Config
Log for #openttd on 22nd January 2020:
Times are UTC Toggle Colours
00:18:06  *** tokai has joined #openttd
00:18:06  *** ChanServ sets mode: +v tokai
00:40:23  *** hythlodaeus has quit IRC
00:47:45  *** Flygon has joined #openttd
01:12:48  *** Wormnest__ has quit IRC
02:18:04  *** Pikka has joined #openttd
02:55:01  *** Hazzard has quit IRC
02:56:26  *** Ammler has quit IRC
02:57:31  *** Osai has quit IRC
02:59:06  *** planetmaker has quit IRC
02:59:11  *** fonsinchen has quit IRC
02:59:11  *** SmatZ has quit IRC
02:59:11  *** Terkhen has quit IRC
02:59:32  *** XeryusTC has quit IRC
02:59:32  *** V453000 has quit IRC
02:59:32  *** tneo has quit IRC
02:59:41  *** Hirundo_ has quit IRC
03:01:12  *** avdg has quit IRC
03:01:12  *** ^Spike^ has quit IRC
03:01:12  *** Yexo has quit IRC
03:11:44  *** XeryusTC has joined #openttd
03:11:49  *** planetmaker has joined #openttd
03:11:49  *** ChanServ sets mode: +o planetmaker
03:11:59  *** V453000 has joined #openttd
03:12:29  *** tneo has joined #openttd
03:12:51  *** Hazzard has joined #openttd
03:13:29  *** ^Spike^ has joined #openttd
03:13:30  *** SmatZ has joined #openttd
03:13:59  *** Terkhen has joined #openttd
03:13:59  *** ChanServ sets mode: +o Terkhen
03:14:29  *** Hirundo has joined #openttd
03:14:59  *** avdg has joined #openttd
03:14:59  *** Yexo has joined #openttd
03:15:59  *** fonsinchen has joined #openttd
03:16:29  *** Ammler has joined #openttd
03:17:00  *** Osai has joined #openttd
04:02:44  *** debdog has joined #openttd
04:05:39  *** glx has quit IRC
04:06:07  *** D-HUND has quit IRC
04:21:53  *** tokai|noir has joined #openttd
04:21:53  *** ChanServ sets mode: +v tokai|noir
04:28:52  *** tokai has quit IRC
04:31:01  *** ZirconiumX has quit IRC
04:31:15  *** ZirconiumX has joined #openttd
05:48:12  *** agentsmith101 has joined #openttd
06:05:55  *** snail_UES_ has quit IRC
06:06:32  *** marimeireles has joined #openttd
06:07:14  *** tokai has joined #openttd
06:07:15  *** ChanServ sets mode: +v tokai
06:09:29  *** sla_ro|master has joined #openttd
06:14:12  *** tokai|noir has quit IRC
06:43:22  *** tycoondemon has quit IRC
06:44:44  *** mindlesstux_ has quit IRC
06:44:44  *** crem has quit IRC
06:44:44  *** Ttech has quit IRC
06:44:44  *** ST2 has quit IRC
06:44:44  *** arron[m] has quit IRC
06:44:44  *** UncleCJ[m] has quit IRC
06:44:44  *** rubenwardy has quit IRC
06:44:47  *** mindlesstux has joined #openttd
06:45:01  *** crem has joined #openttd
06:45:07  *** Smedles has joined #openttd
06:45:18  *** ST2 has joined #openttd
06:45:47  *** rubenwardy has joined #openttd
06:46:24  *** Ttech has joined #openttd
06:50:02  *** Smedles_ has quit IRC
07:03:35  *** tokai|noir has joined #openttd
07:03:35  *** ChanServ sets mode: +v tokai|noir
07:10:32  *** tokai has quit IRC
07:11:39  <DorpsGek_III_> [OpenTTD/OpenTTD] joestringer opened pull request #7954: Feature: Add news notifications for significantly delayed trains https://git.io/JvL0g
07:21:37  *** andythenorth has joined #openttd
07:22:31  *** tycoondemon has joined #openttd
07:24:53  <andythenorth> yo
07:38:37  *** marimeireles has quit IRC
07:39:33  *** marimeireles has joined #openttd
08:04:32  *** Smedles has quit IRC
