IRC logs for #openttd on OFTC at 2016-10-02
⏴ go to previous day
00:53:06 *** Smedles_ has joined #openttd
01:00:32 <Samu> was already using 21 GB memory, using the hdd too much
02:04:58 *** NekoMaster has joined #openttd
02:05:06 <NekoMaster> Evening everybody!
02:05:50 <NekoMaster> I've got a "simple" question about a certain locomotive and its recommended use
02:07:42 *** NekoMaster has left #openttd
02:43:43 *** ChanServ sets mode: +v tokai
03:18:53 *** LadyHawk- has joined #openttd
03:22:54 *** LadyHawk- is now known as LadyHawk
04:44:21 *** NekoMaster has joined #openttd
05:14:41 *** NekoMaster has left #openttd
05:16:01 *** Extrems has joined #openttd
06:22:38 *** NekoMaster has joined #openttd
06:23:08 <NekoMaster> Why do some of my locomotives in the sorting order go out of order?
06:24:41 <NekoMaster> For Example, my EMD E7 Unit has ID 2001, and my New Haven EP5 has ID 4002, but my EP5 appears above the E7-Unit
06:25:25 <NekoMaster> (also the Vehiclesort.pnml has the ep5 near the bottom and EMD E7 is near the top)
06:31:59 *** NekoMaster has left #openttd
06:49:50 *** NekoMaster has joined #openttd
06:50:01 <NekoMaster> Updated the NARS ADD-ON SET
06:55:51 *** NekoMaster has left #openttd
07:50:01 *** andythenorth has joined #openttd
09:32:57 *** sla_ro|master has joined #openttd
09:33:47 *** Alberth has joined #openttd
09:33:48 *** ChanServ sets mode: +o Alberth
09:36:41 <andythenorth> 6am UK time I woke up :P
09:43:59 <Alberth> I did that at 5am UK time, but decided to sleep a bit more :p
09:44:51 <andythenorth> Road Hog ‘Brit’ roster is a bit misleading :P
09:45:36 * andythenorth wonders if there’ll ever be other Road Hog rosters
09:47:49 <andythenorth> except…can I be arsed?
09:47:53 <V453000> idk, I am torn on this concept
09:48:03 <V453000> 1. it feels like it's nicer to make one thing properly
09:48:17 <andythenorth> it’s useful tactic to avoid bloat
09:48:17 <V453000> 2. having the variety of different climates and shit has it's own magic
09:48:23 <V453000> otherwise climate is just reskin of cargoes
09:48:39 <andythenorth> good idea = ‘put that in another roster’, not just add to existing
09:48:48 <andythenorth> but eh, otherwise, no
09:49:01 <andythenorth> for trains it’s worth it, but not RVs
09:54:33 <Alberth> nobody said you cannot make a roster by a sub-selection of your tramz :p
09:54:52 <Alberth> sort of firs-everything vs basic-climate
09:55:04 <andythenorth> considered it :)
09:55:24 <andythenorth> but probably not interesting to me, I like to have all the different types for different cargos
09:56:01 <andythenorth> the idea of rosters was to give a different set of gameplay choices
09:56:10 <andythenorth> fast versus slow and strong or so
09:56:29 <andythenorth> but it’s hard to vary vehicle stats much and still have a good game
09:57:12 <V453000> unless you are nuts? :P
09:57:24 <Alberth> hmm, how fast can RVs go?
09:57:35 <Alberth> give small & fast a new meaning :p
09:58:00 <V453000> hover bus has like 500kmh doesnt it
09:58:31 <Alberth> nice scalable multi-cargo back-bone :p
09:59:12 <andythenorth> the ultimate vehicle is probably 8t, 1/8, 500kmh
09:59:59 <V453000> 65k power, 666 capacity?
10:00:30 <andythenorth> someone draw me one more pax tram
10:03:39 <V453000> just take a random train sprite from nuts
11:18:30 *** frosch123 has joined #openttd
11:29:45 *** Progman has joined #openttd
11:30:47 <andythenorth> every 30 years or so 1870-1990
11:34:09 <Alberth> less is better, imho :)
11:34:47 <andythenorth> less than in the first screenshot?
11:34:50 <andythenorth> or just generally?
11:35:28 <Alberth> the "previous" had a lot of different tramz, even with different sizes
11:36:10 <andythenorth> ah some were mail, confusing sorry :D
11:36:49 *** Biolunar has joined #openttd
11:38:00 <andythenorth> main thing was better visual progression old -> modern :)
11:39:15 <Alberth> pax_tramz has more variation in doors, which make them look more different imho
11:39:35 <Alberth> assuming the top 5 of "previous" were pax
11:39:45 * andythenorth is only about 10,000 pixels away from next beta
11:41:06 <Alberth> dropped the black roof also makes it look better, more fresh
11:41:14 <Wolf01> What could I do today for NRT? Sprites? (it's just a copy/paste structure and function from rails)
11:41:43 <Wolf01> Or we could define the features for the 0x10 prob?
