IRC logs for #openttd on OFTC at 2011-12-29
⏴ go to previous day
00:25:21 <FLHerne> Just seemed a bit quiet on here...
00:27:33 <Mark> It's the days between Christmas and New Year Eve.
00:31:28 <Bawx> When I am trying to install OpenTTD 1.2.0-beta1, it says "Downloading of OpenMSX failed". Why does that happen?
00:54:38 *** mooseman has joined #openttd
01:03:26 *** ChanServ sets mode: +v tokai
01:20:07 <mooseman> Anybody else playing OpenTTD on Ubuntu? :D
01:38:35 *** fjb|tab is now known as Guest22027
01:38:36 *** fjb|tab has joined #openttd
02:32:27 *** AndroUser has joined #openttd
02:33:09 <AndroUser> :l How do I change my nick?
02:40:34 *** AndroUser is now known as Kendued
02:43:34 *** Kendued is now known as ^-_Kendued_-^
02:44:02 <mooseman> Are you on Ubuntu, Kendued?
02:44:10 <^-_Kendued_-^> Lol i i guess you found me playng with /nick
02:44:16 <^-_Kendued_-^> No, Im on android
02:46:11 <^-_Kendued_-^> Anyone else to talk to online?
03:03:56 *** AndroUser2 has joined #openttd
03:05:05 <Afdal> Can someone explain to me how depot pathfinding for maintenance works?
03:06:19 <Afdal> Why do trains service more often when you split depots from the main track using a path signal than with block or presignals?
03:38:58 <Afdal> Do trians only look ahead for 9 tiles when pathfinding?
03:58:13 <Eddi|zuHause> there is a setting for the maximum distance
04:05:30 *** fjb|tab is now known as Guest22042
04:05:30 *** fjb|tab has joined #openttd
05:30:26 *** Kurimus has joined #openttd
05:56:26 *** Eddi|zuHause has joined #openttd
06:25:42 *** thefiler has joined #openttd
06:26:15 *** thefiler has joined #openttd
06:28:59 <thefiler> need some questions answered
06:31:09 <Afdal> I might be able to help, but probably not
06:33:15 <Rubidium> thefiler: experts of what? Does your question really need an expert, or just someone knowing the answer
06:33:40 <Rubidium> in any case, I'm an expert but possibly not in the area you need the expert in
06:37:13 <Afdal> Afdal Can someone explain to me how depot pathfinding for maintenance works?
06:37:15 <Afdal> Afdal Why do trains service more often when you split depots from the main track using a path signal and less with a block or presignal?
06:39:10 <Rubidium> I guess thefiler quit right after asking his meta question... why are people so unpatient?
06:40:00 <Rubidium> Afdal: vehicles try every X ticks to find a depot that is within Y pathfinder penalties (don't know the exact number)
06:40:26 <Afdal> When they're in need of a maintenance, right?
06:40:45 <Afdal> So does a one-way path signal have less penalty than regular signals?
06:40:54 <Afdal> That still doesn't explain how
06:41:03 <Afdal> trains find the depot with regular signals
06:41:09 <Afdal> They can find depots with both
06:41:14 <Rubidium> with path signals it starts looking from the end of the reservation, whereas with normal signals it's from the front of the train
06:41:23 <Afdal> but if you use a path signal they maintain themselves significantly more often
06:42:10 <Afdal> hmm, I don't understand how that would bias the path-signaled depot
06:42:19 <Rubidium> if the signal blocks are large/long, then it is more likely to that the depot is sought when the train's reservation ends near the depot
06:43:04 *** thefiler has joined #openttd
06:43:11 <Afdal> Here's an example by the way
06:43:16 <Afdal> in case I'm not being very clear
06:43:38 <Rubidium> Afdal: it really depends on the situation; with many short signal blocks with path signals it's less likely to find one with path signals
06:44:05 <Afdal> But in this case it's the opposite of that
06:44:15 <Afdal> The path signal favors maintenance
06:45:19 <Rubidium> as I said, it tries to find a depot every X ticks, and does so with a max distance of Y
06:45:54 <Rubidium> now I'd reckon that the distance from depot to main track is very close to Y
06:46:21 <Afdal> If you make the track between the main track and the depot
06:46:29 <Afdal> any longer, trains won't service at all
06:47:48 <Rubidium> so, likely the only place a train can be to 'see' the depot within max distance is the tile with the pre/path signal
06:48:49 <Afdal> I just don't see how that would be any any different between signal types
06:49:11 <Rubidium> now for the path signal, once the train passes the signal before the path signal the train reserves a path up to the path signal. From that moment *all* pathfinding of that train happens "from" the path signal's location
06:49:31 <Rubidium> which means that with the path signal it sees the depot two tiles earlier
06:50:04 <Afdal> So with the path signal it has more time to look for a depot
06:50:24 <Afdal> more time to coincide with the depot check tick
06:50:59 <Rubidium> and this is where the X ticks comes to play as well. If the train tries at the tile just before the pre signal it is unlikely to try at the tile with pre signal due to its speed, as such it doesn't see the depot. With the path signal it would see the depot because of the reservation ending at the signal
06:51:36 <Afdal> Thanks, that makes a lot of sense
06:51:46 <Rubidium> ... changing the maximum distance to depot for maintenance?
06:51:54 <Rubidium> there's a setting for that ;)
06:52:27 <Afdal> Does that mean I could help trains service even better by making the signal block just before the path signal longer?
06:53:16 <Rubidium> yapf.maximum_go_to_depot_penalty or npf.maximum_go_to_depot_penalty depending on the chosen pathfinder
06:54:22 <Afdal> So just so I'm clear on this
06:54:44 <Afdal> The train looks for depots in front of its reserved path?
06:55:38 <Afdal> Agh, now I'm confused again. It sounds like the way you're arguing it, it would only affect if there was a path signal just before the split signal
06:56:10 <Afdal> instead of at the split signal
06:56:34 <Rubidium> well, depends of what's "in front" is ofcourse
06:56:38 <Afdal> Because it reserves the extra track before the split, there's more chance to hit the depot check
06:56:43 <Afdal> but a block signal won't do that
06:57:31 <Afdal> Now I'm confused again :(
06:58:04 <Rubidium> well, to confuse you even more...
06:59:00 <Rubidium> the pathfinding *always* happens from the "front" of the reservation, but with path signals it reserves till the next safe waiting point (signal / end of track), but with block signals it only reserves the track directly under the train
06:59:54 <Afdal> front being the furthest from the train, right?
06:59:59 <Rubidium> front is the reservation furthest away from the train in the direction it is riding
07:00:45 <Afdal> Yeah, the way you're describing it sounds like a path signal BEFORE the split signal is what would matter
07:01:53 <Rubidium> Afdal: but a block signal leading to a path signal reserves a path as well
07:02:19 <Rubidium> yes, but only when there's no other train in that signal block
07:02:36 <Afdal> It doesn't show it like that with track reservation on
07:02:46 <Afdal> Looks like the block signal acts the same as usual
07:02:53 <Afdal> show track reservation*
07:03:17 *** lordnokon has joined #openttd
07:03:24 <Afdal> with the reserved track only directly under the train
07:04:12 <lordnokon> is there no way to custom industries, instead of using on the 2040 tones per month limit
07:04:13 <Rubidium> for me it allocates the whole track from the block signal up to the path signal
07:04:53 <Afdal> Are you on 1.1.4 or 1.20 beta?
07:04:58 <lordnokon> so if i want to have a coal mine produce 1mil tones a month, that would save on the amount of industries needed to be build or funded, which will help not slowing down the game with lots of industries?
07:05:52 <Rubidium> but 1.1.4 behaves the same as the way I'm describing it
07:06:25 <Afdal> The track reservation shows the same before with the block signal before the path signal as per usual block signals
07:06:43 <Rubidium> oh, I'm not using the one way path signal
07:07:01 <Rubidium> but the 'normal' path signal
07:07:45 <Afdal> That doesn't explain why one-way path signals favor servicing still
07:08:34 <Afdal> all over factors being equal
07:08:55 <Rubidium> different pathfinder penalties between the signals I'd guess
07:09:23 <Afdal> Is there a penalty for exit signals?
