IRC logs for #openttd on OFTC at 2016-09-04
⏴ go to previous day
00:04:48 *** mescalito has joined #openttd
00:05:38 <Samu> hmm problem seems to be with 2x2 grid and 3x3 grid
00:08:40 <Samu> static RoadBits GetTownRoadGridElement(Town *t, TileIndex tile, DiagDirection dir)
00:11:41 <_dp_> Samu, and if you're looking for infinitely growing layout there aren't many options really
00:11:41 <_dp_> Samu, because anything that has loops or dead ends is gonna stuck eventually
00:16:45 <Samu> i need to read log, i'm lost in conversation
00:19:38 <Samu> hopefully my computer doesn't crash in the mean time
00:19:57 <Samu> okay, the issue I got is with 2x2 and 3x3 grids, about roadbits growth
00:20:25 <Samu> when extending the road in a direction, it still builds on water
00:21:24 <Samu> I see a rb_template that disregards anykind of water in the tile, it only looks at the slope configuration
00:21:56 <Samu> something tells me this is the faulty function GetTownRoadGridElement
00:22:14 <Samu> something needs to be added to check for water in the tile, but how am I doing that?
00:22:47 <Samu> but i may be wrong, the whole town growth stuff is complicated to follow :(
00:27:53 <_dp_> Samu, and what's the problem with water?
00:28:49 <_dp_> Samu, bridge counts as one road tile so except for town zones it hardly affects anything
00:42:45 <Samu> if (!IsRoadAllowedHere(t1, tile, target_dir)) return; is missing there
01:03:10 <Samu> if (HasTileWaterGround(tile)) return; - this function is returning false for coasts....
01:07:17 *** CompuDesktop has joined #openttd
01:07:19 *** sim-al2 has joined #openttd
01:12:31 *** sim-al2 has joined #openttd
01:12:31 *** Supercheese has joined #openttd
01:12:31 *** mindlesstux has joined #openttd
01:12:31 *** Clockworker has joined #openttd
01:12:31 *** planetmaker has joined #openttd
01:12:31 *** ccfreak2k has joined #openttd
01:12:31 *** charm.oftc.net sets mode: +ov planetmaker planetmaker
02:02:05 <Samu> alright, so houses have been dealt with correctly
02:03:12 <Samu> bridges, I am not quite sure, since they follow the road
02:09:26 <Samu> just want to make sure...
02:20:18 <Samu> /* Prevent towns from building roads where ships can traverse */
02:26:42 <Samu> /* Don't build the bridge if the bridge head at the destination is traversable by ships */
03:02:57 <Samu_> ewmm... i broke something, towns don't want to build bridges over rivers
03:28:21 <Samu_> I can't reproduce the issue, :(
03:30:53 *** mindlesstux_ has joined #openttd
03:34:35 *** mindlesstux_ is now known as mindlesstux
03:48:39 <Samu> can you see the difference?
03:53:59 <goodger> one of them has built a bridge to nowhere at vast expense
04:08:53 <Samu> the bridge to the left of Kedtown
04:09:51 <Samu> with my patch, every tile that contains water won't have anything built in it
04:10:45 <Samu> I find it a bit restrictive with bridges though
04:10:55 <Samu> I'm not sure if I like that
04:14:03 <Samu> well, cyas all take care
05:10:40 *** trendynick has joined #openttd
05:25:41 *** Lejving has joined #openttd
07:09:28 *** sla_ro|master has joined #openttd
08:10:58 *** Alberth has joined #openttd
08:10:58 *** ChanServ sets mode: +o Alberth
10:08:26 *** andythenorth has joined #openttd
10:22:34 *** Progman has joined #openttd
10:36:25 *** tycoondemon has joined #openttd
11:05:34 *** frosch123 has joined #openttd
12:18:51 <andythenorth> such variations, such few pixels
12:22:26 <frosch123> i would still recommend against generating random pixel pattern and trying to filter them for good vehicle sprites
12:24:22 <andythenorth> generative sprites is a nice idea
12:24:34 <andythenorth> how much time will it save?
