IRC logs for #openttd on OFTC at 2012-01-30
⏴ go to previous day
00:20:56 <DDR> Hey, guys, can anyone think of a mod that'll allow me to play starting from 1700 or so?
00:21:02 <DDR> I'm in the mood for a long game.
00:21:19 <DDR> What happened to the horse-and-buggys, anyway?
00:24:12 * Rhamphoryncus votes for his own patch :D
00:24:51 <Rhamphoryncus> Rubidium: There's my first patch sent. Now you can reap what you've sown. ;)
00:50:13 *** kkb110_ has joined #openttd
01:04:23 *** supermop_ has joined #openttd
01:04:53 *** fray`007 has joined #openttd
01:46:30 *** kkb110_ has joined #openttd
02:49:37 *** supermop has joined #openttd
03:15:28 *** Stimrol has joined #openttd
03:31:02 *** Markavian` has joined #openttd
04:19:11 *** Markavian has joined #openttd
04:24:44 *** cmircea has joined #openttd
04:47:14 *** fray`007 is now known as bondau
05:11:55 *** bondmain_ has joined #openttd
05:21:33 *** tokai|noir has joined #openttd
05:21:33 *** ChanServ sets mode: +v tokai|noir
05:28:52 *** ChanServ sets mode: +v tokai
05:56:24 *** Eddi|zuHause2 has joined #openttd
06:04:23 *** roboboy has joined #openttd
06:05:50 *** Eddi|zuHause2 is now known as Eddi|zuHause
06:26:46 *** Markavian has joined #openttd
06:33:19 *** JVassie has joined #openttd
07:20:51 *** mahmoud has joined #openttd
07:24:43 *** Cybertinus has joined #openttd
07:29:11 *** sla_ro|master has joined #openttd
07:47:25 *** Progman has joined #openttd
07:49:36 *** andythenorth has joined #openttd
07:50:31 <andythenorth> a python module that parses nml item blocks into first class objects
07:50:56 <andythenorth> thereby exposing them for templating engines
07:51:43 <andythenorth> the item block looks like a dict anyway....someone stupid like me might just call eval() on it and see what happens :o
07:54:21 <Rhamphoryncus> if they're that close would json work? Much better than eval
07:55:47 <andythenorth> load them as json?
07:56:04 <andythenorth> I've solved it for my set
07:56:16 <andythenorth> just wondering if there's a way to do it with fewer intermediate steps
07:56:30 <andythenorth> and in a way that's easy for other people to pickup and use
08:02:48 *** DayDreamer has joined #openttd
08:07:29 <Eddi|zuHause> andythenorth: if you want an nml parser, nmlc has one of those
08:11:18 <whsck> Hello could anyone help with an error?
08:11:49 <DorpsGek> planetmaker: Don't ask to ask, just ask
08:12:41 <whsck> "Network Synchronization Failed" - Multiplayer Game, with 4 friends over Lan(hamachi) - Everyone else is fine, however around 1960 I always get that message
08:13:07 <whsck> Then every log in, disconects me straight away, works fine in single player
08:16:27 <Eddi|zuHause> then pause the server in that state. the devs probably need "live" access to such a game
08:16:43 <Eddi|zuHause> and also tell us the version you're playing
08:18:02 <Eddi|zuHause> and if possible, keep around autosaves from before this happened
08:19:55 <Eddi|zuHause> with as much detail as you can get
08:21:11 *** cmircea has joined #openttd
08:22:49 <planetmaker> yes, please, what eddi says
08:23:52 <planetmaker> especially we need a savegame
08:25:08 <planetmaker> we need a way to actually trigger the same desync. That probably will prove difficult... does it trigger on any specific action or upon join?
08:26:32 <whsck> Well its fine for about 10years and the other players are fine, Nothing triggers it from what I can see, everytime its happend i've been doing somthing different, after that I try rejoin and it d/c's me straight away with Network Synch Failed
08:26:47 <planetmaker> did you build something the others did not build?