07:09:36 <Afdal> no wait that shouldn't matter
07:09:41 <Afdal> It's the same with each branch
07:10:01 <Rubidium> I don't know by heart what the different pathfinder penalties are
07:10:46 <Afdal> Also the maintenance rate with presignals and regular block signals at the track split is the same
07:10:56 <Afdal> It's only higher with a path signal
07:13:01 <lordnokon> any help on my questions?
07:13:30 *** DayDreamer has joined #openttd
07:15:34 *** DayDreamer has left #openttd
07:15:36 <Afdal> The way you describe it, there should be a significant service rate difference between a two-way and a one-way path signal at the split
07:22:30 <Afdal> So that explains why a two-way path signal will allow better servicing than a one-way
07:22:54 <Afdal> But it doesn't explain why a one-way path signal will allow better servicing over block/presignals
07:32:27 <Afdal> Actually, a second test gives slightly lower maintenance rate for two-way than one-way
07:32:36 <Afdal> I guess I need to run it a lot longer to be conclusive
07:58:18 *** jonty-comp has joined #openttd
08:03:43 *** TWerkhoven[l] has joined #openttd
08:21:48 *** [1]Mark has joined #openttd
08:21:49 *** Mark is now known as Guest22060
08:21:49 *** [1]Mark is now known as Mark
09:21:24 <lordnokon> can anyone help me, my trains are suffering, on hills in nightly, i've tried the advanced options. Am i missing something? i've seen youtube clips where people's trains fly up and down hills
09:24:51 <Terkhen> lordnokon: do you have realistic acceleration enabled? if so, with which slope inclination?
09:24:52 <Terkhen> also, it might just be that people in those videos have trains carrying less cargo or with more power
09:26:22 <lordnokon> train acceleration is set to orginal
09:28:52 <Terkhen> try to set it to realistic, slope 3%
09:28:57 <Terkhen> those are the default settings
09:33:00 <lordnokon> my trains are the same size, with power and speed. still they lose speed, theirs never does
09:33:01 *** andythenorth has joined #openttd
09:34:52 <Afdal> Original acceleration is the default
09:35:38 <lordnokon> ok hang on it works now... but now at the stations everything slows down
09:36:53 <lordnokon> and all corner speeds are fooked lol
09:49:00 *** Progman has joined #openttd
09:51:20 *** LordPixaII has joined #openttd
09:53:19 *** TGYoshi has joined #openttd
09:53:36 *** KingPixaIII has joined #openttd
09:59:49 <Terkhen> Afdal: original was the default :)
10:00:51 <Terkhen> lordnokon: my guess is that he is playing with 0% slopes
10:01:00 <Terkhen> but without a savegame it is just a guess
10:05:47 *** Progman_ has joined #openttd
10:10:45 <planetmaker> hello andythenorth
10:11:36 *** Progman_ is now known as Progman
10:13:34 *** Alberth has joined #openttd
10:13:34 *** ChanServ sets mode: +o Alberth
10:31:50 *** mahmoud has joined #openttd
10:39:41 <TrueBrain> oh noes! It is ohnoes!
10:42:33 <ohnoes> I would like to ask a question. I am still novice but I am facing a problem regarding a coal mine, a power station and the railways I built. For some reason I do not understand, my coal keeps accumulating at the station next to the power station and when I click it there is a "x tonnes of coal en-route from <the station where the coal mine is>"
10:43:26 <planetmaker> do you use transfer or unload orders or does the power plant station not accept the coal?
10:44:35 <ohnoes> I am using full load <coal mine> -> full unload <station near the power station> -> maintain at <some depot>
10:45:24 <planetmaker> just use normal 'goto'. It'll unload, if accepted
10:45:41 <planetmaker> and does the station accept coal?
10:45:42 <ohnoes> I think I tried that before too before I changed to "full unload" but I will give it a try again. thanks!
10:46:08 <planetmaker> if it doesn't unload then, the power plant or station does not accept coal.
10:46:26 <planetmaker> hard to say what when talking w/o savegame
10:47:23 <ohnoes> but the power station says "requires: coal". what do you mean by it "does not accept coal" ? Is there a way I can check that?
10:48:22 <planetmaker> if you don't use industry newgrfs, the power plant will always require coal
10:48:49 <Alberth> I always open the 'build station' window, make station the same size as you have, and hover over the existing station. Then watch the build-station window, it should say 'accepts coal'
10:49:32 <Alberth> there should be an easier way to do this, I agree ;)
10:51:10 <ohnoes> exactly, that's what I did too
10:51:19 <ohnoes> but actually since yesterday I am trying the 32bpp
10:51:31 <ohnoes> newgrfs... is that the cause of the problem ?
10:51:57 <Alberth> could be, do you use ECS ?
10:53:06 <ohnoes> my bad. I just noticed that the power station was partially covered by the train station (without accepting coal)
10:53:07 <Alberth> always nice to solve a problem just by asking questions ;)
10:53:33 <ohnoes> like I said. noob :) thank you everybody
10:53:43 <Alberth> Extending the station a bit should work, I think
10:54:03 <ohnoes> yeah, that's probably what I'll do :)
10:54:17 <ohnoes> I made it parallel, so probably I will just add a new line
10:54:54 <Alberth> existing cargo at the station will not be accepted until you transport it again though
10:55:39 <Alberth> just letting it linger is also an option, it will disappear in time
11:00:20 *** Devroush has joined #openttd
11:08:29 *** JVassie has joined #openttd
11:20:36 *** |Jeroen| has joined #openttd
11:35:56 *** valhallasw has joined #openttd
11:44:47 *** mahmoud has joined #openttd
11:48:23 *** tokai|mdlx has joined #openttd
11:52:07 <Amis> Whoever did the hot air balloooooon newgrf...
11:52:51 <Alberth> it does not say who the author is?
11:53:34 <Amis> I have no idea. I just put together a newgrf pack quickly and was playing it and all of a sudden
11:53:55 <Alberth> or if you can find the topic in the forum, you can find his forum name and send a PM to him
11:54:19 *** Adambean has joined #openttd
11:58:13 <Amis> Haha, I didn't know it's possible :D
11:58:17 <Amis> "Aircraft ran out of fuel"
11:58:44 <Alberth> it was recently added
11:59:44 <planetmaker> well. it was possible before, if you destroyed all airports
11:59:54 <planetmaker> and the aircraft had nowhere to land
11:59:56 *** KingPixaIII is now known as Pixa
12:21:40 *** Markavian has joined #openttd
13:03:32 *** Biolunar has joined #openttd
13:17:33 *** Brianetta has joined #openttd
13:19:48 *** Brianetta has joined #openttd
13:27:26 *** valhallasw has joined #openttd
13:55:35 *** sla_ro|master has joined #openttd
14:01:14 <Belugas> mister Alberth, how are you going?
14:01:25 <Belugas> damned.. fingers are slow today
14:02:06 <Alberth> a bit annoyed with my lack of progress w.r.t. paths in freerct, so I am playing a bit of OpenDune :p
14:02:33 <Alberth> ieks, please keep that there :)
14:02:41 <Belugas> you are? cool :) I should give it a try indeed :)
14:02:51 <Belugas> herm... at OpenDune...
14:03:04 <Belugas> not at sending you da cold
14:19:11 *** HerzogDeXtEr has joined #openttd
14:41:58 <Alberth> looks like an Orcs machine from warhammer 40000 :p
14:54:41 <JVassie> can anyone see anything wrong with ym query pls?
14:54:49 <JVassie> 1064 normally means reserved keyword
14:54:58 <JVassie> but i cant see any that im usign which are reserved
14:58:44 <JVassie> even if i change it to this
15:05:55 <Belugas> JVassie, i would remove the quotes on the INSERT statement. I never use them. But i'm not useing MySQL, so i cannot tell
15:06:15 <JVassie> around the column names or the values in inserting?
15:06:46 <Belugas> INSERT INTO Transferees (title,firstname,lastname...
