IRC logs for #openttd on OFTC at 2019-02-14
⏴ go to previous day
00:13:31 <peter1138> The build # is different.
00:14:31 <glx> hmm build number is correct indeed
00:14:43 <glx> just not yet replaced with date
00:20:57 <glx> maybe a network issue in the precedent pull
00:45:05 *** Thedarkb1-T60 has joined #openttd
01:02:02 *** ChanServ sets mode: +v tokai
02:05:54 <DorpsGek_II> [OpenTTD/OpenTTD] JGRennison opened issue #7226: NPF: Ship: Assertion failure when ship encounters shore https://git.io/fh7CV
02:18:44 *** supermop_Home has joined #openttd
02:48:24 *** snail_UES_ has joined #openttd
03:06:44 *** Thedarkb-X40 has joined #openttd
07:10:28 *** Smedles has joined #openttd
07:34:03 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN opened pull request #7227: Replace Window::scrolling_scrollbar with Window::active_widget https://git.io/fh74H
07:37:20 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN commented on pull request #7209: Fix: volume slider behavior in music gui https://git.io/fh74F
08:05:41 *** andythenorth has joined #openttd
08:10:36 *** Pokka is now known as Pikka
08:42:30 *** Eddi|zuHause2 has joined #openttd
08:58:40 <andythenorth> is it some kind of special day?
09:57:32 *** andythenorth has joined #openttd
10:00:05 <LordAro> peter1138: doesn't matter; got cake?
10:29:08 <debdog> certainly better than cmake
10:36:10 <samu> I failed to trigger that crash
10:43:25 <samu> requires 90 degrees forbidden
10:44:06 <samu> if it happens for ships, i suppose it should also happen for trains, let me test
10:50:02 <samu> hmm nop, doesnt happen for trains
11:03:31 <samu> doesn't consider reversing for trains
11:22:27 *** andythenorth has left #openttd
11:31:25 *** Eddi|zuHause2 is now known as Eddi|zuHause
11:32:26 <Eddi|zuHause> aren't these the kinf of things that hang on the wall for years and then nobody dares to touch it because they forgot how long it was?
11:51:42 <peter1138> Well it looks easy to partially revert that.
11:55:06 *** andythenorth has joined #openttd
11:55:34 <andythenorth> Pikka: HSTs, mail or no mail?
11:56:33 <peter1138> I'd go for no mail.
11:59:47 <andythenorth> trains + sandwiches
12:01:20 <andythenorth> no mail is simplest
12:01:32 <andythenorth> kind of wastes some tile space though :P
12:02:47 *** andythenorth has left #openttd
12:12:01 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN commented on issue #7226: NPF: Ship: Assertion failure when ship encounters shore https://git.io/fh7zk
12:56:55 <samu> no need to revert too much
12:57:28 <samu> seems to be a case of 90 degrees
12:58:16 <samu> need to see how this was before the change
12:58:24 <samu> I think the assert doesn't need to be there
13:00:29 <samu> result.best_trackdir = INVALID_TRACKDIR;
13:00:57 <samu> there is indeed no track, that a reverse should be happening
13:01:34 <peter1138> That's why I said partially revert. Literally junk that hunk.
13:02:07 <peter1138> I can do a PR when I have my lunchbreak.
13:02:08 <samu> for train it doesn't happen, because PBS reserved stuff
13:03:07 <peter1138> Yeah, but white chocolate lion bars are incredibly sweet.
13:03:54 <samu> for road vehicle, the code makes a reverse track special bit
13:05:24 <samu> if (type != TRANSPORT_ROAD || HasBit(subtype, ROADTYPE_TRAM)) return;
13:06:07 <samu> how do i test trams... need newgrfs
13:07:38 <peter1138> Yeah, get a tram set.
13:14:36 <samu> hmm nop, doesn't assert, or i dont know how to
13:21:05 <samu> well, it's just the ship
13:21:28 <samu> maybe don't ask the pathfinder for a track if 90 degrees
13:22:17 <samu> reverse before asking pathfinder
13:22:42 <peter1138> Pathfinder should say no path if there's no path.