08:05:50  *** Smedles has joined #openttd
08:16:30  <Pikka> yoyo
08:17:14  * andythenorth drawing all the hoppers
08:17:50  <andythenorth> only the – view matters right? :P
08:18:36  *** marimeireles has quit IRC
08:18:55  <Pikka> only the diagonal view matters :)
08:18:58  * Pikka bbs
08:27:49  *** Smedles has quit IRC
08:28:33  <DorpsGek_III_> [OpenTTD/OpenTTD] JMcKiern commented on issue #7950: Normal screenshot mode may incorrectly show a huge screenshot warning dialog https://git.io/JvIMx
08:29:03  *** Smedles has joined #openttd
08:31:18  *** andythenorth has quit IRC
08:45:37  <DorpsGek_III_> [OpenTTD/OpenTTD] nielsmh commented on pull request #7954: Feature: Add news notifications for significantly delayed trains https://git.io/JvLur
08:51:33  *** tokai has joined #openttd
08:51:33  *** ChanServ sets mode: +v tokai
08:58:29  *** tokai|noir has quit IRC
09:04:52  *** Smedles has quit IRC
09:07:35  *** Smedles has joined #openttd
09:27:01  *** Samu has joined #openttd
09:27:44  <Samu> hi
09:32:46  *** gnu_jj has quit IRC
09:45:45  *** gnu_jj has joined #openttd
09:54:42  *** gnu_jj has quit IRC
09:57:27  *** andythenorth has joined #openttd
09:59:38  *** gnu_jj has joined #openttd
10:53:31  <DorpsGek_III_> [OpenTTD/OpenTTD] nielsmh dismissed a review for pull request #7868: Feature: NewGRF callback profiling https://git.io/JeddG
10:55:49  <DorpsGek_III_> [OpenTTD/OpenTTD] nielsmh commented on pull request #7868: Feature: NewGRF callback profiling https://git.io/JvL2A
11:23:40  *** marimeireles has joined #openttd
11:39:12  *** andythenorth has quit IRC
12:25:31  *** andythenorth has joined #openttd
13:24:05  <Samu> hi
13:48:03  *** tokai|noir has joined #openttd
13:48:03  *** ChanServ sets mode: +v tokai|noir
13:48:17  *** snail_UES_ has joined #openttd
13:54:51  *** tokai has quit IRC
13:56:55  *** sla_ro|master has quit IRC
14:06:45  <Pikka> hi
14:08:41  *** snail_UES_ has quit IRC
14:09:28  <andythenorth> hi bob
14:10:29  <Pikka> hallo
14:13:01  <andythenorth> haz hoppers
14:13:05  <andythenorth> all solved
14:14:59  <planetmaker> hazzard hoppers, too?
14:17:50  <planetmaker> I wonder for something like uranium ore :D Or live scorpions
14:24:57  <Sacro> Just don't fly directly overhead
14:29:57  <andythenorth> lava
14:30:03  <andythenorth> geysers
14:30:11  <andythenorth> magnetic anomolies
14:30:15  <planetmaker> that's liquids :P
14:30:31  <planetmaker> but sure... a hopper with liquids might be even more fun :D
14:40:49  *** nielsm has joined #openttd
14:52:48  *** Flygon has quit IRC
14:54:25  *** tokai has joined #openttd
14:54:25  *** ChanServ sets mode: +v tokai
14:59:34  *** sla_ro|master has joined #openttd
15:01:24  *** tokai|noir has quit IRC
15:21:24  *** WormnestAndroid has quit IRC
15:22:54  *** WormnestAndroid has joined #openttd
15:24:02  *** WormnestAndroid has joined #openttd
15:24:54  *** WormnestAndroid has joined #openttd
15:49:27  *** agentsmith101 has quit IRC
15:53:18  *** agentsmith101 has joined #openttd
15:55:58  <Pikka> 101% horse then?
15:55:58  *** agentsmith101 has quit IRC
15:56:39  *** spnda has joined #openttd
16:03:16  <spnda> What's the difference between Action1 and 2?