11:42:18 <Alberth> we can, if it involves sprinkles and magic! :p
11:43:19 <Wolf01> Grf should provide catenary for road (trolleybus), I'm sick to see the road icon printed over the road tiles :P
11:43:20 <Alberth> but basically depends if you want to see stuff moving, I guess
11:44:10 <andythenorth> Wolf01: find out why the new grf I made won’t load?
11:44:42 <Wolf01> I tried only the last one with cattenary bit
11:45:13 <andythenorth> Fatal: attempt to use invalid ID (sprite 80)
11:45:38 <Wolf01> I'll try it on the afternoon
11:45:55 <andythenorth> I am taking kids to park for an hour or so now
11:48:36 <Wolf01> I'll be back in 2 hours
11:51:47 <Wolf01> Also I was thinking... why have both "rail" and "elrail" menu items? With different RailTypes you end up having 20 items instead of 5: rail, elrail, third rail, dual power multiplied by slow speed, normal speed, high speed and narrow gauge.. wouldn't it be possible to change the gui (not the spec) to add a powered method compatible with that rail? For example an "add catenary" or "add third
11:51:47 <Wolf01> rail" button in the build rail roolbar?
11:53:25 <Wolf01> It should work like the convert rail tool, but the convert rail converts only between speeds
11:56:13 <Wolf01> In fact, when you add catenary or third rail, you convert it to the other railtype, but without showing to the user that it's another railtype
11:59:22 *** HerzogDeXtEr has joined #openttd
12:11:54 *** supermop_ has joined #openttd
12:17:22 <Milek7_> how to add scrollbar to window?
12:18:35 <Milek7_> i tried adding NWID_VSCROLLBAR widget but now only scrollbar is visible
12:21:06 <Alberth> please show the code in a pastebin
12:22:11 <Alberth> likely, you need a new container, as you need to define where in the window the scrollbar should be added
12:24:29 *** Wormnest has joined #openttd
12:44:48 <Alberth> iirc, WT_PANEL is also a container, so you added the scrollbar as content of the panel
12:45:30 <Alberth> end the panel container before adding the scrollbar
12:45:48 <Alberth> ie insert a "EndContainter()," line
12:46:26 <Alberth> note that ysize = 0 makes the panel very thin :)
12:49:23 <Alberth> panel can be used as just a canvas to draw on, and as a background for more widgets
12:50:59 <Milek7_> hm, but original code don't have endcontainer for panel
12:52:42 <Alberth> technically it isn't, but code closes all container when it reaches the end of the array as a safety measure
13:19:05 *** andythenorth has joined #openttd
13:31:55 *** sla_ro|master2 has joined #openttd
13:49:26 <Wolf01> andythenorth, what is different in the 2 roadtypes grf, other than the second roadtype?
14:00:28 <Wolf01> Also, isn't possible to have a stack frame windows instead of that stupid dropdown in VS2015?
14:02:03 <Wolf01> Ok, found it, its well hidden
14:02:21 <Alberth> second problem solved :)
14:05:19 <Alberth> names are not even familiar to me
14:08:48 <Wolf01> Could it be that I don't even try to load sprites for roadtypes?
14:19:33 <andythenorth> Wolf01: the name, and the label, and the catenary flag are different
14:20:27 <Wolf01> dbg: [grf] [example_roadtype_two_types.grf:4] RoadTypeReserveInfo: Road type property found 0x1B, num info 1
14:20:28 <Wolf01> dbg: [grf] [example_roadtype_two_types.grf:80] RoadTypeReserveInfo: Road type property found 0x1B, num info 1
14:20:42 <Wolf01> Also, what are "4" and "80"?
14:22:58 <Alberth> random guess: sprite number?
