IRC logs for #openttd on OFTC at 2020-08-03
⏴ go to previous day
00:32:57 <b_jonas> what the heck? how did I just crash a train? I wasn't editing the tracks or signals there
00:33:03 <b_jonas> can you crash a train by just reversing it?
00:33:48 <b_jonas> it was stuck in some strange state partway out of the depot, stopped but not at a signal, so I tried to reverse it
00:33:57 <b_jonas> and bang, a train crash
00:35:50 <b_jonas> and apparently the other train it crashed with is still inside the depot, with a red cross next to it indicating that it's stopped
00:37:08 <_dp_> you shouldn't be able to but I think there is indeed some bug involving depots
00:37:35 <_dp_> if you can figure out how to reliably reproduce it it would be awesome
00:38:28 <b_jonas> sadly I don't think I can reproduce it, I'm in the middle of replacing many trains and saved before that, so I can't reproduce the operations
00:39:02 <b_jonas> does it help if I send a save afterwards or something?
00:39:32 <b_jonas> it's not that I mind this train crash much, I have a ton of money
00:39:54 <b_jonas> (time to build a vacuum train network after this probably)
00:40:43 <_dp_> ideal would be save immediately before reverse
00:43:21 <b_jonas> although it looked like the glitch started a bit earlier
00:43:24 <FLHerne> b_jonas: If something is stuck in a strange state, save before trying to fix it :p
00:43:43 <b_jonas> I should have thought of it earlier
00:44:31 <b_jonas> but trains and signals are weird, I don't always recognize what's only my misunderstanding and what isn't
00:48:00 <b_jonas> does it help to reload after I make the save like that? or does it hurt more because if reloading fixes the glitched state then I've lost more info?
00:48:12 <b_jonas> I guess I shouldn't reload. stupid question.
00:48:18 <b_jonas> not initially at least
00:49:50 <glx> usually reversed train will crash in a train waiting at a signal
00:57:38 <b_jonas> I'm using path signals in case it matters
00:57:46 <b_jonas> so I think the depot has a path signal too
00:58:10 <Timberwolf> What happens to `curv_info_prev_cur` when the previous vehicle disappears in a depot? It looks like it gets clobbered to something other than 0, but I'm not sure what.
00:58:54 <Timberwolf> (I'm actually using `curv_info_prev_next` with `vehicle_curv_info`, but it seems to be the disappearance of the previous vehicle that causes weird stuff to happen)
00:59:37 <Eddi|zuHause> Timberwolf: i'm not sure, but here is "vehiclr_is_hidden" that should indicate "in depot/tunnel"
00:59:45 <Timberwolf> Actually, might be `other_veh_z_offset` that's causing my issue.
00:59:48 <b_jonas> now I'll have to wait and see how the steel and goods network works, now that I've separated them to give them even more capacity
01:00:13 <Timberwolf> Now I think about it, that sounds more like something that might get a bit weird if the first vehicle has disappeared.
01:08:04 <Eddi|zuHause> what do you mean?
01:10:10 <Timberwolf> I don't know the inner mechanics, but I have two checks for "should I display 3 sprites or 1 long sprite" in Trains - one is "is this vehicle in a curve" and the other is "is the z-index relative to adjacent vehicles != 0". One of those is not returning the expected result when driving in to a depot, and I have a hunch it's because the first vehicle is hidden and therefore it's not really valid to ask
01:10:16 <Timberwolf> it about curvature or z-index.
01:10:44 <Timberwolf> I suspect just checking whether the front vehicle is hidden is the sensible option.
01:12:11 <Eddi|zuHause> i thought i used the hidden flag in CETS, but i can't find any evidence of it
01:13:47 <Eddi|zuHause> i suppose what held me off was that i needed two more types of slice sprites, which included the front or back half of the vehicle
01:14:36 <Eddi|zuHause> btw, is your code public somewhere?
01:14:51 <Timberwolf> Yes, that's how I solved it. Depending on situation I display the vehicles either as [front, middle, back] or as one long middle vehicle.
01:15:15 <Eddi|zuHause> "you" solved that... right :)
01:15:29 <Eddi|zuHause> that's exactly what CETS did
01:15:40 <b_jonas> you're making extra-long articulated trains?
01:15:52 <Timberwolf> Heh. I took your suggestion, but I created it independently. Might be interesting to see how similar they are.
01:15:52 <Eddi|zuHause> i mean, it would additionally need a "front+middle" or "middle+back" sprite
01:18:52 <Eddi|zuHause> i would link you my code, but seems devzone is out of order
01:19:26 <Timberwolf> I was wrong, it does seem to be curvature info that's the problem. But yay, I'm at a point I can fix the depot issue now.