13:23:10 <samu> but it's a single tile, the tile just next
13:28:35 <samu> u got this at line 478 ship_cmp.cpp /* Can't continue in same direction so pick first available track. */
13:29:03 <samu> maybe should be repeated just before line 500
13:29:43 <samu> if (tracks == TRACK_BIT_NONE)
13:32:26 <samu> hmm opf doesnt have 90 degrees :|
13:33:23 <samu> it really has to ask the pathfinder because of opf
13:44:24 <peter1138> Doesn't have 90 degrees?
14:09:16 *** sla_ro|master has joined #openttd
14:26:17 *** Extrems has joined #openttd
14:27:27 <DorpsGek_II> [OpenTTD/OpenTTD] glx22 commented on pull request #7227: Replace Window::scrolling_scrollbar with Window::active_widget https://git.io/fh7a0
14:31:01 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN commented on pull request #7227: Replace Window::scrolling_scrollbar with Window::active_widget https://git.io/fh7ar
14:33:45 *** snail_UES_ has joined #openttd
14:35:07 *** andythenorth has joined #openttd
14:40:10 <andythenorth> the power->weight on those is insane
14:44:37 <andythenorth> I should go to exeter
14:46:53 <Eddi|zuHause> i remember someone mentioning instances where one of those engines went with a single passenger car
14:47:36 <andythenorth> is that a Ludmilla?
14:51:18 <Eddi|zuHause> the only way you could make that worse is having two of these, because you lack a steering car for push-pull-service
14:53:47 *** Gustavo6046 has joined #openttd
15:04:51 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN opened pull request #7229: Fix #7114: Partial revert of 7d06fec799 for ship pathfinding. https://git.io/fh7V2
15:20:14 <peter1138> When 1.9.0 is released.
15:21:30 <peter1138> You can build your own docs though, of course.
15:23:34 <andythenorth> maybe I should add some more trains
15:23:39 <andythenorth> then it will be 85% again
15:25:27 <Eddi|zuHause> andythenorth: think of it this way: "everybody has to die" is not true, if you count up the number of people who ever have died against the ones that have ever been born, you get a number something along the lines of 95%. if that ever reaches 100% it means mankind is extinct
15:26:04 <andythenorth> it's the exact analogy
15:26:23 <Eddi|zuHause> if your project ever reaches 100% completion it is dead
15:26:51 <andythenorth> OTOH, I do have to release 2.0.0 sometime :P
15:26:54 <andythenorth> 2.0.0 doesn't mean dead
15:27:55 <peter1138> If only you had engine variations...
15:31:16 <andythenorth> I didn't want to mention :P
15:31:22 <andythenorth> so how are we doing that again?
15:32:00 <Eddi|zuHause> so, i haven't solved any of my performance problems, but i haven't had any system freezes when i manually echo high/low to /sys/class/drm/card0/device/power_dpm_force_performance_level instead of keeping it at auto
15:32:43 <peter1138> Have you considered it might be faulty hardware?
15:32:47 *** kapamat has joined #openttd
15:33:19 <Eddi|zuHause> yes, but the alternative hardware is worse/older/less supported/not dx11
15:39:17 <DorpsGek_II> [OpenTTD/OpenTTD] J0anJosep commented on issue #7226: NPF: Ship: Assertion failure when ship encounters shore https://git.io/fh7wz
16:05:00 <DorpsGek_II> [OpenTTD/OpenTTD] nikolas commented on pull request #7209: Fix: volume slider behavior in music gui https://git.io/fh7ru
16:10:41 <samu> Should I use autoreplace or should I not
16:15:30 <andythenorth> is opf deleeted yet?
16:15:37 <andythenorth> or even deleted?
16:20:04 <peter1138> samu, "90 degrees for opf" ... what does that mean?
16:20:15 <peter1138> With OPF, ships do turn 90 degrees.
