IRC logs for #openttd on OFTC at 2012-08-12
⏴ go to previous day
01:36:50 *** Chris_Booth has joined #openttd
01:55:39 *** Biolunar_ has joined #openttd
02:15:53 *** boomslang has joined #openttd
04:15:11 *** HerzogDeXtEr1 has joined #openttd
04:56:17 *** Eddi|zuHause has joined #openttd
05:49:49 *** andythenorth has joined #openttd
06:02:54 <andythenorth> I was losing my game badly
06:03:01 <andythenorth> so I turned HEQS running costs down
06:03:08 <andythenorth> can't do that in real life :P
06:27:41 *** andythenorth has joined #openttd
06:39:38 *** sla_ro|master has joined #openttd
07:28:03 *** Progman has joined #openttd
07:38:24 *** Alberth has joined #openttd
07:38:24 *** ChanServ sets mode: +o Alberth
07:45:48 *** andythenorth has joined #openttd
07:58:16 <Alberth> I have been thinking about the number of industries; could it be useful to define some sort of function in newgrf that gives the total number of industries?
07:58:26 <Alberth> sort of like the snowline
08:04:48 <Alberth> frosch said yesterday that we have varying construction probabilities of industries, but I cannot find them in the spec :(
08:05:44 <Alberth> perhaps I misunderstood him
08:15:58 <Eddi|zuHause> Alberth: i think a semi-proper solution would be to "force" missing industries despite all slots already being used
08:16:17 <Eddi|zuHause> Alberth: this could be a flag of this callback
08:17:30 <Eddi|zuHause> so during a cb22 run, a NewGRF could check "this is a 512^2 map, and there are less than 5 bauxite mines on the map, force the creation of a bauxite mine"
08:20:23 <andythenorth> +1 to what Eddi|zuHause said
08:20:33 <Eddi|zuHause> Alberth: this avoids the issue of "reserving" slots for unusable industries, in turn you might get more industries than intended, if the NewGRF thinks they are essential to the economy
08:21:01 <andythenorth> although....if closure was trying to maintain a fixed number of industries on the map (it doesn't afaik), forcing new types might conflict
08:21:45 <Eddi|zuHause> if closure is enabled, the number of industries will eventually balance out
08:21:48 <andythenorth> I want the ability to force, even though I'm not convinced reserving slots is wrong
08:22:06 <andythenorth> I can reserve slots carefully, then player builds industries and the plan fails :P
08:22:36 <andythenorth> player-can-build means anything that is too tied to specific number per map is dicey
08:22:40 <Alberth> you do have the ability to force, but just to one instance
08:23:03 <Eddi|zuHause> Alberth: apparently that works only for closing, not for opening
08:23:11 <andythenorth> Alberth ...also assuming a map location can be found :)
08:24:12 <Alberth> Eddi|zuHause: it works for opening, and iirc there is some code that prevents closing the last instance too
08:24:27 <Alberth> andythenorth: good luck building non-buildable industries :D
08:25:18 <andythenorth> Alberth: if I (in newgrf) limited the number of instances built, I could try and keep space free
08:25:24 <Eddi|zuHause> Alberth: we were searching for that code that forces opening, but we did not find it
08:25:37 <andythenorth> it's "try very hard" somewhere
08:25:47 <andythenorth> I've seen it, I just don't understand it
08:26:00 <Eddi|zuHause> andythenorth: no, that is the code that searches locations
08:26:33 <Alberth> andythenorth: I believe the newgrf should state shape of the ground, so the game can adapt
08:27:14 <Alberth> also, the "i can build here" should be split in no because of terraforming, or no because of other reasons, probably
08:28:27 <Alberth> or perhaps the terraforming should be taken out of that decision entirely, and assumed to be correct to the requested shape
08:28:30 <andythenorth> so we can retry harder, or not?
08:29:16 <Alberth> to prevent terraforming places where FIRS then refuses to build for other reasons, mostly
08:29:42 <andythenorth> broadly, I think we can ignore the terraforming issue for the 'number of industries' problem
08:29:50 <andythenorth> I think frosch has ideas for it anyway
08:30:21 <Eddi|zuHause> Alberth: problem with that is, if the industry can be built on several slopes, which one to terraform to?
08:30:44 <andythenorth> the one for the selected layout
08:30:49 <Eddi|zuHause> that's why terraforming is disabled completely if the shape check callback is used
08:32:34 <Alberth> terraform to a randomly chosen layout
08:33:33 <Alberth> andythenorth: indeed, terraforming negotiation is mostly for making the 'build industry' process less likely to fail
08:36:04 <Eddi|zuHause> Alberth: i think the problem is, that it first checks whether industry slots are free, and then forces the building of a missing industry type. but it should do that BEFORE checking the free industry slots
08:36:41 <andythenorth> is that quite expensive?
08:37:58 <Eddi|zuHause> andythenorth, Alberth: but industry layout and terraforming shape is not a 1:1 mapping. one layout may be suitible for building on several shapes
08:40:06 <Alberth> I even think you don't want to specify all of them explicitly, you need a more general shape definition
08:41:55 <Alberth> iirc frosch was thinking along the lines of starting at eg the northern corner, and then each tile-edge is a subset of {go up, level, go down}
08:42:09 <Alberth> so you can define a lot of shapes in a compact form
08:48:40 *** TGYoshi has joined #openttd
08:51:04 <Alberth> Eddi|zuHause: do you agree that for industry types that have to be on the map, IndustryBuildData::SetupTargetCount() line 2187 industry_cmd.cpp ensures that its target_count is non-zero ?
08:52:06 <Eddi|zuHause> Alberth: i think the problem is line 2666
08:53:07 <Eddi|zuHause> it needs an exception when an industry is missing
08:55:50 <Eddi|zuHause> i think we can live wouthout that
08:56:07 <Eddi|zuHause> adds some random delay
08:56:23 <Eddi|zuHause> instead of spawning the industries immediately on 1.1.
08:57:53 <Alberth> it basically construct a slowly increasing number of industries, but only when the actual construction can keep up
08:59:28 <Alberth> since there is no control on close down, industries close like they are mad on large maps with dense industries
09:03:11 *** Cybertinus has joined #openttd
09:05:32 <andythenorth> rate limiter on closures?
09:05:42 <andythenorth> maybe configurable for player?
09:05:55 <andythenorth> [also map scaled]
09:09:14 <Alberth> let openttd decide whether or not to close :)
09:09:32 *** Devroush has joined #openttd
09:09:41 <Eddi|zuHause> let's fix opening first, and then handle closing :)
09:11:16 <Alberth> forcing one instance is a very small part of the problem imho
09:11:52 <Alberth> ie you get a 1 industry of each type extra, and that's it
09:11:55 <Eddi|zuHause> that's why i proposed forcing multiple industries by cb22
09:12:26 <Eddi|zuHause> e.g. on small maps with FIRS i get one of each farm, which makes them pretty much unplayable
09:12:49 <Alberth> fixing a broken system by adding more layers just makes it more complicated imho
09:13:04 <Eddi|zuHause> i don't really think the system is broken
09:13:10 <Alberth> imho FIRS is not designed for playing on small maps
09:13:31 <Eddi|zuHause> the map was like 128x256, so not THAT small
09:13:57 <Alberth> perhaps you should read the log of #openttd of yesterday again?
09:14:34 <andythenorth> I think forcing instances via cb22 is useful and a small win
09:14:36 <Eddi|zuHause> i remember the log (mostly)
09:14:37 <andythenorth> the rest of this...