01:20:48 <Eddi|zuHause> it's much older than nml's support for the "other_veh_*" variables
01:21:02 <Eddi|zuHause> so it queries var[0x62] directly
01:22:10 <Eddi|zuHause> this will be replicated with "SLICE" being -1, 0 and 1. and i constantly forget which one is front and back
01:22:31 <Timberwolf> I remember looking at this, I never realised what `slice_group` referred to until now.
01:23:19 <Eddi|zuHause> those are the separated front/middle/back sprites, instead of the full vehicle
01:26:03 <Eddi|zuHause> "other_veh_z_offset(-1) + other_veh_z_offset(-2)" this looks fishy, missing abs()?
01:30:45 <Timberwolf> It's not supposed to be a straight addition but I am pondering whether that's correct - I only do hill climbing/descending sprites based on the difference between front and middle vehicle for lengths > 4, I did work this out with a sheet of paper but now I have doubt over whether that is correct or just happens to work in the scenarios I tried.
01:32:44 <Eddi|zuHause> "switch_{{.}}_height_front" where you only query one offset looks wrong as well
01:32:56 <Eddi|zuHause> it should probably be both
01:33:42 <Eddi|zuHause> i can't judge whether your offsets are correct, since i've never worked with these nml variables before
01:33:45 <Timberwolf> That one is deliberate, I use both for short vehicles but with long ones I found they stay in the "climbing" state too far forward if you include the rearmost vehicle.
01:35:10 <Timberwolf> If one of the vehicle sections is L1 then it gets unreliable and you have to use both, but at that point the vehicle is short enough it doesn't matter so much if it stays "climbing" for a few pixels more.
01:35:14 <Eddi|zuHause> i can't see this working properly if both front and back aren't in sync when deciding whether to draw full or partial sprites... suddenly parts will disappear, or be drawn double
01:41:48 <Timberwolf> There's one situation where it draws double, which is the "hill crest" situation, but it's not noticeable for anyone who isn't carefully moving trains forward with the sprite aligner open.
01:42:26 <Timberwolf> If you find a track configuration which results in a noticeable artifact, that would be very useful :) - means I will have missed something from my sheet of paper.
01:54:04 <Eddi|zuHause> Timberwolf: what about an opposite crest? i.e. \/ (valley) instead of /\ (hill)?
01:54:38 <Eddi|zuHause> (or am i misunderstanding what a "crest" is?)
01:55:30 <Timberwolf> "crest" here is /-. But you're right about the abs(..), I went back to my initial workings out and it should be that, I just want to know "Is either the front or mid unit different".
01:55:46 <Timberwolf> Also I think checking the state of the rear unit will prevent that overdraw.
01:56:12 <Timberwolf> /\ and \/ work fine, need to check with sprite aligner to see it's not overdrawing anything.
01:56:45 <Eddi|zuHause> in all 4 directions?
01:57:51 <Eddi|zuHause> maybe i'm missing something, but your way seems incomplete
01:58:05 <Eddi|zuHause> i haven't actually tried anything, though
02:00:24 <Timberwolf> Yep, now I've added abs() and also checked the rear vehicle, it's fine with nothing getting double-drawn either.
02:08:26 *** Speedy` has joined #openttd
04:34:18 *** Flygon_ has joined #openttd
05:07:35 *** Smedles has joined #openttd
07:16:07 *** snail_UES_ has joined #openttd
07:43:05 *** HerzogDeXtEr1 has joined #openttd
08:05:45 *** gelignite has joined #openttd
08:11:42 *** sla_ro|master has joined #openttd
08:33:31 *** andythenorth has joined #openttd
09:28:47 *** andythenorth has joined #openttd
10:56:11 <andythenorth> frozen date + new vehicles are either 'granted' as GS reward, or 'researched' (with GS offering a choice of which to research)?
10:56:22 <andythenorth> blah blah complications with upgrading wagons and stuff :P
10:56:56 <andythenorth> I want to play a game which develops routes instead of replacing all the trains every 20 years :D
10:57:09 <andythenorth> but still has 'new' sometimes
10:58:11 <planetmaker> can a GS modify the game date?
11:07:29 <Timberwolf> You only escape the current year when every local authority you provide service to rates you as "outstanding".
11:08:42 <_dp_> Timberwolf, that's basicaly what highscoring on btpro cb is like :p
11:09:37 <_dp_> except house luck is more important than authority rating
11:10:16 <_dp_> you get small house - you reset :p
11:24:21 <Timberwolf> I'm now pondering a chaos monkey GS. Randomly sells vehicles, joins bits of your network, deletes bits of track, builds a 16-tile long train and injects it into your network...
