IRC logs for #openttd on OFTC at 2011-12-30
⏴ go to previous day
00:01:15 <Afdal> If you want to play with realistic acceleration
00:01:29 <Afdal> distancing your depots from the main track
00:01:34 <Afdal> really helps prevent jams
00:01:45 <Afdal> since they have to slow down at depots
00:02:06 <Afdal> And when you distance them from the main track, you run into this problem more
00:02:09 <Aali> Afdal: seriously, use service orders
00:02:24 <Afdal> I don't know how to use servicing orders effectively
00:02:41 <Afdal> How do you know when exactly it's the best time to service
00:02:55 <Aali> service orders are skipped if the train doesn't need service
00:03:17 <Afdal> Will it still service somewhere else if it needs to?
00:03:22 <michi_cc> It is impossible to have defaults that fit each and every playing style, but you are of course free to increase the mentioned setting for your games.
00:03:23 <Aali> so just put it after the drop order (so it doesn't go to depot with cargo)
00:03:36 <Afdal> exactly, so how do you know
00:03:53 <Afdal> when its best to place a service order
00:04:13 <Aali> having a service order in the order list prevents the train from servicing on its own
00:04:15 <Afdal> Normally I space my depots about a full maximum screen zoom apart
00:04:52 <Afdal> Yeah, and then the train can go longer than it should without service, resulting in higher breakdown rates
00:05:13 <Aali> you must have some long routes
00:05:26 <Aali> could solve that with waypoints I guess
00:05:51 <Aali> station -> service -> waypoint -> service etc.. -> station
00:06:17 <Afdal> then how is that any different
00:06:23 <Afdal> than just letting them service on their own
00:06:30 <Afdal> when you select so many depots manually
00:06:48 <Yexo> if you give your trains a service order the distance doesn't matter
00:06:49 <michi_cc> Service orders do not care about the tile limit.
00:07:14 <glx> an the train goes to service only where you want
00:07:39 <Afdal> yeah and then when it passes that spot you wanted it to service, because it didn't need to at that time
00:07:47 <Afdal> you can end up with lower reliability
00:08:00 <Yexo> the same happens already, I don't see your point?
00:08:01 <Afdal> if you don't allow it to service at other depots too
00:08:10 <Afdal> Exactly, I don't see the point of servicing orders
00:08:12 <Yexo> yes, hence the waypoints and allowing it to service at other depots
00:08:27 <Afdal> Again, which is pretty much the same as just letting it service manually
00:08:37 <Afdal> If you're going to select a bunch of other depots
00:08:46 <Yexo> except that letting them service automatically you have a limit of tiles the depot can be away from your mainline
00:08:55 <Yexo> with service at depot orders you can circumvent that limit
00:09:11 <glx> no need to click on every depot
00:09:20 <Afdal> Well, that could be awfully tedious with big tracks
00:09:27 <Afdal> But I guess you only need to do it once
00:09:32 <Yexo> but trains don't need to service that often
00:09:40 <Afdal> setting up the order list, I mean
00:09:49 <Yexo> I mean, you don't need a depot every 20 tiles
00:09:55 <glx> and use shared orders too :)
00:10:12 <Yexo> or just play without breakdowns and without servicing :p
00:10:15 <Afdal> Do you really need waypoints?
00:10:19 <Afdal> I don't see how that would help
00:10:23 <Aali> default breakdowns are kind-of a big deal though :)
00:10:35 <Aali> never managed to go beyond ~100 trains with them
00:10:40 <Afdal> I like the extra challenge servicing brings to the game
00:10:45 <Yexo> the "service" orders are evaluated at the time the train leaves the station
00:10:50 <Afdal> Gotta figure out ways to make the best depots, like now :3
00:11:06 <Afdal> Well the "normal" breakdown rate is horrendous, true
00:11:07 <Aali> the improved breakdowns patch made it sane
00:11:14 <Aali> well, a bit more sane anyway
00:11:14 <Afdal> I don't find Reduced that big a deal though
00:11:40 <Afdal> Are you guys ever planning on adding that patch to OpenTTD?
00:11:47 <Afdal> as an advanced setting?
00:11:58 <Afdal> I've never played with a patch before
00:12:29 <Aali> it was pretty dead back when I used it.. which was like.. years ago
00:12:40 <Aali> so no, I dont think that particular patch has any chance
00:13:17 <Afdal> What was whoever saying about waypoints in addition to the goto depot order now?
00:13:22 <Afdal> Why would you need waypoints?
00:13:46 <Afdal> Yexo the "service" orders are evaluated at the time the train leaves the station
00:13:51 <Afdal> Oh, does that apply to waypoints too
00:14:36 <Afdal> But wait, what you end up doing with service orders then
00:14:44 <Afdal> is spacing out the possible check
00:15:07 <Afdal> so it's possible for it to be less efficient, even though the check will always occur
00:15:15 <FLHerne> Can anyone suggest any compile options to cut CPU load ingame?
00:15:25 <Afdal> I guess you could just put the waypoints directly before the depot splits
00:16:02 <FLHerne> This 133MHz processor is irritating, and the game lags a bit...
00:18:28 <Afdal> only problem with this fix is you can't put waypoints on diagonal track
00:18:53 <FLHerne> Is that a compile option? I thought there was a setting in the .cfg for that?
00:20:26 <FLHerne> i.e. max ships = 0 or however it's phrased
00:21:02 <michi_cc> It is a setting for the person sitting in front of the keyboard.
00:21:52 <FLHerne> I am sitting in front of the keyboard, so that's OK
00:22:05 <michi_cc> It means, just don't build any.
00:23:15 <michi_cc> I don't know what OS/hardware you use, but you could test whether the 8bpp-optimized or the 32bpp-optimized blitter is faster (32bpp-anim will very likly be the slowest).
00:24:10 <FLHerne> 8bpp is faster, I tried that on my last build
00:24:48 <FLHerne> This time I'm building without networking or sound support, to try and get a couple more FPS
00:25:13 <FLHerne> I was just wondering what else I could disable :-)
00:26:15 <michi_cc> Networking will likely have no effect, and there's no need to disable sound support, just pass '-s null' and '-m null' on the command line to get the 'do nothing' sound and music driver.
00:29:37 <FLHerne> I'll start, compiling then, hopefully it'll finish before I get up...took about 6 hrs last time...
00:30:54 <Afdal> So does filling your track with path signals
00:31:11 <Afdal> Because trains are constantly checking if they have to service?
00:31:17 <Snail_> hi, I have a question probably for OTTD developers
00:31:48 <Snail_> in the "railtypes" feature, when drawing tunnels, the tunnel head is always the original one
00:32:30 <Snail_> this keeps compatibility with different landscapes, but limits the newGRF authors as to drawing different tunnel entrances across different types
00:33:16 <Snail_> would it be possible to add an option to replace the original tunnel head with a custom sprite provided by the railtype newGRF? of course, this custom sprite would need to be compatible with the landscapes...
00:34:36 <Afdal> Actually... If a path signal always ensures a depot check anyway, how is putting maintenance in orders along with waypoints any different?
00:35:40 <michi_cc> Snail_: Just Action A the original tunnel sprite. It is not possible for a NewGRF to query the base set in use though.
00:36:34 <Snail_> michi_cc: you mean it's not possible for a newGRF to figure out which landscape the game is in?
00:38:08 <michi_cc> You can query for temperate/arctic/tropical etc, but not for the base set. If you could, you'd have an instant desync if both players with the original base set and players with OpenGFX join the same multiplayer game.
00:39:46 <Snail_> I see... so if I decided to "action A" the original tunnel sprite with a different one based on the base set, someone playing with OpenGFX would see that sprite being different from the rest of the landscape?
00:40:21 <michi_cc> Yes. And GRF parameters work for single player, but still fail for multiplayer.
00:40:50 <michi_cc> All that is the reason why you can't provide a custom tunnel sprite in the railtype definition.
00:41:54 <Afdal> If a path signal always ensures a depot check anyway, how is putting maintenance in orders along with waypoints any different? Doesn't seem to make any difference, just tested it on an old save game.
00:42:28 <michi_cc> Maintenance orders will always find the depot, regardless of the length of the track leading to thedepot.
00:42:57 <Afdal> So maybe that would be better if you were using 8 tile+ length trains
00:43:08 <Afdal> But if you're not it really makes no difference
00:44:03 <Afdal> err not "maybe", it would be
00:44:59 <Afdal> So does filling your track with path signals over block signals eat up more CPU because trains are constantly performing depot checks??
00:45:12 <Snail_> michi_cc: so I can't provide a custom tunnel sprite in the railtype definition, but I can still change it through action A, right? so what's the difference?
00:45:35 <michi_cc> You get *one* custom sprite, not one per railtype.
00:46:28 <Snail_> all the railtypes will have the same then
00:47:23 <michi_cc> But even with Action A, as soon as your sprite contains even a bit of grass it will always be a mismatch for about half the players (depending on which base set's grass you take).
