IRC logs for #openttd on OFTC at 2011-07-29
⏴ go to previous day
00:54:44 *** ccfreak2k has joined #openttd
02:26:40 *** rhaeder1 has joined #openttd
04:12:53 *** Adambean has joined #openttd
05:16:19 *** darkequi has joined #openttd
06:07:20 *** DayDreamer has joined #openttd
06:21:48 <darkequi> one "network" too much ^^
06:29:49 *** Br33z4hSlut5 has joined #openttd
06:33:32 *** Born_Acorn has joined #openttd
06:33:32 *** PierreW has joined #openttd
06:33:32 *** ThaAmazonous has joined #openttd
06:38:21 *** sla_ro|master has joined #openttd
06:47:49 <dihedral> <darkequi> omg dihedral :D <- friends may usually leave out the 'god' part :-D
07:10:04 *** TWerkhoven has joined #openttd
07:15:08 <dihedral> and a planetmaker, too - how nice ;-)
07:15:27 <dihedral> looks like this day can start now :-D
07:16:30 * TWerkhoven starts the clock :p
08:02:19 *** Chillosophy has joined #openttd
08:17:48 *** RandomGuest1 has joined #openttd
08:18:48 *** Alberth has joined #openttd
08:18:49 *** ChanServ sets mode: +o Alberth
08:23:57 <DorpsGek> Alberth: anyone was last seen in #openttd 29 weeks, 1 day, 15 hours, 56 minutes, and 26 seconds ago: <anyone> hi
08:26:18 <Terkhen> oh, so he was actually here once? :P
08:29:01 <Alberth> at least once, and since 'hi' is the last thing he said, I doubt he had a long discussion about feature he works on :p
08:29:35 <DorpsGek> __ln__: I have not seen nobody.
08:29:47 <__ln__> well, that guy wasn't here yet then.
08:31:00 <Pikka> I have not seen nobody neither
08:41:33 <DorpsGek> planetmaker: someone was last seen in #openttd 40 weeks, 5 days, 15 hours, 1 minute, and 10 seconds ago: <Someone> indeed
08:41:50 <Terkhen> RandomGuest1: coding, offtopic talk, the usual
08:42:07 <RandomGuest1> o nice...im playing OpenTTD as usual, still trying to grasp it -,-
08:42:15 *** Vikthor has joined #openttd
08:48:19 *** Brianetta has joined #openttd
08:54:10 <dihedral> good luck with that :-P
08:56:10 <RandomGuest1> good luck? you need like an engeneering diploma to win at the game...
08:57:25 <planetmaker> economically it's a dead simple game...
09:01:02 <RandomGuest1> no, im pretty sure you need an engeenering degree. I tried implementing simple signals and it makes one train goes to the wrong station, followed by a crash...
09:01:37 <Noldo_> there were trains in the juction while you were changing it?
09:02:01 <RandomGuest1> no....the train just goes impossibly lost and went to another train's depot station...
09:03:10 <RandomGuest1> so now i build bridges over railway tracks =D
09:04:09 <Alberth> you can only have train crash by messing with the signals while they are used, or by using the 'ignore red signal' buttons afaik
09:04:44 <RandomGuest1> >.> ok...so i did
09:05:29 <Alberth> and for depots, yes at times they head for the nearest one they can find, so you must make sure that they can get back to the right track again
09:05:57 <Alberth> alternatively, you can add depot orders in the order list, and they won't go to some random depot any more
09:09:27 <Alberth> A common error with signals is that people try too advanced things too fast. Try to keep it as simple as possible until you understand it fully, only then try something new.
09:10:57 <Alberth> for example, only use the one-way path signal or only use the basic block signal
10:02:11 *** Chris_Booth has joined #openttd
10:20:19 *** Progman has joined #openttd
10:43:49 *** Eddi|zuHause has joined #openttd
10:44:19 <Eddi|zuHause> sometimes server authentication fails...
10:45:02 <Alberth> I never have that with oftc
10:46:56 <Eddi|zuHause> it may be a random server in the rotation that i haven't accepted the certificate yet...
10:47:41 <Eddi|zuHause> annoyingly, Qt has a bug where the entire program crashes when the connection timed out while the confirmation window is open...
