IRC logs for #openttd on OFTC at 2012-01-27
⏴ go to previous day
00:03:34 *** Asteconn has joined #openttd
00:10:47 *** roboboy has joined #openttd
00:14:54 <Rhamphoryncus> tropic refurbishment set.. has a train that can do 300 km/h.. but no cars that'll do that x_x
00:19:03 <Rhamphoryncus> Heh. Long vehicles has a semi. Should be nice.. but they're not articulated and they're still treated as tiny so they overlap
00:38:38 *** Asteconn has joined #openttd
00:44:01 <Rhamphoryncus> Hah, generated a subtropic map with a 110 tile sea gap down the middle. There is a spot with an island where you could do a 45 and 70 tile bridge though
01:46:04 *** Xrufuian has joined #openttd
02:54:16 *** kkb110_ has joined #openttd
02:59:11 *** Asteconn has joined #openttd
03:53:02 *** cornjuliox has joined #openttd
04:16:23 *** cmircea has joined #openttd
04:52:26 <Rhamphoryncus> Turns out water is worth a lot more if you deliver it to a water tower rather than a factory outside of town x_x
05:09:49 *** tokai|noir has joined #openttd
05:09:49 *** ChanServ sets mode: +v tokai|noir
05:12:43 *** SirWaddles has joined #openttd
05:20:48 *** tokai|mdlx has joined #openttd
05:21:52 *** Rhamphoryncus has joined #openttd
05:21:52 *** theholyduck has joined #openttd
05:21:52 *** glevans2 has joined #openttd
05:21:52 *** Rubidium has joined #openttd
05:21:52 *** Maarten has joined #openttd
05:21:52 *** tparker has joined #openttd
05:21:52 *** ccfreak2k has joined #openttd
05:21:52 *** mikegrb has joined #openttd
05:21:52 *** synthon.oftc.net sets mode: +ovov Rubidium Rubidium orudge orudge
05:21:52 *** Belugas has joined #openttd
05:21:52 *** Born_Acorn has joined #openttd
05:21:52 *** TinoDidriksen has joined #openttd
05:21:52 *** synthon.oftc.net sets mode: +ov Belugas Belugas
05:22:43 *** ChanServ sets mode: +v Terkhen
05:22:43 *** ChanServ sets mode: +v planetmaker
05:22:43 *** ChanServ sets mode: +v Yexo
05:22:43 *** ChanServ sets mode: +v SmatZ
05:56:18 *** Eddi|zuHause has joined #openttd
05:56:18 *** cmircea has joined #openttd
05:56:29 <Asteconn> Dia daoibh a chairde!
05:58:18 <Asteconn> I have a gameplay question
05:58:33 <Asteconn> What's the best method to maximise returns on train passenger transport?
05:59:08 <Asteconn> I mean, I can almost always run a profit, if sometimes that profit is only just above breaking even
05:59:24 <Asteconn> But I'd like to maximise my returns and such
06:16:54 <Rhamphoryncus> Mostly I find passengers dissatisfying. I usually only provide service to pacify the town, not to get a profit
06:18:50 <Rhamphoryncus> Although atm I'm using aviator set with 1/1 speed and turning a decent profit. Unless I broke it by not setting it to 1/1 until after I started (it has a warning about really wanting 1/1 speed)
06:24:03 <Asteconn> I do enjoy providing passenger train transport actually
06:24:46 <Asteconn> Sometimes I get §100000+ yearly (no inflation) profits for some trains, but I have no idea how I accomplish this.
06:24:54 <Asteconn> Aside from, as you suggest, going quickly.
06:25:41 <Asteconn> But then again, I get some trains running at 100mph returning 60k+ profits sometimes, and others running at 165mph with profits of only a couple of thousand.
06:25:51 *** andythenorth has joined #openttd
06:26:19 <Rhamphoryncus> Are they full both ways?
06:26:49 <Rhamphoryncus> There's actually an optimal distance for a given speed
06:27:10 <Asteconn> Well - that doesn't surprise me too much, but please, do tell :D
06:29:19 <Rhamphoryncus> Oh, an optimal *time* in transit. From that you want the furthest distance possible
06:30:32 <Asteconn> A friend of mine I play online with often, he builds airports and sends concordes from one end of the map to the other, makes a fortune doing so xD
06:30:57 <Rhamphoryncus> I love doing that :)
06:35:35 <Asteconn> I play for the trains though. I have a feeling that I'm a closet gricer or something xD
06:37:23 *** JVassie has joined #openttd
06:44:59 <Asteconn> Another pertinent question then - how best to fill my trains up, but still keeping a decent service frequency?
06:46:17 <Eddi|zuHause> make sure your next train arrives around the time the first one gets full
06:47:51 <Rhamphoryncus> If you look earlier in that page you'll see a list of factors for station rating. The ones you can have consistently are high vehicle speed, low days since last pickup (sitting on full load works), amount of cargo waiting, and having a statue
06:48:42 <andythenorth> mako or chameleon?
06:49:12 <Rhamphoryncus> andythenorth: chameleons are for training...
06:50:09 <Rhamphoryncus> Asteconn: So if your train does 321 km/h, has less than 7.5 days between pickups, leaves less than 100 in the station at any time, and there's a statue, in theory you should get 100%
06:50:55 <Eddi|zuHause> i think there's a maximum for the speed, beyond that it's ignored
06:51:05 <Asteconn> Does a train waiting at a station as a result of timetabling count?
06:52:38 <Asteconn> Recently I've started to set all of my passenger trains to wait at the station for a bit
06:52:51 <andythenorth> Eddi|zuHause: is all CETS templating cpp?
06:52:59 <andythenorth> do you use any python templating
06:53:28 <Eddi|zuHause> andythenorth: the properties are in python
06:53:45 <Eddi|zuHause> andythenorth: and the graphics are ... complex
06:53:46 <andythenorth> can you name a file with examples?
