15 <andythenorth> puts the vehicles closer to the industry
17:16:26 <TallTyler> You inspired me to start a new Steeltown game πŸ™‚
17:16:38 <andythenorth>
17:16:44 <andythenorth> look for a result quite like that
17:16:50 <andythenorth> the main layout is very TTD sized
17:16:57 <andythenorth> (the actual blast furnace part)
17:17:39 <andythenorth> then there's secondary plant like the big fans
17:17:57 <andythenorth>
17:17:57 <andythenorth> trying to do stuff like this is so out of scope
17:19:27 <TallTyler> More building options in CHIPS?
17:19:48 <TallTyler> Or perhaps official support for object tiles?
17:19:54 <TallTyler> (Included in FIRS)
17:20:18 <andythenorth> I have considered putting objects into one of the grfs
17:20:23 <andythenorth> the FIRS + CHIPS grf is nice
17:20:29 <andythenorth> but it's out of sync with FIRS
17:20:35 <TallTyler> There’s a FIRS Objects GRF that somebody else made but it doesn’t have the right ground textures
17:20:38 <TallTyler> Yeah
17:21:15 <TallTyler> I bet you could automate creation of object tiles πŸ˜›
17:21:19 <andythenorth> wonder if the industry spritelayouts can be taken wholesale
17:21:39 <andythenorth> I haven't looked at objects, but I bet it could be done
17:22:10 <andythenorth> implementation could be something like:
17:22:10 <andythenorth> * give spritelayouts a name string and object ID
17:22:17 <andythenorth> * render all of those out into objects
17:22:29 <andythenorth> objects are probably in groups or some crap?
17:22:47 <TallTyler> I did object tiles in ITI and it’s really easy
17:24:37 <TallTyler> I would particularly appreciate tiles with stored cargos of various types, since station cargo sprites are finicky and don’t appear if you always keep a train loading to keep station ratings up
17:25:01 <andythenorth> that one has always been a puzzle to me
17:25:30 <andythenorth> some grfs have had the industry change sprites if it's producing well
17:25:35 <andythenorth> but I never bothered in FIRS
17:26:04 <andythenorth> the convention very early was that cargo display = cargo waiting
17:26:07 <andythenorth> which is super logical
17:26:15 <andythenorth> but has problems
17:26:28 <TallTyler> It makes more sense to me for the industry tiles to be full of buildings and machinery, and for the player to build loading/unloading/storage tiles around it using object or station tiles
17:26:39 <andythenorth> - cargo sprites aren't seen unless you make a mistake
17:26:39 <andythenorth> - cargo sprites flicker between states
17:26:48 <TallTyler> That way the whole complex can fit the topography and track design
17:27:15 <TallTyler> Yeah, cargo waiting seems like the wrong choice to me but I see why it was done that way
17:27:17 <andythenorth> what about station tiles displaying cargo per rating? πŸ˜›
17:27:33 <andythenorth> I don't know if we can ever unlearn the current convention
17:27:44 <TallTyler> Sure we can, we just stop following it
17:27:58 <andythenorth> I use it to detect failing stations
17:28:04 <andythenorth> hmm
17:28:11 <andythenorth> lol
17:28:20 <andythenorth> how about recolouring failing stations angry red?
17:28:31 <TallTyler> I guess I do too, but that’s a pretty bad reason to do it πŸ™‚
17:28:31 <andythenorth> I feel we miss some 'this is a game' type thinking sometimes
17:29:11 <andythenorth> I never tried displaying flat (non-isometric) icons floating over stations
17:29:13 <andythenorth> might be a thing
17:29:39 <TallTyler> When I added land value to ITL Houses I recoloured sprites based into the calculated value as a debug tool, and never removed it. It’s fun but probably nobody notices (and there’s not much point to knowing anyway)
17:32:08 <andythenorth>
17:32:26 <andythenorth> kinda broken eh πŸ™‚
17:34:36 <andythenorth> generate a full set of FIRS station tiles also? πŸ˜›
17:34:38 <andythenorth> oof
17:34:43 <andythenorth> the faff of building
17:36:51 <andythenorth> does max mountain height scale with map size?