12:17:12 *** arikover has joined #openttd
13:22:41 *** Carolusclen has joined #openttd
13:57:12 <TrueBrain> found an old Macbook Pro .. it appears I once owned one .. I cannot remember my password for the life of me .. but it still boots
13:57:17 <TrueBrain> it has to be at least 5 years old ..
14:14:55 *** D-HUND is now known as debdog
14:38:34 <andythenorth> TrueBrain reinstall the OS :P
14:38:46 <andythenorth> or do the thing where safe mode lets you reset the user
14:38:50 <andythenorth> such security, such Apple
14:47:47 *** gelignite has joined #openttd
14:58:28 *** b_jonas has joined #openttd
15:10:33 *** snail_UES_ has joined #openttd
15:16:04 <TrueBrain> Yeah, requires an install CD ... that is not happening :D
15:16:44 <LordAro> TrueBrain: and you were so looking forward to being an OSX dev as well
15:17:34 <andythenorth> it doesn't just do the web OS recovery? :o
15:19:27 <glx> could it be 10.12.6 by luck ?
15:38:37 *** WormnestAndroid has quit IRC
15:38:50 *** WormnestAndroid has joined #openttd
17:24:33 <andythenorth> does it boot to disk recovery or anything?
17:37:51 <_dp_> just removing 1.5 multiplier from vehicles would already improve it tremendously :p
17:46:05 *** longtomjr has joined #openttd
17:48:25 *** Wormnest has joined #openttd
18:02:52 <nielsm> I don't even know how it's calculated right now
18:04:57 <nielsm> okay yeah vehicle value * 1.5 is weird
18:05:06 <nielsm> I guess that's just CS logic
18:08:14 <nielsm> hmm I would really like to have actual income and actual expenses on vehicles, i.e. without any transfer credits fakery
18:09:16 <nielsm> or at least revenue and expenses separated out
18:14:06 <supermop_Home_> well now i went to shoot a video of the skipping record i got from discogs, to send to the seller
18:14:12 <supermop_Home_> and it didn't skip..
18:15:20 <b_jonas> nielsm: you can configure your order lists to unload instead of transfer to do exactly that
18:15:50 *** frosch123 has joined #openttd
18:17:06 <nielsm> uh yes I do know how the orders system works
18:17:19 <nielsm> I'm not talking about seeing values for specific vehicles in a specific game
18:17:26 <nielsm> I'm talking about having it available in general
18:35:23 *** HerzogDeXtEr has joined #openttd
18:44:51 *** Progman has joined #openttd
19:18:25 <frosch123> how long until dp is banned from reddit for posting his own discord link under the reddit mod's discord link? :p
19:41:42 *** gelignite has joined #openttd
19:44:17 <andythenorth> transfer credits with cdist do kill some of my vehicles :D
19:44:30 <andythenorth> and being obsessively completionist, I hate seeing the red numbers
19:45:56 <frosch123> do you have siblings? just imagine those vehicles are run by them
19:48:06 <andythenorth> I am busy trying to imagine that train length 4.8 in depot really doesn't matter
19:48:18 <andythenorth> even when the value has odd rounding errors
19:48:38 <andythenorth> hmm, could we just round it up to integer tile length? :P
19:50:57 <frosch123> so, for the true sandbox experience all labels should be editboxes?
20:07:27 *** sla_ro|master has joined #openttd
20:07:31 <_dp_> frosch123, well if they want to be official they should be more welcoming to others :p
20:08:15 <_dp_> also I kinda mention citymania like every other citymania word there citymania and noone seems citymania to bother so far xD
20:08:52 *** ericnoan has joined #openttd
20:09:54 <TrueBrain> frosch123: btw, fun fact, the Discord servers run JGRPP :)
20:11:42 <TrueBrain> and according to them spnda is an OpenTTD developer? Did I miss that?