12:27:37 <Alberth> not so sure what unit of time to attach :p
12:28:23 <frosch123> but it says years :)
12:28:33 <frosch123> it's my favorite time unit :p
12:29:50 <Alberth> you can further configure it by changing the speed you're traveling
12:51:56 <Alberth> Compared to the previous time (still available if you remove the "2"), 060 had the commit message typo fixed, 180 was rewritten, and 200 is merged into it, 190 and all other patches are unchanged
13:10:30 *** HerzogDeXtEr has joined #openttd
13:20:11 *** andythenorth has joined #openttd
13:27:00 <Alberth> thanks for the reviews and useful comments
13:55:46 <andythenorth> think that’s ‘done'
13:56:07 <andythenorth> except for the angles I haven’t drawn on 10 trams :P
14:05:44 *** mescalito has joined #openttd
14:08:01 <Wolf01> Now need only proper lightrailtypes
14:08:59 <Wolf01> Went to town fair this morning, seen donkeys with 5 legs
14:14:41 <Samu> it is using DistanceManhattan to get the coordinates of the closest ship depot near the ship
14:15:55 <Samu> the problem is when there's obstacles in the way
14:16:21 <Samu> or even when the depot is in another water body, innaccessible by the ship
14:19:11 *** sla_ro|master has joined #openttd
14:20:19 <Samu> pathfinding would come next
14:20:30 <Samu> i mean, after finding the closest ship depot
14:20:49 <Samu> the pathfinding will try to direct the ship towards those coordinates
14:22:58 <Wolf01> First you use pathfinding to find a path, then the ship follows the path
14:23:13 <Wolf01> It's called "path finding" for a reason
14:23:35 <Samu> the pathfinder needs the coordinates of the destination
14:24:11 <Samu> but this function is giving it coordinates of possible inaccessible deots
14:24:21 <Wolf01> Andy, I let your immagination do the rest
14:24:51 <_dp_> Samu, only pathfinder can check accessibility
14:25:29 *** gelignite has joined #openttd
14:26:48 <andythenorth> Wolf01: I get the picture :P
14:27:22 <Wolf01> Samu, get a brief tour on how A* works
14:27:52 <andythenorth> Wolf01: actually need roadtypes for mining trucks ;)
14:28:01 <andythenorth> I’ve removed them from Road Hog temporarily
14:28:11 <andythenorth> they make no sense without dedicated road
14:28:41 <Samu> for all depots, pre-pathfind and get favourable results? then give the best favourable coordinates to the real pathfinder
14:28:55 <Wolf01> And pathfind it again?
14:29:16 <Wolf01> Once you find the depot you already have a path
14:29:55 <Wolf01> Also, you don't need to have a patfind to get the coordinates, the pathfinder is used to check if the coordinates are accessible
14:30:28 <Wolf01> For coordinates you can just scan a map sector
14:33:10 <Wolf01> If you use a pathfinder to find something which could be inaccessible you might end with an infinite loop
14:33:22 <Wolf01> Or you can limit the iterations
14:44:50 <DorpsGek> Commit by alberth :: r27633 /trunk/src (3 files) (2016-09-04 14:44:42 +0200 )
14:44:51 <DorpsGek> -Codechange: Extract _saveload_mode use from BuildFileList
14:45:17 <DorpsGek> Commit by alberth :: r27634 /trunk/src (fios.cpp fios.h) (2016-09-04 14:45:11 +0200 )
14:45:18 <DorpsGek> -Codechange: Improve name of the SmallFiosItem struct.
14:45:46 <DorpsGek> Commit by alberth :: r27635 /trunk/src (6 files in 2 dirs) (2016-09-04 14:45:40 +0200 )
14:45:47 <DorpsGek> -Codechange: Move FileType and FileToSaveLoad structure definitions.
14:46:13 <DorpsGek> Commit by alberth :: r27636 /trunk/src (3 files in 2 dirs) (2016-09-04 14:46:07 +0200 )
14:46:14 <DorpsGek> -Codechange: Rename FileType to AbstractFileType.
14:46:35 <DorpsGek> Commit by alberth :: r27637 trunk/src/saveload/signs_sl.cpp (2016-09-04 14:46:29 +0200 )
14:46:36 <DorpsGek> -Codechange: Don't use _saveload_mode for scenario loading detection.
14:47:13 <DorpsGek> Commit by alberth :: r27638 /trunk/src (7 files in 2 dirs) (2016-09-04 14:47:07 +0200 )
14:47:14 <DorpsGek> -Codechange: Move FiosType enum, move and rename SetFiosType function.
14:47:44 <DorpsGek> Commit by alberth :: r27639 /trunk/src (fileio_type.h saveload/saveload.cpp) (2016-09-04 14:47:39 +0200 )
14:47:45 <DorpsGek> -Codechange: Also always set the abstract FileToSaveLoad::filetype when setting a mode.
14:47:58 <Alberth> not so nice memories about FileType ?
14:48:33 <DorpsGek> Commit by alberth :: r27640 trunk/src/openttd.cpp (2016-09-04 14:48:28 +0200 )
14:48:34 <DorpsGek> -Codechange: Remove another use of _saveload_mode in the loading code.
14:50:28 <DorpsGek> Commit by alberth :: r27641 /trunk/src (5 files in 2 dirs) (2016-09-04 14:50:22 +0200 )
14:50:29 <DorpsGek> -Codechange: Fold the _fios_items file list vector into its own class.
14:54:10 <DorpsGek> Commit by alberth :: r27642 /trunk/src (4 files) (2016-09-04 14:54:03 +0200 )
14:54:11 <DorpsGek> -Codechange: FiosGet* file query functions take a destination file list.