16:03:19  <andythenorth> 109% Pikka
16:03:40  <andythenorth> got any torpedo wagons in 1x 8bpp? o_O
16:03:52  <Pikka> afraid not
16:03:55  <andythenorth> nvm
16:04:00  <andythenorth> I shall draw some for you
16:04:02  <Pikka> in grf spnda?
16:04:44  <planetmaker> @spnda, "Action2 is used to group sets of sprites from the previous Action1 together, and make them accessible by a variational or random action 2 (chain) or an Action3. "
16:05:34  <Pikka> draw them in 2x 32bpp and I'll put them in UKRS3 too :)
16:07:00  <spnda> planetmaker: So could I call a Action2 a "spritegroup ()" in NML?
16:08:19  <planetmaker> I guess that's true. But Pikka knows better, I guess. My NFO is *very* rusty
16:08:30  <spnda> It was my first time writing NFO yesterday lol
16:08:45  <planetmaker> he... bridges or stations or why?
16:08:51  <spnda> Testing a new feature
16:09:10  <planetmaker> hm?
16:09:14  <Pikka> yes, in nml "spriteset" is action 1 and "spritegroup" is action 2, afaict
16:09:41  <planetmaker> what are you testing exactly, spnda
16:09:48  <spnda> So I don't necessarily need Action2? I can just bind Action1 sprites to Action0 with Action3?
16:09:57  <spnda> planetmaker: Road Stops
16:10:22  <Pikka> action 3 normally refers to one or more action 2
16:11:05  <Pikka> you'll want action 2 chains if you want to be able to do any sort of variable test to decide what graphics to use
16:12:07  <spnda> Pikka: So should I just define a Action 2 like "1 * 1  02 14 00 01 01 01"? The first two 01 just saying there's just one sprite coming or?
16:13:17  <Pikka> whatever format works for you. bearing in mind action 1/2 IDs are word-sized
16:13:40  <Pikka> https://newgrf-specs.tt-wiki.net/wiki/Action2
16:14:21  <Pikka> when referring to them from action 2s, that is
16:14:27  <spnda> My Action1 ID isn't word sized... It's just "7 * 4 01 14 02 06" Huh?
16:15:00  <Pikka> yep, sorry
16:15:20  <Pikka> I meant when referring to them from another action 2/3. The "entries..." in the spec
16:16:16  <planetmaker> hm, is road stops stations or another, new feature?
16:16:32  <spnda> i did it as a new feature, 14
16:16:38  <spnda> just thought it'd be a nice addition to nrt
16:16:56  <planetmaker> spnda, then you might be quicker in making a small patch for NML than writing it in NFO :D
16:17:25  <spnda> Well, I kinda want to learn NFO aswell
16:17:34  <Pikka> NML still has to be NFO underneath, doesn't it?
16:17:35  <planetmaker> if it's a sane structure to how it works, e.g. like industries, objects or so, it's really easy
16:17:45  <spnda> And NML still basically uses NFO
16:17:48  <planetmaker> there's not really NFO beneath NML. It compiles directly to GRF
16:18:13  <spnda> Though it uses the same data structure
16:18:39  <planetmaker> nah, there's not NFO beneath it. Of course it outputs a (new)grf... which very much is similar to NFO in binary form
16:18:50  *** Wormnest__ has joined #openttd
16:18:51  <Pikka> ya
16:19:22  <spnda> And also having the code here infront of you resembles NFO. I can quite literally read from the docs where and which data and bytes from NFO are read where and when...