09:14:40 <Alberth> progressing from one 'era' to another one is very very difficult
09:14:53 <andythenorth> the more I think about it, the more I want a NewEconomy, made of up of several quite small things
09:14:57 <andythenorth> and exposed properly to GS
09:15:09 <andythenorth> and as a game option
09:15:16 <Eddi|zuHause> that would be NoEconomy :)
09:15:25 <Alberth> and I don't see how forcing additional industries fixes that problem in a good way
09:15:30 <andythenorth> NoEconomy exactly :)
09:15:37 <andythenorth> we have a winner :)
09:16:48 <andythenorth> all those work for me :)
09:17:13 <andythenorth> Alberth: the alternative to forcing in cb22 is to have the game take care of this better
09:17:41 <andythenorth> the issue I want to solve is when wholly new chains arrive; replacing existing types with new types, I don't care about so much right now
09:17:53 <andythenorth> (1) force in cb22
09:17:57 <andythenorth> (2) reserve slots
09:18:05 <andythenorth> (3) have game increase number of industries required
09:18:43 *** frosch123 has joined #openttd
09:20:06 <andythenorth> (3b) have game count how many industry types are a in a grf, and scale required numbers accordingly
09:20:19 <Eddi|zuHause> ok, eras: a mind-scenario - imagine we have a "pre-industrial" era with very many very small industries (like each town has almost all industry types (carpenter, smith, market, ...). and now it shifts to an industrial era, where many small industries close down and fewer large industries replace them
09:20:40 <andythenorth> Eddi|zuHause: that one probably just works
09:20:58 <Eddi|zuHause> now the newgrf needs to a) need a way to force opening the new industries, b) lower the target count of industries, c) force close the smaller industries
09:22:08 <andythenorth> a) can't be done b) limit numbers of types in the newgrf, with map scaling calculated by the author c) use the monthly prod change cb (ugh)
09:22:23 <frosch123> are you discussing again? or are you still discussing? :p
09:22:34 <Alberth> sounds to me that you are asking for control over the total number of industries in the game over time, ie my third line in this channel this morning
09:23:23 <Alberth> frosch123: it feels like new ground to me :)
09:23:50 <Eddi|zuHause> yes, but also no. it's probably a bad idea to expose absolute numbers to the NewGRF. it needs to be scalable by industry density and map size
09:23:52 <frosch123> the 'force build' flag sounds wrong to me :p
09:24:07 <frosch123> we were discussing about disallowing 'force closure', so why add 'force build'?
09:25:36 <andythenorth> if a new type is available, and I knew I could cast-iron trust the game to build it quite quickly, I wouldn't want the cb22 forcing
09:25:49 <andythenorth> but in the short term the cb22 solution is appealingly simple :P
09:26:15 <Eddi|zuHause> andythenorth: i think the simple exception to line 2666 is more simple :)
09:26:29 <Eddi|zuHause> and might render most of the discussion here irrelevant
09:26:37 <Alberth> andythenorth: I don't think a GS has enough room for interesting behavior if the newgrf has to be obeyed
09:27:05 <andythenorth> Alberth: maybe, but then you throw out a lot of newgrf spec. Is anyone brave enough to do that?
09:27:31 <andythenorth> and now your GS needs to handle things like location checks etc
09:27:49 <Alberth> just saying that you cannot give newgrf more control and expect GS to be able to do anything
09:28:26 <andythenorth> newgrf could cede control over cb22
09:28:51 <Alberth> "but respecting result from newgrf"
09:29:23 <andythenorth> it's a proposal, could be changed easily
09:30:01 <NGC3982> I don't understand this. I cant get the production of normal industries to exceed 67%.
09:30:15 <NGC3982> I can manage FIRS, but the normal industries are a pain.
09:30:22 <Alberth> NGC3982: read the game mechanics wiki page
09:35:36 <NGC3982> And im missing out on somethin
09:35:40 <NGC3982> "For an industry that has good (greater than 60%) service: 67% chance of increase, 33% chance of decrease."
09:36:25 <NGC3982> Reading this, i don't understand why my production don't -ever- increases, while stuck at 69%?
09:39:42 <Alberth> what 69%? the percentage transported? that's not the service level
09:40:19 <Alberth> that level is in the town window afaik
09:41:42 <Alberth> although they are related in some way, of course
09:42:02 <frosch123> Alberth: industry producton changes are about percentage transported
09:42:23 <frosch123> anyway, if the probability for an increase is higher
09:42:29 <frosch123> that does not imply that it occurs all the time :p
09:42:38 <NGC3982> Well, it doesn't happen at all?
09:42:45 <frosch123> it just means that it is more likely to increase than decrease
09:42:54 <frosch123> there is nothing said about whether it happens at all
09:43:39 <NGC3982> This industry has been at 36 tonnes for 17 years, and i can't figure out why. :)
09:44:45 <NGC3982> The trains are set to full-load, and the trucks to pickup and transfer.
09:45:06 <frosch123> enable smooth economy then
09:45:13 <frosch123> then all industries change produiction every month
09:45:34 <frosch123> but don't complain about dozen of them closing every year :p
09:46:45 <NGC3982> What am i missing then?
09:47:20 <frosch123> the more you make industries depend on cargo transported, the more unserviced industries (i.e. those which are not connected at all) close down
09:47:23 <NGC3982> It's been so long since i played without industry GRF:s, i can't really remember how to get production to increase quickly.
09:48:43 <frosch123> luckily it does not change quickly :p
09:48:58 <frosch123> else all industries would have 2k production, and your transport network would collapse
09:49:19 <NGC3982> I guess im doing it right, and i simply need to wait.
09:50:42 <Eddi|zuHause> frosch123: i once had a solution for that: ignore industries that belong to a town that has no company ratings
09:51:07 <frosch123> Eddi|zuHause: yeah, i guess something along the lines
09:51:17 <frosch123> stop talking about firs, first fix default industries
09:51:17 <NGC3982> Eddi: I tend to do that too.
09:51:19 <Eddi|zuHause> so unserviced regions stay mostly untouched, until you can get to them
09:51:29 <frosch123> only close if there are not already too few industries of that type
09:51:32 <NGC3982> But mostly for not-wanting-to-bribe reasons.
09:51:39 <frosch123> open if there are too few of them (already done)
09:52:19 <frosch123> very minor chance for production changes for industries in areas without industrialisation
09:54:45 <frosch123> (resp. if industries in unindustrialized areas change, either make them close, or double production)
09:54:58 <frosch123> i.e. only notable changes
09:55:32 <Eddi|zuHause> that will probably worsen the problem of mass-closing industries
09:56:16 <frosch123> only close if there are too many of that type
09:57:08 <frosch123> but yes, maybe big production changes are also weird
09:57:15 <frosch123> so, cross that last idea :p
10:07:23 *** drac_boy has joined #openttd
10:53:56 *** Chris_Booth has joined #openttd
11:10:23 *** michi_cc has joined #openttd
11:10:23 *** ChanServ sets mode: +v michi_cc
11:23:17 *** mahmoud has joined #openttd
11:57:53 *** andythenorth has joined #openttd
12:02:50 <drac_boy> so andythenorth are you going to do anything fun or still keep up same complain about industries mess? :)
12:16:49 *** TheDude has joined #openttd
12:20:30 <drac_boy> I finished breakfast and theres nothing left on the plate sorry :)
12:22:07 <TheDude> to cook a lunch for me, I am quite reluctant to cook it myself
12:22:28 <TrueBrain> andythenorth: je ne parle pas francais
12:23:10 <glx> andythenorth doesn't either :)
12:25:13 <andythenorth> be nice if scenario authors could get game to spawn industry types near seed points
12:25:18 <andythenorth> could be done with signs
12:25:29 <andythenorth> do it by produced or accepted cargo labels
12:25:45 <andythenorth> doing it by industry type - we have no way to ensure that, kind of fragile
12:26:01 <TrueBrain> why not make a GS like script for scenarios
12:26:06 <TrueBrain> where they can predefined what happens at what time
12:26:12 <andythenorth> maybe that's the way to do it
12:26:15 <TrueBrain> one big state machine which triggers on dates
12:26:27 <TrueBrain> where they can start industries etc, maybe even production
12:26:34 <TrueBrain> as the scenario knows the newgrfs etc
12:27:00 <Eddi|zuHause> such a script isn't really reusable
12:27:09 <TrueBrain> it is not meant to be
12:27:10 <Eddi|zuHause> so i don't think many people will bother
12:27:11 <andythenorth> GS library might be?
12:27:14 <TrueBrain> it is to be paired with a single scenario
12:27:20 <TrueBrain> much like the Tutorial AI
12:27:58 <Eddi|zuHause> hmz... the internet knows absolutely no date on when doctor who is supposed to start...
12:28:06 <Eddi|zuHause> it always says "this fall"
12:28:24 <Eddi|zuHause> occasionally it hints "august"
12:28:54 <Eddi|zuHause> one would imagine the program plan for august is already done, so the date would be known...
12:29:35 <V453000> there is some issue with the xUSSR railway set on bananas, when we put it on the server we cant connect to the game, it says that we have to download the newgrf ... when we download it, it rescans newgrfs and wants to download it again
12:29:55 <TrueBrain> update to 1.2.2-RC1
12:30:13 <V453000> iz trunk .. :) but still
12:30:14 <TrueBrain> Windows clients calculate wrong MD5 in several cases
12:30:46 <drac_boy> V453000 if you're going to use russia trains..at least consider some russia planes too? :P
12:31:01 <V453000> im not really going to play that game :p
12:31:36 <Alberth> andythenorth: I once had the idea to be able to assign areas for each industry type where it could spawn new industries
12:31:37 <TrueBrain> DorpsGek: wtf dude?
12:31:40 <TrueBrain> @opentd commit 24416
12:31:44 <TrueBrain> @openttd commit 24416
12:31:48 <DorpsGek> TrueBrain: I don't recognize you.