06:54:02 <andythenorth> I want to check I'm not re-solving something you solved already
06:54:04 <Eddi|zuHause> andythenorth: the properties are handled in write.py
06:54:37 <Eddi|zuHause> andythenorth: the graphics are handled in tree.py
06:54:49 <andythenorth> you're using % mostly?
06:55:40 <andythenorth> I've tried the $ templating
06:55:59 <andythenorth> but including one template into another is clunky
06:56:20 <Eddi|zuHause> not sure what you mean
06:56:36 *** Cybertinus has joined #openttd
06:57:32 <andythenorth> currently it's a horrible mix of cpp and $ templates. I'm replacing the cpp
06:58:15 <Eddi|zuHause> aha, never seen that
06:58:39 <Eddi|zuHause> yes. it'd make sense to get rid of the CPP then
06:58:59 <Eddi|zuHause> besides #include, i use very little CPP
07:00:13 <Eddi|zuHause> for the vehicle-identifiers and for "named constants" (which nml unfortunately doesn't support)
07:00:32 <Eddi|zuHause> and some magic for the vehicle slicing
07:01:26 <andythenorth> ignore the html/xml note - it's fine to work with plain text, not nodes
07:02:17 <Eddi|zuHause> "for truck_id in trucks: truck = trucks[truck_id]" <- i'd use "for idtruck_, truck in trucks.iiteritems()"
07:02:28 <andythenorth> yes that would be better indeed
07:02:43 <andythenorth> my for loops are unsophisticated
07:02:49 <andythenorth> I only recently discovered enumerate
07:03:13 <Eddi|zuHause> yup, handy tool :)
07:03:33 <andythenorth> there are n+1 templating modules for python
07:03:46 <Rhamphoryncus> Huh. While passengers are best at 70 days, for mail it's best at 90
07:03:47 <andythenorth> python templating just smells write for nml - to me at least
07:04:00 <andythenorth> I'm going to rewrite BANDIT without any cpp
07:04:14 <andythenorth> I'm also surprised how fast python is locally
07:04:31 <Eddi|zuHause> Rhamphoryncus: 70 days, or 70*2.5 days?
07:04:49 <Rhamphoryncus> 70. What's the 2.5 for?
07:05:11 <Eddi|zuHause> cargo is "aged" every 185 ticks, which equals 2.5 days
07:05:43 <Eddi|zuHause> so when the code says "cargo_age == 70", then it's 70*2.5 days
07:05:44 *** sla_ro|master has joined #openttd
07:06:12 <Rhamphoryncus> The wiki page talks about days. 0.4%/day probably means the code is actually 1%/2.5 days
07:06:52 <Eddi|zuHause> possible, but the wiki may be inconsistent
07:09:26 * Rhamphoryncus switches gnuplot to histogram with a 2.5 step size
07:10:35 <Rhamphoryncus> Hrm. Technically that's still not right. It should be a saw pattern, going up smoothly, then dropping down every 2.5 days
07:18:13 <Rhamphoryncus> Looks to be a variance of about... 0.2% :)
07:20:24 <Rhamphoryncus> And the peak for passengers seems to be 37.7439. Sorry, that's a factor, not a percent
07:22:33 <Eddi|zuHause> note that newgrfs may nowadays adjust the aging speed, so some newgrfs may provide "long distance" coaches, which screw up your careful calculations
07:24:51 <SpComb> andythenorth: or nml could include its own variable expansion, include mechanisms, ... :)
07:26:44 <Eddi|zuHause> the most important thing missing is constants and include
07:27:16 <Eddi|zuHause> the second biggest is local identifier scope
07:27:58 <Eddi|zuHause> i.e. extending "templates" to more than just realsprites
07:28:59 <SpComb> but on the other hand those things are hard to get right
07:29:19 <SpComb> I've been playing with puppet's DSL a lot now, and the language is sometimes fairly silly
07:29:54 <SpComb> in that, you end up doing stupid stuff like inline_template("... <%= .. %> ...") to implement your "business logic" at times
07:30:43 <SpComb> because the DSL itself doesn't have anything like looping constructs for processing arrays, whereas the templates can use arbitrary ruby code
07:30:48 <andythenorth> SpComb: right now, I like that nml doesn't enforce a templating language :)
07:31:08 <andythenorth> it would be useful for new authors if it provided one, but lack of opinion about which suits me right now
07:31:23 <SpComb> andythenorth: so everybody using nml uses a different templating language, making it harder to share code across projects?
07:31:34 <andythenorth> code sharing is over-rated anyway
07:32:09 <andythenorth> if nml does provide templating, we'd have to decide which ;)
07:32:14 <andythenorth> so might as well try some first
07:32:31 <andythenorth> Eddi|zuHause: it was the local scope that started me thinking about $
07:32:46 <SpComb> or stop thinking of it as "templating" and more like "structured language"
07:33:02 <andythenorth> FIRS is full of variadic macros, which if you analyse them are simply defining an identifier which should be local
07:33:10 <andythenorth> THIS_ID(tile) etc
07:33:25 <andythenorth> after you've seen that 50 times in one file, you realise it could just (literally) be written as $
07:33:31 <andythenorth> it contains zero information
07:34:16 <andythenorth> if you were smart enough you could get rid of it entirely in things like industry layouts, and just use implicit assumption
07:34:25 <SpComb> I've never actually seen any NML code, so I'm just talking :)
07:34:35 <andythenorth> tilelayout tilelayout_fertiliser_plant_3 {
07:34:35 <andythenorth> 0, 1: THIS_ID(tile);
07:34:39 <andythenorth> might as well be written
07:34:52 <andythenorth> tilelayout tilelayout_fertiliser_plant_3 {
07:34:57 <SpComb> but writing python code to generate your NML code doesn't sound right either
07:35:08 <andythenorth> I don't like the generator I've written
07:35:12 <andythenorth> generators are explosive
07:35:22 <andythenorth> dumb templating of code is awesome
07:35:44 <SpComb> "dumb" templating involves some "smart" code to control the templating
07:36:06 <SpComb> at which point half of your set is written as python code and you never see any NML
07:36:52 <andythenorth> nml is 'just' output
07:37:37 <SpComb> it's already one level of meta (it generates grfcodec input?) - adding another level of meta on top of the meta.. hmm :)
07:37:54 <andythenorth> SpComb: seems fine to me :)
07:38:22 <andythenorth> database -> controller -> template -> xhtml + css -> rendered by browser -> jquery madness -> final result
07:38:30 <andythenorth> is my daily world
07:38:49 <andythenorth> config data -> build set -> nml -> compile seems ok to me
07:39:05 <SpComb> set *code* + data -> nml
07:39:48 <SpComb> is where it goes wrong, imo
07:40:25 <andythenorth> I agree, but what to do about it?
