IRC logs for #openttd on OFTC at 2013-02-05
⏴ go to previous day
00:01:58 <Zuu> My main point against that is that you must encode the town id in a such way that ICU cannot think that your town ID is a text character that it may want to replace with a better looking character.
00:02:48 <Eddi|zuHause> no, i mean on the internal side, not what you print on the screen
00:04:31 <Zuu> hmm, yes you could make DrawString strip out that data from the string before calling ICU.
00:05:59 <Eddi|zuHause> then you can add sugar to the chat where it automatically uses the translated name of stations and stuff
00:07:42 <Eddi|zuHause> (and add the vehicle icons to indicate it's a station)
00:09:15 <Eddi|zuHause> (still needs a method to determine whether the user meant the town, or the station with the town's name)
00:10:15 <Eddi|zuHause> Zuu: have you thought about applying this method to the subsidy news?
00:11:04 <Zuu> Nope, but it can probably applied there too once the framework is there.
00:45:46 <peter1138> most of the time nobody types the name in properly anyway
00:46:34 <Eddi|zuHause> they might start to use tab-completion better if they see a useful effect :)
00:50:07 <Eddi|zuHause> maybe instead of tab-completition it should auto-suggest the completition :)
01:11:02 <Bonez305> how do I go about setting up a server on a VPS ?
01:46:14 * Supercheese wonders whether he should "cheat" by distant-joining stations rather than trying to build a new junction
01:47:02 <Supercheese> I like having stations visually close to the cargo source, but sometimes that's just not practical
01:54:01 *** roboboy has joined #openttd
01:54:42 <Eddi|zuHause> make a feeder service
03:10:56 *** Biolunar_ has joined #openttd
04:31:34 *** roboboy has joined #openttd
05:37:45 *** TrueBrain_ has joined #openttd
05:56:16 *** Eddi|zuHause has joined #openttd
06:55:30 <peter1138> oh, i see, building infrastructure is micromanagement...
07:17:26 *** valhallasw has joined #openttd
07:36:17 *** Devroush has joined #openttd
08:07:10 *** ComLock has joined #openttd
08:25:43 *** valhallasw has joined #openttd
08:35:13 *** roboboy has joined #openttd
08:41:04 <Supercheese> Blaaaaah, sandwiched in between two hills, no room to build a ro-ro station
08:43:22 *** valhallasw has joined #openttd
08:43:30 <MNIM> Room can be a terrible thing, eh?
08:44:19 *** andythenorth has joined #openttd
08:48:31 <Flygon> Supercheese: Oh goooooooooooood, I hate that
08:48:36 <Flygon> ESPECIALLY when working with towns
08:48:49 <Flygon> I've basically had to build what I consider narrow gauge branchlines...
08:48:59 <Flygon> Pass stations the size of 4 instead of 7
08:49:20 <Supercheese> I ended up bulldozing to BEHIND the steel mill and putting a ro-ro station a dozen tiles farther back
08:50:22 <Supercheese> I'm going to need a lot of throughput, considering these weak 1890s-era locomotives
08:51:57 <Flygon> Sextuplehead if it's 1845
08:52:12 <Supercheese> Spoilers: Invisible Power Boosting Engine
08:52:24 <Supercheese> although I only use that to climb this one massive hill
08:52:41 <Bonez305> This game will take a bit of getting used to, i suck :/
08:52:50 <Flygon> I just make longer stations, Supercheese
08:53:13 <Flygon> More locomotion, more station, faster loading than having half-a-locomotive peeking out >_>"
08:53:32 <Flygon> The game REALLY needs curved stations
08:53:38 <Supercheese> Flygon: I need to drag 150 tons of food from tileheight 2 to tileheight 6 in a distance of 25 tiles
08:54:02 <Supercheese> with 1890s engines, quadruple-heading might not even work :P
08:54:12 <Flygon> Supercheese: 6 Consolidation Locomotives
08:54:38 <Supercheese> NARS doesn't have autorefit, which is mandatory
08:54:49 <Flygon> I forgot the high tractive effort one
08:55:02 <Supercheese> Yeah, I don't have that engine yet
08:55:14 <Supercheese> the Garatt, whenever that comes, could probably do it singlehanded
08:55:20 <Supercheese> Garratt*, or however
08:55:41 <Supercheese> Yeah, my corrected version is right
08:55:41 <Flygon> Is that the 0-6-6-0 one?
08:57:00 <Supercheese> (at least I think that's the UKRS oen)
08:57:04 <Flygon> Lemme boot up my UKRS gam
08:57:09 <andythenorth> Supercheese: ships
08:57:34 <Supercheese> Might as well just make a waterfall
08:57:42 <Supercheese> run the ships straight up it
08:57:50 <Supercheese> they can do that, right? :P
08:58:26 <peter1138> 08:41 < Supercheese> Blaaaaah, sandwiched in between two hills, no room to build a ro-ro station
08:58:43 <peter1138> should've just copy & pasted a layout all over it
08:59:04 <Flygon> Supercheese: Ah. The Garrett got overlooked by the 177km/h locos x.x
08:59:55 <Supercheese> Oh, the 0-10-0 big bertha might do the trick too, 1,200 hp and almost 200kN tractive effort
09:00:00 <Flygon> Because nothing is more useful than Triple-heading the Mallard in 1938 just for that tractive 185km/h goodness
09:00:02 <Supercheese> comes a bit earlier in 1919 too
09:00:21 <peter1138> freight multiplier 1
09:00:36 <peter1138> no need to worry then :p
09:00:36 <Supercheese> peter1138: cheating! oh wait-
09:01:42 <Bonez305> are 128x256 games popular ?
