IRC logs for #openttd on OFTC at 2017-03-26
⏴ go to previous day
00:14:36 <nekomaster> now i have to find a bunch of trucks...
00:15:02 <nekomaster> and to make things interesting... perhaps a different truck for each kind of truck for each generation
00:15:16 <nekomaster> So the 1950's trucks will all look different
00:15:24 <nekomaster> variety is the spice of life
00:15:37 <nekomaster> I'll just google things like 1950's tanker trucks
00:16:03 <nekomaster> and see what pops up for companies like Kenworth, Ford, Petterbuilt, etc
00:19:49 <DorpsGek> Commit by peter1138 :: r27829 trunk/src/gfx_func.h (2017-03-26 00:19:41 +0100 )
00:19:50 <DorpsGek> -Fix (r27821): Occasional crash caused by mixing int/uint arithmetic.
00:26:21 *** HerzogDeXtEr has joined #openttd
00:33:18 <supermop_home> no good fonts for new resume
00:33:54 <Wolf01> Mi resume "please hire me" in Arial 48, bold
00:34:14 <Wolf01> I'm really tired this night
00:37:57 *** paooolino has joined #openttd
00:39:17 <paooolino> anyone successfully compiled openttd using MS VS2017?
00:41:05 <Wolf01> Not yet, I'm scared of upgrading my VS2015
00:44:01 <paooolino> i'll try to see if I can install VS2015 then
00:56:51 <Wolf01> Bed is calling... must resist
00:56:53 *** Hiddenfunstuff has quit IRC
00:57:19 <Eddi|zuHause> don't forget to switch the time
00:57:39 <Wolf01> Oh, devices do it automagically
00:57:56 <Eddi|zuHause> some do some don't
00:58:27 <Wolf01> Like my alarm clock and car's clock
00:58:49 <Eddi|zuHause> ... or the local church's clock :p
00:58:53 <Wolf01> But I don't use the alarm clock and car clock is already offset to an unknown time
00:59:40 <Eddi|zuHause> and the worst part, on monday i have to be at a meeting earlier than usual
00:59:51 <Wolf01> But I won't be the one in charge to change the time on church's clock... we don't even have a church's clock
01:00:05 <Eddi|zuHause> and "usual" is also already earlier than i'd like
01:00:25 <Eddi|zuHause> so it's now triple early
01:20:33 *** Stimrol has joined #openttd
03:18:43 <Samu> take a look and tell me how bad it is
03:38:51 *** supermop_home has joined #openttd
04:16:56 *** gelignite_ has joined #openttd
05:53:46 *** nekomaster has joined #openttd
05:54:04 <nekomaster> I can't believe I actually went and passed out
07:02:41 *** Biolunar has joined #openttd
07:55:47 *** andythenorth has joined #openttd
08:37:32 *** Hiddenfunstuff has joined #openttd
08:40:26 *** Alberth has joined #openttd
08:40:26 *** ChanServ sets mode: +o Alberth
08:53:21 *** Progman has joined #openttd
09:05:23 <peter1138> Hmm, is there a reason we still need 3 pathfinders?
09:06:16 <Sacro> OCPF, one combining path finder
09:06:34 <Sacro> (only works in darkenness)
09:07:16 <Alberth> only for trade-off between performance and perfection, probably
09:07:46 <andythenorth> don’t we just need YAPF and the original?
09:08:11 <Alberth> something like that could work :)
09:08:59 <peter1138> Original is only there for ships.
09:09:04 <Alberth> although, I do want to implement JPS one day, at least for ships
09:09:26 <peter1138> Does NPF ever perform better than YAPF?
09:09:35 <andythenorth> maybe it lets RVs find depots
09:09:47 <andythenorth> with YAPF, they can’t
09:09:55 <andythenorth> but turning breakdowns off fixes that
09:10:19 <Alberth> I don't know how NPF and YAPF compare
09:10:53 <peter1138> andythenorth, eh? they do for me
09:11:28 * andythenorth never posts FS reports of bugs
09:11:42 <andythenorth> because I always have a bunch of newgrfs that can’t be had from bananas
09:11:53 <andythenorth> and I’ve blatantly been reloading them over and over again
09:18:01 <Eddi|zuHause> peter1138: i have in the back of my head that NPF is usually like 10 times slower than YAPF
09:18:20 <Eddi|zuHause> because of all the virtualisation overhead
09:19:10 <Eddi|zuHause> peter1138: i guess it's kept because there is no real reason to remove it
09:22:04 <nekomaster> wouldn't removing NPF streamline some code though?
09:22:49 <andythenorth> it would increase maintainability
09:23:17 <Eddi|zuHause> virtual functions, i mean
09:23:31 <peter1138> NPF doesn't have any
09:24:22 <Eddi|zuHause> afair it reimplements those, because it's written in C, not C++
09:24:54 <Eddi|zuHause> but i never actually looked at it
09:25:34 <Eddi|zuHause> YAPF was the first true C++ part of the code
09:25:47 <Alberth> nekomaster: hardly, there is a common interface to talk with the path-finder for all the other parts, so basically you throw away one case, reducing eg pathfinder settings from 3 cases to 2
09:26:08 <Alberth> true C++ *cough* *cough*
09:26:12 <Eddi|zuHause> and before the entire project was moved to C++, the C part reimplemented lots of object-oriented patterns
09:26:33 <Alberth> I killed some of those :p
09:26:47 <Alberth> ie entire newgrf code :p
09:27:22 <Alberth> ah, is death at current processors, isn't it?