10:50:24 <Alberth> that's better than 'be', where creating of some data files needed for its operation already fails :p
10:52:35 <krinn> when having a tile flatten at height 4, slope = 0 then, complement slope return 15, but AITile.LowerTile refuse to lower it if tiles next to it are also flat: worst it does report ERR_NONE, and if i just by hand remove one corner, bingo it works and lower the tile
10:53:04 <krinn> shoudln't the AITile.LowerTile lower my tile even it was flatten at first when i use 15 as slope target
10:54:50 <Alberth> so you lower at four positions at the same time?
10:55:01 <Alberth> I'd try just one position
10:55:38 <Alberth> (which I think is the primitive operation)
10:55:40 <krinn> it work if by hand i remove any corner, then it success and get out of my loop (loop that wait a failure or height=what i'm trying to get)
10:56:13 <Alberth> by hand i remove any corner <-- but that's not 4 at the same time
10:56:18 <krinn> so you mean i should remove one corner when i see a flat tile, then redo it with 15 ?
10:56:41 <krinn> nope my ai trying remove the 4 at same time, but keep looping as it return success but doesn't do it
10:57:03 <krinn> i remove one by hand to get my ai out of the loop and the ai is then able to lower it
10:57:14 <Alberth> but by hand you do a different operation, so you cannot compare it
10:57:34 <krinn> ok, so let's not speak about my helping hand ^^ if i don't: ai keep in the loop
10:58:15 <Alberth> I think you can just change a single corner at a time. Did you try that?
10:58:47 <krinn> no, and this should work, but i was thinking i can just gave the compslope to the function as it wish a slope
10:58:48 <Alberth> I can see doing more corners at the same time would be useful though
10:59:16 <krinn> well, you end lower a flatten tile to a flatten tile just one height bellow
11:00:00 <krinn> doc should then be change to disable that
11:01:01 <krinn> doc say i could SLOPE_N | SLOPE_W (=SLOPE_NW), nothing about asking SLOPE_N | SLOPE_W... (all) will return success but fail to do it
11:01:40 <krinn> because that's what complement slope of 0 is returning and in this case the function fail but report success
11:02:32 <Alberth> that sounds like a bug
11:02:58 <krinn> yep, if i use 1 or 3 corners -> working, using 4 -> not working and report ERR_NONE
11:05:25 <krinn> could provide save game & my dirty code trying to do it as sample
11:06:56 <krinn> not a bug easy to see, as noone will try lower a flatten tile surround by other flatten tiles
11:10:05 <Alberth> do you have a small test AI? please submit a bug for this problem
11:10:51 <krinn> erf, not a small one, i suppose i can do one for the purpose
11:17:35 <krinn> * This can only be true, if multiple corners of the start-tile are terraformed (i.e. the terraforming is done by towns/industries etc.).
11:17:35 <krinn> * In this case the terraforming should fail. (Don't know why.)
11:18:26 <krinn> from terraform.cpp someone note it will fail, only he note town/industry will be affect and didn't think ai can also influence all corners at the same time
11:19:05 <Alberth> that note is likely from years before the AI framework
11:19:25 <krinn> yep, but it now create trouble with ai
11:19:39 <krinn> this should be fix, or doc should disallow 4 corners at same time
11:20:25 <krinn> he says about multiple corners, where i see it fail only if all corners are choose
11:21:49 *** Chris_Booth_ has joined #openttd
11:22:08 <Alberth> yeah, something should be changed, but I don't know what exactly
11:22:20 <krinn> the easier will be the doc
11:22:55 <Alberth> but that complicates writing an AI, so imo that is the last option
11:23:35 <krinn> and even noone see it, i suppose town keep failing on it too
11:24:55 <krinn> the test also fail and doesn't return the CMD_ERROR
11:25:32 <Alberth> possibly. I don't know whether they would terraform flat tiles, it does not sound like something that would happen often
11:25:53 <krinn> i don't know why a town would do that yes
11:26:46 <krinn> for me it's easy: take a 16 squares : add 4 tiles in middle at height 3, all others at height 2: now even flat, you may wish to lower them so all the 16 will be at height 2
11:27:16 *** Chris_Booth_ is now known as Chris_Booth
11:28:17 <krinn> i might success doing it by choosing to handle minheight==maxheight last, so i will remove corners ones first (with minheight=2 maxheight=3)
11:42:55 *** Chruker has joined #openttd
12:12:57 <krinn> i've done a map+ai to see the bug if anyone wish it
12:20:55 <Alberth> nice, such things make reproducing the bug, and checking a fix MUCH easier
12:21:28 <krinn> i'm filebugging it with the files
12:28:44 <Zuu> krinn: Good work with ironing out bugs in the terraforming API. I don't know how much people have tried to do terraforming before. Its only when the API is being used that some bugs are found and fixed.