08:27:01 <whsck> Everything i've built everyone has
08:27:25 *** Devroush has joined #openttd
08:28:03 <planetmaker> right... there's something the host could do, but it requires possibly lots of harddisk space
08:28:10 <planetmaker> he could turn on desync debugging
08:28:29 <planetmaker> rcon debuglevel desync=3
08:28:50 <planetmaker> if it's a non-dedicated server he might not even need the 'rcon' part
08:29:22 <whsck> What shall I do once thats happend?
08:29:23 <planetmaker> and then save the game under a easy-to-remember name and load that savegame again
08:29:39 <planetmaker> and then continue playing as normal
08:29:59 <planetmaker> and then make a savegame again when a desync triggers
08:30:32 <planetmaker> and then the host has to provide us with at least: the initial savegame, the last savegame and the command.log which is found in the autosave folder
08:30:40 <whsck> Shall i do this before making the report then?
08:31:11 <whsck> Shall I make the report now or wait to add the Desync debugging
08:31:24 <planetmaker> if you trigger then the desync again, you can add that info the the report you make now
08:32:15 <whsck> Okay thank you, will create the report later with all the information when hes back
08:34:10 <planetmaker> well, as said: there's no harm to create the report now, with a current savegame and amend the missing info later
08:34:43 <whsck> I mean I have the latest save file of the game but it wont trigger the desycnh on single player, only multiplayer
08:34:51 <planetmaker> though the method I described often is the only way to really get hold of a desync
08:35:01 <andythenorth> Eddi|zuHause: nmlc parser :) good point
08:35:06 <planetmaker> a desync cannot trigger in singleplayer
08:35:12 <planetmaker> it's a multiplayer-only bug ;-)
08:35:25 <whsck> oh yeah obviously, dumb me
08:35:39 <planetmaker> which means that server and client have different ideas of what the game state looks like...
08:36:48 <planetmaker> which means that server and client have different ideas of what the game state looks like...
08:36:59 * andythenorth explores in nml module-land
08:38:38 <andythenorth> probably...nmlc parser.py could be used to instantiate python objects for each item, which are then passed to [templating engine of choice]
08:39:00 <andythenorth> meaning that all the properties for a vehicle, house etc are simply defined in nml, no extra boilerplate needed
08:39:20 <andythenorth> the templating engine would fill in the switch chains for each item
08:39:32 <andythenorth> ...but I'm not rebuilding BANDIT to v6 to test that :P
08:40:56 <whsck> Could my issues be because they're on windows 7 and im on XP?
08:42:08 <andythenorth> we could give each item block an additional dict 'extended_properties' or 'arbitrary_properties' or such, these are constants used only during build
08:42:09 * Rhamphoryncus wants a singleplayer desync, just for the wtf value ;)
08:43:07 <Rhamphoryncus> whsck: I wouldn't expect though, especially if you haven't compiled openttd yourself
09:03:26 *** Brianetta has joined #openttd
09:04:33 <andythenorth> Eddi|zuHause: this scopes identifiers to the current vehicle:
09:04:35 <andythenorth> switch (FEAT_ROADVEHS, SELF, ${trailer.id}_cargo_switch, cargo_classes) {
09:05:12 <Eddi|zuHause> andythenorth: that doesn't really offer an improvement over what i have now
09:05:51 <andythenorth> you want nml to handle the local scope for you?
09:09:07 <andythenorth> Eddi|zuHause: if that is your goal(?), the only thing I can think of is that switch identifiers are local to the item by default, and you have to call a global namespace for those that aren't
09:09:48 <Eddi|zuHause> switches aren't part of items
09:10:30 <andythenorth> so it would be a headache :)
09:10:58 <andythenorth> you'd have to add delimiters or set scopes or something
09:11:58 <andythenorth> I can't think of anything except concatenation of (item_identifier + foo_string) - by python, cpp variadic macro, carrier pigeon etc.
09:51:34 *** andythenorth has joined #openttd
09:51:51 <andythenorth> "Eddi|zuHause: switches aren't part of items" <- actually, you know the call chain though?
