IRC logs for #openttd on OFTC at 2011-08-01
⏴ go to previous day
01:16:17 <DorpsGek> Eddi|zuHause: 1.13698630137
01:18:41 <Eddi|zuHause> @calc 1400*83/73
01:18:41 <DorpsGek> Eddi|zuHause: 1591.78082192
01:20:47 <Eddi|zuHause> @calc 1400*1.137
01:20:47 <DorpsGek> Eddi|zuHause: 1591.8
02:26:50 *** rhaeder1 has joined #openttd
04:04:30 *** supermop has joined #openttd
05:39:13 *** sla_ro|master has joined #openttd
05:56:53 *** fjunike has joined #openttd
05:58:48 <fjunike> good morning. can i get the group_id having the group_name?
06:02:36 *** Br33z4hSlut5 has joined #openttd
06:33:00 *** Cybertinus has joined #openttd
06:35:01 *** DayDreamer has joined #openttd
07:02:27 *** Alberth has joined #openttd
07:02:27 *** ChanServ sets mode: +o Alberth
07:47:08 *** Biolunar has joined #openttd
08:01:10 *** Vikthor has joined #openttd
08:37:12 *** Eddi|zuHause has joined #openttd
09:20:33 <planetmaker> Zuu: ad 1) is fine. Ad 2) could rather wait for a review of all transparency options
09:23:30 <planetmaker> Maybe it could rather be added to the list of things shown in the options menu
09:23:46 <planetmaker> hm... are you missing the changes to english language file?
09:28:45 <planetmaker> no. No GUI so far at all :-)
09:39:57 <Zuu> I guess it could get an option in the options menu in the mean time while waiting for the transparency options review (if it ever is gonig to happen)
09:43:09 <Zuu> One could think about generalizing the hiding of other companies to other things than just signs and have it in the transparency window. That said, I will not attempt on that for now.
09:57:10 <planetmaker> though I'm not too convinced about hiding other company stuff than signs
10:01:45 *** Adambean has joined #openttd
10:31:21 *** Chris_Booth has joined #openttd
10:52:45 <Alberth> Good point, I often experience that problem with industries. Ah, I can go this and this way (decided from a partially visible set of industries), then you start laying tracks, and you find many other industries are in the way
10:54:48 <Eddi|zuHause> reminds me of a forum thread that went like "why is my rating totally screwed up, when i only modified empty terrain?" - "Man, the whole map is full of trees!"
10:56:54 <planetmaker> We obviously need the "all mighty transparency button": hide everything ;-)
11:03:12 <Alberth> Noldo_: WHITE (at least if you believe current applications)
11:04:03 <Alberth> someone: any progress on 'cartoon-style' graphics of simutrans?
11:13:04 <planetmaker> Alberth, graphics work is mostly missing
11:13:30 <planetmaker> Like the construction stages need making
11:13:32 <Alberth> pity, it is such a nice idea
11:13:41 <planetmaker> And more houses need conversion from simutrans to OpenTTD
11:13:58 <planetmaker> and now I'm waiting for nml-ification :-P
11:28:30 <Zuu> How about renaming "show only own signs and names" to "hide competitor signs and names", as the option will still show buyos and oilrig stations?
11:29:44 *** Chillosophy has joined #openttd
12:11:31 <Zuu> No objections. I take that as a yes :-)
12:12:35 * Alberth never sees any competitor at all
12:14:29 <Zuu> or should it still be worded in positive words. Eg. "Show competitor signs and names" and be on by default?
12:15:20 * Zuu thinks that a good idea as it keeps more parallellism with the current options in the settings menu.
12:26:07 *** dnicholls has joined #openttd
12:27:15 <planetmaker> sounds better, Zuu
12:28:57 <planetmaker> Zuu, that depends where you add it to the GUI
12:29:03 <planetmaker> and how you interpret the bit
12:29:10 <planetmaker> in the options, texts are positive
12:29:32 <planetmaker> In the adv. settings... I'm not sure
12:29:38 <Zuu> For now I add it in the settings menu where you enable/disable showing of station names, waypoint names etc.
12:29:54 <planetmaker> yes, that's what I meant with 'options' :-)
12:30:03 <planetmaker> to not confuse it with adv. settings :-)
12:30:07 <Zuu> I've moved it there instead of advanced options.
12:30:18 *** TWerkhoven has joined #openttd
12:31:35 *** Brianetta has joined #openttd
12:34:02 *** Kurimus has joined #openttd
12:34:38 <Eddi|zuHause> hm... i can't find a useful formula for "unrealistic prices"
12:36:56 <dnicholls> are there any NFO examples of advanced spritelayouts with register modifiers?
12:38:36 <planetmaker> at least none I know of
12:41:52 <dnicholls> ah that looks useful, thanks
12:43:40 <planetmaker> you do have the last nightly version of grfcodec / nforenum, do you?
12:49:39 <planetmaker> 832 only introduces it
12:50:42 *** equinox has joined #openttd
12:54:42 <dnicholls> so is it not fully supported in grfcodec/renum yet?
12:55:11 <planetmaker> You don't have the latest version of grfcodec / nforenum
12:56:11 <planetmaker> And even then I cannot confidently answer your question :-) Maybe it's not 100% supported yet
12:58:13 <dihedral> anybody to magento performance hints?
12:58:46 <planetmaker> i.e. I'm not sure about station support
12:58:55 <planetmaker> the others most probably work
13:01:10 <dnicholls> ok, I'll see if I can get anywhere with 832, thanks
13:05:30 <planetmaker> depending on where you are and what you already have you might risk a look at NML
13:05:44 <planetmaker> it also supports parametrized spritelayouts
13:13:37 <Zuu> Finally, a new patch for the hide competitors patch has arrived on FS#4701 :-D
13:13:49 <DorpsGek> planetmaker: Error: The command "bugs" is available in the OpenTTD and WT2 plugins. Please specify the plugin whose command you wish to call by using its name as a command before "bugs".
13:18:01 <planetmaker> + case OME_SHOW_COMPETITOR_SIGNS:ToggleBit(_display_opt, DO_SHOW_COMPETITOR_SIGNS), InvalidateWindowClassesData(WC_SIGN_LIST, -1); break; <-- I think there's a typo s/,/;/g @ Zuu
13:18:17 <Eddi|zuHause> i don't suppose i could make buying articulated parts cost anything?