15:08:00 <JVassie> brain acheing, hard to think what i couldve been doing wrong :p
15:08:36 <JVassie> INSERT INTO Persons (P_Id, LastName, FirstName)
15:08:36 <JVassie> VALUES (5, 'Tjessem', 'Jakob')
15:15:05 <andythenorth> JVassie: your column IDs are escaped in strings
15:15:08 <andythenorth> try them without
15:15:23 <JVassie> s'what belugas suggested :)
15:15:35 * andythenorth was just guessing
15:25:07 *** fjb|tab is now known as Guest22088
15:25:08 *** fjb|tab has joined #openttd
15:30:00 *** JVassie_ has joined #openttd
15:48:20 *** Chris_Booth has joined #openttd
15:56:21 *** DayDreamer has joined #openttd
16:06:18 *** mikegrb_ has joined #openttd
16:16:59 *** leeroberson has joined #openttd
16:17:11 <leeroberson> wow, lots of users in this channel, hurray
16:17:15 *** leeroberson is now known as eberon
16:17:26 <eberon> I'm just starting back on OpenTTD after an extremely long hiatus.
16:18:14 *** supermop has joined #openttd
16:18:15 <eberon> can someone tell me, does OpenTTD ship with an AI in the installer, or do I need to find and pick one in order to play with an AI?
16:18:39 <andythenorth> download AIs from in-game content service
16:18:50 <andythenorth> I don't think one comes bundled
16:18:53 <eberon> okay great, I got the sense I needed to do that
16:19:00 <eberon> most of them seem experimental, is there a solid one?
16:20:35 <Zuu> There are several good ones.
16:20:42 <Zuu> I use to play with CluelessPlus and AIAI.
16:21:05 <eberon> durr, my eyes skipped the wiki link on the site, this should be super helpful.
16:21:16 <Zuu> Convoy is also very stable.
16:22:07 <eberon> thank you for the recs, clueless so far seemed to be the one with the most reasonable looking description but I will try a few of them
16:22:10 <Zuu> And there are more, it's just that I tend to only play with 2-3 in a single game.
16:22:34 <eberon> I am so excited at how much OpenTTD has improved in the last couple of years. This is one of my first childhood games, got it on floppies at a computer store that had racks of shareware in ziplock bags.
16:23:40 <Belugas> floppies... long time ago indeed :)
16:23:42 <andythenorth> there's a *lot* of new stuff :)
16:23:44 <Zuu> My other AI, PAXLink is something you could stay away from if you look for rock solid AIs :-)
16:24:09 <eberon> Zuu: I hope to become killer at this game again!
16:24:41 <eberon> andythenorth: one thing that turned me off of OpenTTD was the vast changes. I liked that it improved gameplay issues but otoh some of the changes were a little extreme and felt like cheating. I like the rules manager in the game. I wish it had presets.
16:25:07 <eberon> I've done some work on other FOSS projects in the past, maybe that is something I will try to advocate for or work on in OpenTTD at some point
16:25:31 <eberon> after a play a few dozen games :D
16:25:33 <Zuu> You are more than welcome to contrubute to the OpenTTD project :-)
16:26:45 <Zuu> But starting by playing is probably a good idea to get an idea of the vast changes and improvments compared to TTD :-)
16:27:04 <eberon> Yeah. I just noticed the server list on OpenTTD.org, that's amazing, I guess there's network code now?
16:27:38 <Alberth> nah, we have a 'multiplayer' button at the intro screen just for looks :)
16:28:00 <Zuu> (world wide openttd game day 2)
16:28:20 <eberon> Alberth: ha, in open source games there's a ~50% chance that means you'll get a "coming soon!" screen when you press a button like that
16:28:44 <andythenorth> ottd doesn't do that so much
16:28:47 <Alberth> eberon: our buttons are mostly quite functional :)
16:28:51 <andythenorth> if it doesn't work, it's not shipped
16:29:32 <Alberth> of ocurse, it may not do what you think it does :)
16:29:56 <andythenorth> more often it doesn't do what I think it should :P
16:31:28 <Elukka> Eddi|zuHause: starting to look alright?
16:32:57 <michi_cc> Elukka: Slightly related question: Would you say that drawing such a wagon for x2/x4 zoom in still in 8bpp makes sense and looks good? Or are zoomed in sprites not in 32bpp useless?
16:34:30 <Elukka> i think i'd do it in 32bpp (why not?) but still in the pixel art style if it's meant to fit together with current graphics
16:35:35 <michi_cc> Because the 32bpp-anim blitter is a lot more expensive than all other blitters.
16:36:00 <Elukka> that sprite is for the default zoom level, just zoomed in in photoshop so it's easier to see
16:36:36 <michi_cc> The question is more like, is it possible to draw proper zoomed-in sprites in 8bpp or are the colours too limited?
16:36:55 <Elukka> i haven't tried though
16:37:26 <michi_cc> So it does make sense to implement/allow it :)
16:37:49 <andythenorth> michi_cc: it's possible
16:37:56 <andythenorth> what the style would be I have no idea
16:38:11 <TrueBrain> michi_cc: if it doesnt hurt, there is never a loss by allowing it ;)
16:38:19 <Elukka> could just do the current style at higher resolution, andy
16:38:46 <Elukka> depends on if it's meant to fit in with current graphics or with typical 32 bpp graphics
16:38:54 <andythenorth> Elukka: in that case just zoom in photoshop, with interpolation turned off
16:39:01 <andythenorth> instead of clicking 4x as often
16:39:11 <Elukka> that's not a higher res sprite though
16:39:16 <michi_cc> The current typical 32bpp graphics don't even fit themselves :)
16:39:24 <Elukka> if the sprite was double or four times the resolution there's much more room for detail
16:39:37 <andythenorth> that's a style change :)
16:40:16 <andythenorth> michi_cc: what does it do? Additional realsprites?
16:40:52 <michi_cc> Extended RealSprite (or however you'd want to call it) for multiple zoom levels and/or 32bpp.
16:41:48 <Elukka> hm. i'll do a quick 2x version of one of my sprites (just one angle) to see how it'd look
16:42:13 <andythenorth> michi_cc: looks ok
16:47:53 <Elukka> so, is there/will there be a trunk implementation of higher res sprites for the extra zoom levels?
16:47:56 <Elukka> because that'd be really nice
16:48:39 <TrueBrain> Elukka: I think that is what michi_cc is hinting about ;)
16:53:38 *** bremerjoe has joined #openttd
16:53:58 <bremerjoe> Good evening everybody!
16:56:47 <TrueBrain> "We received instruction from the Debt Management office, The Presidency, to have your fund approved by the contract Review and Payment Committee"
16:57:48 <Elukka> oh man twice the resolution is lovely
16:58:34 <Elukka> so much more room for detail
16:59:03 <andythenorth> there speaks someone who doesn't have about 250 open tickets for newgrf sets :P
17:00:32 <Elukka> technically i'm supposed to do the rolling stock for CETS, i'm just too lazy to get much done :P
17:00:59 <andythenorth> EZ should be rendered not drawn
17:01:11 <andythenorth> and with the extra angles in CETS, you might seriously consider it
17:01:21 <andythenorth> build one model, render all 16 angles
17:02:28 <Elukka> i'd do that even for default resolution sprites because all those angles get really tedious but i don't have the required rendering wizardry skills
17:02:29 <andythenorth> possibly use UV mapping
17:02:46 <andythenorth> with UV mapping, you draw a net and then wrap it to a mesh
17:02:51 <Elukka> i could do the models easily enough, but i can't make a renderer poop out pixel art
17:03:03 <Elukka> it's possible and i've heard someone does it through arcane scripting magics, i just don't know how
17:03:08 <Elukka> i'd love to know, though
17:03:59 <Elukka> it'd make it immeasurably easier and faster... i don't think i'd use uv maps, i'd just get the basic shape and overpaint that
17:04:08 <TrueBrain> what happened to that dude anyway?
17:04:28 <TrueBrain> our voxel / cubical dude
17:04:38 <andythenorth> we drove him away, he was a fool
17:04:57 <Elukka> can we get a non-fool who can help me set up a rendering thing? :P
17:06:07 <TrueBrain> Elukka: I just setup a blender file which renders my lego for me :P
17:06:49 <andythenorth> apparently the best thing is to use something called Qubicle Constructor
17:06:54 <andythenorth> where did I hear that before? :p
17:07:40 *** TomyLobo2 has joined #openttd
17:07:50 <Elukka> is openttd actually isometric?
17:07:52 <Elukka> or dimetric or something
17:07:55 <Elukka> important to get the projection right
17:07:56 <TrueBrain> andythenorth: minecraft for grownups :)
17:08:07 <TrueBrain> Elukka: which app you use?