09:52:01 <andythenorth> switch has an item as caller
09:52:22 <Eddi|zuHause> no, you only know the call chain _after_ you resolved the identifiers
09:52:28 <Eddi|zuHause> that's what they're there for
09:52:33 <Eddi|zuHause> they have no other purpose
09:52:45 * andythenorth made a silly mistake there :)
09:53:05 <andythenorth> file under 'chicken and egg'
09:53:21 * andythenorth thinks of something evil
10:03:43 <xiong> Elu, Mm-m, I'm starting to think it's not as clear-cut as we discussed yesterday. The more capacity on a line, the more cargo that cargodist will assign to it. If player over-serves an unprofitable line; then that may steal cargo from the more profitable, less serviced run.
10:16:53 <Elukka> i haven't noticed that
10:17:20 <Elukka> just thousands and thousands of passengers waiting for the insufficiently serviced route
10:18:14 *** cornjuliox has joined #openttd
10:29:59 *** Arafangion has joined #openttd
11:02:14 *** Markavian` has joined #openttd
11:09:20 *** danielon has joined #openttd
11:15:19 <__ln__> muchas millones de gente, especialmente en españa.
11:15:57 <Eddi|zuHause> actually, spanish is like the third most spoken language on the planet
11:16:37 <andythenorth> ~even chance it will be the most widely spoken language in the USA
11:16:42 <__ln__> Eddi|zuHause: very true, but i assumed español meant nationality rather than language here.
11:17:03 <Elukka> just a little bit unlikely, andy
11:17:26 <__ln__> Elukka: why's that? isn't 30 million a good start?
11:17:27 <danielon> I will use the google translator
11:17:31 <Eddi|zuHause> andythenorth: in the south, probably soon. other regions, i don't think so
11:18:00 <Elukka> it's a significant minority language, naturally since there are a lot of spanish speakers on the continent
11:18:10 <Elukka> it's not going to somehow supplant english though
11:18:47 <danielon> I am trying to create a multiplayer game but I can not be published, I mean I can not get out on the page server
11:20:01 <Eddi|zuHause> you have to do port forwarding in your firewall and router
11:20:04 <DorpsGek> Eddi|zuHause: OpenTTD uses TCP and UDP port 3979 for server <-> client communication, UDP port 3978 for masterserver (advertise) communication (outbound), and TCP port 3978 for content service, a.k.a. BaNaNaS (outbound)
11:20:09 <__ln__> danielon: ¿no te has aprendido inglés en la escuela?
11:21:00 <danielon> esos puertos ya los he avierto pero no lo logro
11:21:23 * MNIM whacks __ln__ and danielon on the head
11:21:25 <Eddi|zuHause> make sure you open _both_ TCP and UDP
11:21:28 <__ln__> danielon: ¿y otras idiomas extrañeras?
11:21:30 <MNIM> don't make me quote pulp fiction
11:21:43 * andythenorth should find the stats on spanish
11:21:55 <Elukka> why's it matter so much
11:22:07 <andythenorth> it's quite qualified - something like 'spoken, but not the primary language'
11:23:00 <Elukka> kinda like swedish here in finland but that gets official language status for some reason :P
11:24:27 <Eddi|zuHause> germany has a few official minority languages
11:25:08 <__ln__> andythenorth: how rapidly does it increase?
11:25:25 <DorpsGek> __ln__: I have not seen wikipedia.
11:25:46 <Eddi|zuHause> "In Deutschland fallen Nordfriesisch, Saterfriesisch, Dänisch, Sorbisch und Romani unter die Definition einer Minderheitensprache."
11:27:22 <Elukka> it would not reach 50% if the entirety of mexico decided to emigrate, if that's what you're wondering :P
11:28:14 <Eddi|zuHause> Elukka: but the spanish speaking population in the USA also has the highest birth rate, not just immigration rate
11:28:47 <Elukka> still zero chance of it becoming a majority language
11:31:56 <Eddi|zuHause> so... why does my kopete not connect for two days?
11:32:09 <__ln__> Elukka: 5% is enough to have all official documents to be issued in two languages, requiring all people to learn the other language, etc.
11:33:10 <Elukka> pretty sure they don't currently issue all documents in spanish and have mandatory spanish at school for everyone, though?