07:40:34 <SpComb> dunno, but if you wanted everyone to use NML, it'd be nice to see people... writing NML
07:41:48 <SpComb> but maybe this is the point where I should just keep quiet, because I don't know anything about NML
07:42:41 <andythenorth> you'll only end up having to write code if you keep talking :)
07:44:36 <Eddi|zuHause> SpComb: the point is not to force people writing nml. in the end, it's much easier to just put values into a table
07:44:46 *** Progman has joined #openttd
07:45:13 <andythenorth> SpComb: currently my pipeline for BANDIT (it's amusingly convoluted I know) is:
07:46:11 <andythenorth> vehicle data in object database on the web -> python cms -> python page templates rendering html with <pre> tag -> copy + paste to filesystem -> python generator script -> c pre-processor defines, includes and #if -> nml -> result
07:46:41 <andythenorth> Eddi|zuHause has google docs -> tsv -> python generator -> c pre-processor defines -> nml -> result
07:46:51 <andythenorth> Eddi|zuHause: do you curl the data from google, or manually copy and paste?
07:47:20 <Eddi|zuHause> andythenorth: the exporting from google is done manually
07:47:42 <Eddi|zuHause> file -> save as -> text (tsv)
07:47:58 <andythenorth> didn't want to use the API?
07:48:37 <Eddi|zuHause> andythenorth: a) didn't bother, and b) doing that automatically may result in incomplete/temporary stuff bleeding in
07:48:47 <SpComb> why not nml -> tsv -> google docs/whatever? :)
07:49:11 <SpComb> or is this a case of "artists must be able to add new vehicles directly via click click"?
07:49:59 <Eddi|zuHause> SpComb: this is a case of "i want to avoid copy-pasting 600 times
07:50:09 <andythenorth> SpComb: in my case I can't be bothered to learn to parse tsv, I know how to get data out of an object database
07:51:10 <andythenorth> I want to decouple data from how the data is structured
07:51:24 <andythenorth> i.e. not fooling with formatting and crap for data structures
07:51:38 <Eddi|zuHause> SpComb: now imagine i'm adding a new column, like recently the vehicle running cost
07:52:03 <andythenorth> you just add a column and put the figures in
07:52:06 <Eddi|zuHause> SpComb: if it weren't a table, i would have to edit like 600 files
07:53:10 <SpComb> simple solution: add google docs online import capability directly to openttd, you just add a http:// url instead of a .grf file!
07:53:34 <Eddi|zuHause> that's just crazytalk
07:53:41 <andythenorth> I'm halfway there :P
08:02:39 <andythenorth> SpComb: so any ideas how to make nml smarter? :D
08:10:25 *** Asteconn_ has joined #openttd
08:12:23 *** Asteconn_ is now known as Asteconn
08:13:28 <Rhamphoryncus> Ouch. It'd only take me around 3-4 years to recoup the cost of scrapping my short-hop train and replacing it with an industrial tram
08:19:05 <andythenorth> I should just declare a python class truck
08:38:00 *** Twofish has joined #openttd
08:45:05 *** DayDreamer has joined #openttd
08:45:55 *** cmircea has joined #openttd
08:45:55 *** DayDreamer has left #openttd
08:49:34 *** Devroush has joined #openttd
09:05:51 <andythenorth> fifth_wheel_truck inherits truck
09:25:04 *** Asteconn has joined #openttd
09:30:37 *** TheMask96 has joined #openttd
10:23:29 *** Progman has joined #openttd
10:45:51 *** andythenorth has joined #openttd
11:17:12 *** George|2 has joined #openttd
11:17:12 *** George is now known as Guest671
11:17:12 *** George|2 is now known as George
11:44:02 *** Twofish has joined #openttd
12:16:32 *** Brianetta has joined #openttd
12:19:59 *** kkb110_ has joined #openttd
12:29:33 *** kkb110_ has joined #openttd
12:29:33 *** andythenorth has joined #openttd
12:29:33 *** theholyduck has joined #openttd
12:29:33 *** glevans2 has joined #openttd
12:29:33 *** ccfreak2k has joined #openttd
12:43:58 *** tparker has joined #openttd
12:43:58 *** mikegrb has joined #openttd
12:43:58 *** Belugas has joined #openttd
12:43:58 *** weber.oftc.net sets mode: +ovov orudge orudge Belugas Belugas
12:43:58 *** TinoDidriksen has joined #openttd
12:43:58 *** Born_Acorn has joined #openttd
12:59:24 *** kkb110__ has joined #openttd
13:41:27 *** Stimrol has joined #openttd
13:59:56 *** saftusx has joined #openttd
14:00:02 *** saftusx is now known as saftus
14:13:24 * Belugas is going to stay at the door of best Buy tomorrow, waiting of them to open. Transformer Prime should be avaialbe
14:16:27 *** TGYoshi has joined #openttd
14:26:47 <Eddi|zuHause> this is getting creepy
14:30:15 <Belugas> damned you Eddi|zuHause... now, I have to find where is Creeping Death
14:32:11 * andythenorth wants to stop doing current job and start writing python templates for nml
14:32:24 <andythenorth> current job is writing python templates :P
14:34:10 <Belugas> it's the subject that is boring, i'd say :)
14:36:29 <andythenorth> we just sold more stuff
14:36:50 <andythenorth> but nml is bugging me :P
14:39:57 *** roboboy has joined #openttd
14:44:34 <Belugas> yuo mean... tempting you?