09:02:07 <peter1138> i like them but that doesn't count
09:02:17 <Supercheese> That's right about minimum size IMO
09:02:26 <Supercheese> I'm currently on a 256x256 game
09:02:34 <Supercheese> very nice, I just hooked up all industries
09:02:47 <peter1138> i like 128x128 with 3 or 4 towns
09:02:55 <peter1138> 64x is pushing it though
09:04:03 <peter1138> A fatal NewGRF error has occurred:
09:04:09 <peter1138> Attempt to use invalid ID (sprite 222)
09:04:19 <peter1138> i guess too-many-railtypes isn't handled :P
09:04:28 <Flygon> I'm using a 2048*2048 game :D
09:04:38 <Supercheese> Flygon: way too big :S
09:04:51 <Flygon> I should have gotten a daylength patch installed too... I barely covered NE-USA by the time it hit 1985
09:04:57 <Supercheese> 1024x512 is excessive for singleplayer, IMHO
09:05:25 <Bonez305> are 128x256 games good for 4 players ?
09:05:26 <Supercheese> for multiplayer, perhaps larger maps are workable
09:05:42 <Flygon> I WISH this map worked in MP
09:05:52 <Supercheese> Bonez305: try 256x256 first, as peter1138 mentioned, that's pretty much the "standard" map size
09:06:03 <Flygon> And nobody in this house plays OpenTTD
09:06:11 <Flygon> 256*256 is a fantastic starting size, yeah
09:06:50 <Bonez305> sorry i meant 256 256
09:08:01 <Bonez305> think about running a server on my VPS, so basically just drop the folder in a "openttd -D ?
09:11:52 <Flygon> Supercheese: I suspect I should get 100/40 to run 2048*2048 MP. Which I'd like. This NA map is FANTASTIC.
09:16:30 <Flygon> Only 24/1 atm, makes MP impossible :(
09:20:58 <Supercheese> I see, yeah, I hear aussie net pricing is blarg :(
09:21:13 * Supercheese resorts to more doubleheading
09:25:06 *** oskari89 has joined #openttd
09:44:09 <Flygon> Supercheese: Pricing isn't the problem
09:44:46 <Flygon> Can UKRS2 trains go over 225km/h? With the High Speed Carriages?
09:48:12 <peter1138> there's a hitachi that goes ~250 too
09:48:30 <Flygon> peter1138, Hitachi trains have only been driven up to 130km/h here...
09:48:57 <Flygon> It'd have been cruised faster, but the curves here don't allow that :)
09:49:33 <Flygon> Yeah, I know the actual Hitachi train in the UK :3
09:49:41 <Flygon> Hitachi's here are becoming rare
09:49:51 <Flygon> It's become a gag that peeps report Hitachi sightings... from Britian
10:00:08 *** valhalla1w has joined #openttd
10:01:12 <peter1138> so i ended up with...
10:01:36 <peter1138> if (r <= .125f) { t = .5 - tanf((.125f - r) * 2 * M_PI) * .5f; } else { t = .5 + tanf((r - .125f) * 2 * M_PI) * .5f; }
10:02:33 <peter1138> (r is 0..1, not radians or degrees)
10:06:54 *** kais58_ has joined #openttd
10:08:23 *** tokai|mdlx has joined #openttd
10:23:03 *** valhallasw has joined #openttd
10:32:35 <peter1138> "It looks like it has the same issue tbh... I still have all the source files for OpenGFX so it should be fairly easy to tweak though.
10:32:40 <peter1138> are they not in vcs?
10:34:23 *** roboboy has joined #openttd
11:03:10 <jonty-comp> bollocks to chicago, coventry is the windy city
11:05:13 <jonty-comp> [3~[3~[3~[3~[3~[3~[3~[3~[3~[3~a
11:05:43 <jonty-comp> [3~[3~[3~[3~[3~[3~[3~[3~[3~[3~a
11:07:40 <SpComb> it's a modified version of brainfuck called perl
11:07:54 <andythenorth> the most human language
11:16:51 *** andythenorth has joined #openttd
11:23:39 <peter1138> Pikka, Pikka Pikka Pikka Pikka Pikka Pikka Pikka Pikka Pikka Pikka Pikka Pikka Pikka Pikka
11:25:09 <Pikka> peter1138 is the hairy lime type man
11:25:37 <oskari89> Pikka: How many OpenTTD-related projects do you have on progress?
11:27:04 <oskari89> Okay, i was thinking that Finnish Trainset could utilise a coder now since Lakie hasn't got time at all for coding...