16:19:31  <planetmaker> but strictly speaking there's no NFO involved anywhere in NML except when you ask it to output NFO. Nowhere else
16:19:45  <Pikka> makes sense
16:20:30  <Pikka> hmm
16:20:35  <Pikka> road stops should probably be stations
16:20:59  <Pikka> or at least have access to the basestations variables
16:21:12  <spnda> or use the same variables
16:21:44  <spnda> I mean, I'm not going to completely redo how road stops are handled. They're still going to be a BaseStation
16:21:44  <Pikka> I guess the stations variables aren't that useful. just the basestations ones that are shared with airports
16:22:27  <Pikka> new feature ID it is then :P
16:23:16  <spnda> yup, 14
16:25:03  <planetmaker> hm, yeah. And it just needs re-implementing the variables as in basestation
16:25:11  <planetmaker> with their same name and numeric ID
16:25:58  <planetmaker> hm...! Then you can indeed borrow the sprite generation and composition like from industries and objects @ spnda
16:26:03  <planetmaker> that would... make it nice :D
16:26:29  <planetmaker> rail stations are... different
16:26:56  <planetmaker> I remember frosch saying / explaining somewhen that the way stations do it is actually 'better' by certain definitions... alas
16:27:00  <planetmaker> it's unique
16:27:06  <planetmaker> and that sucks :P
16:27:36  <planetmaker> https://xkcd.com/927/
16:29:01  <spnda> I'd guessed there's a reason why there's still no NML support for stations lol
16:29:14  <planetmaker> yes. And the same reasoning applies for bridges
16:29:29  <planetmaker> stations are unique. bridges are unique. Everything else follows a common scheme
16:30:10  <planetmaker> (or two. vehicles all one scheme. all other buildings another common scheme)
16:30:23  <spnda> So I will follow that common scheme for buildings aswell!
16:35:41  *** Progman has joined #openttd
16:36:46  <planetmaker> hm... https://dev.openttdcoop.org/issues/2746
16:37:44  <spnda> 5 years ago already? its been a long time with no updates..
16:38:32  <planetmaker> https://dev.openttdcoop.org/issues/1348 for bridges... 8 years :D
16:44:08  <spnda> Object's Action3 looks exactly like Station's Action 3... hmm
16:44:28  <spnda> There's a lot of mentions of default groups and other groups. What are these?
16:44:35  <planetmaker> the differences are - iirc - in how layouts are handled
16:44:51  <planetmaker> groups are the things how it is sorted in the building menu
16:45:31  <planetmaker> cargo stations, rural stations, trackless stations could be groups for stations. For objects like fences, foot paths,...
16:45:44  <planetmaker> light houses etc
16:45:56  <spnda> Ok so just Station Classes and the default group would be the default stations?
16:46:06  <planetmaker> I think so
16:47:44  <spnda> Well actually it seems that the first checks are for cargo ids, then come default groups...
16:47:52  <spnda> RoadTypes and TramTypes don't use the default group at all it seemss
16:48:56  <spnda> Ohhh I think the documentation is actually giving a good explanation...
16:51:25  <spnda> Yeah so I basically don't care about cargo ids for now...
16:56:40  <nielsm> can there be decorative station tiles that don't belong to a specific transport mode?
16:56:48  <nielsm> I guess it would change the rules for station (sign) ownership
16:59:50  *** marimeireles has quit IRC
17:00:26  <FLHerne> That would be nice
17:00:48  <FLHerne> So many of my stations have 'rail' logos on them just because I'm using empty tiles
17:00:54  <nielsm> also station tiles that don't confer catchment area (but could be cheaper in maintenance)
17:01:43  <nielsm> although that's leaning into "newobjects that respond to nearby stations" kind of
17:02:09  <spnda> Well aren't those just objects anyway?
17:02:25  *** rptr has joined #openttd
17:02:35  <rptr> was this the last on copy&paste? https://www.tt-forums.net/viewtopic.php?f=33&t=41259&start=220
17:04:11  <FLHerne> spnda: Objects don't have cargo state
17:04:22  <spnda> Ok that's true..
17:16:38  <spnda> Which sprite is that even showing there on the right?? https://cdn.discordapp.com/attachments/495228944719544324/669589916992405504/unknown.png
17:16:54  <spnda> Is that some docks sprite? I don't recall seeing that anywhere before
17:19:33  <andythenorth> that is a piece of a dock from a newgrf or base set
17:24:46  *** Pikka has quit IRC
17:28:59  *** Samu_ has joined #openttd
17:34:51  *** Samu has quit IRC
17:47:02  *** HerzogDeXtEr has joined #openttd
17:51:31  *** HerzogDeXtEr has quit IRC
18:00:45  *** andythenorth has quit IRC
18:14:53  *** Wolf01 has joined #openttd
18:29:00  *** glx has joined #openttd
18:29:00  *** ChanServ sets mode: +v glx
18:30:17  <spnda> I can't find where the Spec in the GRFFile struct is put into a Class... Is there some function or something?