17:36:54 <DorpsGek> [OpenTTD/nml] glx22 updated pull request #247: Add: Translate ternary operator into varact2 if needed
17:37:01 <andythenorth> 512 x 512 seems particularly pathological for FIRS
17:37:09 <andythenorth> 256 x 256 actually does better
17:38:50 <andythenorth> as does 4096 x 4096
17:40:14 <peter1138> Anything other than 256x256 is illegal.
17:40:43 <frosch> hmm, an object grf with firs industries...
17:40:55 <frosch> how many bug reports would that trigger from multiplayer people
17:41:18 <frosch> when they get trolled by other players, and try to connect to objects, which produce nothing
17:42:17 <andythenorth> lol
17:42:27 <andythenorth> the existing object grf seems popular though πŸ™‚
17:42:32 <andythenorth> for model railway people
17:43:22 <andythenorth>
17:44:50 <andythenorth> hmm
17:45:16 <andythenorth> what if grf industries could flatten land? πŸ˜›
17:45:50 <frosch> or the opposite
17:45:59 <peter1138> What if they could expand or shrink?
17:46:00 <frosch> layout contains slope per tile, and ottd terraforms
17:46:02 <andythenorth> if land could flatten industries?
17:46:10 <peter1138> What if there was any rain forecast?
17:46:17 <andythenorth> there was rain here, but it didn't
17:46:30 <andythenorth> also Apple weather diverged a lot from BBC
17:46:37 <andythenorth> the fruit people got it right today
17:46:58 <andythenorth> frosch: does that work with the convention of '1 tile per industry'
17:47:14 <andythenorth> I guess fancy slopes use more tiles
17:47:28 <andythenorth> 'flat' would all be the same tile πŸ˜›
17:47:57 <andythenorth>
17:47:57 <andythenorth> this placement is coincidentally quite interesting on slopes πŸ˜›
17:48:06 <andythenorth> that counts as 'flat' for FIRS
17:49:03 <frosch> i would put the slope with the tile layout, so tile ids do not matter
17:49:13 <andythenorth> in the spritelayout?
17:49:20 <frosch> i am just unsure how to deal with tiles which have no fixed slope requirement
17:49:21 <andythenorth> would be trivial to render that from current FIRS
17:49:41 <andythenorth> no slope => don't change?
17:49:45 <andythenorth> unless neighbouring tile changes?
17:49:49 <frosch> pretty sure there is something in the irc logs :p
17:50:52 <andythenorth>
17:51:02 <andythenorth> FIRS taking the highest corners of each tile there πŸ˜›
18:14:58 <andythenorth> hmm
18:15:06 <andythenorth>
18:15:06 <andythenorth> splits still look weird to me
18:15:09 <andythenorth> like it's 2 industries
18:18:20 <andythenorth>
18:18:20 <andythenorth> the effect with stations is quite different
18:21:29 <andythenorth> the outposts are more clearly just a subsidiary part
18:21:54 <andythenorth> in fact, with busier / better sprite choices, the 2 x 2 probably looks better than 3 x 3
18:21:56 <andythenorth>
18:22:44 <andythenorth>
19:05:18 *** supermop_toil has joined #openttd
19:13:10 <frosch> does the outpost need to be buildings? it could be cargo piles
19:13:25 <frosch> high rise racks and fork lifts
19:33:21 <supermop_toil> rare slow day at work = use vernier caliper to make 3d models of my camera bodies
19:43:51 *** ely-the-kitsune has quit IRC (Remote host closed the connection)
19:45:15 <Wolf01> <frosch> high rise racks and fork lifts <- and even glitchy conveyors and pipes which connect to the main part
19:45:42 <frosch> why glitchy?
19:46:10 <frosch> but yes, underground belts and pipes would explain the cargo transport between the parts
19:47:18 <Wolf01> Graphics which span over multiple tiles (out of bounding box) were really glitchy, I don't know if it has been fixed
19:47:51 <frosch> well, you have to split them
19:48:14 <frosch> i think there is a truegrf feature request to do that automatically :p
19:51:11 <Wolf01> :)
19:57:18 <supermop_toil> what sorcery are you planning
19:58:23 <Wolf01> Animated cable baskets would be cool too
19:58:41 <supermop_toil> nah just add those as a real waytype in the game
19:59:18 <Wolf01> Between 2 parts of the same industry?