09:27:36 <Eddi|zuHause> anyway, that can't have been easy for optimisation purposes
09:28:41 <Alberth> would be interesting to find out whether all the template stuff is actually improving performance
09:29:05 <Eddi|zuHause> Alberth: even 10 years ago i was taught that you better implement lists as continuous array chunks, because of caching
09:29:28 <Alberth> :O nobody ever told me :p
09:29:44 <Alberth> I am too old, probably :p
09:31:02 <nekomaster> Would it ever be possible to have proper directional single tracks in openTTD?
09:31:37 <Alberth> how don't we have those now?
09:31:48 <Eddi|zuHause> Alberth: the point of all the templating was that you have all the flexibility of NPF available, but resolved at compile-time rather than run-time
09:32:41 <Eddi|zuHause> which resulted in this factor 10-ish improvement
09:33:04 <Eddi|zuHause> back then, NPF was kept for comparison
09:33:27 <Alberth> but at the same time, you make c++ optimization impossible due to all the convoluted constructs
09:34:30 <Eddi|zuHause> what do you mean?
09:34:50 <Alberth> you can up-cast to a derived class with a template, but I don't believe a c++ compiler can't recognize that case by itself
09:35:38 <Alberth> all optimization in a compiler is based on common code patterns, wrapping it all in templates breaks having common patterns, and thus optimization opportunity recognition
09:36:36 <Eddi|zuHause> i'm not sure what you mean... all the template stuff should be resolved by the time optimization happens
09:37:21 <Alberth> if you have a virtual derived method, the compiler recongizes that, and eliminates it
09:37:49 <Alberth> diy at template-unfolding time means the compiler never recgonizes such a thing
09:38:14 <Alberth> ie the templates try to outsmart all compiler optimizations from the moment it was add until eternity
09:38:32 <Alberth> ie sort-of hard-coding the 'optimal' result
09:38:48 <Alberth> except the optimal result is unlikely to be optimal at any point in time
09:39:27 <Eddi|zuHause> well, you could re-implement NPF with actual classes and stuff, and compare performance with YAPF again
09:39:51 <andythenorth> if we delete it, does it get faster or slower? o_O
09:39:59 <Alberth> like I said, it would be an interesting experiment to see if the template stuff is actually working
09:40:02 <Eddi|zuHause> probably neither
09:40:32 <Alberth> compiling the code gets a bit faster, but hardly noticable :p
09:40:55 <Alberth> NPF is only a few small files
09:40:56 <andythenorth> is a non-existent NPF infinitely fast or infinitely slow?
09:40:57 <peter1138> i was more thinking about maintainability
09:41:21 <peter1138> if you need a pf change right now, you need to do 2 or 3 times
09:41:46 <Alberth> would it also kill some old src/core data containers?
09:42:37 <peter1138> doubt it, npf is pre all thawt
09:42:47 <Eddi|zuHause> i think if we removed NPF, all the infrastructure for having multiple pathfinders will decay, and if we ever want to try a new pathfinder again, it would have to be redone
09:42:54 <peter1138> urgh, zbase never got fixed :(
09:42:56 <Alberth> we established NPF is slower?
09:43:14 <Alberth> I have no problem throwing the slowest one out
09:43:19 <peter1138> it probably is, as that was largely the point of yapf
09:44:02 <Eddi|zuHause> it's easy to test... pick a huge coop game, and switch pathfinders
09:44:24 <Alberth> you need some ship tests too, imho
09:44:52 <Eddi|zuHause> there might be sublte behaviour differences because all the penalties are different, which might block a coop network
09:44:59 <Eddi|zuHause> there is no NPF for ships
09:45:15 <Alberth> ok, that's a fast test then :p
09:45:52 <peter1138> there certainly is NPF for ships
09:46:52 <andythenorth> it’s in the settings
09:47:02 <andythenorth> so either the settings lie, or...
09:47:29 <Eddi|zuHause> weird, i never noticed it there
09:47:38 <peter1138> git blame says 2005
09:48:06 <peter1138> ships also never had the original pathfinder removed
09:48:09 <Eddi|zuHause> well, ok, then there is NPF for ships
09:48:49 <Eddi|zuHause> yes, because both NPF and YAPF have problem with large fully-connected areas
09:49:29 <Eddi|zuHause> the problem with original ship pathfinder is that it completely fails with rivers
09:49:49 <Eddi|zuHause> and even in large open waters it may fail to recognize diagonal routes
09:50:06 <peter1138> doesn't it also have an element of chance?
