IRC logs for #openttd on OFTC at 2019-09-27
⏴ go to previous day
00:19:49 *** Lejving has joined #openttd
01:16:55 *** Thedarkb-X40 has joined #openttd
01:17:22 *** glx is now known as Guest3477
02:42:52 *** Sheogorath has joined #openttd
07:08:26 *** HerzogDeXtEr has joined #openttd
08:06:35 *** WormnestAndroid has quit IRC
08:28:03 *** WormnestAndroid has joined #openttd
08:53:51 *** Wormnest_ has joined #openttd
09:31:28 *** ChanServ sets mode: +v tokai
10:53:17 *** andythenorth has joined #openttd
11:01:41 <Samu> multiple rail types in the same tile
11:04:53 <andythenorth> rewrite nml to be fast then? o_O
11:05:11 <Eddi|zuHause> sure, how fast you want it to be done?
11:05:50 <Eddi|zuHause> it's just a compiler. compilers are a solved problem.
11:06:33 <Eddi|zuHause> my compiler construction professor always said this lecture should be called "history of computer science", because all this stuff was basically done in the 70s
11:07:56 <andythenorth> how many nml users are there?
11:08:09 * andythenorth wonders where the utility is in spending time on a rewrite
11:08:14 <andythenorth> is it just me using it?
11:08:50 <Eddi|zuHause> total users, active users, future users, concurrent users?
11:09:46 <Eddi|zuHause> plenty, fewer, dunno, meh?
11:11:18 <andythenorth> I could design around it
11:11:37 <andythenorth> splitting grfs up seems quite plausible
11:11:58 <Eddi|zuHause> for whatever design you can come up with, you'll probably be the only user, ever
11:12:12 <andythenorth> well that simplifies design :)
11:12:25 <andythenorth> Iron Horse Pont Engines
11:12:31 <andythenorth> Iron Horse Pony Wagons
11:12:35 <andythenorth> Iron Horse Pony Coaches
11:12:43 <andythenorth> all individual grfs
11:12:50 <Eddi|zuHause> that makes literally no sense
11:15:42 <andythenorth> will be substantially faster
11:16:04 <andythenorth> and the work to do it will balance very well against the time saved
11:16:08 <andythenorth> so it's logical conclusion?
11:17:39 * andythenorth wonders about the user experience
11:29:51 *** blackli0nxx has joined #openttd
11:49:39 <Eddi|zuHause> user experience will be horrible and your maintenance costs will skyrocket
11:54:48 <andythenorth> why does my code get shorter as I handle more cases, more generically? :P
11:55:14 <andythenorth> long if-else blocks -> loops
11:55:20 <andythenorth> long blocks of declarations -> loops
11:56:06 * andythenorth knows why this is :P
11:56:19 <andythenorth> just procrastinating from a fiddly problem
13:59:20 <Eddi|zuHause> so, how's that procrastinating working out for you?
14:06:26 <Samu> I actually posted a fix for this one but nobody cares
14:09:34 <andythenorth> Eddi|zuHause: fixed all the intermodal container rendering
14:09:43 <andythenorth> and all the rules for which cargo shows which container
14:09:45 <andythenorth> with randomisation
14:09:50 <andythenorth> and class based fallback
14:09:57 <andythenorth> just need to draw the containers now :P
14:10:25 <Samu> I loaded a save, and now my vehicles are using full load orders, because it rerandomized some parameters on load
14:10:38 <Samu> shouldn't be using full load setting
14:10:52 <andythenorth> the final loop for was only 30 LOC, of which 10 LOC are whitespace or comments :P
14:11:27 <andythenorth> discovering what the rules needed to be was like 1 day of thinking, and progressively deleting special cases
14:14:37 <Samu> it's still ultimately, a 'start_date' issue PR
14:14:58 <Samu> is that why nobody did anything about it?
14:16:09 <Samu> i really wish I had my computer
14:16:37 <LordAro> people are busy, and ultimately the problem is not that interedting
14:18:54 *** andythenorth has joined #openttd
15:00:44 <Eddi|zuHause> so, i got this passenger train, which is barely reaching top speed on flat routes, and struggles uphill... options are a) double head it, b) electrify it, or c) ignore it, because it'll run against a red signal from a freight train anyway, and there's no scheduling
15:03:13 <andythenorth> but if (c), does it accelerate slowly?
15:03:28 <andythenorth> recovery time from checks is important w.r.t flow
15:04:06 <Eddi|zuHause> i'll try a) first, i think
15:04:30 <Eddi|zuHause> there should be a shunting option where i add a second engine uphill
15:04:32 <andythenorth> are you contended for train length?