18:34:16  *** andythenorth has joined #openttd
18:35:23  <glx> which Spec spnda ?
18:35:34  <spnda> StationSpec
18:35:52  <glx> StationChangeInfo()
18:37:58  <spnda> Well I saw that, but didn't see any place in there adding a spec into a specific StationClass
18:40:57  <spnda> I only see prop 0x08 which allocates a StationClass and adds it into the cls_id for a StationSpec array in the GRFFile
18:41:34  <glx> yes classes are created if needed by prop 08
18:41:56  <spnda> Still if I would do StationClass::Get(cls_id)->GetSpecCount() I get 0
18:42:12  <spnda> This is not for StationClass, rather for RoadStopClass but it's the same for 0x08
18:42:28  <supermop_work> "Each vessel will be approximately 135,000 gross tons"
18:42:32  <supermop_work> is that big?
18:47:10  <Wolf01> So, I noticed I own Train Simulator 2020 too, that's like Fifa, but it's literally always the same game just changed the year in the name
18:48:51  <spnda> glx: In the BuildRailStationWindow struct the count of stations is assigned by "StationClass::Get(_railstation.station_class)->GetSpecCount()". I do the same for RoadStops but the ones I added by my GRF are not showing up. The class exists but there's no specs assigned.
18:48:51  <supermop_work> yeah but every year the best star locomotives sign with different teams
18:49:47  <andythenorth> supermop_work: 135k is quite big
18:50:00  <glx> https://github.com/OpenTTD/OpenTTD/blob/9d5dd893cdd6ca75d1ed7ad2aa7ca5d98feffa5f/src/newgrf.cpp#L5521 <-- count is updated there
18:51:52  <spnda> "RoadStopClass::Assign(roadstopspec);" I've got it at the same position in there for my feature...
18:52:23  <andythenorth> supermop_work: this is 135k grt http://www.shipspotting.com/photos/middle/9/0/1/2833109.jpg
18:52:35  <andythenorth> https://en.wikipedia.org/wiki/MSC_Daniela
18:52:47  *** mindlesstux has quit IRC
18:53:25  <supermop_work> this one has themed rooms instead of containers
18:53:40  <glx> spnda: code is visible somewhere ?
18:54:06  <spnda> Only locally just now. I could commit it to my repo though
18:54:50  *** mindlesstux has joined #openttd
18:56:36  <spnda> glx: https://github.com/spnda/OpenTTD/tree/roadstops
18:57:52  <spnda> I'm kinda thinking of making a Draft PR for this already
18:58:12  <nielsm> wow that's a very smooth-looking flyover I built here for once https://0x0.st/zCwh.jpg
18:59:21  <nielsm> bridge portion barely any longer than necessary, no huge curves to get back to parallel with the line being joined
19:10:20  <andythenorth> supermop_work: is it a Disney cruise ship?
19:10:56  <supermop_work> i might not be allowed to saw yes
19:11:17  <supermop_work> rather i am allowed to say that disney is building 3 new cruise ships of that displacemnt
19:11:22  <supermop_work> but not much more
19:12:32  <andythenorth> yeah that's on Google, top result for 135k
19:14:23  <spnda> I think I might aswell just open up a draft PR for my feature 14 already
19:15:48  <andythenorth> it's definitely wanted :)
19:15:58  <andythenorth> how easy it is to get reviewed, I don't know :)
19:19:22  <spnda> I'm just gonna have it be a thing
19:53:31  <DorpsGek_III_> [OpenTTD/OpenTTD] LordAro merged pull request #7953: Fix #7952: don't try to access destroyed QueryStrings https://git.io/JvL3g
19:53:32  <DorpsGek_III_> [OpenTTD/OpenTTD] LordAro closed issue #7952: When closing an interface window within the game whilst typing in Japanese with IME keyboard, the game crashes https://git.io/JvLko
19:54:20  *** Smedles has quit IRC
19:55:20  *** Smedles has joined #openttd
20:14:43  <glx> spnda: after a quick read it seems ok, are you testing with a custom GRF ?