12:31:55 <TrueBrain> so what? That function should work!
12:32:32 <Alberth> he's truly ma today :D
12:32:42 <DorpsGek> TrueBrain: The operation succeeded.
12:32:51 <DorpsGek> TrueBrain: Commit by rubidium :: r24425 /branches/1.2 (11 files in 4 dirs) (2012-07-20 19:49:02 UTC)
12:32:52 <DorpsGek> TrueBrain: [1.2] -Backport from trunk:
12:32:53 <DorpsGek> TrueBrain: - Fix: [Windows] Unbreak NewGRF MD5 sum calculation. Macros and side effects do not mix, especially if there is some obscure '#define min' in a windows header that nobody thinks of [FS#5231] (r24416)
12:32:54 <DorpsGek> TrueBrain: - Fix: Disallow removing roadtypes from bridges when not dragging in bridge direction [FS#5221] (r24414)
12:32:55 <DorpsGek> TrueBrain: - Fix: Draw wires under low bridges if the bridge is transparent, not if the wire is transparent (r24403)
12:33:03 <TrueBrain> V453000: first entry ^^ ;)
12:39:51 *** Chris_Booth has joined #openttd
12:44:26 <andythenorth> Alberth: just use signs?
12:44:47 <andythenorth> dirty, probably works
12:45:06 <Alberth> for GS? sure Zuu also uses that
12:45:33 <Alberth> you may want to state a max distance in the sign or so
12:47:59 <andythenorth> make it available to ottd even without GS?
12:58:39 <Alberth> using signs feels hacky then, imho
12:59:18 <Alberth> you'd need to make a generic solution to indicate areas
13:00:06 <Alberth> the TileArea already exists, it'd be mostly GUI for specifying, and of course to actually do something useful with the areas
13:00:54 <frosch123> hmm, i am stuck on lever 12
13:01:05 <Alberth> and likely you want to be able to state what the tile area is intended to be used for :)
13:01:17 <frosch123> old games are harder
13:02:27 <TrueBrain> andythenorth: sure you didnt mean diper wars?
13:02:27 <frosch123> i saw an advertisement for some *innovative* flash game, i thought i know that game, booted dosbox, and now play sokoban from 1984 :p
13:03:03 <Alberth> sokoban is quite difficult indeed :)
13:05:21 <frosch123> flash games are like teeny music
13:05:44 <frosch123> if you are old enough you know all melodies of all *new* songs already in advance :p
13:06:02 * andythenorth used to make flash games
13:06:15 <andythenorth> all were copies of 80s games
13:06:33 <frosch123> andythenorth: and your customers did not know them :)
13:09:10 <andythenorth> Alberth: signs are hacky indeed
13:09:18 <andythenorth> but we already have methods for adding them etc
13:09:29 <andythenorth> new gui = yet more code :)
13:09:50 * andythenorth favours shortcuts today :P
13:10:41 <andythenorth> can signs be placed programmatically?
13:11:27 * andythenorth just had the idea of creating sets of signs
13:11:36 <andythenorth> that can be toggled on/off for visibility
13:12:00 <andythenorth> and used to control things like GS communication, industry seeding etc
13:12:20 <TrueBrain> signs created by GS are like that :P
13:12:29 <TrueBrain> hmm, no, you cannot hide them
13:12:38 <TrueBrain> you should make a new thing called ... 'marker'
13:12:45 <TrueBrain> which are like signs, but different, for that sole purpose
13:15:33 <TrueBrain> and while at that, make them TileAreas, like suggested, and you have a complete setup :D
13:16:11 <frosch123> TrueBrain: zuu wrote a lib, which reads signs on game start, stores them in saveable squirrel structs, and then removes the signs
13:16:32 <TrueBrain> but like said, that is a hack :P
13:16:36 <frosch123> he uses that stuff in his gs, to mark places on the map for the script with signs
13:17:00 <frosch123> it's less a hack than hardcoding the positions into the script :p
13:20:39 <frosch123> gah, sokoban has no undo
13:20:44 <frosch123> not even for a single move :(
13:21:31 <TrueBrain> undos all your moves :D
13:21:48 <frosch123> old games are so cruel
13:22:06 <TrueBrain> I hate that games make it so easy these days
13:22:12 <frosch123> touching a wall marginally is game over
13:22:19 <TrueBrain> you died? Owh no worries, we just put you back 5 seconds before you died and off you go again
13:22:22 <frosch123> that would never happen for today games :p
13:22:46 <frosch123> TrueBrain: yup :p diablo iii is silly
13:23:29 <frosch123> skyrim at least tried to have some story or so :p
13:24:22 <andythenorth> marker, applied similarly to signs
13:24:31 <andythenorth> possibly has drop-downs for values that are enums or such
13:24:43 <andythenorth> and fields for values that are unbounded
13:24:45 <V453000> openttd has no story! :P
13:24:54 <TrueBrain> write a Quest for OpenTTD :P
13:25:32 <andythenorth> Eddi|zuHause: the 'lots of small industry, then few large industry' case - GS could do that yes / no?
13:25:33 <Zuu_> Just randomly checked the logs and found out that I was mentioned. :)
13:25:44 <andythenorth> [apart from 'wtf is the interface between GS and newgrf] :P
13:26:12 <Eddi|zuHause> andythenorth: no it can't, because it is depending heavily on the newgrf
13:26:27 <V453000> Zuu_: I noticed that when I used the "neighbours are important script", it only influences towns generated on map creation, not towns funded in game. Is there any way to go around that and affect all towns?
13:26:37 <andythenorth> Eddi|zuHause: for open / close; what else for?
13:26:57 <Zuu_> Andy: SCP, if krinn finds time to work on it to release a new version.
13:27:01 <Eddi|zuHause> andythenorth: for figuring out whether the industries are small or not
13:27:11 <andythenorth> because industries are black box to GS?
13:27:30 <Eddi|zuHause> andythenorth: for figuring out that there is a difference of "phases" at all
13:27:59 <Zuu_> Hmm, sorry SCP is not really grf - gs. Bad memory of mine.
13:28:00 <Eddi|zuHause> no, an industry set that has such a difference in style, must tell that to the game through some methods, so the game can figure out what to do
13:29:49 <andythenorth> doesn't that entail that the game is capable of understanding every concept a newgrf author might invent?
13:29:54 <andythenorth> or something else?
13:30:25 <Zuu_> V453000, IIRC it should handle that event but it might not be working. You could try to save + load. Also report it in the thread and I'll tke a look at he code when possible (at a train at the moment)
13:31:03 <V453000> ok ... I tried even saving/loading :)
13:31:06 <Eddi|zuHause> yes. the game should be prepared for certain "ideas" that industry authors may have, and be able to react on them. but this can only happen if the NewGRF is capable of telling this to the game
13:31:19 <andythenorth> isn't "small" irrelevant?
13:31:35 <andythenorth> the only information of interest is 'close this type after [date xyz]'
13:31:36 <Eddi|zuHause> "small" wrt maximum production
13:32:09 <andythenorth> we can't just key to industry IDs?
13:32:51 <andythenorth> write a GS that knows the IDs of industry types
13:33:14 <andythenorth> because it doesn't work, or because it's stupid and fragile?
13:33:30 <Eddi|zuHause> mostly the latter :)
13:34:08 <andythenorth> I always end up here when discussing newgrf and GS :P
13:34:28 <andythenorth> either GS controls the narrative, or newgrf controls the narrative
13:34:29 <Alberth> ignore eddi for now, and just see how far you get?
13:34:36 <Eddi|zuHause> imagine you add an industry to FIRS-GRF, how do you add this industry to FIRS-GS?
13:34:46 <andythenorth> Eddi has an unfortunate tendency of being correct :|
13:34:52 <Eddi|zuHause> and then make sure the versions the player uses are compatible?
13:35:04 <andythenorth> Eddi|zuHause: adding a type is a total PITA
13:35:22 <V453000> Zuu_: posted it in the thread :)
13:35:44 <Eddi|zuHause> i really hate being correct all the time :)
13:35:51 <andythenorth> I am only baffled because no-one is prepared to say 'newgrf controls game' or 'GS controls game'
13:36:04 <andythenorth> it's same as when we talked with Pikka about it, only less cross :)
13:36:37 <TheDude> I am with andythenorth, I also miss the ability of GS to access cargos by ID
13:36:46 <andythenorth> I don't have a position :P
13:36:53 <andythenorth> I just want to make an industry set that is fun
13:36:57 <andythenorth> and then maybe write some GS's
13:36:59 <Zuu_> It might work in a Scenario where the GS know for sure which GRFs that are used in the game.