12:29:38 <krinn> lol it depend, i would prefer not fall on them
12:31:31 *** Devroush has joined #openttd
12:32:22 <krinn> this doesn't really make sens to lower a flatten tile, only under certain conditions, so maybe many have use the function but never meet the conditions
12:34:28 <Alberth> or the code fails at some further point, and they never bother to find out why
12:34:59 <krinn> hihi, like the comment in terraform.cpp
13:22:14 *** Kurimus has joined #openttd
13:23:52 *** sla_ro|master has joined #openttd
13:55:22 *** Adambean has joined #openttd
14:12:50 *** Prof_Frink has joined #openttd
14:13:09 *** seandasheep has joined #openttd
14:23:10 <Eddi|zuHause> feature request: the transparency for loading indicators should also trigger transparency of effect vehicles (smoke)
14:24:14 <dihedral> smoke, sparkes, shadows..... wait a second! :-P
14:25:21 <Eddi|zuHause> no really, i can't align steam engine sprites while they are covered in steam :p
14:26:27 <Eddi|zuHause> is it playable yet? ;)
14:26:41 <__ln__> is it screenshottable yet?
14:27:46 <krinn> when a new version will be release ? (lol kidding)
14:28:38 <TrueBrain> Eddi|zuHause: it has been for a while, but yes, very much so
14:31:44 <Alberth> krinn: at the release date of 0.6
14:36:42 <__ln__> what's the point in the game? (i never played the original nor do i know it)
14:38:02 <Eddi|zuHause> it's basically the predecessor of the command&conquer-type games
14:39:39 <TrueBrain> it is one of the very first RTSes
14:39:43 <Pikka> but you still need the old data files, right?
14:39:53 <TrueBrain> yup; like OpenTTD used to be ;)
14:40:05 <Pikka> are the harvesters still magnificently braindead?
14:40:07 <TrueBrain> we need artists to make us an OpenGFX :P
14:41:58 * __ln__ is currently on page 350 of the Dune trilogy
14:43:27 <Pikka> it's been about 15 years, but aren't there about 7 books?
14:43:39 <TrueBrain> guess I should change DNS for libemu or something
14:43:46 <Pikka> about 15 years since I read them, that is
14:45:40 *** Usuario has joined #openttd
14:46:37 <Pikka> anyhoo, goodnight peeps
14:51:19 <krinn> but contrary to dune, openttd never get out in 320x240 :p
14:54:58 * Rubidium ponders whether to refute that
14:57:11 <Rubidium> at least 2 releases with 320x240 as main resolution were made
14:57:27 <Rubidium> and at least two with 240x320 (if the wiki's correct, ofcourse)
14:57:44 <Eddi|zuHause> depends on what you call a release :p
15:00:46 <Rubidium> someone made a fork and releasesd that
15:00:46 <krinn> looking then @ dune gfx is scary
15:03:24 *** Biolunar has joined #openttd
15:14:47 <peter1138> so is there a 32bpp project full of conflicting images and extra zoom crud?
15:15:05 *** askdlj5 has joined #openttd
15:15:23 <askdlj5> Hello :) Can anybody advice please a competitive trains AI?
15:15:39 <planetmaker> nocab, admiralAI, choochoo
15:15:51 <planetmaker> simpleAI and AIAI
15:16:08 <askdlj5> wow, so fast :) Thanks a lot!
15:16:49 <Alberth> the faster the response, the more you have to pay
15:28:28 *** LordAro has joined #openttd
15:44:37 <Eddi|zuHause> orudge: are you aware that wiki.ttdpatch.net gives a 403? and maybe introduce a redirect to tt-wiki.net?