09:50:20 <peter1138> with yapf ships always take the same route
09:50:24 <peter1138> with opf, they don't
09:56:08 <andythenorth> 15 slopes yak-shaved, 3 more to do \o/
09:56:57 <peter1138> hmm, in the arabic translation, some texts are missing their colour codes
09:59:28 <nekomaster> i honestly dont know how some people can get motivated to draw so many sprites
10:00:31 <peter1138> hmm 5°C, that's a bit cold to get the legs out
10:01:28 <andythenorth> nekomaster: mental strength :P
10:01:46 <andythenorth> if you look at RH, it’s mostly copy-paste
10:01:51 <nekomaster> I'm not fond of tedious work
10:03:55 <nekomaster> coding isn't so bad because I can just recycle stuff and change a few settings
10:06:16 <andythenorth> the rest is copy-paste
10:08:32 <andythenorth> if only I could automate it :P
10:08:58 <Alberth> get a few infinite amount of monkeys :p
10:09:22 <andythenorth> you mean evolutionary procedural? o_O
10:09:25 <andythenorth> then pick the best ones?
10:10:10 <nekomaster> Well right now I'm planning on just reusing one model of truck for each class for all cargos
10:10:42 <Alberth> :O generate random combinations, and ask which to iterate in the next generation, hmm, tediously clicking I am afraid :p
10:10:52 <andythenorth> Alberth: yeah, that’s why I didn’t :P
10:11:42 <Alberth> if you have good constraints on what "good" is, you could dramatically reduce the number of vehicles to select from
10:12:06 <Alberth> but it's non-trivial, to say the least :p
10:13:02 <nekomaster> does anyone have any idea when Semi-Trailer Truck Doubles started becoming a thing?
10:14:38 <andythenorth> nekomaster: IRL, 1930s-1950s
10:14:55 <andythenorth> well that’s when I found a picture of one in California
10:15:13 <andythenorth> but there is a difference betwen ‘something unusual that is worth a picture on the internet’
10:15:17 <nekomaster> when I think of Doubles I think of your typical Semi-Tractor
10:15:19 <andythenorth> and 'actual history’
10:15:51 <nekomaster> I imagine that doubles as we know them today probably started becoming a big thing after the success of big powerful Semi-tractors in the 60's and 70's
10:18:42 <andythenorth> nekomaster: you could spend a lot of time in Hank’s Truck Pictures ;)
10:20:14 <nekomaster> looks like Western Express started doing doubles in 1958
10:22:11 <nekomaster> At the moment Articulated Semi-Trucks don't start appearing on my lists until 1975, after they lifted the limit on length of semi-trucks
10:22:22 <nekomaster> so by 1975 long nose trucks started becoming the norm
10:24:31 <andythenorth> my advice: just use IRL to get rough ideas, then make it fit the game
10:25:03 <andythenorth> if it helps, north america is a big continent, and the rules etc for trucks in Florida can be very different to those in BC
10:25:05 <nekomaster> Well in 1975, the first Semi-turck appears
10:25:15 <andythenorth> take what you need to make a good roster
10:25:27 <nekomaster> and I think in 1980 the first doubles will appear
10:25:48 <nekomaster> I'll just take some other Semi-tractors from each generation date, and make them fit
10:26:08 <andythenorth> doubles are a massive jump in capacity, right?
10:26:35 <andythenorth> maybe you should limit them to a turnpike NRT :P
10:26:52 <nekomaster> Well some area's do allow doubles
10:27:18 <nekomaster> I know they apepar here in Oshawa from time to time on major roads but they're usually coming off the highways and going to General motors or other industries in Oshawa
10:27:42 <nekomaster> Like trucks going to and from the Concrete plant here
10:29:56 <nekomaster> I think I might go with a 1980 GMC General for the first Double Semi-Tractor
10:31:59 <nekomaster> I think I'll model the first Double semi-tractor after the Smokey and the Bandit 1980 GMC General
10:34:06 <nekomaster> Also I like how its got a big sleeper cab bolted on the back
11:03:46 <andythenorth> 28 slopes done :P 12 left to do
11:06:37 <andythenorth> I get the names of forests, lakes, rivers, quarries and mines btw
11:07:11 <nekomaster> YOu talking about how you name Road Hog trucks?