13:37:11 <frosch123> TheDude: contrary to industries you can access cargos by id
13:37:52 <Eddi|zuHause> Zuu_: but that's not the normal use case of an industry set
13:38:17 <andythenorth> maybe we should...hmm
13:38:37 <andythenorth> - scripted scenarios (know set of grfs, newgrf industries cede control to GS)
13:38:45 <andythenorth> - freeform play, newgrf controls
13:38:49 <andythenorth> probably bad idea
13:39:14 <Eddi|zuHause> in general, the NewGRF should control nothing, but give hints to the game. which the game can then expose to GS or not.
13:39:44 <andythenorth> Eddi|zuHause: for closing that's possible, with careful use of existing cbs
13:40:02 <Eddi|zuHause> like, the GRF hints to the game the introduction date of a vehicle. the game then decides whether the vehicle should get available (random factor, refitability, etc.)
13:40:25 * andythenorth wonders if grfs should be able to look at some global vars
13:40:31 <andythenorth> like 'current epoch' or such
13:40:41 <andythenorth> avoiding hard-coding to specific dates
13:40:54 <Eddi|zuHause> what's an "epoch" outside model railway context?
13:41:18 <andythenorth> something GS and newgrf authors need to agree on :P
13:41:26 <andythenorth> it's pretty arbitrary
13:41:42 <Eddi|zuHause> honestly, i don't see that happening
13:41:50 <andythenorth> it could even be that each newgrf details publicly what it will do for certain values of certain vars
13:41:58 <andythenorth> like an event subscriber system
13:42:05 <andythenorth> "All problems in computer science can be solved by another level of indirection" :P
13:42:34 <Zuu_> There could be sort of a board/table where gs/GRFs can expose information to eachother.
13:42:36 <andythenorth> let's assume we have some bundle of economy stuff, call it NoCo. Keep it simple.
13:42:43 <andythenorth> Then industries subscribe to NoCo events
13:42:59 <andythenorth> and the newgrf author controls how the newgrf responds to NoCo events
13:43:02 <andythenorth> but details it publicly
13:43:19 <andythenorth> default NoCo events are existing smooth economy (or non-smooth)
13:43:25 <andythenorth> and GS can influence NoCo
13:43:41 <andythenorth> cb 29 and 35 are examples of current NoCo events
13:43:53 <andythenorth> but need a bit of rethinking
13:44:22 <andythenorth> cbs are enough, we don't need actual pub / sub
13:45:28 <andythenorth> give newgrf a way to also trigger NoCo events, which is turned off if a specific GS flag is set
13:45:54 <andythenorth> then eddi can have periods of industry opening / closures in the absence of a newgrf, with specific dates
13:46:12 <andythenorth> but GS can still do stuff with the same newgrf
13:46:24 <andythenorth> also give towns NoCo options :P
13:46:58 <andythenorth> "absence of a newgrf" / "absence of a GS" /s
13:47:53 <andythenorth> meh, the implementation I can't figure out, but there's something in this
13:49:03 *** Eddi|zuHause has joined #openttd
13:49:57 <andythenorth> we *could* get off the fence and start imposing some more details of the economy
13:50:39 <andythenorth> like 'there will be n industrial phases" etc
13:52:08 <Alberth> Any 'replacenew' NML guru around???
13:53:33 <Zuu_> Perhaps a GS could make a graph of the industry chains, and use that to during some 10-20 years focus on different chains.
13:56:05 <Zuu_> Eg start with most production and secondary industries in wood chain and then after 10 years over several years shift over to food chain.
14:00:26 <andythenorth> Zuu_: you have access to industry accepted / produced cargos?
14:01:33 <Zuu_> Yes. Without that AIs would have a hard time.
14:04:42 <Eddi|zuHause> i think we need a plan on what scopes we want to have. e.g. local scope, regional scope, global scope and where the different parts (production, destinations, open/close) interact
14:05:30 <Eddi|zuHause> focus on the default industries, and then add sprinkles for what newgrf industries want to do differently
14:06:10 <Eddi|zuHause> "region" may mean "all industries of a town" or "all industries nearby", or both, or separate...
14:06:31 <Eddi|zuHause> e.g. the idea that industries need power
14:06:45 <Eddi|zuHause> which isn't currently possible by NewGRF means alone
14:09:37 <Alberth> Eddi|zuHause: call it 'supplies'
14:10:33 <Eddi|zuHause> Sacro: he should have read the newspapers
14:15:30 <andythenorth> ProductionCanIncrease() <- does that look at value of production multiplier?
14:16:54 <andythenorth> Eddi|zuHause: so opening / closing / prod change etc might all vary by a scope (region)?
14:18:44 <andythenorth> local scope is town
14:18:49 <andythenorth> global scope is obvious
14:19:28 <andythenorth> then what? regions = squares of arbitrary size? or a set of towns?
14:19:50 <Sacro> Eddi|zuHause: he found out
14:27:40 <Eddi|zuHause> andythenorth: no, local scope is the industry itself
14:28:12 <andythenorth> so town is regional scope? Or something more arbitrarty?
14:28:23 <Eddi|zuHause> regional scope is something that constitutes "nearby", details to be decided
14:29:39 <andythenorth> so it might be a set, composed of n local scopes
14:29:48 <andythenorth> according to arbitrary criteria :P
14:29:51 <Eddi|zuHause> in each scope, the industry set tells the game what it likes to happen (like industry probability callback)
14:30:20 <Eddi|zuHause> and the game or a gamescript can read that and decides whether it wants to listen or do something arbitrary
14:31:20 <Eddi|zuHause> in turn, the NewGRF can read some variables about each scope what actually happens
14:31:58 <Eddi|zuHause> i think it's easiest if "region" means "town" there
14:32:09 <Eddi|zuHause> so local scope, town scope, global scope
14:32:14 <andythenorth> so much of the rest of the game is town scoped
14:32:50 <Eddi|zuHause> so the industry may read a town variable whether the town has a serviced power station
14:34:07 <Eddi|zuHause> the industry set has fairly tight control over the local scope, but next to no control over the town or global scope
14:34:26 <andythenorth> it's a bit of a specific example, but I would handle things like electricity using the secret / magic cargos idea
14:34:31 <Eddi|zuHause> in turn, the game script has no control over the local scope, but lots of control over the town and global scope
14:34:53 <Eddi|zuHause> it's just an example
14:35:00 <andythenorth> so more examples:
14:35:03 <Eddi|zuHause> doesn't need to be happening at all
14:35:06 <andythenorth> - industry wants to close, what happens?
14:36:50 <Eddi|zuHause> industry tells the game it wants to close. game checks global conditions (like protection of last industry) and then decides whether to close or not. if a gamescript is running, game may issue an event to the gamescript, so the game script can decide what happens
14:37:29 <andythenorth> [details] so interface might remain same (return 03 to cb29 or cb35) ?
14:37:42 *** Progman has joined #openttd
14:38:07 <Eddi|zuHause> just the NewGRF may not expect it to actually happen the way it requested it
14:38:39 <andythenorth> - newgrf has a new type available to build, and wants it on the map, what happens?
14:40:11 <Eddi|zuHause> that's a tricky one. currently, only cb22 may tell the game that the probabilities have changed. the question is, whether that is enough
14:41:55 <drush> are there any cross-compiling devs here?
14:42:31 <Eddi|zuHause> cross from what to where?
14:42:45 <andythenorth> putting implementation aside, principle is the same as closing no? Newgrf tells game it wants a new industry type to appear (with extreme prejudice); game decides outcome, with optional event for GS
14:43:13 <drush> Eddi|zuHause: cross from linux to ps3 (I've got the toolchain needed)
14:43:37 <Eddi|zuHause> and what's the problem?
14:44:41 <drush> first, does configure script have a toolchain prefix option?
14:44:57 <drush> second, I really want to avoid -rdynamic flag (unsupported by toolchain)
14:46:04 <TrueBrain> NGC3982: please refrain from such random links; it is not appreciated in this channel (even more without context)
14:46:44 <drush> third, that I can actually solve, is that there's code that looks for asm-generic/ioctl.h, but I have that in my toolchain inside a different directory (that can be changed, but a logic switch would be better)
14:47:27 <drush> such logic switch that if OS=PS3, then use the one supplied in the ps3 toolchain
14:47:37 <drush> otherwise continue as usual
14:48:15 <NGC3982> TrueBrain: sure thing
14:48:36 <Eddi|zuHause> drush: there's a "CC_HOST" and "CC_TARGET", not sure what they do, though
14:49:25 <drush> host is the machine the outcome is supposed to run
14:49:40 <Eddi|zuHause> no, host is the compiling machine
14:49:46 <Eddi|zuHause> for things like strgen and stuff
14:49:50 <drush> isn't build the compiling machine?