16:21:56 <samu> i had a patch request about it, but got rejected
16:23:54 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN updated pull request #7227: Replace Window::scrolling_scrollbar with Window::active_widget https://git.io/fh74H
16:24:16 <peter1138> Yeah we're more likely to remove that option, at least for ships.
16:24:22 <peter1138> And more likely to remove OPF.
16:26:25 <andythenorth> pls, it's healthy to delete stuff
16:26:33 <andythenorth> ok maybe not as much as I want to delete :P
16:30:20 <andythenorth> samu: maybe you could do a PR removing OPF? o_O
16:30:25 <andythenorth> you've probably seen a lot of that code?
16:30:40 <andythenorth> and you could find all the regressions
16:34:41 *** Wormnest has joined #openttd
16:36:48 <andythenorth> dunno if anyone would approve it though :x
16:55:21 <Samu> gonna try dpi aware stuff, brb
16:59:53 <Samu> it was high dpi aware stuff
17:25:34 *** Progman has joined #openttd
17:38:39 *** HerzogDeXtEr has joined #openttd
17:42:56 *** synchris has joined #openttd
18:33:09 <Samu> arf, my AI doesn't work well with groups
18:33:16 <Samu> wasn't built with it in mind :|
18:35:48 <andythenorth> what does Article 13 mean for Bananas?
18:37:37 <nielsm> well I suppose technically it'd mean everything had to be manually reviewed before becoming available
18:38:53 <andythenorth> but how do we know what infringes copyright?
18:39:04 <andythenorth> does this mean Bananas is over?
18:40:19 <nielsm> well isn't part of the uploader agreement that you give bananas a non-exclusive, non-retractable license to distribute the uploaded work forever?
18:41:04 <nielsm> of course you can only grant that if you actually have the rights to grant those rights
18:42:49 <andythenorth> there's no case law obviously
18:42:56 <andythenorth> and limited legal opinons, it's too new
18:43:20 <andythenorth> but there seems to be a requirement to deploy content recognition technology on uploads to prevent infringement
18:43:45 <nielsm> add a feature to bananas to allow people to upload material they do _not_ want available on bananas
18:44:02 <andythenorth> interesting idea
18:44:19 <andythenorth> but we'd need them to delegate us the rights to hold that copyrighted material
18:44:32 <andythenorth> I can think of at least one contributor who won't do that
18:45:17 <nielsm> well, if they won't let the service provider do automatic screening against their material they don't have a good case if someone else uploads their material and it can't be recognized, imo
18:46:34 <nielsm> specifically for GRFs could allow registering just GRFid's of course
18:46:35 <peter1138> Personally I'd just ignore it all.
18:46:45 <nielsm> as long as nobody sues there's no problem
18:47:01 <peter1138> It's clearly an ill-thought out law that's designed to only apply to the likes of Facebook, Instagram and Twitter...
18:47:35 <peter1138> I better go and get some dog food.
18:49:07 <andythenorth> just move Bananas to the UK
18:49:13 <andythenorth> article 13 probably won't apply eh
18:49:16 <andythenorth> we opted out of this stuff :P
18:50:03 <nielsm> the article 11 link tax is the most ridiculous though
18:50:13 <nielsm> why must I pay you to refer people to you?
18:50:34 <nielsm> did I ever pay for the right to suggest someone go to the store and buy a newspaper?
18:53:31 *** frosch123 has joined #openttd
19:18:05 <DorpsGek_II> [OpenTTD/OpenTTD] J0anJosep opened pull request #7230: Fix #7226: No ship track due to "forbid 90 deg turns"-> Do not call pathfinders. https://git.io/fh7Pg
19:21:10 <DorpsGek_II> [OpenTTD/OpenTTD] J0anJosep commented on pull request #7230: Fix #7226: No ship track due to "forbid 90 deg turns"-> Do not call pathfinders. https://git.io/fh7Pa
19:22:24 *** andythenorth has joined #openttd
19:35:42 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh approved pull request #7217: Fix #6599: Disable build and rename button in build vehicle window when no vehicle is selected. https://git.io/fh7PH
19:35:51 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh merged pull request #7217: Fix #6599: Disable build and rename button in build vehicle window when no vehicle is selected. https://git.io/fhQfL
19:35:57 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh closed issue #6599: Can still click on buy button in vehicle selection window even if no vehicle is selected https://git.io/fh7PQ
19:43:23 <Samu> 90 degrees for opf soon!