14:54:36 <DorpsGek> Commit by alberth :: r27643 /trunk/src (4 files in 4 dirs) (2016-09-04 14:54:30 +0200 )
14:54:37 <DorpsGek> -Codechange: FiosGetDrives function also takes a destination file list.
14:54:58 <DorpsGek> Commit by alberth :: r27644 /trunk/src (4 files) (2016-09-04 14:54:52 +0200 )
14:54:59 <DorpsGek> -Codechange: Split GetFiosItem into BuildFileList and FindItem, and move both to FileList.
14:55:27 <DorpsGek> Commit by alberth :: r27645 trunk/src/console_cmds.cpp (2016-09-04 14:55:21 +0200 )
14:55:28 <DorpsGek> -Add: Give console commands their own file list storage.
14:56:00 <DorpsGek> Commit by alberth :: r27646 /trunk/src (3 files) (2016-09-04 14:55:54 +0200 )
14:56:01 <DorpsGek> -Codechange: Move _fios_items variable into the SaveLoadWindow class.
14:56:29 <DorpsGek> Commit by alberth :: r27647 /trunk/src (6 files) (2016-09-04 14:56:23 +0200 )
14:56:30 <DorpsGek> -Codechange: Introduce file operations, and use it to replace most of SaveLoadDialogMode
14:57:02 <DorpsGek> Commit by alberth :: r27648 /trunk/src (5 files in 2 dirs) (2016-09-04 14:56:56 +0200 )
14:57:03 <DorpsGek> -Codechange: Remove remaining _saveload_mode usage.
14:57:25 <DorpsGek> Commit by alberth :: r27649 /trunk/src (5 files in 2 dirs) (2016-09-04 14:57:20 +0200 )
14:57:26 <DorpsGek> -Codechange: Introduce detailed file type enum, rebuild FiosType with it.
14:57:49 <DorpsGek> Commit by alberth :: r27650 /trunk/src (16 files in 4 dirs) (2016-09-04 14:57:43 +0200 )
14:57:50 <DorpsGek> -Codechange: Replace SaveOrLoadMode by FileOperation and DetailedFileType.
14:58:10 <DorpsGek> Commit by alberth :: r27651 /trunk/src (5 files in 2 dirs) (2016-09-04 14:58:04 +0200 )
14:58:11 <DorpsGek> -Codechange: Introduce methods for setting the name and title of _file_to_saveload.
14:58:51 <Alberth> right, now try stuff I wanted to do in the first place :p
15:18:52 * andythenorth has 1 out of 10 trams done :P
15:19:07 <andythenorth> such productivity :P
15:20:35 *** sim-al2 has joined #openttd
15:20:42 <Samu> gonna ruin my patches probably
15:21:56 <Alberth> 10% done already, andy :)
15:22:44 <Alberth> Samu: unless you change things near file loading/saving handling, not likely
15:24:37 <Samu> I did change something related to saveloads for AI scripts
15:25:45 <Samu> add start_date parameter for Random AIs on new game v1 r27603 - still works
15:26:25 *** Flygon_ has joined #openttd
15:27:27 <Samu> saveload.cpp, saveload.h have been changed, grrr
15:28:00 <andythenorth> Alberth: I uncovered a bug with 8 more trams :P
15:28:07 <andythenorth> one step forward…one step back...
15:29:38 <Samu> faster server autosaves v5 r27601 - i am getting weird results, it says it patches everything fine, but also 2 failed hunks for every file, tortoisesvn what will it be?
15:32:04 <Samu> need to look at this more carefully
15:35:03 <Samu> fios.cpp - never touched it
15:38:10 <Samu> toolbar_gui.cpp - i touched it, but then reverted, so i left it back to what it was, that means... im not touching it now
15:38:38 <Samu> dedicated_v.cpp - never touched it
15:38:53 <Samu> fios_gui.cpp - never touched
15:40:47 <Samu> console_cmds.cpp - i touched it but ended up not needing to change anything
15:46:17 *** supermop has joined #openttd
15:49:09 <Samu> openttd.cpp, saveload.h, saveload.cpp, afterload.cpp - only these 4 files that I need to check for conflicts
16:09:33 <Samu> saveload.cpp - Alberth adds about the way it accesses files
16:09:56 <Samu> I add stuff about the way it saves the stuff
16:10:22 <Samu> not really conflictuous, but it touches similar areas
16:10:36 <Alberth> usually, I add stuff, and then remove other stuff :)
16:10:56 <Alberth> but yeah, could be in somewhat the same area
16:12:23 <Samu> what goes into the savegame, more specifically
16:12:34 <Wolf01> Meh, I should give up programming
16:13:53 <Samu> I've also added some other compression methods
16:13:53 <Wolf01> It looks easy but has 24536435734067293620349632e45346467 problems
16:14:06 <Samu> Alberth didn't touch that part
16:15:30 <Alberth> I didn't touch content of savegame, only split information of file names into more separate parts
16:16:52 <Alberth> :o more problems than molecules in the universe?