14:24:45 <Wolf01> As the error shows invalid sprite id 80, yes
14:26:49 <Wolf01> Ok, I have only 1 roadtype in roadtype map
14:27:42 <Wolf01> The second one didn't reserve
14:28:23 <andythenorth> not sure if the grf is valid
14:28:52 * andythenorth has to go out for a few hours
14:29:02 <andythenorth> but the grf should decompile with grfcodec ok
14:31:15 *** sla_ro|master2 has quit IRC
14:33:22 <Wolf01> Ok, seem to be the label
14:33:54 <Wolf01> Always return INVALID_ROADTYPE when trying to allocate a new roadtype because the label i "0"
14:34:37 *** sim-al2 has joined #openttd
14:38:56 <NGC3982> I think i got the jist of factorio
14:52:45 *** luaduck has joined #openttd
15:00:22 *** luaduck has joined #openttd
15:06:53 *** luaduck has joined #openttd
15:13:34 <V453000> this particular sprite appears in strange arrays
15:16:03 <Wolf01> Nice, it reminds me the prime numbers distribution map
15:28:16 <frosch123> V453000: GB(ti->x ^ ti->y, 4, 3)
15:28:56 <frosch123> there are 4 rough land sprites
15:29:10 <frosch123> the graphics are choosen depending on the position on the map, not randomly
15:29:22 <frosch123> we just need a better function at that place
15:29:26 <V453000> but the other types seem to look random
15:29:38 <V453000> the "hole" one is doing -- - - -- --- :)
15:29:40 <frosch123> the decision rough/not rough is random
15:29:52 <frosch123> but if it is rough, then above code decides the 4 versions
15:30:02 <V453000> why isn't it random as well?
15:31:01 <frosch123> wait, there are 8 versions
15:35:13 <frosch123> 182 mib of yeti ...
15:39:02 *** HerzogDeXtEr1 has joined #openttd
15:42:04 <DorpsGek> Commit by frosch :: r27657 trunk/src/clear_cmd.cpp (2016-10-02 15:41:56 +0200 )
15:42:04 <frosch123> try next nightly, whether it looks better
15:42:05 <DorpsGek> -Change: Replace another occurence of the ancient tilehash function with the newer one. This time to make rough land appear more random.
15:44:40 <frosch123> it's possibly that it looks worse :)
15:45:14 <frosch123> the newer hash function has a lower frequency, so neighboured tiles are likely to have the same variant
16:03:59 <Samu> who's a compiler expert?
16:08:06 <Samu> visualstudio code generator defaults to SSE2, which works, but i need to make a non-SSE2 build to be able to run on my old cpu Athlon XP 2700+
16:09:53 <Samu> using openttd_vs140.vcxproj
16:13:25 <Samu> forgot to mention, it crashes on the release win32 build, but not on the debug win32 build
16:13:56 <V453000> let's see frosch123 :) thanks
16:15:08 <Samu> didn't try x64 buids as the athlon xp 2700+ is a 32-bit cpu only t.t doesn't interest me, but i can test it
16:18:24 *** Extrems has joined #openttd
16:19:33 <Samu> testing on fx-8150, debug x64 with /arch:SSE - does not crash
16:20:10 <Samu> testing on fx-8150, release x64 with /arch:SSE - does not crash
16:25:50 <Samu> testing on fx-8150, release Win32 with /arch:SSE - CRASH!
16:27:50 <Samu> testing on fx-8150, debug Win32 with /arch:SSE - does not crash
16:28:06 <Samu> the build that matters crashes, just my luck :(
16:28:23 <Samu> who can help? what can i do?
16:32:27 <Samu> reaches main menu and crashes immediately
16:32:58 <Samu> visual studio points at here: this->LeaveStation();
16:33:40 <Samu> tException thrown at 0x00FD0041 in openttd.exe: 0xC0000005: Access violation executing location 0x00FD0041.
16:41:36 <Samu> this is the next statement to execute when this thread returns from the current function.
16:41:48 <Samu> how do i find out what's the current function?
16:42:30 <Samu> > openttd.exe!00fd0041() Unknown - is this the current function?
16:42:44 <Samu> source code unavailable :(
16:44:12 <Samu> 00FD0041 add byte ptr [eax],al
16:44:21 <Samu> this is where it crashes, no idea what it means
16:48:05 <Milek7_> it crashes in this->LeaveStation() or before?
16:48:49 <Samu> that is the next statement
16:49:41 <Samu> this->PlayLeaveStationSound();
16:49:52 <Samu> gonna try disable sounds, brb
16:50:00 <Milek7_> this empty virtual function
16:50:32 <Milek7_> probably was optimized out
16:51:09 <Samu> tried with no sound packs
16:52:44 <Milek7_> what is value of this?
16:53:06 <Samu> + this 0x0493b688 {tcache={cached_override=0x00000000 <NULL> cached_tilt=false user_def_data=0 '\0' ...} other_multiheaded_part=...} Vehicle * {Train}
16:53:45 <Samu> it has a huge amount of info
16:55:29 <Samu> it's a train at tile 23413
16:57:21 <Milek7_> there should be window with cpu registers in debugger
16:57:47 <Wolf01> Mmmh, InitializeRoadGui() seem to not be called, also for rail there are 2 functions with the same name where changes only the case of Gui -> GUI
17:02:15 <Samu> erm, refresh, it's two images
17:08:49 <Milek7_> maybe check compiling on mingw
17:12:09 <Milek7_> and i just compiled for win32 with /arch:SSE
17:12:30 <Samu> ming is better than visual studio?