14:50:05 <Eddi|zuHause> target is the machine it should run on
14:51:18 <TrueBrain> Eddi|zuHause: sorry to say, but you used the wrong tupple there :)
14:51:24 <TrueBrain> OpenTTD uses the CC_BUILD / CC_HOST pair
14:51:45 <Eddi|zuHause> then i have an outdated wiki page
14:52:04 <TrueBrain> Eddi|zuHause: basically, there are 2 systems to do that .. BUILD/HOST and HOST/TARGET
14:52:07 <TrueBrain> which is confusing like fuck
14:52:12 <TrueBrain> depending on the software another pair is used
14:52:15 <drush> you're telling me TrueBrain
14:52:30 <TrueBrain> well drush, you could just have read the manual
14:52:32 <TrueBrain> --build=BUILD configure for building on BUILD [guessed]
14:52:33 <TrueBrain> --host=HOST cross-compile to build programs to run
14:52:36 <TrueBrain> it is not rocket science ;)
14:52:39 <Eddi|zuHause> TrueBrain: then it was probably changed in the makefile rewrite
14:52:43 <TrueBrain> Eddi|zuHause: possibly
14:52:46 <Eddi|zuHause> meaning the page is like 5 years outdated :p
14:52:46 <TrueBrain> we once used BUILD/TARGET
14:52:49 <TrueBrain> which was even worse :D
14:52:52 <TrueBrain> got to go, back in a few
14:55:42 <andythenorth> so except cb22 (called randomly on an industry type), all current industry cbs are specific to an industry instance
14:55:54 <andythenorth> maybe there should be a regular cb to grf
14:56:10 <andythenorth> called multiple times, allowing return values for various interesting things
14:57:41 <andythenorth> hmm, return value 09, increase global warming (amount in register 0x100)
14:58:46 *** TheDude has joined #openttd
14:58:48 <andythenorth> return value 07, request town to grow / shrink (signed)
14:59:01 <andythenorth> optional news message :
14:59:26 <andythenorth> "due to pollution town shrinks" or "due to employment town grows"
15:00:11 <andythenorth> ach, terrible examples, they are specific to industry instances, not general to the grf
15:07:22 <andythenorth> - please turn on / off cargo type xyz
15:07:51 <Eddi|zuHause> idea: an industry could have an "economy effect" (like, number of jobs)
15:08:10 <Eddi|zuHause> which determines the "size" of the industry
15:08:50 <Eddi|zuHause> so the game(script) would not aim to have a fixed number of industries, but a fixed number of jobs
15:08:58 * Alberth has already a job, namely 'tycoon'
15:09:10 <Eddi|zuHause> and the game script could tie this to town growth
15:09:20 <andythenorth> ach, so also cargo destinations are interesting, haven't thought about those much
15:09:33 <andythenorth> but PAX might prefer a town with 'more tourism industries' or such :P
15:10:09 <drac_boy> Alberth heh, I always have "company leader" a lot of the times :p
15:11:14 <Alberth> drac_boy: that's a good job too :)
15:13:28 <andythenorth> so some kind of regular cb to the industry grf seems wise
15:13:36 <andythenorth> daily probably way too frequent, monthly enough
15:14:41 <andythenorth> with return values including: try to build industry of type x (n instances); turn cargo on / off; news message(s)
15:15:14 <andythenorth> changing cargo payment rates might be nice, but I think the game handles that horribly right now
15:15:31 <drac_boy> Alberth yeah ... and the company value varies from 1DEM to very high values depending :p
15:18:09 <Alberth> I only care about money coming in faster than I can build :p
15:18:55 <drac_boy> Alberth well I don't ... I run light mixed trains with as much single tracks as there could be
15:20:23 <Alberth> single tracks are bad with breakdowns enabled :)
15:24:04 <andythenorth> Eddi|zuHause: so closing many small industries frees up labour pool to build fewer large industries?
15:24:12 <andythenorth> each industry has a labour factor
15:24:20 <andythenorth> therefore game 'knows' about small and large industries
15:24:32 <Eddi|zuHause> exactly, that's the point
15:24:56 <andythenorth> and closure of one primary or secondary due to (poor performance, run out of resources, whatever) - frees up labour for a new industry?
15:29:34 <andythenorth> Eddi|zuHause: scope labour pool to town or global?
15:29:39 <andythenorth> or some intermediate region?
15:30:10 <andythenorth> x% of town popn works
15:30:21 <andythenorth> it could all work
15:30:22 <Eddi|zuHause> town growth may be too fast
15:30:38 <andythenorth> smells a bit of classical economics, but I like the gameplay possibilities :)
15:32:23 <drac_boy> Alberth btw theres a reason theres always at least one possible detour route somewhere :)
15:34:15 <andythenorth> Eddi|zuHause: interesting possibilities, like not enough population in remote areas with natural resources
15:34:44 <andythenorth> combine population requirements with ability to seed areas of map with certain industry types
15:34:49 <andythenorth> gameplay gets a lot more interesting to me
15:35:21 <Eddi|zuHause> "natural ressources" may be a good town scope effect. the game itself wouldn't care, but a game script may
15:35:21 <andythenorth> I think this would be a good [optional] element to NoCo
15:35:53 <andythenorth> I don't know if population is a specific thing or a variant of magic cargos
15:36:01 <andythenorth> I'm thinking of magic cargos as frosch(?) had them
15:36:03 <Eddi|zuHause> the game script will then make "agricultural towns", "forest towns", "mining towns" etc.
15:36:38 <Eddi|zuHause> but the newgrf must be prepared for it
15:36:55 <andythenorth> can't avoid tying newgrf + GS together somehow
15:37:05 <Eddi|zuHause> so there must be a property with a universal set of values
15:37:31 <andythenorth> I recall magic cargos are things like VOLT (electricity) etc
15:37:41 <Eddi|zuHause> "this is a mine", "this is a plantation", "this is a farm"
15:37:54 <andythenorth> industry label scheme :P
15:38:10 <Alberth> cargo types of production should be sufficient, probably
15:38:24 <andythenorth> maintain sets of known cargo labels?
15:38:52 <Alberth> until we get weekly new industry sets, that seems doable to me :)
15:39:00 <Eddi|zuHause> well there are already flags for "extractive" industries etc.
15:39:12 <Eddi|zuHause> just has to be clarified what they should mean
15:39:23 <Alberth> perhaps you can use the classification of the cargo?
15:39:38 <andythenorth> union of type flag and produced cargos
15:41:03 <andythenorth> might be overkill :P
15:49:50 <Eddi|zuHause> Alberth: but one industry could produce a cargo from natural ressources, and antoher from synthetic chemistry
15:50:36 <andythenorth> so look at type and cargo?
15:50:37 <Eddi|zuHause> so you can't judge from the cargo alone
15:51:03 <andythenorth> stuff goes badly when we try to make it too detailed, or so cargo classes suggests :)
15:52:29 <Alberth> perhaps you should look at input too, but I think that you always go wrong at some point, unless you code industry knowledge in the GS
15:53:19 <Alberth> but perhaps first make something that works somewhat before you start optimizing?
15:53:55 <Eddi|zuHause> i'm thinking a property of general industry characteristics. like primary/extractive/surface, primary/extractive/underground, primary/natural/summer, primary/natural/full year, secondary/clean, secondary/dirty, tertiary
15:56:24 <Eddi|zuHause> examples (in the same order): sand pit, coal mine, farm, forest, brewery, refinery, bank
15:56:52 <Eddi|zuHause> a game script may make sure that no two "dirty" industries are in the same town
16:05:43 <Eddi|zuHause> oh, secondary/energy could be one of them
16:06:43 <Eddi|zuHause> hm... must screw up my system now...
16:06:49 <drush> I've ran into a new problem whilst cross-compiling
16:07:55 <drush> I fixed all include and lib paths by hand, used HOST and BUILD equal to the toolchain, this generates a strgen for ppc64
16:08:32 <Eddi|zuHause> BUILD must be the regular linux compiler
16:08:38 <Eddi|zuHause> exaclty for strgen and stuff
16:10:02 <TrueBrain> does Eddi know where the exit button is?!
16:21:52 *** peter1138 has joined #openttd
16:21:52 *** ChanServ sets mode: +o peter1138
16:47:45 <peter1138> Windows update still is not working :-(
16:52:58 <TrueBrain> Windows being Windows :(
16:56:46 <TrueBrain> we werent talking about you Zuu_, I promise :D
16:57:17 <Zuu_> You mean, while I have been away?
16:57:30 <TrueBrain> last time you came because we talked about you :P
16:57:44 <andythenorth> did I miss it getting solved?