13:18:33 <planetmaker> I guess not without patch
13:18:40 <Zuu> planetmaker: It does compile fine here
13:19:14 <planetmaker> I haven't tested compilation, just reading
13:19:34 <Eddi|zuHause> linear cost factor is annoying, because either you lose resolution on the (far more common) lower end, or you set low limit on the highest possible price
13:19:51 <Zuu> It's the same way of copling togeather several statements in the third part of a for-loop head.
13:20:45 <planetmaker> Zuu, in any case I wonder whether the coding style wouldn't require a multi-line way of writing it there
13:20:52 <Zuu> I though, it would make sense on a one-liner but if not, I can change it to a semicolon.
13:21:26 <planetmaker> Zuu, I'm not the biggest expert around :-), but it looks odd to me
13:23:55 <Zuu> Okay, I'll change it. Actually all other case:es in the switch block are one-liners, but they all only have one statement.
13:25:49 <Alberth> indeed, such use of , is not allowed in normal statements
13:27:05 <planetmaker> there, our coding style expert when we need him :-)
13:28:08 <Zuu> planetmaker: coding style fix/change @ FlySpray
13:28:46 <Alberth> the point of code and its style is to make it as unsurprising as possible, so one can concentrate on what it actually does
13:32:03 <planetmaker> I shall give it a go tonight, Zuu
13:32:28 <planetmaker> I think the idea is undisputed ;-)
13:40:38 *** douknoukem has joined #openttd
13:46:41 *** Progman has joined #openttd
15:02:07 *** Vikthor has joined #openttd
15:19:28 <MNIM> GET IN THE BANANAMOBILE!
15:39:19 *** LordAro has joined #openttd
15:40:56 <LordAro> Alberth: your fix to the invalid pointer crash yesterday...doesn't work :L - i added the char *dest = readme_text in (then readme_text = dest after the for loop), but same error :(
15:49:49 <LordAro> you have 10mins before i have to go :)
15:51:20 <Alberth> merge with trunk failed :(
15:55:04 <Alberth> otherwise, shall I send you a PM (to Lord Aro iirc) ?
15:55:24 <LordAro> yes :) that would be fine
15:55:54 <Alberth> I just pulled an update, my system has to do a full rebuild
15:56:07 <Alberth> which takes 5 minutes or so
15:56:17 <LordAro> i (should|)shall be back in 4 hours though
16:16:57 *** Prof_Frink has joined #openttd
16:22:13 *** Zeknurn has joined #openttd
16:43:22 *** andythenorth has joined #openttd
16:43:47 *** Zeknurn has joined #openttd
16:44:09 <Gniarf> hi there. "the longer distance the cargo travels, the more you get paid." it means I shouldn't make too short routes ?
16:46:12 <Gniarf> now how does it works with bus and passengers ? bus stations can be 10 tiles apart
16:47:22 <planetmaker> they can. But you need not do that
16:47:34 <planetmaker> you could for example use inter-city coaches
16:48:09 <planetmaker> And... just check it out. Usually a bus going small circles in a town will nevertheless make a profit
16:50:04 *** andythenorth has joined #openttd
16:50:04 *** andythenorth has left #openttd
16:50:28 *** andythenorth has joined #openttd
17:16:18 <andythenorth> orudge: definitely can't post to tt-forums from my virgin media connection
17:20:49 <Gniarf> how do passengers behave ? do they want to go to the nearest station ?
17:21:07 <Gniarf> I know I can use orders to trick them
17:21:11 <andythenorth> depends if you have yacd or not
17:23:10 <orudge> andythenorth: but you can browse the forum?
17:23:10 <Gniarf> using default version, no patch (1.0.1, a bit old)
17:23:29 <andythenorth> orudge: yes no problem
17:23:38 <andythenorth> You are using IPv6 (2002:5ce9:3804:0:225:ff:fe48:251b).
17:23:55 <orudge> it would seem there's some kind of an issue with IPv6, and I'm guessing it's at the server end in that case
17:24:02 <orudge> since I (sometimes) have the same issue via IPv6
17:24:09 <andythenorth> let me see if I can submit any other forms on the forum
17:24:09 <orudge> what IPv6 provider are you using?
17:24:16 <andythenorth> I can disable IPv6 on my router
17:24:23 <orudge> no, no, that shouldn't be necessary
17:24:26 *** Chillosophy has joined #openttd
17:24:27 <andythenorth> I have no idea who's providing it currently
17:24:32 <andythenorth> how do I find out?
17:24:38 <Zuu> Gniarf: With long-distance buses, my AI CluelessPlus can easily turn 18 times as much income as runnig costs (at least using the eGRVTS NewGRF)
17:24:42 <orudge> you presumably set it up?
17:24:47 <orudge> I don't think Virgin Media offers IPv6
17:24:54 <orudge> although I may be wrong
17:24:57 * TWerkhoven would like to know if they do
17:24:59 <orudge> it just seems unlikely
17:24:59 <andythenorth> that's interesting
17:25:48 <andythenorth> orudge: I've disabled it in OS X
17:26:02 <orudge> well, generally, I would recommend you keep IPv6 on, because it's awesome and all that
17:26:03 <michi_cc> According to whois that is a Virgin Media allocation.
17:27:09 <michi_cc> Duh, it is of course, because it is a 6to4 address, which is probably the most unreliable IPv6 method.
17:27:43 <orudge> has anyone else here who uses IPv6 had any problems posting to the forums, then?
17:28:28 <Rubidium> michi_cc: using the method of RFC1149 will likely make it even less reliable ;)
17:28:38 <Rubidium> orudge: not that I can remember
17:29:44 <Rubidium> sorry that I violated the 'needs 3 words' forum rule ;)
17:29:58 <orudge> see, if I try to reply to that, I get "the connection was reset"
17:30:05 <orudge> I also have issues uploading via SSH over IPv6 :/
17:30:08 <orudge> but downloading is fine
17:30:21 <orudge> and it only seems to affect the forums
17:30:30 <orudge> or, well, that machine, at least
17:31:04 *** Born_Acorn has joined #openttd
17:31:04 *** PierreW has joined #openttd
17:31:04 *** ThaAmazonous has joined #openttd
17:31:05 <Rubidium> I'm using one of those tunnel brokers though
17:32:13 <orudge> now, if I turn off my firewall, it seems fine
17:32:24 <orudge> andythenorth: do you use Outpost Firewall by any chance?