19:59:43 <supermop_toil> sure
20:00:22 <supermop_toil> while we're at it, add passenger concourses and bridges for station platforms
20:00:46 <supermop_toil> and passengers who get lost without them a la rollercoaster tycoon
20:01:09 <supermop_toil> also the ability to pick up passengers and drop them in the ocean?
20:01:21 *** ely-the-kitsune has joined #openttd
20:02:09 <FLHerne> > As a special case, if the first byte of a new industry layout is FEh, then only two bytes follow: the industry number and the layout number. The specified layout of the specified old industry type will be added to the layout list of the current industry.
20:02:15 <FLHerne> (from Action0/Industries)
20:02:54 <FLHerne> just checking, "old industry type" means one of the original default industries, and can't be an industry type defined earlier by the grf?
20:03:15 <frosch> i would think so
20:04:31 <FLHerne> ok
20:05:36 <FLHerne> I had an idea to extend industry layouts to make implementing andy stuff less crazy, so I'm looking for a prefix that can't already exist
20:06:10 <frosch> i would ignore the old property, and add a new one with a different/extensible format
20:06:17 <FLHerne> so in that case, FE FE (or FE <anything bigger than 24>) should be fine
20:06:19 <frosch> similar to the advanced sprite layouts for stations
20:06:30 <FLHerne> well, then someone could try to define both
20:06:44 <frosch> nml would only write the new one
20:06:48 <FLHerne> it already takes two different formats, so why not three :p
20:07:00 <frosch> we have this schema in countless places. deprecated property and new property
20:07:10 <frosch> new property overwrites whatever the old one was set to
20:07:47 <FLHerne> hm
20:30:36 <andythenorth> Offset into another layout? πŸ€ͺ
20:36:24 <andythenorth> Composing layouts, but in nml?
20:36:44 <andythenorth> There are many evil things we could do in nml
20:37:09 <andythenorth> Evil / convenient /s
20:37:44 <andythenorth> Probably fine, nml isn’t slow or anything
20:44:24 *** ely-the-kitsune has quit IRC (Remote host closed the connection)
20:45:13 <andythenorth> what about multi-stage placement? πŸ˜›
20:45:26 <andythenorth> place the primary unit, then try and place satellite(s)
20:45:48 * andythenorth isn't sure what problem that solves
20:46:17 <JGR> It comes back to spec limitations and extensions
20:49:38 <JGR> The graphics and shape check callbacks would probably have to change as well
20:50:46 <andythenorth> I don't know that adding complications wins much
20:50:58 *** ely-the-kitsune has joined #openttd
20:51:19 <peter1138> Does it gameplay?
20:51:33 <andythenorth> isn't it
20:51:34 <andythenorth> no
20:52:01 <andythenorth> comes down to: how much should industries be buildable on every map?
20:52:32 <JGR> Having multiple industry layouts doesn't strictly add to gameplay either, but the variety looks nicer
20:52:48 <andythenorth>
20:52:48 <andythenorth> Vanilla already had to make a decision about industry buildability
20:53:09 <andythenorth> despite some walls of text about "you don't know what you're doing"
20:53:28 <andythenorth> also Samu spent about 1 year trying to solve rock and hard place for vanilla industries
20:53:57 <andythenorth> requirements: (A ! B) && (A & B)
20:54:01 <andythenorth> or something
20:54:19 <andythenorth> I skipped the formal logic module in my philosophy degree πŸ˜›
20:55:38 <andythenorth> also in FIRS, the repeatably problematic industries are ones I'm trying to forcibly co-locate near other industries
20:56:19 <JGR> The industry placement mechanism just tries random tiles
20:56:34 <andythenorth> so if random is random
20:56:40 <andythenorth> and we hit the limit for num tries...
20:57:14 <andythenorth> hmm what about a pre-condition for which random tiles to try πŸ˜›
20:57:23 <andythenorth> that would also aid placing near .... eg. towns, or sea
20:57:34 <andythenorth> reduce the set of tiles
20:57:51 <andythenorth> could probably just take them out of the map array directly?