15:50:57 <orudge> Eddi|zuHause: yes, it was being hammered by spambots
15:51:03 <orudge> I'll put the site back up in a bit
15:55:35 *** supermop has joined #openttd
16:36:25 *** douknoukem has joined #openttd
16:52:33 <CIA-2> OpenTTD: rubidium * r22692 /trunk/os/windows/installer/install.nsi: -Update: [Windows] Installer didn't mention 7 all the time
17:00:22 <planetmaker> apropos wiki, orudge: any chance for an offline version of the newgrf specs wiki?
17:01:46 <Eddi|zuHause> we did discuss that a few days ago, the script possibly needs some adjustments
17:02:10 <Eddi|zuHause> because of serving three different wikis from the same installation, or so.
17:04:04 <orudge> yes, I have a script, just need to adjust it
17:25:25 *** frosch123 has joined #openttd
17:51:01 <flitz> hi, somebody who can quickly answer this: when a vehicle gets its cargo layout changed, it is immediately redrawn in the depot, which call informs the depot-window to redraw ? I tried OnInvalidateData but it didn't work
17:52:48 <flitz> ah, I found it: SetWindowDirty(WC_VEHICLE_DEPOT, front->tile); sry for asking :>
17:53:12 <flitz> didn't know that you can set a window dirty by knowing only its identifier, good to know
17:53:59 <flitz> (so far I worked with passing bool pointers along to child windows and use them to notify their parents of changes)
17:54:59 <Rubidium> in any case, calling On* functions on windows is in 99% of the cases wrong; those On.. functions get called by the framework when they need to be triggered
17:55:39 <Rubidium> e.g. OnInvalidateData should rather be InvalidateWindowData(class, index, data) (or something like that)
17:55:49 <Rubidium> there's a method for invalidating all windows of a class as well
17:57:43 <flitz> hm yes: "void SetWindowClassesDirty ( WindowClass cls )" this looks like what I need
17:58:37 <Alberth> you can also make a single widget dirty :)
17:58:43 <Rubidium> but it's better to dirty as little as needed
18:01:39 <flitz> the single-dirtiing requires the vehicle's tile in my case, I'm not sure if its safe to use it if my vehicle's tile is always 0
18:02:51 <Alberth> normally, you'd move the widget enum to a .h file and use a value from that
18:03:16 <planetmaker> flitz: the you could dirty the widget... you know that number
18:05:49 <flitz> hm, what do you mean ? What I was using for the window descriptor was it's class, this is written in the enum in window.h, yes. but what would be the window number ?
18:06:37 *** Chris_Booth has joined #openttd
18:06:49 <Alberth> widget number != window number
18:07:49 <Alberth> see vehicle_gui.h with a enum used at many places through SetWindowWidgetDirty()
18:09:30 <flitz> ah this you mean, but I need the function as cross-window version so I used SetWindowDirty()
18:11:18 <flitz> to be precise: cmdrefitvehicle does a call 'SetWindowDirty(WC_VEHICLE_DEPOT, front->tile)' I need to do the same, but without having a correct tile for front (in case front is virtual), so I invalidate the whole window class instead if front is virtual
18:15:55 <Eddi|zuHause> well, "front->tile" is the window number, so you just need to replace that with the number of your "virtual" depot
18:16:23 <Eddi|zuHause> i don't think number 0 is used anywhere.
18:16:37 <Eddi|zuHause> 0 is a corner tile that can never have a depot on it
18:16:51 <Alberth> flitz: SetWindowWidgetDirty() is cross-window
18:24:55 <flitz> but again with window number, I will just insert the virtual depot number there for now
18:26:08 <Alberth> yeah, window class + number together is the unique window identification
18:27:17 <flitz> and the number is just in case that there are multiple windows of the same class ?
18:27:29 <Alberth> you can have several depot windows opened, but when clicking at a depot, you need to be able to find the right opened window to highlight it
18:39:31 <Eddi|zuHause> hm... the sprite-sorter gives some bad/unstable results sometimes... vehicles that should be behind another vehicle get drawn on top
18:41:59 <Eddi|zuHause> that was about foundations drawn over vehicles
18:42:15 <Eddi|zuHause> not sure if it's really the same issue
18:42:33 <Alberth> it is also the sprite sorter messing up
18:43:07 <Alberth> although newgrf bounding boxes may also be part of the problem
18:51:38 <frosch123> Eddi|zuHause: vehicles being shorter than 4/8 ?