16:16:57 <andythenorth> Wolf01: try brogramming instead
16:17:05 <andythenorth> of being devloloper
16:17:19 <Samu> you got code at line 270, i got code at line 275, it's really close, and the tortoisesvn patching isn't all that great at putting stuff in the right place
16:17:25 <Wolf01> I would like to take an advanced course but only on what I need
16:18:23 <Alberth> Samu: it has good reason not to be too good, a developer is much better at deciding the right solution
16:19:02 <Wolf01> For example now I'm just adding a "usercontrol" which contains a scrollviewer with an image, I need to pass a property to it (the image source)... no matter of what I try, I always end up with nullPointerException... and I set up everything right
16:19:48 <Wolf01> And it's 4 lines of code and 2 clicks!
16:20:04 <Wolf01> I even used the same names of the demo
16:20:59 <Wolf01> The demo had also the image caption which I don't need
16:21:21 <Samu> next change for alberth is from 2781 and below
16:21:38 <Samu> the whole bulk of my changes are before that
16:22:02 <Samu> only got a change at 2851
16:23:58 <Samu> that's right in the middle of a function you make tons of changes
16:24:25 <Samu> this may be the only real conflict
16:25:03 <Samu> if (_network_server || !_settings_client.gui.threaded_saves) threaded = false;
16:25:48 <Samu> i changed it to if (!_settings_client.gui.threaded_saves) threaded = false;
16:25:48 *** Wolf01 is now known as Guest740
16:25:49 *** Wolf03 is now known as Wolf01
16:26:21 <Samu> but no idea if that would achieve the same expected behaviour with albert changes
16:29:20 <Samu> there should be a 3-way diff viewer
16:29:50 <Samu> revision 27632, my patch and the latest trunk version
16:30:58 <Wolf01> I think there is, at least I used it in PHPStorm
16:34:38 <Samu> faster server autosaves - RIP
16:34:38 <andythenorth> Wolf01: make some roads instead? o_O
16:34:59 * andythenorth needs a break from teeny tiny trams
16:35:11 <Wolf01> Not sure on what to do, I would like to be sure instead of redoing the same again and again
16:35:28 <Wolf01> If you want to follow me on the process I might try it
16:36:10 <andythenorth> I can also test compile
16:36:57 <Wolf01> Let me just some minutes to finalize the image zoom of my feed reader (without the f*****g usercontrol)
16:37:52 <supermop> andythenorth: draw roads?
16:40:39 <Samu> tortoisesvn patched saveload.cpp correctly, but i wonder if the behaviour about threaded saves is still the same
16:42:41 <Samu> openttd.cpp change is totally unrelated, there will be no conflict
16:43:10 <Samu> my change is about adding a start_date for random ais
16:43:42 <Samu> Alberth change is about save and load stuff
16:44:41 <Alberth> why do you constantly highlight me, please stop doing that
16:45:27 <LordAro> ooh, Alberth commit queue
16:46:46 <Samu> i'm thinking out loud saveload.h
16:47:03 <Wolf01> Ok, andythenorth, I'm ready to do something
16:47:32 <andythenorth> 4 trams left, none of which I want to do right now :D
16:47:33 <Wolf01> Not sure what, but something must be done
16:48:21 <andythenorth> I think we can do it piece by piece
16:48:37 <Wolf01> Yes, that's a good idea
16:48:38 <andythenorth> first move the storage to m4, but provide a shim
16:48:46 <andythenorth> so that all old function calls keep working
16:49:08 <andythenorth> let’s try and rebuild the bridge while the trains still go over it :P
16:49:56 <andythenorth> patch will be too big to throw everything off the cliff and have a non-compiling game while we make 80 changes to how tile gets roadtype
16:50:58 <Wolf01> Ok, since today is saveload.h day I'll try to break things too
16:52:20 <Wolf01> I'll just move m7 7..6 to m4 1..0?
16:52:21 <Samu> max_no_competitors = 15 v31 r27601 - it appears there is no conflict with this
16:52:41 <Wolf01> Or do you have another suggestion?
16:53:13 <Samu> we touch the same files, but not the same functions
16:53:29 <andythenorth> use m4 0…3 for road and m4 4…7 for light rail
16:53:45 <andythenorth> and assume for both that 0 = default
16:53:57 <Wolf01> 0 means no road and no tram
16:54:00 <andythenorth> that will change after newgrf support, but is safe for a long time
16:54:18 <andythenorth> ah I proposed to store an integer there corresponding to type ;)
16:54:24 <Wolf01> 0001 0000 = road but no tram, 0000 0001 tram but no road
16:55:20 <andythenorth> so 0 is ‘no [road | light rail] type present’
16:55:26 <Wolf01> And obviously 0001 0001 both
16:55:39 <andythenorth> how do we get 16 types of each if 0 is needed to define ‘none'?
