IRC logs for #openttd on OFTC at 2019-04-26
⏴ go to previous day
00:23:57 <DorpsGek_II> [OpenTTD/OpenTTD] Teletoby12000 commented on issue #7543: suddenly it crashes without any reason https://git.io/fjsVw
00:33:48 *** Supercheese has joined #openttd
02:59:04 *** Wormnest has joined #openttd
06:44:16 *** tokai|noir has joined #openttd
06:44:16 *** ChanServ sets mode: +v tokai|noir
07:14:59 *** sla_ro|master has joined #openttd
08:52:59 <LordAro> peter1138: would be interested in your thoughts on 6605
08:57:34 <peter1138> [11] bin/openttd(_Z15MakeDefaultNameI5DepotEvPT_+0x35) [0x55d5ff80f175]
08:59:14 <peter1138> Depot->town is invalid.
09:04:11 <LordAro> peter1138: thst summarises alberth's investigation 4 months ago, but i'm getting a segfault for some reason
09:19:31 *** chomwitt has joined #openttd
09:20:37 <peter1138> If it was a really old savegame without newgrf information saved in it then it would use your newgame newgrf settings and maybe trigger that, but...
09:20:51 <peter1138> Conversion from OTTD savegame without gamelog: version 21, 0
09:32:49 <V453000> hm, so can I articulate a train with the same train 4 times, and then change length of each of the units by a callback? I tried position_in_vehid_chain and position_in_consist but the switches seem to always return 1.... what would I be doing wrong? https://paste.openttdcoop.org/phfwzu8ui
09:33:51 <V453000> I guess I'll just add a new item which has the different length if I can't, but just wondering if I could do it in a cleaner way here.
09:37:55 <LordAro> peter1138: ah, interesting
09:38:07 <LordAro> probably shouldn't do that :p
09:58:25 <Eddi|zuHause> today is one of those days where i should be doing 10 things, and get none of them done...
10:00:55 <Eddi|zuHause> V453000: i think there is also position in articulated chain
10:12:25 <V453000> I tried to add "return" to all the places
10:16:54 <peter1138> Is that how NML does callbacks?
10:16:58 <V453000> somehow it doesn't even get to the switch, even if I put 0: return 8; to the switch, all parts are still 1 unit long
10:17:09 <V453000> generally I guess it is
10:17:26 <peter1138> Looks like it hides the CB selection from you.
10:17:30 <V453000> the other things work, just the length not
10:19:28 <peter1138> why do you set speed/power/weight etc in the callback section when they are static?
10:24:34 <V453000> I don't exactly remember but
10:24:51 <V453000> I generally do use switches for those very often
10:25:23 <V453000> based on leading engine, track type, ...
10:25:41 <V453000> position in consist too :)
10:25:56 <V453000> why here, I don't know
10:26:26 <V453000> is that harmful in any way?
10:26:41 <peter1138> If they vary, it's fine.
10:27:01 <peter1138> Well, every callback will reduce performance a tiny immeasurable bit.
10:28:43 <V453000> I do wonder what the reason could have been
10:28:55 <V453000> maybe just for code consistency that I always define it there, not sure
10:28:58 <peter1138> A static property is "use this value", a callback is somewhat more involved. They are however only called at certain points, but also every time graphics are evaluated behind the scenes the varaction2 chain is testing "is this a callback"
10:29:22 <peter1138> I'd leave them there but comment them out if they're static.
10:29:28 <peter1138> Also you have different values.
10:29:43 <peter1138> speed: 100 vs speed: 110 km/h
10:29:56 <V453000> this code is 7 years old from a time when I had zero idea :) I could try to do a cleanup at some point but I have like 240 items
10:30:03 <peter1138> weight: 80 vs weight: 150 ton
10:30:09 <V453000> yeah, I only maintain the callback values
10:30:24 <V453000> the ones in the property often a copypaste or outdated
10:31:25 <peter1138> Well that's bad practice :p
10:32:16 <V453000> I have no justification, yes it is :D
10:32:17 <peter1138> I wonder if NML does/could detect a static CB result and optimise it out? Probably a bit beyond it currently :-)
10:32:33 <V453000> I just always cared more if it works right, not how the code looks
10:33:07 <Eddi|zuHause> <peter1138> A static property is "use this value", a callback is somewhat more involved. They are however only called at certain points, but also every time graphics are evaluated behind the scenes the varaction2 chain is testing "is this a callback" <-- i don't think there is any GRF where it doesn't do that...