18:52:37 <frosch123> well, everything will go better some day :p
18:53:44 <Eddi|zuHause> the weird thing is, i have not found a bug in my "GUI extra callback info" patch yet...
18:59:45 <Zuu> How about having a AICONFIG_AI_DEVELOPER flag for ai settings that tell that an AI setting will only be visible if the user have the AI developer tools active?
19:00:10 <Zuu> (I just started to poke with the code)
19:01:39 <Zuu> So that all debug settings can be hidden from normal users.
19:02:30 <Alberth> you mean a compile-time flag?
19:03:05 <Alberth> so we have 2 binaries instead of one?
19:04:08 <Alberth> and no way to stop users from downloading the debug version :p
19:12:22 <Zuu> not a compile time flag, but a flag just as AICONFIG_INGAME but that controls visibility instead of ability to edit the setting.
19:16:33 *** JVassie has joined #openttd
19:17:26 <Alberth> _settings_client.gui.ai_developer_tools a flag like that?
19:20:12 <Zuu> A setting with AICONFIG_AI_DEVELOPER would only be visible if also _settings_clinte.gui.ai_developer_tools is true.
19:21:40 <Alberth> so a AI developer has to change his source code to debug
19:21:58 <Alberth> and then also set that ai_developer flag in openttd?
19:22:42 <Alberth> I don't understand why you want to have a second flag for controlling output
19:24:14 <krinn> Debug on/off will only be visible if AICONFIG_AI_DEVELOPER is set to 1
19:24:44 <krinn> and "normal" user will still see other settings but not "debug" ones (like UseBus on/off) ...
19:25:32 <Alberth> and how is that different from _settings_clinte.gui.ai_developer_tools ?
19:26:06 <krinn> with that on/off it just enable the gui, but ai still will show debug option even off
19:26:25 <krinn> (if i get it right of what zuu was saying)
19:26:55 <Alberth> so you want that flag to be available in the AI ?
19:27:48 <Zuu> Each AI can add settings that they find suitable for debug/development but do not want to clutter the interface for regular users with.
19:27:56 <krinn> bof -> the french yes/no/don't care
19:28:38 <Zuu> I recently added a setting to CluelessPlus to enable/disable clocking of its performance. That's a setting that I don't think anyone except someone knowing what to observe in the AI debug window will have use for.
19:29:39 <krinn> i see what you mean zuu, but after adding many taunt in the debug ai windows, i realize that windows is just never use by a "normal" user ^^
19:30:30 <Zuu> I could of course instead try to write a release-script that disable these settings, but that would be a bit fragile.
19:30:43 <krinn> so you can even insult players in it, except curious and devs, nobody will likely see it (lol except when the ai crash)
19:31:27 <krinn> i have a debug setting because enable it, not only enable debug message, but eat more ticks for debug purpose
19:31:41 <Zuu> I sometimes take notice if someone has made the logs look pretty, but duno really how important that is other than for your own understanding :-)
19:32:31 <Alberth> ah, I see what you are doing now. yes that could be useful
19:34:26 * frosch123 still does not understand how that is different to _settings_client.gui.ai_developer_tools
19:34:42 <frosch123> or is the problem that you cannot check settings in the info.nut?
19:35:15 <krinn> no it's a problem everyone see your debug settings if you put them in the info.nut to enable/disable them for debug
19:35:45 <frosch123> well, but if you could check the setting in info.nut, you could hide some settings
19:35:52 <krinn> zuu was tied their visibility to gui.ai.developper_tools == 1 else you'll only see normal ai settings
19:36:00 *** douknoukem has joined #openttd
19:36:21 <krinn> we couldn't hide info.nut after the ai is init
19:37:05 <krinn> but we could indeed tied the return value already to gui.dev setting and disable what a user tried to enable ^^
19:37:37 <krinn> still if i get it right, zuu want fully remove the button
19:38:47 <DorpsGek> frosch123: Hyronymus was last seen in #openttd 1 week, 6 days, 10 hours, 22 minutes, and 45 seconds ago: <Hyronymus> if you do the latter, people can always suggest changes
19:40:27 <frosch123> krinn: yes, but wouldn't that be able, if the settings could be accessed from info.nut?