17:32:32 *** Brianetta has joined #openttd
17:32:54 <orudge> well, turning off my firewall fixed it for me, it seems!
17:33:10 <orudge> but that's not necessarily a good fix
18:00:35 <planetmaker> hm, Zuu? Your hide sign patch fails to apply on most of its hunks...?
18:00:41 <planetmaker> Did you start with a patched version?
18:01:14 <Zuu> yes, though I didn't think they would collide.
18:02:15 <Zuu> It could also be an issue with line endings?
18:02:41 <Rubidium> all but Windows' patch should handle that relatively gracefully
18:03:00 <planetmaker> but it was exactly that
18:03:02 <Zuu> When I open the patch in gVim, I see MS line endings at the end of the code lines.
18:03:21 <planetmaker> after dos2unix it applies without any inconvenience
18:04:01 <planetmaker> dos2unix command is made for that ;-)
18:04:11 *** andythenorth is now known as Guest4425
18:04:12 *** andythenorth has joined #openttd
18:04:19 <Zuu> I use that one sometimes.
18:04:52 *** frosch123 has joined #openttd
18:13:37 <planetmaker> Zuu: in signs_gui.cpp you define OwnerVisibilityFilter(...)
18:14:08 <planetmaker> and in FilterSignList() you call it only when the visibility bit has already been checked
18:14:14 <planetmaker> seems a bit double tested
18:14:38 <planetmaker> Rb suggested to replace the bit check in FilterSignList by an assertion instead
18:16:26 <Zuu> The check in OwnerVisibilityFilter could be turned to an assertion, but not the other way around.
18:17:02 <planetmaker> ehm... that's what I meant to write
18:17:38 <Zuu> But you are right that the check in OwnerVisibilityFilter is not needed. (I wrote the code in the order that I wrote the filter function first and then didn't see that the check become unneccsarry. But I guess that is what code review is for ;-)
18:18:00 *** |Jeroen| has joined #openttd
18:19:01 <planetmaker> don't worry to update
18:36:05 <planetmaker> jaja. Always these evil trailing white space ;-)
18:38:49 <Alberth> stupid editors that do not even show it
18:39:23 <frosch123> yeah, those editors are only beaten by those insering tabs and spaces at random
18:39:58 <planetmaker> :-) It could show it. But... it's not done a good way; thus I don't use it most of the time
18:40:50 *** Xrufuian has joined #openttd
18:41:32 <CIA-2> OpenTTD: planetmaker * r22708 /trunk/src/ (6 files in 3 dirs): -Feature [FS#4701]: Display option to hide competitors' signs and station names (Zuu)
18:42:18 <Zuu> Thanks to everyone who has helped with the review.
18:44:53 <frosch123> wow, ttdpatch got a usable bug tracker
18:45:22 <frosch123> i might actually dump some stuff into it :)
18:45:28 <Alberth> now they only need a developer :)
18:49:00 <Zuu> planetmaker: Have you taken any look at FS#4704? (the one-liner AIAirport noise fix)
18:51:01 <planetmaker> I looked at it. But I didn't yet spend time on understanding what it exactly means
18:51:25 <planetmaker> and I'm not familiar with that part of the code to know that from that patch alone
18:51:39 <Zuu> ok, I guess it needs some time to get into it if you are not familar with the concepts.
18:52:27 <Zuu> Maybe I also could make a better job on explaining what the patch tries to do?
19:02:34 <planetmaker> so, it basically changes the function thus that it also returns the noise level for unavailable airports?
19:03:55 <Zuu> I just added a more estensive comment at FS
19:04:34 <Zuu> Although it requires that the AI have at least one airport built of that type in order to get any noise level data.
19:05:41 <Zuu> I mean, for available airports, the AI don't need an airport to use it, but for unavailable airports, the requirment is that the AI must have one airport in order to get any data (just -1)
19:06:48 <planetmaker> so... it only means it additionally gets info on not-anymore available airports?
19:08:47 <Zuu> Eg. if it got a small airport after it has become unavailable.
19:08:49 <planetmaker> as it checks a specific tile
19:09:37 <Zuu> The tile doesn't have to be an airport tile.
19:10:13 <planetmaker> where does the "AI has to have that airport type" come into play then?
19:10:37 <Zuu> The change from IsValidAirportType => IsAirportInformationAvailable
19:11:11 <Zuu> The later returns true also if an AI have an airport of the given airport type, while the former does not.
19:12:22 <Zuu> (I'm just building OpenTTD with the patch and will make a small AI to just verify that the patch works - from code review I can't see any reason why it wouldn't but a verification can't hurt)
19:12:51 *** DayDreamer has joined #openttd
19:20:24 <krinn> planetmaker, if i drop a station, townrating is check before or after the action? (i mean the airport might destroy plants tile while been build, so before = no effect, and after, my rating could be bellow the acceptable value because of tree destruction)
19:22:13 <planetmaker> rating is checked and then construction as a single step takes place
19:22:20 <Alberth> sounds like a question for the source code, but what difference does it make?
19:22:20 <planetmaker> (for each click you do, of course)
19:22:41 <krinn> Alberth, if check before, say tr = -200 i could plant the station
19:23:02 <planetmaker> checking afterwards makes little sense IMHO :-)
19:23:05 <krinn> Alberth, if check after say tr=200 and loosing 50 because of loosing tree -> now tr -250
19:23:23 <andythenorth> they can spot unrealism at 50 paces
19:23:50 <andythenorth> but when asked for unrealism, they contribute ideas for realism :P
19:23:56 <Alberth> krinn: but then you have no way of finding out how much damage you'll do beforehand, so just build it is still the only sane solution
19:24:36 <krinn> human maybe, ai can count all tiles taken and rating lost before the action :)
19:24:46 <Alberth> andythenorth: much like you ask for simplification of signals, and they propose completely new signals besides what we have already :)
19:25:16 <planetmaker> krinn: but AIs are treated like humans without eyes ;-)
19:25:22 <andythenorth> I ask for a simple newgrf feature, and suddenly we reinvent half the spec :P
19:25:34 <planetmaker> they get no special treatment, especially not in the game mechanics
19:25:46 <krinn> andythenorth, i think one should make a thread "for openttd realism = godwin)
19:26:06 <andythenorth> I once pointed out something unrealistic to a model train fan
19:26:12 <andythenorth> his answer was quite brutal
19:26:44 <andythenorth> and pointed out some basic differences between real life and a toy
19:26:52 <andythenorth> based around, you know, scale, and stuff like that
19:27:09 <andythenorth> this from a guy who is known to spend £tens of thousands per year on model trains
19:27:19 * Alberth notices he has done industries in the wrong order, I skipped PBI
19:27:37 <krinn> bah, some women were burn at some epoch because it wasn't "realistic" or just possible to see a girl with red hair
19:27:58 <andythenorth> who wrote the story about the map fallacy? Lewis Carrol?
