IRC logs for #openttd on OFTC at 2012-04-20
⏴ go to previous day
00:01:33 <Ammler> heffer: I do that for suse
00:09:24 <Mazur> morten: Have you looked at /usr/local/bin ?
00:19:38 *** Biolunar has joined #openttd
00:26:18 *** Djohaal has joined #openttd
00:47:07 *** XeryusTC has joined #openttd
00:47:37 *** planetmaker has joined #openttd
00:48:08 *** planetmaker is now known as Guest1005
01:08:15 *** Stimrol has joined #openttd
01:11:11 *** HerzogDeXtEr has joined #openttd
01:30:06 *** HerzogDeXtEr1 has joined #openttd
01:43:38 *** brambles has joined #openttd
02:00:18 *** HerzogDeXtEr has joined #openttd
03:13:02 *** supermop has joined #openttd
04:43:23 *** andythenorth has joined #openttd
04:45:12 <DDR> Heck yeah... random refinery just built up opposite to where I was shipping oil to another refinery, so I switched the truck feeders to that refinery to transfers and now my ships are taking oil both ways for double profit! :D
04:47:00 *** andythenorth has left #openttd
04:56:18 *** Eddi|zuHause has joined #openttd
05:12:21 *** telanus has joined #openttd
05:29:37 *** roadt__ has joined #openttd
06:04:48 *** mahmoud has joined #openttd
06:32:45 <Rhamphoryncus> Eddi|zuHause: words escape me
06:32:58 <Rhamphoryncus> Well actually, I can think of a few words, but none I want to say ;)
06:34:11 *** kkb110_ has joined #openttd
06:47:30 *** TWerkhoven has joined #openttd
06:54:17 *** fonsinchen has joined #openttd
07:20:55 <Rhamphoryncus> fun fact: evil mode can be cleared by imposing a temporary speed limit
07:21:37 <Rhamphoryncus> This spaces out the trains that aren't there yet, letting the backlog clear and full speed resume
07:29:13 *** TGYoshi has joined #openttd
07:58:41 *** roadt__ has joined #openttd
08:24:01 *** sla_ro|master has joined #openttd
08:43:32 *** TWerkhoven[l] has joined #openttd
08:55:29 *** goodger has joined #openttd
09:40:43 <NGC3982> ah, NGC3982 has some competition.
09:41:21 <NGC3982> NGC4258 (M106) is fantasticly profiled by CXC with the blue x-rays and red IR
10:15:50 *** drac_boy has joined #openttd
10:28:00 *** Guest1005 is now known as planetmaker
10:39:23 *** FLHerne has joined #openttd
11:09:41 *** HerzogDeXtEr has joined #openttd
11:43:49 *** roboboy has joined #openttd
11:46:14 *** fonsinchen has joined #openttd
11:56:53 *** Firartix has joined #openttd
12:58:35 <Ammler> [14:56] <vuntz> grr, tarball is openttd-1.2.0-source.tar.xz (not the usual scheme) <-- just you know, I am not the only guy not liking the uncommon -source
12:59:03 <planetmaker> what is usual? *-src-*
12:59:22 <Ammler> the source bundle is usually called package-version.tar.gz
13:01:10 <planetmaker> aha. And how do you distinguish it from a binary package?
13:01:28 <Ammler> the binary package does have a arch in the bundle name
13:01:32 <peter1138> by the lack of a binary indicator
13:02:27 <peter1138> but we've done it this way since forever
13:02:40 <peter1138> don't think it makes a big difference :)
13:03:26 <Ammler> well, it needs custom config
13:04:44 <Ammler> the question is, is more complicated to fix it and so current configs needs to change to default or always need to setup special config for openttd
13:04:56 <peter1138> custom config of what?
13:05:29 <Ammler> to get upstream version
13:05:38 <Ammler> to prepare source bundle
13:06:00 <peter1138> you're talking about something specific without saying what
13:06:12 <peter1138> i don't require any 'config' to grab the latest source and compile it
13:06:23 <Ammler> e.g on rpm, you need to specify the package name as it is not default
13:06:58 <Ammler> like %setup -n openttd-version-source instead just %setup
13:07:21 <peter1138> once that line is there, how many times do you need to reenter it?
13:08:18 <Ammler> as said, hat is the question, what is easier to change existing configs or keep "wrong config"
13:08:45 <Rienzilla> can a dedicated server go on pause if there's no client connected?
13:09:08 <peter1138> it's not that uncommon either
13:09:23 <peter1138> Rienzilla, check the settings
13:09:32 <Ammler> you know a 2nd project using that uncommon file sheme?
13:24:14 <blathijs> Ammler: planetmaker: I am also slightly discontent with the -source postfix, FWIW
13:24:54 <blathijs> Though it just means that I manually need enter the package version whenever I import a new OpenTTD version, since git-buildpackages autodetects the version as 1.2.0-source
13:39:13 <Ammler> wow, that is even worse as for us :-)
13:40:14 <Ammler> well, the lack of customation for such version grep is prove to me that it _is_ uncommon
13:43:03 *** kkb110_ has joined #openttd
13:46:53 *** TWerkhoven has joined #openttd
13:50:55 <blathijs> Ammler: Well, it's not so bad. I can handle a few extra keystrokes every few months, really :-)
13:58:18 <Ammler> it is less for me, it is just that evryone who setup it needs to customize it
14:09:16 <blathijs> Ammler: You mean everyone who wants to compile the rpm?
14:10:32 <Eddi|zuHause> more like everyboy who writes a specfile
14:11:20 <Eddi|zuHause> which i don't think happens often... people will just copy an existing one
14:11:42 <Eddi|zuHause> but i don't think i fully understood the problem anyway
14:24:56 <Ammler> blathijs: no, everyone who setup build
14:25:05 <Ammler> or monitoring version change etc,.
14:27:33 <Ammler> well, if you fix it, it is work for me, but still I would prefer it :-)
14:57:27 *** Prof_Frink has joined #openttd
15:09:36 *** Progman has joined #openttd
15:21:49 *** andythenorth has joined #openttd
15:34:22 <drac_boy> hmm vineyards or not
15:40:19 *** supermop has joined #openttd
15:57:59 *** tokai|noir has joined #openttd
15:57:59 *** ChanServ sets mode: +v tokai|noir
16:02:44 <supermop> more alcohol in ottd
16:03:04 <supermop> better throw in poppy fields and marijuana plantations too
16:11:56 *** oskari89 has joined #openttd
16:14:24 <telanus> just making certain: da{P g e} would give me dag and dae depending on singular vs multiple
16:19:16 <Eddi|zuHause> vineyards produce fruit, not alcohol :)
16:19:45 *** Biolunar has joined #openttd
16:21:47 <planetmaker> Eddi|zuHause: depends on how ripe you pick the grapes ;-)
16:24:32 <FLHerne> I've had grape juice accidentally ferment in my fridge before...the carton burst in the end, made a sticky mess :P
16:27:28 <Eddi|zuHause> i've had juice ferment outside the fridge... but IN the fridge? how many years did it stand there?!
16:29:27 *** Devroush has joined #openttd
16:31:09 <supermop> well lager yeast could work at fridge temperature
16:31:26 <supermop> not sure i'd want a fruit just fermented with them though
16:31:34 *** Chris_Booth has joined #openttd
16:34:32 *** Chris_Booth is now known as Guest1087
16:34:38 *** Chris_Booth has joined #openttd
16:35:41 <FLHerne> how many years did it stand there?! Not that long actually. The fridge is a bit unreliable though...#
16:36:42 <ffpp> does anybody else think that it would be very helpful to have additional information on cargo you deliver in the game ? like average travel time on a given multi-hop route or something like this
16:37:16 <NGC3982> how can cargo affect average travel time?