20:11:51 <frosch123> don't they have multiple servers? or is the channel specific about one of them
20:13:27 <_dp_> afact it's not very formal
20:13:43 <_dp_> i think jgrpp and vanilla servers are run by different people actually
20:13:59 <TrueBrain> I am always amused by how scattered communities really are
20:14:10 <TrueBrain> just a general observation btw, not a specific one
20:14:29 <TrueBrain> under one label, many different things happen, without anyone really knowing what exactly
20:14:44 <TrueBrain> it does confuse me most of the time :)
20:15:04 <TrueBrain> I notice this more with reddit communities; main reason I do not really do anything with reddit
20:15:53 <TrueBrain> takes a really deep dive to understand who is part of what where and how :)
20:16:54 <_dp_> well, sure, doesn't change the fact that I for example can't even imagine what half of that stuff does :p
20:17:07 <TrueBrain> but in general OpenTTD has been very focused: make the game. And push external everything else
20:17:10 <nielsm> <TrueBrain> and according to them spnda is an OpenTTD developer? <-- well he had three PRs merged
20:17:14 <nielsm> so technically co-author
20:17:26 <TrueBrain> the reason we are not running the forums, for example
20:17:39 <TrueBrain> or any official OpenTTD server, for that matter :)
20:18:20 <TrueBrain> reddit communities often do a lot of things a bit .. and that makes my head spin to understand where the line starts and stops :P
20:18:41 <TrueBrain> and to be clear, it is a me-problem, more than anything else :)
20:20:58 <TrueBrain> _dp_: and that is honestly a good point, some README.mds could use a touch-up to make sure people understand what they are reading :)
20:21:10 <nielsm> really as long as the separate communities understand whether any of the core developers are actually active in their corner and don't assume wrong things if they aren't, fine
20:22:06 <frosch123> nielsm: how is your interview going?
20:22:28 <nielsm> haven't received a reply yet
20:22:43 <TrueBrain> nielsm is going for it? :D
20:22:46 <nielsm> sent my initial response a few hours
20:24:59 <TrueBrain> why did I start with writing a Visitor in Rust, instead of an hello world
20:25:04 <TrueBrain> this is .. difficult to wrap your head around :)
20:25:56 <_dp_> doing hello world is boring after a fifth time xD
20:26:23 <frosch123> tell that to the people who did gnu hello
20:26:37 <TrueBrain> sure, what is their email? :P
20:26:45 <nielsm> hello world is just to smoke test your toolchain
20:28:52 <TrueBrain> but the lack of inheritance in Rust is breaking my brain :(
20:29:54 <TrueBrain> and the fact that enum variants have no type, is just bloody annoying
20:38:59 <andythenorth> frosch123 YES! :) All strings editable, on demand :)
20:39:08 <andythenorth> so I can earn or score whatever I want!
20:47:06 <frosch123> can you compile again?
20:49:34 <andythenorth> no, but I do need to fix that
20:49:52 <andythenorth> I did consider patching the depot train length number :P
21:15:13 *** ChanServ sets mode: +v tokai
21:35:22 <TrueBrain> [8034] Load { register: C, address: Emulation { emulation: U8(12), native: PointerBank { bank: Zero, address: Addition { left: U8(12), right: RegisterD } } }, arg_size: U8 }
21:35:23 <TrueBrain> [8036] BranchCondition { flag: Z, is_set: true, address: PointerBank { bank: PBR, address: U16(8034) } }
21:35:29 <TrueBrain> I am feeling like a mad-man :D
21:36:15 <frosch123> hmm, you have a C register? is that some 8080?
21:41:08 <frosch123> i like how it has way more index registers than arithmetic registers
21:41:15 <frosch123> indirection is important :)
21:41:37 <TrueBrain> it is one WEIRD CPU
21:41:50 <TrueBrain> many ... very ... weird edge cases
21:42:07 <frosch123> all cisc cpu are weird
21:42:44 <TrueBrain> take the Emulation node above .. it picks emulation only IF emulation is on (duh), if the DL is non-zero, and only for a certain commands
21:43:01 <TrueBrain> and the only change? How it wraps in the address-space (8bit vs 16bit wrapping)
21:43:46 <frosch123> no, that's the wrong one
21:43:53 <TrueBrain> the above snippet is also annoying in another way .. it is basically an infinite loop .. it is waiting for an interrupt, where the interrupt handlers changes the value
21:44:09 <TrueBrain> it is the 15th command of the ROM .. you NEED to implement interrupts ASAP when emulating :P
21:44:17 <frosch123> what was the name for the 64kb just after 1MB that you got when you disabled the 1MB wraparound?
21:44:37 <TrueBrain> but yeah, something like that
21:45:14 <frosch123> HMA, high memory area
21:45:51 <TrueBrain> left: Box::new(Node::U8(Box::new(U8 { value: offset }))), <- I hear you like boxes
21:59:02 <Eddi|zuHause> is that what happens when you make assembler "human readable"?
22:24:12 *** virtualrandomnumber has joined #openttd
22:41:52 <_dp_> openttd really misses some extra zoom levels for minimap on 4k
22:43:14 <_dp_> other than that it plays just fine on 4k with just a bit of tweaking
23:04:38 *** Smedles has joined #openttd
23:28:14 <andythenorth> I need a name for a train
23:29:04 <andythenorth> supermop_Home_ ^ picture is bad, but seems to include a pacer?
23:32:16 <andythenorth> did I use that already?
23:40:40 *** Smedles has joined #openttd
23:51:40 *** Smedles has joined #openttd
continue to next day ⏵