16:57:53 <andythenorth> sometimes, peter1138 solves it
16:59:10 <Zuu_> Hehe. I followed the logs, but feared that I would lost my connection too many times so I quit from IRC.
16:59:54 <Zuu_> Now Im back at the main lines and should be better off. :)
17:00:14 <TrueBrain> I can kick you, just to point out you aren't :P
17:02:40 <Zuu_> Hehe. The current train even have a 3G connection that it maintains and offer through WiFi.
17:04:09 <Zuu_> Eg more stable but usecure aainst evasdropping from fellow passengers. :)
17:13:46 * andythenorth should organise all these NoCo ideas into sanity
17:27:56 <andythenorth> I think introduction of 'workforce' is the most insane idea for a while
17:28:02 <andythenorth> might be insane enough to be good
17:29:37 <andythenorth> is it even true to say GS can trigger a newgrf cb?
17:29:48 <andythenorth> someone explained how GS worked to me once, but I forget :P
17:31:29 <TrueBrain> it _can_ do everything, that is of little issue :P
17:31:53 <Alberth> openttd is like a big set of functions to a GS
17:32:56 *** Eddi|zuHause has joined #openttd
17:33:38 <Eddi|zuHause> ok, i successfully screwed up my system
17:33:45 <Zuu_> And a GS is like a thread that runs continuesly. (actually it isnt)
17:34:26 <TrueBrain> Eddi|zuHause: I already asked if you even knew where the exit button was .. don't scare us like that please :P
17:34:57 <Eddi|zuHause> on boot, the driver says something like "no supported video devices found"
17:35:04 <Zuu_> Eg it may be busy doing something while openttd puts an event in its inbox.
17:35:16 <Eddi|zuHause> i now booted in a backup system
17:35:24 <Eddi|zuHause> that i totally forgot how to boot into :p
17:36:19 <TrueBrain> that is the important thing :P
17:38:25 <Eddi|zuHause> now i still have no idea what to do, though
17:39:01 <Eddi|zuHause> everything is configured the wrong way
17:40:00 <Alberth> oh, I should remove my kernel, forgot about that, thanks :)
17:40:38 <Eddi|zuHause> it's not depending on kernel, i tried booting into multiple ones
17:40:47 <Eddi|zuHause> from 3.1.10 to 3.5.1
17:41:13 <Zuu_> Btw, is setting up a jail complicated? ( for someone who has not done it before)
17:41:20 <Alberth> my 3.5.[01] kernel crashes Xorg on the nouveau driver
17:42:08 <Zuu_> I though about setting up some AI battles in jail.
17:42:38 <TrueBrain> why? they are already contained?
17:44:41 <Zuu_> Or maybe a vm is easier.
17:45:04 <TrueBrain> why? they are already contained?
17:45:21 <CIA-4> OpenTTD: translators * r24464 /trunk/src/lang/ (korean.txt slovak.txt):
17:45:21 <CIA-4> OpenTTD: -Update from WebTranslator v3.0:
17:45:21 <CIA-4> OpenTTD: korean - 5 changes by telk5093
17:45:21 <CIA-4> OpenTTD: slovak - 19 changes by klingacik
17:46:09 <Zuu_> The AIs are, but not OpenTTD or the script to run the battles.
17:46:27 <TrueBrain> normally you use a jail if it is possible to escape an application
17:46:35 <TrueBrain> for security reasons
17:46:41 <TrueBrain> but running a battle, how can that be an issue?
17:48:25 <Alberth> Zuu_: run them as a different user?
17:49:04 <Alberth> although I agree with TB in not seeing the need
17:50:02 <Zuu_> A single battle would not be an issue, but if you start to automate it and let it run over time it doesent hurt to secure it from accesseng the rest o the system.
17:50:24 <TrueBrain> as a jail only protects against directory access
17:50:35 <TrueBrain> *BSD does also against process access, kinda
17:50:40 <TrueBrain> but it is easily avaded
17:51:44 <Eddi|zuHause> i'll try to un-screw my system now...
17:52:11 <Zuu_> Ok. Dont hol your breath though. Im thinking about prioritizing a new tutorial chapter for the moment.
17:53:10 *** andythenorth has joined #openttd
17:53:22 <andythenorth> are apple laptops waterproof?
17:53:56 <TrueBrain> waterresistent at best
17:54:02 <TrueBrain> waterproof is a myth in many cases :P
17:54:09 <andythenorth> splashproof perhaps?
17:54:39 <Rubidium> they are pretty water proof. At the slightest proof of water your warranty is void
17:54:43 <andythenorth> having seen the insides, I suspect the keyboard might be the weak spot for splashproofness
17:55:18 <TrueBrain> I know the latest models have a pretty clever solution there
17:55:22 <TrueBrain> you can spill water on it
17:55:24 <TrueBrain> and it will still work
17:56:22 <andythenorth> well my toddler is testing it for me anyway
18:03:31 *** Eddi|zuHause has joined #openttd
18:20:20 <andythenorth> it's not really one thing yet
18:20:53 <andythenorth> markers are a nice idea, but probably aren't NoCo
18:22:33 <Eddi|zuHause> might simply have the GameScript read signs on startup (and immediately delete them)
18:23:10 <Alberth> like some Z person has already implemented? :)
18:23:34 <Eddi|zuHause> i thought the idea sounded familiar :)
18:24:52 <Zuu_> Some Z person even packed that as a lib.
18:24:59 <Eddi|zuHause> anyway... had to downgrade the graphics driver (either they really screwed up the release, or they removed support for fairly recent hardware already) and had to downgrade kernel, because the old driver wouldn't build
18:26:10 <Eddi|zuHause> now i have to either edit the menu.lst, or remember that second part the next time i boot :p
18:26:52 <andythenorth> there were some good ideas today / yesterday, did I miss any from that NoCo list?
18:27:00 <andythenorth> more is better, we can edit later
18:27:31 *** TheDude has joined #openttd
18:27:52 <Eddi|zuHause> the thing about describing the industry type with a few universal properties
18:28:04 <Eddi|zuHause> which the game script can then use for placement rules
18:28:05 <Alberth> andythenorth: newgrf requesting all kinds of things does not make a lot of sense to me, I like the idea expressed by Eddi a few days ago, the game has mechanisms, and newgrfs fill in settings and tweaks to customize behaviour
18:28:18 <Eddi|zuHause> instead the newgrf implementing them on construction callback
18:29:58 <TheDude> Eddi, to the cargo in GS, I mentioned earlier, I admit I was wrong, but what I though is that GS cant set a goal for any cargo, only for cargos with town effect, and that is what I miss there
18:30:28 <andythenorth> Eddi|zuHause: added properties idea
18:30:33 <Alberth> TheDude: what do you mean "goal for a cargo" ?
18:31:09 <Eddi|zuHause> Alberth: "deliver 1000 coal to this town"
18:31:23 <andythenorth> Alberth: 'requesting' is more or less current behaviour so I find it easy to describe :) Can you help describe the alternative idea?
18:31:46 <Alberth> you can do that for any cargo
18:33:12 <TheDude> I mean this SetCargoGoal (TownID town_id, GSCargo::TownEffect towneffect_id, uint32 goal)
18:33:45 <Zuu_> But you then need to manually grow the town. If the goal is men fo town growth.
18:34:07 <Zuu_> If.using ra carges without town effect.
18:34:28 <Alberth> andythenorth: a newgrf cannot request opening nor decide closing, as it has no overview, it can just provide information to the game to allow it to make good decisions
18:34:31 <TrueBrain> Zuu_: I think the line is bad, your english is breaking up :D
18:34:49 <Eddi|zuHause> Zuu_: your keyboard is broken
18:35:08 <Zuu_> Yep. See you soon from a real keyboard. :)
18:35:41 <TrueBrain> TheDude: although GS has a clean API for TownEffects and its growing of a town, which is easy and fast to implement, it also has a much lower level control
18:35:50 <TrueBrain> it does require you to do more work, but then you can make it grow for what-ever reason
18:35:53 <TrueBrain> including any cargo
18:36:06 <TrueBrain> but even noise level, amount of houses, and for all I care: your name or company value
18:36:29 <Alberth> if name == "myname": win!
18:36:36 <andythenorth> Alberth: do you have any ideas for which information is needed?