11:07:23 <nekomaster> that might work out for North american stuff too
11:08:28 <nekomaster> 1975 Petrolia [SEMI TANKER]
11:08:55 <nekomaster> 2025 Kedron Triple
11:13:21 <nekomaster> so now that I have some stuff laid out, I think I should get to work on the sprites
11:23:09 *** ChanServ sets mode: +v tokai
11:32:21 *** FLHerne has joined #openttd
11:34:50 <nekomaster> time for for the painful tedious work of sprit work
11:36:08 <Samu> hmm when the ship is going against the pier and wants to find a ship depot, it fails at finding it, it should not fail
11:36:26 <Samu> pathfinder doesn't try reverse finding
11:44:49 <Alberth> ships don't turn around
11:46:26 <Alberth> likely, it should explore other feasible directions
11:48:46 <andythenorth> ‘paste from viewport’ is nuts :)
11:48:51 * andythenorth wonders who uses it
11:49:44 <Eddi|zuHause> i don't think the feature is that crazy... but i don't really use viewports
11:50:00 <Eddi|zuHause> it's possibly a bit misnamed
11:50:29 <andythenorth> I could make a section of map with ready-made stations on
11:50:31 <andythenorth> and just paste them in
11:52:40 <Alberth> it just jumps to the VP location, afaik
11:53:07 <Samu> when the ship leaves depot, it goes the wrong way now, it didn't use to be like this, or at least I don't recall it doing like this
11:55:19 * andythenorth thought paste from viewport would paste infra
11:55:31 <Samu> okay, the problem might be caused by me after all, gonna try fix it
11:56:05 *** gelignite has joined #openttd
12:01:37 <Samu> bool bDest = IsShipDepotTile(n.m_segment_last_tile) && GetShipDepotPart(n.m_segment_last_tile) == DEPOT_PART_NORTH;
12:01:53 <Samu> ships only stop at the northest tile of a ship depot
12:02:02 <Samu> i was letting them stop at either
12:03:07 *** Wormnest has joined #openttd
12:03:26 *** ChanServ sets mode: +v Alberth
12:03:26 *** ChanServ sets mode: +v orudge`
12:05:50 <Alberth> andy, that would require the copy/paste patch :p
12:08:28 <Samu> ships have to reverse at times
12:08:58 <Samu> but the pathfinder says path not found, hmm
12:09:43 <Samu> once it reverses, path is found
12:10:18 <Samu> how do i make the pathfinder aware that it will eventually reverse the ship'
12:13:27 <Wolf01> Is it me or webster didn't update the clock?
12:17:42 <Wolf01> Samu, if you make the pathfinder ignore the ship direction? When stationally most ships can turn around in place, so once the pathfinder finds a path, the ship should automatically rotate
12:17:49 <Eddi|zuHause> Wolf01: is it maybe set to UTC?
12:18:32 <Samu> order 1 was skipped because ship is "lost", pathfinder doesn't find a path, but the ship would reverse anyway
12:18:43 <Samu> and then it would find the path
12:19:07 <Wolf01> Ships should avoid 90° turns only when moving
12:19:32 <Wolf01> But I don't see why they can't turn 180° when stationary
12:19:50 <Wolf01> Like trains at stations
12:21:26 <Alberth> imho, only as a last resort
12:23:39 <Wolf01> <Eddi|zuHause> Wolf01: is it maybe set to UTC? <- yeah, time zones and DST always ruin my biorythm... I don't see why not to use UTC everywhere, you can have high noon even at 20:30, even now clocks aren't in sync with the sun anyway
12:23:58 <Eddi|zuHause> i've seen ships get hopelessly lost after docking in a river, because they couldn't turn around
12:25:02 <Eddi|zuHause> Wolf01: well, the general idea of DST is that the clock should move with sunrise (but changing the clock every day to adjust would be unwieldy)
12:25:27 <Eddi|zuHause> it's rather unimportant for your biorhythm when noon is
12:25:41 <Wolf01> Lets use roosters then
12:37:27 *** frosch123 has joined #openttd
12:40:26 *** nekomaster has left #openttd
12:52:46 <Samu> what are tile hases for? I see this many times, but still not sure what they're doing
13:01:07 <Alberth> quick finding of a tile
13:02:10 <Alberth> if you have a million tiles randomly picked, and you need to know if a given tile is in that set, you don't want to check each of the million tiles
13:03:13 <Alberth> so you compute a mostly unique value for each tile, and save each of the million tiles according to its unique value
13:03:51 <Alberth> for a given tile that you need to find, compute also the unique value, and only checks the tiles stored on that value
13:05:06 <Alberth> for a good hash, the latter is a few tiles at most
13:05:29 <Alberth> so you save checking of about 999,998 other tiles or so
13:07:42 <Samu> typedef CNodeList_HashTableT<CYapfShipNodeExitDir , 10, 12> CShipNodeListExitDir;
13:28:30 *** FLHerne has joined #openttd
13:39:36 <Samu> m_max_search_nodes = 10000 is quite a low value
13:40:02 <Samu> can't detect a ship depot a few tiles away on open sea
13:41:12 <Samu> wish i could visualize what tiles it has worked on during search
13:41:39 <Samu> how further 10000 can reach
14:07:36 <Wolf01> Try caching all ship depots, order them by nearest and then try to pathfind towards the nearest one, if unreachable then try the next one
14:08:49 <Samu> it's a hard setting that value m_max_search_nodes = 10000
14:09:37 <Samu> user defined, actually, in config
14:10:27 <Samu> when invoking pathfinder, it would still only search up to 10000
14:11:43 <Samu> if (!QueryNewTileTrackStatus()) return TryReverse();
14:11:53 <Samu> querynewtiletrackstatus returns false
14:12:07 <Samu> tryreverse, wondering what happens
14:13:11 <Samu> i see, TryReverse only reverses for non-tram road vehicles
14:13:29 <Samu> i think i know where i got to fix
14:18:18 <Samu> ships might be doing reversings unexpectedly, i wonder
14:18:29 <Samu> nothing like testing and see what happens
14:21:22 <Samu> yay, can't believe i'm doing well
14:22:43 <Samu> this has also solved the ship is lost message, it's no longer lost
14:24:43 <Samu> must try some 90 degree forbidden stuff
14:27:23 <Samu> ah i see, the ship is too smart now
14:27:47 <Samu> this fix is creating unexpected reversals
14:28:09 <Samu> Alberth: how do i make a OnlyReverseAsLastResort
14:32:26 <Samu> it should go straight because there's still water ahead
14:56:56 *** orudge` has joined #openttd
14:56:56 *** ChanServ sets mode: +o orudge`
15:04:47 *** a_sad_dude has joined #openttd
15:05:35 <Wolf01> a Web page is now like
15:05:35 <Wolf01> "V’Ger from the first “Star Trek” movie, a piece of technology
15:05:35 <Wolf01> that we once understood but can no longer fathom" true... that's really true
15:06:37 <Wolf01> WTF copying from PDF is always so shitty?