11:27:34 <oskari89> And that won't possibly change on near future
11:27:55 <oskari89> If you are intrested, pm DanMacK
11:28:33 <Pikka> danmack's been very quiet lately
11:28:49 <oskari89> Yes, he has got some other projects going
11:29:40 <oskari89> There's nearly all specs on excel file
11:30:34 <Pikka> specs in excel is the fun and easy bit
11:30:52 <oskari89> And about 60 % of graphics ready for coding :P
11:31:38 <oskari89> All related files here
11:32:11 <oskari89> Nicely on proper folders
11:32:17 <oskari89> And makefile there too
11:38:54 <planetmaker> we should write OpenTTD in assembler ;-)
11:38:57 <Pikka> but then I guess we all just like to keep doing things the way we've been doing them
11:39:29 <Pikka> OpenTTD is vastly more complex than even the most complex grf, planetmaker
11:40:27 <Pikka> proper programming languages make sense for OpenTTD. but most grfs are so simple that nml seems massively overkill. But if it's what people feel comfortable with, then I'm not one to judge. :)
11:41:02 <planetmaker> I wouldn't write a "hello world" programme in assembler either ;-)
11:41:14 * peter1138 ponders committing some random stuff without review
11:42:22 <V453000> I need a minion to code signals to be aware what kind of track they are on :(
11:43:04 <peter1138> ttp://git.fuzzle.org/openttd.git/commit/?h=railtype_nocombine
11:43:20 <V453000> I would like to make tracks which have signals directly on them, similar to current road signs
11:44:04 <peter1138> (ships stopping instantly from 150 knots is... silly, mind you)
11:44:50 <peter1138> and then you need an advanced setting to disable and reduce ship crashes
11:44:55 <Pikka> I've been doing a fair bit of research into late 19th century shipping, for non-openttd reasons lately
11:44:57 <V453000> trains stopping instantly from 600kmh ;)
11:45:09 <Pikka> and it seems for merchant ships of that period, sinking within 10 years of being launched was the norm
11:45:32 <peter1138> FISH ships don't get maintained
11:45:43 <Pikka> they have 0 reliability decay
11:45:43 <peter1138> their reliability is unmatched
11:46:38 <oskari89> They have no planned obsolescense
11:47:25 <Flygon> Pikka: Let's design Tiremes and push them more than 1 tile from the coast
11:47:29 <peter1138> (it's silly how much effort i put in to rgb recolours which is undoubtedly never going to hit trunk)
11:47:35 <planetmaker> FS5454 could use quite a bit more doxygen at least
11:47:35 <Flygon> And give it a 50% chance of sinking :D
11:48:21 <peter1138> planetmaker, that one is a crude hack to test the concept, defintely not ready
11:48:46 <Pikka> let's just bung it all in and call it 2.0 \o/
11:50:30 <peter1138> cos it doesn't need to be 32 bits. rubidium suggested it, heh
11:50:57 <peter1138> all but the last commit is sensible at least though, imho
11:54:32 <planetmaker> you want railtypes to provide combined ground+track sprites?
11:55:19 <peter1138> some people have complained that the sprite combining method doesn't allow them to draw the junctions as they want them
11:55:32 <peter1138> that should allow it without introducing more variables
11:55:46 <planetmaker> in what way exactly?
11:55:49 <peter1138> although i haven't tested it cos nobody drew anything yet
11:56:14 <peter1138> sprite combinding is a bit awkard for 3rd rail/metro style stuff, i believe
11:56:52 <planetmaker> yes, ok. But what sprites are provided by this... it talks awefully lot about ground sprites. Which... would be bad, if railtypes need to provide ground sprites
11:57:02 <peter1138> no, it's the overlay
11:57:14 <planetmaker> they should provide underlay. Which in turn is drawn on the ground. which is drawn below the actual tracks
11:58:07 <planetmaker> ok, so new overlay sprites for all junctions
11:58:31 <peter1138> for simplicity it's for all flat layouts, except bare (hence the -1)
11:59:06 <peter1138> i don't really know if it's a good idea or not, i just coded it :p
11:59:36 <peter1138> but i figured that would be less cpu-intensive than adding a variable for newgrfs to check and decide what to draw
11:59:42 <peter1138> (which is what pikka asked for a while back)
12:00:01 <Pikka> but I'm not really bothered :)
12:00:09 <V453000> I think it could be pretty good for some situations
12:00:27 <peter1138> i think Eddi|zuHause was quite vocal but as he doesn't make grfs (afaik) he doesn't matter ;)
12:01:13 <planetmaker> I think on the account that he doesn't make NewGRFs, but well :-)
12:02:43 <Eddi|zuHause> i don't make grfs, i make scripts that make grfs :)
12:03:04 <peter1138> as an advantage junction tiles would actually be less cpu-intensive to draw than now ;)
12:03:28 <planetmaker> peter1138, the order of sprites would be as determined by the track bits value of the tile? Or how?