19:28:02 <peter1138> andythenorth, yeah, it's a mismash of OO/HO, heh
19:28:17 <andythenorth> peter1138: this guy was into p4 :P
19:28:29 <andythenorth> and custom etching his own parts
19:28:44 <krinn> the problem is that "realistic" guys are always pationate guys, hence brutal answer when you tickle it
19:29:06 <peter1138> andythenorth, cool :)
19:30:13 <peter1138> though not cool in the "gets all the ladies" sense...
19:30:25 <krinn> planetmaker, in wiki none note rating loose for terraforming (except because of loosing tree on tiles altered), there's none except that so ?
19:35:43 <Zuu> planetmaker: I've now play-tested the patch with a test AI that each day displays the noise level for the small airport. On 1.1.2-RC2 on the other hand, the expected bug-ish behaviour is reproduced.
19:36:16 <CIA-2> OpenTTD: alberth * r22709 /trunk/src/fileio.cpp: -Doc: Doxyment fileio.cpp.
19:37:03 <krinn> Zuu, why care about old airport noise level? if you upgrade airport, you shouldn't end with a so old airport that you lost its noise level info no?
19:37:31 <Alberth> krinn: usually, because you didn't upgrade yet :)
19:37:43 <Zuu> andythenorth: after my NUTS post it got to me that my post was maybe too realistic. ;-)
19:37:51 <krinn> in my tests, i only end with such case because the airport couldn't be upgrade at all (no place for it)
19:38:32 <krinn> Alberth, when a new airport get out, old one isn't remove, need more delay than that, and meanwhile, you upgrade them
19:38:50 <Zuu> krinn: You need to get info about how much noise the old airport generate in order to see how much more noise the new airport generates. No point on sending away aircrafts and start upgrading if you already by noise level can see that it will not work.
19:39:35 <Alberth> krinn: that's what you do, but not every author does the same
19:40:19 <krinn> Alberth, if the ai cannot upgrade, it will never need to know that info too, as it will never upgrade its airports
19:40:57 <krinn> Alberth, might be a problem if the ai v1 doesn't do that, save map, load ai v2 that handle it, this time yep, the ai might try upgrade the airport blindly
19:40:58 <Zuu> In some air-only games I've tested, all AIs had trouble turning profit and were not able to upgrade all airports by that time. It can partly be blamed by the fact that CluelessPlus does not have a system for reserving money for future investments.
19:41:57 <Alberth> Zuu: AI authors should do something themselves too :)
19:42:19 <krinn> what! we're AI authors! Used to be assist
19:42:27 <krinn> the computer does the job, not us
19:43:33 <Zuu> Still, with noise levels turned on, not all towns allow upgrading to large airports. The commuter does not appear until after the small airport has become unavailable.
19:43:41 <krinn> Zuu, the aircraft test is good yep, i was unable to upgrade mostly because at certain size, it's nearly impossible to get that much tile flat, so now terraforming add
19:43:44 <andythenorth> Zuu: UnrealisticAI!
19:44:22 <Zuu> andythenorth: Asking for YAZA - yet another zuu ai?
19:44:51 <andythenorth> I actually want VerboseAI
19:45:00 <andythenorth> it builds routes that spell out popular quotes
19:45:24 <andythenorth> shakespeare, ghandi, the tao etc
19:45:38 <Zuu> spelled out using road/rail as letters?
19:47:29 <krinn> just to get a "unrealistic! no train company would be rail as a quote!"
19:47:54 <Alberth> Zuu: yeah, I never understood the purpose of the commuter airport
19:48:12 <andythenorth> just to do something awesome :P
19:48:45 <Zuu> The commuter airport is actually quite good performance wise. I only would like to swap the controller tower with the depot to better work with my depot orders. :-)
19:49:43 <Zuu> And as krinn, says, AIs are not (yet) very good at terraforming, so therefore airports with a small size are much easier to handle.
19:50:06 <krinn> Zuu, nope i said "wasn't", now more than good :)
19:50:18 <planetmaker> Zuu: shall I swap depot and tower?
19:50:31 <planetmaker> Only side effect would be servicing would then take place in the tower :-P
19:50:38 <Alberth> NO, it's unrealistic! :p
19:50:42 <Zuu> planetmaker: It is good if you order vehicles to visit the depot before the station.
19:50:48 <krinn> planetmaker, must be a code problem as your depot is a uperleft = also answer to getlocation()
19:50:51 <Zuu> Otherwise, the current setup is better.
19:51:25 <planetmaker> well, we need NewAirports and newgrf-able state machines
19:51:30 <Zuu> Having the depot at the entrance is however more consistent with the international airport.
19:51:56 * Alberth ponders how to do NewAirports
19:52:14 <krinn> you can still get airport location, but need to check its tiles and note use AIStation.GetLocation()
19:53:15 <Zuu> krinn: You do higlight a good point, several AIs might break on an airport with the hangar at the north corner.
19:53:20 <krinn> planetmaker, your desing is making AIStation.GetLocation()==AIAircraft.GetDepotLocation() (or function like that)
19:53:52 <krinn> didn't because i simply remove commuter as buildable airport
19:53:55 <Zuu> I probably have some things that would explode :-)
19:54:14 <krinn> but still, might happen with some newGRF stilling a "classic" airport
19:55:04 <Zuu> Yep, that will need some code cleenup/improvements in the AIs to handle it.
19:55:25 <krinn> i think it's not solvable, only by the grf maker
19:55:26 <Zuu> But I don't think it will be impossible for AIs to handle as orders have a property that say if they are station or depot orders.
19:56:18 <Zuu> And as long as the airport have at least one non-depot tile, you can use that one to send to API functions that want a station/airport tile.
19:56:27 <krinn> it's doable if the author know the weakness, or changing getlocation code, but this will brake the rule "return topleft tile, build from topleft..."
