IRC logs for #openttd on OFTC at 2014-05-23
⏴ go to previous day
00:27:40 *** Prof_Frink has joined #openttd
02:22:11 *** Pereba_ has joined #openttd
02:27:01 *** Pereba_ is now known as Pereba
02:36:57 *** fkinglag has joined #openttd
03:51:11 *** LSky_net has joined #openttd
03:52:19 <LSky_net> I woke up to this, anyone have an idea what that relates to?
03:52:21 <LSky_net> Message: Assertion failed at line 127 of .../src/core/smallstack_type.hpp: PooledSmallStack::CanAllocateItem()
04:18:41 *** tycoondemon2 has joined #openttd
05:01:30 *** Flygon_ has joined #openttd
05:36:55 *** Hazzard has joined #openttd
05:51:04 *** tycoondemon has joined #openttd
05:54:04 *** tokai|noir has joined #openttd
05:54:04 *** ChanServ sets mode: +v tokai|noir
05:57:24 *** pthagnar has joined #openttd
06:23:57 *** Flygon_ is now known as Flygon
06:24:04 <Flygon> But what do planets do?
06:25:12 <planetmaker> they make places Yeti like to poo on ;)
06:25:28 <planetmaker> while they moo, of course
06:38:25 *** Aristide has joined #openttd
06:44:01 <V453000> serious yeti business
06:44:01 <V453000> and final details could be later :)
06:44:01 <V453000> some graphical bugs but that only takes re-rendering now :)
06:49:37 *** Aristide has joined #openttd
06:59:24 *** Flygon_ has joined #openttd
07:02:15 <V453000> nice, it is noticeable :D i wasnt sure
07:10:43 <Aristide> TCL install many Parkings xD
07:12:55 *** Pensacola has joined #openttd
07:18:45 *** Flygon_ is now known as Flygon
07:33:05 <LSky`> im guesing thats the vehicle factory?
07:33:26 <Xaroth|Work> I sure hope it's not the food processing plant
07:53:49 <V453000> it is the machinery factory yes :)
07:54:10 <V453000> machinery increases effectiveness of workers for primary industries, something like supplies
07:59:32 *** Myhorta has joined #openttd
08:18:34 *** killertester has joined #openttd
08:29:14 <Aristide> V453000: I like Hybrid bus :D
08:29:20 <Aristide> But cost 380 000 Euros :x
08:36:04 <peter1139> What a horrible noise.
08:38:59 *** montalvo has joined #openttd
08:40:55 <V453000> yeti busses aint making no noise (:
08:41:39 <peter1139> Road vehicle running sounds!
08:51:51 *** montalvo has joined #openttd
09:53:02 *** jjavaholic has joined #openttd
09:53:12 <V453000> you are seriously ill
09:53:17 <V453000> there isnt even anything to be seen
10:35:45 *** tycoondemon has joined #openttd
11:32:36 *** Supercheese has joined #openttd
11:39:12 *** Klanticus has joined #openttd
11:41:03 <Rubidium> LSky: I hope you're using a patched version that throws that assertion. It means that something tries to allocate a pool item without checking whether there is space for a new pool item
11:42:05 <V453000> you cant build swimming pools in openttd.
11:50:11 <planetmaker> so not true, V453000 !