12:03:50 <peter1138> yeah, it's just track bits - 1
12:04:14 <planetmaker> might make sense. Would also allow curvy tracks
12:04:29 <peter1138> and the original split up parts are still needed for half-tile corners and pbs reservations
12:04:38 <planetmaker> though that might better require reading adjacent track bits
12:04:41 <peter1138> and any non-flat tile
12:04:56 <peter1138> yeah, curvy tracks needs adjacent info
12:05:09 <planetmaker> ui... looks like world ends here now ;-) What a downpour
12:05:12 <peter1138> plus a state machine to actually work :p
12:05:25 <peter1138> and more vehicle angles
12:05:29 <peter1138> meh, not openttd any more ;)
12:05:41 <planetmaker> more vehicle angles is already feasible
12:06:23 <peter1138> i mean without cets-like hacks
12:06:43 <peter1138> but our vehicle movement code is not up to the task
12:10:11 <Pikka> I'll buy more vehicle angles once the size of the vehicles is the same in different directions ;)
12:10:54 <peter1138> it needs sub-pixel positioning to work right
12:11:38 <peter1138> well, sub-unit, they're not pixels
12:12:53 <Eddi|zuHause> yes, the "units" in horizontal/vertical movement need to change from 4px to 3px (roughly)
12:13:30 <Eddi|zuHause> or we need to switch to hexagonal raster :)
12:13:35 <peter1138> much finer than that please
12:15:31 <Pikka> he means for getting the vehicles the same size (approximately) :)
12:16:11 <peter1138> Eddi|zuHause, yeah, more hacks though :-(
12:16:33 *** HerzogDeXtEr has joined #openttd
12:17:10 <Bad_Brett> Are you talking about me? :D
12:17:13 <peter1138> needs maybe an extra byte, and treat it as a fixed-point decimal
12:17:34 <Eddi|zuHause> peter1138: we have that extra byte already, it's called "progress"
12:18:07 <peter1138> Eddi|zuHause, it's reset in a few places
12:18:08 <Eddi|zuHause> but it applys to the whole consist
12:18:20 <peter1138> it appies to the front
12:18:28 <peter1138> cos it's not meant for positioning as such
12:18:39 <peter1138> and also it's depending on direction
12:19:33 <peter1138> Eddi|zuHause, so i want independent x/y subposition, along with finer rotation!
12:19:47 <Eddi|zuHause> CODE IT THEN! :)
12:20:09 *** andythenorth has joined #openttd
12:20:18 <peter1138> didn't we have fixed-point types somewhere?
12:20:25 <Eddi|zuHause> instead of OpenTTD 2.0 we then call it OpenLomo 1.0 :)
12:20:27 <peter1138> istr celestar being involved somewhere
12:20:50 <peter1138> Eddi|zuHause, oh and smooth gradients!
12:20:56 <Eddi|zuHause> peter1138: that was the economy rewrite thingie, i think it wasn't included
12:21:21 <peter1138> "trucks" i think USians would call them
12:22:27 <Eddi|zuHause> i need to go now :/
12:27:28 <peter1138> "But when there is long, wide lines with heavy traffic, many many track-changing crossings on line and not very much obstacles around, copy-paste is a must."
12:27:37 <peter1138> i honestly nave never ever ever felt the need for C&P
12:31:09 <Pikka> every junction needs to be a huge cloverleaf you've copied off the internet
12:31:16 <Pikka> don't you know how to play openttd?
12:31:54 *** Supercheese has joined #openttd
12:35:41 <Flygon> THAT'S A COMENG DESIGN D:
12:35:48 <Flygon> And Comeng made shit that just worked
12:38:17 <peter1138> Flygon, unrealistic
12:38:25 <peter1138> the carriages are too long!
12:39:07 <Flygon> We have HST's in Australia, silly
12:39:21 <Flygon> ...it's just that they never reach their design speed due to track design >_>
12:39:33 <peter1138> they're at least a full tile long, maybe 1.5
12:40:04 <V453000> anyway, I have this https://dl.dropbox.com/u/20419525/signals_concept.png as signals on my new rails ... problem is, on slopes I need different sprites and on diagonals the positions differ between diagonals of the same direction. Would anyone be interested in adding awareness for signals which exact track they are on?
12:40:20 <Flygon> Are you familiar with the trains used under the Intercity 125 brand?
12:40:21 <V453000> very conceptual but should illustrate the idea
12:40:30 <Flygon> The Mk. 3 carriages that BR used?
