IRC logs for #openttd on OFTC at 2008-08-07
⏴ go to previous day
00:05:54 <spence> We're Now Seeking For A New IRCop, For More Info Type /Server -m IRC.P2PChat.Net -j #Morpheus
00:10:18 <ben_goodger> there's so little to do...
00:31:40 <Eddi|zuHause> if you ever think there is nothing to do, go to the suggestions forum, pick a random thread, and start implementing what they request there :p
00:33:04 *** Eddi|zuHause has joined #openttd
00:51:24 <ben_goodger> well, I can't actually write C
00:51:44 <Fennec> C code. C code run. Run, code, run. plz?
00:52:10 <ben_goodger> now, give me a python API and I'll be all over it :P
00:52:56 <Eddi|zuHause> languages are all the same...
00:52:57 * Fennec should stare at the openttd source some time :)
00:53:23 <Fennec> Eddi|zuHause: ha! Try using anonymous closures in C and see where you get! :)
00:56:14 <Eddi|zuHause> i'm not quite sure what "anonymous closures" are right now, but the openttd code uses all kinds of concepts that were not even invented when C was defined
00:56:32 <Fennec> earlier today I got to see the Perl debugger crash on some insane overloading semantics :P you never have to deal with those in C :)
00:57:49 <Fennec> Eddi|zuHause: little things like, oh, $a = 4; my $b = sub { return $a; }; $a = 5; print $b->(); # prints 4
00:58:58 <Fennec> great for metaprogramming.
01:00:01 <Eddi|zuHause> what does the "my" do there?
01:01:13 <Fennec> yes. my creates a local variable. as opposed to local, which does not. yes, this is quite possibly the stupidest single thing in Perl. :P actually the first $a could have a 'my' on it to. but anyway.
01:02:51 <Eddi|zuHause> i'm stil not sure what you are trying to tell me
01:04:46 <Fennec> sub wrap_maker { my ($tag) = @_; return sub { my ($html) = @_; return "<$tag>$html</$tag>"; } my $bold = wrap_maker("b"); print $bold->("i am in bold"); # prints "<b>i am in bold</b>"
01:04:53 <Fennec> I'm saying, in reference to <Eddi|zuHause> languages are all the same...
01:04:59 <Fennec> that there's no trivial analogue to that in C.
01:07:19 <Fennec> obviously, the same results can always be achieved, but they involve hanging on to more data structures instead of having the compiler implicitly hang on to them for you.
01:09:36 <Eddi|zuHause> yes, but i expect someone that says "i can ONLY program in language XYZ" to be less theoretically sophistically and not very abusive of extreme language features
01:11:20 <Fennec> heh. "abusive of extreme language features" ---> don't look up Test::Resub on CPAN, it'll scare ya :P
01:11:23 <Eddi|zuHause> and that last line of yours can easily be done with C++-templates
01:12:04 <Fennec> hmm? can it be done with said templates when I get my tag name and formatting metadata from user input or a database setting somewhere instead of at compile-time?
01:13:48 <Eddi|zuHause> you can disable any argument by changing the requirements on the fly... that does not make it a good discussion style
01:13:55 <Fennec> I mean, there's certainly ways to go about doing it (make yourself a new class which stores an appropriate interface and has an invocation technique). it's just an entirely different way of going about doing it.
01:14:30 <Eddi|zuHause> and you're certainly not going to sell me perl...
01:14:33 <Fennec> therefore, I simply contend, all languages are not the same.
01:15:10 <Eddi|zuHause> they are, with a sufficient level of abstraction
01:15:11 <Fennec> naah, I'd reccomend more Ruby-ish if you're doing something new and are otherwise interested in that direction I think.
01:16:08 <Fennec> with a sufficient layer of abstraction, a fish and a banana are the same. :)
01:16:26 <Eddi|zuHause> yes, that is what i am saying
01:16:31 <Eddi|zuHause> you can eat both
01:17:11 <Eddi|zuHause> there is no inherent reason why you would say "i can only eat bananas"
01:19:37 <Belugas> what about : i have no teeth?
01:21:05 <ben_goodger> or if you were a vegan.
01:21:11 <Eddi|zuHause> i would actually believe you :p
01:21:21 <ben_goodger> or a vegetarian, or if you were allergic to fish
01:22:11 <Eddi|zuHause> and if you are allergic to C, you should probably stay far away from here ;)
01:22:39 <Eddi|zuHause> otherwise you are just making up excuses not to try to learn it
01:23:36 * Belugas was once allergic to C, but thanks to ttrs2, he got kinda cured...
01:27:47 *** welshdragon has joined #openttd
01:28:40 <CIA-5> OpenTTD: belugas * r14011 /trunk/src/graph_gui.cpp:
01:28:40 <CIA-5> OpenTTD: -Codechange: not required to define an enum which was just the representation of another.
01:28:40 <CIA-5> OpenTTD: If you want to customize it more easily, why not a simple const of said enum value?
01:28:50 <Eddi|zuHause> anyway, a "bold=tag_factory('b')" is not a revolutionary programming pattern invented by perl...
01:56:55 *** welshdragon is now known as welshdra-gone
02:00:30 <Yexo> is it just me or is bugs/svn/hg .openttd.org down?
02:00:44 <Belugas> Yexo, we gave TTD Dude the same answer :D
02:01:00 <Yexo> that should teach him to read hext time
02:04:16 <Belugas> maybe a momentary lapse of IP-whatever
02:04:39 <Belugas> no, bugs is really down
02:04:51 <Yexo> I can use svn from the command line, but not the web interface
02:15:21 <Belugas> so it means the web services are down, but not the www one
02:15:43 <Belugas> looks like a job for TrueBrain that must be sleeping
02:16:03 <Yexo> hg isn't accessable at all
02:22:36 <Belugas> sorry, nothing i can do :(
02:22:57 <Belugas> and damned... DoDrawString is a TINY little bit EVIL
05:33:25 *** Vikthor has joined #openttd
06:02:23 *** Sir-Bob has joined #openttd
06:50:18 <Celestar> hows it going peter1138 ?
06:51:52 <Celestar> any new developments (=
06:51:59 <peter1138> Buttons for sorting
06:55:13 <peter1138> Makes the destination view much more useful, sorting by cargo waiting.
06:55:35 <Celestar> pulling a rebuilding
06:58:16 <Celestar> did you get a chance to do some profiling?
07:05:30 <Celestar> peter1138: oh .. profiling works now
07:07:59 <Celestar> > gprof openttd | grep Routing_t | head -n3 0.07 25.70 0.02 1764518 0.00 0.00 Routing_t::FindNextHop(unsigned short, unsigned short) const 0.07 25.74 0.02 1669320 0.00 0.00 Routing_t::UseVehicle(Vehicle const*, unsigned short, bool) const 0.04 26.59 0.01 1782248 0.00 0.00 Routing_t::RecomputeCache(unsigned short) const
07:08:12 <Celestar> there are the three most expensive routing functions
07:08:22 <Celestar> at 0.07, 0.07 and 0.04 %
07:08:35 <peter1138> C++ is inefficient... remember?
07:08:53 <peter1138> At least, everyone always said so, heh...
07:08:58 <Celestar> 9.12 2.54 2.54 7189643 0.00 0.00 Blitter_8bppOptimized::Draw(Blitter::BlitterParams*, BlitterMode, ZoomLevel)
07:09:29 <Celestar> this takes about ONE HUNDRED times as much time as the routing system (with 200 active vehicles and all cargo routed)
07:10:07 * Celestar stops thinking about optimizing Routing_t at once
07:10:14 <Suisse> waaaa i'm not lucky with damn assert today :'(
07:10:17 <Celestar> peter1138: the sortable cargo view is awesome
07:10:20 <Celestar> Suisse: what are you doing?
07:10:31 <Suisse> 2 assert in less than 5 minutes ^^'
07:10:33 <Celestar> peter1138: you forgot to adjust the TODO file (=
07:12:37 <peter1138> So the real question is why did the other patches have performance problems, and why is the rest of the code so slow? :P
07:12:51 <Celestar> we spend more time in GetCharacterWidth than in the entire routing system :P
07:13:17 <Celestar> peter1138: or we're just better :P
07:14:53 <Celestar> peter1138: I'm trying to use a bigger game
07:16:33 <Celestar> peter1138: I dunno, but the three real C++ components: YAPF, YAPP and Cargodest have no noticable performance impact on the game :P
07:21:38 <Suisse> someone know about the presignal entry bug?
07:23:09 <planetmaker> Suisse: what kind of behaviour do you refer to as bug?