17:20:01 <Samu> am i really downloading a 2012 version?
17:21:09 <Samu> but we're in 2016 is this guide still valid?
17:21:12 <Milek7_> newer should also work
17:21:40 <Milek7_> and also newer libraries
17:26:08 <Milek7_> if it works i can compile version with other patches
17:29:19 <Samu> windows smartscreen does not like your .exe lol
17:31:03 <Dakkus> Whoa... I am visitin my sister at her music studio... And in one of the studio rooms a friend of ours is playing seriously good music that sounds JUST like TTD music.
17:31:25 <Dakkus> Is it any use if I ask him to donate a few songs for OTTD?
17:31:40 <Dakkus> Because I think he just might. And damn does that music sound like authentic OTTD!
17:31:42 <Samu> sec, lang pack file stuff, is this based on 1.6.1?
17:32:37 <Dakkus> (Wonder if this similarity has something to do with me being in Bristol, around Chris Sawyer's home region)
17:33:56 <Samu> but dpi settings kinda suck
17:37:54 <Samu> it's a windows 10 compatibility issue
17:43:39 <Milek7_> what patches do you need?
17:43:43 <Samu> Select "Use pre-packaged repository catalogues". there is no such package
17:45:29 <Samu> ahm, none at the moment, I'm not going to use the athlon xp anytime soon, maybe in 4 months
17:45:48 <Samu> that rig is on my old house, i won't go there anytime soon
17:48:52 <Samu> i was only testing if i could build a SSE openttd
17:52:06 <Milek7_> what VS version you used?
17:52:48 <Milek7_> i also compiled this exe with it
17:54:56 <Samu> microsoft visual studio express update 3
17:55:20 <Samu> microsoft visual studio express 2015 version 14.0.25420.01 update 3
17:56:15 <Samu> 11,5 MB (12.133.376 bytes)
17:56:51 <Samu> 13,0 MB (13.670.400 bytes)
17:57:31 <Samu> gonna try building 240 ais here
17:57:45 <Samu> think that version is not 32-bit
17:59:39 <Milek7_> openttd_mc_win32.exe: PE32 executable (GUI) Intel 80386, for MS Windows
18:00:56 <Samu> building a 32-bit sse now
18:01:09 <Milek7_> 64bit with /arch:SSE?
18:01:56 <Samu> i built your 240 companies as 64-bit with default architecture, which is SSE2
18:02:10 <Samu> now i'm building it as 32-bit with SSE
18:03:05 <Samu> there is a lot of performance warnings
18:03:14 <Samu> but i don't think it matters, it compiles
18:03:34 <Milek7_> i will fix these stupid warning in next version
18:03:41 *** gelignite has joined #openttd
18:03:43 <Milek7_> it don't like (bool)something
18:03:49 <Milek7_> it want something != 0
18:04:33 <Milek7_> my exe is compiled with this patch
18:04:51 <Samu> my build 11,5 MB (12.132.864 bytes)
18:05:00 <Samu> yours 11,5 MB (12.133.376 bytes)
18:05:17 <Samu> it's always different, no matter what
18:06:35 <Samu> so it crashes on trunk, but not on 1.6.1
18:07:17 <Samu> gonna try the real trunk
18:08:33 <Milek7_> msvc linker is terribly slow
18:11:08 <Samu> there are 3 solutions where I change the architecture instructions thing to SSE
18:11:24 <Samu> openttd, settingsgen and strgen
18:12:52 *** andythenorth has joined #openttd
18:13:33 <Samu> release win32, almost there... zzz so slow
18:15:23 <Samu> Exception thrown at 0x00FD0041 in openttd.exe: 0xC0000005: Access violation executing location 0x00FD0041.