14:47:02 *** lollercaust has joined #openttd
14:47:03 * Hirundo ponders inline python in NML
14:49:15 <andythenorth> or - have no opinion on a templating engine
14:49:24 <andythenorth> but don't block any
14:49:38 <andythenorth> FWIW I'm going to try rewriting BANDIT (again) with chameleon templating
14:49:50 <andythenorth> because I know it and it's robust and not too clever
14:50:11 <Eddi|zuHause> hm... weirdness: a gamescript-debug window popped up during game start, and now it's blocked by the main toolbar, i can't close or move it
14:50:30 <andythenorth> Hirundo: I had pretty good results yesterday testing pythons built in Template module $ replacement
14:50:43 * andythenorth has to go brief lawyers
14:50:51 <andythenorth> wonder if I'll get biscuits?
14:51:47 <Hirundo> andythenorth: what about something like .. item template foo(param1, param2) .... item bar extends foo(42, 31)
15:01:41 <planetmaker> Hirundo, templates for items? Sounds.... tasty
15:01:54 <peter1138> ah... the yes album
15:01:56 <planetmaker> and for switches, please ;-)
15:02:35 <planetmaker> probably even more important
15:04:57 <Hirundo> I agree switch blocks blocks can be tiresome to write, but I'm not sure if switch templates are the best solution
15:05:27 <Hirundo> switch in itself is a too nfo-ish construct IMHO
15:05:58 <Hirundo> too limited, requires too many quirks to get things done with
15:08:18 <planetmaker> hm. if's basically a case construct in other languages. Not sure it's nfo-ish
15:09:02 <Hirundo> Imagine writing C++/whatever with nothing but switch-statements to work with
15:09:07 <planetmaker> of course it could be replaced by a real function-like thing ^^
15:09:30 <planetmaker> like then in the callback block: default: graphics_switch
15:10:32 <planetmaker> and there if (year < 1990) return image1; case (position) { 0: return image2; 1: return image3; return image4; }
15:10:42 <planetmaker> feasible. But... much harder to parse, I guess
15:10:55 <planetmaker> or rather harder to validate
15:12:09 <Hirundo> if graphics_switch is a function that accepts parameters, you have a lot of templating power
15:12:22 <planetmaker> but a template function like function yearswitch(img1,img2, img3) { case (year) { 0 .. 1990: return img1; 1991..2000: return img2; return img3; } }
15:22:44 *** Xrufuian has joined #openttd
15:36:26 <Eddi|zuHause> hm.. the ET831/ET87 sparks from the wrong place...
15:40:50 *** TGYoshi_ has joined #openttd
15:54:10 *** Prof_Frink has joined #openttd
16:16:45 *** andythenorth has joined #openttd
16:18:19 *** supermop has joined #openttd
16:38:47 *** |Jeroen| has joined #openttd
16:43:42 *** Brianetta has joined #openttd
16:57:53 *** sla_ro|master has joined #openttd
17:22:32 *** Asteconn has joined #openttd
17:34:06 *** Phoenix_the_II has joined #openttd
17:41:00 *** chester has joined #openttd
17:44:17 *** PhoenixII has joined #openttd
17:45:41 <CIA-1> OpenTTD: rubidium * r23857 /trunk/known-bugs.txt: -Document [FS#4997]: some SDL misbehaviour
17:47:47 *** Phoenix_the_II has quit IRC
17:49:00 <Eddi|zuHause> # backe, backe Kuchen, der Bäcker hat gerufen
17:56:28 *** Stimrol_ has joined #openttd
17:58:39 *** Alberth has joined #openttd
17:58:40 *** ChanServ sets mode: +o Alberth
18:16:10 *** Chris_Booth has joined #openttd
18:29:22 *** sla_ro|master has joined #openttd
18:36:07 *** andythenorth has joined #openttd
19:03:02 <Eddi|zuHause> crazy feature request: allow vehicles to be length 0
19:05:09 <Eddi|zuHause> fewer headaches when refitting for different amount of wagons
19:06:16 <Eddi|zuHause> e.g. the A2/ETA177 has two end parts of length 7, and an optional middle part of length 7
19:06:37 <Eddi|zuHause> length 7 vehicles are currently implemented as 2/3/2
19:07:04 <Eddi|zuHause> so i have an articulated vehicle of 9 parts, 2/3/2+2/3/2+2/3/2
19:08:00 <Eddi|zuHause> it would be way easier to code this as 2/3/2+0/0/0+2/3/2 instead of 2/1/1+1/1/1+2/3/2, and reworking the var 61/62 checks
19:08:03 *** HerzogDeXtEr has joined #openttd
19:10:37 <Eddi|zuHause> there might be guards to disallow <1 length for the front vehicle
19:11:25 <Eddi|zuHause> which might otherwise make the vehicle inaccessible in the depot
19:13:57 <Eddi|zuHause> there can't possibly be non-spam behind a topic "Research - History - Politics - Licenses - AI - Econimics"
19:14:38 *** Hero_of_Time has joined #openttd
19:15:59 <Alberth> request a lock of the topic :)
19:16:22 <Eddi|zuHause> Alberth: but to report it, i first have to open it
19:17:27 <Alberth> actually it is non-spam, just the usual "I want to expand the game to <topic> areas"
19:17:52 <planetmaker> I think I approved the posting ;-)
19:18:24 <planetmaker> approved as in "allowed it to show in the forums"
19:18:35 <planetmaker> not the contents ;-)
19:18:43 <Alberth> yeah, but it won't go anywhere :)
19:19:04 <planetmaker> it's the suggestions forum after all
19:19:27 *** frosch123 has joined #openttd
19:19:34 <planetmaker> and obviously he never tried ECS or FIRS. Hasn't heart about bribing. Never heart of goal scripts ;-)
19:19:39 <Alberth> sure, not a problem to post suggestions there, even if they are completely useless :)
19:20:05 <Alberth> or infra-structure costs :p
19:20:35 <Alberth> Although I must admit I haven't used them either ;)
19:21:53 *** Asteconn has joined #openttd
19:22:02 <planetmaker> oh, and he isn't aware that it's also up to him to make a better train AI ;-)
19:22:33 <planetmaker> Though IMHO the most daring statement is "More micromanaging will make players cherish (...)"