19:56:46 <Zuu> One interesting thing I found out is that the Heliport have a hangar according to the API.
19:57:17 <planetmaker> all except the one-tile heli port(?)
19:57:41 <Zuu> heliport == the one-tile heliport
19:57:54 <Zuu> The others have names like helidepot and helistation.
19:58:15 <planetmaker> I just recently checked the definitions in the source and there it has no hangar
19:58:17 <krinn> Zuu, i'm not sure, but i think the plaform also answer to depot query
19:59:02 <planetmaker> the only one missing that feature actually
19:59:04 <Zuu> planetmaker: IIRC helicopters that visit the heliport will get serviced. So it might be something related to that.
19:59:12 <krinn> should retest that, it's easy with ai and sign
20:03:43 *** LordAro has joined #openttd
20:04:05 <planetmaker> Zuu: in the airport definitions the heliport is its on airport type
20:05:05 <planetmaker> and the comment reads "/** AirportSpec definition for airports without any depot. */ "
20:05:16 <planetmaker> but... might not exclude servicing
20:05:33 <planetmaker> though it'd feel wrong
20:08:25 <krinn> nope, no hangar report by the heliport and the platform
20:15:24 <CIA-2> OpenTTD: frosch * r22710 /trunk/src/ai/api/ (ai_airport.cpp ai_airport.hpp): -Fix [FS#4704]: Allow AIAirport::GetNoiseLevelIncrease() also for expired airports.
20:15:35 <frosch123> sorry, zuu, forgot the "(zuu)"
20:16:41 <krinn> erf, we couldn't blame him if it doesn't work
20:16:59 <Zuu> frosch123: No problem, such things happens.
20:17:44 <Zuu> I just posted the result of my playtest of the patch to the bugtracker. It contains nothing that holds against inclusion of the patch.
20:18:19 <Zuu> It did however prove me wrong on one thing. All AIs can get information about no longer available airports no matter if they have or even had one of the given type.
20:18:51 <Zuu> This is however also true for eg. AIAirport.GetAirportWith(airport_type)
20:19:39 <DorpsGek> __ln__: Darkvater was last seen in #openttd 1 year, 15 weeks, 3 days, 8 hours, 41 minutes, and 35 seconds ago: <Darkvater> good ol' days :)
20:19:50 <Zuu> But that is not a big issue as there exist a function that AIs should use to test if an airport type is really buildable as well.
20:20:47 <Zuu> Knowing that an airport that you can't build and doesn't have (anymore) is 4 tiles wide and give 3 in noise is not any information that will give any huge benefit to an AI. :-)
20:21:42 <krinn> not for the ai airport, but it's good information if you look at an airport of someone else, not that i do that, but who knows
20:22:47 <Zuu> hmm, yes. As long as an airport can uniquly be idenified by its size, you can now get the noise level of an opponents old acient airport. ^^
20:23:13 <planetmaker> Zuu: it'd not be a good idea to rely on unique airport sizes
20:23:20 <planetmaker> I can even change and break that now
20:23:49 <krinn> as long as the airport have a noise level info, you can stare at it until opponent die and then knew town accept X more noise level (as example)
20:24:21 <planetmaker> and actually city + metropolitan are same size
20:24:50 <planetmaker> no, they're different
20:25:04 <Zuu> I don't rely on that information, it was just a crazy idea on how to abuse the API :-)
20:25:05 <krinn> i was thinking they were same size
20:25:30 <planetmaker> I re-programmed each tile layout as newgrf ;-)
20:25:33 <LordAro> Alberth: PM read :) will do stuff tomorrow, thanks again :)
20:25:35 <planetmaker> commuter was more tiles
20:25:55 <Alberth> ok, hopefully it works
20:26:14 <Zuu> planetmaker: Sounds promising, is that included in OpenGFX+ Airports?
20:26:52 <planetmaker> Zuu: what exactly? You mean different sizes for (default) airports?
20:27:09 <Zuu> no, the re-programmed tile layouts.
20:27:20 <planetmaker> It also has some issues - which don't make it easy to change that while using the same state machine
20:27:24 <krinn> can newgrf change airporttype or introduce new ones ?
20:27:29 <planetmaker> Zuu: that's the only way you can re-define airports
20:27:40 <planetmaker> so, yes. Each airport there has a re-programmed tile layout
20:27:46 <planetmaker> tile layout != state machine though
20:28:18 <planetmaker> at least only way, when I want snow ;-)
20:28:35 <Zuu> tile layout is basically a 2D array (or modelled as a 1-dimensional array) with tiles?
20:28:54 <planetmaker> it's a 2D array with relative positions wrt the Northern tile
20:29:02 <planetmaker> and defines which tile is at the given positions
20:29:03 <krinn> planetmaker, airtype is fix in newgrf (must re-use one existing) ?
20:29:10 <planetmaker> I can also leave gaps or so
20:29:30 <planetmaker> krinn: airport type: no.
20:29:33 <planetmaker> I can define my own
20:29:49 <planetmaker> What I cannot do is use my own movement pattern - there I have to re-use an existing one
20:29:49 <Zuu> krinn: No, the rotatable airport in current OpenGFX+ appear at index 10
20:30:25 <Zuu> But NoAI works with it using index 0 (AT_SMALL) without problems as far as I have seen.
20:30:39 <krinn> i only see 0-8 + 255 invalid
20:30:40 <Zuu> Other than that it can't build the rotated view.
20:30:53 <planetmaker> krinn: do you have opengfx+ airports loaded?
20:31:11 <planetmaker> we talk of newgrf airports ;-)
20:31:14 <krinn> but it's not about opengfx, it's because i cannot knew that info
20:31:28 <krinn> as i cannot get the AT list
20:31:45 <planetmaker> what should an AI do with that?
20:31:53 <Zuu> krinn: Calm down, the API is not ready for more airport types. Some work need to be done there before we as AI authors need to wory.
20:32:04 <krinn> and to create an airport you need AirportType to build one, and you can only get AirportType info... by querying a tile with an airport :)
20:32:15 <krinn> AirportType list are AT_...
20:32:21 <planetmaker> Zuu: it's somewhat difficult. I *can* define new airport types
20:32:39 <planetmaker> But it's currently somewhat pointless... as I cannot define my own movement pattern
20:32:41 <krinn> if one build a new airportType, we will be like the refit vehicle size problem
20:33:11 <planetmaker> And... I actually *might* include my own interpretation of an airport in OpenGFX+ Airports at some stage
20:33:20 <Zuu> krinn: you can iterate from 0 to < 255 and check IsValidAirport for each value to see if there are any new airports. In theory. In practice I've found that indexes without airport types looks valid too.