11:51:05 <planetmaker> heck, even the default 'houses' have a fountain in which you surely can swim :)
12:01:27 <planetmaker> pew. Lucky me that you agree so easily. Or I would have had to make a swimming pool object newgrf ;)
12:03:22 <planetmaker> but would be a nice easter-egg thing for one or another grf :)
12:11:06 *** sla_ro|master has joined #openttd
12:21:45 *** Eddi|zuHause has joined #openttd
12:34:33 *** Klanticus_ has joined #openttd
12:38:32 <Flygon> Is a method of having day and night
12:38:45 <Flygon> Shows couples swimming in the pool at night
12:39:59 *** kais58__ has joined #openttd
12:50:39 *** kais58__ is now known as kais58|AFK
12:53:46 *** Aristide has joined #openttd
12:54:39 <Aristide> V453000: I like inside of trolleybus :x
12:57:11 <V453000> Flygon: just make a newgrf which changes everything and all sprites turn to dark every odd year
12:57:37 <V453000> idk if possible for landscape
12:58:08 <Flygon> V453000: Don't worry, I wasn't completely seriou
12:58:47 <Flygon> I do think a day-night system with those day patches that actually give a 24 hour clock into the game would be awesometastic though
12:59:52 <V453000> no need to interconnect that
13:00:02 <V453000> having day for 1 year and night for another year would be nice enough
13:01:11 <Flygon> A standard gauge HST line from Melbourne to Broken Hill, assuming HST = 250km/h max
13:01:19 <Flygon> Via Bendigo and Swan Hill
13:01:22 <Flygon> Others would call me mad
13:01:25 <planetmaker> V453000, that actually is not entirely possible. You cannot change infrastructure and landscape sprites. They don't have the required facilities to check for time
13:01:30 <Flygon> I call it a method of printing money in OpenTTD
13:01:48 <V453000> that is what I was saying pm :) "idk if possible for landscape"
13:01:57 <planetmaker> sorry, slow reading :P
13:02:02 <V453000> but having those things able to check for that would be nice too
13:02:08 <V453000> especially for e.g. randomizing landscape
13:02:11 <Eddi|zuHause> planetmaker: which is a problem since TTRS switched road sets... why has that not been adressed yet?
13:02:32 <planetmaker> Eddi|zuHause, it has. There's a fix for that in the openttd repository
13:02:39 <planetmaker> for desyncing multiplayer by that means ;)
13:03:07 <planetmaker> but the fix is differently... disabling the time dependence for MP entirely :P
13:03:59 <planetmaker> V453000, yes, it would. I totally agree. Such feature has nasty thorns, though
13:04:18 <planetmaker> Especially thorns from the cpu usage twitch ;)
13:05:26 <planetmaker> but then... nearly every feature takes cpu :)
13:06:38 <planetmaker> indeed. solves many problems :P
13:06:59 <planetmaker> your sprites can have any colour you like. Provided it's black :P
13:08:27 <planetmaker> also they can choose that any time at at any frequency you like ;)
13:19:25 *** HerzogDeXtEr has joined #openttd
14:17:28 *** Myhorta has joined #openttd
14:24:13 *** Klanticus has joined #openttd
14:36:17 *** tycoondemon has joined #openttd
15:04:39 *** Hazzard has joined #openttd
15:22:47 *** Alberth has joined #openttd
15:22:47 *** ChanServ sets mode: +o Alberth
15:26:05 *** kais58|AFK is now known as kais58__
15:34:34 *** TheMask96 has joined #openttd
15:50:46 *** Progman has joined #openttd
15:57:36 *** Phreeze has joined #openttd
16:00:20 <Phreeze> just noticed that theres a ottd or ttd screenshot :)
17:06:39 *** gelignite has joined #openttd
17:07:40 *** fkinglag has joined #openttd
17:17:02 <Rubidium> planetmaker: transparent is a more useful color for speeding up drawing ;)
17:21:24 *** Klanticus_ has joined #openttd
17:29:34 *** andythenorth has joined #openttd
17:46:10 <DorpsGek> Commit by translators :: r26607 /trunk/src/lang (6 files) (2014-05-23 17:45:58 UTC)
17:46:11 <DorpsGek> -Update from WebTranslator v3.0:
17:46:12 <DorpsGek> catalan - 25 changes by juanjo
17:46:13 <DorpsGek> simplified_chinese - 3 changes by xiangyigao
17:46:14 <DorpsGek> croatian - 17 changes by VoyagerOne
17:46:15 <DorpsGek> czech - 7 changes by Eskymak
17:46:16 <DorpsGek> lithuanian - 7 changes by Stabilitronas
17:46:17 <DorpsGek> russian - 8 changes by Lone_Wolf
17:46:38 *** frosch123 has joined #openttd
17:52:51 <DorpsGek> Supercheese: I have not seen cat.