15:04:42 <andythenorth> (a) is best if not contended, lower capital cost
15:04:45 <Eddi|zuHause> i have some room spare for length
15:04:49 * andythenorth has no idea how TF works
15:05:15 <Eddi|zuHause> well, you can make trains as long as you want, and it gives you options for platform length
15:05:24 <Eddi|zuHause> ... which you can increase with mods
15:07:30 <Eddi|zuHause> in my current game i'm using 400m as my baseline platform length, and this train is about 300m long
15:09:09 <Eddi|zuHause> tbh, electrifying is ridiculously cheap in this game, and would be probably the best choice from raw gameplay perspective
15:09:48 <Eddi|zuHause> but lack of shunting also would mean electrifying most of the network at once
15:11:34 *** firewire1394 has joined #openttd
16:22:07 *** Wormnest_ has joined #openttd
16:22:26 *** firewire1394 has joined #openttd
16:59:29 *** HerzogDeXtEr has joined #openttd
17:10:07 <Samu> road vehicles should slow down to breakdown, like trains do
17:13:16 <Samu> ships could perhaps have a breakdown-mix between that of a road vehicle and that of an aircraft
17:13:42 <Samu> moves at reduced speed during breakdown, and lasts only a few seconds
17:14:07 <Samu> while emiting smoke or so
17:17:12 <Eddi|zuHause> a ship of any decent size doesn't really slow down if it "breaks"
17:17:38 <nielsm> at least it takes a long while for it to
17:18:33 * peter1138 mumbles about flight dynamics of Elite Dangerous...
17:19:30 <Eddi|zuHause> there's preservation of momentum, and ships have excessive amounts of momentum
17:20:32 <Eddi|zuHause> so if anything, a broken down ship would become uncontrollable and ram something in its path
17:22:20 <Eddi|zuHause> add to the other bunch of PRs that, realistically, i'm never going to finish? :p
17:22:56 <andythenorth> do I need lunch?
17:23:25 <Eddi|zuHause> i've built this sloped route, and have no clue where to put the station into the mountain
17:25:05 * andythenorth awaits Eddi|zuHause twitch stream
17:25:23 <Eddi|zuHause> ... if anyone wants to watch me play games at 3-7fps?
17:25:49 <Eddi|zuHause> (not accounting for additional load of compressing videos)
17:26:09 <Eddi|zuHause> away for the weekend
17:37:27 *** firewire1394 has joined #openttd
17:52:47 *** Progman has joined #openttd
18:41:39 *** frosch123 has joined #openttd
18:42:30 *** WormnestAndroid has quit IRC
18:44:08 *** WormnestAndroid has joined #openttd
18:48:01 *** Wormnest has joined #openttd
18:56:14 *** WormnestAndroid has quit IRC
18:56:54 *** WormnestAndroid has joined #openttd
19:01:15 <Samu> some kind of vehicle that walks underground
19:01:40 <Samu> submerges water and land
19:02:57 <Samu> makes the xyz coordinates relevant
19:03:36 <Samu> instead of a X or Y tunnel, it can also walk in Z tunnels
19:05:07 <Samu> maybe call it submarine if it walks on water
19:05:13 <Samu> elevators and submarines :p
19:06:20 <Samu> negative levels for water
19:07:02 <Samu> negative heights, they can move ultra fast, kinda like aircraft speed
19:07:19 <Samu> they submerge at the speed an aircraft lift offs
19:10:30 <Samu> there may be a hybrid engine that can be a submarine and elevator at the same time, walks on both terrains
19:12:01 <Samu> subport seems like the correct name
19:14:49 *** Wormnest has joined #openttd
19:17:13 <Samu> generally, hybrid vehicles move faster under water, aircraft speed kind of speed, and move slower under ground, truck speed levels
19:38:05 <Samu> x and y tunnels would be pretty much water tunnels, i've heard ppl tried water tunnels for ships somewhere
19:39:10 <Samu> the difference would be that the entrance would be in a negative-height coordinate
19:41:07 <Samu> at least graphic-wise, then they move up to height 0, already hidden from view
19:42:20 <Samu> and the tunnel then goes to either direction from there: x, y or z, maybe zx and zy too for climbing terrain underground
19:43:02 <Samu> hum, no enthusiasm around here
19:45:19 <nielsm> would require the big landscape rewrite to be RCT-like instead of a flat map
19:48:18 <Samu> just having fun thinking
19:49:39 *** firewire1394 has joined #openttd
19:50:27 <Samu> walking above the ground you say?
19:50:43 <Samu> i wasn't thinking of that
19:50:56 <Samu> it's all under the ground or under water
20:22:50 *** Wormnest has joined #openttd
21:21:00 <nielsm> the servers listing webpage, what's the criterium for which version gets listed first? just which one has the most servers?
21:22:51 <frosch123> so yeah, looks like that has not been mirgrated to git :)
21:22:59 <nielsm> it's listing 1.9.2 servers first atm, and then everything else after a break
21:23:32 <frosch123> ok, let's look at the source :p
21:23:51 <glx> it's still using old architecture too
21:26:29 <frosch123> oh, it's hand coded into the website ..