15:26:07 *** Progman has joined #openttd
15:26:21 *** sim-al2 has joined #openttd
15:31:44 <Eddi|zuHause> PDF is only marginally better than CSS :p
15:37:36 *** tokai|noir has joined #openttd
15:37:36 *** ChanServ sets mode: +v tokai|noir
15:40:24 *** supermop_home has joined #openttd
16:06:48 <supermop_home> dont have univers on here anymore
16:07:02 <supermop_home> also new ID seems not to have spell check
16:10:40 *** HerzogDeXtEr has joined #openttd
16:14:21 <Samu> i added that to follow track
16:15:42 <Samu> ship is turning left to reverse
16:15:53 <Samu> instead, i want it to follow the water
16:17:01 <Samu> i only want it to reverse if there really is no other alternative
16:22:04 <Samu> it needs to know if there are other exits
16:22:13 <Samu> if there is, then don't reverse
16:22:19 <Wolf01> <Eddi|zuHause> PDF is only marginally better than CSS :p <- indeed, in PDF is the whole structure which is a pita
16:22:45 <Samu> use the node which has an exit
16:23:06 <Wolf01> "ship too smart now" XD
16:24:27 <Alberth> try the reverse exit last
16:24:49 <Alberth> or give it a high penalty
16:26:32 <Samu> reverse exit last? gonna see if such thing exists
16:30:06 <Samu> CYapfOriginTileTwoWayT ?
16:33:41 <peter1138> Yeah, big penalty for reverse
16:46:33 <Samu> gotta see how's it done for rails, rails have it working best
16:55:47 <Samu> bool ret = pfnFindNearestDepotTwoWay(v, origin.tile, origin.trackdir, last_tile, td_rev, max_penalty, YAPF_INFINITE_PENALTY, &fdd.tile, &fdd.reverse);
16:58:36 <Samu> nice way to solve things
17:05:45 <Samu> pfnFindNearestDepotTwoWay isn't what I want
17:08:16 *** andythenorth has joined #openttd
17:18:07 *** smoke_fumus has joined #openttd
17:25:54 <Wolf01> Last time I've laughed so much it was when I read the same rants about Delphi and it's weird conversion of values to boolean
17:26:53 <andythenorth> there’s a guy with a list of poor PHP design choices somewhere
17:27:06 <Wolf01> Yeah, I bookmarked that
17:28:04 <peter1138> lol @ song lyrics : bedding taylor swift, every night in the oculus rift
17:37:40 *** marrenarre has joined #openttd
17:38:57 <marrenarre> If I install OpenTTD via the system package manager on Debian, is it bad to install online content via OpenTTD’s package manager?
17:39:34 <marrenarre> Especially since e.g. opengfx is installed by APT as a dependency.
17:40:03 <frosch123> there is only a bare minimum in the package manager
17:40:37 <supermop_home> hmm need to remember how to put stuff on my website
17:41:26 <marrenarre> frosch123: Okay so no worries about any weird mixup? I’ll just go ahead and install stuff?
17:42:51 <frosch123> the content download is only supposed to offer stuff that is compatible
17:42:54 <Alberth> since online content is done as normal user, it'll end in your $HOME, not in /usr/somewhere, where the packages stuff is
17:53:04 <peter1138> marrenarre, it will not conflict :) openttd will not alter the system installed content, and will use the newest content when possible.
17:54:43 <marrenarre> frosch123, peter1138: Okay, thank you!
17:55:32 *** marrenarre has left #openttd
17:55:55 *** orudge` has joined #openttd
17:55:55 *** ChanServ sets mode: +o orudge`
18:02:34 <Snail> peter1138: any chance we could raise the number of RailTypes in a game to 32?
18:02:47 <peter1138> I had a patch for that...
18:03:10 <Snail> nice… where can I find it?
18:04:14 *** sla_ro|master has joined #openttd
18:04:18 <Rubidium> it's from this millenium, so it's not that old
18:04:20 <peter1138> I think there's a more recent one someone else did on the forums.
18:04:33 <peter1138> Rubidium, they all are! decade, on the other hand...