16:37:51 <ffpp> not because it affects it, just to see how long it takes for it to travel
16:38:35 <ffpp> in large networks where cargo has to take many hops, I find it often difficult to judge which routes lack efficiency
16:39:12 <ffpp> vehicle income and profit is always skewed by transfer vs. final income percentages
16:52:57 <FLHerne> stupid plasma-desktop
16:55:36 <FLHerne> I try to move one icon, and then the whole thing goes berserk again :@
16:58:55 <supermop> yeah don't touch a plasma
16:59:10 <supermop> im sure it will ionize your skin
17:00:22 *** Alberth has joined #openttd
17:00:23 *** ChanServ sets mode: +o Alberth
17:02:13 <supermop> i desperately need to find someone to go with me to the opera tomorrow
17:03:37 <andythenorth> supermop: this may not be the best place to advertise ;)
17:03:44 <andythenorth> unless friend-of-a-friend kicks in ;)
17:03:53 <andythenorth> tried Facebook :o
17:05:23 <supermop> also need someone to go to record store day in my place
17:08:42 *** TWerkhoven2[l] has joined #openttd
17:12:55 *** theholyduck has joined #openttd
17:15:25 *** |Jeroen| has joined #openttd
17:16:25 <andythenorth> ^ this is happening because I removed some strings for station names?
17:16:30 <andythenorth> thereby breaking savegames?
17:17:00 <andythenorth> he's changed the bug report :P
17:19:29 * andythenorth might have broken all savegames
17:19:43 <andythenorth> and in my experience, removing strings causes ottd crashes
17:19:53 <andythenorth> [unreproducibly]
17:20:17 <Alberth> how can you change a newgrf sitting at my harddisk?
17:21:04 <planetmaker> spooky action at the distance, of course!
17:21:40 <Alberth> especially, as it was only virtually at my disk :p
17:21:54 <andythenorth> I can break your game with the aid of bananas :P
17:23:17 <planetmaker> andythenorth: nope. OpenTTD won't update existing savegames
17:23:24 <planetmaker> unless you manually force that
17:24:00 <Eddi|zuHause> in that case, change at least the min-compatible-version
17:24:40 <Eddi|zuHause> btw. openttd should try harder to prevent you from (manually) updating to a not compatible version
17:26:10 <Alberth> nah, you have to be able to shoot yourself in the foot every now and then :p
17:26:34 <Alberth> like I just did, my 'hg fetch' of firs is broken
17:26:55 <andythenorth> so ottd won't update a FIRS 0.7.2 savegame to 0.7.3 ?
17:27:01 <planetmaker> Eddi|zuHause: how should it do that?
17:27:06 * andythenorth never plays the game so doesn't pay attention
17:27:27 <planetmaker> doesn't it prevent you already quite hard to update to an incompatible one?
17:27:34 <planetmaker> just reload_newgrfs won't do it
17:27:41 <planetmaker> if you overwrite the file
17:27:48 <Alberth> andythenorth: it tries to find an exact match first
17:27:56 <Eddi|zuHause> planetmaker: no. i mean if you edit the newgrf config from the gui
17:28:21 <Eddi|zuHause> planetmaker: it curently makes no difference if you change to a compatible or incompatible grf
17:28:26 <planetmaker> on a running game? Why should it? You're then a developer and know what you do
17:28:33 <planetmaker> by definition ;-)
17:28:41 <Eddi|zuHause> planetmaker: no, why would i know that? it doesn't say in the GUI
17:28:44 <andythenorth> so I don't have to leave all the legacy strings in place?
17:29:13 <planetmaker> Eddi|zuHause: a non-developer can't change it in the GUI
17:29:21 <planetmaker> (on an existing map)
17:29:23 <Alberth> andythenorth: I don't think so, the savegame contains the exact version newgrf that is used, and it will reload that
17:29:23 <Eddi|zuHause> planetmaker: but i'm a developer
17:29:32 <Eddi|zuHause> planetmaker: i'm talking specifically about me
17:29:50 <michi_cc> andythenorth: Make sure to bump the compatible version, and then you can just respond with 'changing NewGRFs in game is not supported'.
17:29:55 <Eddi|zuHause> planetmaker: i get no notice whether my replacement now is to a compatible or incompatible version, even though the game could easily test it
17:29:59 <planetmaker> andythenorth: legacy strings are only useful for people who want to update the grf on a running game
17:30:20 <planetmaker> Like you. And maybe a couple of other people
17:30:32 <Alberth> Eddi|zuHause: perhaps you should protect you from yourself and refrain from using the newgrf dev setting
17:30:34 <planetmaker> "Normal" users don't update NewGRFs on existing maps
17:30:52 <Eddi|zuHause> Alberth: that's not the point
17:30:53 <planetmaker> But removing the strings, of course, requires you to up the min_compatible_version, andythenorth
17:31:02 <Alberth> planetmaker: "Normal" devs also don't :)
17:31:23 <andythenorth> savegame compatibility is confusing :P
17:31:52 <andythenorth> why do I attempt to preserve compatibility if the game ignores that?
17:31:54 <Alberth> andythenorth: just make every version incompatible with every other version :)
17:32:15 <andythenorth> I try quite hard to not require version bumps, on the basis that it doesn't ruin people's saves
17:32:21 <andythenorth> seems to be incorrect assumption
17:32:58 <Alberth> andythenorth: it only becomes important if I don't have the exact version, I guess
17:33:17 <TWerkhoven> doesnt openttd automatically retrieve old versions (if available on bananas) if a savegame/server requires it?
17:33:23 <planetmaker> (and if the exact version is not on bananas)
17:33:32 <Alberth> which is a non-issue, as every version is at bananas
17:33:45 <planetmaker> andythenorth: it is important for people who play with nightly versions or who actually DO change newgrfs on their savegames
17:33:48 <Alberth> TWerkhoven not automatically
17:34:21 <Alberth> but you can click 'find missing content' :)
17:34:23 <planetmaker> andythenorth: you might want to do that to do the favour for yourself. To keep your testing savegames
17:34:40 <Eddi|zuHause> maybe the game should make it easier for non-developers to "update" (not add) grfs?
17:36:18 * Alberth is not getting within 20 miles from that can of worms
17:36:59 *** Chris_Booth has joined #openttd
17:37:23 <TWerkhoven> well, thats what i meant
17:37:45 *** Biolunar_ has joined #openttd
17:39:59 <Eddi|zuHause> well, currently the min-compatible-version has exactly no use
17:40:26 <Eddi|zuHause> non-developers can't use it because they can't change newgrfs, and developers can't use it because it's completely hidden from them
17:40:46 <Alberth> so what remains is removing it entirely :p
17:41:25 <andythenorth> I don't really understand the system
17:41:31 <andythenorth> min-compatible version
17:42:14 <andythenorth> I understand the core issue, but not the solutions we have
17:42:31 <supermop> i overheard two people on the street last night talking about lawyers using github
17:42:52 <supermop> for domething? i guess drafting documents?
17:43:17 <planetmaker> andythenorth: savegame: grfID + md5sum --> identical file, if accessible
17:44:07 <planetmaker> grfID + version + not matching md5sum --> use the file with the highest rev which still has the min_compatible_version which is lower than version
17:44:34 <planetmaker> will only be used if a newgrf with matching md5sum can't be found nor downloaded
17:44:39 <andythenorth> so...I should leave the legacy strings in?
17:44:42 <andythenorth> for maximum compatibility
17:44:54 <andythenorth> I can leave them in english only?
17:45:01 <andythenorth> it might cause bug reports, but not crashes
17:45:12 <planetmaker> Remove all or none IMHO
17:45:18 <planetmaker> but not translations only. That sucks
17:45:21 <CIA-1> OpenTTD: translators * r24156 /trunk/src/lang/ (afrikaans.txt italian.txt):
17:45:21 <CIA-1> OpenTTD: -Update from WebTranslator v3.0:
17:45:21 <CIA-1> OpenTTD: afrikaans - 36 changes by telanus
17:45:21 <CIA-1> OpenTTD: italian - 12 changes by lorenzodv
17:45:28 <planetmaker> it'll show as untranslated strings, also
17:45:33 <planetmaker> and would be very annoying
17:45:34 <andythenorth> now I have to go and find a changeset and forward merge it somehow :(
17:46:23 <andythenorth> and it's a big commit, not atomic :P
17:51:01 *** andythenorth has joined #openttd
17:54:07 *** Miguelzinho has joined #openttd
17:54:56 <Alberth> hmm, zou een language file met line wrapping niet veel handiger zijn?