16:55:58 <Wolf01> You get 15 types where 0 is none
16:56:08 <Eddi|zuHause> i don't think you need a "none" type
16:56:13 <andythenorth> or we could use the track bits?
16:56:15 <Eddi|zuHause> you just have no road bits
16:56:21 <Wolf01> Or we don't deprecate the m7 7..6 and get full 16 types
16:57:03 <andythenorth> shouldn’t need them :)
16:57:15 <andythenorth> use track bits, infer presence of road / light rail
16:57:18 <Eddi|zuHause> if there are no road bits present, the road type is ignored
16:57:28 * andythenorth needs an acronym for light rail
16:57:31 <Eddi|zuHause> and if road bits are present, the road type must be valid
16:57:47 <Wolf01> Or as eddy says, we could do some math involving m3 0..3
16:58:18 <Eddi|zuHause> if "is not zero" is "math"?
16:59:32 <supermop> andythenorth: me. i still have so melb-ish roads and tramways i rendered while i was living there
16:59:33 <andythenorth> m5 0..3 for road as well as m3 0..3 for tram
16:59:40 <Wolf01> We'll need to change also the level crossing stuff, now it always assume road is present for tram
16:59:44 <Eddi|zuHause> what's the current meaning of m7 7..6?
17:00:04 <andythenorth> present road types
17:00:16 <andythenorth> could already be inferred from m5 0..3 and m3 0..3
17:00:22 <andythenorth> maybe there are performance reasons not to
17:00:24 <Eddi|zuHause> that's probably redundant
17:00:42 <Eddi|zuHause> i don't think performance is relevant here
17:00:43 <andythenorth> seems redundant to me
17:01:22 <Wolf01> Ok, so 0 is normal road|tram
17:01:36 <Eddi|zuHause> afair the reason for the crossing limitation was that there are no sprites for trackbeds of tram & rail
17:02:02 <andythenorth> ach, so m7 isn’t redundant, because of level crossings
17:02:26 * andythenorth wondered if we could refactor that first to clean house
17:02:41 <Wolf01> You need to have both m3 and m5 bits because you can have different layouts for road and tram in the same tile
17:02:53 <Eddi|zuHause> Wolf01: so that breaks existing tram track grfs, which don't provide the additional sprites?
17:03:09 <Samu> Error C2011 'FileOperation': 'enum' type redefinition (compiling source file ..\src\os\windows\win32.cpp) openttd C:\Program Files (x86)\Windows Kits\8.1\Include\um\shobjidl.h 1852
17:03:24 <Wolf01> We should find a workaround or fix the existing tram track grfs
17:03:48 <andythenorth> substitute a sprite
17:03:52 <Samu> visual studio does not like r27651
17:04:21 <Eddi|zuHause> Wolf01: there are some existing workarounds for similar stuff, like a grf providing too few shore sprites
17:05:05 <Eddi|zuHause> particularly, the shore sprites where only one corner has shores, but no edge touches a water sprite
17:05:43 <Alberth> what's this um\shobjidl.h file? it's not of openttd
17:06:07 <Samu> line says typedef class FileOperation FileOperation;
17:06:34 <Samu> line before says #ifdef __cplusplus
17:06:56 <Wolf01> Samu, don't include default windows libraries, they have classes which conflict with ottd useful ones and also with ottd
17:07:39 <andythenorth> Wolf01: I think I would start by adding something like HasLightRail and HasStreet to the tile functions
17:07:56 <Samu> i just tried to build openttd r27651
17:08:05 <andythenorth> then replace calls in the drawing and construction code to HasBit() that look at m7 0..1
17:08:15 <andythenorth> then we have some abstraction
17:08:23 <andythenorth> then we can move the storage of type to m4 later
17:08:35 <andythenorth> road_cmd.cpp L1106 is an example
17:10:47 <DorpsGek> Commit by alberth :: r27652 trunk/src/saveload/saveload.cpp (2016-09-04 17:10:41 +0200 )
17:10:48 <DorpsGek> -Fix(r27650): Use the file operation being performed to set the _sl.action variable.
17:11:30 <Eddi|zuHause> i'm out of ideas how to handle this Baldy's Boss guy. he seems to have a roleplaying approach to the game, but then projects that roleplaying onto game mechanics that don't exist
17:11:31 <Wolf01> Yup, just refactoring, no new features
17:11:46 <andythenorth> Eddi|zuHause I think he can be safely left alone :)
17:11:52 <andythenorth> I think he is happy in his own little world
17:11:58 <andythenorth> the questions aren’t really questions
17:12:35 <Alberth> it's fun tinkering with his save game :)
17:13:16 <Wolf01> road_map.h is a good place?
17:14:16 <Alberth> it can always be moved later :)
17:14:26 <V453000> Eddi|zuHause: he is retarded, you can't do much with that
17:14:39 <andythenorth> iirc correctly, you end up having to import road_map.h to places it wasn’t imported before
17:14:49 <Eddi|zuHause> V453000: i don't think that's the right word
17:14:55 <V453000> he might hold the record of opening completely idiotic threads
17:15:10 <Wolf01> Just like me :D:D:D:D:D:D
17:17:51 <Eddi|zuHause> V453000: we had a SirXavius.