10:33:16 <peter1138> If you saw the varaction 2 (compile to NOF? :D) you'd see what I mean.
10:33:30 <peter1138> Eddi|zuHause, well if it doesn't use callbacks...
10:34:09 <Eddi|zuHause> i haven't seen anyone use no callbacks at all
10:34:49 <peter1138> I imagine a lot of DBSetXL vehicles don't use callbacks. Some do, of course.
10:35:06 <peter1138> Unless he actually updated it :p
10:35:31 <Eddi|zuHause> dbset is full of refit callbacks, capacity callbacks and stuff
10:35:39 <V453000> anyway, any ideas about the length callback please?
10:36:31 <peter1138> length: return switch_railstrong3_length;
10:36:36 <peter1138> is that return meant to be there?
10:36:42 <peter1138> It's not that for the line above.
10:36:55 <Eddi|zuHause> that return seems wrong
10:37:01 <peter1138> Is return correct at all?
10:37:15 <peter1138> or should it be 0: 1
10:37:18 <Eddi|zuHause> return is a valid reserved word
10:37:49 <peter1138> Eddi|zuHause, I mean in the context of cb results.
10:37:53 <peter1138> Switch blocks, etc.
10:38:04 <Eddi|zuHause> i don't know if it's necessary, but it might be needed to avoid confusion
10:38:47 <V453000> I had it out and it didn't seem to work either but I'll remove it :)
10:39:51 <Eddi|zuHause> peter1138: my switches seem to all have return in those places
10:40:20 <V453000> I vaguely remember there's been a NML version which made the returns not required
10:40:27 <peter1138> Ok. I did not know as it wasn't there for the other lines.
10:40:49 <Eddi|zuHause> but the "return <switch name>" thing is very likely wrong
10:41:00 <V453000> for example there speed: return 100 + (100 * count_veh_id(150) / ((num_vehs_in_consist-2) /2) ); it is, but I guess that is not a switch but just a function
10:41:53 <Eddi|zuHause> V453000: nml will create a switch out of that formula
10:42:57 <V453000> ooh something just changed
10:42:58 <peter1138> LordAro, of course, the issue you get with NewGRFs loaded is probably also valid :-)
10:44:39 <V453000> 8 2 2 2 1 gave me train length 0.7 in the depot
10:44:49 <V453000> I'd expect like 0.9 but eh :D
10:45:24 <Eddi|zuHause> V453000: your articulation callback seems odd
10:45:53 <V453000> I use this one all over the place :d
10:46:04 <V453000> oh it shoulnd't be all 1
10:46:06 <Eddi|zuHause> you have 4 times the same line
10:47:31 <V453000> thank you very much, I didn't imagine I could be that stupid
10:47:58 <LordAro> peter1138: doesn't affect the title game, at least, or is that too early for it to matter?
10:48:20 <peter1138> LordAro, depends on what vehicles are used.
10:48:22 <Eddi|zuHause> LordAro: title game never has newgrfs
10:48:42 <peter1138> Eddi|zuHause, it does if you load it as a regular game, and have newgame-newgrfs set up.
10:49:04 <peter1138> Well, the one-true-titlegame, anyway :-)
10:49:12 <peter1138> Obviously the newer ones are fake anyway.
10:49:24 <Eddi|zuHause> nobody needs thise
10:49:43 <LordAro> not sure what it should do instead though, for older games that didn't have grf information
10:49:53 <peter1138> That's not the bug.
10:50:18 <peter1138> The bug is calling delete v during load updates group statistics before the group stuff is all set up.
10:50:33 <Eddi|zuHause> LordAro: there is no correct way to handle old savegames without newgrf information.
10:50:50 <peter1138> Eddi|zuHause, there is incorrect, and less incorrect.
10:51:02 <peter1138> Eddi|zuHause, not-crashing is definitely less incorrect.
10:51:19 <Eddi|zuHause> that would be helpful, indeed :p
10:51:33 <LordAro> peter1138: would forcing the group statistics to be loaded before deleting vehciles be the correct solution?
10:52:25 <LordAro> or does it need a more general "change the order things are loaded" ?
10:53:10 <Eddi|zuHause> without any context, i'd attempt "modify deleting vehicles so it can work without messing with groups"
11:05:59 <V453000> hmmm somehow I expected the offsets to be relative to the center of the bounding box not a corner :D so of course the same sprite on 8/8 and 2/8 aren't in the same spot even if the centers of the bounding boxes are the same
11:06:15 <V453000> guess I'll just define the sprites again with different template :)
11:07:52 <peter1138> Yeah it's somewhat messy
11:07:59 <V453000> in fact that's the only method how I show them now so that's fine, no duplicates at least
11:08:04 <V453000> meh, easy enough to change
11:08:28 <peter1138> That's what led to all the reversing short wagons issues :p
11:08:40 <V453000> reversing short wagons issues?