00:48:38 <Snail_> how about making tunnel entrances as sprites that only contain the entrance and no landscape? a little bit like the track sprites, which contain only the tracks and the ballast underneath, but no landscape...
00:49:34 <michi_cc> The only sane way IMHO is to introduce a parameter for the tunnel sprite that by default (for multiplayer compatibility) is set to "use base set tunnel sprite".
00:50:08 <michi_cc> The grass would overlap with the tracks and the train because you'd need to know where to cut a "hole" into the sprite.
00:51:38 <Snail_> but the hole could remain the default one in terms of size... so OTTD would still know where to cut
00:52:47 <Snail_> re. the parameter, yes I agree, but still it would be awkward to draw a custom tunnel head that would suit both XIXth century NG rails and modern TGV tracks
00:55:09 *** JVassie_ has joined #openttd
00:56:29 <michi_cc> Someone™ would need to draw all needed sprites for the original base set.
00:57:00 <michi_cc> (I.e. all landscape types in all orientations.)
00:58:30 <Snail_> you mean if one wanted to create custom tunnel heads?
00:58:55 <Snail_> as a matter of fact I already drew some, for my NG tracks
01:00:08 <Afdal> Does filling your track with path signals over block signals eat up more CPU because trains are constantly performing depot checks?
01:05:01 <michi_cc> Snail_: I mean grass sprites with a fitting hole.
01:05:09 <Afdal> Last question I swear :3
01:06:09 <michi_cc> Afdal: The depot check is only expensive if it decides to search for a depot. As long as the train does not need servicing, it will be very quick.
01:06:28 <Afdal> ah, but it will use more CPU then won't it
01:06:38 <Afdal> just less if you have your depots spaced properly
01:06:49 <Afdal> err, less the less space you have between them
01:08:17 <Snail_> but the grass could be that of the default sprite, on which the new head would be the overlay...
01:08:33 <planetmaker> that's difficult.
01:08:45 <planetmaker> tunnel heads most likely have different slopes than the tile slopes
01:08:49 <Snail_> just like in the normal tracks... the grass would be around the tunnel head, but the tunnel head itself would be replaced
01:08:57 <planetmaker> just look at the ttd tunnels - they have grass on top
01:09:09 <planetmaker> thus it looks awkward then, too
01:09:25 <michi_cc> I doubt you can even measure the extra check for the no service case. And do remember that only a very limited amount of tiles is searched for a depot, so a depot search is still cheap in comparison to a full pathfind at a junction tile.
01:09:49 <Snail_> yes, perhaps we could restrict such an overlay to approximately follow the original head. I could live with that
01:10:01 <michi_cc> Snail_: It would need grass tiles with a proper transparent area, otherwise the grass would overlap the vehicles.
01:10:28 <Afdal> Yeah but if a train needs to service
01:10:44 <Afdal> It's going to check every 2 tiles for a depot when path signals
01:11:08 <Afdal> and with a lot of trains and a lot of track, that could add up
01:12:04 <planetmaker> thus it would need terrain sprites for all 4 slopes, one for each of the track sides. Thus two per each of the slopes and terrain type. For both base sets
01:12:06 <michi_cc> So more or less a tile that is all grass and has the same shape and outline as the default tunnel portal. If a set wants to provide a tunnel portal wider on the inside, it would have to hide the extra grass with some kind of darkness.
01:12:07 <planetmaker> then it *might* work
01:12:45 <Snail_> planetmaker: that would be great
01:12:57 <planetmaker> well. Go ahead and create those sprites
01:13:02 <planetmaker> would be a first step ;-)
01:13:14 <planetmaker> Please do for both basesets
01:13:20 <michi_cc> Or maybe where the portal hole is a bit bigger, under the assumption that no one would like to draw a tunnel portal just a pixel "thick".
01:13:58 <planetmaker> still, the problem is where to place the hole
01:14:11 <planetmaker> now the entry portal can be anywhere on the tile
01:14:18 <Snail_> I already have a set of custom tunnelheads
01:14:23 <Snail_> you want me to PM it to you?
01:14:28 <planetmaker> Snail_: not custom tunnel heads
01:14:31 <planetmaker> the grass sprites
01:14:56 <planetmaker> where stuff can be overlayed onto
01:15:46 <planetmaker> I wonder though... how much it needs on the top side of the slopes...
01:15:51 <planetmaker> or whether at all
01:15:54 <michi_cc> planetmaker: Totally arbitrary tunnel portals wouldn't work, graphics artists would just have to draw sprites that fit onto the default hole (which would roughly the shape and size of the default hole).
01:16:29 <planetmaker> hm.... I'd make that bigger by quite a bit
01:17:45 <michi_cc> Some, but a bigger hole means the portal itself has to be more massive as well.
01:18:12 <planetmaker> hm... would it work to always draw the default background, then a custom background on top, then the train, the default top and then a custom top?
01:18:25 <planetmaker> that'd not allow bigger portals, though
01:18:35 <planetmaker> thus it needs a bigger portal
01:18:58 <michi_cc> Oh, and the tunnel sprite actually consists of two parts, one drawn behind and one drawn over the vehicle sprite (Base set sprites 2365 and following).
01:19:43 <planetmaker> a back where the track and the train is drawn onto. And then overpainted by the front
01:20:12 <planetmaker> where the front must include the top
01:20:47 <michi_cc> planetmaker: The default tunnel sprites have shading that suggests the roundness of the tunnel tube, so that wouldn't really work as a base sprite.
01:21:59 <planetmaker> I know. And they even have no grass. Depending on track type
01:23:21 <michi_cc> Anyway, that results in 48 sprites per base set to be drawn. Get to work, someone™ :)
01:23:31 <planetmaker> Thus what it needs is like the untouched part of the tunnel sprites
01:23:39 <planetmaker> And the player needs to supply the additional back part and the additional front / top part
01:24:11 <planetmaker> @calc 2 * 4 * (4+2)
01:24:31 <planetmaker> but I see no other way to implement it otherwise
01:24:48 <planetmaker> and it's not like it's not feasible. But someone [TM]
01:26:08 <Snail_> planetmaker: I'm willing to give it a try if you need "someone" to draw ;)
01:26:18 <Snail_> I can use the sprites I already drew as base
01:26:44 <Snail_> do I need to break them in a certain way? or can I just send you the final result and you can try and break it up?
01:27:02 <planetmaker> Snail_: mind, it needs those sprite parts which are NOT the tunnel of the tunnel tile
01:27:21 <planetmaker> and yes, I'll need the two parts separately. I don't need finished tunnels
01:27:52 <planetmaker> as that breaking up is IMHO the hard part. After all I don't want new tunnels. But the default tunnels properly split up into parts that the actual portal can be replaced by a custom one
01:28:31 <planetmaker> i.e.: no custom portals. But the existing tunnels broken into "back grass", "portal" and "front + top grass"
01:28:41 <planetmaker> hm... no, "front grass"
01:28:49 <planetmaker> portal was extra and includes top grass
01:30:56 <planetmaker> let me best make a dirty example...
01:30:58 <Snail_> you only need the default parts, so that a custom portal can later replace the default one
01:31:11 <Afdal> You guys know your wiki is kinda messed up right now, right?
01:31:22 <Afdal> It hasn't been displaying pages properly the last several days
01:31:23 <planetmaker> yes. For this to work we first need to change the default to a way that it can be modified
01:31:45 <planetmaker> Afdal: works for me.
01:31:59 <planetmaker> so: no, I don't know nor experience troubles
01:32:04 <Afdal> Well I'm not blocking any ads or scripts
01:32:18 <Afdal> And openttd's wiki is the only wiki I go to that's not displaying correctly
01:32:18 <planetmaker> there are no ads except a link to our sponsor
01:33:00 <Snail_> planetmaker: ok. Will it work if I take a screenshot from a game and then split up the tunnel from there?
01:33:09 <Afdal> disabling Adblock fixed it
01:33:19 <Afdal> I wonder what I'm blocking
01:33:46 <Afdal> No wait, it fixed itself just now
01:33:49 <Afdal> without my doing anything
01:33:50 <michi_cc> Snail_: Decode trg1r.grf and look at sprites 2365+.
01:33:55 <Afdal> Weird, I wonder what the problem was
01:35:58 <michi_cc> And also the appropriate plain slope sprites, but they are distributed over all base GRFs.
01:36:02 <Eddi|zuHause> has the railtype yet an option to choose whether to use the rail/monorail/maglev tunnel?
01:36:48 <Snail_> those sprites are split already, but we'll need to split them even further, right? the tunnel head needs to be separated from the grass above the tunnel?