11:33:35 <Eddi|zuHause> Elukka: in the USA that is managed on state level.
11:33:43 <Noldo> do you live there Elukka?
11:34:00 <Eddi|zuHause> Elukka: so states with a significant minority of spanish speaking people do have spanish as official language
11:34:28 <Elukka> eddi, didn't think of that
11:34:30 <Eddi|zuHause> Elukka: the USA as a whole don't have an official language at all
11:35:06 <Eddi|zuHause> english is a de-facto official language on federal level, but there's no law defining it
11:44:33 <andythenorth> Eddi|zuHause: I may still be making silly mistakes - but namespaces? Forget concatenating the identifiers, just write them out (each must be unique within the project), but scope them to the calling item?
11:45:22 <Eddi|zuHause> i don't get what you mean
11:45:54 <andythenorth> nvm, I'm probably smoking crack
11:46:15 <andythenorth> it's not even a problem I need to sold, and I'm not good at comp sci :P
11:51:17 * andythenorth should read nml src instead of guessing
11:53:12 * andythenorth is presuming that a mapping could be constructed of {item: (switch_identifiers_called_by_item)}
11:53:27 <Elukka> and now i have to download space engine from a glacially slow russian server...
11:56:03 <Elukka> everything is going to be a painful mess of broken links if this keeps going
11:58:03 <andythenorth> with a mapping, the identifiers could be rewritten relatively trivially
11:58:27 <andythenorth> but this is all guesswork
12:00:34 <Elukka> procedurally generated universe pretties
12:05:32 *** TGYoshi has joined #openttd
12:17:09 <peter1138> new landscape generator!
12:23:57 *** roboboy has joined #openttd
12:39:58 <andythenorth> multi-stop docks!
12:40:13 * andythenorth has been too busy writing actual code to make tiresome feature requests :P
12:41:32 <Eddi|zuHause> for various definitions of "actual" and "code" :p
12:43:09 * andythenorth struggles for comeback to that
12:51:24 * peter1138 commits andythenorth
12:56:22 <Rhamphoryncus> Hrm, expanding to a double mainline worked a little too well. Now all my stations are bogged down with extra trains that used to be bogged down on the mainline.
12:58:42 *** tokai|mdlx has joined #openttd
13:18:36 <Arafangion> Are they in-out stations?
13:24:56 <Elukka> space engine looks amazing
13:24:59 <Elukka> i urge everyone to try
13:28:40 <__ln__> not very photorealistic
13:29:30 <Elukka> a sun is rising from behind another sun
13:33:35 <Rhamphoryncus> Arafangion: roros, 8 or 12 wide
13:34:05 <andythenorth> revert andythenorth
13:34:48 <Rhamphoryncus> and now I'm switching a bunch of my path signals to 1-way because one confused little train managed to choose a path backwards, halfway across the map (no exaggeration), across both sides of my double mainline
13:52:46 *** Asteconn has joined #openttd
13:54:58 *** mahmoud has joined #openttd
14:02:07 <planetmaker> <Elukka> a sun is rising from behind another sun <--- there are confirmed planets out there which orbit a multiple star system
14:02:34 <Elukka> i've never seen a space thing as pretty as this program
14:02:46 <Elukka> random desert planet with thick cloud cover
14:02:50 <Elukka> there are dry riverbeds on the surface
14:03:23 <planetmaker> they call it an image of Mars
14:04:34 <planetmaker> the image you linked could be of Mars
14:04:50 <Elukka> it's procedurally generated
14:06:19 <Elukka> i think it has about the real universe's worth of content :P
14:30:33 <blathijs> wth, looks terrible to script ;_)
14:47:57 *** bondmain_ has joined #openttd
15:12:22 *** TWerkhoven[l] has joined #openttd
15:30:11 *** DayDreamer has joined #openttd
15:30:49 *** DayDreamer has left #openttd
15:40:51 *** sla_ro|master has joined #openttd