17:08:07 <andythenorth> Elukka: dimetric
17:08:37 <Elukka> i've used kerkythea for general rendering work but it's been having issues
17:08:54 <TrueBrain> for blender: Isometric, 30 degrees, at a 45 angle
17:09:10 <TrueBrain> 12.5 in 'width' is 1 tile
17:09:32 <TrueBrain> (on a 128x64 resolution)
17:09:52 <Elukka> thanks, i've gotta try it out
17:10:01 <TrueBrain> took me a while to find the right values ...
17:10:03 <Elukka> i'm thinking maybe i'll make my models out of pixel-sized blocks
17:10:07 <TrueBrain> the .blend on te wiki (and tt-forums) is wrong
17:10:09 <TrueBrain> annoying as fuck :P
17:10:56 <Elukka> 24 angles by hand is... tedious as hell
17:11:23 <TrueBrain> single change, and redo them all :(
17:11:50 <TrueBrain> atm I just hit RunScript, and 2 minutes later I have new files :P
17:12:20 <Elukka> you have script for it?
17:12:45 <TrueBrain> render all faces, save them correctly, runs pngcodec over them
17:13:28 <Elukka> if i can have that you'll get all the cookies
17:13:33 *** TomyLobo2 is now known as TomyLobo
17:13:46 <TrueBrain> power of Python in Blender :P
17:14:00 <TrueBrain> very specific for my case, but you get the point :)
17:14:22 <TrueBrain> hope to finish my 'master' blender soon; will make sure to publish it
17:14:30 <TrueBrain> (linking to other blend files is AWESOME)
17:19:13 *** LordPixaII has joined #openttd
17:19:53 <Yexo> bremerjoe: your article is named "Ladebuchten", it should be named "Ladebuchten/De"
17:20:47 <Yexo> I've now moved it for you
17:21:32 <Yexo> and I see it's time to run the language script again
17:22:21 <Yexo> if you add a new translations and link it from the english page, the script adds links from all translated pages to the new translation
17:23:02 <Yexo> so in the case above: You add it to "Loading bays", script adds the link to "Ladebuchten/De" from "Estaciones_para_vehículos_de_carretera/Es" "Aire_de_chargement/Fr" and "Laadperrons/Nl"
17:23:17 <Yexo> last time I ran that script was somewhere in the summer I think, completely forgot about it
17:23:31 <TrueBrain> should we run it every night?
17:23:34 <Yexo> would be best if it ran every night on the server
17:23:42 <bremerjoe> THX Yexo! Could you tell me how you edited the article name?
17:23:43 *** fjb|tab is now known as Guest22099
17:23:44 *** fjb|tab has joined #openttd
17:23:56 <TrueBrain> Yexo: either do it under your own user, or I can move it to www-data or something if you like
17:24:10 <TrueBrain> or do it under your own user, and I will move it :P
17:24:11 <Yexo> bremerjoe: on the top next to "page" and "edit" there is a "move" button
17:24:33 <Yexo> TrueBrain: ok, currently I run it from my laptop. I'll run it again first and later upload it and notify you again
17:24:55 <bremerjoe> learning step by step. Thanks again Yexo!
17:25:47 <bremerjoe> I understand correctly that I need to link mine to the english version by adding |de=Ladebuchten there, right?
17:26:22 <Yexo> yes, but you've already done that
17:26:41 <Yexo> you don't need to edit all other languages, I'll do that (with a script) later
17:26:50 <bremerjoe> OK. Just wanted to make sure I know that this is necessary and not messing up anything
17:32:15 <JVassie_> doing a mysql query 27k times in php sucks ass
17:32:28 *** Chris_Booth is now known as mdf
17:32:46 <TrueBrain> doing (..) php sucks ass
17:33:12 *** mdf is now known as Chris_Booth
17:33:40 * JVassie_ shall play RIFT whilst it trundles along
17:33:42 <andythenorth> JVassie_: what are you doing / why?
17:34:30 <JVassie_> i have a table of soem 27k rows
17:34:37 <JVassie_> which is a copy of another table
17:34:55 <JVassie_> the new table is using the old tables unique id as a foreign key
17:35:06 <JVassie_> so the new table needs a new unique id
17:39:02 *** supermop has joined #openttd
17:39:58 <andythenorth> 27k doesn't sound that much
17:41:07 * andythenorth knows very little about mysql
17:41:32 <andythenorth> couldn't you just drop the old data to csv, then write an import script, using mysql's auto-inc when you add the new record?
17:45:47 <Belugas> ... and you're doing youre query once per row???
17:47:05 <andythenorth> also, why copy the table?
17:47:10 <andythenorth> why not just add a column?
17:49:30 * andythenorth learns about database normalisation
17:50:28 * Belugas learned a ling time ago that sometims, normalization is the opposite of performance
17:51:27 <Belugas> so, sometimes, for sake of performance, y9u have to sacrifice a bit of normalisation, or put TONS of indexes. Which sometime sucks on inserting
17:51:55 <Belugas> but is fast on reading. since I do more inserts than reads... well...
17:53:25 *** Biolunar has joined #openttd
18:03:44 <Belugas> -16 :( i will not go out shopping a new guitar today :S
18:08:13 *** encoded has joined #openttd
18:08:51 <encoded> i love TTD, i was playing it off and on till... yesterday
18:09:17 <encoded> only now i found out about openttd
18:09:50 <encoded> it's great... but i can bearly see stuff needs a few more zoom levels
18:10:59 <TrueBrain> even more zoom levels?
18:11:09 <TrueBrain> I thought 4 times zoomed in should be sufficient, even for a 60" screen
18:12:12 <Belugas> what stuff are you talking about?
18:13:55 <encoded> even at 1280x768 it could be bigger
18:14:50 <encoded> i cant belive im the first to comment about it
18:15:29 <andythenorth> it's never been mentioned before ever ever ever :)
18:15:52 <TrueBrain> You are absolutely the first, and this issue hasn't been adressed at all!
18:16:02 <TrueBrain> TO THE BATMOBILE! (*tune starts here*)
18:16:33 <encoded> how do i get a screenshot?
18:16:38 <eberon> I think I'm going to be spending a lot of time lurking in this channel from now on ;)
18:16:46 <TrueBrain> put your screen on a photocopier
18:16:53 <andythenorth> encoded: '?' then 'screenshot'
18:17:02 <andythenorth> unless you have opengfx, then I have no idea :)
18:17:28 <TrueBrain> eberon: you and the other 114 people in here :D
18:17:29 <andythenorth> encoded: if you were playing TTD, I figure you have the original graphics?
18:17:52 <encoded> i went with the openfx
18:18:10 <andythenorth> well it seems to also be '?' in opengfx
18:18:55 <andythenorth> also if you go to 'advanced settings' -> 'interface' -> 'display options' -> 'maximum zoom level'
18:19:01 <andythenorth> what value do you have?
18:19:14 <TrueBrain> andythenorth: start first by asking which version he is playing ;)
18:19:23 * andythenorth forgets these things
18:19:46 <andythenorth> hg pull -u; make run -j13 is like a daily thing for me
18:20:08 <DorpsGek> TrueBrain: topic [<channel>]
18:20:13 <DorpsGek> TrueBrain: topic [<channel>]
18:20:20 <TrueBrain> how does this work?
18:20:22 <DorpsGek> TrueBrain: 1.1.4, 1.2.0-beta1 | Website: *.openttd.org (translator: translator, server list: servers, wiki: wiki, patches & bug-reports: bugs, revision log: vcs, release info: finger) | Don't ask to ask, just ask | 'Latest' is not a valid version, ever | English only
18:20:47 <TrueBrain> "'Latest' is not a valid version, ever"
18:20:49 <TrueBrain> that part, to be clear ;)
18:20:53 <Belugas> [13:16] <TrueBrain> put your screen on a photocopier <-- lol!!!! Made my day :D I DO LOVE YOU, It's now official !
18:21:41 <eberon> mm, abundant room to look at the terrain
18:21:56 <andythenorth> encoded: looks lovely :)
18:22:09 <andythenorth> there is a way to get big gui, I forget how
18:22:54 <Aali> its a grf, its in bananas
18:23:00 <Aali> also, long time no see #openttd
18:23:00 <encoded> but youre saying i can get more zoom levels in latest beta?