20:33:25 <planetmaker> Like a real dirt airport. Small without buildings
20:33:56 <planetmaker> or maybe a different interpretation of city
20:34:06 <planetmaker> but that might be too big
20:34:32 <planetmaker> and could actually also be done as another layout / view for the small or city respectively
20:35:05 <Zuu> I would opt for a situation when views only have different rotations, but still have the same performance etc.
20:35:31 <krinn> city = just a big town or it have other attributes ?
20:35:32 <planetmaker> Zuu: that will always be the case
20:35:39 <planetmaker> A view means the state machine doesn't change
20:35:45 <planetmaker> Which means: same performance
20:35:55 <planetmaker> maybe just different rotation - but that's of no importance
20:35:57 <Zuu> planetmaker: good to hear. That will simplify the API significantly.
20:36:24 <Zuu> city == a town that grows quicker IIRC
20:36:26 <planetmaker> everything else IMHO would be a new airport and make the distinction between view and airport pointless
20:37:06 <Zuu> planetmaker: Good that my hope is the same as how the specs are layed out.
20:37:49 <krinn> we should have a kind of airport capacity if you introduce new airport type
20:38:21 *** Twerkhoven[L] has joined #openttd
20:38:38 <krinn> if i iter and query all answer, what will gave me an idea if the airport is better than another one, except assuming bigger = better, but newgrf authors might not assume that (like build only 1 road on a big airport)
20:38:56 <Zuu> we also need AIAirport.GetNumHoldingAircrafts(airport_tile), as AIs can no longer hardcode the holidng patterns of the existing airports.
20:38:58 <planetmaker> krinn: that's not clear for the defaults either
20:39:11 <planetmaker> the intercontinental is not the best, not even by throughput
20:39:15 <krinn> yep planetmaker but default are know
20:39:36 <krinn> and could be test, and you make your human choice, leaving ai without having to decide
20:40:05 <planetmaker> krinn: but humans don't know either
20:40:07 <krinn> but with newgrf, choices need to be taken by the ai
20:40:27 <krinn> yep until i play it, but ai can play with it, while i never play it myself
20:40:27 <Zuu> Anyway, I might take a look at the API and see what can be improved on the airport side. I don't know much about NewGRFs or NewGRF Airports (yet) on the technical level.
20:41:12 <krinn> if airport have a value (througput or something) ai could choose
20:41:14 <planetmaker> well, currently: nothing which is of interest to AIs except availability dates
20:41:33 <Zuu> It will probably be possible to expose the number of runways to the AI. (yes I know, the # of runways are not the most important for performance)
20:41:52 <krinn> would be a not bad start yes
20:42:02 <planetmaker> yes, that might be possible as well as loading places and hangars
20:42:08 *** Brianetta has joined #openttd
20:42:11 <planetmaker> and heli stations
20:42:25 <planetmaker> and of course large vs. small vs. heliport type
20:42:26 <krinn> if one make a 12x12 airport full of maiz with 1 runway, at least ai could see it might not be a good one
20:42:43 <planetmaker> oh, and important is the catchment size
20:42:45 <andythenorth> forgot to draw 2 angles of forklift :P
20:42:58 <planetmaker> that's actually quite important
20:43:13 <krinn> catchment is what ? how many planes can land at same time ?
20:43:26 <krinn> or time to clear runway
20:44:19 <planetmaker> so... currently: availability, dimensions, size (wrt. aircraft types), noise, catchment,
20:44:32 <planetmaker> later: runways, depots, heliports, loading bays, hangars
20:44:42 <krinn> depots is easy to get now
20:44:46 <andythenorth> airports should have a 'snowed in' disaster
20:44:57 <planetmaker> rotation (view) will also need to be exposed
20:45:25 <planetmaker> krinn: I just list what I consider important. I didn't read current API ;-)
20:45:41 <Alberth> andythenorth: good suggestion :)
20:46:07 <krinn> yes, this thing can let ai decide, and be able to pick itself new airport never seen
20:47:54 <planetmaker> that's how an AI should do it, yes
20:48:13 <planetmaker> otherwise it'll fail on airports sooner or later
20:48:25 <andythenorth> depends on airport closure :P
20:48:25 <planetmaker> currently rather later, though
20:48:35 <andythenorth> I would like to be able to use disasters
20:48:49 <planetmaker> disaster newgrfs ;-)
20:49:05 <krinn> can only fail: use hardcode preference (that type is > that type), or guess with poor info (bigger size = better), newer = better...
20:49:25 <Zuu> planetmaker + frosch123: I just realized, that unless you did that, I forgot to include updates to the NoAI changelog.
20:49:50 <krinn> but as i said, size could be wrong with a "fancy" maz airport with 1 runway but huge, and newer = commuter is sample it might not be the case (or heliport)
20:50:08 <frosch123> Zuu: i considered it a fix
20:50:16 <frosch123> which is not to be documented
20:50:56 <frosch123> i extended the doc though with an @pre
20:51:02 <Zuu> Good, I completely forgot about it until now. I guess as it is a fix it should neither have updates to the compatibility layers for old API versions.
20:52:10 <krinn> Zuu, bah, old version might test return = valid or -1, and apply some heuristic with -1, now, they will work as they get the right result like if the airport isn't dead
20:53:57 <Zuu> I don't think it will cause any trouble to old AIs. Hopefully :-)
20:54:56 <krinn> me too, they get the right result, and never now get -1, even one was handling -1 case won't happen that's all
20:58:54 <Zuu> planetmaker: Can views have different introduction dates?
20:59:15 <Zuu> Or maybe I should seek the NewGRF specs rather than bothering you. ;-)
20:59:22 *** andythenorth has left #openttd
21:00:48 <krinn> :) i also think planetmaker is openttd internal wiki, sorry planetmaker to bug you always with questions
21:01:43 <planetmaker> Zuu: no, one airport, one introduction date
21:01:53 <planetmaker> view is just that: a different view with a rotation
21:02:21 <planetmaker> I can change graphics depending on that, but...
21:02:41 <planetmaker> but changing that is not unthinkable
21:03:22 <krinn> openttd support landing north/south ? (as if view could switch airport direction)
21:03:32 <planetmaker> krinn: sorry, what do you mean with 'openttd internal wiki'?