15:45:27 *** bondmain_ has joined #openttd
16:18:11 *** frosch123 has joined #openttd
16:43:08 *** TWerkhoven[l] has joined #openttd
16:47:35 *** supermop has joined #openttd
16:49:15 *** bondmain_ has joined #openttd
16:52:10 *** Biolunar has joined #openttd
16:59:50 *** |Jeroen| has joined #openttd
17:28:39 *** Brianetta has joined #openttd
17:30:13 *** theholyduck has joined #openttd
17:40:36 *** bondmain_ has joined #openttd
17:43:49 *** Asteconn has joined #openttd
17:45:49 *** cmircea has joined #openttd
18:16:56 *** Prof_Frink has joined #openttd
18:31:42 *** TheMask96 has joined #openttd
18:46:11 <CIA-1> OpenTTD: translators * r23872 /trunk/src/lang/ (10 files): (log message trimmed)
18:46:11 <CIA-1> OpenTTD: -Update from WebTranslator v3.0:
18:46:11 <CIA-1> OpenTTD: simplified_chinese - 1 changes by chenwt0315
18:46:11 <CIA-1> OpenTTD: finnish - 1 changes by jpx_
18:46:11 <CIA-1> OpenTTD: french - 23 changes by OliTTD
18:46:13 <CIA-1> OpenTTD: german - 2 changes by planetmaker
18:46:13 <CIA-1> OpenTTD: hungarian - 12 changes by IPG
18:49:51 *** lollercaust has joined #openttd
18:50:08 *** bondmain_ has joined #openttd
18:51:45 *** lollercaust has joined #openttd
18:57:52 *** lollercaust has joined #openttd
19:06:32 *** lollercaust has joined #openttd
19:06:42 *** DayDreamer has joined #openttd
19:06:48 *** DayDreamer has left #openttd
19:08:51 *** HerzogDeXtEr has joined #openttd
19:11:22 *** JVassie has joined #openttd
19:16:51 *** andythenorth has joined #openttd
19:18:02 *** andythenorth has joined #openttd
19:19:48 <andythenorth> Yexo: did you have any design for file inclusion in nml? Or just an idea that it might be useful?
19:20:45 <Yexo> nothing yet except the idea
19:21:28 <andythenorth> the common case I see for it is including switch chains - sound right?
19:42:01 <supermop> sorry i keep interjecting with these totally random ideas,
19:42:45 <supermop> but does fuel have any ingame effect other than picture next to the buy-list sprite?
19:43:19 <supermop> like, is there any difference between steam and diesel other than smoke animation?
19:46:22 <frosch123> steam/diesel has no effect by itself
19:46:39 <frosch123> but usually sets will offer different running cost, power and te
19:51:12 <supermop> but there is no way for a set to say 'this vehicle has property coal' and then provide a separate curve for running cost to grow over time vis-a-vis 'oil' or 'diesel'
20:07:41 <andythenorth> supermop: that's the domain of the newgrf
20:07:45 <andythenorth> pikka does it in NARS 2
20:08:13 <andythenorth> I might do it in FISH, but only if I can be arsed
20:09:13 <supermop> but it is possible for those who can be bothered coding it
20:13:09 <SmatZ> I mean, why would anyone give such a photo on a dating site...
20:13:52 <andythenorth> supermop: yes it's possible
20:14:08 <andythenorth> Yexo: file inclusion could use a notation something like: include({base : foo_vehicle}, "sprites/nml/some_file.nml")
20:14:31 <andythenorth> 'base' could then be used to concatenate with any identifier using a control character, maybe $ or such
20:15:57 <andythenorth> switch chain identifiers are then sorted
20:16:14 <andythenorth> without the headache of a full templating / macro language
20:29:50 *** theholyduck has joined #openttd
20:34:20 *** brendan10211 has joined #openttd
20:51:37 *** tokai|noir has joined #openttd
20:51:37 *** ChanServ sets mode: +v tokai|noir
20:57:21 <planetmaker> SmatZ, dear SmatZ, should you visit dating sites?
20:58:57 <V453000> this is basically just another facebook
20:58:59 *** DayDreamer has joined #openttd
21:07:14 <frosch123> there are dating sites with public pictures?