20:58:03 <andythenorth> try random from sea tiles
20:58:16 <andythenorth> try random within n of town
20:58:39 <andythenorth> the map array is linear though? πŸ˜›
20:59:13 <JGR> Is that functionally different from adding a "is within n of town" to the very start of your placement callback?
21:00:06 <JGR> The number of tiles to try has to be bounded somehow
21:00:46 <andythenorth> the candidate set of tiles would have to meet the pre-condition
21:00:56 <andythenorth> so there's a reduced chance of not placing
21:01:08 <andythenorth> they might still not be valid
21:01:13 <JGR> How would you form this candidate set?
21:01:28 <andythenorth> landscape class or something
21:01:30 <andythenorth> wavey hands
21:01:39 <JGR> You'd most likely still be doing random sampling then filtering
21:02:02 <andythenorth> for performance reasons?
21:02:13 *** Flygon has quit IRC (Quit: A toaster's basically a soldering iron designed to toast bread)
21:03:42 <JGR> Yes. it wouldn't make much sense to computer "distance to nearest town" for every tile on the map just to form a huge list of candidate tiles
21:04:50 <andythenorth> ok, well ideas come and ideas go πŸ™‚
21:06:56 <JGR> If you want an industry to be placed next to another one or next to a town, it would make sense to have a special mode to pick the "random tile" differently though
21:07:51 <JGR> e.g. pick a random industry of the desired type, or a random town, then add some random offset to that within the desired radius
21:08:37 <andythenorth> hmm
21:08:46 <andythenorth> should I just do that in GS?
21:08:58 <andythenorth> that was another route I considered, just doing all this in GS
21:09:18 <andythenorth> I've been advised not to try and generate all industry placement in GS though
21:09:44 <JGR> GSs are kind of a pain from an actual player point of view
21:10:38 *** dP has joined #openttd
21:10:38 <dP> citybuilder scripts do it in gs
21:10:49 <JGR> I'd be inclined to just add an extra property and then implement the required functionality in the main codebase (but that is my stock answer to a lot of things)
21:12:05 <andythenorth> might be the right answer in this case
21:12:11 <andythenorth> callback not property πŸ˜›
21:12:30 <andythenorth> pre-defined range of valid returns
21:12:39 <dP> grfs aren't suitable for this kind of general logic, you'll have to add special generation for every usecase
21:13:03 <dP> i.e. next to town, lake, whatever someone comes up with next
21:13:38 <JGR> The existing placement callback is the general case already
21:13:46 <dP> in my cb I have powerplants spawn only in a specific frame around the town and it's not even symmetrical
21:13:50 <andythenorth> lol imagine feeding every tile to a callback, and letting grf do the filtering
21:13:57 <andythenorth> a huge iterator
21:14:02 <DorpsGek> [OpenTTD/nml] glx22 updated pull request #247: Add: Translate ternary operator into varact2 if needed
21:14:38 <JGR> The problem is that variables which are typically used in placement callbacks are often rather expensive
21:15:11 <DorpsGek> [OpenTTD/nml] andythenorth commented on pull request #247: Add: Translate ternary operator into varact2 if needed
21:15:35 <dP> problem is when you need too specific placement rules so random choice takes too long
21:22:28 <andythenorth> so "random offset from defined entity" seems quite good
21:22:40 <andythenorth> would solve a couple of common FIRS cases at least
21:23:31 <andythenorth> providing an industry as the constraint is problematic if the count of that industry on the map is 0 πŸ™‚
21:23:39 <andythenorth> FIRS has to control the order of certain industries via ID
21:27:08 <JGR> Other than the first instance, industries aren't all generated in ID order
21:27:35 <DorpsGek> [OpenTTD/nml] glx22 commented on pull request #247: Add: Translate ternary operator into varact2 if needed
21:28:09 <JGR> The code just seems to randomly sample a cumulative probability distribution to get the industry type for each attempt
21:32:48 <andythenorth> yeah it's the first one that I had to account for
21:40:01 <andythenorth>
21:40:01 <andythenorth> wonder if it's worth teaching the sub-layout to allow building at different tile height to main part
21:40:18 <andythenorth> not sure it will vastly increase number of valid places
21:40:59 <TallTyler> I think it would
21:41:12 <TallTyler> "Vastly" is subjective though πŸ™‚
21:42:55 <andythenorth> it's probably not hard, just faff
21:43:39 <andythenorth> this chlor-alkali plant is particularly suited to split layouts
21:43:41 <andythenorth>
21:43:55 <andythenorth> it's got strong company colour and is quite distinctive shape
21:44:01 <andythenorth> and it's really dense appearance
21:45:42 <andythenorth>
21:47:21 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
21:48:23 <DorpsGek> [OpenTTD/nml] JGRennison commented on pull request #247: Add: Translate ternary operator into varact2 if needed
21:53:42 <andythenorth>
21:53:42 <andythenorth> allow non-flat chlor-alkali plants? πŸ˜›
21:54:16 <andythenorth> sprites look ok πŸ˜›
21:56:25 <andythenorth>
21:56:25 <andythenorth> very swiss?