11:08:49 <peter1138> Offsets being weird and wonky.
11:09:09 <V453000> just wonder which issues do you mean
11:09:22 <peter1138> And then you get stuff like zBase where the offsets are all wrong anyway, and they're just standard length vehicles o_O
11:09:34 <peter1138> s/reversing/flipping/
11:09:40 <Eddi|zuHause> V453000: there's a feature that you can ctrl+click a (non-articulated) vehicle in the depot and it switches directions
11:10:03 <Eddi|zuHause> V453000: that is broken, if you naively do it with shortened vehicles
11:10:26 *** gelignite has joined #openttd
11:10:52 <V453000> I remember when it was moved to a flipping flag :P lost my shit
11:10:57 <Eddi|zuHause> V453000: that's why it's now disabled for shortened vehicles, unless the newgrf explicitly sets a flag that it handles the situation
11:11:35 <V453000> I don't understand what the 28px long thing is for
11:11:46 <peter1138> V453000, original vehicles. TTD original is 28px.
11:12:05 <peter1138> And, indeed, DBSetXL is mostly 28px
11:12:20 <peter1138> I'd just make the depot always be 32px myself.
11:12:32 <peter1138> But apparently people complain when the spacing is too wide.
11:12:37 <peter1138> Especially for default vehicles.
11:12:53 <V453000> seeing the trains intact is nice
11:13:01 <peter1138> I wonder if we should be doing it on 28*4, i.e. subpixels.
11:13:15 <peter1138> Then at least the situation with vehicle 1 & 3 in that picture might be resolved.
11:13:54 <peter1138> Rounding errors add up :/
11:14:06 <peter1138> Could use floats :p
11:15:05 <Eddi|zuHause> ... or we could scrap 32px and enforce proper rotational dimensions
11:15:22 <Eddi|zuHause> which would be more like 24px
11:15:26 <peter1138> I'm aware 32px is wrong :-)
11:15:52 <Eddi|zuHause> and probably invalidate all existing newgrfs, savegames, and everything :p
11:16:11 <peter1138> There was talk of fixing it a while ago when Pikka did the rendered 4x sprites correctly.
11:22:00 <Wolf01> F is not a game, is a computer
11:22:03 *** andythenorth has joined #openttd
11:24:55 <Wolf01> Ahahah, my friend got evacuated from the hotel because a fire alarm early this morning
11:27:44 <planetmaker> sounds like the thing you definitely want to experience when in a hotel :P
11:35:13 <Wolf01> Like when we destroyed the pier when in cruise ship, 14 hours delay to patch the ship and 1 place not visited
11:55:06 <V453000> fuck I realized that in order to do the articulation hack I'd have to duplicate the definition for all 1st gen rail wagons :D
11:55:34 <V453000> I'll see how awful it looks if I put the locomotive on the second 4/8
11:55:44 <V453000> some of these early steamers look to be there anyway
11:56:09 <peter1138> The default short engine does that, iirc.
11:57:52 <peter1138> The lovely MJS 250.
12:00:47 *** Thedarkb-T60 has joined #openttd
12:21:37 <andythenorth> I need to repro the MJS 250
12:22:09 <andythenorth> it looks like a pickup truck though :P
12:27:11 <andythenorth> I did kinda draw the MJS250
12:27:18 <andythenorth> but I can't show anyone because I broke Jenkins
12:29:22 <Artea> I going need learn how to add NewGRF into running dedicated server
12:32:46 <peter1138> You can't add it to a running game.
12:35:40 <planetmaker> you need to edit the cfg and then start a new game from it
12:36:05 <planetmaker> however make sure that the cfg is not overwritten on terminating OpenTTD - thus better use a cfg with a non-default name and ask openttd to load from that upon startup
12:36:20 <planetmaker> ./openttd -c mycfg.ini
12:37:03 <Artea> my server is now on VPS
12:37:30 <Artea> already set a dedicated.cfg
12:38:02 <Artea> just don't know how to setup them
12:38:17 <Artea> already figure out how to download content via cmd
12:42:32 <peter1138> Doesn't matter what the config name is, if you edit it while OpenTTD is running, and then exit OpenTTD, it will overwrite it.