17:52:58 <Supercheese> Here has never been cat
17:55:41 *** Phoenix_the_II has joined #openttd
18:02:26 <frosch123> no, but i believe we have a cow
18:06:59 <Alberth> V is clearly the expert :)
18:08:43 <planetmaker> nah, I'd not claim that. He just took foreign language courses
18:10:29 <andythenorth> are we making anything?
18:10:57 * Alberth just made a new patch
18:11:06 * planetmaker works on building nml-py3.
18:11:18 <Alberth> and /me is wondering about making some tea
18:12:20 <andythenorth> maybe I’ll play the game
18:12:37 <andythenorth> or maybe I’ll fix the UI elements that are failing colour contrast
18:14:38 <Wolf01> paint it like one of your french girls
18:15:28 <Xaroth|Work> planetmaker: for unix or win?
18:17:32 <planetmaker> Xaroth, unix was done weeks ago
18:17:50 <Xaroth|Work> then what's wrong with win?
18:17:54 <planetmaker> and actually I 'just' need to get it done for the right wine version as we have on the server
18:18:04 <planetmaker> Xaroth, cx_freeze working under wine...
18:18:21 <planetmaker> and configuring the CF to like that
18:19:02 <Xaroth|Work> why cx_freeze? does it require some strange lib?
18:19:40 <planetmaker> if you give me another means to create a simple-to-use exe file for windows users... feel free
18:19:58 <Alberth> for some reason, MS doesn't install a useful program like Python on their machine :(
18:20:09 <planetmaker> yeah. Nor provides a package manager to do so
18:20:10 <Xaroth|Work> hrnf, if they have python3 installed, iirc, you don't -have- to make it an exe
18:20:34 <planetmaker> Xaroth, then not. But then it still requires setting up all paths etc. pp
18:20:41 <Alberth> Xaroth|Work: obviously, most don't have that :)
18:20:58 <planetmaker> I can meanwhile also provide an installer which works with a py3 install. But as alberth says.
18:21:18 <planetmaker> And those require possibly installers tailored for the specific python version. So many more packages which need building
18:21:53 <planetmaker> yeah. The world is lame in some parts
18:22:25 <Xaroth|Work> yeh, I'm used to being able to tell people to install libraries mine needs :P
18:25:43 * Alberth is pretty sure, my Linux DVD has what Xaroth|Work needs :p
18:27:01 *** tycoondemon has joined #openttd
18:34:35 <andythenorth> people are going to love these
18:35:06 <Eddi|zuHause> 2010-12-21 01:30:50 <Eddi|zuHause> and a black cow says "yo man, moo man!" <-- how did that ever come out of my brain?
18:38:43 <Alberth> this will be the first newgrf I want to play at 4x zoom :)
19:15:29 <V453000> so how do I code an industry newgrf? :D
19:16:05 <andythenorth> you get planetmaker to do it
19:17:07 <V453000> that sounds ... reasonable but hm :)
19:17:30 <V453000> I am not very sure they want to do the minionizing :P
19:18:59 <andythenorth> V453000: do you want to do anything clever / fancy?
19:19:15 * Alberth considers moving pictures fance
19:20:07 <V453000> primaries have base production based on game date (is possible?), primaries do not produce anything without workers, delivered workers increase production, delivered machinery increases effectiveness of workers
19:21:14 <V453000> is that a reasonable mechanism?