17:55:47 <Eddi|zuHause> that is very borderline amount of english in that line :p
17:56:11 <Alberth> transl: Would'nt a language file with multi-line wrapping be much more usueful?
17:56:26 <Alberth> Eddi|zuHause: I realized that too 30 seconds after entering it :p
17:59:08 <telanus> andythenorth: is there an afrikaans translation of FIRS?
18:00:22 <planetmaker> telanus: not currently. But you're certainly free to provide it
18:00:42 <andythenorth> Alberth: wrapping lang files :o
18:00:53 <andythenorth> freedom from horizontal scrolling :P
18:02:08 * andythenorth bumps the savegame
18:03:12 <andythenorth> I need a commit prefix for 'broke your savegames' :P
18:03:25 *** TheMask96 has joined #openttd
18:05:22 <Alberth> -Change: Let's not support old games any more.
18:05:39 *** valhallasw has joined #openttd
18:07:37 <andythenorth> yeah that's roughly what I chose :P
18:08:35 <andythenorth> "Change: broke all your savegames because I can't be bothered to diff a changeset" :P
18:12:15 * andythenorth wonders what he started with this FIRS US / Brit English thing
18:14:56 <Alberth> 'prevent opening' <-- what string is used for that text?
18:19:15 <Eddi|zuHause> andythenorth: US and GB english have two different language IDs, so you can offer both
18:20:31 <andythenorth> but all the rivet counters have popped up with opinions :)
18:20:52 <andythenorth> is there a language ID for 'northish'?
18:21:45 <Alberth> or 'pirate language'? :)
18:22:01 <andythenorth> google has a number of silly languages
18:23:27 <FLHerne> andythenorth: US v UK English is because there are differences :P
18:24:11 <FLHerne> And people sometimes find it irritating when the 'wrong' spelling/name is used
18:26:53 <andythenorth> I've been through that debate ;)
18:26:53 <andythenorth> I've made the set mostly US, it irritates people
18:28:34 *** snack23 has joined #openttd
18:28:34 <Alberth> indeed, pirate language should be default!! :p
18:28:57 <FLHerne> andythenorth: You did, yes. It irritates me sometimes :P
18:29:19 <Alberth> FLHerne: make a better industry newgrf
18:29:43 <planetmaker> nah, just help PaulC sort out the differences BE / AE / AU :-)
18:29:51 *** th_gergo has joined #openttd
18:30:02 <andythenorth> declare yourself your own language
18:32:24 * andythenorth can't make FIRS tickets be 'fewer'
18:32:30 <andythenorth> I close one, write 4 more :P
18:33:07 *** frosch123 has joined #openttd
18:33:33 <andythenorth> so how do I add new translations?
18:33:50 <planetmaker> Just copy the lng file to the folder
18:34:00 <andythenorth> I don't need to do any quality control?
18:34:06 <planetmaker> make sure the language header is right
18:34:15 <planetmaker> well. sure. You should test it for sanity
18:34:37 <planetmaker> i.e. become a native speaker of that language and start the game with it :-)
18:34:51 <planetmaker> you may skip the pre-requisit with knowledge of the language ;-)
18:34:59 <planetmaker> but not the deduction :-P
18:35:14 * andythenorth envisages someone trolling the grf with lots of 'your mum' jokes instead of industry names :P
18:35:53 <andythenorth> we're very trusting about all this code we share around freely :P
18:36:05 <andythenorth> I'll compile e.g. patches without reading them
18:36:41 <andythenorth> be trivial to do something like a makefile attack with rm * -r
18:36:51 <planetmaker> who says that they're not read?
18:37:38 <andythenorth> I don't read things I apply ;)
18:37:45 <andythenorth> I just curl and patch :P
18:39:00 * Alberth check whether he can add his trollings into the FIRS repo
18:40:03 <andythenorth> italian doesn't compile
18:40:13 <DorpsGek> andythenorth: I have not seen Snail.
18:40:19 <DorpsGek> andythenorth: I have not seen snail.
18:40:42 <andythenorth> issues with plurals
18:41:35 <DorpsGek> Alberth: I have not seen snails.
18:42:01 <andythenorth> why would my shell crash :P How silly
18:44:16 <FLHerne> Linux on Apple hardware - Apple build quality, Linux not-awful-ness :D
18:44:27 <andythenorth> one day Linux will be done
18:44:32 <CornishPasty> My mac works just fine on OS X
18:44:39 <andythenorth> the problem with OS X is, it's done, but is being undone
18:44:49 * andythenorth has become a conservative crusty in his old age
18:44:54 <FLHerne> My macs work just fine on Linux :D
18:44:59 <andythenorth> Feature uncreep :P
18:45:22 <CornishPasty> 2154 is the year of Linux on the desktop
18:46:00 <Alberth> andythenorth: mine builds ok, but I didn't clean it first
18:46:23 <Alberth> CornishPasty: yes, at the two desktops that still exist then :p
18:46:34 <FLHerne> CornishPasty: 2012 is :P
18:46:34 <andythenorth> Alberth: your copy of snail's translation? or your translation?
18:46:39 <CornishPasty> FLHerne: lolrofl
18:46:47 <Alberth> andythenorth: firs trunk
18:47:01 <andythenorth> I didn't commit italian yet
18:47:05 <andythenorth> as it's broken :(
18:47:25 <FLHerne> Linux derivatives are already the dominant OS on smartphones; soon will be on tablets
18:50:50 <andythenorth> I have no idea how to fix plurals, esp. for a language I can't read :)
18:50:51 <CornishPasty> They're the dominant OS on smartphones because Android is the new Symbian
18:51:20 <andythenorth> Android is, according to my friends who have to develop for it, a low-quality raging mess
18:51:43 <andythenorth> whereas iOS is a relatively stable predictable environment
18:51:59 <andythenorth> except that Apple will shaft you at any point with an update that breaks your app
18:52:22 <CornishPasty> Or by not allowing your app onto the app store for <insert arbitrary reason>
18:52:44 <andythenorth> at least there are only 2 predictable problems
18:52:58 <andythenorth> whereas some of them now have over 1,000 variants to QA for
18:53:05 <andythenorth> which is obviously impossible :P
18:53:57 <krinn> querying if (!vehlist.IsEmpty()) print("aircraft="+cEngine.GetName(vehlist.Begin())+" r_dist="+distance+" e_dist="+AIEngine.GetMaximumOrderDistance(vehlist.Begin())); <<<-- i get e_dist=495616 for name=Douglas DC-8 so... bug ?
18:54:07 <Eddi|zuHause> "it's stable. except when they break it on purpose"
18:54:38 <andythenorth> rock, hard place
18:54:39 <krinn> thru gui this model is set to 704 cases
18:57:28 <krinn> just it's more than origin->target as this is a full go and back
18:57:48 <andythenorth> if I'm breaking savegames, 0.8.0 makes more sense than 0.7.3
18:59:15 <krinn> planetmaker, i now understand what you mean with the a->b->c trouble that a->c is too far
18:59:41 *** oskari892 has joined #openttd
18:59:52 <krinn> but this is awful, this then not limit aircraft distance, this limit aircraft usage
19:00:01 *** telanus1 has joined #openttd
19:00:03 <planetmaker> andythenorth: not necessarily
19:00:12 <planetmaker> I'd do versions by features
19:00:28 <planetmaker> depends on how you want to do it
19:00:40 <andythenorth> I'm kind of easy, but I'd expect 0.x.y to be compatible with other 0.x
19:00:42 <planetmaker> you could just break things in trunk and not do it in 0.7 branch
19:00:51 <planetmaker> I'd expect that, too :-P
19:00:57 <andythenorth> no branches are in use afaik
19:01:26 <andythenorth> so 0.7.3 or 0.8.0?