18:36:37 <TrueBrain> always-grow, but yes :P
18:37:36 <Alberth> andythenorth: information about how to build it (ie terraform stuff), and how happy it is
18:37:44 <Eddi|zuHause> Alberth: that's awfully close to "if (username == admin && password == backdoor)"
18:37:47 <Alberth> seems to be the most interesting candidates
18:37:59 <TrueBrain> Eddi|zuHause: what is wrong with that? :D
18:38:07 <TrueBrain> I once saw it in Open Source software, hidden in de code
18:38:10 <TrueBrain> made me laugh my ass off :D
18:38:23 <Alberth> Eddi|zuHause: "close"? :D
18:38:53 <andythenorth> happiness is a range
18:39:47 <andythenorth> so controlling when industries appear is taken out of scope for newgrf?
18:39:49 <Alberth> perhaps some indications of what other industries it needs, but that may be pretty obvious from the cargoes
18:40:50 <Alberth> as in "can only appear after 19XY?" seems a good form of information to me
18:41:29 <Alberth> oh probability of appearance would be useful too perhaps and the total number of industries
18:41:31 <andythenorth> k, I think 'request' was a bad choice of word
18:41:57 <andythenorth> the problem is that I want to remove 'guarantee' which is what newgrfs currently get wrt closing etc
18:42:33 <andythenorth> spec ensures that a newgrf author can control closing currently
18:42:50 <Alberth> yeah I know, it sucks, badly
18:42:52 <andythenorth> I think that should be delegated to game, and implicitly GS
18:43:46 <andythenorth> which in my simple brain just means changing how result 03 is handled for cb 29 / cb 35
18:43:59 <andythenorth> 'happiness' is a little more sophisticated
18:44:33 <andythenorth> 'can only appear after xyz date' is just cb22 in current form, unchanged
18:44:50 <andythenorth> that includes probability of appearance
18:45:27 *** Fawksie has joined #openttd
18:46:12 <andythenorth> I'll just change 'requests' to 'indicates' :P
18:46:17 <andythenorth> it comes out about the same then
18:46:57 <Alberth> why not express 'indication to open' as a non-zero probability for cb22?
18:47:10 <andythenorth> that's how it works :)
18:47:32 <Alberth> perhaps with control of the total number of industries as well
18:47:45 <andythenorth> that's the missing part
18:47:59 <Alberth> so you can decide whether to expand the number of industries or do it by replacement
18:48:46 <andythenorth> expanding the number of industries sounds like another case for my regular cb to the newgrf level (rather than to individual industries)
18:49:18 <andythenorth> actually, it might just be a flag on cb22 return result?
18:49:38 <andythenorth> allow increase total number of industries
18:49:48 <Alberth> a table like the snowline table is not sufficient?
18:50:32 <Eddi|zuHause> a table is difficult when it should become dynamic (depending on date)
18:51:16 <Alberth> I see it as a function on date, for every year
18:51:33 <Alberth> eg 1 value for each month from 1950 to 2050
18:51:38 <Eddi|zuHause> so it has 5000000 entries?
18:51:55 <Alberth> although you may want to have a more compact notation for that :)
18:52:34 <Eddi|zuHause> no, a table is a bad idea
18:52:40 <Alberth> most newgrfs do nothing the last 4800000 years :)
18:52:59 <andythenorth> newgrf can count the number of industries at any point
18:53:05 <Eddi|zuHause> but there is the 500 years in the future project°!°°!
18:53:27 <Alberth> ok, the last 4300000 years then :D
18:53:38 <andythenorth> this could just be calculated at run time
18:54:13 * andythenorth is favouring letting newgrfs do more emergent stuff and less scripted stuff
18:54:18 <Eddi|zuHause> it's still a bad idea. a callback could do it in 1 range check
18:54:22 <andythenorth> let the GS do the scripted stuff
18:55:19 * andythenorth considers removing all intro dates from FIRS, and starting a GS
18:55:22 <Alberth> Eddi|zuHause: I'd like to separate the number of industries cb from other cbs
18:55:33 <Alberth> whether or not it is a real table is not that important
18:56:17 <Eddi|zuHause> and is it the job of the newgrf to scale by map size, or does the game apply scaling?
18:56:42 <andythenorth> there are some issues with that though
18:56:47 <andythenorth> was thinking about it earlier
18:57:17 <andythenorth> for a set with 50 industries, if the example map size is 256x256, I'd just set each industry to have 1 instance
18:57:18 <Alberth> density and map size are not orthogonal
18:57:20 <andythenorth> which scales stupidly :P
18:58:03 <drush> ladies, gentlemen, and everyone else,
18:58:37 <drush> 3rdparty/squirrel compiles for ps3, next stop: solving tcp.h issues
18:59:03 <Alberth> andythenorth: I can see that a newgrf is in a better psition to compute such a scaling
18:59:11 <michi_cc> andythenorth: NewGRF :) At least aircraft range was implemented this way.
18:59:34 <andythenorth> scaling things by map size is dull
18:59:39 <andythenorth> ask whoever deleted it from FIRS :P
19:00:29 <Alberth> don't implement it, and the game will do it for you :p
19:00:45 <andythenorth> not in the case of clustering ;)
19:00:57 <andythenorth> no matter what map size, you'll get the same number of clusters in nml FIRS :)
19:02:56 <andythenorth> Alberth: wrt industry cbs, what's your aim?
19:03:00 <andythenorth> I missed that comment above
19:03:16 <andythenorth> separate how / which? :)
19:04:45 <Alberth> I see the method to obtain the desired total number of industries as a separate call, not something you can stick to cb22
19:04:58 *** valhallasw has joined #openttd
19:05:19 <Alberth> but given my major involvements with newgrf extensions, I may be wrong here :p
19:06:06 <andythenorth> I think it's something separate
19:06:29 <andythenorth> I am quite favouring one utility cb to the industry newgrf
19:06:51 <Alberth> anyways, I wish all of you a few nice hours of evening, and see you again soon-ish
19:07:25 <andythenorth> utility cb to industry newgrf would be similar to how cb36 works for vehicles
19:07:58 <Eddi|zuHause> it needs to be a global callback
19:08:25 <andythenorth> global = to the newgrf, not the industry type / instance?
19:08:51 <andythenorth> or do you mean to all active industry newgrfs?
19:08:56 <Eddi|zuHause> yes, i think it should just return the total number of industries
19:09:13 <andythenorth> I am +1 to that way of doing it
19:09:22 <Eddi|zuHause> well, global callbacks are chained. the first one to return a valid return value wins
19:10:00 <andythenorth> maybe it should offer various values, e.g. 'fewer', 'more', 'specific number'
19:10:17 <Eddi|zuHause> that doesn't make a lot of sense
19:10:44 <andythenorth> you could count them and offer 'fewer' anyway
19:10:45 <Eddi|zuHause> if you use the callback, you know the exact numbers
19:10:55 <Eddi|zuHause> if you don't use it, well, you don't use it
19:10:59 <andythenorth> trying to provide ways to make the newgrf less specific :P
19:11:05 <andythenorth> hints, not instructions
19:11:41 <andythenorth> I guess newgrf just says "I need this many" and hopes the game or GS agrees
19:19:59 <Fawksie> ning folks, I've just installed 1.2.1 from the Debian repo and I'm not getting any content appearing in the online content downloader
19:20:02 <andythenorth> Eddi|zuHause: labour force <- core game, or strictly newgrf?
19:20:11 <Fawksie> no such problems with 1.2.1 on Windows on the same network
19:20:41 <Eddi|zuHause> andythenorth: no idea.
19:21:03 <andythenorth> core game could use some kind of interesting new feature :)
19:21:29 <Eddi|zuHause> Fawksie: firewall blocking UDP or something?
19:21:33 <Zuu> Fawksie: Have you edited the iptables rules?
19:22:16 <andythenorth> frosch123: was magic cargos your idea for electricity etc?
19:22:33 <Zuu> The defaults should probably be ok, but if you have tighten up the defaults, that may be the problem.
19:23:51 <Fawksie> Eddi|zuHause: none running, and it would be affecting the Windows machine as well if that were the case
19:27:43 <Eddi|zuHause> question: how the hell do i open a windows-16bit program in ida? if i open it as ms-dos, i only get the "must be run in windows" part, and if i try to open it as PE (windows) then it says the file type is not recognized
19:29:22 <frosch123> andythenorth: there was a idea about making houses and industries produce "town effects"
19:29:57 <andythenorth> can we 'simply' make it an automagic cargo system?