17:18:08 <V453000> don't remember him luckily
17:19:38 *** Wormnest has joined #openttd
17:20:09 <Samu> who's an expert windows compiler for openttd, :o
17:20:26 <Samu> gonna try compite r27632 again
17:21:42 <Alberth> V: not remembering the 36 or so pages he once posted about a new kind of openttd game? :p
17:23:40 <Samu> will try compiling every version from r27632 to r27652 to see where it fails
17:23:58 <Wolf01> andythenorth, we already have HasTileRoadType(TileIndex t, RoadType rt)
17:25:25 <andythenorth> Wolf01: that needs splitting imho
17:25:27 <andythenorth> if (HasTileRoadType(ti->tile, ROADTYPE_TRAM)) {
17:25:29 <Wolf01> if (HasBit(rts, rt)) { is just if (HasTileRoadType(tile, rt)) {
17:26:25 <andythenorth> ROADTYPE_TRAM corresponds to the m7 bit for tram
17:26:47 <andythenorth> we need another way to know if there is tram on the tile
17:29:15 <Alberth> what visual studio version do you have?
17:30:36 <Samu> Microsoft Visual Studio 2015 Shell (Integrated)
17:30:51 <Alberth> ok, should be new enough
17:31:02 <Samu> version 14.0.25420.01 Update 3
17:33:36 *** tokai|noir has joined #openttd
17:33:36 *** ChanServ sets mode: +v tokai|noir
17:35:04 <andythenorth> then we have to track all the uses of that down
17:35:11 <Alberth> 27647 will fail as first revision, I think
17:37:23 <frosch123> some windows header already uses that name
17:37:53 <Alberth> what about LoadSaveOperation ?
17:38:19 <Alberth> any way to test that without a commit?
17:38:20 <frosch123> oh, you were already on the subject
17:39:02 <frosch123> i think commit is the easiest option, unless you have a win compile environment
17:39:22 <Wolf01> Pass him the patch and let him do it for you
17:40:46 <frosch123> Alberth: "SaveLoad" feels more common to me than "LoadSave"
17:41:40 <Samu> oh i had no idea u got stuff to test these things
17:41:58 <LordAro> Samu: you should learn how to bisect
17:42:13 <LordAro> absolutely no need to test every single commit
17:42:25 <Wolf01> Binary tree, this long unknown friend
17:42:53 * LordAro divides Wolf01 into 2 equal parts
17:43:22 <Wolf01> And then checks in which one the stupid is, then divides it again
17:44:27 <LordAro> things aren't going to change that much
17:44:41 <LordAro> and backwards compatibility is important too
17:48:06 <Wolf01> andythenorth, I think this refactoring is pointless, I can't find a place where bits are used, and it just needs to change "HasTileRoadType(TileIndex t, RoadType rt)" or "GetRoadTypes(TileIndex t)"
17:51:16 <Wolf01> Personal continuous integration guy
17:51:33 <Alberth> very good in counting too
17:52:17 <andythenorth> Wolf01: I can’t find bits either
17:52:50 <Samu> Severity Code Description Project File Line Suppression State Error C3861 'SHSaveLoadOperation': identifier not found openttd D:\OpenTTD\trunk\src\ini.cpp 110
17:54:02 <Samu> rebuilding usually takes more time to build
17:54:33 <Samu> got the error already, but it's still building the rest
17:56:51 <Samu> 5 warnings, but those warnings have been there for years
17:57:17 <Wolf01> andythenorth, what if I start with defining the new entries on the enum but pointing to the old values?
17:57:37 <Alberth> Samu: interestingly, that file builds for me :)
17:58:06 <Alberth> change line 110 back to SHFileOperation(&shfopt);
18:03:58 <Samu> i built debug x64, let me try the others
18:06:56 <DorpsGek> Commit by alberth :: r27653 /trunk/src (15 files in 4 dirs) (2016-09-04 18:06:50 +0200 )
18:06:57 <DorpsGek> -Fix(r27647): Rename FileOperation enum and values to SaveLoadOperation to avoid nameclash with windows compiler toolkit.
18:10:21 <andythenorth> Wolf01: sounds plausible
18:10:35 <andythenorth> my brain goes ‘bzrt’ when I deal with enums
18:10:52 <Samu> release win32, it builds
18:10:53 <andythenorth> I keep thinking they’re some kind of list :P
18:11:24 <Alberth> andy, but they are, a list of constants :)
18:12:27 <Samu> tested with openttd_vs140.vcxproj
18:12:35 <andythenorth> seems more like a lookup table to me
18:12:45 <Samu> i wont test vs100, visual studio always complains about versions
18:12:45 <andythenorth> don’t ask me to explain how my brain works :P
18:13:00 <andythenorth> to me a list is a thing that can have pop, push, etc :P
18:13:20 <Wolf01> Enums are just a mean to give names to numbers
18:13:52 <Wolf01> And lists which could be traversed
18:13:57 <andythenorth> why not just use C pre-processor constants?