01:38:04 <planetmaker> I'd split it into three parts... then the actual normally sloped grass part can be made a bit bigger to allow for more custom tunnel portals
01:38:20 <Eddi|zuHause> it needs 4 parts
01:38:34 <Eddi|zuHause> ground-front, ground-back, entrance-front, entrance-back
01:40:03 <planetmaker> ground and back are the same
01:40:29 <Eddi|zuHause> "ground" meaning the grass, and "entrance" meaning the wall/roof
01:40:38 <michi_cc> The base set would only need the grass though, if no custom tunnel sprites are defined for a railtype the code would just fallback to drawing the default tunnels like now.
01:41:51 <michi_cc> The railtype NewGRF would then supply only the back and front sprites with the portal itself. And it does *not* need an extra back part, because that can be attached to the tunnel track sprite that is already used.
01:41:53 <Eddi|zuHause> i believe the 32bpp-ez tunnel entrance is cut wrong...
01:42:37 <michi_cc> So a NewGRF would supply one back sprite with rails + back part of tunnel portal and a new sprite with the front portal overlay.
01:42:46 <Snail_> s o, the sprites where the tunnel head is facing the viewer need to be in three parts: track underlay, portal overlay, grass overlay?
01:44:08 <michi_cc> The base set needs sprites like the two I linked (just tweaked a bit better to allow as many different portal outlines as possible), just in all four directions for all 6 different grass types.
01:44:39 <Snail_> any link to the other grass types?
01:45:02 <michi_cc> Basically I don't know if a bigger or a smaller cutout is better, simply because I don't draw tunnel portals :)
01:46:12 <Snail_> hehe, the only ones I drew so far have a smaller cutout, but a bigger portal itself (NG has smaller loading gauge)
01:46:13 <michi_cc> For the original TTD sprites you need to decode all trg*.grf and search the slopes in there, for OpenGFX there is probably also some nice page with them.
01:47:10 <planetmaker> there's a gimp file with all terrain sprites for OpenGFX
01:47:29 <michi_cc> And just a note on my examples: The sprites are very crude combinations of the tunnel sprites with the portal erased and the slope grass to fill them up.
01:48:21 <planetmaker> but same as michi, I'm not sure on the exact way a good cut would need to be
01:48:38 <michi_cc> planetmaker: One thing they do need to show is some shading to indicate the tunnel "bulb", because if you really just use the slope it would look weired.
01:48:54 <michi_cc> Oh, and the cutout needs the identical between TTD and OpenGFX of course :)
01:49:17 <planetmaker> probably that's best...
01:49:27 <planetmaker> doesn't make it easier, though
01:49:36 <Snail_> I guess that if one supports custom portals, the grass over the tunnel would be drawn *before* the custom portal itself, right?
01:50:02 <Snail_> so that if one supplied a larger portal (like the massive ones in the XIX century) with lots of masonry around the cutout, it would be drawn over the grass
01:50:15 <michi_cc> Yeah, drawing order would be back grass, then tracks with back portal, front grass, front portal overlay.
01:50:18 <planetmaker> michi_cc: but showing the "bulb" implies already a terrain-covered thing... though... maybe not, as it can be overdrawn. So you might be right that it's a good thing to have
01:50:30 <Snail_> michi_cc: sounds great
01:51:04 <michi_cc> planetmaker: The point is that you can overdraw/hide the bulb, but you couldn't add it in the portal overlay to match the grass.
01:52:04 <michi_cc> So the shading needs to be just so to not look off. Quite a challenge to draw.
01:53:13 <Snail_> what do you mean by tunnel bulb? the part of the slope that is actually horizontal coz there's the tunnel underneath?
01:53:29 <Snail_> like, the semicircle emerging from the slope?
01:54:31 <michi_cc> The right sprite in my example has a lighter area at the top which indicates that the area in the middle isn't sloping like left and right.
01:56:59 <Snail_> you mean the "bulb" has to be separated? why can't it be drawn in the same sprite as the slope?
01:57:23 <michi_cc> I mean it has to be present so you can't just take the plain sloped grass.
01:58:07 <Snail_> anyway I did a quick example for one orientation only. Can I show you guys so that I see if I'm on the right track?
02:03:37 <michi_cc> Snail_: You are allowed to, but you have to decide for yourself if you can :p
02:04:08 <Eddi|zuHause> typical german answer :p
02:04:18 <Snail_> I meant across irc (I'm not an expert of this)
02:04:33 <planetmaker> just post a link to what you mean ... ?
02:04:58 <planetmaker> I mean... we did, why wouldn't you be allowed to?
02:05:15 <Snail_> ok, well I have no space to upload it to
02:05:20 <Snail_> I'll just PM it to you guys
02:05:38 <planetmaker> just use imagebin as I used
02:08:28 <Snail_> thanks for the tip about imagebin :)
02:09:05 <planetmaker> well. that back part of the grass is not large enough imho
02:09:22 <planetmaker> it should not contain the cut-out of the "bulb", I guess...
02:10:06 <planetmaker> you might not need create the portals itself. That'd only be needed for an actual railtype NewGRF
02:11:07 <planetmaker> And the ground/back (with tracks)...not sure the walls are needed. Just the ground
02:11:47 <Snail_> but perhaps new walls might be supplied by the newGRF
02:12:05 <michi_cc> Snail_: The back grass part should also have grass in the area where the tracks are to allow tracks that are smaller than default.
02:12:12 <Snail_> for instance, a portal made of bricks would have reddish walls, while one made of stone would have greyish walls
02:12:13 <planetmaker> yes. But this task is not a newgrf task, but what the base set must supply
02:12:49 <planetmaker> i.e. those parts which the newgrf must not / need not include
02:12:52 <Snail_> michi_cc: got it, you have a point
02:13:13 <Eddi|zuHause> planetmaker: btw, do you have anything to say on the report i linked earlier where people have problems with the opengfx signals?
02:13:41 <Snail_> planetmaker: ok... well, ideally a newGRF should either provide its own walls, or no walls at all, and then the custom walls would be used
02:13:46 <planetmaker> I'm not aware of any signal issue and OpenGFX has signals of both types, Eddi|zuHause
02:13:52 <michi_cc> And I guess the cut-out of the left-most sprite should be a bit bigger as otherwise it would be impossible to have portals that are "bigger" than default on the inside.
02:14:04 <planetmaker> though I never play with British, thus it's long ago I checked
02:14:34 <Eddi|zuHause> planetmaker: as far as i read out of it, opengfx shows british semaphore signals regardless of the traffic-side setting
02:15:10 <michi_cc> Snail_: A NewGRF would provide either no tunnel sprites at all, in which case the current (uncut) tunnel sprites are shown, or it would provide the two portal sprites (i.e. second-left and right sprite in your example).
02:15:15 <planetmaker> I don't play with semaphore either. So... don't know currently either
02:15:33 <planetmaker> but I shall check. But not before sleep
02:15:48 <planetmaker> but what's a "british" semaphore and what a German, Eddi|zuHause?
02:16:04 <Snail_> michi_cc: ahh I understand. So those "cut" sprites (1st and 3rd column in my drawing) would *only* be used if a newGRF provided custom portals
02:16:35 <planetmaker> Snail_: and those sprites would have to be supplied by that very newgrf
02:16:35 <Eddi|zuHause> planetmaker: british semaphores point to the left if red, and downwards if green, german semaphores point to the right if red and upwards if green
02:16:41 <Snail_> therefore walls would not be needed, because the newGRF would need to provide them. Same for the portals themselves. So all I need to draw are my 1st and 3rd columns, perhaps with a larger bulb to allow for larger tunnel heads?
02:17:02 <Eddi|zuHause> planetmaker: which is problematic alone for the fact that they're pointing into the track if they are drawn on the right side
02:17:26 <michi_cc> Which is why the cut-out needs to have a good shape (I don't know if good here means large or small, you'd probably have to test that with your french portals) to allow the most flexibility for NewGRFs.
02:18:05 <planetmaker> Eddi|zuHause: at least there are sprites of both types
02:18:26 <Snail_> mich_cc: yes. Well, we would need a "one size fits all", both for NG and SG. I'm not gonna draw two bulb sizes :D
02:18:37 <planetmaker> in sprites/png/infrastructure/semaphores.png
02:18:47 <Eddi|zuHause> planetmaker: i don't have opengfx
02:19:00 <Snail_> perhaps I'll just enlarge this a little bit. Would look a little bit weird for NG. But considering the vast majority of sets out there uses SG, the latter needs priority
02:19:03 <michi_cc> Snail_: Yes. And that is the difficult part of the job :)
02:19:46 <planetmaker> Eddi|zuHause: whatever you use, it's an easy thing to check as you can switch base sets easily
02:19:54 <planetmaker> if you know what to look for
02:19:56 <Snail_> the grass in the underlay is not a joke too, cz it needs to be shaded (a little bit darker)
02:20:05 <michi_cc> A bigger cut-out doesn't necessarily mean a big tunnel inside though, just that the portal would have more walls/decoration to cover the space.
02:20:08 <Eddi|zuHause> planetmaker: and the bandwith to get it and keep it up to date?