07:23:21 <Suisse> if you are trying do drag you mouse on a train line where is a presignal entry, the next signal will be presignal entry instead of current signal
07:24:20 <Rubidium> Celestar: retry without newgrfs ;)
07:24:52 <Celestar> Rubidium: retry what?
07:25:11 <Celestar> Rubidium: only if you have a large newgrf-free savegame (=
07:25:19 <Rubidium> Celestar: Pile Transport
07:25:38 <Rubidium> Suisse: I've got no idea what I need to do to reproduce that
07:25:43 <Celestar> Rubidium: where is it?
07:25:51 <Suisse> (but i didn't remove newgrf in the r14005 ^^')
07:26:21 *** lobster_MB has joined #openttd
07:26:47 <Celestar> peter1138: so what remains? the vehicle-type view and the station size thingy
07:27:39 <Suisse> i can have the same problem with r14005
07:28:51 <Suisse> Rubidium linux or mac?
07:29:01 <Rubidium> Linux, but that doesn't matter
07:29:11 <Suisse> maybe a wrong predefine :O
07:29:48 <Rubidium> as if the game code runs on defines?
07:30:11 <Suisse> run no, compiled yes :o
07:30:17 <Celestar> Rubidium: there is something very very very wrong with that game
07:30:21 <Celestar> Rubidium: non of the vehicles move
07:30:34 <Rubidium> remove your default newgrfs
07:30:52 <Rubidium> it's a game from before the "save newgrfs in savegame", so it takes the newgrfs from your default config
07:31:07 <peter1138> Yeah, signal dragging copies pre-signals now.
07:31:37 <Celestar> Rubidium: that's better (=
07:31:45 <peter1138> Should be simple to fix.
07:31:49 <Suisse> just the presignal entry
07:32:05 <Suisse> and i don't want to search by myself x)
07:34:11 <Celestar> Rubidium: not that you transport much cargo in Pile transport with destinations activated :P
07:35:34 <Celestar> Rubidium: bah... they did only 117kunits of cargo on that one ..
07:35:45 <Celestar> my YAPP-test game had 2.5 million
07:36:06 <Celestar> that's about 20 times as much
07:36:17 <Rubidium> hmm, so it isn't their biggest one?
07:36:59 <planetmaker> Moin, Rubidium: we had bigger games.
07:37:01 <peter1138> Suisse, no, copies all pre-signal types...
07:37:15 <planetmaker> but not newgrf free games
07:37:47 <planetmaker> peter1138: usually dragging only makes sense with pbs signals, pbs one-way signals and normal one-way...
07:38:15 <peter1138> planetmaker, yes, I'm talking about a bug...
07:38:22 <Rubidium> planetmaker: according to your website it is "our largest game ever"
07:38:32 <planetmaker> Rubidium: at that time (2005) :)
07:38:51 <Rubidium> planetmaker: then fix that and tell us what is the largest game ever!
07:39:08 <Celestar> Rubidium: no noticable impact of cargodest (=
07:39:41 <planetmaker> but much less compact than your game
07:39:46 <Suisse> only normal signal work :/
07:41:47 <peter1138> It's copying the signal type for everything, a normal signal will obviously be copied as a normal signal
07:42:08 *** Sir-Bob has joined #openttd
07:42:12 <Celestar> how do I obtained the population of a town tile
07:42:52 <Celestar> I want some system that generations an equal amount of passengers (statistically) from A->B than from B->SA
07:46:17 <Celestar> Ammler: do you have a grf list for that one?
07:46:43 <Ammler> Celestar: I am checking, if current coop pack works with it
07:48:58 <Celestar> peter1138: for the station sizes concerning passengers and mail, we could just use the amount of passengers/mail that was generated, right?
07:48:59 <planetmaker> Celestar: you're routing everything so far, or just PAX?
07:49:05 <Celestar> planetmaker: configurable
07:49:10 <CIA-5> OpenTTD: peter1138 * r14012 /trunk/src/rail_cmd.cpp: -Fix (r13936): Don't copy pre-signal type when dragging signals.
07:49:23 <planetmaker> and nice commit :)
07:49:35 *** ChanServ sets mode: +o Bjarni
07:49:42 <Suisse> ho so it's not a new feature :O
07:49:52 <Celestar> planetmaker: currently only two items left on the TODO list
07:49:58 <Celestar> plus one or two minor things in the code
07:50:26 <Celestar> 1) Improve destination generation (bias the PRNG by station size)
07:51:02 <planetmaker> Sounds like pretty quick progress :)
07:51:23 <planetmaker> Hehe. All people from Munich central going to Kleinkleckersdorf village station :P
07:51:26 <Celestar> 2) Minimap a) color the beams by vehicle-type
07:51:39 <Celestar> planetmaker: about. yeah (=
07:52:19 <Ammler> Celestar: seems to work nice with GRFPack 7.1
07:52:27 <Celestar> Ammler: where is it?
07:52:49 <Celestar> peter1138: we need to adapt the configure script so that it searches for adjacencylist.hpp
07:53:03 <Ammler> well, you will have the most already
07:55:52 <Ammler> but only 290 trains and 350 rvs/trams
07:56:37 <Ammler> Hiroshima Tramset might also be important
07:58:15 <Celestar> Ammler: don't get Routing CPU usage above 1% with that one either
07:59:10 <Rubidium> yeah, but is the routing actually the part that takes the CPU power or is it the fact that there are many many more packets?
08:00:37 <Celestar> Rubidium: I don't think the amount of packets have a huge influence on CPU power, but memory consumption
08:02:02 <Rubidium> loadunload vehicle and age cargo are certainly negatively influenced
08:02:36 <Rubidium> and improved loading seems to get messed up with cargopackets (at least when I tried it a while ago)
08:02:48 <Ammler> those stations have waiting pax around >10k
08:03:32 <Ammler> we played that game right after you broke the 4k limit, we had then sometimes 100k or more waiting pax there... :-)
08:07:27 <Rubidium> oh noes... a limit has been broken? Lets fix it ;)
08:07:56 <planetmaker> first learn the rules, then bend the rules, then forget the rules.
08:09:22 <Ammler> Rubidium: you introduced then the soft- and hardlimit :-)
08:09:41 <Celestar> Rubidium: we need to merge cargopackets
08:09:49 <Celestar> Rubidium: based on age
08:09:50 *** [com]buster has joined #openttd
08:10:07 <Celestar> Rubidium: the most expensive thing is when you have millions of cargopackets
08:10:29 <planetmaker> transported cargo in that game approx 500000 pax per month
08:11:20 <Celestar> Rubidium: do you know anything about the improved loading code?
08:13:37 <Celestar> Rubidium: would you mind to have a look at it?
08:14:01 <Rubidium> I know nothing of the whole routing stuff
08:14:18 <Rubidium> and improved loading is dead simple
08:14:30 <peter1138> Improved loading fails a bit with destinations :(
08:14:30 <Rubidium> 1) you have an ordered list (on arrival time) of vehicles
08:14:51 <Rubidium> 2) you have a list with the amount of cargo of each kind
08:15:25 <Celestar> peter1138: where does it fail?
08:15:45 <Rubidium> 3) you for over list #1 with #2 as parameter and you decrease the amount of cargo whenever full load has been selected on that vehicle by the number of "items" it can store in the vehicle
08:16:02 <peter1138> Cargo quantity is allocated but destination is ignored.
08:16:14 <Rubidium> 4) when amount of cargo (list #2) is 0 you can't load anything
08:16:18 <Yexo> in one of the previous cargodestination patches, the loading was dead slow. IIRC thsi was because on every tick (or every few ticks) only one cargo packed is loaded, so if you have to load lots of small cargo packes, it takes a lot longer
08:16:39 <Celestar> we don'T have that problem (=
08:18:28 <peter1138> LoadUnloadStation()
08:18:32 <peter1138> int cargo_left[NUM_CARGO]
08:21:34 <peter1138> and economy.cpp:1508
08:23:37 <peter1138> We need to reserve cargo on a per-destination basis :(
08:25:28 <Celestar> Rubidium: peter1138: we also need to merge the packets. dunno how but we need to do it, otherwise the cargolists become so long that the looping give a noticable impact and the game "stutters" even tho CPU load is far from 100%
08:25:40 <Celestar> peter1138: plus the whole thing in Routing_t as well (=
08:26:56 <Celestar> paste the current code in NoRouting_t (=
08:27:59 <Celestar> peter1138: we need to reserve cargo on a per-nexthop basis I think, don't we?