18:04:40 <Snail> it would be nice itf it were put in trunk
18:05:13 <peter1138> Well it was a bit nasty, and might break stuff like NRT...
18:05:34 <Rubidium> ... make a folder with ancient patches in ^/trunk ...
18:05:57 <Rubidium> ... and then not provide or update them. We did put the patch in trunk, right?
18:06:54 <peter1138> ahhh the old opengl patch
18:07:04 <Rubidium> π <- there, all of my patches
18:07:25 <peter1138> i mean... what even is a pocket pc these days?
18:09:12 <Rubidium> given that pockets are growing in size as of late... I guess they're getting bigger
18:09:19 <Snail> how many road types do we have with NRT?
18:11:32 <frosch123> Snail: 16 road + 16 tram
18:11:45 <frosch123> so, essentially, 16 items in every dropdown
18:12:12 <Snail> right… but considering that rails have additional levels of complexity (gauges, electrification types)
18:12:14 <frosch123> Snail: why do you need more railtypes? do you want to vary them like new objects? or do you want to provide combinations of mixed properties?
18:12:16 <Snail> we could have 32 for them
18:12:42 <Snail> I need to provide combinations of rail types (light, heavy, highspeed…) with electrification systems
18:12:52 <frosch123> 18:12] <Snail> right… but considering that rails have additional levels of complexity (gauges, electrification types) <- that does not sound as if you have taken a look at the existing nrt grfs :p
18:13:13 <frosch123> Snail: consider adding grf parameters
18:13:27 <peter1138> consider gameplay vs realism
18:13:44 <frosch123> a dropdown with "a" "b", "a+b", "c", "a+c", "b+c", "a+b+c" just makes no sense
18:13:52 <Snail> even with parameters, I’d need more than 16 in a game… especially if the game spans from 1840 to today
18:15:11 <Rubidium> oh yes... we definitely need 4 feet 8.5 inches and 1435 mm gauges ;)
18:15:22 <Snail> frosch123: it does make sense. A player would need to choose between “catenary”, “third rail”, “third rail + catenary” across different track types (2 gauges, 3 axle weights, different speeds)
18:15:43 <Snail> Rubidium: no, but we can’t put 1435 and 1000mm on the same track :p
18:15:46 <frosch123> Snail: no, it makes no sense, because there is no train grf which would use all those combinations
18:15:57 <Snail> frosch123: there will be
18:16:48 <Snail> frosch123: it’s like the chicken and the egg. There are no train grf’s like that, because the rail types don’t allow. It’s not a valid argument to not expand rail types
18:17:17 <frosch123> it's a first world problem, you cannot have everything
18:17:20 <Alberth> having 32 different train types that cannot use each other rail, is nuts
18:17:38 <Snail> Alberth: there will of course be compatibilities
18:18:03 <Alberth> why have different railtypes if they can both use 1?
18:18:03 <peter1138> does NUTS have rail types?
18:18:06 <Snail> trains with both 3rd rail and catenary would be able to use “3rd rail”, “catenary” and “3rd-rail + catenary” types
18:19:13 <Alberth> Snail: people will simply build the most common railtype
18:19:57 <Rubidium> super high speed, super heavy 3rd rail catenary... aka "universal rail"
18:20:00 <peter1138> and complain that autoreplace doesn't work
18:20:15 <Snail> Alberth: but they would evolve with time, and with their needs
18:20:44 <Alberth> 200 year game play, 32 railtypes, 6 years for a new railtype?
18:20:48 <Snail> say, if the best vehicle for climbing a mountain is a 3rd-rail engine, they would build that track, or a “3rd rail + catenary” track to allow for more compatibilities
18:21:35 <Snail> Alberth: not necessarily, because different railtypes would be available at the same time. Users would choose the best depending on cost and their capabilities
18:22:55 <Alberth> you'll end up with 3/4 of the track types unused, so why have them in the first place?
18:23:20 <Alberth> condense choices, and give player a reasonable amount of useful alternatives
18:24:06 <Snail> this is a way to avoid the problem rather than dealing with it
18:24:43 <Snail> I’d need to force my players to buy a heavier track if they need 3rd rail… even if they only use light vehicles
18:25:22 <Alberth> most players don't care about such details, imho
18:25:39 <Alberth> unless you have trains as a hobby or prefosseion
18:25:51 <Snail> they’d have to pay more for a track that’s heavier and has a higher topspeed than they actually need… because we don’t want to increase the number of rail types
18:27:14 <Alberth> any number is insufficient, by definition
18:27:33 <Snail> so why not going back to 2 :p
18:27:58 <Alberth> I could even play with 1
18:28:06 <peter1138> well we had 3 originally...