12:43:21 <peter1138> Anyway, perhaps the easiest way to set up NewGRFs is to do it with a scenario on a desktop PC, and then upload that to the server.
12:46:23 <V453000> yeah what's what we do, make a savegame/scenario with NewGRFs, upload to server, load
12:46:54 <V453000> I wonder if the server can load the game with older versions of newgrfs from bananas and obtain them automatically though
12:47:14 <V453000> you need to have them downloaded first, but i don't know if you can download the non-latest that way
12:47:20 <Artea> you can update NewGRF with content cmd
12:47:35 <V453000> yes, to get latest versions
12:47:44 <V453000> a sidetrack, sorry :)
12:50:33 <planetmaker> peter1138, will it overwrite the cfg it loaded from or openttd.cfg?
12:51:01 <planetmaker> anyway, I agree, the easiest way to setup a server with a good map & newgrf: transfer a savegame to load
12:51:40 <planetmaker> @V453000, you can download newgrfs from banananas via rcon... but it's not exactly an excellent user experience
12:52:14 <planetmaker> you can only download the 'latest' iirc, though
12:53:44 <Artea> is hard to download via cmd
12:54:22 <peter1138> planetmaker, the running config, obviously.
12:54:57 <peter1138> planetmaker, I was clarify your point, not disagreeing ;)
12:58:52 <Artea> downloaded OpenGFX+ Tree from content cmd
13:02:42 <Artea> There were 193 failed login attempts since the last successful login.
13:02:51 <Artea> seems bots found out my login :(
13:10:08 <V453000> yeah that's what I wondered pm, if I could download the older versions ... sometimes I did get load errors due to non-latest versions just because that grf hasn't been on the server before
13:10:22 <V453000> (before the latest one has been added to bananas)
13:27:26 *** Thedarkb-T60 has joined #openttd
13:39:16 *** supermop_work_ has joined #openttd
13:43:02 *** tycoondemon has joined #openttd
13:49:31 *** supermop_work has joined #openttd
13:50:00 *** supermop_work_ has quit IRC
13:50:48 *** supermop_work has joined #openttd
13:52:02 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN commented on pull request #7540: Change: New layout for the Station view window https://git.io/fjsMH
13:54:29 <peter1138> LordAro, I'm being picky now, but I think ZOOM_LVL_BASE is fine to be a ZoomLevel, just not ZOOM_LVL_SHIFT. If that's possible :/
13:59:35 <nielsm> lots of things are put in enum that should be plain const
13:59:42 <nielsm> because of C heritage?
14:02:55 *** AKTheKnight has joined #openttd
14:03:05 <peter1138> LordAro, maybe? I dunno.
14:03:14 *** ToBeFree has joined #openttd
14:03:34 <peter1138> maybe it should be static const ZoomLevel ZOOM_LVL_BASE = ...
14:03:35 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh commented on pull request #7540: Change: New layout for the Station view window https://git.io/fjsDk
14:03:42 <peter1138> Is that even possible? :p
14:03:46 <LordAro> i read something the other day about how gcc was deciding to optimise out some enum value checks because it couldn't possibly be a member of the enum
14:04:03 <LordAro> it was a gcc bug report, but...
14:17:27 <Artea> has been so long I dont even remember the Anime
14:21:24 <Artea> this town have a good suburbs
14:22:23 *** tycoondemon has joined #openttd
14:51:23 *** Thedarkb-T60 has joined #openttd
15:59:44 * peter1138 downloads the route plan for tomorrow... assuming it's not pissing down :/
16:04:13 *** gelignite has joined #openttd
16:17:22 <peter1138> Looks like there's nothing else to watch at the cinema for a while :p
16:23:23 *** Wormnest has joined #openttd
16:35:02 <andythenorth> 75mph wind tomorrow?
16:35:58 <peter1138> Hmm, 33 mph gusts. Maybe.
16:36:14 <peter1138> Maybe I should go swimming...
16:36:31 <andythenorth> maybe I should go hot tubbing
16:36:42 <andythenorth> it's big enough that my kids can swim endlessly around it in a circle :P
16:36:45 <peter1138> Maybe I should sit on my PC all day and... contemplate NRT.
16:37:57 <peter1138> I should resquash the last bits and then we can sit on it for another year.
16:40:15 <peter1138> Maybe I should unsquash it all.
16:40:17 <andythenorth> strictly you can approve the PR :P
16:40:21 <andythenorth> because I opened it
16:52:39 <peter1138> Is it necessary/useful for AIs to know if a road type is road or tram?