08:29:09 <Celestar> so maybe make a list of nexthops
08:29:17 <Celestar> (which is easy, it's just the edge list)
08:30:22 <Celestar> and make a std::vector out of cargo_left
08:31:03 <Celestar> LoadUnloadVehicle then gets the nexts only the cargo_left[v->nexthop] something
08:37:16 <peter1138> Right, yes, nexthop, not destination...
08:38:01 <peter1138> What about cargo without a destination?
08:39:18 <Celestar> doesn't matter, it's only transitional when enabling the function in game
08:39:46 <Celestar> so assign to the first nexthop in the list?
08:40:35 * Celestar is just browsing through the other paxdest patch
08:40:43 <Celestar> return (uint32)sqrt(DistanceSquare(t0, t1));
08:40:46 <peter1138> Well, it matters if you don't have destinations enabled :)
08:40:55 <Celestar> peter1138: then don't use the nexthop
08:41:12 <Celestar> peter1138: yes, sqrt :)
08:41:20 <peter1138> Does it have a "GoSlowOnceADay()" function?
08:41:26 <peter1138> Or was it monthly? I never tried it.
08:41:38 <peter1138> I looked at the source and hated it ;)
08:42:03 <peter1138> I mean... all that extra junk saved in a new chunk for cargo packets... how odd.
08:42:59 <Celestar> it totally lacks const
08:44:26 <Celestar> for (o = first; o->dest != st->index; o = o->next) ; <= error o->GetType() maybe != OT_GOTO_STATION
08:44:54 <Celestar> oh it's checked later :o
08:50:23 <Celestar> peter1138: honestly, I dunno why the previous patch was slow
08:50:35 <Celestar> except that it looped all the passenger cargopackets once each month (=
08:51:50 <Celestar> peter1138: so are you working on the gradualloading thingy?
09:02:33 *** Doorslammer has joined #openttd
09:13:45 <peter1138> You have received a PGP Universal Secured Message from:
09:13:49 <peter1138> To read this message securely, please click this link:
09:13:59 <peter1138> Defeating the point, somewhat :o
09:14:06 <peter1138> Plus the link doesn't work, heh...
09:14:19 <planetmaker> now, that's secure indeed :P
09:35:15 <Wolf01> anybody who wants to help me on drag&drop road stations? (so they will support also extending adjacent stations like rail stations do)
09:37:43 *** Sir-Bob has joined #openttd
09:38:10 <ln> someone holds down the mouse button while you move the mouse?
09:38:57 <peter1138> Wolf01, I can't help at the moment but I will support the idea. Just code it nicely ;)
09:39:18 <peter1138> Placing multiple road stops is quite painful currently.
09:39:22 <Wolf01> ln: maybe I'm trying to do this for 2 weeks without getting something working?
09:40:58 <Wolf01> thank you peter1138 ;) the main problem about them is the 16 bays limit, but I have enough troubles with drag&drop to solve before that :P
09:41:09 *** thgergo has joined #openttd
09:41:33 *** Singaporekid has joined #openttd
09:41:45 <Celestar> Wolf01: it's a problem?
09:42:08 <Celestar> Wolf01: static const uint LIMIT = 16; ///< The maximum amount of roadstops that are allowed at a single station
09:42:11 <Celestar> there's your problem (=
09:42:42 <Wolf01> no, if I raise the limit I only move away the problem
09:42:47 <Celestar> it's just there because DV insisted it to be there
09:43:21 <Celestar> just raise to 64 if you want to (=
09:43:49 <Wolf01> I want to keep all the things as they are now, maybe I'll put a max bays patch options like the max station size
09:44:34 <Wolf01> but the problem is with the drag&drop, I need to count how many bays are already joined with the station I want to extend
09:44:54 <peter1138> That seems quite easy.
09:45:05 <peter1138> There must already be a counter...
09:45:56 <Celestar> I should convert multistop into a std::list
09:46:31 <peter1138> What is it currently?
09:46:42 <Wolf01> and the other 2 problems are: 1) if I drag an area of 5x4, where it will put the bays? 2) how does the drag&drop work, I was able to make a lot of features with that, but seem that something changed and I'm not able anymore
09:47:18 <Rubidium> peter1138: linked list inside the roadstop struct
09:47:22 <Celestar> peter1138: no a linked list
09:47:29 <Celestar> peter1138: std::list-by-hand
09:47:29 <peter1138> Wolf01, for directional, limit to 1-dimension
09:47:32 <Rubidium> like is done for vehicles and orders
09:48:32 <peter1138> For drive through, 5x4 is fine.
09:48:58 <Celestar> bus stop on a 8-lane instate?!
09:59:31 <Celestar> it shows the station input, output and throughput
09:59:59 *** Martijn has joined #openttd
10:00:15 <blathijs> Celestar: And why are there six numbers for passengers then?
10:00:43 <peter1138> Celestar, from the other patch?
10:00:46 <peter1138> I think it's confusing.
10:00:54 <peter1138> No explanation of the data.
10:01:42 <Martijn> Does anyone here know how to set a money limit to a game?
10:02:00 <Martijn> I'm running a dedicated OpenTTD server and I think this would be a nice feature, but I cant seem to find the setting for it
10:02:18 <peter1138> fonso, is it possible for that to be Tile_t::iterator or TileIndex::iterator, like other (stl) C++ iterators?
10:02:42 <fonso> I'm not looping over a single tile
10:03:14 <fonso> but we have different ways of iterating
10:03:20 <fonso> we need different iterators
10:03:29 <peter1138> Certainly an iterator is nicer than a macro.
10:03:51 <fonso> but you can still use the macro in critical places
10:04:06 <fonso> because of virtual members
10:04:19 <peter1138> Does it need virtual members?
10:04:33 <fonso> otherwise we don't get a common base class
10:04:45 <peter1138> Right, but do we need a common base class?
10:04:56 <peter1138> The operations are not interchangable are they?
10:05:00 <fonso> it would be nice in terraform_cmd.cpp
10:05:01 <peter1138> Or maybe they are...
10:05:19 <fonso> depending on a flag I construct one iterator or the other
10:05:25 <fonso> and then use them interchangeable
10:07:15 <peter1138> If you used the OrthIterator directly would it still be virtual? Probably, I guess...
10:07:28 <Celestar> peter1138: it's not finished
10:07:37 *** Progman has joined #openttd
10:07:44 <peter1138> Celestar, what isn't?
10:07:56 <Celestar> blathijs: one set current month, one set last month
10:08:01 <fonso> if the compiler is stupid it might look for further subclasses
10:08:10 <Celestar> peter1138: dunno, I've made with patch for r870 (old repo) years ago
10:09:16 <Celestar> also a vehicle counter per month
10:09:47 <fonso> peter1138: if you call OrthIterator::operator++ explicitly like it's not virtual
10:10:25 <Celestar> peter1138: I'll need the data for dest generation anyway, the question is whether to display it or not
10:10:37 <peter1138> So it could be fast in normal use... Wonder if it's inlinable too...
10:11:03 <peter1138> Celestar, do you gain anything by seeing it?
10:11:56 <Celestar> peter1138: dunno, you could optimize your stations possibly
10:11:57 <blathijs> Celestar: Shouldn't input and output sum up to throughput?
10:12:07 <Celestar> blathijs: throughput is the stuff that it transferred
10:12:14 <Celestar> peter1138: but possibly not, no
10:13:01 <blathijs> Celestar: Ah, input and output is stuff that entered/left the stations through the doors, not what entered/left over track :-)
10:13:25 *** [com]buster has joined #openttd
10:13:37 <dih> Martijn, you can set the start loan, but you cannot define the max value a company may have!
10:15:18 <Martijn> But there seem to be servers that do have this feature!
10:15:48 <Bjarni> at least I haven't seen anything like that
10:16:00 <Bjarni> and I find it hard to know how they would have implemented it
10:16:23 <Bjarni> or do you mean max loan?
10:16:32 <Martijn> I dont know if its company value... I think its just the money you currently have
10:16:46 <Martijn> i've seen some servers that end the game if a player reaches $10M for example
10:18:26 <Bjarni> I guess that's more like a stop criteria than a limit
10:18:50 <Martijn> Its a bit like the game restarting when it reaches 2051
10:19:18 <Martijn> But instead the game restarts when a player has a set amount of money
10:19:54 <Bjarni> sounds like a script that keeps track of the players and sends a restart command if the value is higher than a certain amount
10:20:20 <dih> Martijn, there are scripts that wrap around the server console
10:20:30 <Martijn> Like autopilot you mean?