18:28:24 <Snail> right, so we play in a different way. My point is, I don’t want to force anyone to use lots of railtypes if they don’t want to
18:28:44 <Snail> I’m just suggesting the possibility of using more than 16 for those who need to/want to
18:31:02 <Alberth> reality is broken, don't need to replicate it in openttd, imho
18:31:15 *** ChanServ sets mode: +v orudge`
18:31:21 <Rubidium> but if you want to go so far into the nitty gritty of things, then railtypes are basically bad. You need something to configure the perfect railtype for your particular situation
18:31:27 <peter1138> +return (RailType)(GB(_m[t].m3, 0, 4) | (GB(_m[t].m1, 7, 1) << 4));
18:31:34 <peter1138> ^ that was the nasty bit about the patch :p
18:32:34 <Rubidium> So you don't have 32 railtypes, no you have dozens of gauges, dozens of voltages, dozens of weights, dozens of forms, dozens of types of steel and so on to choose from
18:34:05 <Alberth> use a railway simulator to simulate a railway
18:34:26 <Snail> that’s like putting to an extreme. I’d be fine to have a combination of 2 / 3 gauges, max weight, max speed, and electrification type to choose from
18:34:30 <Alberth> it's not, it's a transport game
18:34:37 <Snail> a bit like locomotion when electrification was an overlay to a track
18:34:54 <Snail> but it’s not possible in OTTD to allowing for more railtypes would be a way to get the same effect
18:35:03 <Rubidium> because for reality, I really need NP46, UIC54E1, UIC54E3, UIC54E5 and UIC60
18:35:30 <peter1138> don't forget uk loading gauge!
18:36:06 <Rubidium> and 1500 VDC <2500 A, 1500 VDC 2500-4000 A, VDC > 4000 A, 3000 VDC, 15kV AC, 25 kV AC
18:36:31 <peter1138> just add a byte to the map array
18:37:15 <Alberth> less if you have more values for a property than 2 :p
18:37:33 <Rubidium> and not to mention the safety systems that might be available (which is, ofcourse, a bitmask): ATB EG, ATB NG, ATB VV, ERTMS level 1, ERTMS level 2, ERTMS level 3, ERTMS level 3 hybrid, PZB, LZB, ...
18:37:57 <Snail> ok, so let’s put everything on 1 railtype
18:38:01 <Rubidium> oh, and the type of sleepers ofcourse... concrete, wooden, composite
18:38:08 <Snail> monorail, maglev, electrified…
18:38:14 <Snail> but why not road vehicles and trams too
18:38:38 <Rubidium> heh... by I haven't started on the hardnesses/brands of rail yet...
18:39:56 <Snail> even cable cars… and airplane runways too… who needs different roads/rails anyway
18:41:08 <Rubidium> there are at least 14 "qualities" of rail and at least 44 "brands" of rail that we need to model properly
18:41:24 <Rubidium> can we ever be realistic for a huge country like the Netherlands
18:42:43 <peter1138> ok we can use 4 bits of m6
18:42:48 <Alberth> :O that many? lots of old tracks I guess
18:43:15 <Snail> Rubidium: if we took things to an extreme, we could talk about either 1 type for all, or thousands of types. My point is just to have the possibility to expand current things a bit
18:43:49 <Rubidium> but where it too much?
18:45:21 <Rubidium> and we already expanded it two bits
18:46:28 <peter1138> actually it was always 4 bits
18:47:35 <frosch123> peter1138: we need a dropdown with a text filter :p
18:47:45 <Snail> one could argue that 16 tram tracks are also “too much”… after all there is less variation for trams than there is for trains… but that was done
18:48:07 <Alberth> Snail: it's like the map size. I have yet to see a usefully filled map at 1024x1024, yet people wanted 2048x2048. Later, people wanted 4096x4096. Patchpacks still have 8192x8192, it is never enough, even though all those maps are just more empty space
18:48:32 <Alberth> yet I still haven't seen anyone filling a map of 1024x1024
18:50:28 <peter1138> ps, why the game run so slow?
18:50:38 <peter1138> make it use all cores!
18:54:02 <Alberth> gpu doesn't work either
18:56:21 <peter1138> michi_cc has a patch for that!
19:14:55 <andythenorth> Snail: why weight, specifically?
19:15:45 <andythenorth> also do you have speed limits?
19:32:44 <Snail> I have both axle weight limits and speed limits
19:33:10 <Snail> axle weight limits are important in the early years, when you can run light rolling stock on lighter rails (cheaper)
19:33:29 <Snail> and before using the large steamers of the late XIX century, you need to upgrade your rails
19:34:37 <Snail> having more railtypes allows to use early electric rolling stock, without having to buy heavier tracks
19:35:48 *** Supercheese has joined #openttd
19:45:47 <DorpsGek> Commit by translators :: r27830 /trunk/src/lang (3 files) (2017-03-26 19:45:37 +0200 )
19:45:49 <DorpsGek> catalan: 6 changes by juanjo
19:45:50 <DorpsGek> italian: 1 change by lorenzodv
19:45:51 <DorpsGek> russian: 6 changes by Lone_Wolf
20:00:13 *** Arveen2 has joined #openttd
20:07:55 * andythenorth finds constraints useful
20:08:02 * peter1138 ties up andythenorth
20:08:10 <peter1138> Wait that sounds a bit too kinky.
20:08:17 <andythenorth> can you unsay that?
20:09:05 <V453000> omg you guys tying each other?