02:20:37 <planetmaker> all those wasted bits. I'm really sorry
02:20:45 <Snail_> michi_cc: yes it's true. I'll do some tests
02:20:55 <Eddi|zuHause> i'm also really sorry that i'm living in the most underpriviliged village around
02:21:17 <Eddi|zuHause> but seriously, my whole garden hose bandwidth is already used for the whole day now
02:21:52 <planetmaker> last time I heart you talk about bandwidths it was in terms of movies / hour ;-)
02:22:43 <Eddi|zuHause> the last time i spoke about bandwidth was that it doesn't suffice for viewing a low-res livestream like the openttd-yogscast
02:23:24 <planetmaker> which still is a lot compared to 3MB for opengfx
02:23:44 <Eddi|zuHause> which is 3MB of something else i won't get instead
02:23:47 <Snail_> I think this encompasses all the climates for OpenGFX. It shows the final result of how my NG tunnels should look like. Are those all the openGFX climates I need?
02:25:21 <planetmaker> a base set is a base set and all need to supply exactly the same sprites
02:25:37 <Eddi|zuHause> the sprites look like the wall is standing in the middle of nowhere
02:26:13 <planetmaker> anyway, I'm off to bed. Have a good night, folks
02:27:09 <Snail_> Eddi|zuHause: yep because I didn't edit it much
02:28:13 <Snail_> and that would be the inconveniences of a "one size fits all" base I guess... it needs to accommodate both old-style tunnel heads like the one I drew (which includes a big wall) as well as more modern heads without a wall around the opening
02:29:11 <planetmaker> you start to understand why it wasn't (yet) done ;-)
02:30:43 <Snail_> anyway I think I found the toyland sprites, I've got a file with all the openGFX infrastructure
02:31:06 <Snail_> I'll just need to decode the original sprites and give it a try... I still think it's gonna be better than nothing
02:34:31 *** Biolunar_ has joined #openttd
03:00:10 <Snail_> hmm, I just decoded trg1r.grf, I could find the sprites for the normal and snow-covered tunnels, but not the others (desert and toyland)
03:00:26 <Snail_> perhaps they're in another *.GRF file?
03:13:13 <Snail_> nevermind I found them
03:20:23 *** adamkex has joined #openttd
03:44:58 <Eddi|zuHause> yes, the climates are in the other grfs
03:48:45 <adamkex> Can somebody help me explain how a a feeder service with busses and trains work? x1 is a trainstation outside of city x. x2 is a busstation inside city x. y1 is a trainstation outside of city y. y2 is a bustation inside city y. What I want is that passengers travel by bus from x2 to x1, then by train from x1 to y1, and then by bus from y1 to y2. i seem to have set this up but my trains seem to be loading the same passengers as it unloade
03:52:06 <Eddi|zuHause> transfer only works one-way
03:52:39 <Eddi|zuHause> pr you need really complex systems with more than one station
03:53:05 <Eddi|zuHause> one station only handling incoming passengers, the other one only handling outgoing passengers
03:55:02 <adamkex> so my best bet is just to place trainstations very close to the city?
03:55:58 <Eddi|zuHause> yes, at least close enough so the trains can unload without transfer
03:56:37 <Eddi|zuHause> and use the trams to haul passengers to the stations with transfer and leave empty, so more pople use the long distance trains
04:02:52 <adamkex> Eddi|zuHause: but how would the arriving passengers enter the city if all busses are set so they leave the train station empty?
04:55:55 *** fjb|tab is now known as Guest22163
04:55:56 *** fjb|tab has joined #openttd
05:56:18 *** Eddi|zuHause has joined #openttd
07:25:23 *** Twofish has joined #openttd
07:48:43 *** encoded has joined #openttd
07:50:41 *** Alberth has joined #openttd
07:50:41 *** ChanServ sets mode: +o Alberth
08:04:10 <encoded> can i just unzip the nightly build over an installed stable?
08:10:31 <Alberth> sounds like a good way to make a big mess of your file system to me :p
08:16:51 <encoded> its in D:\games\openttd not program files if thats what you mean
08:18:08 <encoded> can someone explain how to get 2 trains in 1 track? i never did it like that in original TTD
08:18:20 <Alberth> no, I meant that now you have no way of knowing which file is used and which one is not.
08:20:05 <Alberth> there is also a Signals wiki page, which you may read after the tutorial
08:23:50 <planetmaker> encoded: the best choice is to not unpack *over* another installed version, but just unpack the game into a new directory
08:26:57 *** DayDreamer has joined #openttd
08:38:25 <dihedral> Eddi|zuHause, what about LTE? did you check the coverage of that?
08:43:42 *** sla_ro|master has joined #openttd
08:56:08 *** Devroush has joined #openttd
09:02:12 *** TGYoshi has joined #openttd
09:18:02 <Eddi|zuHause> dihedral: i did, but they limit bandwidth to what i have right now after ridiculously low amounts of traffic
09:21:58 *** valhallasw has joined #openttd
09:26:11 *** |Jeroen| has joined #openttd
09:27:31 *** Progman has joined #openttd
09:43:49 *** Brianetta has joined #openttd
09:50:05 <dihedral> Eddi|zuHause, that is a sad story indeed
09:50:42 <dihedral> i just got myself internet from "kabelbw"
09:51:05 <dihedral> as the standard dsl line would not go beyond 2Mbit/s
09:57:14 *** andythenorth has joined #openttd
09:58:17 *** TWerkhoven has joined #openttd
09:59:39 <encoded> wtf are you guys doing up?
10:00:32 <andythenorth> there speaks a man without children
10:00:35 <dihedral> it is 6am in YOUR timezone
10:01:23 <Terkhen> it's 11 AM here, and I'm busy enjoying my holidays :)
10:03:21 * Alberth wonders how you can be busy doing that :p
10:05:22 <Terkhen> being with the family, slacking and playing games take all of my time :P
10:07:15 <Terkhen> it's almost time to go back, though
10:11:12 *** DayDreamer has joined #openttd
10:50:47 *** fjb|tab has joined #openttd
10:56:21 <andythenorth> there's one a bit like that already in the game
10:56:40 <andythenorth> those are standard gauge
10:58:38 <Alberth> likely they picked whatever was easily available
11:00:59 *** TWerkhoven2[l] has joined #openttd
11:01:50 * andythenorth ponders industrial rail vehicles
11:01:53 <andythenorth> maybe a railtype
11:02:06 <andythenorth> meh - there would have to be signals and things
11:02:20 *** Chris_Booth has joined #openttd
11:11:36 <Alberth> man included at the front :p
11:20:38 *** mahmoud has joined #openttd
11:25:13 *** TWerkhoven2[l] has quit IRC
11:27:11 *** JVassie_ has joined #openttd
11:27:20 *** KenjiE20 has joined #openttd
11:48:53 *** tokai|noir has joined #openttd
11:55:50 *** fjb|tab is now known as Guest22186
11:55:50 *** fjb|tab has joined #openttd
12:01:40 *** Devedse has joined #openttd
12:03:12 *** TWerkhoven has joined #openttd
12:07:30 <CIA-6> OpenTTD: rubidium * r23687 /trunk/readme.txt: -Fix [FS#4922]: document where the game script directory is
12:16:39 *** lordnokon has joined #openttd
12:18:13 <lordnokon> i need a bit of help beter understanding min profit under the detailed performance rating
12:18:35 <lordnokon> i do i get a 100% rating on it?
12:18:57 *** DayDreamer has joined #openttd
12:19:23 *** DayDreamer has left #openttd
12:19:47 *** vargadanis has joined #openttd
12:21:26 <vargadanis> it looks to me that from the depo there should be a free way to the 3rd and 4th track of the station yet the trains wouldn't go
12:21:30 <TWerkhoven> you mean the transparancy?
12:21:33 <vargadanis> any tips why that could be?
12:21:39 <TWerkhoven> or waiting for free path
12:21:45 <vargadanis> waiting the free path
12:22:17 <Rubidium> missing electrification
12:22:17 <vargadanis> I use the path signals and one way path signals for this station
12:22:41 <Rubidium> that's even clear from this image
12:23:08 <vargadanis> after I start the game it seems that it defaults back to the non electric one
12:23:17 <vargadanis> I'll pay attention when loading the game
12:23:35 *** LordPixaII has joined #openttd
12:24:12 <Rubidium> then set that setting for default railtype to "most used"
12:25:33 <vargadanis> that would be a smart thing to do, wouldn't it? So I cannot do that - smart solutions are not my thing in TTD :) hehe
12:26:27 <Alberth> you only need more practice :p
12:28:28 <Alberth> (we have many lords here nowadays)
12:32:22 <lordnokon> i've read that before, but still sit on 0/100
12:32:37 <lordnokon> even thou i have no vechile making negative income
12:34:29 <Alberth> so the lowest profit is too low
12:42:45 <lordnokon> al my income are about 100 000ponds
12:43:20 <lordnokon> per train, ship, aircraft of verchile
12:44:25 *** TWerkhoven2 has joined #openttd
12:53:10 <Alberth> I don't know then. Perhaps you can post your save game at the forum and ask it there?