10:20:32 <dih> they can check for money or company value
10:20:39 <dih> pause - say a few lines, issue newgame
10:20:53 <dih> though i think Kurt uses a python script
10:21:06 <Martijn> I dont know who Kurt is :+
10:21:08 <Celestar> peter1138: the "loading iterator" takes up to 2 million cycles :(
10:21:19 <Celestar> peter1138: economy.cpp:1600
10:21:21 <dih> Kurt is the guy who runs the server 'Kurts hard limit' :-P
10:21:35 <dih> that is the server that finishes at 10Mio
10:21:42 <Bjarni> it wouldn't be feasible to code a limit to company value. It would be like "can't place track here. Company is too big"
10:21:57 <dih> Bjarni, that would be funny :-D
10:22:09 <Martijn> Does Kurt ever come by here?
10:22:26 <Martijn> It seems like I need the script that Kurt is using, right?
10:22:36 *** welshdra-gone is now known as welshdragon
10:23:16 <Bjarni> if you want that feature
10:23:32 <Martijn> Lets see if I can find out how to contact this Kurt
10:23:35 <dih> Martijn, Kurt is not oblidged to hand out his script
10:23:48 <Bjarni> you can still ask him for it though
10:23:56 <Bjarni> but if he says no then you have to accept that
10:24:00 <dih> the best bet to get in touch with him is to go on his games or visit his website
10:24:07 <dih> also, if you feel like doing some coding
10:24:12 <dih> you can use autopilot :-P
10:24:31 <Martijn> Does autopilot/ap+ also include this feature?
10:24:41 <Martijn> Or is that what you mean by "doing some coding"? :P
10:24:50 <peter1138> Celestar, our new one?
10:25:01 <dih> that is what i mean by 'doing some coding'
10:25:06 <peter1138> Celestar, that'll be better if you combine packets, right?
10:25:22 <dih> you have a good wrapper to the console, all you need to do is implement that feature
10:25:42 <Martijn> Too bad I've never coded TCL before though
10:27:04 <Celestar> peter1138: MUCH better
10:27:14 <Celestar> peter1138: it'll just be much much much fewer packets
10:27:28 <Celestar> peter1138: I have another idea
10:28:17 <dih> Martijn, it's more simple than odd
10:28:34 <dih> you should be able to just copy past from examples in the code itself
10:28:53 <Martijn> You're making it sound attractive :P
10:28:56 <peter1138> Celestar, store packets by nexthop?
10:29:07 <Celestar> peter1138: problem solved ;)
10:29:18 <Celestar> peter1138: I've gotten the time for that loop from O(n) to O(log n)
10:30:05 <Celestar> by adding a dozen lines
10:31:10 <Celestar> well not quite from O(n) to O(log n) but something in that magnitude
10:31:38 <Celestar> peter1138: the thing is, it makes no sense to call UseVehicle over and over and over again, with u and cp->target being the same for a shitload of packets
10:34:09 <Celestar> peter1138: so I'm caching that result locally in the loop (=
10:36:04 <Celestar> meh how do I print a '%' in fprintf?
10:36:50 <Tefad> have you guys not switched to streams yet?
10:37:03 <Celestar> Tefad: unfortunately no :(
10:37:08 <Tefad> or is the whole thing C with classes
10:37:29 <Tefad> btw, i'm relearning C++ as of late
10:37:40 <Tefad> funny how learning java breaks all my C++ knowhow : x
10:38:15 <Tefad> NO you can't create new objects inside a function argument list that accepts references.
10:38:52 <Tefad> because you have to keep track of the pointer so you can free it later.. silly memory management.
10:38:56 <Celestar> Tefad: currently, it's basically mixed C/C++
10:40:49 <peter1138> Celestar, ah... handy.
10:42:20 <Celestar> peter1138: that's the resolt
10:44:32 <Celestar> peter1138: it slows down the smallish loop
10:45:00 <peter1138> You mean it's slower if there a only a few packets?
10:45:01 <Celestar> but it gives a huge boost (about an order of magnitude) if the packet count is noticable (like 1000, sometimes even more)
10:45:07 <Celestar> but that's hardly noticable
10:45:24 <Celestar> I'm doing an overall test
10:45:33 <peter1138> Hmm, maybe have both paths, and use the cache only if there are a lot of packets.
10:47:54 <Celestar> averaging over 100000 loops altogether
10:51:03 <Celestar> 90000 cycles with everything cached
10:51:04 <peter1138> Instead of 2 million?
10:51:23 <peter1138> That's quite a bit faster :D
10:53:12 <Celestar> 2 million wasn't really averaged
10:53:17 <Celestar> just over a few loops
10:53:22 <Celestar> trying again uncached
10:54:54 <Celestar> wait about 8 minutes
10:55:01 <Celestar> (because that what about 100k loops will take)
10:59:17 <fonso> #define INVALID_TILE MapSize()
10:59:17 <fonso> Am I the first one who wants to do something like that?
10:59:42 <Celestar> peter1138: 700000 cycles without the cache
11:00:14 <Celestar> peter1138: but the thing is, the really big loop make the game stutter, because they can take well over a few milliseconds apparently
11:00:36 <peter1138> Celestar, so it's still no good? :(
11:00:43 <blathijs> fonso: Don't we have an INVALID_TILE defined somewhere?
11:00:54 <peter1138> fonso, what's wrong with its current value?
11:01:17 <blathijs> And your proposal isn't?
11:01:46 <fonso> I'm not the first one, obviously
11:02:10 <peter1138> INVALID_TILE is not a define :)
11:02:56 <peter1138> Oh, had you not seen it at all? heh
11:03:17 <fonso> Somehow I must be blind today
11:12:16 <Celestar> peter1138: I gain nothing from not caching
11:12:26 <Celestar> (difference is below background noise)
11:17:46 <Celestar> the cache helps much much more than the packet merge :D
11:17:48 <peter1138> Wolf01, that's a blank file.
11:17:59 <peter1138> Celestar, but both together?
11:18:39 <Wolf01> ok, now it should work
11:24:40 <Celestar> (NO SIGnificant change)
11:24:48 <Celestar> the loop it 10 times slower than without routing
11:25:02 <Celestar> I take about 8500 cycles without routing
11:25:06 <Celestar> and 90000 cycles with routing
11:27:34 *** Wezz6400 has joined #openttd
11:27:44 <Celestar> well I used 700000 cycles before, so 90000 is a good thing (=
11:33:40 <Celestar> peter1138: pushed onto server
11:34:42 <Celestar> peter1138: so what to do with improvedload?
11:35:10 *** Brianetta has joined #openttd
11:38:56 <dih> is there anything that could be done to extend the length of lines one can chat in the game
11:39:02 <dih> 64 chars is amazingly short
11:39:23 *** SquireJames has joined #openttd
11:39:27 <dih> i know that most strings are limited to the same length
11:39:33 <SquireJames> Howdy boys, long time no see
11:39:36 <Celestar> Rubidium: do you have a sample savegame where improved load fails
11:39:38 <dih> but for the chat it has actually only become an annoyance
11:40:02 <SquireJames> Yup, i got bit by the TTD bug again
11:40:25 <SquireJames> Played "Trainz" for a while but, short of paying 3 quid(!) per loco, not much UK content
11:40:34 <Celestar> peter1138: when do you think it's time to go public? ;)
11:41:20 <SquireJames> Love the new PBS, been waiting for her for a while, just need programmable signals and i'll be in OTTD perfection
11:41:35 <Celestar> what do you expect programmable signals you do?
11:41:56 <SquireJames> Not demanding them, just seem like a good idea
11:42:29 <SquireJames> Devs agreed when I posted about em long ago, and some form of route priority or filtering would be nice with all these lovely cargo specific platforms
11:45:45 <Rubidium> Celestar: I'll try to find one
11:47:37 <SquireJames> So, without wishing to pry, whats the newest aim of OTTD?
11:47:56 <SquireJames> PBS and NewIndustries/NewCargos largely implemented these days
11:48:13 <dih> SquireJames, it's a secre
11:48:19 <SquireJames> sounds good, general tidying is always welcome
11:48:28 <SquireJames> ah, yes the AI thingy, i heard about that
11:48:42 <planetmaker> you have to play them...
11:48:51 *** grumbel has joined #openttd
11:49:16 <SquireJames> Ah, so some have been completed?
11:49:24 <planetmaker> one could even say too fast for a single person...