19:22:41 <planetmaker> I'll not subscribe to that ;-)
19:22:42 <Alberth> I pondered to post about that, but in the end could not be bothered to hit 'reply'
19:22:51 <planetmaker> hehe. You neither?
19:23:22 <Alberth> I am sure someone will tell him about AI's :)
19:24:12 <Alberth> maybe tomorrow or Sunday, if nobody did, and I am terribly bored :)
19:29:19 <Alberth> I am somewhat pondering about adding warnings for the user about things he might miss
19:29:33 <Alberth> For example, not being able to build anything for X months
19:29:47 <Alberth> Not being able to transport some type of cargo
19:30:14 <Alberth> Forgetting to refit, so he attempts to load non-existing cargo at some station
19:30:32 <planetmaker> you mean a kind of mentor which checks the sanitiy of actions?
19:30:46 <Alberth> well, I load FIRS but default vehicles -> I cannot transport EMNS
19:30:56 <planetmaker> ah, newgrf sanity
19:31:11 <planetmaker> That will need sandbox testing in the newgrf dialogue
19:31:17 <planetmaker> and that makes much sense
19:31:43 <planetmaker> I think frosch123 has some probably more details ideas there...
19:32:18 <andythenorth> Alberth: sounds like grf-topia?
19:32:27 <Alberth> at the moment, I am doing these things in-game
19:32:31 <Alberth> andythenorth: it does?
19:32:45 <planetmaker> what it needs is: testing each vehicle type (train, road, water, air) whether it can transport all cargos
19:32:50 <planetmaker> maybe air can be skipped
19:33:05 <frosch123> we need to decouple newgrf.cpp from the rest of ottd
19:33:14 <Eddi|zuHause> "lower-saxony: state attorney says: you may call the president a liar"
19:33:19 <Alberth> I was pondering to do it as part of world generation
19:33:27 <frosch123> so you can process a newgrf configuration without affecting pools or other gamestates :)
19:33:53 <planetmaker> Alberth, yes... though then it's too late
19:33:54 <Alberth> ah, that's how it is grf-utopia :p
19:34:12 <planetmaker> imho the grf config dialogue needs a box like "sanity check"
19:34:51 <planetmaker> but where exactly that check is called is very minor to making it feasible in the first place ;-)
19:34:52 <Alberth> it reports "you are insane" <OK> with me :p
19:35:13 <frosch123> the only step i did for now was restructuring the temporary variables during newgrf load
19:35:18 <frosch123> i.e. the stuff which is now in _cur
19:36:00 <frosch123> another first step would be to display "your newgrf configuration has conflicts" if the number of newgrfs is > 20
19:36:02 <Alberth> I am blissfully unaware of what _cur contains :p
19:36:14 <frosch123> Alberth: it's even commented :p
19:36:54 <Alberth> ok, so it may be penetrable for me too :)
19:37:17 <Eddi|zuHause> possibly it should be completely decoupled from global variables
19:37:20 <Alberth> (20:40:48) frosch123: another first step would be to display "your newgrf configuration has conflicts" if the number of newgrfs is > 20 <-- much easier solution, just reduce the limit to 19 :D
19:37:28 <frosch123> basically i want an interface between newgrf.cpp and the outside world :p
19:37:36 <Eddi|zuHause> so you create like an object NewGRFContext()
19:37:39 <Alberth> sounds like a good step indeed
19:40:51 <Alberth> and a whole lot more complexity :p
19:42:30 <frosch123> alternatively we can execute openttd again as childprocess and pass the newgrf config via commandline :p
19:44:51 <Alberth> good, cmdline at windows still limited at 128 characters? :)
19:45:29 <frosch123> well, we can also use stdin and stdout
19:45:37 <Alberth> could be, it has been > 10 years I last used Win*
19:45:40 <frosch123> if we compile win binaries as console applications
19:45:54 <frosch123> 255 would sound like a nice dos number
19:46:21 <Eddi|zuHause> it's probably >15 years since i tried long command lines the last time
19:46:51 <Eddi|zuHause> i DID hit the limit in DOS back then
19:48:38 <frosch123> i also hit in on linux
19:49:08 <frosch123> sometimes rm * does not work, if there are too many files in the directiory
19:49:45 <frosch123> bash has a limit of 32k arguments or so
19:51:46 <Eddi|zuHause> looks like the DOS limit is 127
19:58:39 <Eddi|zuHause> my DOS 5.0 manual says: "Sie können beliebig viele Befehle in eine einzige Zeile eingeben, vorausgesetzt, daß die Gesamtzeilenlänge 128 nicht überschreitet."