21:47:16 *** lollercaust has joined #openttd
22:16:02 <Yexo> andythenorth: a much simpler solution would be to make identifiers file-local (at least optionally)
22:16:38 <Yexo> for example if you do switch (..., $my_switch) you can use $my_switch everywhere in the same file, but in other files it'd be invalid (unless you also specified $my_switch there too)
22:17:34 *** theholyduck has joined #openttd
22:24:12 <andythenorth> Yexo: if I understand that correctly, it's pretty neat
22:25:27 <andythenorth> this is within the same nml file?
22:26:14 <Yexo> what I basically meant was: let nml replace $ (or another sign) by the filename if it's used as prefix for an identifier
22:26:58 <Yexo> $my_switch would become spritesnmlsome_filemy_switch, or sprites_nml_some_fille_my_switch or something like that
22:27:11 <Yexo> or more likely some internal name that will never clash with user defined names
22:29:26 <andythenorth> definitely improves on the current situation
22:29:33 <Arafangion> Ie, a poor man's namespace.
22:29:43 <andythenorth> what about when an nml file is assembled by cpp?
22:29:54 <andythenorth> does it fail that case?
22:30:04 <Yexo> dunno, depends on the implementation
22:30:13 * andythenorth thinks cpp smells wrong for nml, but it's common useage
22:30:24 <Eddi|zuHause> hm... in hg, what's the easiest way to select individual chunks for (not-)commiting?
22:30:34 <Arafangion> You could invent a concept of modules, ala python.
22:30:50 <Arafangion> Everything's global - but only within that one file.
22:31:06 <Yexo> Arafangion: can't do that, the final result has to be compiled to nfo
22:31:17 <Yexo> so there are quite a lot of limitations, like everything stored in parameters is always global
22:31:21 <Yexo> I can't limit that per-file
22:31:39 <andythenorth> Yexo: the thing I really missed initially was repeating varaction2 IDs in templates that are included
22:31:47 <andythenorth> the variadic macros are one solution
22:32:05 <andythenorth> I've got a solution I like now
22:32:17 <Arafangion> Yexo: Is the limitation making it accessible to C++, or accessible to other nfo's?
22:32:29 <Yexo> if those repeated varaction2's are always the same you could try including them only once
22:32:45 <andythenorth> they would usually have stuff inserted with cpp constants
22:32:52 <Yexo> Arafangion: nml is compiled to a grf file, so it's bound by the limitations in the grf spec
22:33:03 <Arafangion> Ah, your hands are tied, then. :)
22:33:06 <andythenorth> tbh, I think it was a pattern for nfo + cpp
22:33:11 * Arafangion heads off to work, to find another place where he can pretend to be an expert.
22:34:20 * Rubidium heads off to some other kind of "work" ;)
22:36:10 <Eddi|zuHause> simply "repeating" varaction2-chains for each "item-template" may quickly run into the next limitation, so it may not be of much use
22:37:20 <Eddi|zuHause> for example i currently need 3 separate action2-chains per vehicle
22:37:38 <Eddi|zuHause> one for displaying the vehicle as itself
22:38:05 <Eddi|zuHause> one for displaying the vehicle as a slice of the adjacent vehicle (same structure, but using var61/62)
22:38:21 <Eddi|zuHause> and one for purchase list (avoiding some variables)
22:40:30 <Eddi|zuHause> the generator that makes this duplication is currently one of the least elegant parts of the code
22:40:37 <andythenorth> Eddi|zuHause: which file?
22:41:48 <Eddi|zuHause> night to all blitzquitters.
22:41:50 *** Eddi|zuHause has left #openttd
22:42:01 *** Eddi|zuHause has joined #openttd
22:43:12 * andythenorth adventures in CETS code
22:43:28 <Eddi|zuHause> the code i spoke about is in tree.py
22:44:14 <Eddi|zuHause> if you find a point that should be better documented, please tell
22:44:46 * andythenorth has some reading to do :P
22:45:14 <Eddi|zuHause> that file could probably use some splitting
22:45:18 <andythenorth> any particular lines in this code to look at?