16:52:55 <peter1138> I wonder if compatibility is enough.
16:53:27 <peter1138> Hmm, building tram track is different though.
17:04:36 *** sla_ro|master has joined #openttd
17:09:58 *** Progman has joined #openttd
17:18:51 <Corns> trams can't overtake so that may be important
17:21:54 <Corns> peter1138 as far as i know, companies can share roads by default but not tram tracks
17:24:06 <andythenorth> tram track can't have dead ends in the same way
17:24:11 <andythenorth> it will just be broken
17:24:16 <DorpsGek> andythenorth: pikka was last seen in #openttd 2 weeks, 1 day, 4 hours, 1 minute, and 51 seconds ago: <Pikka> if the separate sprites had cc masks, I wouldn't have made them separate sprites :)
17:24:39 <andythenorth> if pikka teaches his AI to build tram tracks, then we can add support :P
17:24:51 <andythenorth> samu has an AI? o_O
17:39:13 *** erratic has joined #openttd
17:56:59 *** HerzogDeXtEr has joined #openttd
18:32:15 <Artea> if you are bored, come to my server :P
18:32:43 <Samu> i can't, im running an ai tournament :( haven't stopped yet
18:34:37 <Samu> currently on round 9 / 47
18:35:31 <Artea> my server is little over 825 years
18:38:07 <peter1138> Yay, come home and start eating straight away. </fatty>
18:38:41 *** daspork_ has joined #openttd
18:39:12 <peter1138> I ate a biscuit and a bit of chocolate.
18:39:32 <peter1138> Hmm, maybe I should make RoadTypeList and TramTypeList for AIs
18:39:43 <peter1138> Currently there is only RoadTypeList which lists both.
18:40:41 <peter1138> Ah, I should do it like VehicleList stuff.
18:41:18 <peter1138> Hmm, not quite, that's a lot of different list types. Hmm.
19:08:00 *** Gustavo6046 has joined #openttd
19:10:23 <peter1138> Tunnels/bridges as stations, what?
19:10:40 <Eddi|zuHause> i've thought about that concept in the past
19:11:17 * Eddi|zuHause on must go outside - but doesn't want to
19:11:55 <Artea> why isn't Theme Hospital OST in OTTD :(
19:12:12 <nielsm> because we don't have an XMI loader
19:12:22 <andythenorth> boom boom, cycling time
19:12:54 <nielsm> but if you can find your own software to convert the XMI files to standard MID files it's not hard to write a baseset file to get it in
19:13:32 <nielsm> also the TH music isn't available under a free license (as far as I know) so it can't be made available as a download anyway
19:17:33 <peter1138> TURTLES HAVE SHORT LEGS
19:17:54 <peter1138> Gotta love Krautrock.
19:19:11 <Eddi|zuHause> we could implement a loader, and then offer a .obm file that only works if you have that game installed
19:25:35 *** supermop_work_ has joined #openttd
19:26:14 *** supermop_work_ has quit IRC
19:26:19 *** supermop_work_ has joined #openttd
19:30:19 <peter1138> Now make an obm :-)
19:30:34 <nielsm> make a copy of openttd/baseset/orig_win.obm and name it something appropriate
19:30:48 <nielsm> the most annoying part is filling in the MD5 sums
19:31:08 <Eddi|zuHause> that's just one command, and pasting the results?
19:31:21 <nielsm> depends on your OS and text editor :P
19:31:45 <Eddi|zuHause> of course, there's always ways to make it needlessly complicated :p
19:32:00 <nielsm> I didn't say it was hard just annoying
19:33:49 <Artea> I don't know how to do it
19:34:18 <nielsm> you don't know how to copy a file?
19:34:25 <nielsm> don't know how to open a file with a text editor?
19:34:29 <nielsm> don't know how to use a text editor?
19:34:37 <nielsm> don't know how to make md5 sums of files?
19:35:45 <Artea> I'm making the copy of orig_win.obm
19:55:16 *** andythenorth has joined #openttd
20:09:38 *** frosch123 has joined #openttd
20:12:51 <DorpsGek_II> [OpenTTD/OpenTTD] Gymnasiast commented on issue #7530: Crash in kdtree.hpp when removing track in development version https://git.io/fjs7C
20:13:59 <Artea> don't know what I'm doing wrong now
20:33:25 <Artea> looking at OTTD with Theme Hospital OST
20:35:55 <LordAro> peter1138: wrt #6605, apparently tile 37809 is actually a house, that the game also thinks is a depot
20:36:48 <V453000> it counts the number vehicles in the consist and draws them in an alternating way. For some reason the rear ones just won't draw.