19:45:44 <LordAro> Samu: it's like you totally ignored what everyone said
19:45:58 <Eddi|zuHause> that sounds like my experience
19:48:32 <Eddi|zuHause> <nielsm> the article 11 link tax is the most ridiculous though <-- the worst part about that is that they passed the exact same law in germany, and it turned out to be completely useless
19:50:23 <Eddi|zuHause> half of the news publishers went "what? that's crazy even for our standards, google can just link our news for free"
19:50:27 <Eddi|zuHause> and for the other half google went "what? that's crazy, so we shorten our links to not contain any content, see how many people click on that"
19:50:49 <Eddi|zuHause> then they tried to force google to show the full details (and get paid), which didn't work
19:51:08 <Eddi|zuHause> and after like 2 years they finally gave in and said "fine, you can also show it for free"
19:52:39 <Eddi|zuHause> so in the end, google (which is 99% of all links) pays nothing, and smaller news aggregators can't compete with google anymore. nobody gets paid.
19:53:11 <Eddi|zuHause> well, except for the lawyers
19:59:57 <andythenorth> kinda wonder how it's supposed to work
20:00:10 <andythenorth> these EU-wide laws
20:00:14 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh approved pull request #7224: Change: Use selected group as parent when creating a new group. https://git.io/fh7Xm
20:00:17 <andythenorth> that seem to benefit nobody
20:02:35 <nielsm> laws that think to protect people who don't understand technology against technology, enacted by people who don't understand technology
20:04:05 <andythenorth> that benefit technology giants
20:06:39 <andythenorth> which means we've all opted in to data collection, through all those popups
20:10:23 <peter1138> Hi, I have dog food now.
20:10:26 <peter1138> And I have raw milk.
20:10:38 <peter1138> And the missus is cooking a giant dinner.
20:11:18 <peter1138> nielsm, is the AI/GS framerate stuff near ready?
20:11:59 <nielsm> well, it works as-is, question is mostly about presentation I suppose
20:14:25 <andythenorth> peter1138: are those the ingredients? :o
20:19:06 <peter1138> I should start removing old branches, yes.
20:25:27 *** Thedarkb1-T60 has joined #openttd
20:26:37 <peter1138> ^ Just a rebase, nothing new since... the other day, I guess.
20:44:58 <andythenorth> bananas, multiple people can upload the same grf right?
20:45:09 <andythenorth> I have uploaded my grfs from both coop and my account
20:45:30 <peter1138> Does it not at least check the GrfID?
20:45:49 *** gelignite has joined #openttd
20:47:18 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN opened pull request #7231: Add: Add flag and railtype sprite type to draw pre-combined ground sprites. https://git.io/fh71z
20:47:43 <peter1138> I just discovered the Draft feature... nice.
20:48:42 <peter1138> Ah but the list doesn't show Draft status.
20:50:17 *** Thedarkb-T60 has joined #openttd
20:52:49 <andythenorth> eh I have FIRS uploaded from 2 accounts on bananas
20:53:01 <andythenorth> but I can't upload FIRS 3 from my account, only coop
20:53:05 <andythenorth> don't really understand that
20:54:44 <peter1138> Just turn it into a free-for-all.
20:54:53 <peter1138> Problem is, that person expects it to be curated.
20:56:56 <andythenorth> person is not native english speaker
20:57:01 <andythenorth> so I'm not just dismissing as a fuckwit
20:57:32 <peter1138> Well, it's not a stupid request, it's just not what Bananas is.