18:14:02 <Alberth> compile farm will do vs100, perhaps
18:14:22 <Wolf01> Constants could not be traversed :P
18:15:01 <Wolf01> Look at them like a constant list
18:16:48 <Wolf01> With enums, if you are writing a switch (MyEnum value) a lot of ides automatically provide all cases for the "MyEnum"
18:22:23 <andythenorth> so we’ll have a new RoadTypes enum?
18:22:46 <andythenorth> and RoadType also
18:26:12 <Wolf01> Not sure about the 0x7
18:26:28 <andythenorth> I’m sure I didn’t write that patch :)
18:26:33 <andythenorth> someone gave me most of that paste
18:27:43 <andythenorth> FWIW, I think it’s fine to only have 15 each of tram / street
18:27:51 <andythenorth> and use one value for ‘none’
18:27:58 <andythenorth> but also not necessary
18:28:41 <LordAro> those *SubtypeHasCatenary functions are horrible
18:28:49 <Wolf01> The static inline bool HasRoadType(rt) could just check if has m5 3..0
18:30:47 <LordAro> well that's ok then :)
18:31:23 <andythenorth> 3 trams to draw :(
18:31:30 <andythenorth> 12 bugs fixed :P
18:47:16 <Samu> i'm back to that depot choice problem
18:52:10 *** Wormnest has joined #openttd
19:05:53 <Alberth> Samu: deciding what the desired result is (in the general case), would be a start
19:09:46 <Samu> desired result is any acessible depot
19:11:19 <Wolf01> How did it get the blocked depot service order? If from pathfinding then there's a problem in pathfinding, if by map scn, then you should run a pathfind routine for each found depot to check if accessible
19:11:50 <andythenorth> Wolf01: that’s how I thought it would work
19:11:59 <andythenorth> we’ll need an equivalent setter for construction
19:12:41 <Alberth> simple approach is to find the closest non-examined depot on the map, and try to do a path find operation
19:13:08 <Wolf01> Not really sure on checks, wolves are not really good at math
19:13:13 <Alberth> if it works, you have a path, if it fails, find the next closest non-examined depot, and try that one
19:13:27 <Alberth> etc, until you find a depot, or you run out of depots
19:14:15 <Alberth> smart approach is to run a path find from all depots to the ship
19:14:48 <Wolf01> Implement multithreading
19:15:07 <Alberth> smart approach is not that much worse than from the closest reachable depot
19:15:45 <Samu> how do i "invoke" the pathfinder ? :(
19:16:09 <Alberth> no idea, never done that
19:16:56 <Alberth> look for a pathfind routine that looks like it computes a path, then find how it is used
19:24:28 *** Progman has joined #openttd
19:26:31 *** ChanServ sets mode: +o orudge
19:27:17 *** planetmaker has joined #openttd
19:27:17 *** ChanServ sets mode: +o planetmaker
19:27:27 *** Clockworker has joined #openttd
19:27:40 <_dp_> Alberth, that's num_of_depots*shortest_path even in good case, smartly checking from closest to furthest will almost always be better
19:29:15 <Alberth> _dp_: ... ? You know A* does not expand further-away nodes until needed, right?
19:30:21 <Samu> there are 3 pathfinders for ships t.t
19:31:47 *** ccfreak2k has joined #openttd
19:31:55 <Samu> i'm trying to compare how openttd computes the nearest road vehicle depot to how it's done for ships
19:32:45 <_dp_> Alberth, ah, forgot how it works, nvm then, checked wiki, it does pretty much the same as what I meant by smartly checking.
19:34:17 <Alberth> "from depots" is the smart trick here :)
19:35:44 <_dp_> Alberth, nah, important part is having heuristic :p
19:36:05 <_dp_> pretty sure starting from ship is actually better coz of cashing
19:37:42 <Alberth> yeah, but then you need to go through all depots for the heuristic, on each step
19:38:20 *** mender27 has joined #openttd
19:38:30 <mender27> hello fellow OpenTTD players :)
19:38:35 <Alberth> I would like to try JPS one day too, I played with it, but the article assumes a normal grid instead of our track-based grid
19:38:57 <mender27> I'm trying to add some files to Fedora Linux and I'm wondering where does OpenTTD normally store its music files?
19:38:58 <Alberth> mender27: not sure how many that are here, probably only a few at best :p
19:39:39 <mender27> I already tried the "gm" and "data" directories, but no go. The game doesn't recognize the openmsx.obm file when either of those :(.
19:42:31 <mender27> Alberth, yes, I wish that was the case, but I've already tried the baseset directory :).