19:01:30 <andythenorth> I am writing changelog now
19:01:59 <planetmaker> so what's new compared to 0.7.2?
19:02:09 <planetmaker> I'd think not much?
19:02:22 <krinn> planetmaker, just to make it clear : having an aircraft range of 10 cases, i suppose i could jump from airport to airport if they are within limit of 10, but will in fact fail at the 3rd airport because total range would be 20, outside the aircraft range
19:02:52 <planetmaker> unless you make orders to fly back the same way
19:02:58 <krinn> so it's not a distance limit, it's a usage limit: i could only use 2 airports
19:03:40 <planetmaker> if your airports have a mutual distance of 15 in a triangle, you'll have no issue with a plane range of 20
19:03:45 <krinn> oh i get it, your explain, i have take as a->b->c a-c fail when in fact c->a might fail
19:03:58 <planetmaker> but if they're in a line, the distance c->a is 30, thus 20 will not suffice
19:04:28 <andythenorth> 0.7.3 smells better :P
19:04:57 <krinn> planetmaker, i won't as i have already spent all my amount of time to cry yesterday, else i would again !
19:06:18 <planetmaker> krinn: it's not hard to also check distance of last order -> first order, is it?
19:06:38 <planetmaker> it's a ring anyway
19:07:25 <krinn> does openttd take 495616 or the 704 as a->b check ?
19:07:36 <andythenorth> if i.delta > foo, raise :P
19:08:22 <krinn> andythenorth, except if i.delta < foo because foo= foo**2 :)
19:10:20 <frosch123> krinn: both AIOrder::GetOrderDistance and AIEngine::GetMaximumOrderDistance() return squared euclidian distance
19:10:30 <frosch123> so, you can compare them without any conversion
19:10:46 <krinn> except i'm not comparing them
19:10:55 <frosch123> (for aircraft; for ships it is actually manhattan distance)
19:11:10 <krinn> as i compare the max distance an aircraft could do vs my manhattan distance from airport a to b
19:12:37 <planetmaker> still, easier to compare squares
19:12:45 <frosch123> use AITile::GetDistanceSquareToTile then :)
19:13:02 <frosch123> but, granted, we should document the unit of those vehicle/order distance functions
19:13:36 <krinn> yep, my fisrt thinking as you see would be compare ab bc ca and not check abc
19:18:43 *** fonsinchen has joined #openttd
19:23:53 *** supermop has joined #openttd
19:28:33 *** Rhamphoryncus has joined #openttd
19:32:11 <Alberth> I have not seen someone for a long time
19:32:29 <TrueBrain> I have seen many people not for a long time
19:36:07 <michi_cc> krinn: Always use AIOrder::GetOrderDistance to query the distance.You can pass tiles that either are part of a station or are not, it will automatically calculate the right thing.
19:37:01 <krinn> you can solve the problem in two ways: force a bad distance and check AIOrder validity next. Or check the distance and not force AIOrder as you already know the distance will be reject
19:37:50 <krinn> because the 1st one is rought, while the 2nd way allow me to pickup an aircraft that will indeed be able to travel that distance
19:37:50 <michi_cc> AIOrder::GetOrderDistance does not depend on the actual orders of any vehicle. You pass two tiles to it.
19:39:35 <michi_cc> AIEngine::GetMaximumOrderDistance and AIOrder::GetOrderDistance complement each other, and you can always use > or < on the returned values without knowing if it is square, manhatten or whatever that is applicable for the vehicle type.
19:40:43 <TrueBrain> its like someone thought about it when designing it :D (wasnt me, thats for sure; I dont think things through in general :D)
19:40:49 <krinn> well, except i need to know the two points, while knowing the distance is easy
19:41:09 <krinn> eheh i know this might looks weird at first ok
19:41:34 <TrueBrain> your loking for a place to put your second <whatever> down?
19:41:40 <michi_cc> How do you build an airport at the destination without knowing a tile? :)
19:42:25 <michi_cc> But if you really want to emulate it, for aircraft the distance is the squared distance between the northern tile of each airport.
19:42:33 <krinn> that's that, i don't have the airport, it's just to handle the engine, so the X & Y aren't important and not pass, only x->y is useful
19:43:08 <krinn> but the solve appears easy, just compare distance**2 and i'm done
19:44:52 <michi_cc> Anyway, only AIOrder::GetOrderDistance is guaranteed to have the right result of something changes in the future, for anything else you have to watch out yourself.
19:45:16 <frosch123> "I personally prefer FIRS, but that's mainly because it only takes 1 GRF slot while ECS can take up to 7, and the total number of GRFs one can load at once is rather limited." <- haha
19:46:00 <krinn> well, i'm not making magic there, and i don't really see why only the AIOrder::GetOrderDistance is guaranteed vs the MaxOrderDistance michi_cc
19:46:18 <Terkhen> we need 255 NewGRFs per game :P
19:46:29 <krinn> vehlist.Valuate(AIEngine.GetMaximumOrderDistance); + vehlist.KeepValue(distance*distance)
19:46:52 <frosch123> Terkhen: yup, and 256 cargos with 512 industry types or so
19:47:20 <Terkhen> I agree, otherwise the game feels unnecessarily limited
19:47:23 <krinn> but we certainly need more than just 1 nogo...
19:47:50 <Eddi|zuHause> well, if you load single-aircraft and single-object grfs, plus every station set out there, you really get to the limit fast
19:48:58 <krinn> if you think about all nogo scripts as "funny script", 1 is enough, you run that mode or that mode of playing, but if you think of nogo helper script, 1 is clearly too limited
19:50:11 <Eddi|zuHause> krinn: the script author can load use the other scripts as modules
19:50:23 <glx> it's a goal script, not a helper script
19:51:04 <krinn> well, nogo helper script could do wonder, even better than some newgrf as it leave a chance for AI to speak with a NOGO script, while newGRF don't
19:52:07 <krinn> I've seen Eddi|zuHause speak in forum with someone (wow him again!) about newGRF limit per aircraft type, a newGRF that will do that will broke all AI, a NOGO script can do that easy and can speak to AI to tell them what are the limits
19:52:12 <Eddi|zuHause> *hüstel* callback 18!
19:52:41 <frosch123> krinn: i don't understand your latest question about GetTownAuthority. that functions doe not return a bool
19:53:33 <krinn> frosch123, ah yes my bad, but the it return the townID, and the bool in my mind was compare result vs your check town :/
19:54:07 <SilPho> Not sure if this is a bug on my end or a general bug: Buying land doesn't seem to have a cost. Only the cost of bulldozing. Is that right?
19:54:08 <frosch123> still i don't get your question. are there maybe some negations too few or too many?
19:54:22 <krinn> frosch123, i'll try to make myself clearer
19:54:45 <frosch123> maybe you have an example ai :)
19:54:52 <krinn> frosch123, i wish find a tile that townX handle, so if i build my airport at that tile i then must get sure townX own the airport
19:55:06 <krinn> frosch123, plenty, any AI using airport get the problem
19:55:33 <frosch123> why do you actually want to find a tile within the authority of a town?
19:55:45 <frosch123> i could understand the reverse, but not this way
19:55:56 <krinn> because i just want build an airport for townX and not for townY
19:56:06 <krinn> this cause a lot of trouble else
19:56:29 <frosch123> but townauthority is about town rating
19:56:36 <frosch123> not about "belonging to a town"
19:56:47 <Eddi|zuHause> so you want to valuate a tile list by which town a station would get
19:57:00 <krinn> then that's what i'm seeking, a belonging to a town
19:57:17 <krinn> Eddi|zuHause, yes, just that, but this is impossible to get
19:57:20 <Rubidium> SilPho: that sounds about right
19:57:22 <Eddi|zuHause> make a delaunay triangulation :)
19:57:37 <frosch123> you can use GetClosestTown
19:57:42 <Eddi|zuHause> (i have code :))
19:57:45 <frosch123> that will tell you the name which the station will have
19:58:04 <Eddi|zuHause> (it isn't quite up to date though)
19:58:06 <frosch123> well, i have no idea what you want then :)
19:58:54 <SilPho> Rubidium: Really? Just seems strange that you can "buy" bulldozed land for free
19:59:17 <krinn> frosch123, i just want build an airport in a town, and make sure it will belong and be handle by that town
19:59:53 <krinn> that's seem so easy, but in fact, anything i have try end with a "could chance it work"
19:59:58 <frosch123> what do you mean with "belong" and "handle"?