19:59:34 <Eddi|zuHause> apparently, you could separate several commands by pressing ctrl+t
20:01:17 <frosch123> 1) implement a shell that can execute binaries
20:01:58 <frosch123> 2) allow adding arguments to commands using / to separate the command and the arguments (no space)
20:02:18 <frosch123> 3) add directories to your filesystem and use \ because / is already used for command/argument separation
20:02:41 <Alberth> in particular since /was already used for unix :)
20:03:33 <frosch123> it's all the fault of us date format
20:03:52 <Belugas> good for him :) One of the devs in here just gave resignation. He's going for a better job. Happy for him, sad for us. he's been here for 10 years. I've trained him, worked with him and had great time doing so
20:04:08 <frosch123> if they would not use such a silly character like / for date separation, it would be not such a popular key on the us keyboard
20:04:35 <Eddi|zuHause> how do you explain the usage of ~ in win95 then? :)
20:04:40 <frosch123> and as such it would not have been used for arguments, or searching in various viewers or crappy - yet popular - editors
20:05:38 <Alberth> Belugas: openttd dev resigning? :o
20:05:39 <frosch123> Eddi|zuHause: oh, i forgot to mention
20:06:00 <Alberth> usually they just leave :)
20:06:15 <frosch123> dos used a normal character to prefix filenames in the filesystem if the files were deleted
20:06:29 <frosch123> i think it was a "." (cannot remember)
20:06:54 <frosch123> later they wanted to use the character in filenames, so they added another speical symbol to represent the "." (or whatever it was)
20:07:10 <Eddi|zuHause> hm, don't remember that
20:07:37 <Eddi|zuHause> i always believed it was a <0x20 character
20:07:44 <frosch123> it's really impressive how dos is screwed up, though they only copied features from unix and wang, while trying to make it simpler; but still adding the stuff later
20:08:21 <frosch123> dpmi is another great story :p
20:08:54 <Eddi|zuHause> oh dpmi caused great fun
20:08:55 <frosch123> dpmi 0.9 was a pre-release which needed shipping early. it contained several low level functions which should not be accessible for normal applications
20:09:00 <Belugas> naaa work@work resigning, Alberth
20:09:01 <Eddi|zuHause> with buggy loaders in commercial games
20:09:22 <frosch123> 1.0 then removed those functions. but since applications were already using them, it was dropped and 0.9 remained active
20:17:02 <frosch123> ok, wiki prooves me wrong. it was not the "." character which marked deleted files
20:17:23 <frosch123> but some >80h character; so it also became important when extending filenames beyond ascii
20:17:50 <frosch123> "0x05 means, the character is actually 0xE5" :p
20:18:00 *** Devroush has joined #openttd
20:19:19 <Eddi|zuHause> it was somewhat different in Novell DOS, there you could restore files with the complete name, instead of guessing what the first character was
20:19:45 <frosch123> yeah, novell dos reps. drdos was a lot better
20:19:51 <SmatZ> [21:09:19] <frosch123> 1.0 then removed those functions. but since applications were already using them, it was dropped and 0.9 remained active
20:20:01 <SmatZ> ^^^ interesting knowledge, I was never aware of that
20:20:06 <frosch123> all drivers and such had a lot smaller memory footprint
20:20:13 <SmatZ> I thought it just never reachedthe "stable" state
20:20:34 <frosch123> sad was though that lots of games screwed up when they could not find mscdex
20:20:45 <frosch123> so somewhen i copied nwcdex to mscdex :p
20:21:38 <Eddi|zuHause> i never properly got my CD drive to work...
20:21:48 <frosch123> SmatZ: the dos release cycles were so long, that is was common that applications worked around bugs, and then relied on the buggy behaviour; so it was not possible to fix the original bug without breaking stuff... which would then be considered a bug :p
20:22:13 <Eddi|zuHause> my tower was so screwed up, you had to lay it down because the 5 1/4" slots were vertical
20:22:26 <Eddi|zuHause> and i already had 2 HDDs, so i would have to disconnect one
20:22:40 <frosch123> well, "release cycle" is maybe the wrong word. i mean you did not update your os regulary, but kept it on the machine when bought
20:22:47 <Eddi|zuHause> and i never figured out how to use the CD port on the sound card
20:23:32 <frosch123> oh yeah, i think my first sound card had like 3 different connectors for various non-atapi cdroms :p
20:24:48 <Eddi|zuHause> but none of them worked with the drive
20:24:59 <Eddi|zuHause> or it needed special drivers which i didn't have
20:24:59 <frosch123> hmm, actually let's see whether the ralph brown intlist is still maintained :)
20:25:10 <Eddi|zuHause> and there was no internet to get drivers
20:25:32 <michi_cc> I wonder just who came up with the idea to add cdroms to sound cards and not proper storage adapters.
20:25:59 <frosch123> no idea either, i would assume they were only used for audio
20:26:19 <frosch123> in the early days there were such weird things like video or image cds
20:26:34 <frosch123> everyone invented his own format to burn on cds :p
20:26:35 <michi_cc> It's not as if PCs didn't already had one or more of the different floppy and hdd connectors.
20:26:39 <Eddi|zuHause> there's a separate audio connector
20:27:28 <michi_cc> Yeah, my first cd drive still had play/stop and skip buttons on the front.
20:28:13 *** Rhamphoryncus has joined #openttd
20:28:53 <Eddi|zuHause> it had even TRIPLE SPEED!
20:30:15 <frosch123> then you had one before me
20:30:29 <frosch123> i already had quad speed :p
20:37:34 *** Progman has joined #openttd
20:40:53 <Alberth> michi_cc: but thsoe floppy/hdd things were parallel, not serial
20:51:52 <Eddi|zuHause> MAN ... the edit history of google docs is FUCKING USELESS!