19:44:48 <krinn> as it is now, if you click on configuration for on ai, you may see three boolean "bus" "debug" "train"
19:44:59 <krinn> what zuu wish is normal user see "bus" "train"
19:45:15 <krinn> and with gui.debug.... the 3 display
19:46:01 <krinn> for now, we could disable it if a user enable it, but we cannot remove it from the configure part, as the info.nut is use to init the ai
19:46:40 <frosch123> somehow we are talking in circles
19:46:52 <frosch123> is there no "yes" or "no" to my question
19:47:19 <krinn> i didn't really understood it, doesn't make sens to me
19:49:36 <frosch123> info.nut defines the settings. so if you could do something like 'if (AIGameSettings.GetValue("gui.ai.developper_tools")) { AddSetting(...' you could conditionnaly hide settings
19:49:57 <krinn> that's what zuu was wishing yes
19:50:06 <frosch123> otoh, there is already a 'flags' parameter to it
19:50:18 <frosch123> not sure, whether checking other settings makes sense
19:50:37 <frosch123> as it would make configuring the game harder, if you first set ai settings and then change other game settings
19:53:54 <krinn> dunno if zuu will make it, for me, i would be resume to sed the line with settings to add // in front before packaging the ai
19:54:44 <krinn> and even if the option is visible, i'm not sure users are scared when seeing you can enable some debug
20:28:13 <planetmaker> (or a new one for that matter, dunno ;-) )
20:31:20 <Zuu> Will probably happen whenever I get to do rail.
20:31:50 <Zuu> Unless someone decide to make a standalone libarary for that which I might just use.
20:31:55 <krinn> what is special about that ukrs train ?
20:31:56 <planetmaker> Do you have any idea how much your libs are used?
20:32:27 <Zuu> planetmaker: Not exactly. I can't see which AIs use it has dependencies in the bananas manager.
20:33:28 <Zuu> But I have four versions of it in my OpenTTD online content, saying that there are four banans-AIs that use different versions. That said, that could very well be four of my AIs :-)
20:34:19 <Zuu> I know that AIAI uses the code that I wrote for replacing rail crossings with bridges.
20:34:25 <krinn> imo you should have made small libs and not a big one
20:34:40 <Zuu> small libs can't have circular dependencies
20:34:43 <krinn> your superlib looks like glibc :)
20:35:02 <planetmaker> ah, that's the reason, I just wondered about that, too ;-)
20:35:38 <Zuu> Parts of it could be separated, but I've been lazy and just stacked everything there as well :-)
20:38:34 <krinn> say i need your tile lib
20:38:45 <krinn> i will get all the bloat for using tiles
20:38:48 <Zuu> Then, only the tile lib need to be imported.
20:39:00 <krinn> i think people would have like better small libs aim at 1 task
20:39:12 <Zuu> All other parts will reside in SuperLib.*
20:39:14 <krinn> else they will just rip tile.nut and look for the function they seek no?
20:39:51 <Zuu> IIRC Tile, Direction and Helper are inter-dependent.
20:41:22 <krinn> you might have break them into small part
20:41:41 <krinn> it's easier for you because reusing tool in direction to handle tile, looks logic i know
20:41:49 <Zuu> I tried that route, but gave up. See the SuperLib topic for more details.
20:42:14 <krinn> but for user, you're adding direction where they might just wish AITile.IsAStationHere() function (lol fake func name)
20:44:18 <planetmaker> as sub libs can be imported... I see no big issue
20:45:03 <krinn> for starters i think it look complex to use it
20:45:44 <Zuu> The library exist mainly to save me from maintaining the same code in more than one place. As a secondary benefit, other AI authors can load the library and use selected parts if they wish. I have put some effort into not poluting the global scope more than necessary but the route to split it into many separate libs ended up unfeasible as I soon needed to duplicate some functions in order to make the libs non-cyclck, violating 1).