18:15:36 <Samu> well, so 1.6.1 is fine, but trunk is not
18:16:21 <Samu> gonna try removing the opntittle.dat
18:16:38 <Samu> there will be no savegame conversion
18:19:33 <Samu> well that means i suck at configuring this thing to make SSE builds
18:19:50 *** sla_ro|master has joined #openttd
18:21:20 <Samu> just needs dpi compatibility
18:23:13 <Samu> without opntittle.dat mine doesn't crash, let me try build a vehicle or so, brb
18:25:18 <Samu> [00] openttd 0x0057E371 SQVM::Call + 362 (d:\openttd\trunk\src\3rdparty\squirrel\squirrel\sqvm.cpp:1484)
18:26:32 <Samu> gonna test 1.6.1 on map generation
18:32:12 <andythenorth> bloody mail tramz, hard to provide visual variety
18:32:18 <andythenorth> got 3 generations, all v. similar
18:33:05 <Wolf01> andythenorth, I don't know why it doesn't load, it seem a problem with the label, but I'm not sure
18:33:28 <Samu> 1.6.1 doesn't crash on map generation
18:33:47 <Wolf01> Also, I'm committing some changes
18:35:10 <andythenorth> Wolf01: in a bit, I can try changing props one at a time to isolate the issue
18:35:15 <andythenorth> feeding the kids right now ;)
18:35:34 <Samu> generating a 4096x4096 with 1.6.1 sse build
18:35:43 <Wolf01> I think you should be able at least to test gui sprites and catenary
18:35:52 <Wolf01> But for sprites I'm not sure
18:36:00 <Samu> feedback Error reading register value.
18:37:02 <Samu> nevermind, this is for r27657
18:37:44 <Samu> map generated, AIs started playing
18:41:23 <Samu> but on r27657, map was generated, then shortly after... crash
18:42:00 <Samu> > openttd.exe!SQVM::Call(SQObjectPtr & closure, __int64 nparams, __int64 stackbase, SQObjectPtr & outres, unsigned __int64 raiseerror, unsigned __int64 can_suspend) Line 1484 C++
18:44:02 <Samu> Milek7_: how did you configure SSE on visual studio?
18:44:08 <Samu> maybe i'm doing it wrong
18:46:49 <Milek7_> C/C++ -> Code Generation -> Enable Enhanced Instruction Set -> /arch:SSE
18:49:39 <Samu> strange i did the same, did you make it to openttd only?
18:50:38 <Samu> configuration: all configurations
18:55:36 <andythenorth> Milan apparently
19:12:30 <Samu> i have no idea what i'm doing wrong
19:13:04 <Samu> can i report this crash, or it doesn't matter?
19:14:38 <Samu> what the hell, it didn't crash now? :(
19:16:35 <Milek7_> how to use Scrollbar?
19:17:15 <Milek7_> there is SetCount, but where i should set element heigth?
19:17:58 <Samu> oh milek, i can't manually set 240 companies when creating a multiplayer game via gui
19:18:06 <Samu> doesn't let me enter 3 digits
19:21:08 <Samu> i think my vcxproj was corrupt, i created a new one
19:21:44 <Samu> 11,6 MB (12.266.496 bytes) - mine
19:21:53 <Samu> 11,7 MB (12.268.544 bytes) - yours
19:22:17 <Wolf01> andythenorth, I'm going to try to set fire to the house, if you are able to do some tests meanwhile... ah for sprites please look at the spec in road.h, while the cursor has offset by 1 because I tried to load the current sprites you provided
19:22:20 <Samu> yours say OpenTTD norev000
19:23:23 <Milek7_> yes, but mine is from git master
19:25:43 <Milek7_> Samu: network/network_gui.cpp:1158
19:25:57 <Milek7_> ShowQueryString(STR_JUST_INT, STR_NETWORK_START_SERVER_NUMBER_OF_COMPANIES, 3, this, CS_NUMERAL, QSF_NONE);
19:25:59 <Milek7_> ShowQueryString(STR_JUST_INT, STR_NETWORK_START_SERVER_NUMBER_OF_COMPANIES, 4, this, CS_NUMERAL, QSF_NONE);
19:26:36 <Samu> i suppose git master is above git gud meme
19:49:00 <Samu> damn Wolf01 i just read your message, lool
19:56:29 <Samu> there's problems with your build
19:56:49 <Samu> some AIs are erroring saying Build 0
19:57:09 <Samu> trans need 18520 and your is 0
19:57:37 <Samu> OpenTTD ver 1.7 Rev.0 Build 0
19:57:50 * andythenorth sorted those out
19:59:50 <Samu> OpenTTD ver 1.7 Build.0 Rev.27657
20:00:07 <Samu> yeah, trans works on mine :p
20:01:43 <Wolf01> It's useless to draw trams without NRT ;)
20:01:59 <andythenorth> it’s useless to draw mining trucks without NRT :P
20:13:11 <andythenorth> Wolf01: openttd.cpp L950?
20:14:03 <Wolf01> How stupid I am from 10 to 10?