20:15:08  <spnda> yeah I wrote it myself. Maybe I made some mistakes in my NFO code. Idk
20:15:24  <spnda> I'm opening a draft pr btw
20:19:34  <DorpsGek_III_> [OpenTTD/OpenTTD] spnda opened pull request #7955: Feature: NewGRF Road Stops (Feature 14) https://git.io/JvLQL
20:19:54  <spnda> nonono that was suppossed to be a draft
20:19:57  <andythenorth> screenshots? o_O :)
20:21:07  <LordAro> spnda: just add WIP to the title
20:21:13  <spnda> ok, will do
20:21:34  <glx> the new commit checker is so nice :)
20:21:54  <spnda> andythenorth: https://cdn.discordapp.com/attachments/495228944719544324/669591112696070185/unknown.png is the latest one I have from the GUI
20:21:55  <LordAro> sooper fast
20:22:05  <andythenorth> spnda: add that
20:22:06  <spnda> Still not finished and buggy, but something atleast
20:22:11  <glx> 2 clicks and you see clearly why it fails :)
20:22:11  <andythenorth> it just helps to get attention
20:22:13  <andythenorth> makes it real
20:22:18  <LordAro> test GRF couldn't hurt either
20:22:55  <spnda> my own test grf doesn't work anyway I think so yeah
20:24:10  <spnda> glx: Oh does that commit checker automatically bail when there's no confirmed review?
20:24:46  <glx> no, it always checks
20:25:04  <glx> same for the compile farm, each time you push changes it builds
20:28:23  <TrueBrain> glx: and just realise that if we port the rest to Actions, warnings/errors of GCC etc are also shown where it happens :o :D
20:28:35  <LordAro> :o
20:28:50  <TrueBrain> well, a simple annotation file would also do that for commit-checker :P
20:28:55  <TrueBrain> something for the weekend I guess :D
20:29:10  <glx> anyway commit checker needs to be improved, if you have trailing whitespaces and a bad commit message only the whitespaces are shown
20:29:20  <glx> commit message check is skiped
20:29:54  <glx> oh and it could also tell on which commit it's working
20:30:36  <TrueBrain> go for it :)
20:31:54  <glx> https://github.com/OpenTTD/OpenTTD-git-hooks/blob/master/hooks/check-commits.sh is only used for CI and not by git ?
20:34:19  <TrueBrain> frosch knows, but it looks that way, yes
20:44:14  <TrueBrain> https://github.com/OpenTTD/OpenTTD/contribute
20:44:15  <TrueBrain> funny page
20:45:22  <TrueBrain> https://github.blog/2020-01-22-browse-good-first-issues-to-start-contributing-to-open-source/ for more info
20:50:45  *** spnda has quit IRC
21:24:47  *** marimeireles has joined #openttd
21:25:26  <TrueBrain> https://www.openttd-cdn.org/ <-okay, first building blocks of the CDN are also in place. Lets see if I can provision that tomorrow :)
21:25:33  <TrueBrain> should be a lot better than the DO CDN :)
21:26:48  <SpComb> dragged your own hardware out to a couple POPs across the globe?
21:27:03  <TrueBrain> that is what we used to do :P
21:27:27  <TrueBrain> kinda trying to move away from that :)
21:28:41  <andythenorth> TrueBrain: there's nothing there!