20:46:12 <Zuu> 1) = not maintain the same code twice
20:47:06 <Zuu> as planetmaker said, it uses sub libraries, and you can choose to just import eg. SuperLib.Tile and then only use that.
20:48:14 <krinn> i'm not against it zuu, or saying it's bad, i'm just telling you the feeling it gave me, bloat & complex, that's what i was feeling when i start building an ai
20:49:41 <Zuu> I can see that, and not very much that can be done against it looking big or complex. It _is_ big. Heck, now it even contains high-level functions to build road/air stations in towns :-)
20:50:36 <krinn> can you build airport north of town ?
20:50:37 *** Chris_Booth has joined #openttd
20:51:08 <Zuu> I'm sure not everyone want to give away the control over that, and use their own implementaions for that. It's just a convinient place for me to place code that I might reuse in eg. TutorialAI.
20:51:52 <Rubidium> pff, even the API is big and bloated. After all I doubt no AI uses (nearly) all functionality
20:52:11 <planetmaker> :-) I doubt that, too
20:52:39 <Rubidium> hell, even OpenTTD is extremely bloated
20:52:42 <Eddi|zuHause> "dbg: [sprite] [85] Loading corrupted sprite from /cets at position 1535213" <-- how the hell did that happen?
20:52:48 <Eddi|zuHause> reload fixed that...
20:53:08 <Rubidium> changing grf while running?
20:53:19 <krinn> zuu airport build north of town might exibit another bug i found with IsWithinTownInfluence
20:53:22 <Eddi|zuHause> using reload_newgrfs command
20:54:04 <planetmaker> krinn: only the Northern most tile is tested for these tests
20:54:12 <krinn> testing north tile of an airport might answer false because of the airport lenght, its first tiles might be outside town influence, while lower part of airport within, but you'll get false
20:54:20 <krinn> yep planetmaker that's what i saw
20:54:49 <Zuu> krinn: IIRC what matters is where the staiton sign is.
20:55:09 <krinn> it mater if you try to find to which town the airport is attach and what town it influence
20:55:13 <Zuu> all other tiles of the station sholud be irrelevant.
20:56:10 <krinn> if you then check your airport is well place, you'll keep moving it because it is well place but seems not ^^
20:56:16 <Zuu> I would expect AIStation.GetLocation(id) to return the tile which has the station sign.
20:56:33 <Zuu> The station sign tile, is from where cargo transport distance is calculated among other things.
20:56:40 <krinn> but the station sign might be outside the town influence
20:57:03 <Zuu> Therefore, I would expect that the station influence the town in which the station sign belong.
20:57:21 <krinn> so do i, until i see it's not
20:58:10 <krinn> it's about tiles cargo accept/product > 8
20:58:25 <Zuu> Anyway, I store which town/industry the station is "attached" to in the station name, so I don't have to deal with that (anymore) :-)
20:59:25 <krinn> you maybe, but not your api :)
21:11:30 <krinn> Zuu, how about getting contrib to your lib ?
21:12:22 <planetmaker> is superlib on the DevZone already?
21:13:04 <krinn> like glibc, it sucks to build small programs, but when building a big ai, your lib should show its strengh
21:14:07 <Zuu> I wouldn't mind patches or help, but would need some control over it as it is a fundamental part of my AIs.
21:15:29 <krinn> right now i think you lib evolve as when you seek a function for your ai you add it to the api
21:15:43 <krinn> it would be adding functions you might not use yourself, or not yet
21:16:18 <Zuu> That is something I would welcome in general.
21:16:23 <krinn> but what if all ai maker contribs their own function (because everyone build ones that ease that or this thing)?
21:17:40 <planetmaker> I guess that's where the maintainers decide what's going to be part of the lib and what not ;-)
21:18:11 <Zuu> and if you dissagree you can always fork or create your own lib ;-)
21:18:15 <krinn> :) that's the task he should made, refuse & accept patch
21:19:20 <planetmaker> Though generally the code base might improve if it should become more than a one-person project :-)
21:19:35 <planetmaker> trallala :-) It's up to Zuu, of course
21:20:07 <planetmaker> But I have the feeling that in NoAI-land still a lot of synergies can be found, if people would start writing a joint AI
21:20:48 <planetmaker> that sounds awesome as a name
21:22:37 <planetmaker> and... from my own newgrf experience: it bears a lot more potential for long-term motivation :-)
21:23:30 <krinn> sounds a good idea, you should get an ai that handle everything fast
21:23:34 <Zuu> Indeed, that's part of why my AI projects get more care than my truly one-man-project: Junctioneer.