11:49:35 <Rubidium> yes, completed in the sense that OpenTTD is completed
11:50:09 <Celestar> Rubidium: is it planned to merge NoAI in trunk at some point
11:50:34 <Celestar> SquireJames: well, NewGRF_ports is one thing thats around, fixing bugs/problems and another huge feature that peter1138 and I are working on at the moment (=
11:50:39 <Rubidium> given a few preconditions, then yes
11:50:47 <Celestar> Rubidium: those being? ;)
11:50:54 <Rubidium> it being finished and such
11:50:58 <SquireJames> working as in "work but always room for improvement" :)
11:51:19 <Celestar> Rubidium: what's it's status? is there a wiki for it or anything?
11:51:27 <SquireJames> well, good to see you gentlemen are as dedicated as ever, makes a nice change given some of the game modding/dev circles i've been in of late
11:53:12 <Rubidium> but currently they're getting rid of the scripting engine and replacing it with another one that does more sane things
11:55:51 <Celestar> Rubidium: who's on it currently anyway?
11:56:20 <Rubidium> currently mostly TB and glx with help from Yexo
11:57:16 <Celestar> shouldn't train support be *much* easier now with YAPP
11:57:18 <Rubidium> Tefad: the squirrel engine is one of the major causes of bugs and especially the nasty type of bugs
11:57:43 <Tefad> i knew there was a reason i didn't like squirrel : x
11:58:11 <Rubidium> Celestar: no; it didn't remove complexity in building railways
11:58:43 <Celestar> Rubidium: but it removes the complexity in the signal networks
11:58:47 <peter1138> Rubidium, but building safe/non-blocking routes might be simpler.
11:58:57 <Celestar> Rubidium: the networks are just as complex as before
11:59:00 <peter1138> Not a huge benefit though.
11:59:05 <Rubidium> that is an issue for the AI writers, NOT the API
11:59:26 <Celestar> why is my cluster quota 10GB
11:59:29 * Celestar goes uploading again
11:59:38 <Rubidium> it actually makes the API more complex as there are now more states a vehicle can be in
12:00:26 <Rubidium> and the main thing we want to avoid is passing trackdirs and such around
12:00:26 <blathijs> Rubidium: What scripting engine will they be using instead?
12:00:26 <Celestar> peter1138: I'm not yet fully sure how to tackle the improvedloading problem
12:01:01 <SpComb> Yet Another Scripting Language :(
12:01:02 <Celestar> peter1138: something that is again transparent to economy.cpp and done in the routing classes
12:01:15 <peter1138> NAIL? What happened to whatever it was?
12:01:18 <Rubidium> SpComb: noo... Not Another Interpreted Language
12:01:23 <blathijs> Hmm, google doesn't seem to know the nail scripting language
12:02:23 <Rubidium> peter1138: it'll get dumped, primarily because it has a hook that tells when stuff goes out of scope that... uhm... doesn't always tell when stuff goes out of scope
12:02:38 <peter1138> Squirrel and C++ originally :o
12:02:45 <blathijs> Yipes, their building their own scripting language?
12:03:39 <Rubidium> lua, perl and python are all relatively *slow*
12:03:48 <glx> squirrel is full of strange design decisions
12:04:05 <Rubidium> factor 4 slower than nail/squirrel on e.g. fibonacci
12:04:19 <Celestar> Rubidium: the netgame we tested, did your company have the improvedload problem there?
12:04:40 <Rubidium> it was likely not turned on anyways
12:06:25 <blathijs> If they really want to have their own language, they should try to make it one where safety can be proved an the result can be JIT'ed to native code
12:06:49 <blathijs> Just-In-Time compilation
12:07:07 *** planetmaker is now known as publicserver2
12:07:18 *** publicserver2 is now known as planetmaker
12:13:05 <Celestar> thanks Rubidium I'll have a look at it right now
12:13:15 *** KillaloT has joined #openttd
12:13:31 <SpComb> hmm, the nail code seems surprisingly concise
12:14:26 <SpComb> if whoever's writing it managed to understand their needs right, and the code is already nearing completion, then I guess it's ok
12:14:40 <SpComb> but it seems there's no support for libraries, including other modules/files?
12:15:04 <SpComb> and it does seem to have an interpretive VM
12:15:15 <SpComb> I thought for a moment that it compiled into machine code :)
12:15:42 <Celestar> we've written a compiler :P
12:16:31 <Rubidium> if only it would compile to the native code of all architectures we support
12:16:37 <SpComb> Celestar: are you serious or joking? It's not in the hg repo
12:16:54 <SpComb> oh, right, Pawn also uses an abstract machine
12:17:15 <Celestar> Rubidium: ah so the other trains just sit around because they think they have reserved stuff but can't load it anyway?
12:17:22 <Forked> how did they compile the first compiler!?!? (ok I know they just wrote it in machine code, but that answer sucks)
12:17:31 <SpComb> have the NAIL people looked at e.g. Pawn?
12:17:45 <SpComb> I've never looked at Squirrel, so I don't really know what it is like
12:18:19 <Noldo> SpComb: last time I checked Pawn it had some 64-bit issues
12:18:27 <Celestar> Rubidium: ok .. the vehicle need to remain in the station because they have full load
12:18:34 <Celestar> Rubidium: just the loading order is poor at best, right
12:19:05 <SpComb> true, it says "pawn is a ... 32-bit extension language with a C-like syntax" on the box
12:19:09 <Rubidium> the loading order is "right" if improved loading is turned off, because then it's full load the vehicle in any order you like
12:19:17 <SpComb> (hmm... I copy-pasted that as "2-bit extension language")
12:19:21 <Rubidium> with improved loading it is first-in-first-out
12:19:28 <Celestar> Rubidium: but not with destinations?
12:20:14 <Rubidium> well, the third train for the SAME destination leaves first, that's not right
12:20:26 <SpComb> so what happens with NAIL if someone writes a library of utility code and other people want to use it in their NAIL scripts?
12:20:40 <Celestar> Rubidium: I'll have that issue fixed next.
12:20:44 <Rubidium> with destinations it should be FIFO per cargo per destination
12:24:26 <peter1138> destination or nexthop? heh
12:26:14 <Rubidium> the destination as shown in the station's cargo list GUI thingy window
12:28:20 <Celestar> peter1138: logically, nexthop is enough afaik
12:33:18 <Celestar> peter1138: because cps get loaded by nexthop as well
12:42:09 *** LilDood has joined #openttd
12:44:08 <KillaloT> What happends when a vechicle is too old?
12:44:16 <KillaloT> I don't see any changes
12:45:04 <glx> except station rating is already low
12:52:16 <Belugas> KillaloT, iirc, it will break down more often
12:52:47 <Belugas> thus loosing profit since time for delivery will be longuer
12:58:26 <Celestar> peter1138: I still have this 'stuttering' in one of my games
12:58:39 <Celestar> peter1138: thing is. I have it without routing as well
12:59:01 <Celestar> I have it in tcsh too
13:08:05 <CIA-5> OpenTTD: rubidium * r14013 /trunk/src/rail_cmd.cpp: -Fix [FS#2202]: do not allow extending signals by dragging in any direction other than the track direction.
13:19:48 *** Dred_furst has joined #openttd
13:30:23 *** ChanServ sets mode: +v tokai
13:50:16 <Celestar> peter1138: without a more sophisticated passenger generation system, I see no real way to make routes usage symmetric
13:51:56 <SpComb> what's the status on passenger destinations now?
13:52:42 <hylje> like a traveller pool per city which increases when people destined there arrive, increases based on population, decreases over time
13:53:06 <Celestar> way too complicated for now
13:53:09 <Celestar> SpComb: mostly finished
13:53:24 <SpComb> what's the scope of it?
13:53:32 <SpComb> as in, what does it do
13:54:10 <SpComb> or "how much of the stuff considered to be "passenger destinations" does it implement?"
13:55:19 <Celestar> SpComb: what do you consider passenger destinations
13:56:42 <SpComb> something that picks destinations for passengers, picks a route involving multiple vehicles, and then gets the passenger to said destination
13:56:54 <SpComb> whereby there's a million different possible algorithms for choosing destinations and routes
13:58:57 <Belugas> "routes usage symmetric"... watzdat?
13:59:14 <hylje> Belugas: about as many people come and go in a route
13:59:18 <Celestar> Belugas: it means that there is an equal amount of pax from A to B as from B to A
13:59:41 <Celestar> Belugas: besides nettest, it's the last big item on the todo
14:00:19 <SpComb> Firefox can't establish a connection to the server at galadriel.td.mw.tum.de.
14:00:21 <blathijs> Celestar: connection refused?
14:00:53 <SpComb> hmm... how does the passenger generation algorithm work?