20:57:53 <andythenorth> it's a confusing and inconsistent request
20:58:16 <andythenorth> I should probably stay out of it, I don't really have any use for bananas
21:08:03 <nielsm> bananas are useful for making desserts
21:08:17 <andythenorth> I just upload stuff
21:08:25 <andythenorth> I downloaded some pikka grfs years ago
21:08:46 <andythenorth> sometimes I have to test some savegame from GH or something
21:08:51 <andythenorth> then find all the missing content
21:08:59 <andythenorth> I don't really use grfs tbh
21:09:51 <peter1138> I used to always use them.
21:09:59 <peter1138> Now I just can't be bothered to set them up.
21:10:05 <peter1138> Even though it's easier than it ever was.
21:10:16 <peter1138> A GUI! Even for parameters! It's magic.
21:12:56 <andythenorth> I made 6 of them :P
21:22:41 <andythenorth> so station tiles with weird catchments peter1138? o_O
21:24:32 <Eddi|zuHause> make catchment area work the intuitive way that everyone thinks it does?
21:24:55 <Eddi|zuHause> also, make the game more euclidean
21:25:30 <peter1138> When do oil rigs appear?
21:25:59 <Eddi|zuHause> somewhere in that range
21:26:12 <Eddi|zuHause> never at mapgen, though
21:26:12 <peter1138> Do they appear on new game during that date?
21:26:31 <peter1138> Well, at least FFWD is... fast :)
21:27:01 <Eddi|zuHause> scenario editor? fund?
21:28:00 <andythenorth> I was just thinking bigger / smaller :P
21:28:52 <Eddi|zuHause> andythenorth: that is pretty meaningless as long as just the largest radius counts
21:29:33 <andythenorth> dunno how it would work anyway
21:29:39 <Eddi|zuHause> and if you want beautification tiles that don't increase area you got newobjects
21:29:52 <andythenorth> I was thinking something like, rail tiles are radius 1
21:30:03 <andythenorth> eye candy tiles get radius 20 or something
21:31:03 <andythenorth> need to figure out *something* so I can make use of industry > more cargos out
21:31:29 <andythenorth> stations are quite limiting there
21:33:40 <nielsm> random thought: would be cool if different railtypes (and roadtypes) could add their own icons to station signs
21:33:55 <nielsm> so e.g. traditional rail and monorail could have different icons
21:34:02 <peter1138> That could be doable.
21:34:13 <andythenorth> no steam train icon? :o
21:34:31 <andythenorth> hmm, dump truck icon for HAUL
21:35:36 <peter1138> Hmm, yeah, it's doable, but these are text glyphs rather than sprites.
21:35:37 <nielsm> also, again I'm going to suggest having some kind of compressed map block storage that could be used anywhere
21:35:45 <peter1138> Just a little bit fiddly.
21:35:45 <nielsm> (which could move some map data out of map arrays even)
21:36:13 <nielsm> e.g. a compressed bitmap (either by eliding blank blocks or using RLE) of a station's catchment area
21:36:39 <nielsm> so you do a rough check for stations possibly in catchment, then check the bitmaps for each of those stations for accurate result
21:38:00 <nielsm> or well, complex catchment areas with per-tile catchment could actually be calculated and cached on demand
21:38:29 <peter1138> I... I'm about to do that last thing anyway.
21:38:32 <Eddi|zuHause> each station gets a hashmap of which tiles it covers
21:38:38 <peter1138> Stations have tons of off-map storage anyway
21:40:00 <Eddi|zuHause> on industry production, you loop through each station and find a non-empty intersection between station area tiles and industry tiles
21:40:15 <Eddi|zuHause> on vehicle delivery you loop through all catchment area tiles
21:40:22 <Eddi|zuHause> or something along those lines
21:40:45 <Eddi|zuHause> it's neither crazy memory nor crazy runtime heavy
21:41:39 <peter1138> Yeah, I'm... about to do that ;)
21:41:46 *** Thedarkb-T60 has joined #openttd
21:41:59 <peter1138> Although I admit I was going to do an x,y array unless you wrote hashmap... D'oh.