20:36:54 *** gelignite has joined #openttd
20:37:05 <V453000> I'm using the same idea elsewhere also with articulated vehicles so that can't make the difference
20:37:21 <V453000> the counting works, it's visible that with 3 units I start getting 2 at the front
20:38:04 <V453000> but the switches with position_in_consist_from_end just don't work (yet they do draw the wagons) and I don't see why
20:39:17 <V453000> everything in the screenshot is the same engine item, no real wagons
20:42:48 <frosch123> mixing position_in_consist and count_veh_id is weird
20:43:12 <peter1138> LordAro, yeah, the savegame is corrupt somewhere, but we don't know why :(
20:43:15 <frosch123> one counts all vehicles, the other only vehicles of one id
20:43:51 <peter1138> LordAro, might be fixable by looping depots and checking their xy is a valid depot tile.
20:44:05 <V453000> frosch123: but it worked when I only used it from the front :d
20:44:08 <peter1138> LordAro, I wonder if there's some depot tile on the map without a depot, heh.
20:44:25 <V453000> the count_veh_id is just deciding which switch to pick, and it succeeds there
20:44:31 <LordAro> peter1138: how would i know if it's a valid depot tile? isn't that done by looking up m2?
20:44:48 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh commented on issue #7530: Crash in kdtree.hpp when removing track in development version https://git.io/fjs71
20:44:49 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh closed issue #7530: Crash in kdtree.hpp when removing track in development version https://git.io/fjO1b
20:45:02 <peter1138> Er, check if it's a depot tile instead of, say... a house? :P
20:45:32 <LordAro> what tile type should depots be?
20:45:45 <frosch123> there is IsDeoptTile or something
20:46:00 <frosch123> they are diferent tiles for each transporttype
20:46:16 <frosch123> LordAro: anyway, can it be caused by moveable depots?
20:46:54 <LordAro> frosch123: moveable depots? i don't follow
20:47:07 <andythenorth> World of Tanks Blitz, or Hog feldbahn?
20:47:10 <frosch123> V453000: your vehicles are articulated
20:47:25 <frosch123> the position counts articulated vehicles
20:47:33 <frosch123> so only the last articulated part is at the end
20:48:11 <V453000> yes, and that part still has the same drawing and uses the same item
20:48:23 <V453000> I use this in chameleons already, just without the count_veh_id
20:48:25 <frosch123> LordAro: isn't there a recent feature to delete a depot and rebuild it
20:48:31 <frosch123> like was only possible with stations before
20:48:40 <LordAro> i don't think that got merged
20:48:45 <LordAro> this issue is much older than that, anyway
20:50:24 <LordAro> nielsm: finally a nice reproducible save for the kdtree issue :)
20:51:01 <andythenorth> is there anything I can do to help fix coop jenkins?
20:51:27 <frosch123> V453000: well, maybe check with the newgrf inspect tool
20:51:31 <andythenorth> currently I can only build on Azure
20:51:36 <andythenorth> and I can only publish on coop
20:51:38 <frosch123> position_in_consist/from_end is var 40
20:51:41 <andythenorth> and the two don't talk
20:51:58 <V453000> newgrf inspect tool? :D
20:52:49 <LordAro> ok, that's a bit more interesting - printing out "depots" which aren't valid depot tiles results in 5 tiles
20:53:10 <V453000> anyway, I tried to make the item read switch_railmedium_9_graphics_both_draw4 directly and I'm getting what I expect
20:53:21 <V453000> so indeed the counting is breaking it somehow
20:53:49 <frosch123> count_veh_id only counts vehicles starting from the current one
20:53:57 <frosch123> you need to use PARENT instead of SELF to count from front
20:54:28 <nielsm> LordAro yeah, and yes it is as JGR said, the cause is the DeleteOilRig function removing the dock tile and airport facilities from the station so it becomes empty, but the viewport coordinate of the station changes when it has an oil rig type airport, so the viewport coordinate of the station changes without the kdtree being updated
20:56:17 <nielsm> the alternate solution to caching the actual sign coordinate would be to just assume that the oil rig station is never shared with anything else
20:56:38 <nielsm> station_cmd.cpp:3969 is: if (!st->InUser()) delete st;
20:56:57 <nielsm> but that may as well be: assert(!st->InUse()); delete st;
20:57:12 <V453000> frosch123: works, thank you so much :)
20:57:28 <LordAro> #6605: all the invalid depot tiles are next to rail lines in positions which could've contained depots at some point in the past
20:58:43 <V453000> frosch123: can I have question number 2 please? I need to make a switch which decides between two drawing modes based on whether the last vehicle in the consist is this specific engine. Can I do that somehow?