21:28:45  <andythenorth> I'll check back soon
21:28:48  <andythenorth> \o/
21:29:08  <TrueBrain> you bettah! :P
21:29:26  <TrueBrain> this is now CloudFlare -> CloudFront -> S3
21:29:31  <TrueBrain> too many layers of indirection
21:29:43  <TrueBrain> but it is fully encrypted and verified connections from start to finish
21:30:30  <TrueBrain> enough for one day; sleep well all
21:32:33  <andythenorth> :)
21:36:30  <rptr> so why wasn't the clipboard feature added? it looks really fancy on youtube
21:36:35  <rptr> i thought it would be crude but it's incredibly fancy
21:50:08  <glx> hmm DorpsGek doesn't watch OpenTTD-git-hooks
21:50:49  <glx> https://github.com/OpenTTD/OpenTTD-git-hooks/pull/12
21:52:41  <glx> oh there was already https://github.com/OpenTTD/OpenTTD-git-hooks/pull/5
22:06:41  *** nielsm has quit IRC
22:12:41  <DorpsGek_III_> [OpenTTD/OpenTTD] stormcone updated pull request #7886: Group management enhancements https://git.io/JeAdk
22:14:57  <DorpsGek_III_> [OpenTTD/OpenTTD] stormcone commented on pull request #7886: Group management enhancements https://git.io/JvLFU
22:18:46  <DorpsGek_III_> [OpenTTD/OpenTTD] James103 commented on pull request #7955: WIP Feature: NewGRF Road Stops (Feature 14) https://git.io/JvLF3
22:21:06  *** marimeireles has quit IRC
22:23:25  <glx> TrueBrain: for DorpsGek to watch a repo it just needs a .dorpsgek.yml ?
22:35:51  <DorpsGek_III_> [OpenTTD/OpenTTD] stormcone commented on pull request #7919: Lifetime profit https://git.io/JvLFd
22:40:42  *** sla_ro|master has quit IRC
22:41:36  <Samu_> I have a question, should priority queues and a star always deliver the cheapest path?
22:41:56  <Samu_> because I'm not getting the same path costs :(
22:42:13  <Samu_> between binary queue and fibonacci queue
22:43:23  <rptr> a star should
22:43:32  <rptr> that's the whole point of the algorithm isn't it?
22:47:34  *** Wolf01 has quit IRC
22:49:29  <Samu_> i better test the different implementations of these priority queues
22:50:04  <Samu_> fibonacci is faster than binary
22:50:25  <Samu_> but i wonder if it's bugged
22:53:18  <rptr> which one is getting a higher path cost?
22:54:51  *** Progman has quit IRC
22:55:41  *** andythenorth has left #openttd
22:56:18  <TrueBrain> glx: yes, and I need to assign permission .. I will do that tomorrow if you do the file :)
22:59:03  <Samu_> binary has higher path cost
23:02:38  <glx> TrueBrain: https://github.com/OpenTTD/OpenTTD-git-hooks/pull/13 :)
23:04:27  <_dp_> fib heap is only faster in theory
23:07:41  *** tyteen4a03 has quit IRC
23:08:08  *** tyteen4a03 has joined #openttd
23:11:23  *** Laedek has quit IRC
23:17:04  <Samu_> it's faster here for AIs
23:17:29  <Samu_> so I tested 3 different Fibonaccis
23:17:35  <Samu_> they all result the same
23:17:56  <Samu_> there's only 1 version of binary heap to test
23:19:09  <Samu_> https://bananas.openttd.org/en/ailibrary/
23:19:24  <Samu_> then there's Queue.PriorityQueue which is utterly slow
23:30:08  <Samu_> it's weird, sometimes cost is cheaper, sometimes it's not
23:31:10  <Samu_> meh, screw it, I shouldn't look much into it
23:31:36  <Samu_> I accept that different queues return different paths, but I didn't expect different costs :(
23:32:31  <Samu_> if A star always returns the cheapest path, I don't understand why it didn't happen regardless the queue
23:37:16  <LordAro> Samu_: if you looked into how the A* works, you might get why different queues might mean that it doesn't always return the optimal result
23:39:11  <LordAro> though hmm, perhaps the A* heuristic is not quite "correct"
23:58:27  <Samu_> https://i.imgur.com/BzxLEl0.png binaryheap vs fibonacciheap vs priorityqueue
23:59:35  <Samu_> binary has the cheapest findPath cost
23:59:49  <Samu_> 45490

Powered by YARRSTE version: svn-trunk