20:52:47 *** supermop has joined #openttd
20:55:25 <andythenorth> Eddi|zuHause: google docs is a half-done solution
20:55:33 <andythenorth> I can't decide yet which
20:55:43 <andythenorth> but excel kicks it's ass in so many respects
20:57:55 <planetmaker> on the coast of being bloated ;-)
21:01:58 <Rhamphoryncus> Google docs has the advantage of being online. Mostly easy to share with people. That's all it has though.
21:02:30 <planetmaker> yes. It also has IMHO the better interface than the online version of excel
21:02:56 <Rhamphoryncus> there's.. an online excel?
21:03:39 <planetmaker> IIRC I once saw one... I don't recall the URL
21:06:13 <planetmaker> it needs login and stuff, though
21:06:22 <planetmaker> though one can share links similar like google docs
21:14:49 <CIA-1> OpenTTD: rubidium * r23858 /trunk/ (7 files in 2 dirs): -Change [FS#5003]: some updates and improvements of the base font (PaulC)
21:17:13 <Eddi|zuHause> ah, so "skydrive" is a microsoft product...
21:17:22 <Eddi|zuHause> that explains the extensive "product placement"
21:20:08 <Rhamphoryncus> "it sounds cool and it's kinda.. cloud-ish.." -- Microsoft marketer picking the name
21:21:25 <Eddi|zuHause> do we have roman numerals in the ¹²³ fashion in openttd?
21:22:05 <Eddi|zuHause> particularly an I
21:22:59 <Rubidium> is there an unicode codepoint for it?
21:23:46 <Rubidium> just not in superscript
21:24:27 <Rubidium> you could (simply) add the glyphs you need in your NewGRF
21:24:55 <Eddi|zuHause> we could offer a {SUPERSCRIPT} tag and write it in small font
21:25:16 <Rhamphoryncus> ahaha, Qalculate has a RNG in it.. and since the calculator is symbol the random number is irrational :D
21:25:18 <Rubidium> or just make the glyphs you provide superscripty ;)
21:26:55 <Eddi|zuHause> there's only a superscript i
21:26:59 <Eddi|zuHause> not a superscript I
21:27:58 <Eddi|zuHause> have an example that adds a glyph in a newgrf?
21:28:23 <Rubidium> /trunk/media/extra_grf/chars.nfo
21:29:54 <Rubidium> Eddi|zuHause: got three options already here: ᴵ ᴵ ᶦ
21:34:16 <Rubidium> don't forget to scroll down to the "other superscript and subscript characters" section
21:34:34 <Eddi|zuHause> i don't find these in kcharselect
21:34:47 <Eddi|zuHause> but honestly, their search kinda sucks
21:37:53 <__ln__> merry christmas, andythenorth
21:41:24 <frosch123> 1D35 seems to be one of those
21:41:30 <Eddi|zuHause> hm, so U+1D35 could be the "right one"
21:43:21 <Eddi|zuHause> and U+2C7D for V
22:01:03 <andythenorth> anyone got a favourite python templating module that can handle repeats
22:03:54 * andythenorth suspects mako is the correct solution for nml templating
22:03:57 <andythenorth> but I don't know it
22:04:04 <andythenorth> and it looks dangerously powerful
22:09:02 <CIA-1> OpenTTD: frosch * r23859 /trunk/src/order_cmd.cpp: -Fix: Inserting conditional orders for ships checked the wrong orders wrt. maximum distance.
22:17:08 <andythenorth> nml doesn't care about extra newlines in switch blocks?
22:17:53 <Alberth> it's a context-free parser, so it shouldn't
22:17:56 <andythenorth> can I rely on that persisting?
22:18:09 <andythenorth> it will make butt-ugly code, but it's a limitation of the templater I'm using
22:18:20 <Alberth> ply is not likely to go away anytime soon :)
22:18:28 <andythenorth> I can have extra newlines
22:18:36 <andythenorth> or loc > 80 char long in some cases
22:19:03 <Alberth> ply doesn't mind either/both
22:19:16 <andythenorth> smells like the wrong templater to me though :P
22:19:28 <andythenorth> although I can't see how others would handle it differently
22:19:39 <Eddi|zuHause> CPP inserts plenty of newlines
22:19:50 <andythenorth> if I write newlines in a template, I'd usually need them respected
22:19:58 <andythenorth> just occasionally it would be useful to not
22:20:36 <Eddi|zuHause> any #if/#ifdef will result in lots of empty lines if the condition is not met
22:20:37 <Alberth> that's the mess with templates; they tend to use the same newline character both for output and for template formatting :p
22:21:02 <andythenorth> and not doing that just litters template with magic characters
22:21:18 <Eddi|zuHause> just escape the newlines :)
22:21:21 <Alberth> which lead to ditching all those template stuff at work, for me
22:21:56 <andythenorth> Alberth: yesterday I got some nice results with python's default Template class and $ ${} substiution
22:21:58 <andythenorth> which means no deps
22:22:11 <andythenorth> but today I have been trying chameleon because I know it
22:22:20 <Alberth> unfortunately, I use Java at work :)
22:22:45 <Eddi|zuHause> Java, the master of all bloatware, doesn't have a useful template library?
22:22:52 * Rhamphoryncus doesn't see how a trainer mech would help with templating ;)
22:23:34 <frosch123> aren't templates part of jave 6?
22:24:00 <frosch123> (resp. 2.6 if you use the traditional numbering)
22:24:24 <Wolf01> "I used to be a captain like you, but then I took a lifeboat in the knee."