14:01:01 <Belugas> you know, i don't think it is important. A can be far more popular than B and does not mean that whomever goes to B from A must go back to A after whatever needed to be done in B
14:01:07 <SpComb> I recall seeing something about it only generating passengers to destinations that are part of the network?
14:01:12 <Belugas> ho my god... it sounds luile realism...
14:01:33 <SpComb> does it just pick a destination station at random, or is there some kind of logic re small towns/big towns central station/suburbs?
14:01:54 <Celestar> blathijs: sorry. port 8000
14:01:59 <hylje> i dont think it takes population/station size into account yet
14:02:19 <hylje> ive had a massive flood of people going from a large terminus to a minor town
14:02:38 <Celestar> SpComb: nearness (and a size-based prototype on my local copy)
14:02:51 <hylje> the distance to the target station
14:03:23 <SpComb> so it distributes the available passengers to destination stations based on the inverse distance?
14:03:35 <blathijs> Celestar: Why not use hg.openttd.org, btw?
14:03:45 <hylje> blathijs: local development
14:03:53 <Celestar> blathijs: we're developing locally at the moment
14:04:10 <Celestar> hylje: well I must admit that it works like a char (=
14:04:28 <blathijs> Celestar: Yeah, I meant pushing to a celestar repository on hg.openttd.org
14:04:32 <SpComb> what criteria does it use to pick the route?
14:04:42 <blathijs> Celestar: But I presume that galadriel is the machine you are sitting behind?
14:04:48 <Celestar> blathijs: yeah kind of
14:04:52 <Celestar> my workstation @ work
14:04:59 <Celestar> it's on 24/7 anyways
14:05:06 <Celestar> unless I'm on vacation
14:05:12 <SpComb> writing OpenTTD code at during work hours
14:05:21 <blathijs> Or perhaps not sitting behind, but at least compiling on :-)
14:05:36 <Celestar> blathijs: nope, I'm compiling on and sitting behind my laptop (=
14:05:46 <Celestar> better display, better keyboard, better CPU, more RAM, better GPU
14:06:09 <Celestar> only thing workstation is good for is visualizing stuff, because it has two TFTs connected
14:06:09 <blathijs> Celestar: In that case, instead of pushing to galadriel, why not push to hg.openttd.org instead?
14:06:19 <Celestar> blathijs: less setup work
14:06:50 <Celestar> SpComb: the repo contains a "TODO" file
14:06:50 <blathijs> Celestar: Lazy ass :-p
14:07:53 *** frosch123 has joined #openttd
14:10:12 <blathijs> peter1138: Is there any particular reason that there are "merge from <peter|celestar>" commits in the hg repos? Can't hg just import individual commits?
14:10:39 <SpComb> those are conflict-merges
14:10:50 <SpComb> or even non-conflict merges
14:11:19 <SpComb> so it imports the history, and then merges the two together
14:13:05 *** welshdragon has joined #openttd
14:13:14 <blathijs> SpComb: But it seems that (at least in some cases) these merges really just merge a single commit
14:13:23 <blathijs> Why not reuse the original commit message then?
14:13:54 <blathijs> I find that one of the advantages of git, if I pull in changes from your repos, I get individual commits with the original commit messages
14:14:08 <SpComb> if you take two tips and merge them together into one tip, then that new tip is differnet from the earlier tip, and thence, it's a different commit/revision
14:14:36 <SpComb> you'll notice that Celestar's repo also contains e.g. the svn commits, followed by a "Merge with trunk"
14:14:53 <SpComb> which indicates that petern merged his local copy with trunk, and then Celestar later pulled in petern's stuff as well
14:15:11 <blathijs> SpComb: Yes, but you could take all the commits in one branch (up until where those branches merge again) and commit then apply those commits, one by one, to the other branch
14:15:24 <SpComb> but I've only read the hg book, and not actually used it all that much, so I don't have very much pratical experience yet
14:16:51 <peter1138> It's a bit irritating.
14:16:55 <peter1138> I'm probably doing it wrong ;)
14:18:56 <blathijs> That's something that's particularly cool in git, but totally breaks when other people have branched your branches
14:19:20 <blathijs> that's probably why there are these merge commits, the alternative I described is essentially rebasing
15:11:48 *** FauxFaux has joined #openttd
15:21:35 <Eddi|zuHause> what exactly are you trying to tell us?
15:22:02 <FauxFaux> That I need to file a bug report about the line passing below zero, if that was at me.
15:22:46 <Eddi|zuHause> no, i was talking with the other 20 people that said anything during the last hour.
15:23:19 <FauxFaux> I recently re-joined (I assumed there was a netsplit or something), I may have missed something. :)
15:25:28 <FauxFaux> Which appears to be down from here. Oh, maybe it's sourceforge migration time.
15:35:49 <Eddi|zuHause> webservices having trouble today
15:38:13 <Rubidium> lighttpd was for some reason gone
15:39:26 <rortom> are you using the new lighttpd version? or the latest stable?
15:40:57 <rortom> we are using lighttpd-1.5.0
15:41:05 <rortom> it has doubled our speed
15:41:14 <rortom> when we switched from 1.4.x
15:42:01 *** peter1138 has joined #openttd
15:42:01 *** ChanServ sets mode: +o peter1138
15:42:02 <rortom> server.network-backend="linux-aio-sendfile" is much faster
15:44:08 <Rubidium> why do python applications require so much memory/leak that badly?
15:44:12 <peter1138> 'The Linux Game Tome' have changed us to OpenTTD :D
15:45:00 <Zuu> What was it before "Open Transport Tycoon Deluxe"?
15:45:07 <peter1138> "Open Transport Tycoon"
15:45:30 <Zuu> Cool, they didn't see the 'D' even :)
15:45:56 <rortom> Rubidium: depends o the python application
15:46:11 <Fennec> OpenTTD! The TTD doesn't stand for anything! Just like HD Radio! :P
15:47:13 *** Eddi|zuHause has joined #openttd
15:47:44 *** Wezz6400 has joined #openttd
15:48:13 *** Wezz6400 has joined #openttd
15:49:00 <Bjarni> Anonymous on Linux Game Town wrote the comment "Racist sexist racist game"
15:49:16 <Bjarni> idiots are everywhere on the internet
15:49:53 <Eddi|zuHause> i should know... i have one right in front of me :p
15:50:12 <Bjarni> I have an internet right in front of me
15:50:29 <Bjarni> you wanna trade internets?
15:50:46 <Eddi|zuHause> yeah. definitely...
15:52:10 <Bjarni> Ok, this is a wonderful game, but I think the team should put more efforts in making a full 100% graphics replacement for at least one scenario (all needed terrain, all kind of tracks, all specific vehicles...), the original game is no more for sale and abandonware is NOT legal. <-- this guy actually have a point
15:52:30 <Bjarni> I'm not asking for the Ultimate Wonder in OpenTTD visuals - just a working graphics, no matter if unpolished, and a basic soundset to play this game out of the box without infinging law. <-- from the same guy
15:53:03 <Bjarni> I can't draw on computers :(
15:54:12 <Fennec> eh, drawing on paper is more awesome :) but less useful I guess
15:54:26 <Bjarni> do you have a scanner?
15:57:08 * Fennec has to go riiight about now though
15:57:19 <Fennec> my train approacheth its destination
16:06:03 <Bjarni> there is a hand drawn simutrans graphic pak
16:18:39 <Zuu> Yea, never tried it but saw it some time ago. :)
16:22:16 <Bjarni> took a while since SimuTrans suffered a server crash and lot a lot of stuff :(
16:24:09 <Eddi|zuHause> i don't suppose this would be adaptable to openttd ;)
16:24:15 <Bjarni> it's actually way better than I imagined when I first read about it
16:24:19 <peter1138> orudge, what happened to the simutrans' forum idea? heh
16:26:01 <peter1138> They should've used Zernebok ;)
16:26:10 <orudge> the Simutrans forums people didn't reply to my messages after their initial enquiries
16:26:14 <orudge> and it seems they went with a different route
16:26:31 <orudge> even though I offered them hosting for all their attachements and posts, etc
16:30:06 <Belugas> i don't know , Eddi|zuHause, maybe it is, i really do not know
16:30:46 <Eddi|zuHause> i mean more from a license perspective than from a technical perspective
16:34:29 <Fennec> that hand-drawn stuff /is/ nice
16:36:08 <glx> SmatZ usually does these compile tests
16:41:22 <michi_cc> yeah, any ancient gcc will probably do
16:43:53 *** lobster_MB has joined #openttd
17:11:02 <Ammler> I made some research for free ssl today, https://www.ammler.ch should work without annoying warnings, maybe you like to make the cert for openttd.org there too? (www.startssl.com)
17:13:23 *** Wezz6400 has joined #openttd
17:13:29 <Rubidium> Ammler: startssl doesn't provide the type of ssl certificate we need for free
17:13:53 <Rubidium> yes, something like that
17:14:31 <Ammler> I just don't know a browser which trust cacert..