20:31:22 <Wolf01> Ok, I don't know what else to do now, I miss some.. things.. which need a revision of the code, but before doing those I would like to have at least this part working
20:38:40 <andythenorth> Wolf01: I am compiling
20:40:28 <Wolf01> DrawRoadBits() is a mess, I need to find a way to be able to define sprites on the grf without defining sprites for all the climates, also layered ones (but I need frosch help here)
20:42:20 <andythenorth> Wolf01: is it now defined to draw catenary *only* if the bit is set? [default tram tracks have no catenary]
20:48:04 <frosch123> does andy's grf have sprites?
20:51:55 <frosch123> can i see it somewhere? :p
20:59:35 <frosch123> Wolf01: do you have something analog to RailTypeMapSpriteGroup?
21:00:16 <Wolf01> There should be RoadTypeMapSpriteGroup
21:02:43 <Wolf01> dev, If I didn't change it wrongly
21:03:57 <Wolf01> All the VS projects have the same name, and I have at least 4 projects
21:04:55 <frosch123> can you load andy's grf without fatal error?
21:06:01 <frosch123> so, that's the next goal then :)
21:06:58 <Wolf01> Yes, but I don't know if is the grf or the code
21:07:50 <Wolf01> The code looks right, but can't allocate a new roadtype
21:12:43 <andythenorth> but I have an nml fork also
21:12:47 <andythenorth> with the example grf in it
21:12:56 <frosch123> i decoded it to nfo :p
21:15:23 <andythenorth> might not be legit
21:15:45 <frosch123> how do i checkout a branch in git?
21:15:52 <frosch123> or did i even clone it correctly?
21:16:15 <andythenorth> git checkout [branch name]
21:16:24 <andythenorth> or git co if you have the magic config for that
21:17:17 <frosch123> i did git checkout -B dev
21:17:25 <frosch123> i think that created a new branch "dev"
21:17:34 <frosch123> how can i check whether i even pulled it?
21:17:39 <Wolf01> I use git clone --recursive
21:18:15 <andythenorth> tells you branches
21:18:32 <andythenorth> usually you get them all if you do a straightforward git clone
21:18:33 <frosch123> so i did not clone it
21:19:25 <Alberth> you need a clone for that :p
21:19:43 <frosch123> it only gave me master
21:20:05 <Alberth> git checkout origin/dev if dev is the branch
21:20:34 <Alberth> you get in detached head state, meaing it won't save your commits
21:21:11 <Alberth> git branch dev # make local label
21:21:24 <Alberth> git checkout dev # checkout local label
21:21:40 <Alberth> latter 2 can be combined, but no idea how
21:22:42 <andythenorth> I never figured that out
21:22:47 <Alberth> git branch -a # list all branches, including remote ones
21:22:59 <frosch123> i was worried i had to reclone :p
21:23:09 <andythenorth> my most common git irritation is typing ‘git co [new-branch-that-doesn’t-exist-yet]’ :P
21:25:49 <andythenorth> Wolf01: the catenary bit is working
21:26:02 * andythenorth doing one change at a time in the grf
21:27:49 <andythenorth> Wolf01: the grf fails on sprite 4 if I change the label
21:27:59 <andythenorth> with just one roadtype defined in the grf
21:28:23 <Wolf01> So it's related to the label as I thought initially, but how?
21:28:34 <Wolf01> I copied that from rails
21:34:02 <Wolf01> frosch123, AllocateRoadType() in road_cmd.cpp:118, is there something wrong there?
21:38:33 * andythenorth wonders how to debug C++ :P
21:39:53 <andythenorth> let’s skip that for today
21:41:12 <Wolf01> Oh it's not complicated, just "put a breakpoint there and F10 / F11"
21:45:03 <frosch123> well, you only have 2 roadtypes, and they are both used by the default road and tram
21:46:00 <frosch123> let's try "ROADTYPE_END = 16"
21:49:01 <Wolf01> I didn't looked at how the railtype enum was structured
21:49:42 <Wolf01> assert_compile(lengthof(road_owner) == ROADTYPE_END);
21:50:14 <Wolf01> I think that check should be done in a different way
21:55:23 <frosch123> lots off stuff to change to increase ROADTYPE_END :)
21:55:35 <Wolf01> Uh oh... problems with roadtype owner
21:55:56 <Wolf01> Should I change the switches to loops?
21:56:08 <Wolf01> And ensure only 2 roadypes are on a tile?