21:26:33 <Zuu> but joint projects can also end up like Transport Empire ...
21:27:38 *** Brianetta has joined #openttd
21:28:26 <krinn> dunno what is transport empire
21:28:42 <planetmaker> they can certainly. But I don't know its history, only its current state: dead
21:29:02 <planetmaker> it needs the individual people wanting to pull of the thing. At worst also alone
21:30:14 *** HerzogDeXtEr1 has joined #openttd
21:30:18 <Zuu> The problem when I was in that project was that it was IRC meetings, discussions and writing meeting minutes but nothing got decided for real and not much code got written.
21:31:13 <planetmaker> oh, you were part of that? :-O
21:31:27 <Zuu> But we invented a raise-hand-system for the IRC meeting and IIRC someone wrote a bot to help with that as well. :-)
21:32:07 <Zuu> planetmaker: Yep, 5 years ago or so.
21:32:25 <planetmaker> What you describe seems like "all managers, no work-force"
21:32:58 <planetmaker> small projects need people doing stuff - and discussions not killing the incentive
21:33:01 <Zuu> If OpenTTD wouldn't have came, it might have slightly better chances, but the problem was more structural than that.
21:33:15 <krinn> that was looking nice, but the screenshot looks like ogre demo
21:34:07 *** Chris_Booth_ has joined #openttd
21:34:12 <krinn> i always think starting things like that are hard, as everyone have a vision of final product, but none can see it
21:34:47 <krinn> openttd would be easier per example, as everyone see what ttd was and what opentdd should becomes
21:35:05 <krinn> if you can show to people, better have a dictator as maintainer to push the force forward :)
21:35:13 <Zuu> I guess the problem is also that many thinks it has to be perfect from start and not thinking in terms of incremetal improvements.
21:35:55 <Prof_Frink> krinn: OpenTTD *started* with working code.
21:36:10 <krinn> and you can't really see "all improvments" as your screen still could only see code and not a game
21:36:24 <krinn> Prof_Frink, oh, that ease even more the process so
21:36:38 <Zuu> krinn: If you know Swedish there was a realy long article about lude some time ago.
21:37:20 <krinn> i only speak chinese under torture, english badly, french nice, and sometimes gaelic when drunk
21:37:23 <Zuu> (lude is the guy who wrote the "working code")
21:37:49 <Zuu> (what he did was to port the asembler code to C)
21:38:26 <krinn> by assembler i suppose you mean reverse eng the code ?
21:39:19 *** Chris_Booth_ is now known as Chris_Booth
21:39:43 <krinn> eheh, wasn't he a crazy dictator ? and got his balls kick when project was enough mature that everyone start to think it might work ,
21:39:51 <Zuu> He is now starting to get famous for uTorrent and Spotify.
21:40:57 <krinn> not really related to game
21:44:12 <krinn> ah scummVM that speak to me
21:46:43 *** seandasheep has joined #openttd
21:47:31 <krinn> woo, he manage to reverse & convert it in a year
21:48:31 <Eddi|zuHause> the way i heard it was that the conversion was done pretty quickly, and then he had it lying around for a year because he was unsure whether to release it
21:52:47 <krinn> that's look impressive work for one man
21:54:47 <krinn> lol it's sweet to read wikipedia about openttd
21:55:46 <krinn> looks at all the feature ottd have, it's impressive, i know them, but never seen all of them makes a big project and improvments
21:56:34 <krinn> ok guys it's 0:00 time to go home for me :)
21:56:39 <frosch123> ottd 0.1 already had a slightly different string system than ttd as far as i could figure out
22:08:40 *** Chillosophy has joined #openttd
22:23:27 <CIA-2> OpenTTD: frosch * r22693 /trunk/src/ (5 files in 2 dirs): -Fix [FS#4691]: The override managers were not reset in some cases like creating a new scenario.
23:16:12 *** mikegrb_ is now known as mikegrb
continue to next day ⏵