17:15:04 <Rubidium> but that's likely Debian's effort
17:20:50 <blathijs> Don't we just need a cert for secure.openttd.org?
17:21:12 <Rubidium> technically yes... but
17:22:55 <blathijs> True, but those are also available through secure.openttd.org/svn, right? (at least the first one, the second should just redirect anyway)
17:23:09 <Rubidium> yes, but the redirect will fail
17:23:23 <Rubidium> because the SSL certficate doesn't match
17:24:30 <blathijs> But you can't block the https port depending on the domain, because you only know which domain is requested after the SSL handshake
17:24:32 <Rubidium> and it won't redirect before the SSL certificate gets accepted and it will *always* complain because the domain doesn't match
17:24:48 <blathijs> Then again, people just shouldn't try to use those urls :-p
17:25:09 <Rubidium> it's all for "ease of use" of the website
17:25:33 <blathijs> There is some HTTPS extension to support HTTPS vhosting IIRC, which already has some browser support
17:26:26 <blathijs> Rubidium: How is it handled currently, btw? I don't get a certificate error, though the certificate seems to be issued to secure.openttd.org only
17:26:37 <blathijs> or is there a seperate ip address for secure. ?
17:26:48 <Rubidium> it's some magic certificate
17:26:54 <Rubidium> it's also for bugs.openttd.org
17:28:48 *** Eddi|zuHause has joined #openttd
17:29:11 <Rubidium> blathijs: some browser support == latest opera
17:32:33 <Ammler> Rubidium: did you add SSLCertificateChainFile and SSLCACertificateFile to apache conf, I forgot that, first time.
17:33:04 <Rubidium> don't use apache (for ssl)
17:33:58 <blathijs> Rubidium: Hmm, I thougth there was more than that
17:34:05 <blathijs> Rubidium: Hmm, magic certificates :-)
17:34:33 <peter1138> Has the reason for the morning brokenness been found?
17:34:55 <peter1138> And has the broken whitespace pre-commit hook been fixed?
17:35:03 <blathijs> Rubidium: Hmm, wget does complain about a non-matching hostname
17:35:04 <Rubidium> peter1138: should be
17:36:34 <blathijs> Rubidium: Ah, there are additional "Subject Alt Names" in the certificate
17:36:51 *** Alberth has joined #openttd
17:37:05 *** Celestar|GSC has joined #openttd
17:37:29 <Ammler> blathijs: seems not to work with my cert, either
17:39:13 <Celestar|GSC> peter1138: do we wanna rewrite the passenger generation system for towns/stations or not?
17:40:17 <Eddi|zuHause> in what way you mean rewrite?
17:41:00 <Celestar|GSC> start from scratch
17:41:08 <Celestar|GSC> otherwise we won't get symmetric loads
17:41:32 <Rubidium> but that'd mean changing steel mills too
17:41:48 <Rubidium> which'd break newgrf compatability
17:42:14 <Celestar|GSC> simple example: you have a network with only two stations: A) which has about 10000 inhabitants in its CA, B) which has 50 inhabitats in its CA. you have a single train running from A to B. we'd need a system where the nofpax generated at A is about the same as the nofpax generated at B
17:42:16 <Eddi|zuHause> how about generating passengers based on the amount of delivered passengers (with some random distribution)?
17:42:19 <Celestar|GSC> Rubidium: no I'm talking passengers only
17:42:47 <blathijs> Rubidium: According to that page, the TLS extension (SNI) is supported by IE7, Firefox 2, Opera 7.6. Server implementation is lacking, mainly
17:43:06 *** SmatZ41 has joined #openttd
17:47:25 <Celestar|GSC> mucht_work: Mucht: you there?
17:48:12 <Celestar|GSC> Rubidium: Eddi|zuHause basically I can imagine all the houses in town adding their passengers to a "passenger pool" thingy
17:48:30 <Celestar|GSC> Mucht: do you anything about the subfig package?
17:48:54 <Eddi|zuHause> Celestar|GSC: didn't the old patch do something like this?
17:49:03 <Eddi|zuHause> send back passengers where they came from?
17:49:34 <Celestar|GSC> Eddi|zuHause: I can implement something like that without too much hassle, the question is: do we WANT that?
17:49:40 <Celestar|GSC> I mean in the first iteration
17:52:22 <peter1138> What's the current problem?
17:53:24 <Celestar|GSC> 19:42 < Celestar|GSC> simple example: you have a network with only two stations: A) which has about 10000 inhabitants in its CA, B) which has 50 inhabitats in its CA. you have a single train running from A to B. we'd need a system where the nofpax generated at A is about the same as the nofpax generated at B
17:53:42 <Celestar|GSC> peter1138: so the question is: rewrite the pax generation system for cargodest now, or think about doing it later on?
17:53:57 <peter1138> Why do we need that?
17:54:08 <peter1138> Is symmetry needed?
17:54:34 <Celestar|GSC> it's not "MUST", neither "SHOULD", but it's on my "NICE TO HAVE" list
17:55:15 <Eddi|zuHause> yes, i think passengers should be (roughly) symmetrical
17:55:29 <Eddi|zuHause> no, i don't think it should be done in the first iteration
17:56:05 <Celestar|GSC> or rather: have it before the trunk merge or after the trunk merge
17:56:56 <Celestar|GSC> peter1138: should have some something like size-based passengers generation. it's mostly finished, I just need some fine-tuning
17:57:37 <Celestar|GSC> er ... that was a question, sorry :P
18:08:50 <blathijs> How far can I mis the "C" key :-)
18:09:02 <blathijs> approximately 4cm, it seems
18:09:12 <Eddi|zuHause> "all the way from a to z" :p
18:09:42 <Celestar|GSC> peter1138: so I'll add it once it's finished and tested
18:09:45 <Celestar|GSC> peter1138: we're almost done :D
18:10:05 <Celestar|GSC> peter1138: mainly improvedloading and then nettesting
18:10:08 <peter1138> You want me to add another minimap view?
18:10:29 <Celestar|GSC> there was talk about coloring the bars by vehicle type, wasn't there?
18:10:32 <Celestar|GSC> dunno if you wanna do it
18:10:41 * Celestar|GSC doesn't really care about that feature, it's just been suggested
18:11:21 <CIA-5> OpenTTD: peter1138 * r14014 /trunk/src/widgets/ (dropdown.cpp dropdown_type.h): -Codechange: Add support for automatically sizing drop down lists to the widest list item.
18:13:54 <Celestar|GSC> peter1138: heh. color by cargotype when > 1 types are displayed, color by vehcile type if only one cargotype is selected :P
18:14:09 <Celestar|GSC> (that was a joke)
18:14:37 <Celestar|GSC> BAH why do I get malformed packets from this stupid DNS server
18:26:28 <CIA-5> OpenTTD: peter1138 * r14015 /trunk/src/widgets/ (dropdown.cpp dropdown_type.h): -Codechange: Add facility for a drop down list to always close (if requested) when the mouse button is released.
18:29:13 * Belugas salutes the work of mister Nelson. Quite nice :)
18:33:04 *** SmatZ41 has joined #openttd
18:33:30 <hylje> Celestar|GSC: also fatten bars depending on volume
18:36:16 *** SmatZ42 has joined #openttd
19:05:44 *** SmatZ41 has joined #openttd
19:12:01 *** SmatZ42 has joined #openttd
19:13:55 *** blathijs has joined #openttd
19:21:12 <peter1138> fonso, inline is fine for performance critical stuff.
19:31:07 <fonso> Is it terribly important to have exactly four big explosions when you clear an area? I'd generally reduce them to two (in the corners you actually select).
19:34:06 *** DJNekkid has joined #openttd
19:35:42 <DJNekkid> a train with action0 prop 1A \b*193 should be underneeth one with 1A \b*187 ... shoult it not?
19:36:10 *** welshdragon has joined #openttd
19:39:34 <frosch123> DJNekkid: IIRC you specify a vehicle ID in that property and not a key value for sorting. I.e. setting the property to 193 means to sort this vehicle in front of the vehicle with ID 193
19:53:55 *** planetmaker_ has joined #openttd
19:55:51 <peter1138> It's a bloody silly way of sorting but that's the way it is :(
20:07:54 <DJNekkid> frosch123: but when all vehicles have an 1A
20:08:04 <DJNekkid> in a range from 0 to 240
20:12:19 <frosch123> does that change anything?