13:00:35 *** TWerkhoven[l] has joined #openttd
13:01:33 <lordnokon> fook i hate this country i stay in
13:03:03 *** TWerkhoven2 is now known as TWerkhoven
13:03:53 <MNIM> lordnokon: have you taken a look at the finances window yet?
13:06:17 <lordnokon> MNIM: yes i have, not strange there
13:12:13 *** andythenorth has joined #openttd
13:17:28 <Eddi|zuHause> lordnokon: min profit is the vehicle with the lowest income
13:19:30 <lordnokon> Eddi|zuHause: per anual year?
13:19:47 <Eddi|zuHause> yes, including vehicles stopped in depot, etc.
13:20:20 <Eddi|zuHause> you can sort the vehicle list by income
13:24:16 <lordnokon> Eddi|zuHause: i just want to confirm, no vechile is suppose to have any negative income? and has to make at least a min of 10 000 for that fin year?
13:25:23 *** Chris_Booth has joined #openttd
13:25:41 <Eddi|zuHause> we can't double check that unless you give us the savegame
13:30:13 *** Biolunar has joined #openttd
13:32:11 <lordnokon> where can I upload my savegame to?
13:32:53 <Eddi|zuHause> either the forums, or a filehoster of your choice
13:33:46 <Rubidium> what the ... a "how to get 32 bits graphics" tutorial completely filled with just changing some fracking settings that have absolutely nothing to dow ith 32bpp graphics
13:34:06 <Rubidium> (e.g. changing the autoreplace money limit)
13:34:50 <Rubidium> that's definitely helping those 32bpp graphics get going
13:44:20 <Elukka> Eddi|zuHause: in attempting to move towards rendering the intermediate views, a friend tried to render a box to fit the templates using an isometric projection at a 30 degree angle from the ground, and the sprite angles you've listed for the CETS templates
13:44:39 *** ZirconiumX has joined #openttd
13:44:52 * planetmaker guesses "the ceiling"
13:45:05 <Alberth> Elukka: 30 degrees is not right afaik
13:45:24 <Elukka> that's what truebrain was using
13:45:58 <Eddi|zuHause> the angle is arctan(1/2), so ca. 26.5°
13:51:28 <TrueBrain> did I say 30? Sorry, has to be 35 :P
13:51:36 *** Biolunar has joined #openttd
13:51:37 <TrueBrain> Orthographic, 35 degrees for the camera
13:51:53 <TrueBrain> 128x64 for resolution, aspect ratio 100/100 (as long as they are the same I guess)
13:52:11 <TrueBrain> camera is at 0/-60/3 (not really relevant tbh)
13:52:22 <TrueBrain> -30 degrees X, -225 degrees Z (y = 0)
13:52:37 <TrueBrain> and then most likely another 90 degrees X rotation to see your object :)
13:53:11 <TrueBrain> Elukka: basically, there are 3 places you need to set things to get it "just right"
13:53:30 <TrueBrain> the Scene has to be in resolution of 128x64 (at least, I like that most)
13:54:01 <TrueBrain> then I have an "Empty" entry, to bias the camera so I can do stuff easier. It is at 0,0,0 with rotation -30, 0, -225 (XYZ Euler)
13:54:19 <TrueBrain> in that is the camera, at 0,-60,3, with rotation 90,0,0 (XYZ Euler)
13:54:38 <TrueBrain> the camera itself is Orthographic, with 35 'scale'
13:55:09 <TrueBrain> also make sure to set your clipping from 0.1 to some insane value .. 500 or so
13:55:12 <TrueBrain> default value can be a bit tight :P
13:55:35 <TrueBrain> that should be the total information to generate images correctly for OpenTTD from blender :)
13:57:42 <TrueBrain> @calc 360 - 225 - 90
13:57:51 <TrueBrain> owh, lol, that is not 35 :P *fail*
13:58:46 <TrueBrain> owh, Elukka, and then a tile is 12.5 <units> in both directions
13:58:56 <Elukka> thanks, i'll try that out some time
13:59:12 <Elukka> coincidentally, with the 35 degree angle i can already render correct sprites with qubicle \o/
13:59:16 <TrueBrain> so I guess the camera is at 45 degrees from the Z, and 60 degrees from the X
13:59:19 <Elukka> the lengths are wrong but that's fixed by cutting out some pixels in the middle
14:01:53 <TrueBrain> PS: the above values work, as in, it is in my blender that generates a valid landscape set. There are a lot of other values on the wiki and I dunno where more; most of them are wrong (for blender at least) :P
14:02:16 <TrueBrain> including the .blend file on the wiki; it generates images that are off by 2 pixels :(
14:03:01 <TrueBrain> and for blender, dont forget to store the PNGs as RGBA :P (RGB is default)
14:11:15 <TrueBrain> so I guess BaNaNaS is almost 3 years old now .. how nice :)
14:11:35 <TrueBrain> 42M downloads, hehe
14:16:33 *** Chris_Booth_ has joined #openttd
14:17:36 *** HerzogDeXtEr1 has joined #openttd
14:18:00 *** Chris_Booth_ is now known as Chris_Booth
14:21:42 <Belugas> 49 million downloads???
14:45:16 *** Adambean has joined #openttd
14:49:58 <Belugas> sorry, my glasses are not what they used to be :)
15:31:31 *** LordPixaII has joined #openttd
15:35:01 *** KingPixaIII has joined #openttd
15:44:05 *** LordPixaII has joined #openttd
16:35:30 <Aali> so thats what happened in Eschede..
17:07:42 *** TomyLobo2 has joined #openttd
17:08:24 *** LordPixaII has joined #openttd
17:12:10 *** Chris_Booth has joined #openttd
17:13:33 *** TomyLobo2 is now known as TomyLobo
17:23:31 *** andythenorth has joined #openttd
17:26:46 <andythenorth> michi_cc: what was the blocker for yacd? was it frequently re-calculating the link graph for all cargo packets waiting on stations?
17:30:07 <michi_cc> Industries on the other hand generate big packets seldomly, so they are no problem at all.
17:30:34 <andythenorth> the initial route finding is to establish the destination?
17:31:30 <michi_cc> The code needs to find out if there is a route to the destination at all, and if yes, which source station to place the packet at.
17:32:24 <andythenorth> so you don't get packets on the network without any possibility of routing?
17:32:45 *** encoded has joined #openttd
17:33:16 <michi_cc> In the worst case this equals to a search of the complete link graph because otherwise you can't know if there really is no route.
17:34:19 <andythenorth> and viable routes can't be cached? (I know you explained this before - sorry) :|
17:34:30 <Alberth> you are looking for the optimal path, or did you tune it to find any path?
17:34:34 <andythenorth> just trying to figure out if we could do 'less' somehow
17:36:26 <Alberth> what if a house has a limited number of destinations?
17:36:43 <michi_cc> If you want to cache more or do less, you have to give up features. e.g. balancing over multiple links by waiting cargo or the different travel 'classes' (fast/cheap/etc).
17:37:30 <michi_cc> The proper solution to the problem is simply to not use A* but instead to use an algorithm similar to what is used for packet routing in real-world networks.
17:39:20 <TrueBrain> michi_cc: what did you use to generate that graph? :)
17:39:56 <michi_cc> callgrind (from the valgrind suite) + kcachegrind
17:40:16 <TrueBrain> one nasty bug: itis bad to add ... to cut off labels .. but okay :P
17:42:39 <andythenorth> what if houses generated large packets infrequently?
17:43:33 *** Markavian has joined #openttd
17:46:05 <michi_cc> andythenorth: Then you probably get the same problems like you always complain about with supplies :) It's going to use less CPU though.
17:46:36 <andythenorth> which problems do I complain about? (there are so many) :P
17:47:08 <michi_cc> Big packets mean you can't really distribute the among several destinations with low demand.
17:53:29 <andythenorth> michi_cc: how about 'PAX will only make n changes before refusing to travel' ?
17:53:41 <andythenorth> then you stop evaluating if a route isn't found within n links
17:53:49 <andythenorth> the maths is proabably still against that :P
17:54:41 <michi_cc> Already implemented :) But you'd be surprised how far you can come in the graph even with a low limit.
17:54:43 *** encoded has joined #openttd
17:55:20 <andythenorth> nah, not really :)
17:55:38 <andythenorth> it's factorial on the number of links per node or something
17:55:41 <andythenorth> my maths is poor :P
17:56:28 <andythenorth> but I do really like YACD, and I miss it in my current game
17:56:40 <CIA-6> OpenTTD: michi_cc * r23688 /trunk/src/train_cmd.cpp:
17:56:40 <CIA-6> OpenTTD: -Codechange: No need to check if a train needs servicing if we only extend a
17:56:40 <CIA-6> OpenTTD: train's path without actually calling the pathfinder. The path extension will
17:56:40 <CIA-6> OpenTTD: stop when hitting a junction tile, so it is impossible to miss a depot this way.