18:23:12 <andythenorth> openttd 1.2.0-beta1 has extra zoom levels
18:25:50 <encoded> thx i'll check it out later
18:29:16 <eberon> guys, I have a question of clarification here
18:29:51 <eberon> Is this trying to state that when a train needs to move on an angle, it slows down (seems obvious, you can even watch it happen) or that one type of track that orients in one direction is actually slower than another (seems absurd).
18:31:02 <eberon> if the former, I'm not sure "move slower on angled track" is exactly precise, not to be a nitpick, probably a more accurate way of saying it is that trains slow down for tracks that turn/change direction
18:31:51 <TrueBrain> diagonal track != angled track
18:31:55 <TrueBrain> I guess there you go wrong
18:32:13 <SmatZ> In OpenTTD, trains move slower on angled track (thanks SmatZ!),
18:32:21 <SmatZ> yay, I am thanked to there! :)
18:32:41 <eberon> yeah, I was going to say, IMO the header there confuses me too
18:33:09 <eberon> are we talking about tracks that are not perpenticular to the isometric view of the terrain or tracks that change direction?
18:33:16 <eberon> sorry if I'm coming off nutty
18:34:09 <TrueBrain> angled track, 2 pieces of track with an angle
18:34:13 <TrueBrain> diagonal track, track that goes diagonal
18:35:47 <SmatZ> on || or == tracks, the train moves slower
18:35:59 <SmatZ> from your point of view
18:36:04 <eberon> it has the same max speed but it has lower acceleration?
18:36:19 <eberon> so you mean pixel distance on the screen
18:36:27 <SmatZ> yeah, something like that
18:36:35 <SmatZ> the problem is that it slows down the train behind it
18:36:46 <SmatZ> eg. when you have two 10-tiles long trains
18:37:02 <SmatZ> the second one is waiting behind the first one, the first one is stopped
18:37:08 <SmatZ> and you start the first one
18:37:31 <eberon> the signaling page both excites and brings great fear to me
18:37:33 <SmatZ> they will soon keep some constant distance
18:37:52 <eberon> ah, so openttd's contributors do consider it a flaw
18:38:15 <SmatZ> but once the first train enters a || or == track (long enough), you will see the second trains will get more near to the first one, possibly slowing down or even stopping
18:38:21 <Aali> its not something you need to worry about 99% of the time
18:38:31 <SmatZ> it's been that way since the dawn of times
18:39:11 <andythenorth> do some newgrfs really have trains that change direction according to orientation of route?
18:39:14 * andythenorth calls BS on that
18:39:27 <andythenorth> length / direction /s
18:39:40 <SmatZ> on heavily used tracks, like in #openttdcoop games, it sometimes is a problem
18:40:05 <andythenorth> "The following examples have been created under the assumption that the length of the train also changes on diagonal track. This seems to be true for only some NewGRFs! (Iirc, it is true for the JapanSet and wrong for the default trains.)"
18:40:07 <SmatZ> andythenorth: length of a vehicle can't change outside of a depot
18:40:33 <SmatZ> andythenorth: maybe there are some news I don't know about :)
18:41:05 <SmatZ> but... just imagine an engine suddenly gets longer... so the wagon behind it has to hop back, or whatever
18:41:21 <andythenorth> smells of more wiki crap to me
18:41:34 <andythenorth> curvature you can get
18:41:40 <andythenorth> orientation is not a varact2 afaik
18:41:45 <andythenorth> unless it's in the 80+ vars
18:41:58 <SmatZ> actually, a newgrf can do that
18:42:06 <Aali> sure its not some rounding "error"?
18:42:16 <SmatZ> but it triggers a "newgrf bug" in openttd
18:42:22 <Aali> since default wagons are all half-tile
18:42:27 <Aali> and I assume japset isn't
18:42:27 <DorpsGek> glx: 'Latest' is not a valid version, ever
18:43:20 * andythenorth roams the 80+ vars
18:43:32 <andythenorth> I could change graphics depending on where a ship/plane is trying to get to :P
18:44:16 <andythenorth> there is a var for direction
18:44:55 <Belugas> there is an app for that
18:45:09 <SmatZ> you are not the first to ponder about "smooth" vehicle rotation, by supplying different sprites depending on the "angle" of the vehicle
18:45:36 <CIA-6> OpenTTD: translators * r23685 /trunk/src/lang/ (norwegian_bokmal.txt swedish.txt vietnamese.txt welsh.txt):
18:45:36 <CIA-6> OpenTTD: -Update from WebTranslator v3.0:
18:45:36 <CIA-6> OpenTTD: norwegian_bokmal - 67 changes by mantaray
18:45:36 <CIA-6> OpenTTD: swedish - 25 changes by tobjork
18:45:36 <CIA-6> OpenTTD: vietnamese - 1 changes by nglekhoi
18:45:38 <CIA-6> OpenTTD: welsh - 50 changes by kazzie
18:45:52 <andythenorth> SmatZ: I have no intention of smooth rotation :P
18:46:00 <andythenorth> there are curvature props anyway. CETS uses them
18:51:18 <andythenorth> Cargo transit time, in +185 ticks
18:51:29 * andythenorth wonders which cargos catch fire if not shipped in time
18:52:13 <andythenorth> new industry: iceberg harvester
18:52:20 <andythenorth> new ship: iceberg tug
18:54:42 <bremerjoe> andythenorth: there are chemicals that need to be cooled as they would burn at regular outdoors temperature, at least my former gf told me so (she: Dipl. chem)
18:55:23 <bremerjoe> so if the dry ice used for cooling is used up after certain time the stuff would burn
18:59:34 * andythenorth ponders road vehicles that wheelie whilst overtaking
18:59:46 <andythenorth> whatever happened to newgrf smoke?
19:00:40 *** ZirconiumX has joined #openttd
19:01:26 * ZirconiumX is having issues with Squirrel
19:02:15 <glx> TrueBrain was faster than me :)
19:02:18 <ZirconiumX> AITown::GetTownCount() returns and int
19:02:51 <ZirconiumX> yet when I put it in a for() loop:
19:03:24 <ZirconiumX> for (int loop_a = 0; loop_a < AITown.GetTownCount(); loop_a++)
19:03:31 <bremerjoe> squirrel + nut = squirrel taking of with nut, so no more programming
19:03:38 <andythenorth> do while i < 1: i=0;
19:03:49 <ZirconiumX> Unless you are watching Ice Age
19:04:12 <TrueBrain> it mostly gives you a bit more, then a broken nut
19:04:36 <glx> one day you'll learn it :)
19:04:44 <ZirconiumX> Your script made an error: comparison between '0' and 'function'
19:05:04 <TrueBrain> ZirconiumX: odd. You sure you have the ()
19:05:07 <TrueBrain> anyway, better to use AITownList
19:05:55 <glx> for (local loop_a = 0; loop_a < AITown.GetTownCount; loop_a++) { // error
19:06:10 <TrueBrain> I love it when my first guess is right
19:06:19 <TrueBrain> too bad your first copy/paste did have the ()
19:06:31 <TrueBrain> but using AITownList will speed up your script by a lot
19:06:43 <Simon_> Whenever I select "Unload if accepted" the game deselects the order to unload.. net effect I'm only able to Unload All
19:06:55 <Simon_> am I missing something?
19:08:00 *** Chris_Booth has joined #openttd
19:08:18 <bremerjoe> Simon: that is normal for the game, nothing missed. If possible it will unload all
19:08:41 <bremerjoe> if cargo is not accepted at station it will leave again
19:08:58 *** andythenorth has joined #openttd
19:09:01 <Simon_> The game will unload cargo that isn't accepted at the station, then reload the cargo to the train
19:09:34 <Simon_> exactly, so does that drop down box not work?
19:11:09 <ZirconiumX> @TrueBrain - So squirrel represents an array by usng an int?
19:11:25 <TrueBrain> your script is missing a (), that is why it fails
19:11:29 <TrueBrain> so you can add a (), and it will work
19:11:35 <TrueBrain> but using for-loops like that is inefficient
19:11:41 <TrueBrain> using AITownList with foreach is much more efficient
19:12:02 <ZirconiumX> Ah - you didn't say that
19:12:11 <TrueBrain> I did, but that is not really important
19:12:28 <andythenorth> is it possible to write an unclosed for loop in squirrel?