20:13:11 <DJNekkid> the order gets scrambled
20:13:30 <DJNekkid> however, not the way i want it to ;)
20:13:33 <frosch123> note, 1a is called an 'action' not a property. i.e. it is executed in the order it is defined
20:14:20 <frosch123> if you refer to a vehicle id that does not exist, it is ignored
20:14:46 <DJNekkid> i think u actually might be on to something there ...
20:14:53 <DJNekkid> i think i got an idea
20:15:20 <DJNekkid> because, when i think of it, it seems like the ones with the value < 116 seem to worke as intended
20:20:53 <DJNekkid> it seems like it actually worked frosch123
20:21:12 <DJNekkid> 7 of them got at the bottom, but that is prolly my fault ;)
20:21:59 <DJNekkid> i just had to define them with an action0 06 00
20:22:05 <DJNekkid> at the start of the file
20:22:12 <DJNekkid> at vehicle ID 0 -> 240
20:22:15 <frosch123> hmm, last time Belugas gave me a nice idea what to listen to, but this time...
20:23:15 <Belugas> can't always please you :D
20:23:39 *** birdspider has joined #openttd
20:24:14 <Belugas> but if you want to be on the safe side, i'll recommand a high does of Mogwai ;)
20:26:22 <frosch123> hehe, 'safe side', actually they appeared quite boring to me :p
20:26:47 <frosch123> sorry, I know you like them a lot
20:27:16 *** [com]buster has joined #openttd
20:29:37 <Belugas> i'd say that they are a bit special, granted
20:29:45 <Belugas> mostly instrumentals,
20:30:16 <Belugas> and whenever they "sing" , it's as if they are just mumbling
20:30:26 <Belugas> the energy they put on :)
20:30:46 <Belugas> but granted, it's not something easily reachable
20:31:14 <Belugas> i just can't wait to see them live :)
20:31:36 <frosch123> sounds like you need a baby sitter
20:31:41 <Yorick> hm, I can't be away for 8 days or someone commits YAPP :)
20:32:08 <Belugas> ho no... my wife hates been in crowded places, even more in rock shows ;)
20:32:20 <Belugas> so i'll be attempting the show alone
20:32:29 <Belugas> does not bother me :)
20:32:59 <Yorick> and flyspray got updated, trac got added
20:33:10 <Belugas> last time, i went to see Tea Party, all alone, and i had quite a good time :)
20:33:31 <Belugas> just watching the band, seeing how they did the songs, etc...
20:53:31 *** blackjak has joined #openttd
20:58:20 *** Yorick_ has joined #openttd
20:58:22 *** Yorick_ is now known as Yorick
21:00:08 <peter1138> Shit, Yorick is back, we'll have to remove YAPP.
21:00:33 * Prof_Frink is watching the openttd codeswarm
21:00:52 <Bjarni> what is it with Yorick and YAPP?
21:02:11 <Prof_Frink> Crikey, what happened January 2007?
21:03:50 <rortom> nice link Prof_Frink, thx for sharing :)
21:04:36 <peter1138> Hey rortom, any news on a 64bit Linux version?
21:04:38 *** Brianetta has joined #openttd
21:04:47 <rortom> peter1138: yes, its building :D
21:05:51 <rortom> mhm whats the generic platform compiler?
21:06:07 <Belugas> no no no, he's not happy, it's just the smilies on his jammy
21:06:44 <rortom> should i build optimized versions for x64?
21:06:56 <peter1138> Assuming you mean architecture, not compiler.
21:07:33 <peter1138> Well, when you're building x86_64 binaries you might as well optimize them for it too ;)
21:07:47 <peter1138> It's not like you lose any backward compatibility.
21:08:16 <peter1138> Let me know when you need me to test it ;)
21:08:36 <rortom> should be in about a month or less i hope
21:08:58 <peter1138> Wow, that's a very slow compiler...
21:09:14 <rortom> the features are not fully ready
21:09:45 <Rubidium> Prof_Frink: makefile rewrite and C->C++
21:10:28 <Belugas> fonso, will try to check ... no promises
21:10:44 <Prof_Frink> Rubidium: Thought that was probably it. Big explosion in the swarm.
21:11:26 <fonso> thanks. I couldn't find any more problems, but I'll happily correct things that you might find.
21:13:06 <Belugas> good night in the mean time
21:13:41 <peter1138> rortom, sorry, when you said "building" I assumed you actually meant the compiler was running...
21:13:51 <peter1138> I don't quite understand the problem in building it :o
21:14:06 *** welshdragon is now known as Guest717
21:20:57 *** SmatZ41 has joined #openttd
21:21:59 <Rubidium> Prof_Frink: can you make a movie out of the codeswarm thing?
21:23:54 <Prof_Frink> Rubidium: You can make it output each frame as an image, then moviefy them.
21:24:28 <Rubidium> I dislike to install half the world just to compile codeswarm
21:36:09 *** welshdragon has joined #openttd
22:00:00 *** welshdragon has joined #openttd
22:09:24 <Bjarni> Prof_Frink: how did you get codeswarm?
22:09:43 <Bjarni> this might be a silly question but it keeps on stalling when I try to checkout
22:09:49 <rortom> svn by google code site
22:09:52 <Bjarni> leaving me with a broken checkout
22:10:02 <peter1138> Brianetta, pb_hovs_bus and grvts?
22:10:02 <rortom> oh, just got the checkout without problems :/
22:10:30 <peter1138> You can do that in trunk, but not 0.6.2 :)
22:10:31 *** Digitalfox has joined #openttd
22:12:10 <Bjarni> ok google's svn officially hates me
22:12:20 <Bjarni> it gives me 20-30 files and then it disconnects
22:12:26 <Bjarni> so I have to start over
22:12:39 <Bjarni> it just did that 8 times in a row
22:16:24 <Brianetta> I can;t get the bugger not to load hovs )-:
22:16:54 <rortom> Bjarni: i can upload the files for you if you want
22:17:01 <Bjarni> I mean I just tried to ssh to a remote computer to check out and it worked right away
22:21:08 <Bjarni> but I didn't change anything
22:21:42 <Prof_Frink> Bjarni: openttd's a fairly boring swarm.
22:23:09 <rortom> quite some people working together
22:25:14 <Prof_Frink> rortom: That's the last month of KDE /trunk.
22:25:52 <Bjarni> how do I make a config file of the trunk?
22:26:10 <Bjarni> or any config file for that matter
22:26:18 <Brianetta> It was a bug on my site
22:26:33 <Bjarni> glx: I meant a config file for codeswarm ;)
22:26:59 <Prof_Frink> Bjarni: If you just want white dots, then just `ant run` and fill in the svn address and username
22:27:31 <Prof_Frink> Else, edit data/sample.config if you want multicoloured dottery
22:29:23 <Bjarni> null pointer exception
22:30:28 <Bjarni> clearing the cache and trying again seem sot do the trick
22:39:53 <Bjarni> everybody shows up and fades away
22:40:21 <Bjarni> looks like I'm the only one to be there from the start and the end and never fades away even once
22:40:40 <Prof_Frink> Reliable Bjarni is reliable.
22:40:48 <Zuu> So you are saying you will stay online and talkning for ever? ;)
22:41:00 <Prof_Frink> Zuu: Not just talking, coding.
22:41:35 <Bjarni> and not just coding, I will also commit
22:45:25 <Zuu> I hope to commit a snore soon :)
22:46:37 <Zuu> What you in a commic would write Zzzz
22:47:07 <Zuu> My online dictionary says it's called to snore.
22:49:42 <Bjarni> btw it's spelled "Zzz"
22:54:18 <Zuu> Depends, sure three Z are most common but could be even more.
22:55:14 <Prof_Frink> Depends how loud you snore.
22:58:35 <Bjarni> I know because I never heard it
22:58:49 <Bjarni> and I would be present to hear it if I did ;)
23:01:11 <Zuu> So with that logic, The world is beeping because I hear it right in my head. :)
23:08:04 <Bjarni> now spammers wants to sell me sex improving pills (not viagra... something else)
23:08:18 <Bjarni> it's not like it's new, but this time they are doing it in German
23:09:38 <Bjarni> why would anybody want to tell me anything in German???
23:09:46 <Bjarni> (besides this channel, that is)
23:17:32 *** lobster_MB has joined #openttd
23:19:27 *** welshdragon has joined #openttd
23:48:12 *** Dred_furst has joined #openttd
continue to next day ⏵