12:40:37 <Flygon> Each DMU car is around the same length
12:40:47 <peter1138> that's a signal? :S
12:41:59 <V453000> well it can look differently, but it indeed can work as a signal
12:42:08 <V453000> in case of presignals it should be a lot more obvious what works how to new people
12:42:10 <peter1138> flygon is missing the point :-(
12:42:21 <Flygon> Also, peter1138, if you're at any point abusing my inability to tell a troll from a user, let me know please
12:42:30 <peter1138> 12:39 <@peter1138> they're at least a full tile long, maybe 1.5
12:42:38 <peter1138> plus being in #openttd
12:42:38 *** andythenorth has left #openttd
12:42:49 <peter1138> i think it's pretty obvious i'm comparing it with ttd graphics
12:42:50 <V453000> the same thing can hang on a big board above rails, but it still would make use of the new specs
12:43:16 <Flygon> British Sprinter vs Victorian Sprinter
12:43:22 <Flygon> Victorian Sprinter plzkthx
12:44:04 <Flygon> Though, arguably, the Victorian Sprinters are actually remotely derived from the VR DRC cars which are clearly inspired by American Budd Diesel Railcars
12:44:52 <Flygon> peter1138: And I can reassure you, Sprinters fit inside 0.5 tiles
12:45:27 <peter1138> not when they're drawn like that
12:46:24 <peter1138> of course, the best trains are the pacers
12:49:08 <Flygon> Also, just to confuse you
12:49:45 <Flygon> peter, that must be unsafe
12:49:49 <Flygon> Anyway, gotta brb 30 mins
12:50:13 <peter1138> yeah the door's on the wrong side ;p
12:51:31 <peter1138> It is powered by a 200hp 6 cylinder Leyland TR11 engine driving through a fully automatic gearbox. A flexicoil suspension system supposedly allowed the unit to be used at speeds up to 100mph!
13:01:31 <Pikka> it might get that fast if you drop it off a cliff, peter1138
13:06:03 *** oskari89 has joined #openttd
13:08:27 <Bonez305> is there a factor that makes loading and unloading go by faster ?
13:08:46 <Eddi|zuHause> Bonez305: vehicles can define how much is loaded in each step
13:09:13 <Bonez305> for instance trains.
13:09:28 <Eddi|zuHause> Bonez305: it means you can change it by NewGRFs
13:10:30 <Bonez305> Eddi|zuHause: I dont get it, mind you I'm new to this game
13:11:39 <V453000> each vehicle has some speed in which it loads, defined by its code
13:12:07 <Bonez305> Thats what I wanted to know. :D
13:14:14 <Eddi|zuHause> V453000: so you're calling that "NUSS" then? [that's german for NUTS :)]
13:14:43 <Eddi|zuHause> (or actually NUT, but whatever :))
13:15:07 <Eddi|zuHause> V453000: i mean the "NUTS Unrealistic Signal Set"
13:15:09 <V453000> or signals or whatever
13:15:25 <Eddi|zuHause> "NUSS Unrealistic Signal Set"
13:15:26 <V453000> didnt think about a name yet really
13:16:09 <Eddi|zuHause> you gotta have a backronym before you start the project!
13:16:30 <V453000> I will need the specs to allow it first though :( but I might try to make the best ouf of the current situation too
13:16:33 <Eddi|zuHause> its the mostest imprtantestest thing in the world!
13:17:02 <Eddi|zuHause> (missing apostrophe is for stylistic completeness and thus intentional)
13:17:03 <V453000> well, I am trying it inside NUTS itself :P
13:17:10 <V453000> then it could eventually be released as separate track set
13:17:46 <Eddi|zuHause> V453000: could make it 32bpp as a semitransparent track overlay :)
13:18:05 <V453000> why semitransparent :)
13:18:22 <Eddi|zuHause> because of the tracks (and vehicles) underneath :)
13:18:51 <V453000> vehicles are drawn under signals
13:18:58 <Flygon> Eddi: Lemme guess, build up speed, drop panto, coast, raise panto?
13:19:22 <V453000> well that screws things up I guess
13:19:34 <Flygon> Or do they share voltage?
13:19:39 <Eddi|zuHause> Flygon: yes, that's what the blue sign says. although technically the crossing is refitted to switch voltage
13:19:55 <Flygon> Ah, crossings here aren't refitted
13:19:55 <Eddi|zuHause> only the railway chose not to use that
13:20:00 <Flygon> Too dangerous in peak hour
13:20:33 <V453000> hm I guess having signals to get a spec whether they are drawn after or before a train is not the way to go?
13:20:35 <Flygon> Blowing up a tram with accidental 1500V DC or breaking down a train (and substation) with 600V DC is nobody's idea of fun
13:21:02 <Eddi|zuHause> V453000: it would need specifying the signal's bounding box
13:21:05 <Flygon> The best solution is to just avoid crossings in the first place :)
13:35:50 <Eddi|zuHause> <insert obligatory russian reference here>
13:38:12 <Flygon> Stupid Serial Escalation :U
13:47:57 *** TrueBrain_ is now known as TrueBrain
14:07:44 *** ntoskrnl has joined #openttd
14:56:53 *** ComLock has joined #openttd
15:07:01 *** goodger has joined #openttd
15:39:59 <mouse> hi.. my computer asks for 3rd party files?? any help??
15:40:07 <peter1138> well that's unfortunate
15:40:27 <Rubidium> mouse: what is the exact message?
15:41:17 <mouse> Failed to find a graphics set. Please acquire a graphics set for OpenTTD. See section 4.1 of readme.txt.
15:41:32 <mouse> and 4.1 says something about 3rd party files
15:41:42 <Rubidium> so... you're using Mac OS X?
15:42:17 <Rubidium> mouse: have you read 4.1.1?
15:43:03 <mouse> what am i supposed to do with the file i just downloaded??