21:56:23 <frosch123> hmm, i guess ROADTYPE_END was wrong then
21:56:35 <Wolf01> default: NOT_REACHED();
21:56:35 <Wolf01> case ROADTYPE_ROAD: return (Owner)GB(IsNormalRoadTile(t) ? _m[t].m1 : _me[t].m7, 0, 5);
21:56:37 <frosch123> we need ROADTYPE_END = 2, STREETTYPE_END = 16 or so
21:57:25 <Wolf01> All functions for ownership of road have roadtypes hardcoded
21:57:37 <frosch123> yeah, ownership is based on road/tram
21:57:44 <frosch123> we need STREETTYPE_END = 16
21:58:22 <Wolf01> STREETTYPE_END = 16 and TRAMTYPE_END =16 too
21:58:59 <frosch123> that woudl then requrie splitting _roadtypes into _tramtypes
21:59:48 <frosch123> sounds like chicken/egg :p
22:01:13 <frosch123> so, if we keep road/tram ids separate, the reservation of the label already needs to nkow the type
22:02:01 <frosch123> andythenorth: how does the grf define whether a type is road or tram?
22:03:22 <frosch123> maybe property 0x08: Define label for roadtype, 0x0X: define label for tramtype?
22:04:00 <frosch123> meh, i thought we could postpone the compatibility-stuff to the end :p
22:04:00 <Wolf01> Or compatibility set with base roadtypes?
22:04:05 <Milek7_> how to use scrollbars?
22:04:59 <andythenorth> frosch123: so far there is no definition of that :)
22:05:06 <andythenorth> I figured it would have to be a property
22:05:06 <frosch123> yeah, i think we need two properties to define labels
22:05:12 <andythenorth> not just a flag?
22:05:19 <frosch123> one to create a roadtype, one to create a tramtype
22:05:44 <frosch123> andythenorth: question is whether you are allowed to define a roadtype with the same label as some other tramtype
22:06:05 <frosch123> are labels unique across road/tram, or only within their type?
22:06:56 <andythenorth> unique within type?
22:07:11 <andythenorth> they are otherwise totally orthogonal w.r.t vehicles?
22:07:26 <frosch123> so you can have both a WOLF road, and a WOLF tram
22:07:40 * andythenorth envisages that if overlapping, we end up arguing about for 6 months in forums
22:08:04 <andythenorth> “all labels for tram must be ’T___’”
22:08:16 <andythenorth> “all labels for road must be ‘R____’"
22:08:22 <andythenorth> except on Tuesdays
22:08:39 <andythenorth> and then someone will use T for Trolleybus road
22:08:52 <andythenorth> “zomg sky has fallen, major newgrf author rage quits"
22:09:05 <Wolf01> For road should be RO__, to avoid confusion with RAIL
22:09:23 <andythenorth> for tram, should always be TRAM to be explicit :P
22:10:10 <Wolf01> Nope. Only 4 chars -> uint32
22:10:43 <andythenorth> see the problem?
22:13:48 <frosch123> i wonder, should there be two GRF features?
22:14:10 <frosch123> usually you identify stuff by feature+id
22:14:20 <frosch123> but not we add this road/tram as subfeature
22:14:26 <frosch123> i think road should be 12 and tram 13
22:14:31 <frosch123> two grf features :)
22:16:14 <frosch123> currently I think separate GRF features are best :p
22:16:49 <Samu> trams are trains on roads
22:17:38 <andythenorth> frosch123: +0.5 to separate grf features
22:18:12 <andythenorth> it’s quite a lot of fragmentation, same features, implemented twice
22:18:26 <frosch123> you can share the code
22:18:30 <andythenorth> unless we unify all the functions, but I dunno, I would be lost about unifying parser functions
22:18:37 <frosch123> train/road/ship/aircraft also share lots
22:18:43 <andythenorth> might be andythenorth lack of programming ability :P
22:18:58 <Samu> there are trucks separate from buses
22:19:55 <Samu> why do you separate trams from roads?
22:20:03 <Samu> i mean, what's the idea behind
22:21:48 <andythenorth> frosch123: cleaner to just split them to different features, surely?
22:21:54 * andythenorth is now convinced
22:22:10 <andythenorth> opens the way to splitting ROAD_VEH in future maybe also
22:22:28 <andythenorth> although that might not be beneficial
22:22:36 <andythenorth> eh it ruins my current templating :P
22:22:59 <Samu> while you're at it, separate intercontinental from large airports
22:23:23 <Samu> along with some aircraft vehicles that look too big to be on a large aiport
22:23:32 <Samu> but not on intercontinental
22:23:38 <frosch123> considering all the compatibiilty properites, which list labels of other things
22:23:55 <frosch123> it makes all the difference whether you can list labels of other things :)
22:25:32 <andythenorth> seems cleaner to me to enforce the separation
22:26:54 <frosch123> same as 8, but road/tram swapped
22:27:09 <Wolf01> But props start from 0x08, is it safe?