20:00:43 <krinn> i mean says townX: when built the airport should affect townX
20:01:20 <krinn> town and station affect each over you know :)
20:01:23 <Rubidium> SilPho: buying land is a nuisance anyway
20:01:49 <krinn> airport affect noise and rating in the town
20:01:49 <frosch123> krinn: you are asking way to broad. i have no idea what kind of affection you mean
20:02:02 <frosch123> ah, so you want to know about noise
20:02:18 <SilPho> Rubidium: For the buyer or for other players?
20:02:44 <krinn> i know that already frosch123
20:02:54 <Rubidium> everyone (for other reasons)
20:03:13 <krinn> frosch123, you never build an airport for townX that was in fact affecting townY ???
20:03:33 <frosch123> again... affect what?
20:03:47 <frosch123> an airport affects a town in about a dozen ways
20:03:48 <SilPho> Rubidium: I guess I'm not really disputing that, but I'm sure there used to be a cost for it. Maybe it was intentionally removed
20:03:56 <frosch123> iirc noise even affects multiple towns, not a single one
20:04:09 <krinn> frosch123, in all the ways you wish, i don't care the way, i only care WHO it will affect
20:04:23 <frosch123> then your question is ill-formed
20:04:35 <krinn> and you cannot get it, when per example, two towns are close each other, and might share same tile influence
20:04:47 <frosch123> the various effects on towns all have different radii and rules
20:05:43 <krinn> ok let me try the other way, think about an AI
20:06:31 <krinn> you check AITile.IsWithinTownInfluence for townX
20:06:47 <krinn> it take the tile and townID as param
20:07:03 <krinn> you have sometimes return value > 0, mean townX influence that tile ok
20:07:48 <krinn> but if you try to build the airport, you may per example get "Error too many station" or "townY doesn't accept more noise" for that tile
20:08:10 <krinn> see the problem: that tile is influence by townX, but in fact, an airport at that tile will belong to townY
20:08:58 <krinn> while there's 0 station in townX (to explain the "too many station" error), and townY indeed as too many stations
20:11:06 <frosch123> so, about noise: the town affected by building an airport is returned by AIAirport::GetNearestTown. you can check the effect using GetNoiseLevelIncrease and GetAllowedNoise
20:11:22 <frosch123> wrt. town rating, the town being asked is returned by GetTownAuthority
20:11:54 <frosch123> and AIAirport::GetNearestTown may indeed return a different town than GetTownAuthority or GetClosestTown
20:12:09 <krinn> please look at the airport sample in the screenshot in the bug report, you might understand better visually see it
20:12:42 <frosch123> your screenshot does not help me about understanding what you actually want
20:12:56 <frosch123> you now asked me twice about town noise, which i answered
20:13:10 <krinn> i told you twice i know how to handle noise
20:13:12 <frosch123> if you want to know some other town effect you have to ask a proper question
20:13:30 <frosch123> there is no such thing as "airport belongs to town"
20:13:37 <frosch123> it is different for each effect
20:13:53 <Eddi|zuHause> maybe that is already the biggest bug
20:14:31 <krinn> that's it Eddi|zuHause, as is, i endup trying to build two airports in the same town because openttd mistake me about the tile owner
20:15:11 <krinn> i cannot get a valid answer from GetTownInfluence, nor from GetTownAuthority, nor from GetClosestTown!
20:15:13 <Eddi|zuHause> krinn: so you only care about which label the station will get?
20:15:36 <krinn> yes! as this really mean that town will handle that station
20:16:10 <Eddi|zuHause> that would mean GetClosestTown should handle it
20:16:17 <Rhamphoryncus> SilPho: think of it this way: bulldozing itself buys, but unless you plant a sign you soon abandon it
20:16:19 <krinn> Eddi|zuHause, it doesn't :(
20:16:33 <frosch123> krinn: because your assumptions are wrong
20:16:41 <krinn> Eddi|zuHause, you sometimes have a small town next to a big town that have expand close to it
20:16:50 <frosch123> GetClosestTown is the correct answer wrt. name in the station sign
20:16:58 <frosch123> but it is not the correct answer for noise
20:17:12 <SilPho> Rhamphoryncus: I could think of it that way, I'm just wondering if this is desired behaviour. The wiki seems to disagree
20:17:41 <Rhamphoryncus> SilPho: I can think of behaviour I'd like better, but this isn't a bug, no
20:17:42 <Eddi|zuHause> krinn: which effect is wrong when you rely on GetClosestTown?
20:17:58 <SilPho> Rhamphoryncus: Fair enough. Thanks :)
20:19:13 <krinn> Eddi|zuHause, all effects that influence a town: if you remove something there, townY will get crazy, while you are handling townX, same, you have check noise for townX, but anytime you try build it there, townY keep yelling it doesn't want more noise...
20:20:03 <Eddi|zuHause> krinn: so you need to loop over all tiles of the station layout?
20:20:24 <frosch123> Eddi|zuHause: no, he just has to use two functions
20:21:15 <frosch123> GetClosestTown and AIAirport::GetNearestTown are two functions he wants
20:22:04 <krinn> see the airport, it's name is Chateauroux Est while i was expecting build one for Babyonne (all the 2 are influence tile found for bayonne)
20:22:29 <Eddi|zuHause> krinn: so which function did you use there?
20:23:06 <krinn> 2: all 2 are influence over bayonne, all 3 in the screenshot are town authority for over bayonne
20:23:15 <krinn> as you see the 3 are only a few
20:23:43 <Eddi|zuHause> yes, town authority is smaller than the town for very large cities
20:23:51 <Eddi|zuHause> as it does not grow with city size
20:23:56 <frosch123> krinn: your screenshot lacks GetClosestTown
20:24:53 <Rhamphoryncus> So what you want is one that incorporates the rings?
20:25:21 <krinn> a way to find a tile in that ring that really belong to the target town and not another town !
20:25:33 <Eddi|zuHause> krinn: if you didn't use GetClosestTown, then why do you think it is wrong?
20:25:48 <krinn> because i have test it already, and it fail too
20:26:02 <krinn> getclosesttown could tell you the closest town is X while you will have airport build for Y
20:26:41 <Rhamphoryncus> Conceptually, take the distance for each town, subtract the size from the largest ring from each, then take whichever is closest. Note that the distance may be negative
20:27:04 <Eddi|zuHause> krinn: like frosch123 said. GetClosestTown will tell you which name it will get, Airport::GetNearestTown will tell you which noise is affected
20:27:35 <andythenorth> anyone want to build FIRS 0.7.3 and tell me the md5?
20:28:00 <Eddi|zuHause> krinn: so you need to check both to get the "right" airport
20:28:55 <andythenorth> I have 01a3ed84fc554101c44cc5de1a4f7956
20:29:00 <krinn> I can limit it by checking them both, but still both can mistake
20:30:37 <andythenorth> md5s don't match the bundle server
20:30:59 *** drac_boy has joined #openttd
20:32:25 *** Chris_Booth has joined #openttd
20:32:35 <andythenorth> why so many md5s?
20:33:00 <andythenorth> if I use grfid -m I get one md5
20:33:10 <andythenorth> make bundle_zip has written a different md5
20:33:19 <andythenorth> and the bundle server has yet another
20:33:53 <drac_boy> andythenorth I dunno what to tell you, I have *never* ever used these ever :-/ sorry
20:34:03 <andythenorth> I can't release while they're different
20:34:13 <andythenorth> as it tells me the grf is broken
20:34:17 <andythenorth> how / why I have no idea
20:34:55 <andythenorth> frosch123: would you mind md5-ing FIRS 0.7.3?