15:43:52 <mouse> 4.1) (Required) 3rd party files
15:44:15 <Pikka> put the game's data files into a baseset/
15:45:06 <mouse> i´m sorry, but i am really a dummy.. How do i make a baseset??
15:45:43 <Pikka> I am not an os x user, but basically, find where it installed openttd
15:45:52 <Pikka> there will be a directory there called baseset
15:46:57 <mouse> something is called: ogfx1_base.grf
15:47:25 <planetmaker> a base set consists of several files
15:47:50 <planetmaker> the one you quote is one of 7 you need. But you don't need to unpack it, if you got it as a *.tar file
15:48:16 <mouse> the file is called: ogfx1_base.grf
15:52:58 <mouse> now i´ve found the baseset, but nothing is in it?????
15:55:15 <planetmaker> Exactly that's your problem. The dir is empty. And you need to unzip the file from the download location I linked you into that very dir
15:57:11 <mouse> IT WOORKS!! Thank you so much:)
15:58:28 <peter1138> wow, it's all documented and everything
16:09:27 <peter1138> which one is the parameter, 00?
16:09:52 <peter1138> where is it, callback or graphics chain?
16:10:33 <Pikka> industries don't have a graphics chain, but it's outside any specified callback
16:10:57 <peter1138> outside, so always seen?
16:11:19 <peter1138> some of the callbacks are called with no tile
16:12:18 *** TheMask96 has joined #openttd
16:12:51 <peter1138> not sure if that's true :p
16:13:18 <peter1138> industry_available callack but then industry is null too
16:17:06 <Pikka> actually, it's after a callback check, so it's not seen by the availability or location callbacks anyway.
16:17:41 <peter1138> CBID_INDUSTRY_DECIDE_COLOUR, CBID_INDUSTRY_INPUT_CARGO_TYPES, CBID_INDUSTRY_OUTPUT_CARGO_TYPES are all called without tile
16:36:00 <Pikka> if I put in another check so it skips that part of the grf if it's the first of the month, it doesn't crash
16:36:20 <Pikka> it's definitely a problem with trying to check var 61 before the industry is properly initialised
16:38:09 <Pikka> and if I then try and prospect a coal mine on a date other than the 1st, it crashes :D
16:43:30 <Pikka> I suspect even if the crash is fixed, it still may not be possible to use that var in this circumstance, so I might just switch to using another quasi-random var instead ;)
16:54:44 *** Prof_Frink has joined #openttd
17:03:45 <Pikka> your patch also affects other variables, eg var 65
17:03:56 <Pikka> yet I am finding that var 65 works :)
17:04:04 <Pikka> I hope it will not make it unwork
17:08:39 *** ComLock has joined #openttd
17:12:53 <Rubidium> peter1138: the 0x60 case, does that really need it? I don't see this->tile being used there
17:26:32 <peter1138> Rubidium, you're right. i saw tile...
17:26:45 <peter1138> Pikka, 65 needs a tiel
17:26:57 <peter1138> so in the cases where there's no tile, it won't work
17:27:24 <Pikka> well, it seems to be working
17:27:35 <peter1138> maybe it's not crashing :p
17:27:47 <Pikka> it isn't, but it's also working ;)
17:27:53 <Pikka> so hopefully your changes will not change that
17:28:43 <peter1138> the result is meaningless
17:29:50 <Pikka> anyway, we'll see, if it causes me problems I'll just have to change things around a bit, no dramas
17:29:51 <peter1138> alternatively we can ask why is tile invalid?
17:30:05 <peter1138> i->location is set in these cases
17:30:52 <peter1138> i'm looking at DoCreateNewIndustry()
17:31:01 <peter1138> we have a tile, but pass INVALID_TILE
17:31:07 <peter1138> i'm assuming there's a good reason for it
17:31:10 <Rubidium> something in the specs?
17:50:03 *** |Jeroen| has joined #openttd
18:27:53 *** valhallasw has joined #openttd
18:32:43 <Wolf01> hel... wait for it ...lo
18:40:34 <Eddi|zuHause> that was legen...
18:46:13 <DorpsGek> Commit by translators :: r24969 /trunk/src/lang (danish.txt korean.txt) (2013-02-05 18:46:08 UTC)
18:46:14 <DorpsGek> -Update from WebTranslator v3.0:
18:46:15 <DorpsGek> danish - 17 changes by Hellwing
18:46:16 <DorpsGek> korean - 48 changes by telk5093
19:19:29 *** Alberth has joined #openttd
19:19:29 *** ChanServ sets mode: +o Alberth
19:20:16 *** Progman has joined #openttd
19:57:41 * Rubidium will stop fixing bugs and committing if you strive for equal bug fix and commit counts ;)
19:59:30 <Wolf01> once at work I had to make 6 commits for a bugfix, one for the core and one for each other 5 applications which used that feature
20:00:42 * Rubidium has seen bugfixes in OpenTTD with hundreds of commits ;)
20:09:39 *** Rhamphoryncus_ has quit IRC
20:19:29 *** Rhamphoryncus_ has joined #openttd
20:24:57 <Eddi|zuHause> "-Fix: behaviour of passengers" :p
20:26:17 <Rubidium> nah, that's not that many commits
20:28:34 <Rubidium> make the UI scale to the translation... that was hundreds of commits
20:28:58 <Supercheese> and right-to-left stuff, I'd wager
20:29:40 <Rubidium> nope, that's simpler
20:31:57 <Rubidium> only a part of the windows need custom code, the rest is done generically
20:32:12 <Rubidium> for the UI scaling all windows need to be (basically) rewritten
20:34:47 *** HerzogDeXtEr1 has joined #openttd
20:35:48 <Eddi|zuHause> Rubidium: and it still cannot scale the sprite font?