21:42:49 <nielsm> nah std::set is probably not good
21:42:50 <Eddi|zuHause> i did 3 days worth of coding with STL containers, i'm now an expert
21:43:23 <peter1138> nielsm, I'll start off with whatever works, and then consider performance later.
21:43:33 <nielsm> for caching I think an RLE bitmap is the most efficient
21:44:01 <nielsm> in terms of cpu/memory balance
21:44:08 <Eddi|zuHause> what are we sending faxes or something? :p
21:44:15 <peter1138> Forget about space. Dedicated servers don't come with 1KB RAM...
21:44:15 <nielsm> a straight bitmap would be the fastest in straight cpu
21:44:41 <nielsm> yeah but if you have 4000 stations with large catchment areas now you're suddenly using rather much regardless
21:44:46 <Eddi|zuHause> i think CPU time is the more important one to optimize
21:45:00 <nielsm> also be wary of data structures that cause memory fragmentation
21:46:39 <peter1138> 4000 stations can't all cover the same area.
21:50:10 <Eddi|zuHause> i don't see memory fragmentation becoming an issue. if the station area is modified, you can easly scrap the cache and rebuild it from scratch
21:50:42 <Eddi|zuHause> it only happens on user interaction, so can be a more expensive operation
21:52:37 *** HerzogDeXtEr1 has joined #openttd
21:53:20 <_dp_> hi! does anyone know why msvc here gives error C2061: syntax error: identifier 'Polyline' ?
21:53:23 <_dp_> static bool FindPolyline(const Point &pt, const LineSnapPoint &start, Polyline *ret)
21:53:46 <_dp_> it's defined as a struct earlier in the same file and it compiles fine with g++
21:54:17 <nielsm> well I don't know if you suggested std::set as a joke, but it would get you a ton of tiny heap allocations (it's a balanced tree) so likely have poor cpu cache locality and cause potential trouble for the heap allocator if some of the elements are freed or moved around during building
21:55:14 <Eddi|zuHause> _dp_: is that the complete message?
21:55:46 <Eddi|zuHause> _dp_: almost always "syntax error"-style messages mean something went wrong BEFORE that point
21:56:07 <Eddi|zuHause> it's just at that point the compiler got confused enough to give up
21:58:51 <Eddi|zuHause> _dp_: could be somthing interfering from a MSVC-specific header
21:59:43 <nielsm> you have an #include at line 968 well past the usual include block, try moving that
21:59:45 <_dp_> Eddi|zuHause, yeah, I though of that but couldn't find anything
22:00:16 <_dp_> also I compiled 1.8 version with studio before
22:00:35 <Eddi|zuHause> _dp_: have you tried just the polyline patch individually?
22:01:12 <nielsm> viewport.cpp(2920): warning C4551: function call missing argument list
22:01:21 <nielsm> 2920: Polyline polyline;
22:01:31 <nielsm> it absolutely thinks Polyline is the name of a function, not your struct
22:01:39 <ST2> _dp_: I've also had issues with msvc project files, but sidework made me postpone this :|
22:02:32 <_dp_> Eddi|zuHause, no but it's a bit modified patch already, don't have it separate
22:03:04 <_dp_> also experimenting with azure so it's not quite easy to try random guesses :(
22:03:42 <_dp_> since it don't have studio myself and can only run pipelines
22:04:32 <_dp_> nielsm, I'll try renaming it than, thanks!
22:04:36 <Eddi|zuHause> definitely move those stray #includes
22:04:49 <Eddi|zuHause> so your struct definition will happen later
22:05:58 <Eddi|zuHause> (and people wonder why we didn't include some patches in master)
22:06:22 <Eddi|zuHause> (after all, they were "well tested" for years)
22:06:33 <_dp_> Eddi|zuHause, stray includes come from vanilla openttd btw
22:07:17 <_dp_> well, it's is tested... as long as it compiles :p
22:07:52 <peter1138> Okay, so with #7230
22:07:53 <nielsm> hmm true yes, those includes are in master
22:08:04 <peter1138> The question is... *should* the 90 degree checks be moved outside of the pathfinder?