21:03:56 <planetmaker> the holding patterns of airports are automatically adjusted.
21:04:15 <planetmaker> It's nothing a player or AI can change or influence
21:04:15 <krinn> openttd internals wiki, with s might be clearer, was thinking as "if one ask question about how something works, planetmaker have the answer"
21:04:44 <krinn> the code itself too, but you need to search all codes to find the same answer, hence planetmaker = wiki answer, faster result
21:05:38 <planetmaker> krinn: I no virtually nothing about the AI API
21:05:58 <krinn> eheh, but most ai trouble doesn't comes from the api itself :p
21:06:38 <krinn> api is far more limited than openttd, and api use openttd mechanic at best, so knowledge of openttd code is stronger
21:06:59 <krinn> i don't think any api functions does an action openttd doesn't do, or very few
21:07:15 <planetmaker> NoAI cannot cheat
21:07:25 <planetmaker> everything an AI can do a player can also do.
21:07:34 <planetmaker> That's the philosophy behind it. And vice versa
21:07:42 <planetmaker> but that's not all implemented yet
21:07:53 <planetmaker> for that it needs people like Zuu ;-)
21:09:43 <planetmaker> as it usually works best, if people can work on both sides of an API ;-)
21:10:42 <krinn> i'm sure, i see that right here :)
21:11:09 <Zuu> But it has taken me long time to get used to the OpenTTD code and doing more than one-liners. :-)
21:11:25 <krinn> you do 2 lines comment now ?
21:11:41 <Zuu> I was refering to one line patcehs.
21:11:42 <Gniarf> can you rotate the map/view by 90 ?
21:12:00 <Zuu> Gniarf: Only if you also rotate your screen 90 degrees ;-)
21:12:20 <Gniarf> I guess it was in simcity 2000
21:12:37 <krinn> that's why you have transparency option
21:12:55 <Zuu> And from today, we even have a new transparency option :-D
21:13:23 <Chris_Booth> nice what is that?
21:14:17 <krinn> except tiles, i don't see what could be transparent
21:14:38 <Alberth> it's a setting that toggles whether some part of the game world are transparent
21:15:21 <krinn> lol, tiles so, to see thru hills? but tunnels aren't really "build" no ?
21:15:52 <Alberth> industries, houses, bridges
21:17:36 * krinn is waiting the nude in house patch from zuu
21:18:32 <planetmaker> Zuu: that's natural. And it's not like OpenTTD's code base is small...
21:19:26 <Zuu> (agrees that it is not small)
21:19:38 <krinn> lot of code and functions, one ever count total lines of code ?
21:20:06 <Zuu> Last time I counted (long time ago) it was about 300 000 lines of code.
21:21:28 <Zuu> I guess the bump in 2009 is NoAI/Squirrel.
21:21:29 <planetmaker> krinn: easy thing to do: for i in 'hg st -A'; do wc -l
21:21:37 <krinn> look at graph, 100k in a ~year
21:22:20 <V453000> unimaginable for me :D
21:22:27 <planetmaker> 2.5 years, krinn ;-)
21:22:28 <krinn> wc -l will count also comments... but i suppose all should count, as comment is also eating time to write them
21:23:09 <krinn> ah yes, oops, didn't saw it was 2 years scale
21:23:29 <krinn> that's still 50k / year
21:25:02 <Zuu> As for comparison, the largest code base I've written just by myself is 30 000 lines, which is of course not tiny, but still much smaller than OpenTTD :-)
21:29:45 <Zuu> night planetmaker & Terkhen
21:30:46 <krinn> good night planetmaker (or was it for Terkhen?)
21:32:05 *** HerzogDeXtEr has joined #openttd
21:43:59 <krinn> Zuu, you know a way to cheat to lower my town rating to check it against my code (when i use my ai to destroy things, i get demolition limit after ~12 tiles delete)
21:44:38 <Eddi|zuHause> krinn: enable magic bulldozer?
21:45:14 <krinn> nope, still have the limit!
21:45:32 <krinn> just enable me to destroy anything, but i'm limit by # of destruction / (dunno seconds)
21:45:34 *** Chillosophy has joined #openttd
21:45:46 <Eddi|zuHause> oh, that is a setting
21:46:12 <krinn> heard of such setting, but never saw it, it's an ai setting only ?
21:46:48 <Eddi|zuHause> try "list_settings clear" and "list_settings terraform" on the console
21:47:31 <krinn> get 64k_frames=100 & frame_burst=10
21:47:49 <Eddi|zuHause> that's a little low
21:47:56 <krinn> that's really low, but it's the aircraft test map, must have been set to bug the ai
21:48:09 <Eddi|zuHause> default is 4M and 4k
21:48:45 <krinn> hmmmm, how should we handle that ? i can grab that settings, but should i adapt my ai to that, making huge pauses ?
21:49:09 <Eddi|zuHause> i can't help you with that ;)
21:49:26 <krinn> damnit! wiki is sleeping !
21:50:06 <krinn> lol didn't know the buldozer is also enable for the ai
21:50:19 <krinn> lol just lost an industry ^^
21:50:36 <Eddi|zuHause> it does that if you keep it enabled :p
21:50:38 <Zuu> krinn: Figure out how the furst thing etc. work and insert Sleep-calls to comply with the rules. Or see if there exist a specific error code for this error that your AI can detect and insert sleep calls.
21:51:10 <Zuu> the magic buldozer is even enabled for towns, so they can destroy industries.
21:51:12 <krinn> Zuu, i'm digging the ai doc to see if there's a kinda "error too fast..." to answer sleep to that
21:52:22 <krinn> bad-> DemolishTile could only answer ERR_AREA_NOT_CLEAR
21:52:23 <Zuu> If you get such an error message in the GUI, it should be not too hard to add it to the API if it is missing there. Maybe something to start with for you? (given that the API doesn't have it yet of course)
21:53:22 <krinn> strange settings, might be useful for evil multiplayer gameplay, but putting that on an ai ?
21:53:56 <krinn> if the ai is evil, it will do it 1 time and user remove it, not like human that could have random action
21:54:02 <Zuu> AIs can play on multiplayer, so the AI can get into trouble if it doesn't handle it.
21:54:34 <krinn> yep, i'm just surprise ai have to deal with that and is affect by that without a specific error to handle it
21:54:44 <Zuu> Also, AIs play by the same rules as humans.