22:27:30 <frosch123> we just decided for option 3 :p
22:27:36 <frosch123> so, we do not need 7
22:29:59 <Wolf01> So, id 12 = road and 13 = tram
22:30:18 <frosch123> but i think adding ROADTYPE_ROAD/TRAM to GetRoadTypeByLabel and AllocateRoadType is stilll good
22:30:29 <Wolf01> I could put some conditions on the RoadTypeReserveInfo() to avoid entire duplication
22:30:52 <frosch123> essentially we get RoadtypeInfo _roadtypes[ROADTYPE_END][STREETTYPE_END]
22:34:37 <andythenorth> can we do tram later :P
22:34:48 * andythenorth doesn’t want to extend spaghetti into nmlc tonight :)
22:35:39 <frosch123> still needs finding all the places which need to use StreetType instead of RoadType
22:36:40 *** sim-al2 has joined #openttd
22:40:28 <Wolf01> Ownership could be left as is, it doesn't need to explode the exact type
22:40:54 <frosch123> it even must stay as is :p
22:41:05 <frosch123> you have no more map bits
22:41:31 <Wolf01> You have them, but Milek7_ could not do anymore 240 players :>
22:45:15 <frosch123> i thought you wanted to store 16 road and 16 tram owners per tile
22:46:26 <Wolf01> That would be cool too, but I would like more different types per tile
22:47:55 <andythenorth> elevated tramway? o_O
22:48:33 <Wolf01> If you want to rewrite the state machine to allow elevated vehicles
22:48:51 <andythenorth> just offset them
22:48:58 <andythenorth> it’s a third roadtype :P
22:49:09 <Wolf01> It won't behave well in curves
22:50:09 <andythenorth> or when crossing catenary
22:52:18 *** andythenorth has left #openttd
22:52:42 <Wolf01> Do frosch wants to continue or bed too?
22:53:32 <Wolf01> Ok, so I could start anime night and go to see it on bed too :P
23:06:15 <Eddi|zuHause> <andythenorth> elevated tramway? o_O <- idea is not that bad, but needs transition tiles between elevated and normal tram (and tram to subway)
23:06:57 <Eddi|zuHause> also, it's a bit tricky with the amount of road types on the tile
23:07:26 <Eddi|zuHause> so you could not have an elevated tram pass over a normal tram with road
23:07:45 <Eddi|zuHause> a true new tile layer might be a better idea
23:09:29 <Wolf01> I have some ideas to have locomotion style rails too, which could be used even for roads
23:12:24 <Wolf01> You need to store start and end point of the segment, and the railtype define which segments you can place
23:12:41 <Eddi|zuHause> well, the idea i had was that you place bridge segments individually, where you can differentiate between single tile bridge segments with low height limit, or larger bridge segments with height limits only where the pillars stand
23:13:13 <Eddi|zuHause> so you have more flexibility if you're only one or two levels above the ground
23:13:30 <Eddi|zuHause> but to span larger valleys you need suspension bridges which can only be straight
23:19:47 <Eddi|zuHause> so a suspension bridge would consist of a starting segment from the bridge head to the pillar, a middle segment from one pillar to the next pillar (possible to repeat this segment), and the reverse of the start segment
23:20:10 <Eddi|zuHause> the segments themselves are fixed (maybe length is configurable like stations)
23:20:17 <Eddi|zuHause> but you can combine the segments as you wish
23:21:43 <Wolf01> Don't forget signals in case of rail
23:22:04 <Eddi|zuHause> the two main questions with this is: how do you store this in the map, and how do you provide the newgrf interface
23:22:47 <Eddi|zuHause> side questions are signals, crossing trackbits between parallel segments, stations, game interface
23:23:02 <Wolf01> Eh, add another array to map
23:23:25 <Eddi|zuHause> one array doesn't suffice
23:23:56 <Wolf01> And you'll get 400MB savegames
23:24:25 <Eddi|zuHause> an idea was that you make flexibly allocated "supertiles" of like 16x16 area, which stores all trackbits of a certain heightlevel
23:25:38 <Eddi|zuHause> you want adjacent trackbits close by in memory, to aid caching effects when doing pathfinding and stuff
23:25:57 <Eddi|zuHause> so you can't just randomly allocate each bridge segment individually
23:26:32 <Eddi|zuHause> but you also can't allocate a whole map's worth array for every heightlevel, most of which will be air anyway
23:26:57 <Wolf01> Save every track type for every player as bitmap
23:27:20 <Eddi|zuHause> how is a bitmap different from an array?
23:33:44 *** mindlesstux has joined #openttd
continue to next day ⏵