18:01:55 <Belugas> maybe one day, andythenorth, maybe one day...
18:05:16 <encoded> how do i use TTD graphic set after i used openfx ?
18:06:42 <Eddi|zuHause> i think TTT had a limit of two transfers
18:06:58 <Eddi|zuHause> with three you already get really far
18:07:33 <Rubidium> encoded: in the game options menu there's a dropdown which you can use to select the base "graphics"
18:08:00 <encoded> yes but only openGfx shows up
18:09:37 <Eddi|zuHause> __ln__: so they're using a DOS version from ca. 1993?
18:10:01 <Rubidium> encoded: then you don't have the original graphics in the right (see readme) location
18:11:11 <Alberth> Eddi|zuHause: DOS is much more stable and needs less resources than newer Windows :p
18:11:33 <Eddi|zuHause> Alberth: sure, until someone needs to maintain it :)
18:14:10 <Alberth> embedded software engineers are likely to be much happier without windows :p
18:19:08 <encoded> readme mentions some .grf files but in my TTD install the .grf files have diferent names
18:19:44 <Rubidium> how are they called there?
18:20:21 <encoded> tredit.grf trend.grf trhcom.grf trhcom2.grf
18:21:25 <Rubidium> Transport Tycoon; the original one, not the Deluxe one
18:21:48 <encoded> ah, and here i was thinking i was playing TTD all these years
18:22:39 <glx> main difference I remember between TT and TTD is one way signals
18:25:43 <encoded> so.... where do i get *cough* some original grfs *cough*
18:26:32 <glx> but you won't get any link here :)
18:27:20 <Alberth> oh, you can get a link too: www.lmgtfy.com
18:29:17 <michi_cc> Eddi|zuHause: I'd consider 4 as the useful lower limit: bus - local train - intercity train - local train - bus
18:31:27 *** supermop has joined #openttd
18:31:52 <andythenorth> does cargodist have the same CPU issue?
18:31:57 <Rubidium> so then me going to Berlin by plane would already be (almost) impossible ;)
18:32:33 <Alberth> andythenorth: it does calculations in a separate thread
18:32:49 <encoded> hey look i think i found it
18:33:13 <Rubidium> bus - local train - ic - aircraft - crappy bus (yay TXL) - local train
18:33:20 <michi_cc> andythenorth: cargodist isn't really fast, but thanks to the design that has no fixed destination for each packet, the most expensive part of the calculations can be moved to a different thread.
18:33:48 <andythenorth> for PAX, would it be more appropriate to take that approach?
18:33:50 <michi_cc> Rubidium: YACD default is 5, so: works :)
18:34:06 <michi_cc> andythenorth: In my personal opinion: no.
18:34:13 * andythenorth really likes the demand in YACD
18:34:42 <michi_cc> Where's the challenge if pax/cargo only goes where you let your trains run?
18:34:50 <Rubidium> train is more relaxed though.. local train - ic, or morning bike ride - ic
18:35:19 <andythenorth> michi_cc: agreed - for cargo
18:35:26 <andythenorth> I'm 50:50 on which is better for pax
18:36:55 *** andythenorth is now known as Guest22209
18:36:55 *** andythenorth has joined #openttd
18:39:57 <encoded> how do i reset openttd settings?
18:43:25 * Belugas would erase the config file
18:43:36 <Belugas> AFTER CLOSING The program...
18:45:31 <CIA-6> OpenTTD: translators * r23689 /trunk/src/lang/ (norwegian_bokmal.txt simplified_chinese.txt swedish.txt):
18:45:31 <CIA-6> OpenTTD: -Update from WebTranslator v3.0:
18:45:31 <CIA-6> OpenTTD: simplified_chinese - 23 changes by chenwt0315
18:45:31 <CIA-6> OpenTTD: norwegian_bokmal - 236 changes by mantaray
18:45:31 <CIA-6> OpenTTD: swedish - 9 changes by tobjork
18:53:55 <Belugas> and give me the old one!!
18:53:59 <Belugas> Plllllllllllllllllllllllllease
18:54:18 <encoded> im having problems with the music pack
18:54:39 <encoded> only NoMusic shows up now
18:55:48 <glx> get openmsx or the original music files
18:57:24 <encoded> i already put original files in
18:57:37 <encoded> and tried to download openmsx from the game
19:09:47 <encoded> ok i used installer to fix issue
19:09:57 <encoded> now copying over nightly
19:23:17 <andythenorth> including yacd for non-pax cargo only is not really viable?
19:23:30 *** Markavian` has joined #openttd
19:24:53 *** mahmoud has joined #openttd
19:29:21 <andythenorth> michi_cc: (apologies for questions) - is the link graph constructed by evaluating orders of current vehicles, or some other method?
19:45:14 *** lordnokon has joined #openttd
19:58:33 *** andythenorth is now known as Guest22225
19:58:34 *** andythenorth has joined #openttd
20:00:34 <michi_cc> andythenorth: The graph is (more or less) identical to the order graph.
20:02:07 <andythenorth> the order graph doesn't change very often (relatively)?
20:03:15 <michi_cc> The graph itself changes only if (implicit) orders change, but the cost of each link changes basically each tick.
20:04:36 <andythenorth> so in principle whether node A is connected to node X could be cached in a (massive) lookup table?
20:04:43 <andythenorth> but not the actual route
20:04:57 <Alberth> hmm, does it make sense to do such calculations then? By the time the cargo reaches the link, the cost is changed anyway
20:05:45 <michi_cc> <michi_cc> The proper solution to the problem is simply to not use A* but instead to use an algorithm similar to what is used for packet routing in real-world networks.
20:06:53 <michi_cc> Packet routing is a hard™ problem though.
20:07:33 <Alberth> andythenorth: not sure it counts as cheating. Finding the optimal route for each and every packet is big overkill, parcel delivery companies don't do that either
20:08:15 <andythenorth> I understood that the initial problem was simply 'can a packet be generated to destination x'
20:08:34 <andythenorth> which means checking if nearby stations are connected to x...
20:09:11 <michi_cc> The initial problem is: can a packet be generated to destination X, and if yes, which nearby station should it be delivered to.
20:09:35 <andythenorth> so a lookup table solves the first, but tells nothing of use for the second
20:09:42 <michi_cc> Which means that you need some kind of cost for each option.
20:11:19 <andythenorth> sounds like packet routing :P
20:14:11 *** andythenorth is now known as Guest22227
20:14:11 *** andythenorth has joined #openttd
20:14:12 *** andythenorth has left #openttd
20:14:27 *** andythenorth has joined #openttd
20:14:43 * andythenorth wonders how stable links are in a typical game
20:16:00 <michi_cc> Quite stable if a typical game has no conditional orders and goto orders are preferably non-stop.
20:17:17 * andythenorth tries to devise some horrible system of storing weightings at each node
20:18:11 <andythenorth> I should do trucks
20:18:19 <andythenorth> maths - i'm way out of my depth :P
20:19:27 <andythenorth> give the first node a starting score of -255, the packet tries to get to any node with a higher score
20:19:42 <andythenorth> scores are initially estimated, then updated when cargo packets arrive at destination
20:19:53 <andythenorth> dunno what problem that is solving
20:20:32 <andythenorth> instead: BANDIT trucks may refit to n trailers (using subtypes). Is there any harm in unifying the subtype across vehicles?
20:20:40 <andythenorth> it may cause undesirable results in some cases
20:21:12 <andythenorth> e.g. a truck with 3x9t trailers might be upgraded to 3x30t trailers
20:21:34 <andythenorth> if there are say...20 trucks in the group, you suddenly have way too much capacity on that route :P
20:21:53 <andythenorth> by a factor of...lots
20:28:43 *** andythenorth is now known as Guest22228
20:28:43 *** andythenorth has joined #openttd
20:30:51 <andythenorth> michi_cc: it would be nice if we could run the whole simulation for, say, 6 months, on ffwd before generating each packet :P
20:31:02 <andythenorth> but trying to know the future is definitely cheating
20:32:12 <andythenorth> it's nice how the ants optimise for shorter links simply because they can travel them faster
20:32:28 <andythenorth> thereby dropping more pheromone on that link
20:32:56 <TrueBrain> ugh, dont talk about Ant Optimizations :P Too many nightmares :P
20:33:09 <michi_cc> I wasn't really thinking about ant behaviour, but more what is described at the section "How MUTE Routes Messages"
20:34:36 <andythenorth> the routing is not dissimilar to the score idea I had
20:34:48 <andythenorth> both node A and node X are known
20:35:16 <andythenorth> node X starts with score 0 and assigns 1 to all connecting to nodes. They then assign 2 to all their connected nodes that don't already have a score etc.