19:12:45 <andythenorth> those tend to be performance sucks :P
19:17:51 <ZirconiumX> so foreach (val in townlist1) {
19:17:57 <ZirconiumX> or something like that
19:18:09 * ZirconiumX apologises for being noobish
19:19:37 <ZirconiumX> Also, Shouldn't there be a 1.1.4 API version?
19:22:44 * ZirconiumX realises no-one wants to talk to a noob like me
19:23:03 <bremerjoe> you do not want answers from a noob like me...
19:23:32 <ZirconiumX> Well, they are better than no answers at all
19:26:09 <bremerjoe> not really, just showing you that the channel is still being updated, so no internet connection issue or program freeze...
19:36:04 <bremerjoe> in spite of several users timeouts
19:37:50 <bremerjoe> btw: extra zoom levels in 1.20beta rock, finally able to really see the signals on 1920x1080 (24")
19:38:02 <bremerjoe> THX to all who made that possible!
19:46:54 <Elukka> yeah it's pretty great
19:52:09 <Belugas> yup, it's pretty and it's great
19:56:23 <bremerjoe> would not exactly call it pretty with regular 8bpp ;) but I am still grateful that it exists, makes some tasks much easier and when I want to enjoy the results I zoom out further anyway
20:00:37 <ZirconiumX> @encoded OpenTTD is in C and uses SDL or Allegro (or other stuff)
20:00:52 <ZirconiumX> But AFAIK it doesn't use GL
20:04:51 <TrueBrain> OpenTTD is very much in C++
20:05:47 <TrueBrain> and for Windows GDI is used, much faster than DX I can only imagine :)
20:07:33 * Belugas finds it darn pretty in 8bpp. been like that for so many times! 32bp is so... realistic . grumble grumble
20:09:46 <encoded> ok so does it use SDL or not?
20:10:08 <CIA-6> OpenTTD: michi_cc * r23686 /trunk/src/water_cmd.cpp: -Fix [FS#4921] (r23413): Infrastructure count of canals/locks/ship depots wasn't updated properly when a company went into bankruptcy or was taken over.
20:10:44 <ZirconiumX> Quote from configure
20:10:52 <TrueBrain> encoded: it is one of the possible video drivers
20:10:55 <ZirconiumX> checking for SDL: no
20:11:06 <ZirconiumX> checking for Allegro: no
20:11:08 <TrueBrain> encoded: it is a modular system; other video outputs can be used, depending on system (OS, libs, ..)
20:11:14 <encoded> ZirconiumX, you suck, you dont know what language its written in
20:11:38 <TrueBrain> for anything not Windows SDL is very common
20:11:42 <TrueBrain> for Windows it is almost always GDI
20:11:48 <TrueBrain> for DOS it has to be allegro
20:12:01 <TrueBrain> Windows can use SDL (not recommended)
20:12:02 <michi_cc> for OS X it is Cocoa
20:12:07 <TrueBrain> ah, good one michi_cc :)
20:12:15 <TrueBrain> I believe OSX can no longer use SDL :P
20:12:41 <michi_cc> It can compile with SDL, but reports say the result isn't very pretty.
20:12:53 <TrueBrain> SDL needs to be bootstrapped, which is silly at best :P
20:13:01 <TrueBrain> owh, I thought we just dropped the support; easier :P
20:13:27 <TrueBrain> but like using SDL on Windows: dont :D
20:13:48 <encoded> i wish Windows(R) windows worked like TTD windows
20:14:05 <TrueBrain> TT windows annoy the crap out of me :P
20:14:11 <encoded> they always know where to pop up
20:14:15 <TrueBrain> X on the wrong side ... OK/Cancel always reversed ... :P
20:14:53 <TrueBrain> well, 'wrong' and 'reversed' are of course up for debate, obviously :)
20:15:00 <ZirconiumX> @TrueBrain - that is a good feature - becuase you always click Yes accidentally
20:15:24 <ZirconiumX> Also, I note that a lot of smileys are being used :p
20:15:49 <encoded> i still dont understand wtf is up with graphics
20:16:06 <TrueBrain> encoded: turn them around?
20:16:12 <encoded> SDL can use gdi and opengl and dx in windows
20:16:54 <encoded> if youre using SDL for linux, why would it be bad to use it for windows?
20:17:11 <ZirconiumX> OMFG - Hal the computer is running an OpenTTD AI
20:17:15 <TrueBrain> encoded: such questions go so deep, I wonder if you have the time to sit down for it to listen to the complete answer :)
20:17:22 <TrueBrain> encoded: I wonder if I can do it off by replying: because it is faster
20:17:30 <TrueBrain> encoded: because it is faster to not use SDL
20:17:31 * ZirconiumX pictures the pixelated men running around
20:18:21 <encoded> but but but... Crysis and Skrim and assasins creed and such use DX
20:18:30 <TrueBrain> then again, they are 3D games
20:18:39 <TrueBrain> so it is like comparing a chair with a plane
20:18:52 <ZirconiumX> Windows GDI is direct to the screen - SDL has to communicate with Windows to show the picture
20:18:56 <TrueBrain> you don't see them using SDL, do you? :)
20:19:23 <encoded> ok, i'll investigate that SDL claim later
20:19:34 <ZirconiumX> GL is not much use either - IIRC you cannot create buttons
20:19:41 <TrueBrain> libSDL: if possible, avoid
20:20:01 <TrueBrain> ZirconiumX: lol; trolling much; that or you have no clue what you talk about :)
20:20:37 <ZirconiumX> I know what I'm talking about
20:20:44 <TrueBrain> you could have fooled me there :)
20:20:48 * ZirconiumX digs out SDL for Dummies
20:21:15 <TrueBrain> encoded: the only reason I guess we still use SDL for Linux, is that 90% of the systems have it, or can easily install it
20:21:33 <TrueBrain> I haven't found a (good) video library yet that does the same
20:22:05 <TrueBrain> SDL is nice and all, for a small project; to quickly make something. If you don't want to bother with performance too much (or if your application is much slower than your blitter)
20:22:18 <TrueBrain> it builds easy. Works the same for many platforms
20:22:20 <TrueBrain> very useful for that
20:23:06 <encoded> ok, relax you threw the ball to SDLs corner, i'll investigate them later
20:23:25 <TrueBrain> a few other problems exist with SDL
20:23:32 <TrueBrain> the ypromise us 2.0 for ... years? now
20:23:35 <andythenorth> ottd windows have the controls in the correct places
20:23:36 <encoded> i'll tell them openttd is talking shit about them ;p
20:23:43 <TrueBrain> I don't think anyone believes it will ever come out :P
20:23:50 <TrueBrain> can't remember which next version they promised :)
20:23:54 <TrueBrain> it should solve "all" problems
20:23:59 <TrueBrain> but for that it has to be released :D
20:24:36 <ZirconiumX> SDL is still in 1.0
20:24:50 <TrueBrain> and that is why we all use 1.2.14+ I guess
20:25:01 <TrueBrain> encoded: I wouldn't call it 'shit'. SDL is good for what it does
20:25:06 <TrueBrain> GDI just does a better job :)
20:25:13 <TrueBrain> (but only rusn on Windows)
20:25:18 <ZirconiumX> @TrueBrain Not fair - you've got a time machine!
20:26:13 <ZirconiumX> 1.3 is supposed to be the new version
20:26:23 <ZirconiumX> (still not released)
20:27:09 * ZirconiumX bursts out with laughter
20:27:13 <glx> 1.3 is dead since main dev is gone I think
20:27:21 <ZirconiumX> SDL 1.3 is currently under active development and beta is anticipated Summer 2011
20:28:14 <TrueBrain> encoded: in regards to DX, there is not really a performance gain to be expected by using a 3D API over a 2D API (which GDI is). Of course we can be wrong, not sure if anyone ever tried :)
20:28:39 <TrueBrain> we do have OpenGL implementations, but they have either never been stable, or very slow
20:28:51 <TrueBrain> (again, drawing 2D on a 3D canvas is a bit silly :D)
20:28:56 <ZirconiumX> @TrueBrain - 32,000bpp for DirectX?