19:43:57 <Alberth> works for me bin/baseset/orig_win.obm
19:44:21 <mender27> Alberth, however, not it seems I have omitted a key detail in the readme attached to the -sources tarball.
19:45:52 <Alberth> you can try ./openttd -d 9 >& logfile.txt
19:46:04 <mender27> Alberth, thanks for pointing me in the right direction :)
19:46:05 <Alberth> then wait until intreo screen, en quite again
19:46:30 <Alberth> that gives a log of all the stuff openttd does
19:47:12 <Alberth> for less insane amounts of output, try a lower number :p
19:48:09 <_dp_> Alberth, there should be a heuristics that doesn't need all of them ^^
19:48:32 <_dp_> Alberth, can't think of a good one as we speak though :(
19:48:54 <Alberth> maybe only the ones that are actually reachable ? :)
19:49:02 <Alberth> as in not too far away
19:49:31 <Alberth> no point in finding a depot, if the order is refused :p
19:51:27 <mender27> Alberth, I found the mistake. It turned out that within the -source directory there was a duplicate directory that actually contained the midi files. I was copying from the wrong directory to begin with :P.
19:51:54 <Alberth> ah, that explains a few things :p
19:56:25 <mender27> Alberth, obviously, now the .rpm created by me works as intended. The music files went into "gm" per readme.
19:56:46 <_dp_> Alberth, oh, I know, closest depot can be pre-computed for every tile
19:56:57 <_dp_> it's a lot of memory though
19:57:11 <mender27> Alberth, now I need to tinker with timidity++, but that's an entirely separate thing
19:57:22 <Alberth> _dp_: store direction :)
19:57:49 <Alberth> mender27: alright, nice you solved it.
19:58:08 <mender27> Alberth, thanks a lot for help. Over and out! :)
20:02:18 <_dp_> Alberth, lol, I was actually thinking about heuristics for A*, didn't notice it solves the problem by itself))
20:03:19 <Alberth> always nice, software outsmarting humans :p
20:04:19 <_dp_> Alberth, but with direction you'll need to rebuild entire cache with every new depot
20:04:49 <Alberth> and every change in water tiles
20:05:24 <Alberth> no idea how you'd do it, tbh
20:05:50 <Alberth> maybe with a floodfill algorithm flooding from all depots, or so
20:06:05 <Alberth> compute direction to nearest depot for the entire map
20:06:55 <Alberth> I was thinking too much in single origin point :)
20:07:50 <andythenorth> can this not be generalised to ‘all ship pathfinding is hokey’?
20:10:12 <_dp_> fun thing with storing direction is that ship doesn't even need to know which depot it's going to)
20:17:11 <andythenorth> if there was a DAG for all possible ship routes…
20:17:25 <andythenorth> how big would the DAG be, before recalculating it when landscape changes
20:17:37 <andythenorth> takes longer than ship pathfinding every n ticks?
20:20:39 <Alberth> there is no 1 dag, you'd get 1 for each destination
20:21:55 <Alberth> and that runs away quite fast, O**3 or worse
20:22:20 <Alberth> number of station * all tiles of the map
20:22:31 <Alberth> latter is quadratic on map size
20:22:31 <Wolf01> Fetch \a n bits from \a x, started at bit \a s. <- why are there all the \a?
20:23:19 <Alberth> gives nice italic font in the documentation :)
20:23:27 <Wolf01> Pretty unreadable on IDE
20:24:09 <Alberth> yeah, like xml is any better :p
20:25:09 <Samu> there is no YapfShipFindNearestDepot function :( will try to create it somehow, lol
20:43:10 *** umgeher has joined #openttd
21:06:03 <Samu> okay i dont know how to code this... time to give up
21:11:22 <LordAro> there does come a point where you do need to know the language you're trying to write
21:32:54 <V453000> baking 1M polygons to 8k texture
21:32:58 <V453000> let's see how this ends up XD
21:51:35 <V453000> 10/10 progress worth the soul sacrifice
21:58:00 *** sim-al2 has joined #openttd
22:01:03 <andythenorth> previous edition was quite lame
22:01:41 <andythenorth> also now haz epic cargo support
22:01:56 * Wolf01 can't see no differences :P
22:04:20 <Wolf01> Yes, they look better now
22:13:42 *** andythenorth has joined #openttd
22:54:24 <Samu> before my computer dies, i better post everything I've been working with
22:54:48 <Samu> i have a feeling my rig won't last, it came back to life after 5 days refusing to even POST
23:00:55 <Samu> i remember a patch being rejected because it was posted on 1st april i think of last year or 2014, must find it
23:04:36 <Samu> Unstuck Ship when Leaving Depot
23:25:34 <Samu> I don't think i got anything else to post
23:34:39 <debdog> sound is annoying, though
23:34:45 <Wolf01> Also I already knew the last one
23:35:12 <Wolf01> It helps with the atmosphere :P
continue to next day ⏵