20:35:42 <andythenorth> a limit is applied
20:36:34 <andythenorth> node A does same, but starting at 255 and scoring upwards
20:36:54 <andythenorth> packets then route to any node with score > than current cnode
20:37:03 <andythenorth> it's probably horribly inefficient computationally
20:37:19 * andythenorth wrote a lot of flash games with dumb solutions in that worked
20:37:22 <TrueBrain> you are kinda describing Dijkstra :P
20:37:38 <andythenorth> I'm hardly original :)
20:37:51 <TrueBrain> just a few years late to come up with it :P
20:38:22 <andythenorth> Dijkstra is slower than A*?
20:38:38 <Aali> what if you just gave the packet to a random station, if that packet gets delivered, increase the likelyhood of that station getting another packet with that destination?
20:38:39 <Eddi|zuHause> A* is a specialized Dijkstra
20:38:46 <TrueBrain> what Eddi|zuHause says :P
20:39:18 <andythenorth> Aali: you end up with unpleasant side effect of cargo hanging around on stations, undelivered :)
20:39:54 <Aali> maybe the industry could "reallocate" it after a while
20:40:51 <TrueBrain> easy solution for YACD btw: just pick a random town to go to :p Connected or not :D
20:41:00 <TrueBrain> forcing full networks FTW :D
20:41:11 <TrueBrain> (mind you: town, not station)
20:41:18 <Aali> isn't there a patch like that already?
20:43:08 * andythenorth ponders routing cargo based on tiles, not stations
20:43:29 <andythenorth> with 8x8 tilegroups
20:43:37 <andythenorth> and the ability to wormhole cargo between stations :P
20:43:58 <andythenorth> also solves part of infrastructure sharing ;)
20:50:59 <TrueBrain> michi_cc: how does YACD determ the from/to atm? Or can you point me to where that is calculated? (just curious)
21:00:23 <TrueBrain> lol; you really went all out feature wise, didn't you? :D
21:05:27 <michi_cc> The most important part of a cargo *destinations* patch is the destinations part because that is what directly defines game play.
21:09:58 <TrueBrain> so if I read this right, a house generates a cargo to a town, depending on weight etc, and after that it checks if there is a route towards that destination?
21:19:14 <TrueBrain> not ever really bothered with cargodest or cargodist patches, I have to say: not what I expected (at all), but much better / full / more gameplay changing :)
21:22:04 <michi_cc> Yes, when a cargo packet is generated, a destination from the list of possible destinations of the source town/industry is chosen. For a town destination a 4x4 tile square as the destination tile is then chosen by weighted random. After that, the code checks if a connection between the source and the destinations exists and is inside the given limits (transfers etc.) and if yes, moves the packet to a station. If no, the packet is dropped.
21:22:37 <TrueBrain> I am mostly impressed you even wrote auto transfer (if I read it correct, of course)
21:23:18 <michi_cc> Auto-transfer is quite trivial if you've already have the link graph infrastructure.
21:23:54 <michi_cc> And frankly, cargod?st without auto-transfers is quite useless.
21:24:07 *** Chris_Booth has joined #openttd
21:24:36 <TrueBrain> I btw wonder, if I have a station with trains and planes, which connect both to another station (which happens to be the dest for all), will they distribute theirself over those 2, or will they only take train/plane (because of the penalties)?
21:26:08 <TrueBrain> how I read the code, it picks either one, and the other one will leave empty
21:26:16 <TrueBrain> but patch over patch over patch, hard to get a total picture :D
21:27:21 <michi_cc> For a single packet: yes. But there are pathfinder penalties e.g. for cargo amount waiting on a specific link or time since the last vehicle. So the best route can change after each cargo packet is routed.
21:27:53 <michi_cc> Additionally, cargo packets have different "personalities", so some will prefer a fast connection and others an especially cheap connection.
21:28:05 <TrueBrain> yeah, that is a really nice touch btw :)
21:28:40 <TrueBrain> its funny; because you went all out on this, and wanted so much, you also have to pay by having to resolve a full graph :D
21:29:17 <michi_cc> The personalities are currently assign randomly, but you could coupled them e.g. with source house is office (accepts goods) or residential.
21:30:00 <TrueBrain> I once joked about implemented OSPF in OpenTTD. You are not far off :P
21:30:18 <TrueBrain> and now I am tempted to suggest BGP :D
21:32:13 <michi_cc> The problem with something like BGP is that each node (== station) needs a full routing table to every other station.
21:32:17 <andythenorth> <TrueBrain> much better / full / more gameplay changing
21:32:20 <andythenorth> it really is ^ :)
21:32:42 <TrueBrain> andythenorth: I am seriously surprised. I always expected YACD to be simple, but doing too much calculations. It is not simple. Not by far :P
21:32:54 <andythenorth> it's full featured and highly polished
21:33:00 <andythenorth> it's just a shame that it can't be shipped :(
21:33:08 <TrueBrain> andythenorth: well, I found code-wise errors, but that is not relevant ;)
21:33:24 <andythenorth> there are still oddities :)
21:34:04 <michi_cc> My local HD has a slightly more advanced state than the publicly available YACD 2.3.
21:34:13 <andythenorth> does it build with current trunk?
21:34:23 <TrueBrain> michi_cc: earlier this evening, when I just read you and andythenorth talking about it, I was just wondering: why not make a full matrix for all stations
21:34:36 <TrueBrain> a relative big map will have, what, 500 stations?
21:34:39 <andythenorth> what does that entail?
21:34:42 <michi_cc> It might if you fix all the save/load version conflicts.
21:34:49 <andythenorth> it it an n dimensional lookup table?
21:34:51 <TrueBrain> but then I realised that people tend to build LOTS of bus stations :(
21:35:05 <andythenorth> with YACD you build a lot MORE bus stations :P
21:35:12 <andythenorth> I think we should cheat and give PAX legs
21:35:39 <TrueBrain> andythenorth: they have
21:35:40 <andythenorth> having to build bus routes covering every tile in every connected town is not that much fun by the second or third game
21:35:51 <TrueBrain> then they are tired :P
21:35:55 <andythenorth> I'll walk 45 mins rather than get a bus
21:36:06 <TrueBrain> haha; but that is not the idea of a TRANSPORT game :P
21:36:06 <andythenorth> some station grfs have car parks
21:36:13 *** LordPixaII has joined #openttd
21:36:18 <michi_cc> andythenorth: Modify station catchment :)
21:36:31 <TrueBrain> I wish station catchment was not a square
21:36:37 <TrueBrain> that is in my top 10 of wanted features :(
21:36:58 <TrueBrain> so many things I want to code
21:37:10 <andythenorth> code a code generator :P
21:37:13 <TrueBrain> I still want to write a quad-tree to improve TileClosestByTown
21:37:19 <TrueBrain> just to see if it does improve
21:37:26 <TrueBrain> I want to write city borders
21:37:31 <TrueBrain> so you can see how big a city is
21:37:41 <TrueBrain> a view to see catchment area
21:37:42 <andythenorth> cargo specific catchment areas?
21:37:49 <TrueBrain> signals on bridges :P
21:37:53 <TrueBrain> omg, I have to stop talking now
21:38:30 <andythenorth> meh to underground
21:38:39 * SpComb photoshops some .png sprites and declares it done
21:38:44 <TrueBrain> michi_cc: if you forget about the '5 leg' constraint, you can make your FindRoute thingy much faster btw
21:38:52 <TrueBrain> but I guess you really want that feature (mostly as it is really nice)
21:39:40 * Rubidium wonders where "a GUI defined by a game script" comes ;)
21:39:57 <TrueBrain> Rubidium: as you might have noticed, I havent written a single line of new code since commit :P
21:40:09 <TrueBrain> bit burned out by all the lobbying and politics tbh :)
21:40:21 <andythenorth> there's politics :o
21:40:32 <TrueBrain> you should read the dev-channel from time to time ..
21:40:39 <Rubidium> I don't blame you either
21:41:08 <TrueBrain> Rubidium: well, it annoys me tbh, as I really want that feature :)
21:41:26 <Rubidium> that's also somewhat why I asked where it is on the "top 10" list
21:41:27 <TrueBrain> but I enjoy blender too much :D
21:41:41 * andythenorth probably isn't allowed in the dev channel :D
21:42:11 <TrueBrain> Rubidium: fair question; and I will write it soon (tm). I also have a few GS related issues to fix :)
21:42:54 <TrueBrain> michi_cc: that is "almost" a quad-tree yeah :p It misses children I guess :)
21:43:10 <TrueBrain> it is the child, so that is 50% of the work? :)
21:44:07 <michi_cc> But an almost-quad-tree might be already good enough if you store the town in all squares that are single-owned and fall back to the old search for multi-owned tiles.
21:45:04 <TrueBrain> that I don't follow, sorry?