20:29:27 <glx> <TrueBrain> it builds easy. Works the same for many platforms <-- even on OSX ?
20:29:50 <TrueBrain> it needs the silly bootstrap
20:29:52 <TrueBrain> but so does Windows :)
20:29:53 <encoded> ive been here a short time, but.. dunno... i find ZirconiumX extremly annoying
20:29:59 <glx> I remember the tweak for windows yes
20:30:19 <ZirconiumX> @encoded - you haven't seen anything yet.
20:30:24 <glx> but it's mainly because dune2 is timer based
20:30:32 <TrueBrain> hmm, I cannot find our benchmarks in regards to SDL and GDI. Sad :(
20:30:43 <TrueBrain> ZirconiumX: I would take encoded advise, and not see it as a challenge
20:32:11 <encoded> anyway im gonna try openttd a bit on ReactOS, if it uses gdi it might reveal a bug or 3!
20:32:13 <ZirconiumX> @Mek - That is very good
20:34:08 <Mek> :) it isn't very usable yet, not interaction possible etc
20:35:17 <TrueBrain> encoded: good luck ;) Let us know!
20:35:27 <TrueBrain> Mek: so that makes you alive, nice to know your trigger-words :P
20:35:54 <ZirconiumX> @Mek - who cares - judging by encoded's response you have a fan club dedicated to just looking at it
20:36:54 <Mek> yeah :) so far it is mostly a "I'm bored during the holidays, lets have some fun with opengl" project
20:38:10 <TrueBrain> encoded: yup; I am waiting for the day it is more commonly available :)
20:40:28 <encoded> i'll idle until i come back
20:40:53 <TrueBrain> oeh, so we can fill your backlog with highlights like thisone encoded
20:40:58 <TrueBrain> that can be very funny encoded
20:41:03 <TrueBrain> or better yet: I just encoded my lego pngs
20:41:12 <TrueBrain> and I wonder how to get my encoded pngs in a tar or grf
20:41:19 <TrueBrain> I read michi_cc was working on getting such encoded pngs there
20:41:26 <TrueBrain> would be quiet epic to have, dont you agree, encoded?
20:42:21 * ZirconiumX decides to have TrueBrain encoded into a .tar.gz file
20:42:48 <ZirconiumX> which is a tar file encoded with Gnu zip
20:43:04 * ZirconiumX bursts out with laughter
20:43:30 <ZirconiumX> wouldn't you agree that I am being a bit of a <encoded>
20:43:48 <ZirconiumX> At least I am not the only one
20:46:33 <Yexo> TrueBrain: that rts/test.html works fine for me (under linux)
20:46:43 <TrueBrain> Yexo: it works in special conditions :)
20:46:48 <TrueBrain> I believe Firefox + NVidia
20:47:01 <TrueBrain> either ATI or Nvidia is still blacklisted, can't really remember :P
20:47:10 <Yexo> I did test in firefox, and I do have NVidia :)
20:47:17 <TrueBrain> well, there you have it :)
20:47:19 *** Brianetta has joined #openttd
20:47:40 <TrueBrain> on Windows it works under "almost" all browsers/GPUs .. with the clear exception of Intel GM GPUs :P
20:47:44 <TrueBrain> those are not GPUs :)
20:47:47 <Yexo> about the for-loop before to iterate over towns: it's not only slower, it's (more importantly) completely wrong
20:48:12 *** Brianetta has joined #openttd
20:48:20 <TrueBrain> Yexo: lol; nowedays town array is not sequential, is it? :)
20:48:39 <TrueBrain> something I can never get used to :(
20:49:35 <TrueBrain> never? You sure? :D
20:55:25 <Belugas> y9u mean there are holes in it?
20:55:27 <TrueBrain> awh, SVN doesnt go back far enough :'( Sad sad panda! But okay, you are right :) I still somehow keep forgetting ;)
21:03:12 <andythenorth> when he's here, there's usually more arguising
21:03:38 <TrueBrain> people like him come and go. White noise, I guess.
21:04:01 *** Chris_Booth has joined #openttd
21:50:36 <eberon> hi everyone, I'm running 1.1.4 and I can't seem to assign a bus to one of my bus+train stations, no matter what part of it I click on
21:51:04 <eberon> is it maybe possible there's a square or two of road right inside/next to the city owned by a competitor that I'm not allowed to use?
21:51:25 <MNIM> eberon: are you sure it's a bus station and not a truck station?
21:52:09 <eberon> blah! it is a truck loading area
21:56:40 *** Phoenix_the_II has quit IRC
22:10:13 <eberon> another question -- is there any way to quickly tell elevation? I kind of hoped it would be in the 'about this square' tool
22:13:18 <eberon> owner, cost to clear, coordinates, and local authority is all I get =\
22:13:32 <Afdal> So can anyone explain to me why trains service for maintenance more often when you
22:13:44 <Afdal> have a path signal leading them to their depot?
22:14:32 <eberon> Belugas: thanks, actually, now that I think about it, the elevation is there, it's just the third coordinate
22:15:07 <eberon> Belugas: but I wouldn't have really thought about it hard enough to realize it until seeing that person's forum post confirming it's in there.
22:15:42 <Afdal> Someone tried explaining to be earlier that the way pathfinding works, it looks for a depot ahead of the reserved track
22:16:18 <Afdal> and a block signal before a two-way path signal allows a track to reserve the whole block between them
22:16:52 <Afdal> meaning there's more time to coincide with the tick where it looks for a depot, and more probability to service
22:17:01 <Afdal> But that doesn't explain a one-way path signal
22:17:42 <Afdal> And the service rate between a one-way and a two-way should be different, but I haven't found any significant difference
22:20:51 <Afdal> Did Belugas leave just after I typed that :(
22:30:07 <Aali> Afdal: do you have trains going through the depot who are not targeting it per se?
22:31:01 <Aali> might be that the penalties sometimes shift in favor of the depot track
22:31:35 <Aali> servicing without a service order was always unpredictable anyway
22:32:08 <Afdal> At least not on the path signalled example
22:32:12 <Afdal> I checked one time for that
22:32:21 <Afdal> they all said going to such-and-such for maintenance
22:33:34 <Aali> but, if anything, the PBS solution should make them service a tiny bit less
22:33:51 <Aali> since a train can block the path to the depot with its own reservation
22:38:49 *** bremerjoe has left #openttd
22:40:20 *** valhallasw has joined #openttd
22:45:37 <michi_cc> Afdal: A train searches a depot for automatic servicing in exactly two situations: a) When the pathfinder is called to extend the path or choose between tracks on a junction tile, and b) when the game time jumps to the next day.
22:47:35 <michi_cc> Afdal: This search is limited by how many tiles (or more precisely by pathfinder cost) the appropriate config variable defines. The search starts always at the last reserved tile furthest away from a path.
22:49:18 *** mahmoud has joined #openttd
22:49:55 <michi_cc> Afdal: In your example that means that in the first image, the depot search (if service is needed) is done either when entering the junction tile leading to the depot (and also starting at this tile) or whenever a new day rolls over, starting from the tile of the train engine.
22:51:37 <Afdal> you mean the last reserved tile furthest from the front of the train right?
22:51:52 <michi_cc> Afdal: In the second image, the new day search will start at the end of the reserved path, which means that for the area between the last and the second last signal before the junction tile the starting tile is the junction tile *and not the train's engine tile*, so the depot is closer pathfinding wise with path signals.
22:53:00 <Afdal> trains only reserve track on a block signal before a two-way path signal
22:53:08 <Afdal> My example uses a one-way path signal
22:54:39 <Afdal> Don't you mean the area *on* the junction tile?
22:56:08 <Afdal> Let me see if I'm understanding this correct: for the purposes of pathfinding, track reservation only counts entire tiles, and pathfinding happens BEFORE the actual track-specific reservation happens?
22:56:29 <michi_cc> For path signals pathfinding is initiated when entering the tile the signal is on.
22:57:24 <Afdal> So.... How is that different for the two signals
22:57:31 <michi_cc> For non-path signals nothing happens when the signal tile is entered.
22:57:34 <Afdal> I'm sorry if I'm still not understand this, it's confusing :(
22:58:02 <Afdal> Am I right about this then?
22:58:08 <Afdal> Let me see if I'm understanding this correct: for the purposes of pathfinding, track reservation only counts entire tiles, and pathfinding happens BEFORE the actual track-specific reservation happens?