20:36:10 <Rhamphoryncus> andythenorth: I'll do it
20:36:33 <frosch123> i am already building :)
20:36:43 <Rhamphoryncus> I'm doing a fresh clone
20:37:23 <andythenorth> 2 data points :)
20:38:06 <Rhamphoryncus> unrelated: how much of firs' design is dictated by the 3 in/2 out limit for industries?
20:39:59 <andythenorth> Rhamphoryncus: quite a lot of it
20:40:15 <frosch123> Rhamphoryncus: grfcodec
20:40:58 <andythenorth> frosch123: so none of my md5s match yours :(
20:41:33 <frosch123> oh, i can update nml as well
20:42:16 <frosch123> wow, i was on nml 0.2 branch
20:44:18 <andythenorth> 95bc342fc014a0ce1a91e996cd449f11 matches yours now
20:44:37 <frosch123> no changes for me after nml update
20:44:46 <frosch123> the 2 changesets it pulled did not affect the result
20:45:18 <andythenorth> anyway, release time
20:45:41 <Rhamphoryncus> Guess I have to update nml too
20:47:09 *** sla_ro|master has joined #openttd
20:47:09 <Rhamphoryncus> I also make sure to make clean every time. FIRS' makefile is not so reliable :/
20:47:20 <frosch123> Rhamphoryncus: hg purge --all
20:47:26 <andythenorth> I have aliased a 'make clean && make install
20:47:32 <frosch123> but make sure to not have any diffs or similar in the firs directory
20:47:39 <frosch123> it will delete everything not from the repo
20:48:00 <Rhamphoryncus> I do make || make clean when working on firs itself, as that seems to work right. However, this is an update of nmlc
20:48:07 <andythenorth> building without the makefile is often much faster
20:48:38 <Rhamphoryncus> I match now too :D
20:49:02 <Rhamphoryncus> You can build without the makefile?
20:49:10 <telanus1> when I started Afrikaans was 95% translated, now 99.4% translated
20:49:20 <Rhamphoryncus> telanus1: woot :)
20:49:36 <telanus1> just 21 lines, then 100%
20:49:53 <frosch123> quick! add new strings :p
20:50:22 <telanus1> Then FIRS translation (hopefully)
20:50:31 <Rubidium> Afrikaans used to be 100% translated
20:50:39 *** TWerkhoven2[l] has quit IRC
20:50:53 <Eddi|zuHause> Rubidium: by google translate :)
20:51:21 <telanus1> a lot by google translate
20:52:24 <frosch123> telanus1: do you maybe also speak traditional chinese?
20:52:29 <andythenorth> most of FIRS 0.7.3 is patches by other people :P
20:52:38 * andythenorth has become a patchpack maintainer, not an author :P
20:52:54 <frosch123> andythenorth: or a boss :p
20:53:04 <andythenorth> Rhamphoryncus: you can build without the makefile if you're only updating graphics :P
20:53:08 <frosch123> only holding a speech when it is done
20:53:14 <andythenorth> it's *way* faster
20:53:33 <Rhamphoryncus> andythenorth: That's what happens when you have a good quality, interesting project that's open to outside help :)
20:53:41 <Rhamphoryncus> ah, never done graphics
20:54:50 <Eddi|zuHause> andythenorth: but if you stopped using "make clean", the makefile would figure out that you're only doing graphics, and only call grfcodec
20:55:17 <andythenorth> but I have to use make clean because the dep check doesn't work reliably
20:55:31 <Eddi|zuHause> andythenorth: you're doing it wrong then
20:55:56 <andythenorth> I have to do a mental dep check instead "what did I change"
20:56:03 <telanus1> frosch123: I only speak Afrikaans & English, and understand Dutch 90% of the time, and able to read very-basic german
20:56:10 *** th_gergo has joined #openttd
20:57:03 <frosch123> well, traditional chinese would have been 94.1 % :)
20:57:57 <frosch123> maybe ottd was banned from the chinese web
20:58:44 <Eddi|zuHause> i'm sure it was due to the controverse copyright :p
21:00:10 <frosch123> too sad that we cannot reach 1/3 of the earth population because the marathi translation is incomplete
21:00:51 <frosch123> hmm, maybe i should suggest that in the "marketing"-topic
21:01:03 <Rubidium> frosch123: interestingly Traditional Chinese has zh_TW
21:01:16 <frosch123> oh, did i confuse them?
21:01:24 <frosch123> i though simplified was TW
21:01:30 <Rhamphoryncus> Hrm. This map will be very profitable once setup.. industries are scattered quite well. Unfortunately starting is looking almost impossible ;)
21:02:47 <Eddi|zuHause> Rhamphoryncus: i've had that problem with YACD... it's really only sensible to start with passengers.
21:03:05 <Rhamphoryncus> This is a 512x512 with 5 towns
21:03:45 <andythenorth> does YACD build against trunk currently?
21:04:34 <Rubidium> andythenorth: ofcourse it does
21:04:50 * andythenorth hasn't tried for some time :P
21:05:03 <Rubidium> ... for the right revisions of trunk
21:05:17 <frosch123> andythenorth: you should ask on the forums whether someone can build 1.2+yacd for you :p
21:05:38 <Rubidium> frosch123: that's easy
21:06:36 <andythenorth> does YACD build against trunk tip *right now* ? :P
21:06:38 <frosch123> sounds like you applied distributive law
21:07:09 <Rubidium> after all, the last version of YACD was against 1.2.0.22532 (or something similar)
21:07:29 <Eddi|zuHause> didn't someone do an update?
21:09:00 <krinn> ah i found why i get mad with it, i remember now, basically any town north another town invalidate mostly all tiles within both town if you check getclosesttown, because the station sign is north too
21:10:02 * andythenorth ponders a python conversion of FIRS
21:10:08 <krinn> that's game mechanic liimit
21:10:24 <Rhamphoryncus> andythenorth: how much could be converted?
21:10:31 <andythenorth> tradeoff: codebase I don't understand (current), but others do (but are busy)
21:10:39 <andythenorth> versus: codebase *only* I understand :P
21:11:18 <supermop> how can a newgrf be written in python?
21:11:21 <Rhamphoryncus> But most of it would still be in nml?
21:11:29 <Terkhen> with lot sof workarounds, probably :P
21:11:45 <Rhamphoryncus> So you'd be replacing the CPP with python?
21:11:56 <krinn> andythenorth, do it in squirrel, you will love other newGRF doing crazy things you cannot see :)
21:12:03 <Rhamphoryncus> With a vague possibility of hacking into nmlc later since it's also written in python..
21:12:24 <andythenorth> Rhamphoryncus: meh
21:12:30 <Rhamphoryncus> very very vague? :)
21:13:04 <andythenorth> nml modules can be loaded
21:13:37 <Rhamphoryncus> The question is if it gains anything, since you're still limited to the underlying newgrf API
21:13:46 <Rhamphoryncus> or VM as I should say
21:13:47 <Zuu> krinn: I hear your pain with airport tiles. I haven't really found the problem that you have, but probably just because I haven't seen it yet. As it should for sure manifest in probably both PAXLink and CluelessPlus.
21:14:27 <andythenorth> I gain code I can work on
21:14:41 <andythenorth> I lose the other developers and/or annoy them :(
21:14:53 <Rhamphoryncus> I should mention that I have LOTS of experience in python
21:14:55 <andythenorth> I also gain a new round of conversion bugs to fix
21:15:26 <andythenorth> Rhamphoryncus: have you looked at how BANDIT works? Or CETS?
21:15:34 <andythenorth> both python, both different but viable
21:16:02 <andythenorth> CETS compiles directly, BANDIT uses templating
21:16:04 <krinn> Zuu, the real problem appears with big airports, by this time, places are short, towns are bigger and many town have "shared" influence tiles, once there, the problem get so tied, because town have many stations and refuse your while you have check some other town noise/rating...