19:21:25 <andythenorth> you’ll probably want templating
19:21:28 <planetmaker> yeah, try and see
19:21:33 <V453000> Alberth: thanks, that looks helpful
19:21:41 <V453000> I meant reasonably possible to code
19:21:52 <andythenorth> yes it’s reasonable, but not short
19:22:11 <andythenorth> you could look at FIRS, to see what not to do :)
19:22:18 <andythenorth> FIRS has a lot of history
19:22:27 <V453000> dont think I can wrap my head around FIRS code
19:22:34 <Alberth> nml needs a macro expander :)
19:22:46 <V453000> templating sounds good since I want to make all industries work super similarly, at the same time I think I would kill more time than save
19:23:17 <Alberth> V453000: that code is probably as basic as it gets
19:23:20 <andythenorth> Alberth: yexo was mostly -1 to that I think
19:23:52 <planetmaker> Alberth, "re-inventing the wheel"
19:23:59 <planetmaker> but not sure I agree
19:24:22 <Alberth> yeah, making people install cpp is friendlier :)
19:24:37 <andythenorth> also ‘nml has other more important features to add'
19:25:04 <Alberth> I agree, if only I could understand it all :)
19:26:12 <Alberth> for me, it would be better to start from scratch, but that would take a few years probably
19:27:52 <andythenorth> Alberth: start nml from scratch?
19:28:34 <Alberth> do you know a better way to understand nfo & nml?
19:31:04 *** bdavenport has joined #openttd
19:45:47 <Eddi|zuHause> Alberth: reimplement it in C++ while you're at it :)
19:51:03 <Alberth> actually, I had plans of making a linkable newgrf format :)
19:51:10 <LordAro> Alberth: also, you're busy doing more important (tm) things ;)
19:53:48 *** talebowl has joined #openttd
20:06:00 <andythenorth> linkable newgrf format \o/ :)
20:06:47 <andythenorth> I wondered if a pure python format could be managed :P
20:07:00 *** Supercheese has left #openttd
20:07:02 *** Supercheese has joined #openttd
20:07:25 <andythenorth> just define classes for all entities
20:07:45 <andythenorth> then write out nfo or newgrf format directly
20:08:32 <andythenorth> probably not general-purpose enough :P
20:09:01 <Eddi|zuHause> # Und Gott sprach zu der Menge:
20:09:05 <Alberth> sounds like a bit too general purpose to me :)
20:09:05 <Eddi|zuHause> # IHR SEID SO LEISE
20:09:09 <planetmaker> andythenorth, linkable does not necessarily mean to define everything as classes or so
20:09:27 <planetmaker> just that you can define and build each vehicle / industry / ... separately and link it later
20:10:20 *** Phoenix_the_II has quit IRC
20:12:50 <andythenorth> Alberth: you’d change the newgrf format, or the meta language, or both?
20:13:47 <Alberth> you get an intermediate format just above newgrf
20:14:46 <Alberth> although if you use a textual format, you'll probably end up with something nfo-ish
20:16:09 <andythenorth> would this be transparent to current nml authors?
20:16:15 <andythenorth> or would it require a change in authoring?
20:17:23 <planetmaker> ideally it should be.
20:19:44 <andythenorth> what (how?) analysis would be done to identify what needs re-compiled and linked?
20:19:53 <Alberth> you should be able to do partial compilation
20:20:38 <planetmaker> andythenorth, *that*, I guess, is the interesting question
20:20:54 <Alberth> that's what make has been doing for a long time (not sure if it still does that), but currently it's useless
20:21:13 <Alberth> and linking, you need to link everything, don't you?
20:22:24 <Alberth> ie you can run a compile for just one industry
20:22:33 <Alberth> (theoretically at least)
20:23:04 <Alberth> do that for every industry, and then link all pieces into firs.grf
20:23:11 <planetmaker> I wonder how to deal with the ID issue, though. Whether to replace them on the run. Or whether to keep two 'pools', a global and local ones
20:23:34 <Alberth> set during linking, i think
20:23:34 <andythenorth> global and local scopes :P
20:23:41 <andythenorth> make the author do it right
20:24:20 <planetmaker> set during linking means we need to run at least an ID replacement in the object code
20:24:46 <Alberth> you need to do that anyway, to connect stuff together
20:25:00 <Alberth> also you need to write out the result, so you need to load all data
20:27:33 <andythenorth> don’t we need to then inspect all IDs?