20:59:38 <V453000> if position_in_consist_from_end == vehicle_id: both_end_drawing_switch, else other_drawing_switch
20:59:59 <V453000> is that the dreadful store_temp thing that I never managed to understand? :D
21:00:24 <LordAro> (deleting the depots while loading apparently isn't enough, still crashes)
21:00:26 <V453000> would you be so kind and write this one switch for me please?
21:00:44 <V453000> I'm staring at the rainbow slugs and what you did there with the store temp and I don't understand
21:00:51 * andythenorth wonders about showing V453000 Horse nml :P
21:01:03 <andythenorth> maybe too scary at this time of night
21:01:27 <DorpsGek_II> [OpenTTD/OpenTTD] Gymnasiast commented on pull request #7540: Change: New layout for the Station view window https://git.io/fjs77
21:02:23 <frosch123> V453000: STORE_TEMP(position_in_consist_from_end, 0x10F), var[0x61, 0, 0xFFFF, 0xC6]
21:03:22 <frosch123> that gives you the vehicle id
21:03:44 <frosch123> so you can then put the 47 into the switch
21:03:58 <andythenorth> V453000: ^ Horse doesn't do what you're looking for afaict
21:04:30 <andythenorth> is it any surprise Horse compiles so frigging slow :D
21:04:38 <andythenorth> 1024 lines for just 1 vehicle
21:05:17 <andythenorth> the dangers of generated code are that is very low cost to make a lot
21:05:19 <V453000> I have about a quarter lines per vehicle :D
21:05:26 <andythenorth> until complexity cost kicks in :P
21:06:06 <glx> maybe the compiler is not optimised
21:06:07 <andythenorth> also it is absolutely *vital* that pantographs go down if the rail is not electrified :P
21:06:28 <V453000> oh my god frosch123 it works, first try, I can't express how thankful I am :) again, and again.
21:06:47 <V453000> oh that kind of switches andythenorth :P
21:07:04 <andythenorth> two liveries, controlled on depot flip
21:07:05 * LordAro wonders what happened to DorpsGek_II
21:07:17 <andythenorth> specific details depending on consist position
21:07:22 <andythenorth> and pantograph layer
21:07:26 <andythenorth> and rear lights layer
21:07:30 <andythenorth> and opening doors in stations
21:07:45 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh commented on pull request #7540: Change: New layout for the Station view window https://git.io/fjs7A
21:07:59 <andythenorth> they are really helpful
21:08:04 <LordAro> oh ok, it just doesn't work for me
21:08:04 <glx> LordAro: it's there it seems
21:08:13 <V453000> I didn't get to do that yet and I have planned to for SO LONG
21:08:22 <LordAro> just being a bit slow
21:08:23 <V453000> might do so tomorrow :) for info overlays in depot
21:08:43 <andythenorth> screenshots if you do v
21:09:04 <V453000> won't be much fancy, a background and an icon on top
21:09:17 <LordAro> Artea: distributing copyrighted music is usually frowned upon
21:10:01 <Artea> hence I didn't put in forum
21:10:16 <LordAro> peter1138: think that #6507 & #6605 are duplicates?
21:10:27 <LordAro> Artea: what makes you think we're more accepting of it?
21:11:07 <Artea> just advertise in case anyone wanna try
21:11:12 <peter1138> LordAro, likely. Unknown cause :(
21:11:57 <glx> the date in the screenshot being 2043, it may be an oilrig closing
21:14:43 <LordAro> peter1138: maybe i should try recreate it with something prior to SLV_128
21:16:01 <V453000> thanks again frosch123, this lets me work on this more tomorrow :) good night people
21:20:58 * LordAro notices there's a function at the bottom of saveload.cpp that's been commented out since 2005
21:23:55 <andythenorth> IT MIGHT BE IMPORTANT!
21:24:44 <peter1138> static void FixDepots()
21:26:40 <peter1138> So that's chocolate brownie with custard for dessert. Gone.
21:26:56 * LordAro attempts to compile r17490
21:28:05 <LordAro> there are a lot of warnings
21:28:09 <LordAro> and almost as many errors
21:28:28 <nielsm> get a contemporary linux distro installed in a VM and compile via that?