20:11:15 <Supercheese> more like such railtankers
20:11:22 <V453000> but big time python mayhem
20:11:47 <V453000> doing some pretty mad shit with python blender :) all of my rendering is now handled by python scripts now
20:12:24 <andythenorth> saves pressing the buttons though
20:12:31 <andythenorth> also proof that python is dangerous
20:12:34 <andythenorth> anyone can use it :P
20:12:51 <V453000> it's not just about that, with blender you can do "some" degree of automation, but this is so much more flexible
20:13:00 <Supercheese> and it's Factorio
20:13:03 <Supercheese> so automate everything
20:13:22 <andythenorth> factorio should have a script API
20:13:26 <V453000> for example, ever since I started using Blender I always wanted a feature which would allow me to override all materials EXCEPT some in a render pass ... blender only allows to override everything, which has stupidly many cases when it completely breaks
20:13:44 <V453000> andythenorth: some guy already wrote fbasic, he controls the factory with it
20:13:49 <frosch123> sounds like we will see yaks in factorio before we see slugs
20:13:50 <V453000> he made a self-replicating base that plays itself
20:13:50 <andythenorth> the real factory?
20:14:07 * andythenorth considers F as general-purpose automation software
20:14:18 <andythenorth> could have a real furnace somewhere
20:14:22 <andythenorth> super pro edition
20:16:23 <V453000> but yeah, stuff's fun
20:16:34 <V453000> also what is it with yaks? :D
20:16:48 <frosch123> they produce a lot of wool
20:17:03 <frosch123> which is an ingredient to firs or something
20:18:14 <andythenorth> FIRS Tibet economy
20:18:26 <andythenorth> yak farm: butter, wool
20:18:34 <andythenorth> windchime factory
20:32:35 <andythenorth> so in Photoshop CC 2017, the paint bucket has anti-alias on permanently
20:32:52 <andythenorth> NFI what Adobe think they’re doing
20:32:58 <andythenorth> this seems like the worst photoshop ever
20:33:09 <supermop_home> andythenorth well I will be sticking with 2016 then
20:33:37 <frosch123> isn't there some checkbox to uncheck somewhere?
20:34:08 <andythenorth> it’s permanently checked and greyed out
20:34:34 <andythenorth> I have to switch to RGB mode, uncheck it, and switch back to indexed
20:35:22 <supermop_home> that sounds even worse than if they had even removed the option to uncheck
20:35:30 <frosch123> ah, yeah, you are probably their only indexed user
20:35:34 <supermop_home> if they make you jump through a hoop
20:35:45 <supermop_home> but I'm not on 2017
21:00:15 <peter1138> does "+ 1 - x" get nicely thrown away?
21:04:24 <supermop_home> now iD has frozen up while trying to print a pdf
21:05:05 <Alberth> + (1 - x) I would expect to be surely being computed at compile time
21:05:34 <Alberth> but, if you want to win the war by merging constants, you have basically already lost :p
21:06:31 <frosch123> where is lordaro to talk about constexpr? :p
21:06:47 <Alberth> but it likely does common sub-expression elimination etc, so likely it computes something completely different than you think :p
21:07:14 <DorpsGek> Commit by peter1138 :: r27831 trunk/src/widget.cpp (2017-03-26 21:07:08 +0200 )
21:07:15 <DorpsGek> -Fix (r24577): Misaligned resize icon (due to widget bounds being inclusive)
21:09:50 <peter1138> Thought I'd messed up. I was running the wrong copy :po
21:10:15 <peter1138> It would've been a "you had one job" moment...
21:10:48 <peter1138> (but yeah, most of the time left = 10, right = 20 ... 10 pixels... not in ottd land! that's 11 pixels...
21:12:40 <peter1138> I also spent too long verifying that I wasn't imagining it was different.
21:12:54 <peter1138> Turns out original TTD doesn't even have them.
21:26:23 <Alberth> yep, left and right are messy inclusive
21:26:52 <Alberth> I killed a lot of that in the gui rewrite :p
21:30:22 <peter1138> It simplifies some things. But it's unusual.
21:33:06 *** supermop_home has joined #openttd
21:49:41 <peter1138> Width 65 x 22... results in the widget being about double the height it should be
21:49:46 <peter1138> (didn't actually measure it)
21:53:51 <Alberth> 0 wide or 0 high widgets are used a lot :)
21:54:13 <peter1138> it's just a minimum size
21:54:50 <peter1138> i don't see why, if the width is set correctly (and is narrower than the widget ends up being) it makes the widget taller
21:56:12 <Alberth> weird, width and height aren't related, except in some special widgets, afaik
22:27:21 *** gelignite has joined #openttd
22:29:58 * andythenorth makes better vineyard
22:30:37 *** sim-al2 has joined #openttd
22:33:36 *** Hiddenfunstuff has quit IRC
22:34:50 <andythenorth> is just base set sprites
22:42:59 <Eddi|zuHause> the building is too tall
22:44:59 *** FLHerne has joined #openttd
22:45:52 <andythenorth> the building is ugh
22:46:46 *** andythenorth has left #openttd
23:58:37 *** Biolunar has joined #openttd
continue to next day ⏵