20:36:19 <Eddi|zuHause> ("scale" as in provide bigger sprites)
20:36:23 <Rubidium> that's something else
20:39:49 <Rubidium> the UI can scale, just there's nothing for determining the height of the sprite font
20:43:56 *** chester_ has joined #openttd
20:47:23 <peter1138> there's probably some grfs that add specific characters in the normal size
20:48:06 <Supercheese> Huh, the New Map Features patchpack crashes when loading the 1.2.3 title game...
20:48:24 <peter1138> that's a good start
20:48:35 <Supercheese> works fine using the standard nightly title game
20:50:00 <Supercheese> Yep, crashy crash
20:51:22 <Eddi|zuHause> peter1138: yes, CETS adds some superscript numbers
20:51:38 <Supercheese> Well, I hope it doesn't crash any further...
20:51:46 <Eddi|zuHause> 0-9 and I, V and possibly X, i forgot
20:52:25 <Eddi|zuHause> default font only has 1-3 and they look ugly
20:52:31 <DorpsGek> Commit by rubidium :: r24970 trunk/src/3rdparty/squirrel/squirrel/sqvm.cpp (2013-02-05 20:52:28 UTC)
20:52:32 <DorpsGek> -Fix [FS#5346]: crash when infinite loop occured during loading a script
20:53:29 <DorpsGek> Commit by rubidium :: r24971 trunk/bin/ai/regression/ (2013-02-05 20:53:26 UTC)
20:53:30 <DorpsGek> -Fix: ignore screenshot folder in regression
20:54:43 *** LordAro has joined #openttd
21:07:26 <DorpsGek> Commit by rubidium :: r24972 trunk/src/water_cmd.cpp (2013-02-05 21:07:24 UTC)
21:07:27 <DorpsGek> -Fix [FS#5258]: do not make overbuilding rivers with canals insanely expensive; there is no real need to fill the river with dirt first, just excavate it a bit and build borders
21:10:11 <Rubidium> argh... LordAro ... why?
21:10:21 <Rubidium> don't you want a 1.3?
21:10:48 <Eddi|zuHause> we haven't released a beta in a while
21:10:51 <LordAro> pfft. you got nearly 2 months to fix bugs :P
21:15:28 <oskari89> Concerning CETS, should someone take a look at GPL-licensed sprites from other sets? Even though someone needs to draw additional 8 directions.. ?
21:16:00 <Eddi|zuHause> oskari89: there's very little advantage you get out of there
21:16:38 <Rubidium> given the amount of time I had free over the last few months, I doubt I have much time in the coming 2 months
21:16:46 <Eddi|zuHause> oskari89: but if you have some sprites, go ahead and put them into the appropriate template :)
21:29:29 <peter1138> who needs releases when you have nightlies :D
21:33:39 <Alberth> peter1138: releases are just for MP servers, so they have a version that everybody agrees on to use
21:38:41 <DorpsGek> Commit by peter1138 :: r24973 trunk/src/newgrf_industries.cpp (2013-02-05 21:38:38 UTC)
21:38:42 <DorpsGek> -Fix [FS#5462]: Prevent access to tile-based variables when tile is invalid.
21:41:35 <DorpsGek> Commit by peter1138 :: r24974 trunk/src/openttd.cpp (2013-02-05 21:41:32 UTC)
21:41:36 <DorpsGek> -Fix [FS#5279]: Honour pause_on_newgame setting when running as a dedicated server.
21:54:03 *** ProfFrink has joined #openttd
21:55:44 *** ProfFrink is now known as Prof_Frink
22:06:33 <peter1138> hmm, with a compiler optimise consecutive writes to a pointer?
22:10:57 <glx> it the 3 values are constant I guess it does
22:11:01 <__ln__> glx: do you feel it was the right thing to do to allow women to wear trousers in paris?
22:11:09 <peter1138> not constant at all :)
22:11:43 <glx> __ln__: it was forbidden since 18XX IIRC
22:13:12 * peter1138 compares with objdump
22:14:42 <glx> but I hope it does optimise it, because it should be simpler to do all the math before the write in memory
22:16:01 <glx> you are comparing the generated code with *p = (foo + bar) * baz; ?
22:16:33 <Eddi|zuHause> i've never looked at the generated asm
22:16:36 <peter1138> using a separate (register) variable
22:16:48 <peter1138> (and yes, it is performance critical)
22:18:07 <peter1138> awkward, the loops are unrolled i think
22:18:26 <peter1138> good for performance but harder to compare :p
22:23:55 <peter1138> it's an audio plugin, so while individually it doesn't, they tend to get added together
22:31:02 <michi_cc> peter1138: Is any of the foobar stuff a pointer itself? If yes, probably not because of aliasing (unless you use (__)restrict or pass the appropriate flags).