21:28:31 <LordAro> error: ‘getcwd’ was not declared in this scope
21:28:48 <LordAro> i've already added -fpermissive to cflags
21:30:38 *** Wormnest has joined #openttd
21:32:53 <frosch123> hmm, according to todays fff, maybe ottd mods should be made in f
21:36:19 <nielsm> and I'm not sure if my fix is safe for those
21:40:38 <LordAro> ./configure CFLAGS="-fpermissive -Wno-error=narrowing" --without-icu --without-png
21:40:53 <LordAro> with only one minor modification to fileio.cpp
21:41:13 <LordAro> apparently the standard library stopped #including unistd.h in gcc4.7
21:41:16 <frosch123> adding a "(char*)" cast?
21:41:48 <peter1138> LordAro, submit a PR? ;)
21:42:28 <LordAro> frosch123: ha, there are many of those
21:42:35 <LordAro> there are 3689 lines of output
21:42:51 <LordAro> so with about 4 lines per warning...
21:42:55 <frosch123> i have old checkouts of 0.6, 0.7, ... etc
21:43:04 <frosch123> they have some modifies which made them compile at some point
21:43:17 <frosch123> but they also break again later :)
21:44:04 <LordAro> Error: Failed to find a sounds set. Please acquire a sounds set for OpenTTD.
21:44:16 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh opened pull request #7544: Fix #7481: Don't modify oil rig stations during removal https://git.io/fjs5R
21:44:42 <glx> path changed since 17490 ;)
21:44:44 <LordAro> original_windows: Original Transport Tycoon Deluxe Windows edition sounds (unuseable: 1 missing file)
21:45:22 <frosch123> it's 17490 before support for more basesets?
21:45:26 <frosch123> so it only accepts original
21:45:33 * peter1138 listens to Clannad... very very 80s.
21:46:09 <frosch123> hmm, it's from 2009, so 0.7 stuff
21:46:47 <frosch123> 0.7 already has basesets
21:46:59 * nielsm is looking through for old oil rig bugs
21:47:07 <LordAro> ah yes, it just doesn't pick up ~/.local/share
21:47:59 <glx> data split happened later
21:48:13 <frosch123> oh, infamous xdg-basedir
21:48:56 <frosch123> rb had to fix some bugs in libxdg-basedir before it worked for ottd
21:49:59 <peter1138> Uh oh, tiling wm spotted
22:13:46 <peter1138> Oh. My VNC isn't starting up properly :(
22:15:20 *** Wormnest has joined #openttd
22:15:22 <peter1138> And now it is. Odd.
22:21:16 <peter1138> "Script took too long in the Load function"
22:21:31 <LordAro> 1b84d893e0c2acb34d880a4c1d68f9d2dc95af8e (svn r27573) is the earliest commit to compile with modern g++ with no other modifications
22:21:43 <LordAro> (with the exception of disabling icu)
22:44:55 <LordAro> compiled and ran 0.7.5
22:45:05 <LordAro> immediately died when i pressed "generate"
22:53:13 <Samu> lordaro, plz update your ai, it lost a game due to crashing early
22:55:30 <LordAro> it's an extremely trivial change though, you could fix it for the purpose of running your games
23:11:50 <DorpsGek_II> [OpenTTD/OpenTTD] Gymnasiast commented on pull request #7540: Change: New layout for the Station view window https://git.io/fjsda
23:20:02 *** Wormnest has joined #openttd
23:39:42 <LordAro> interestingly, 0.4.8 is easier to compile than 0.7.5
23:41:14 <LordAro> "Configure Patches" :D
23:41:30 <Zuu> Great to see in the commit log that the buy button now refits if you filtered by cargo type.
23:41:48 <Zuu> And some other useful aditions and several fixes as well.
23:42:06 <Zuu> I saw you visited my linkedin page so I came here. :-p
23:42:47 <LordAro> gotta be careful what you do on linkedin, they record everything :p
23:43:16 <Zuu> and try to make you pay to see more about your visitors.
23:43:39 <glx> and they send emails when you are not on linkedin
23:47:16 <andythenorth> email is the worst :P
23:48:16 <LordAro> peter1138: nope, no idea what caused the corruption - 0.4.8 & 0.7.5 save games (with deleted depots) load fine
23:48:44 <frosch123> if it is that old, maybe magic bulldozer?
23:49:17 <LordAro> hrm, could be actually
23:49:36 <LordAro> both the savegames are with other players (or, the original AI)
23:49:43 <frosch123> though that would still call the regular clear command
23:56:49 <peter1138> Why would tunnels/bridges as stations make sense? o_O
continue to next day ⏵