21:26:32 <glx> easy to spot for me, it's using the openttd logo with ipv6 mark
21:26:37 <frosch123> i did not know this needs manual mainteneance
21:26:58 <frosch123> i wonder who updated it the past years then
21:27:30 <frosch123> ah, no, it was automated in the past
21:27:50 <frosch123> someone (likely tb) inserted the hard coded number when disabling part of the old infrastructure
21:28:44 <frosch123> let's see whether django autoreload works
21:28:48 <frosch123> or whether i need to kick something
21:34:29 <nielsm> start 1.9.3 with dmusic driver on softsynth and leave it playing all night (speakers off)
21:34:34 <nielsm> see how well it works in the morning
21:40:21 *** supermop_pdx has joined #openttd
21:46:58 <DorpsGek_III> [OpenTTD/OpenTTD] nielsmh commented on pull request #7639: Fix: Total reset of DMusic driver once every hour https://git.io/JeZCP
21:47:42 <frosch123> nielsm: 1.9.3 is now at the top
21:47:54 <frosch123> i had to kick something
21:48:15 *** WormnestAndroid has quit IRC
21:48:53 <nielsm> hope to close that PR tomorrow
21:55:24 *** WormnestAndroid has joined #openttd
22:43:51 <andythenorth> Unable to allocate ID for [random]switch, sprite set/layout/group or produce-block. Try reducing the number of such blocks.
22:48:52 <andythenorth> what did I do? :P
22:51:22 <peter1138> Disturbs you very much?
22:52:15 <andythenorth> goes it throw out?
22:53:47 <Samu> the ppl here say I listen to terrifying music :(
22:55:39 <milek7> it happens with C too ;P
22:55:46 <milek7> fatal error C1061: compiler limit: blocks nested too deeply
22:56:19 <frosch123> andythenorth: reorder your switching
22:56:52 <frosch123> put switching referencing each other closer together
22:57:22 <andythenorth> am I out of varaction 2 ids? :P
22:57:40 <frosch123> order them depth-first, not breadth-first
22:58:16 <Samu> 11 views since 2014... I really must have weird taste in music
22:59:00 <frosch123> nmlc info: Concurrent spritegroups: 241/256 ("generated/iron-horse.nml", line 59783) <- at line 59783 you have 241 switches which were defined before line 59783 and which are referenced after lint 59783
22:59:37 <frosch123> move them closer together, so you do not need to reference a switch from ages ago
22:59:55 * andythenorth frowns at template loops
23:00:14 <frosch123> looking at the generated source
23:00:25 <frosch123> you probably define all spritegroups first, and then all switches
23:00:31 <frosch123> make them instead alternate
23:00:38 <frosch123> interlave spritesets and spritegroups
23:00:44 <andythenorth> the spritegroups are shared globally, not per vehicle
23:00:48 <andythenorth> due to compile time
23:01:00 <andythenorth> but the switches are per vehicle
23:01:47 <glx> ignore compile time if you can't compile at all ;)
23:02:02 <frosch123> so, all vehicles have the same graphics?
23:02:18 <andythenorth> it's a layer showing containers
23:02:23 <andythenorth> containers all same for all vehicles
23:02:49 <frosch123> why do you have those switches for each vehicle then? shouldn't they also be global
23:03:13 <andythenorth> I might have been confused about random triggers
23:03:15 <frosch123> and shouldn't there only be one global switch "containers" in the end?
23:03:41 <andythenorth> I need the random trigger on new load
23:03:49 <andythenorth> does that work with a global switch?
23:04:08 <Samu> WormnestAndroid: what kind of map size do you run your AI tests?
23:04:54 <Samu> I can't reproduce WormAI outperforming other AIs... ever
23:05:00 <andythenorth> ok thanks, I'll refactor
23:06:55 <Samu> "ever" is a strong statement, but I'm retrying your AIs on that screenshots
23:07:06 <Samu> WormAI oscillates 4th and 5th in profits
23:11:04 <andythenorth> a lot of these random switches only have one entry :P
23:11:30 <frosch123> yes, and you need to put the switch that is referenced by that entry directly before it
23:12:31 <frosch123> switch/spritegroup/spriteset, whatever is referenced there
23:13:06 <frosch123> or you could teach nml to reorder them when they are not separated by "if" or similar
23:13:23 <frosch123> but that's probably complicated :p
23:16:18 <Samu> well, i didn't start in 1980, maybe that's why
23:16:45 <DorpsGek_III> [OpenTTD/OpenTTD] abmyii opened pull request #7752: Prevent sounds being produced by inactive industries https://git.io/JeZlF
23:44:09 *** WormnestAndroid has quit IRC
23:44:38 *** WormnestAndroid has joined #openttd
23:46:15 <Samu> preferred max distance 2921, is that manhattan? that's absurdly too high
23:46:35 <Samu> there's breakdowns enabled
23:48:06 <andythenorth> hmm that might have worked
23:49:46 <andythenorth> Concurrent spritegroups: 155/256
continue to next day ⏵