22:58:44 <Afdal> that doesn't seem to make sense actually
22:59:49 <michi_cc> Track reservation doesn't count anything. Pathfinding happens if needed, and if the train is (or is entering) a signal block with path signals, the result of the pathfinding is marked as the reserved path.
23:00:47 <Afdal> So the path signal adds the length of its possible track reservation to the servicing pathfinding?
23:01:49 <Afdal> I still don't see how that would bias it to service more often
23:02:58 <michi_cc> The originating tile of the depot search is the last tile of the reserved path furthest away from the train engine. For a signal block not having any path signals, this is equivalent to the tile the engine itself is on.
23:04:41 <Afdal> Uh, so a train on the track with a path signal before the depot split will activate its pathfinding *sooner*?
23:04:53 <Afdal> If so, how would that make it service more often
23:06:00 <Afdal> Originally when this was explained to me I was thinking that the path signal gives more of a chance for a depot checking tick to coincide with the train while its on that path
23:06:21 <Afdal> twice as much time, in fact
23:06:52 <Afdal> But how does that work if a train reserves only one of the pathways anyway
23:07:37 <michi_cc> A path signal does not lead to more chances for finding a depot, it only changes the time/location the search might be done.
23:08:17 <Afdal> How does it end up with more servicing
23:08:45 <Afdal> I've tested this pretty thoroughly too,
23:08:50 <Eddi|zuHause> a path signal may prevent a depot visit in some cases where the distance between the signal and the depot is very high
23:09:05 <Afdal> the example with a path signal definitely results in significantly higher servicing
23:09:18 <Eddi|zuHause> because if a path is reserved beyond the depot, it cannot be changed to go into the depot afterwards
23:09:35 <michi_cc> Are you sure it actually does end up with more servicing, e.g. do the trains are actually in need of servicing when passing the last signal or passing the junction tile, respectively?
23:10:12 <Afdal> Yeah, an easy way to test this without counting is to place another depot right on the track after the first depot
23:10:50 <Afdal> well, I'll count it again just to be sure
23:11:15 <michi_cc> And, also important, how close are the trains on the track? Is it possible that the non-path signal is still red when a depot search (via the new day route) is initiated? A red non-path signals has a pathfinder penalty which a path signals has not.
23:14:22 <Afdal> the servicing rate with a path signal is nearly twice as high as with block or presignals
23:14:51 <Afdal> Is it possible that the non-path signal is still red when a depot search (via the new day route) is initiated? Which signal?
23:15:27 <Afdal> the signal at the split?
23:16:19 <Afdal> those depots are already at the maximum possible distance from their main track
23:16:38 <Afdal> If I separate them any further trains don't service at all
23:16:44 <Afdal> Because they can't find the depots
23:17:14 <Afdal> so the depot search must be happening right on or very close to the split
23:20:34 <Afdal> a successful depot search, that is
23:24:47 *** andythenorth has left #openttd
23:26:38 <Afdal> On the presignaled version,
23:27:15 <Afdal> if a train is making the exit signal on the main track red when another train gets to the junction, it will take the depot path regardless if it needs to service or not
23:27:30 <Afdal> Whereas with the path signaled version, it won't
23:27:50 <Afdal> And yet the path signaled version *still* has higher train to depot rates
23:28:51 <michi_cc> Savegame and which version?
23:30:01 <michi_cc> Yes. I don't remember any major changes between 1.1.4 and 1.2.0-beta1 that should affect depot searching.
23:32:44 <Afdal> It's better to test the disparity between servicing on those depots with the cyclotrons
23:33:06 <Afdal> to be absolutely sure there's no involvement with the pathfinding from those priority merges just ahead ont eh track
23:39:30 <Afdal> For example, over a 2 year period, for two of those depots
23:39:43 <Afdal> 14 trains in total will service
23:39:56 <Afdal> With a path signal in front, they'll service 27 times
23:40:04 *** FLHerne has joined #openttd
23:44:43 <Afdal> the presignals before a depot are a bad idea now that I think of it, because of this
23:44:45 <Afdal> Afdal On the presignaled version,
23:44:47 <Afdal> Afdal if a train is making the exit signal on the main track red when another train gets to the junction, it will take the depot path regardless if it needs to service or not
23:45:03 <Afdal> Even so, the servicing rate between block signaled and presignaled depots is roughly the same
23:45:17 <michi_cc> The answer is actually very easy, I'm just not sure if it really is intended behaviour or a bug.
23:46:22 <michi_cc> The check for serving on pathfinding is only done if a path reservation happens but not without path reservation. This means the non-path signal version has to rely purely on the "new day" check.
23:47:19 <Afdal> Also interesting is how that check was balanced so well with the new day check
23:48:11 <Afdal> So it has nothing to do with probability then?
23:48:37 <Afdal> the difference between the two, I mean
23:49:10 <Afdal> Are you an OpenTTD developer?
23:49:22 <Afdal> We should see if this is a bug or not
23:50:06 <michi_cc> The non-path signal variant searches for a depot (if the train actually needs service that is) whenever a new game day starts. The path signal variant does the search *additionally* when extending the reservation at a path signal.
23:51:02 <Afdal> Thanks a lot, so nice to finally understand this
23:51:02 <Yexo> I think the reason for that was that if you have a bit longer signal blocks a train with a path reservation would never find a depot since it would always have a too long path
23:51:27 <Yexo> the exception is when it extends the reservation, since at that point it can start searching from the head of the train instead of from the next signal
23:51:36 <Afdal> You might want to add it to both signal types then
23:51:51 <Afdal> Unless you're fine with path signals being the only one to use for servicing
23:51:54 <Aali> trains can still block the way to the depot with their own reservation
23:52:03 <Yexo> no, for block signals the train can change paths after going through the signal, that's a big difference
23:52:04 <Aali> and never find a depot even though they should be able to
23:52:25 <Yexo> Afdal: before we had path signals servicing already worked fine
23:52:38 <glx> for block signals pathfinding is done at each split
23:52:42 <michi_cc> It's not really a bug per se, because originally there was never a service check on pathfinding. The only problem is that I can't remember anymore (even if I wrote all that path signal crap) why I added the check only for path reservation and not for all cases.
23:52:57 <Aali> dont know how to solve *that* except allowing a train to change its reservation on the new day trigger
23:53:01 <Afdal> yeah, but path signals result in better servicing
23:53:15 <Afdal> so now that you have them, you should always use them for depots because of this
23:53:45 <Yexo> it really depends on the layout of your rails
23:54:01 <Yexo> if you have a signal rail with a depot connected directly to it, looking for it everyday is often enough
23:54:05 <michi_cc> But I guess I didn't add it for block signals it could be expensive if it is really called on each and every junction tile.
23:54:13 <Yexo> even with block signals a train will almost never miss the depot in that case
23:54:42 <Yexo> the "problem" occurs when you put your depots away from the mainline so the trains don't hold up traffic, in combination with a smallish maximum penalty to the depot
23:54:56 <Afdal> A Lev4 can easily go through 8 tiles (the default distance to check for a depot)
23:55:06 <Aali> service at nearest or service at X orders "fixed" this problem already :)
23:55:33 <Yexo> Afdal: are you sure it's 8 tiles? I thought it was 16 or so
23:55:46 <glx> yes it's way better to tell where you want to be serviced if needed
23:55:47 <Yexo> same problem will still happen for faster trains of course
23:56:17 <Afdal> I know because that's the maximum distance you can put your depot
23:56:18 <Yexo> but there is no perfect solution for that. Increasing the maximum search distance to depots can lead to trains going too far out of their way to go to a depot
23:56:21 <Afdal> away from the main track
23:56:31 <Afdal> any further than that and trains will never service period
23:56:56 <glx> prevents your trains to go to places from where they can't reach their normal schedule (on very bad layouts)
23:57:03 <Yexo> pf.yapf.maximum_go_to_depot_penalty = 20 * YAPF_TILE_LENGTH
23:57:21 <michi_cc> The default is 20 diagonal tracks, but e.g. curves and slopes also count towards that limit.
23:57:40 <Aali> does the depot itself count?
23:57:41 <Afdal> Then why can't trains find depots any further out than 8?
continue to next day ⏵