22:08:31 <peter1138> forbidding 90 degree turns is basically a feature OF the pathfinder, so...
22:09:05 <nielsm> it looks like that's the only call site of that function
22:09:15 <nielsm> and where it's used is pretty much already pathfinding code
22:09:19 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN commented on pull request #7230: Fix #7226: No ship track due to "forbid 90 deg turns"-> Do not call pathfinders. https://git.io/fh7Dk
22:09:53 <Eddi|zuHause> peter1138: really, ships shouldn't forbit 90° turns
22:10:03 <Eddi|zuHause> there is no sane reason for doing it
22:10:03 <peter1138> Eddi|zuHause, that isn't the point
22:11:43 * peter1138 rebases his patch which removes it for ships.
22:11:45 <Eddi|zuHause> the diff looks sane though
22:13:07 <Eddi|zuHause> but why do we need both a DiagDir and a TrackDir? that seems redundant
22:14:47 <peter1138> TrackDir identifies upper/left or lower/right track bits.
22:15:12 <peter1138> DiagDir is just the 4 directions.
22:15:21 <peter1138> So they are not equivalent.
22:16:17 <Eddi|zuHause> i mean, you could infer the diagdir from the trackdir
22:17:27 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN opened pull request #7232: Change: Don't apply forbid 90 deg turn settings to ships. https://git.io/fh7DC
22:19:50 <peter1138> Looks like OPF and NPF didn't forbid 90 deg turns anyway?
22:19:59 <peter1138> Anything in NPF mentions trains.
22:20:03 <peter1138> And OPF just doesn't.
22:21:02 <peter1138> Hmm, actually forbidding 90 deg turns now makes ships faster, as on-the-spot rotation makes it stop.
22:21:35 <peter1138> Hmm, I guess increase penalty for 90 degree turns?
22:29:50 <peter1138> Ok, where's the function to get Direction from Trackdir, anyway? :/
22:35:41 <peter1138> remove forbid 90 degrees.
22:35:48 <Samu> it was only opf that didn't
22:35:52 <peter1138> Onlly removing it for ships, not trains.
22:36:12 <Samu> it's ... hmm somewhere in npf, driveable trackbits i think
22:39:57 <peter1138> Oh, of course. I already patched that out!
22:40:11 <peter1138> Only checks for TRANSPORT_RAIL :)
22:42:45 <peter1138> Yeah, sorry. Forgot I'd taken care of it.
22:44:56 <Eddi|zuHause> peter1138: something about "trackdir to exitdir" pops into my head
22:45:52 <peter1138> Nah, that's a DiagDir
22:51:46 <Eddi|zuHause> that's what we wanted, i thought?
22:54:59 <peter1138> No, I wanted a Direction.
22:57:14 <peter1138> Hmm, well, adding a penalty doesn't seem to work :/
23:34:09 <DorpsGek_II> [OpenTTD/OpenTTD] J0anJosep commented on pull request #7230: Fix #7226: No ship track due to "forbid 90 deg turns"-> Do not call pathfinders. https://git.io/fh7Sk
23:48:38 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN commented on pull request #7230: Fix #7226: No ship track due to "forbid 90 deg turns"-> Do not call pathfinders. https://git.io/fh7Sc
23:49:08 <peter1138> Hmm, so the penalty *is* working...
23:49:12 <peter1138> Just not on the last segment.
23:50:32 <DorpsGek_II> [OpenTTD/OpenTTD] J0anJosep commented on pull request #7232: Change: Don't apply forbid 90 deg turn settings to ships. https://git.io/fh7SB
23:54:29 *** Supercheese has joined #openttd
continue to next day ⏵