21:16:47 <andythenorth> BANDIT has a main build script that creates first class python objects for each vehicle
21:16:51 <Zuu> krinn: Good news skip the intercontinental as the international is better
21:16:59 <andythenorth> these are then in scope for the templater that writes out nml
21:17:07 * Rhamphoryncus grumbles about dev.openttdcoop.org wanting flash
21:17:37 <krinn> Zuu, well, the size increase chance of failure, but even the smallest one can endup bad
21:18:37 <Zuu> well, I guess that is the joy of all complexity of OpenTTD :-)
21:18:59 <Zuu> Noone has said that making an AI for OpenTTD is easy. :-)
21:19:07 <krinn> at least, it's game mechanic, something "strong written"
21:19:20 <krinn> not like this newGRF hell :)
21:19:44 <Zuu> heh, though also NoGos currently are a black box to AIs I think.
21:20:42 <krinn> no Zuu i already think about it ! NoGo can take over an AI, so nogo can "speak" with an AI
21:20:51 <krinn> so the AI will not get blind like with a newGRF
21:21:07 <andythenorth> krinn: you're on a losing battle ;)
21:21:22 <andythenorth> the default approach to features is 'that should be done in newgrf' wherever possible
21:21:55 <andythenorth> so newgrf has an enormous capability to create insanity for AI authors
21:21:58 <krinn> say you make a NoGo script that limit chopper to 20: your AI can put a sign at his HQ, says "nogo chopper" and nogo read it, and alter it to "chopper 20" -> now your AI know chopper limit is set to 20
21:22:09 <Zuu> krinn: Though, I don't know if NoGo know if a given company is an AI. Placing cryptic signs that are part in NoGo<->AI communication for players would be wierd.
21:22:17 <krinn> with newGRF: your AI will get an error after the 20rd choppers without an explain
21:22:48 <krinn> Zuu, nogo doestn't place the sign, the AI place it to query nogo, nogo check it exist and answer
21:23:11 <krinn> nogo and AI can speak, AI & nogo cannot speak with newGRF
21:23:14 <supermop> an unnamed social site i used just made me a moderator without my input....
21:23:19 <Zuu> yes, with your proposal it is like that. I wrote the message before reading your proposal.
21:23:29 <Rhamphoryncus> andythenorth: I hate to say it but bandit looks inscrutable :(
21:23:41 <andythenorth> which parts did you try to scrute?
21:23:51 <Rhamphoryncus> nml is inscrutable to begin with, and templating just adds more confusion
21:23:59 <supermop> there seems to be some script whereby if you've been around long enough and no one complains about it makes you a moderator?
21:24:02 <krinn> of course the AI need to handle it, but at least, you can (and i would say) easy made it, maybe some better communication could be even made
21:24:34 <krinn> but with newgrf, you cannot even know one is load :)
21:24:40 <Zuu> krinn: For that query system to work, it would help greatly to have a common protocol.
21:24:43 <Rhamphoryncus> andythenorth: templates mostly
21:24:49 <andythenorth> read build_bandit.py
21:24:59 <Zuu> Or AIs would have to try all different hand shaking messages until NoGo responds :-)
21:25:15 <krinn> that's just a standard to set
21:25:28 <andythenorth> specifically the two main Classes
21:25:50 <krinn> but it could be done, that's why nogo & AI can work really good, while nogo & AI cannot with newgrf and are blind
21:26:43 <Rhamphoryncus> andythenorth: that and the cfg are good
21:26:55 <krinn> this is also why i think nogo limit to 1 script will show its limit fast, or we should build one big nogo that could do anything
21:27:03 <andythenorth> newgrf has a defined spec, but no defined ways to use it, so it won't ever be reliable in the way you want
21:27:46 <andythenorth> Rhamphoryncus: truck_template.pynml is straightforward
21:28:07 <andythenorth> specifically the item{} block
21:28:08 <Rhamphoryncus> Yep, that one's good
21:28:09 <krinn> andythenorth, right now, for AI & nogo, it would be like if newGRF have no callback with openttd : blind
21:28:25 <andythenorth> it will stay blind
21:28:39 <andythenorth> there's simply no way to specify what a newgrf is doing reliably
21:29:10 <andythenorth> you could extend the hinting system and hope authors use it....
21:29:21 <Zuu> krinn: A such protocol would probably need either to cover all possible NoGo usages, or it could have a rather slim standard behaviour and leave the rest up to the NoGo to define.
21:30:04 <Zuu> Basically there could be "nogo shortname" to get the short name of the NoGo and "nogo version" to get the version and then leave the rest up to the NoGo to decide.
21:30:30 <krinn> Zuu, i was thinking a basic protocol that define how to speak with each other, and a reserved word someone took and we add it as a valid word, a bit like event, if AI doesn't know it, it won't be able to query it anyway
21:30:53 <Zuu> It should then be packed as an AI and GS library :-)
21:31:16 <krinn> Zuu, this would also gave nogo a chance to register who is AI and who is human
21:31:36 <andythenorth> if Newgrf authors could bundle a helper script, then that could interface with AI + NoGo the same way
21:31:50 <Zuu> Though, only AIs that have the AI beacon active :-)
21:31:51 <krinn> and this could be valuable info for nogo, don't ask human what AI can do, don't ask AI what only human can do
21:32:27 <Zuu> Sounds like an interesting project :-)
21:32:27 <krinn> Zuu, yes of course, but most AI would handle it fast i think, just they would handle more or less fast the new "api"
21:32:56 <krinn> i'm pretty sure anyone would agree to add the "speak" part, specially because we could just made a lib that handle it :)
21:33:42 <krinn> we could try define the easy, register your AI with a nogo script part to see if you wish ?
21:33:53 *** andythenorth has left #openttd
21:33:59 <krinn> and see the number of AI agreeing to handle it
21:34:50 <Zuu> I just scanned the nogo list and though that perhaps add it to the TransportGoal GS for starter. It is a very simple GS that is more a proof of concept than anything. Could be a good start.
21:35:42 <Zuu> An AI aware of it could value cargo capacity of engines higher for example.
21:36:18 <krinn> yep, also make some GS that dosn't ask goal to AI but only human...
21:36:36 <Zuu> One problem I realize, is that it is common for NoGo to put messages to players on the HQ, so probably a different tile has to be used.
21:36:54 <Zuu> One could use (1, 1 + company ID)
21:37:09 <Zuu> Eg, put the signs in a corner of the map
21:37:12 <krinn> well, we could define to put the sign at tile 0 per example, as the sign is per player
21:37:26 <krinn> they don't overlap no ?
21:37:59 <krinn> i mean player 1 build sign at 0, if player2 build one also at 0, you can see one or the other but both exist right ?
21:38:10 <Zuu> Hmm, yes, in plain OpenTTD I think you can put several signs ontop of each other.
21:38:26 <Zuu> Its only in SuperLib where I have tried to work around that and limit it to one sign per tile.
21:38:48 <krinn> and nogo can already knows the owner of the sign, so it should works
21:39:17 <Zuu> So we could put it at tile 0, only thing to watch out for is that depending on some advanced setting the border of the map is reserved for storage.
21:39:27 <Zuu> So using (1, 1) is probably safer.
21:39:53 <Zuu> Though some dev could probably inform us if we use a dangerous tile, and that is easy to tweak.
21:40:31 <krinn> i think as long as the sign are close to an edge it won't disturb too much
21:40:47 <krinn> and the AI/lib should discard it when read
21:41:23 <Zuu> And for those who have a problem with AIs giving "wierd" names to things, there is now a setting to hide competitor signs :-)
21:42:03 <krinn> Ah good, so we could use station names for saving so :)
21:42:59 <Zuu> hehe, and a NoGo could easily identify AIAI and CluelessPlus based on station names.