22:24:48 <andythenorth> chameleon might be overkill
22:24:56 * andythenorth shows restraint, unusually
22:25:16 <andythenorth> but it means can do stuff like ${foo_string.upper()} or whatever
22:25:45 <andythenorth> and also repeat lines trivially
22:26:31 <andythenorth> cookie for first person to tell me result of this: <tal:test repeat="test_repeat (1,2,3)"> ${repeat.test_repeat.number}</tal:test>
22:27:31 <andythenorth> it's a slightly dumb example
22:27:39 <Alberth> frosch123: I don't know, the stuff I tried was supposed to be good for converting models (a collection of objects) to text. However, the manual stated that if you used it for eg Java, you better used the Java formatter afterwards.... which makes you wonder how you'd make such a formatter in Java :p
22:28:13 <andythenorth> less dumb: <tal:test repeat="test_repeat ('cat', 'mouse', 'dog')"> ${test_repeat}</tal:test>
22:36:08 *** supermop has joined #openttd
23:05:53 * andythenorth adventures in templates
23:06:48 *** Chris_Booth has joined #openttd
23:07:14 <Eddi|zuHause> anyone ever wrote an adventure in sed?
23:07:24 <Eddi|zuHause> i've seen sokoban in sed
23:08:18 <andythenorth> Template("${go}", "north")
23:08:52 * andythenorth wonders if it's total overkill to create each vehicle as a python object
23:08:57 <andythenorth> but it's kind of neat
23:11:32 <andythenorth> I could almost as well stick dicts in a list though
23:14:45 <Mazur> Making this chart does special things to one's mind. I'm seeing Main Grills, where I make hamburgers, Diary Farms where teenage girls secretively write, and Adorable Farms.
23:14:59 <andythenorth> you need a better font ;D
23:15:38 <Mazur> There are more of those. :-)
23:15:43 <Eddi|zuHause> andythenorth: i'm having dicts in a dict
23:16:54 * andythenorth will now rewrite BANDIT for the n(?)th time
23:17:24 <andythenorth> nml saves so much time it can be wasted on ever-better templating
23:27:24 <Mazur> Btw, did you guys choose these particular colours for the chain widgets?
23:27:30 <Mazur> And the small map, of course.
23:28:53 <planetmaker> which particular colour?
23:29:06 <planetmaker> you mean for the individual cargos and industries?
23:29:26 <planetmaker> they were iirc chosen such that it remains somewhat distinguishable on the minimap
23:29:51 <andythenorth> minimap comes in multiple colours also
23:30:39 <Mazur> Because some colours get used a lot, offwhite 5 times, I think, and yellow or near yellow, I forget which, 4 times.
23:30:46 *** Cybertinus has joined #openttd
23:31:07 <andythenorth> colours that work in town are very limited
23:31:26 <Mazur> It would be preferable, I think, to have distinct individual colours.
23:31:55 <andythenorth> there are limits
23:33:01 <Mazur> Having just made boxes for all of them.
23:33:12 <Mazur> And shoving them every which way.
23:33:31 <Mazur> I can ount the diverse cargoes as well.
23:33:43 * andythenorth wonders if python args for a call can be done with list comprehension on a dict
23:34:41 * Mazur nose no OO programming.
23:34:50 <andythenorth> i = dict[i] for i in dict?
23:35:26 <Mazur> Someone will have made a dict_lookup()
23:35:59 <andythenorth> python encourages writing out explicit args for calls
23:36:00 <Yexo> andythenorth: dict.items() or something like that
23:37:46 <Yexo> dict.keys() or dict.values() if you only want the keys or the values
23:37:55 <Yexo> not sure what you want to achieve though
23:38:10 <Mazur> How many sugar beats to the minute are there in a sugar rush?
23:38:18 <andythenorth> got a template that gets vars as args
23:38:43 <andythenorth> template(foo='bar', ham='eggs') etc
23:38:57 <andythenorth> I could get all the args from a dict
23:39:01 <andythenorth> or I could not bother
23:39:11 <andythenorth> it's not important, just fooling around
23:39:23 <Yexo> def myfunc(*args, **kwargs):
23:39:30 <Yexo> kwargs is a dict with all keywords arguments
23:40:07 <andythenorth> the interesting thing is that I think there are >1 viable routes for python templating nml
23:40:09 <Yexo> if you want to pass those to another function you could do it manually or use: otherfunc(**dict)
23:40:16 <andythenorth> I'm going to eliminate cpp from BANDIT
23:41:00 <andythenorth> the only difficult thing so far is deciding how much code generation is acceptable
23:41:20 <andythenorth> and how much to just have *lots* of nml switches written out longhand
23:42:00 <andythenorth> I should have results by Monday maybe
23:43:24 <andythenorth> delay is due to baby duties, not difficulty of problem
23:44:57 * Mazur did not know babies had duties?
23:45:11 <Mazur> How muhc is the import tax?
23:45:27 <andythenorth> your entire life's worth
23:46:07 * Mazur is funnier than a turd on a roller-coaster.
23:46:17 <planetmaker> Mazur: why are bauxite and iron ore mine switched?
23:46:59 <planetmaker> and forest and sawmill should be left of biorefinery
23:47:08 <planetmaker> Mazur: not in colour or so. But placement
23:47:19 <planetmaker> they should be switched for dis-entanglement
23:50:43 <Mazur> ... which is why I chose this toll to make the initial chart.,
23:51:11 <andythenorth> is there a python equivalent of #include?
23:51:13 <Mazur> 4 seconds and it's fixed.
23:51:15 <andythenorth> import seems overkill
23:51:31 <andythenorth> eval(file.read())
23:55:07 <Rhamphoryncus> from foo import *
23:55:33 <Rhamphoryncus> Generally discouraged, yet quietly done anyway in various circumstances
23:56:22 <andythenorth> I want to use dicts that are in external files
23:56:41 <andythenorth> e.g. they're data structures that should be separate from code
23:57:28 <Rhamphoryncus> There are other methods, depending on what exactly you're doing. setattr() loop on the module object, direct insertion into the globals dict (if you were using exec())
23:59:38 * Mazur worlks, Glass Works, Iron Works, Brick Works.
23:59:40 <andythenorth> looks like importing a dict just works
continue to next day ⏵