20:27:59 <andythenorth> or does using pools prevent colissions?
20:28:04 <LordAro> clearly you should be writing nnml
20:28:12 <planetmaker> andythenorth, not necessarily. if the objectfile specifies the localIDs just and indicates those which are global and need remain unmodified
20:28:14 <LordAro> new newgrf meta language
20:29:02 <Alberth> I am not convinced you have any global IDs
20:29:21 <andythenorth> chameleon forces local for IDs (var names). Global IDs have to be explicitly defined, and are disliked
20:29:24 <Alberth> just claim you need one, and the linker will give you one
20:29:26 <andythenorth> like many other languages
20:30:23 <Alberth> note in my book, an ID is just a name, rather than a number
20:30:42 <andythenorth> can’t we do something brutal, like just force declaration of all global IDs at top of file?
20:31:00 <planetmaker> Alberth, I'm conviced you have global IDs
20:31:12 <planetmaker> e.g. the grf parameters, they are necessarily global IDs
20:31:45 <planetmaker> newgrf parameters? We cannot do without them, can't we?
20:32:00 <planetmaker> how would we treat parameters which set vehicle costs?
20:32:22 <Alberth> name it "vehicle_cost_parameter_setting"
20:32:50 <Alberth> the action 14 declares it as provider, and some other code declares it as needing it
20:33:05 <Alberth> does it matter what number it actually has?
20:33:22 <planetmaker> not. As long as it's the same in every linked module
20:33:50 <Alberth> the linker ensures that everybody using that name will get the assigned number
20:34:17 <Alberth> and even after the last use, it may be re-used for other purposes
20:34:36 <Alberth> the linker can figure out what the last user is, I think
20:35:07 <Alberth> ie the intermediate format uses symbolic numbers, rather than real ones
20:38:06 <andythenorth> unrelated, but I wondered about teaching nmlc to compile from multiple fles (from a manifest)
20:38:43 <andythenorth> seems like a simple thing that would help some authors split up their code better
20:44:48 <Alberth> hmm, that would be before the scanner, it's quite feasible, actually
20:45:34 <andythenorth> I would have thought so
20:45:44 <andythenorth> I have python code in my compiles for it
20:46:25 <Alberth> it needs some error reporting fixing perhaps
20:47:01 <Alberth> and Python doesn't exactly like long string manipulations
20:48:02 <andythenorth> can’t we just append to an object for readlines in files or something?
20:48:08 <andythenorth> that may make no sense
20:48:50 * andythenorth plays actual ottd
20:50:13 <Eddi|zuHause> ... but... cpp already handles connecting files and error reporting locations...
20:50:17 <Alberth> it doesn't read lines, unless you modify the scanner
20:51:03 <Eddi|zuHause> nmlc even has code for understanding cpp error location output
20:51:25 <andythenorth> teach authors cpp?
20:51:36 <Alberth> otherwise people get hopelessly lost with an error
20:51:42 <Eddi|zuHause> not sure if it also resolves the "included from ..." part
20:51:44 <andythenorth> everyone has to install gcc?
20:52:04 <planetmaker> doing without cpp would not be a bad idea
20:52:05 <Eddi|zuHause> andythenorth: everyone has to install make?
20:52:51 * Alberth doesn't understand the need for make, except as an easy way to bundle scripts in one file
20:52:59 <Eddi|zuHause> planetmaker: but do you want to reimplement _every_ feature of cpp? if you don't, people who use those features will just stick with cpp anyway
20:53:16 <andythenorth> I only use make because of devzone
20:53:20 <Rubidium> maybe a stupid idea... but doesn't NML need PLY? What about /usr/share/pyshared/ply/cpp.py ?