21:43:08 <krinn> i also notice the good, remove all sign on AI reload, nice
21:43:34 <krinn> Zuu, well, i have put (dictatorai) in my name for that, i don't think AI should hide their status
21:43:46 <krinn> But i suppose cluelessplus would register itself with nogo :)
21:44:13 <Zuu> Yes, I would update my AIs and NoGos to use this when we have invented the library. :-)
21:44:41 <krinn> sounds a good thing to do
21:45:06 <krinn> the faster it will be done, the faster new nogo script will handle it, AI will follow, i think at good speed
21:45:35 <krinn> specially nogo & AI share their forum
21:46:14 <krinn> once nogo & ai can speak, we could make them yell both at newGRF :) (kidding)
21:46:38 <Zuu> So, first we need a name :-)
21:48:34 <krinn> the idea, each ant work for the group :)
21:48:34 <Eddi|zuHause> but it's either AI and Goal or NoAI and NoGo
21:48:39 <Zuu> ANSI = Ai Nogo Standard Interface
21:49:00 <krinn> lol you're good at finding name zuu
21:49:23 <Zuu> well, andy has set the standards high on that point
21:51:20 <krinn> :) aren't trying to do the opposite
21:53:02 <krinn> ComPro just the short for communcation protocol
21:53:55 <Zuu> In the end, I kind of like library names that actually describe what they do :-)
21:54:24 <Zuu> ComPro make me think of HiPro, an interface to program hearing aids.
21:55:51 <krinn> hihi SCP script communication protocol (and people will think they could copy using ssh with nogo/noai)
21:56:30 <krinn> in fact, i like it too, it's simple
21:56:47 <krinn> and script include both nogo & noai
21:57:01 <Zuu> and future additional scripts that might be supported
21:57:23 <krinn> You never know with those mad devs !
21:58:20 <Zuu> Though, none of them shown any mad behaviour at the r20k party.
21:59:01 <Eddi|zuHause> not mad behaviour? were you at the same party as me?
22:00:01 <Zuu> Eddi|zuHause: Not mad as in angry. But perhaps mad as in crazy people takling about OpenTTD in a garden all night? :-)
22:02:41 <krinn> Zuu, two bytes should be enough for base : one to query/answer, another one for the query itself, and then anything need to feel the needs of the query
22:03:43 <Zuu> krinn: I will not have much time to start coding this weekend, so if you want to start coding this weekend you can do so.
22:04:40 <Zuu> Sounds nice with some vacation.
22:04:58 <krinn> but i think i could start this weekend to read nogo API better
22:05:52 <Zuu> A similar libary to what we want to do, but still not really is TileLabels.
22:07:00 <Zuu> It reads all signs at start that match a pattern and store them in memory (and remove the signs from the map). It then provides an interface for the GS to ask for the tile index that has a given name/label.
22:07:44 <Zuu> I use it to have signs in the Tutorial save game to declare the location of where the user should build the canal rather than hard coding the tile indices into the NoGo.
22:09:13 <Zuu> If we write the library for AIs and then just translate it to GS, we can re-use the translation script of SuperLib, though it is not really rocket sciense to translate between NoAI and NoGo.
22:10:14 <Zuu> But that we'll see if it is easiest to use a single code base or two separete code bases. Though I would guess a single code base for both AI and GS will help.
22:10:33 <krinn> i was thinking two would be easier
22:10:49 <krinn> with a shared file that record/define the keywords
22:11:24 <krinn> but we could do one for both yes
22:12:06 <Zuu> it needs a script to "bake" the other version, but that is how I do it for SuperLib and it works quite well.
22:12:47 <krinn> so we will do like that, specially that both should shared a lot of code
22:12:58 <Zuu> Though in our case we'll probably want to have at least one source file that is unique for each library defining the specific behaviour of the AI / Goal script edition.
22:13:51 <krinn> maybe, or we could also fake the functions and redefine them
22:13:56 <Zuu> But if you see that it is hard to do and feel like you want to do it the other way, its up to you if you start coding it. :-)
22:17:01 <krinn> isn't it easier to just extends the class ?
22:17:47 <krinn> class NCPAccounting extends AIAccounting and another file with extends GSAccounting
22:18:04 <krinn> so we just use any function with NCP in front
22:18:23 <Zuu> You need to substitute the API calls with the other prefix.
22:18:42 <krinn> yes, but we will make the call :)
22:19:18 <Zuu> You could of course define APIRoad <- AIRoad etc. for AI and with <- GSRoad for NoGo and load the correct defition depending on scipt type.
22:19:37 <Zuu> However, that pollutes the global scope so it should probably be _SCP_APIRoad or so.
22:20:23 <krinn> yep, but it would be easy, we need just two different files that define the GS or AI extends
22:21:00 <Zuu> Hmm, actually the code that reads/writes signs should probably be different as they need to be different enough as NoGo need to handle companies etc. while AI not.
22:21:33 <krinn> this could be easy handle, with default
22:22:01 <Zuu> I trust you to come up with a solution :-)
22:22:39 <krinn> function NCPSign.GetName(SignID, companyID=NCSCompany.COMPANY_SELF)
22:23:15 <krinn> nogo will pass companyID, AI won't and it will be COMPANY_SELF)
22:25:07 <Zuu> I think the SCP library should abstract away sign ids etc. and just allow a query interface. So the library will need to adress that the NoGo edition contains a scan-function that probes for incomming messages which will need to loop over all companies.
22:26:56 <Eddi|zuHause> LEX/YACC or FLEX/BISON?
22:27:05 <krinn> ah yes, so both could share the same GetName(SignID) and the nogo part lookup the company itself
22:31:05 <Zuu> Both ends might use some sort of callback mechanism to register a function to be called when the Scan method finds a message. This way a NoGo can register queries with a function to call when there is a message. I know some AIs use a callback mechanism. The GS Tutorial have a CodeStep that uses a callback function. But these things are extra when the basic library is there.
22:33:09 <Zuu> For the NoGo-side the callback could return the response.
22:33:20 <Zuu> Then the NoGo will after seting up the library only need to call the Scan function in the main loop and it should be able to handle all registred commands.
22:33:59 <krinn> well, i'm not sure of what you have in mind there
22:34:15 <krinn> the scan function can just trigger some parsing function
22:34:57 <krinn> ah yes, sorry, i see now, yes, a callback so the ai and nogo can directly handle the event
22:36:29 <krinn> you want leave the user to choose himself what name the function will be when trigger
22:38:37 <krinn> ok i'm going bed, the big lines are there, i will think about it this we, and start monday
22:40:27 <krinn> last even it appears good, the gslib-tilelabels doesn't have a license
22:41:08 <Zuu> TileLabels should be GPL v2. and I've written all code so it shouldn't be a problem. :-)
22:41:44 <krinn> ok seen, it's in the headers of the files
22:42:10 *** Devedse has joined #openttd
22:44:39 *** Djohaal has joined #openttd
22:54:58 * drac_boy is still having a hard time making sense of some of that topic
22:55:01 <drac_boy> maybe its just me -_-
22:59:56 * drac_boy goes back to more sprites work
23:37:33 <GT> In general, or about this post
23:40:11 <drac_boy> a bit of both maybe, 8bpp here forever :)
23:49:14 <Rhamphoryncus> GT: it's a bikeshed argument
23:49:40 <GT> Well, 32bpp can deliver stunning graphics, but 8bpp has its own, nostalgic atmosphere. Everyone is free to choose what he likes best.
23:50:27 <GT> What does a bikeshed argument mean?
23:50:50 <Rhamphoryncus> Google for it, but basically it's an analogy about a bike shed for a nuclear reactor
23:51:26 <Rhamphoryncus> Everybody understands a bikeshed so they all have their 2¢ to put in. Very few understand the nuclear reactor so they all stay quiet
23:51:51 <Rhamphoryncus> Yet the nuclear reactor is extremely important to get right, the bikeshed is harmless
23:53:05 <Rhamphoryncus> In this case everybody can nitpick the art (I can too), but few are capable of actually doing it
23:53:06 <drac_boy> GT I have another reason but I don't know about going into it here tho :p
23:57:26 *** zpmorgan has joined #openttd
continue to next day ⏵