22:45:45 <andythenorth> I have to see if your problem is template shaped. Otherwise you'll shout at me when I recommend one :P
22:45:47 <Eddi|zuHause> i.e. separating the "conversion" functions
22:46:21 <Eddi|zuHause> andythenorth: since it's a tree, it's likely not solvable with templates
22:47:21 * andythenorth is busy being impressed by it
22:47:40 <Eddi|zuHause> andythenorth: the input is a string of nested expressions of the form "condition:(value1:expression1,value2:expression2,...)"
22:48:07 <Eddi|zuHause> where the leafs reference actual image files
22:48:15 <Eddi|zuHause> by some naming convention
22:49:58 <Eddi|zuHause> conditon can be "date" (aka date of last service), "epochN" (referencing the parameter setting) and some select others which i forgot
22:51:15 <Eddi|zuHause> and "value" can be a single number, a range in the form of "a..b" or in the form of ">a" (and other comparison operators)
22:51:33 * andythenorth is mentally reconstructing what actually comes out of the file.write lines
22:51:58 <Eddi|zuHause> each node generates one switch
22:52:11 <Eddi|zuHause> or three, in the case of "epochN"
22:52:53 <Eddi|zuHause> "prefix" and "postfix" are some magical values that get passed through to make the identifiers unique
22:54:13 <andythenorth> so the number of switches needed is deterministic, but variable according to conditions?
22:54:35 <Eddi|zuHause> prefix is something like "gfx_switch" or "purchase_switch"
22:54:49 <andythenorth> and the postfix is a number?
22:54:57 <Eddi|zuHause> postfix is something that is appended on each tree level
22:55:11 <Eddi|zuHause> so the first level enumerates the children as _0 to _n
22:55:40 <Eddi|zuHause> and the second level makes _0_0 to _0_k and _1_0 to _1_l etc
22:56:55 <andythenorth> this relies an a python object being in scope with various attributes defined
22:57:12 <andythenorth> <tal:foo stuff> is just a naming convention in the templater
22:58:20 <Eddi|zuHause> input strings are something like: "date:(default:pruss,1927..1941:DRG,1942..1949:pruss,>1949:epoch3:(DB:DB,DR:DR,default:DB))"
22:58:27 <andythenorth> I've got the same prefix / postfix pattern as you :P
22:59:09 <Eddi|zuHause> so this creates a switch for the date, with 4 children
22:59:18 <andythenorth> and you pass those to translate_date_short etc?
22:59:31 <Eddi|zuHause> and the 4th child creates another switch which checks the epoch3 parameter
23:00:00 <Eddi|zuHause> well, first this is shoved through parse.py
23:00:05 <Eddi|zuHause> to create the tree objects
23:00:31 <Eddi|zuHause> which is your usual ply stuff
23:01:13 <Eddi|zuHause> pretty simple if you ever have attended a computer science lecture on compiler construction :)
23:01:56 * andythenorth got failed for the pascal course in 1st year engineering
23:02:00 <Eddi|zuHause> the "translate" functions are called for the "value" parts of the string
23:02:08 * andythenorth then quit engineering
23:02:28 <Eddi|zuHause> i.e. it translates ">1949" into something that the nml-parser would understand
23:02:54 <andythenorth> so how many types of switch get generated by this?
23:02:57 <Eddi|zuHause> in this case something like "date(1,1,1949)..date(1,1,5000000)"
23:03:08 <andythenorth> I mean significantly different, not trivially different
23:03:42 <Eddi|zuHause> typically the "translate function" and the checked variable/expression are exchanged
23:04:01 <Eddi|zuHause> possibly the reference (SAME/PARENT)
23:04:19 <Eddi|zuHause> and occasionally a random_switch is generated
23:05:04 <Eddi|zuHause> this is handled by calling the constructor for Switch, ParentSwitch or RandomSwitch
23:06:01 *** eQualizer has joined #openttd
23:07:02 <andythenorth> CETS won't build
23:07:14 <andythenorth> how dull - I wanted to see the processed nml
23:07:37 <Eddi|zuHause> you need ply, we discussed that the last time
23:08:26 <andythenorth> maybe I miss something else
23:14:43 <Eddi|zuHause> andythenorth: dunno what that error means, never seen that before
23:15:01 <andythenorth> Eddi|zuHause: what's your python version?