21:56:56 *** supermop_toil has quit IRC (Read error: Connection reset by peer)
21:57:31 <DorpsGek> [OpenTTD/nml] glx22 commented on pull request #247: Add: Translate ternary operator into varact2 if needed
21:59:34 <reldred> andythenorth: Maybe needs some custom foundation sprites? Like some industrial’ish looking ramps up/down
22:00:04 <reldred> Would only have to be the leading faces
22:00:16 <andythenorth> industries on slopes are banned because they are impossible to serve with long trains
22:00:19 <andythenorth> πŸ˜‰
22:00:27 <reldred> Sure, if you’re a bitch
22:00:29 <reldred> πŸ˜›
22:00:51 <reldred> I route 10+ length trains through mountains all the time
22:01:01 <reldred> You know what sorta maps I play
22:03:21 <JGR> Road vehicles also exist, if the terrain is too steep you can use those to shuttle cargo back and forth
22:04:59 <reldred> I’ve also done that in particularly spicy areas
22:06:02 <glx[d]> oh the industries in 2 latest screenshots are not that bad, there's room to serve them
22:07:01 <glx[d]> and the 2 parts seem to help too
22:07:05 <reldred> Yeah, allowing them to be built over hills though like that I do see producing some ugly results though, so sub buildings on different height levels I’d say is okay but otherwise flat.
22:09:19 <glx[d]> it's better than having an industry surrounded by hills with transmitters on top πŸ™‚
22:11:48 <TallTyler> I like it!
22:13:01 <andythenorth> FIRS used to do this
22:13:06 <andythenorth> it's nice if you want eye candy
22:13:10 <andythenorth> it's crap for gameplay
22:13:28 <andythenorth>
22:14:05 <andythenorth> it's often pathological trying to station build, because the industry prevents most nearby terraforming
22:15:02 <andythenorth> I guess I'll have to encode the flatness into each sub-layout πŸ˜›
22:15:03 <andythenorth> oof
22:15:04 <glx[d]> yeah this one is hard
22:15:55 <glx[d]> but if all tiles are accepting it's doable
22:17:25 <dP> slopes that steep are kinda crap for the gameplay either way
22:19:07 <JGR> Having some genuine obstacles makes things more interesting, in my opinion at least
22:20:44 <andythenorth> hmm encoding the flatness requirement is non-trivial πŸ˜›
22:20:58 <andythenorth> it needs an offset to the north-most corner of the layout
22:21:20 <andythenorth> currently it just uses 0,0 as a proxy, which can be inaccurate, but is simple
22:23:12 <andythenorth> oh there's a relative_coord(x, y) expression
22:23:15 <andythenorth> that will help
22:24:23 <glx[d]> ideally you'd want each sub building on almost flat spot, but allowing them to be on different heights and totally ignore how the space between them is
22:24:31 <andythenorth> yes
22:24:44 <glx[d]> I imagine the pain to implement that πŸ™‚
22:25:39 <andythenorth> I think it's 3 switches and some python
22:25:49 <andythenorth> there are only 2 sub layouts to consider
22:26:00 <andythenorth> but I proceduralised all the location checks πŸ˜›
22:26:21 <andythenorth> so I have to un-proceduralise some part of it, because it's unique to each industry
22:28:26 <andythenorth> tomorrow πŸ™‚
23:02:31 *** nielsm has quit IRC (Ping timeout: 480 seconds)
23:31:22 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)