21:54:59 <krinn> but that settings is emotional
21:55:03 <Zuu> The old AI cheated and was criticed for that.
21:55:16 <Zuu> The old AI had free terraforming.
21:55:57 <krinn> i knew, but it's not a free terraforming, it's don't terraform as much as you wish
21:56:44 <krinn> you know what i mean zuu, except program to do random evil things, the ai when terraforming as a plan, and not some emotional reaction to bug another user
21:58:54 <krinn> hihi -> Area isn't clear, try to demolish the building on it. for ERR_AREA_NOT_CLEAR
21:59:11 <Zuu> I see what you mean, but I dissagree to losen the idea that humans and AIs should play on equal rules.
22:00:07 <krinn> that's not really the same imo
22:00:15 <Zuu> That the API doesn't report a relevant error for the burst-thing, however, I think can be changed. Especially if humans get a relevant error.
22:00:43 <krinn> i don't know, i get the error, but i was playing as my ai to lower its town rating
22:00:49 <Zuu> But not removeal of the burst limit.
22:01:31 <krinn> and that number is more than low (it's the aircraft save), i suppose none would be so restrictive in normal game
22:01:52 <Zuu> That said, there are certain settings in OpenTTD that will get most AIs into trouble.
22:02:00 <krinn> frame_bust=10 ~10 delete and error appears, try to clear a 12x12 area for your airport with that ^^
22:02:15 <Zuu> As AI author you have to choose if you want to support all settings or just the most common ones.
22:02:27 <Zuu> Supporting all settings is however probably impossible ;-)
22:02:46 <Zuu> Just testing all settings will take very long time.
22:03:21 <krinn> i don't know, for that case, never saw my ai face it, and i just can't see the error
22:03:34 <krinn> but returning area_not_clear when trying to clear it ^^
22:03:48 <Zuu> One example that I don't think many AIs handle good is turning of joining of distant stations. If you do that, you can't pass a station id to the BuildStation functions.
22:04:06 <krinn> my ai will assume an opponent object is in the way (that's why it cannot be demolish)
22:05:29 <krinn> too much settings, good for human fine tuning their playing, but for ai authors... too hard to handle them
22:05:44 <Zuu> Anyway the point is, AIs and Humans should play on equal rules. So if you can figure out that a human can get a more relevant error message than your AI, file a bug report on that or even better try to code a fix.
22:07:20 <krinn> i will see if futher things should be done
22:07:22 <Zuu> If the humans get the same level of error message as AIs, one can always "fix" it by providing both AIs and humans with the same better error message.
22:07:49 <krinn> that would be good for ai author that wish handle it
22:08:31 <krinn> but it would be faster to just ignore it and accept random failure with such hard settings
22:09:35 <krinn> i don't do my ai to be playable in all unusual situations that human can met and handle easy (just waiting in that case, while human action are slower in all case)
22:10:36 <krinn> doing such an ai would be support the 300k lines ^^
22:17:44 <Zuu> you do not need to support the GUI code which is probably a rather significant portion of those lines.
22:18:10 <krinn> yep, and sound, graphics... but still that would be a lot :)
22:19:08 <Zuu> I think what is amazing is how stable Convoy has proven to be over the years even though it has been quite long time sinece last release of it.
22:19:41 <krinn> well, easy, few code, less complex = less error
22:20:48 <krinn> stabilizing nocab is certainly harder
22:29:01 <krinn> i think it could be code in newgrf to tell what runaway is use to land and what is use to get off
22:29:07 <krinn> and that info is important
22:29:20 <krinn> more landing runaway = faster queue remover
22:29:32 <krinn> more getoff runway = faster cleaning of loading stations
22:30:09 <krinn> in case one build 2 airport with 3 runways we should be able to base our choice on that
22:30:49 <krinn> if you look at airport, there's one with 2 runway next to each other that are use to land and getoff
22:31:17 <krinn> and the biggest one use 2 but they are opposites, effect -> a landing aircraft is not blocking one that wish getoff
22:31:30 <krinn> faster handling of vehicle, but still they only have 2 runway
22:31:52 <krinn> so if we could knew where they are, or detect runway tiles, we could see efficiency of runway position
22:32:03 <Zuu> Please sumarize and write a proposal in the thread so it will not get forgeten.
22:32:40 <krinn> will screenshot them, i don't think my pov will be understood else ^^
22:35:39 <krinn> will use wiki links to them, should be ok
22:45:17 <Zuu> Hmm, for the first time, reading a NewGRF spec that start to really make sense :-)
22:48:49 *** zachanima has joined #openttd
22:50:21 <krinn> in the sense: i know why they do crazy things, because newgrf allow them to ? :)
22:51:04 <Zuu> No, I mean starting to grasp the spec at a higher level.
22:51:56 <Zuu> You know, the NFO stuff with bits all over :-)
22:54:12 <krinn> i know 0 about newgrf :P
22:54:45 <krinn> just that they do stupid things with prize in it, they have mostly nice graphics, and are way to hard to identify in bananas :)
22:55:28 <krinn> oh, and that i never knew if adding one to get more vehicle i will end with new vehicle but loosing the original ones, lol, maybe ending with fewer than without it
22:55:51 <Zuu> I haven't coded any NewGRF either. I know a little bit about it from what I've heard aronud the forums but have tried to hide away from it. :-)
22:56:39 <krinn> yeah, sad, i was just looking to add some new vehicles, i endup give up with all those problems
22:56:52 <Zuu> I did however draw the old fizzy drinks factory in OpenGFX :-)
22:57:02 <krinn> i only have a few to test the ai with them, but don't play with it
22:57:39 <Zuu> Just don't change NewGRFs in the middle of a game and you should be fine.
22:58:08 <krinn> hmmm, i play ttd original graphism, i'm too use with the icons of the menu, i'm lost with opengfx ones
22:58:09 <Zuu> If you want to play a game with NewGRFs without figuring out which to combine, join #openttdcoop.
23:14:28 *** duckblaster has joined #openttd
23:19:37 <__ln__> like c'mon, 'leopard' is 'luipaard' in dutch. who's going to take that kind of an animal seriously?
23:19:57 <Zuu> krinn: On forums, please use capitalization and punctnation of your sentenses.
23:20:14 <krinn> i'm editing it for that right now :)
continue to next day ⏵