20:54:07 <Eddi|zuHause> Rubidium: you mean like nmlc just runs cpp internally?
20:54:17 <andythenorth> and because “ain’t broke, don’t fix"
20:54:18 <planetmaker> Eddi|zuHause, no. But which of those features are actually used?
20:54:19 <Alberth> Rubidium: I looked at that, but iirc there was something not useful about it
20:54:29 <Rubidium> Eddi|zuHause: exactly ;)
20:54:46 <Eddi|zuHause> planetmaker: i certainly use some pretty crazy features
20:55:28 <Eddi|zuHause> planetmaker: just take the VEH_ID macro(s)
20:55:29 <planetmaker> Eddi|zuHause, no-one will stop you doing that. And it won't become impossible
20:56:10 <Eddi|zuHause> no sane person would invent that :p
20:56:16 <planetmaker> but the average user might already profit much when a few basic cpp capabilities become possible
20:56:29 <planetmaker> like include or some simple macros
20:56:44 <Alberth> in particular multi-line macros
20:57:27 <Eddi|zuHause> "but who needs THAT??"
20:57:54 <planetmaker> Eddi|zuHause, the argument that everything needs to be mimiced is not one which cuts
20:58:16 <planetmaker> Nor what is needed. But providing more options for basic use helps to push the border to where more
20:58:24 <planetmaker> sophisticated parsers are needed further
20:58:32 <Eddi|zuHause> planetmaker: possibly, but wherever you make the cut, you will get complaints
20:58:48 <planetmaker> of course. But that's no difference to now
20:58:59 <planetmaker> and like always, every journey starts with a first step
20:59:13 <planetmaker> going like "you need everything" is not useful for that
21:00:17 <Eddi|zuHause> there is also the other direction: things that are done with cpp are often not the optimal solution, there may be better solutions (like avoiding double include and stuff)
21:00:37 <planetmaker> and if we always add the most-needed thing, we'll get there^
21:00:40 <andythenorth> is anything done with cpp an _optimal_ solution? o_O
21:00:59 <planetmaker> andythenorth, that's indeed a valid question. Possibly #include is. Not sure about the rest
21:01:13 <andythenorth> constant substitution is fricking awesome
21:01:26 <Eddi|zuHause> andythenorth: depends on your metric
21:01:32 <andythenorth> we got a long way in the first cpp FIRS with constant substitution
21:01:59 * andythenorth wonders about python built-in $ templating
21:02:15 <talebowl> I'm assuming that in general, if something isn't covered in the coding guidelines, it shouldn't be used, but I figured it wouldn't hurt to ask anyway: (a) should/can the member initializer list be used; -- (b) should/can uniform initializer syntax be used?
21:02:16 <Eddi|zuHause> you don't need constant substitution if the compiler has proper constant folding
21:02:50 <Eddi|zuHause> optimizing out "parameters" if they never change
21:04:52 <Eddi|zuHause> as such, i think "#define" is one of the least optimal solutions (the quick-and-dirty kind)
21:06:04 *** tycoondemon has joined #openttd
21:06:31 <andythenorth> in what perspective?
21:06:35 <planetmaker> Eddi|zuHause, but it helps for the legibility of code, if you can like #define some_stuff = var1 - var2 + 2 * var3 * param0
21:06:55 <andythenorth> Eddi|zuHause: I have a view, but why least optimal?
21:08:11 <Eddi|zuHause> planetmaker: but then you differentiate between the value of the term or the textual representation of the term. those are very different things
21:45:17 *** montalvo_ has joined #openttd
21:46:33 *** Progman has joined #openttd
23:29:28 *** Pereba_ has joined #openttd
23:33:50 *** Pereba_ is now known as Pereba
continue to next day ⏵