19:30:13 <frosch123> i.e. producing town effect not via accepting certain cargos, but via production of specific industries
19:30:40 <andythenorth> town acts as clearing house (single point) for exchanging supplied / accepted cargos like VOLT, _GAS etc
19:30:42 <frosch123> but the idea is not completely thought through
19:30:53 <andythenorth> make a new pool for these cargos
19:31:08 <andythenorth> put a flag on the cargo definition to put it in this pool, not the regular pool
19:31:19 <Eddi|zuHause> andythenorth: for starters, you should stop calling them "cargos" :)
19:31:26 <andythenorth> allow industries to produce / require these cargos via normal means
19:31:38 <andythenorth> I want them to be cargos
19:31:47 <andythenorth> then just use the existing machinery for produce / accept etc
19:31:59 <andythenorth> houses can accept / produce them
19:32:10 <frosch123> the plan was somewhat to mark cargos as town effects
19:32:13 <andythenorth> town can require certain levels for growth (positive or negative)
19:32:20 <frosch123> so, they would not be delivered to stations, but directly to the town
19:32:25 <andythenorth> frosch123: and deliver via vehicles, or magic?
19:32:39 <frosch123> vehicles transport cargos
19:33:27 <frosch123> cargos can be transformed into town effects by either being accepted by houses stating to process certain cargos into towneffects, or by industries producing town effects as output
19:33:52 <michi_cc> Eddi|zuHause: 16-bit windows exe are NE (New Executable) format, maybe IDA offers that somewhere.
19:34:00 <frosch123> towneffects cannot be transported, but appear immediately at the town
19:34:22 <frosch123> though, stuff might check towneffects provided in the past
19:35:30 *** drac_boy has joined #openttd
19:36:12 <Eddi|zuHause> michi_cc: it only has NE device drivers, not executables...
19:36:28 <andythenorth> frosch123: I think it comes to about the same as the thing I'm thinking of
19:36:45 <andythenorth> I'm only calling it cargos because I'm familiar with how those work
19:37:03 <michi_cc> It has? Because device drivers (VxD) were usually LE (linear executable, originally used by OS/2) format.
19:37:18 <andythenorth> Would houses be able to generate town effects directly? You wouldn't turn houses into mini-industries requiring input to get output?
19:38:04 <Eddi|zuHause> michi_cc: it says "PE/LE/NE Device Drivers"
19:38:12 *** Phoenix_the_II has joined #openttd
19:46:40 <frosch123> andythenorth: houses have no state
19:46:52 <frosch123> so they have to produce towneffect immediately on acceptance
19:47:36 <frosch123> though i am not sure who shall specifiy what cargos turn into what towneffects
19:47:38 <andythenorth> can they produce on any other loop (i.e. nothing arrived)
19:47:40 <frosch123> the house, or the cargo :)
19:47:56 <andythenorth> via a horrible mapping :P
19:47:57 <frosch123> houses can produce cargos
19:48:04 <frosch123> so they would also be able to produce towneffects
19:48:16 <andythenorth> they can produce n cargos via that cb that loops n times?
19:49:02 <andythenorth> the only problem I see with my idea is having houses consume directly
19:49:22 <andythenorth> I wanted a loop to run which sums total demand, and total supply for each effect
19:49:31 <andythenorth> and then nets the result
19:49:43 <andythenorth> I guess town can kind of do that anyway :P
19:49:48 <Eddi|zuHause> frosch123: the cargo defines a town effect, and the industry may choose to issue that town effect or not. houses always issue town effect.
19:50:31 <frosch123> Eddi|zuHause: industry should not produce towneffecty by accepting certain cargotypes
19:50:32 <andythenorth> leaves a lot of control with the industry newgrf no?
19:50:50 <Eddi|zuHause> yes, during production callback
19:51:00 <frosch123> just as output cargos
19:51:00 <Eddi|zuHause> industries will not issue town effect by default
19:51:20 <frosch123> the industry window shall just say: this industry produced electricity
19:51:34 <Eddi|zuHause> industry can produce town effect from totally unrelated cargos as well
19:51:51 <andythenorth> I was trying to figure out if this could also handle 'workforce', but I think it can't
19:51:54 <Eddi|zuHause> or from no cargo at all
19:52:36 <frosch123> just like a industry can produce cargos out of other cargos, or without anything
19:52:44 <Eddi|zuHause> andythenorth: town effect "jobs"?
19:53:13 <andythenorth> industry needs to consume it
19:53:24 <Eddi|zuHause> no, the industry produces "jobs"
19:53:25 <andythenorth> tying it to actual cargo confuses me
19:53:38 <Eddi|zuHause> no cargo needs town effect "jobs"
19:53:45 <andythenorth> so how is it defined?
19:53:45 <Eddi|zuHause> because the industry can make it up on the fly
19:53:54 <andythenorth> town effect is defined by cargos
19:54:07 <Eddi|zuHause> no, town effect is defined by town effect (new feature)
19:54:15 <Eddi|zuHause> a cargo has an associated town effect
19:54:21 <Eddi|zuHause> but it does not define it
19:55:27 <andythenorth> that makes a little more sense to me
19:56:47 <Eddi|zuHause> alternatively, the town effect is defined implicitly. i.e. cargo has a property with a town effect, and industry has a property with one or more town effects
19:57:14 <Eddi|zuHause> and just by setting that property, the town effect will be defined (if not defined before)
19:57:34 <Eddi|zuHause> then the industry can not produce any other town effects than the ones described
19:57:55 <Eddi|zuHause> but that sounds like a bad idea
19:59:10 <andythenorth> the main issue for me is how to ensure cooperation
19:59:27 <andythenorth> across house newgrf, town control newgrf, industry newgrf, GS
20:03:14 <Eddi|zuHause> __ln__: maybe it's a secret military base and they fucked up the editing/censoring? :)
20:04:00 <NGC3982> Why do i never learn to not build on trafficed rails.
20:09:16 * drac_boy is a bit confused why you have so many trains?
20:10:45 <drush> that factory is out of reach of the station
20:11:01 <drush> so I think it's a passenger+mail and maybe goods station
20:11:19 <drac_boy> drush..unlikely...thats a pax-only train
20:14:49 <Eddi|zuHause> talking about construcive answers :p
20:15:08 <Eddi|zuHause> drac_boy: "pax" is short for "passengers"
20:15:16 <drush> I can give some opportunities for that
20:15:24 <Eddi|zuHause> (not sure why, honestly)
20:15:39 <drush> neither do I know Eddi, I thought pax is "passenger annex exchange"
20:15:50 <drush> but anyways, here's the opportunity:
20:15:58 <drush> there's a huge fail in cstdio and cstring in the ps3 build
20:16:20 <drush> a lot of stuff is undeclared
20:16:21 <drac_boy> NGC3982 you probably still haven't told us why you have so many trains?
20:16:37 <drush> I'll just make a pastie
20:16:42 <Eddi|zuHause> drush: missing an include file?
20:16:51 <drush> Eddi|zuHause most likely
20:16:54 <NGC3982> drac_boy, drush: I sent trains to the wrong station.
20:17:53 <drac_boy> NGC3982 oh how many trains were actually meant to call to that station? :/
20:25:10 <drush> I can tell there is definitely some fundamental deficiency here
20:25:42 <drush> with regards to c++ libraries
20:26:15 <Eddi|zuHause> you might want to ask that to your library source
20:33:04 *** andythenorth has left #openttd
21:04:20 <drush> I hear this is a network issue
21:04:27 <drush> so for now I'm just going to disable network
21:04:37 <drush> just to see if that thing builds
21:04:50 <Eddi|zuHause> md5 is used for checking consistency of NewGRFs
21:05:07 <Eddi|zuHause> i.e. the file on disk matches with the one expected by the savegame
21:05:22 <Eddi|zuHause> so disabling network likely won't help
21:06:03 <drush> then why did that part require a tcp.h to be imported from newlib?
21:07:52 <drush> wait, not that particular one
21:10:48 <Rubidium> what? md5.cpp requires tcp.h?
21:15:53 <drush> tcp.h was required by os_abstraction.h
21:16:38 <drush> whatever was done there before md5.cpp, it went smooth with ps3toolchain using newlib's netinet, netinet6 and net directories (containing headers ofc)
21:17:21 <drush> but I suspect this was dumb because those are not headers for gameos, so I'll disable network for now
22:24:01 <Eddi|zuHause> was there ever a reason why one can't remove their own statue?
22:45:11 *** Chris_Booth has joined #openttd
22:46:03 <drush> Eddi|zuHause: because the game doesn't want players to deny that they're egoists?
22:52:30 <drush> will need fontconfig, or disable it
22:55:46 <drush> Eddi|zuHause remember those errors I posted?
22:55:58 <drush> gone for some reason. what magic is this?
22:59:03 *** Djohaal has joined #openttd
23:03:06 *** tokai|mdlx has joined #openttd
23:39:36 *** Djohaal_ has joined #openttd
23:40:02 <drush> goddamnit unicode enable doesn't work
23:48:44 <drush> ladies, gentlemen, and others,
23:49:20 <drush> because I'm going to bed now and posts from now may not appear in my scrollback
23:49:27 <drush> thank you and goodnight
continue to next day ⏵