22:32:00 * NGC3982 is now an insurance agent.
22:32:44 <__ln__> NGC3982: double-oh-3982?
22:32:59 <NGC3982> Actually, i have applied to internal affairs service.
22:33:05 <NGC3982> But that will take at least a year
22:33:17 *** NGC3982 is now known as Double-O-3982
22:34:06 <peter1138> yeah, using the local var cuts processing time a lot
22:34:18 <Double-O-3982> I just love how insurance companies (like RSA) are literaly huge. They basicly have their own school system (in-wich i'm using right now) with a small faculty of teachers and employees.
22:35:45 <Bonez305> is there a way to get my company back someone bought 75% of it
22:35:46 <peter1138> each of the other pointers is used only once per sample
22:35:52 <Bonez305> i forgot to set a password :/
22:35:56 <peter1138> so i don't think i can improve on that
22:36:39 *** oskari89 has joined #openttd
22:37:46 <michi_cc> Yeah, an explicit local var is basically just the equivalent of telling the compiler that writes to *p will not change the input.
22:38:19 <michi_cc> If you're processing more than one sample at once, use MMX/SSE instructions.
22:39:21 <peter1138> beyond my knowledge :p
22:40:04 <michi_cc> Compiler intrinsics make it relatively easy. Does the plugin interface define any data alignment?
22:41:00 <michi_cc> Which CPU and which data type?
22:43:27 <peter1138> hmm, compiler has using xmm registers and scalar (not parallel) instructions
22:43:35 <Eddi|zuHause> Bonez305: that doesn't have anything to do with each other
22:44:36 <peter1138> but anyway, i need to process each input, buffer and output on a sample-by-sample basis
22:44:59 <Eddi|zuHause> Bonez305: someone who owns shares in your company can't control your company, and someone who joins your company because you didn't set a password doesn't need to own any shares in it
22:50:44 *** ProfFrink has joined #openttd
22:51:16 <michi_cc> I think the gcc intrinsics are identical to the MSVC ones, but I could be wrong :)
22:51:55 <michi_cc> Oh, sorry, the first _mm_store_ss should be _mm_store_ps
22:52:28 *** ProfFrink is now known as Prof_Frink
22:52:30 <Bonez305> Eddi|zuHause: Oh, I thought it had something to do with them owning my company lol
22:53:04 <Eddi|zuHause> Bonez305: only an administrator can kick out the second person from your company
22:53:22 <Eddi|zuHause> (or you ask him kindly to leave)
22:54:08 <michi_cc> If the output isn't immediately processed again (and 16 byte aligned) _mm_stream_ps instead of store will skip L1/L2 caches.
23:03:43 *** roboboy has joined #openttd
23:05:50 *** Markavian has joined #openttd
23:07:09 <peter1138> shit, forgot to copy the object between benchmarks
23:07:13 <peter1138> so that was useless :p
23:07:32 <peter1138> got it down to 40 ms, from 220ms originally
23:08:58 <peter1138> object code is smaller too o_O
23:12:26 <peter1138> yeah, um, sorry about that
23:12:40 <peter1138> the foo/bar/baz was just an abbreviated example
23:13:29 <michi_cc> Yeah, but it should still illustrate the concept.
23:14:03 <peter1138> i can't do 8 samples at a time
23:14:10 <michi_cc> Hmm, it would probably be better to load all _*1 before _*2
23:14:12 <peter1138> due to interactions between buffers
23:14:29 <peter1138> well, possible interactions
23:14:39 <peter1138> buffer length could be as low a 1 sample
23:14:45 <Eddi|zuHause> constructs like "*out++" scare ne
23:15:13 <peter1138> thoguh it still uses the whole buffer, the write pointer would be 1 sample behind the read pointer in that case
23:15:37 <peter1138> and the difference between them will be different for each buffer
23:15:51 <peter1138> i am happy with 220ms -> 40ms
23:16:40 <michi_cc> If you can query the buffer length (and buffers >= 8 samples are common), have two code paths (i.e. SSE for > 8 samples and a simple loop for the remainder).
23:22:16 <peter1138> i can process 26 seconds of audio in 270ms
23:23:14 <peter1138> oh wait, i calculated that wrong :S
23:24:13 <peter1138> 10 seconds of audio in 36ms
23:25:36 <peter1138> now do i commit this all at once as optimisation
23:26:09 <peter1138> changed a couple of data structures to enable loops to be reused
23:29:29 <peter1138> i should check it still works acutlly :p
23:43:24 <peter1138> pointer/array deference in the wrong order
23:46:13 <peter1138> that's disappointing!
23:46:23 <peter1138> brings it up to 48ms for 10 seconds of audio
23:46:56 <Eddi|zuHause> darn it! programs that work are slower than programs that don't work :p
23:50:18 *** roboboy has joined #openttd
23:58:41 *** Celestar_ has joined #openttd
continue to next day ⏵