23:16:18 <andythenorth> "The p_error(p) rule is defined to catch syntax errors. See the error handling section below for more detail."
23:16:32 <Eddi|zuHause> that one's irrelavnt
23:16:50 <Eddi|zuHause> can you do the following
23:16:59 <Eddi|zuHause> "python -i scripts/generate.py"
23:17:33 <Eddi|zuHause> "p tables.table[key]"
23:19:32 <Eddi|zuHause> you skipped one :)
23:20:33 * andythenorth is looking for a missing ''
23:21:35 <Eddi|zuHause> i don't see anything that could be wrong there
23:22:41 *** Rhamphoryncus has joined #openttd
23:25:24 <Eddi|zuHause> not entirely sure where the "":"" comes from, though
23:25:53 <andythenorth> I've not seen these structures before
23:26:01 <andythenorth> it's like a dict with prefixes on the keys?
23:27:49 <Eddi|zuHause> there's also r ("raw", don't process escapes)
23:29:00 <Eddi|zuHause> it's easier to just write u"ö" instead of doing all kinds of conversion stuff
23:31:37 <Eddi|zuHause> "ö" varies depending on locale/encoding, while u"ö" always results in the same string, no matter what your encoding is
23:32:07 <Rhamphoryncus> "ö" is actually sort of a bug. It should never have been allowed by the language
23:32:29 <andythenorth> yup, the python cms I use, we often encode in unicode, otherwise sadness occurs
23:34:11 <Rhamphoryncus> It's a good idea to explicitly mark each file as utf-8 too
23:34:41 <Eddi|zuHause> well, python rejects the file if no encoding is stated
23:35:25 <Rhamphoryncus> hrm. Next time I'll spend more than 30 seconds checking my facts. ;)
23:36:28 <andythenorth> Eddi|zuHause: if I remove the value {...} from that structure, I don't get a syntax error from it
23:36:33 <andythenorth> but I don't know if that's the issue
23:37:19 <Eddi|zuHause> andythenorth: can you check, in read.py change the "value" into u"value"?
23:38:32 <andythenorth> also the first key is an empty string, is that allowed?
23:38:56 <Eddi|zuHause> "it never complained before"
23:39:22 <Eddi|zuHause> some keys are not identifier-able, that never stopped it before either
23:39:46 <andythenorth> my python is 2.6
23:39:57 <Eddi|zuHause> "values" is actually a self-reference
23:40:36 <Eddi|zuHause> it's kind of important that it is
23:42:48 <andythenorth> Eddi|zuHause: I need to go to sleep :)
23:42:55 <andythenorth> maybe this is only an issue for me?
23:43:08 <andythenorth> devzone builds cets ok?
23:43:10 <Rhamphoryncus> Got a link so I can look at it?
23:46:36 <andythenorth> Rhamphoryncus: specifically my python is unhappy with 'values' : {....}
23:46:54 <Rhamphoryncus> You're extracting it in to keyword arguments?
23:48:15 <Rhamphoryncus> And you already have a keyword argument named values
23:48:15 <andythenorth> looks like it: process(key, **tables.table[key])
23:48:36 <Mazur> Maybe you should feed i eggs or mice, instead.
23:48:49 <Rhamphoryncus> solution: don't pass the dict as keyword arguments when it's arbitrary like that
23:48:56 <Rhamphoryncus> So remove the ** on the call and in the functions
23:49:54 * Rhamphoryncus mentions that 30 seconds thing again ;)
23:50:50 <Rhamphoryncus> I wouldn't use unicode for a variable name, not in 2.x
23:51:25 <andythenorth> Eddi|zuHause: if it's bugging me enough tomorrow, I might install python 2.7 and see what happens
23:51:43 <andythenorth> if I can be bothered to revisit the fun of having multiple python versions :o
23:52:43 <Rhamphoryncus> I used to have tons, but ubuntu removed the really old ones on me. Something about no longer patching security flaws ;)
23:56:07 *** Markavian has joined #openttd
continue to next day ⏵