21:45:49 <encoded> is there some way to make the map generator not put 5 factories near each other
21:46:10 <Rubidium> disable multiple industries in a town?
21:46:45 <Rubidium> somewhere in the advanced settings
21:48:21 <andythenorth> encoded: use an industry newgrf :P
21:48:32 <encoded> i only see allow multiple industries of the same type
21:48:41 <michi_cc> TrueBrain: Make an map overlay that for each N-by-N square either stores the town or a special value if more than one town has influence over the tiles in the square.
21:48:58 <TrueBrain> michi_cc: and that overlay is missing :P
21:49:08 <TrueBrain> either in a quad-tree or another solution, doesn't really matter :)
21:49:15 <TrueBrain> but TileMatrix alone is not sufficient ;)
21:49:40 <TrueBrain> but I guess you are right ;)
21:49:59 <TrueBrain> will be a heavy CPU vs memory trade-of
21:50:01 <michi_cc> No, but as town center tiles rarely change, calculating and keeping that map up-to-date isn't that big a challenge.
21:51:00 <michi_cc> Depends on how big squares you choose. We're talking about 0.5MB for 4x4 squares.
21:51:06 <michi_cc> for a 2048x2048 map.
21:51:07 <TrueBrain> so you suggest putting the vehicle hash on top of TileMatrix ;)
21:52:44 <TrueBrain> so many fun things to try! :P
21:55:07 <TrueBrain> michi_cc: btw, just thinking out loud tbh, would a solution for your YACD be if you make 'regions' in the route links. What I mean is:
21:55:18 <TrueBrain> if you consider all stations in / near a town to be a single route ina bigger graph
21:55:32 <TrueBrain> that graph will be at most <townsize> big, so a matrix would be <townsize>^2 big, at most
21:55:43 <TrueBrain> within a town you can easily calculate a route on the fly each time
21:55:51 *** LordPixaII has joined #openttd
21:56:03 <TrueBrain> that should (in theory) recude the CPU load by a huge factor (at the expense of memory, of course)
21:57:07 <encoded> what does the Variety Distribution parameter in New Game Dialog do?
21:57:08 <michi_cc> As long as we are still limited by single-core performance and measure our memory use in MB, the proper choice is always memory over CPU.
21:57:38 <TrueBrain> well, then you can just make a <station-count>^2 matrix :P
21:59:03 <Aali> encoded: it distributes variety :)
21:59:23 <Aali> low setting means uniform landscape (macro level), high setting means varied landscape
21:59:56 <TrueBrain> michi_cc: btw, single core or not, 50% CPU to a single function is never a good thing I guess :D
22:00:41 <TrueBrain> I also guess this would be a typical thing that can be threaded; as a path taken does not influence the path another packet will take (in other words: the pathfinding is fully read-only)
22:00:54 <Belugas> HAPPY NEW YEAR EVERYONE. I won't be back until i'llbe back :)
22:02:53 <TrueBrain> anyway, tnx for explaining how YACD works michi_cc :) At least I now understand what you are talking about when it comes to the issues of it :D
22:03:23 <Rubidium> Belugas: but... you're 12 hours early ;)
22:03:33 <Rubidium> in any case, have fun
22:04:10 <andythenorth> TrueBrain: that was my suggestion! :P (grouping regions) :P
22:04:16 <michi_cc> TrueBrain: But a packet moved to a station affects the cost of the link, so it does influence all later route choices. And cargo generation and vehicle unloading is fully integrated into the tile loop/vehicle tick loop.
22:04:23 <andythenorth> only yours is more...better
22:04:31 <andythenorth> but still...I demand my compsci degree :P
22:04:43 <encoded> where is Multiple industries per town option??
22:06:01 <Terkhen> economy -> towns IIRC
22:06:20 <andythenorth> or economy -> industries
22:08:20 <TrueBrain> michi_cc: hehe; like I said: you went all out, and now you are in trouble :D:D (it is a positive thing btw, let that be clear :))
22:08:45 <andythenorth> and also yacd is highly playable - on my laptop at least
22:08:51 <andythenorth> until it flattens the batter anyway :P
22:09:17 <andythenorth> also it's a little...warm
22:11:32 <TrueBrain> ugh; I have a flu, and it is annoying. when you are in bed, you start to cough. when you are sitting up your head hurts. MAKE UP YOUR MIND BODY! Ffs ...
22:13:33 <Rubidium> too bad flu is ambiguous and usually means the lesser undiagnosed variant
22:14:55 <TrueBrain> Rubidium: yeah, it is most likely just a cold. But you can never be sure at this stage :P
22:15:25 <andythenorth> flu makes you hallucinate
22:15:33 <andythenorth> I have had it maybe 3 times in my life
22:15:42 <TrueBrain> it _can_ make you hallucinate maybe :P
22:15:48 <TrueBrain> it is not part of the diagonise :P
22:16:04 <andythenorth> flu makes _me_ hallucinate. ymmv
22:16:50 <Rubidium> well... the medicin I got for my (in)flu(enza) is known to make people hallicunate and jump of the roof of tall buildings (so when I got it I was also more or less on suicide watch)
22:17:12 <TrueBrain> that is one hell of a flu :P
22:17:41 <TrueBrain> it makes you want to fly
22:18:06 <andythenorth> Rubidium: tamiflu?
22:18:27 <andythenorth> the side effects are quite rare ;)
22:18:32 <andythenorth> I just reading about it
22:19:07 <andythenorth> meanwhile...flu tends to cause delirium, and hallucinations can be part of that, but not necessarily
22:19:17 <TrueBrain> also rare for children who got injected with the "mexikaanse griep" (what is the english?) to get a neurological disfunction .. yet more than a few got it :(
22:19:17 <andythenorth> it's quite interesting wiki page
22:20:03 <andythenorth> "In the latter state, a person may be awake and immediately aware and responsive to many stimuli, and capable of coordinated movements, but unable to perform any meaningful mental processing task at all."
22:20:11 <andythenorth> ^ this is a typical morning for me :P
22:20:21 <Rubidium> TrueBrain: I'm always wondering who "studied" that
22:20:23 <TrueBrain> isnt it for everyone andythenorth? :D
22:20:33 <TrueBrain> Rubidium: studied what?
22:20:56 <Rubidium> that the vaccin causes neurological problems
22:21:27 <TrueBrain> several countries, and the supplier of the vaccin admitted it
22:21:45 <TrueBrain> and I doubt a supplier does that when it is not set in stone :P
22:21:58 <andythenorth> last time I had actual flu, my house was being rebuilt. I remember climbing out of a window onto scaffolding to decide some building issues with the builders. This was fun - I was having time skips at that point (like watching slow video)
22:22:04 <TrueBrain> also mostly means they already knew
22:22:38 <Rubidium> TrueBrain: the list of things you could get from medicine is more a list of things that people got during the trials than actually being certain it was caused by that
22:22:56 <TrueBrain> Rubidium: I guessin this case the N got so huge, it is pretty clear :D
22:24:04 <TrueBrain> I am mostly very ciritcal when it comes to statistics, as I know how easy you can manipulate them; but in this case I tend to go with them :P
22:24:10 <Rubidium> but how do you study mental issues? You generally don't know whether the problem already existed before
22:24:41 <TrueBrain> I guess it accelerated the effect, sure
22:24:54 <andythenorth> I am familiar with this
22:24:58 <Rubidium> TrueBrain: or... people started looking for it
22:25:01 <TrueBrain> but if a "large" portion of a huge group gets a certain condition after they all have been injected with something
22:25:16 <Rubidium> for example, now many many more people have dyslexia
22:25:34 <TrueBrain> Rubidium: everyone on earth has either dyslexia or discalculi
22:25:45 <TrueBrain> (by strict definition)
22:26:13 <TrueBrain> the only clause avoiding that is: and is limited in his ability because of it
22:26:18 <TrueBrain> which goes for most diagnoses
22:28:11 <Rubidium> yeah, kids get ADHD from the TV
22:28:17 <TrueBrain> I can imagine with sponsbob
22:28:22 <TrueBrain> I even turn totally ADHD over it
22:28:47 <TrueBrain> who ever thought it would be educational to show a hyper-active thingy on television, and promote it as good behavoir, should be shot :P
22:28:50 <Rubidium> (or is it the parents seeing all those busy kids in TV programs saying they've got ADHD, so their kid has it as well)
22:29:27 <TrueBrain> Rubidium: like said, with many conditions it goes: you don't have it until others have negative effects from it
22:29:34 <Rubidium> it's (IMO) by a large part cause by upbringing
22:29:42 <TrueBrain> if the latter holds, many diagonses start to hold :P
22:31:06 <TrueBrain> well, after spending 2 minutes with a "claimed" ADD person and a real ADD person, you know who is the "claimed" :P
22:32:56 <TrueBrain> Rubidium: I guess what we are both saying is: the world is doomed if you believe everyone out there :D
22:33:55 <TrueBrain> want to hear another 'fun' story .. got a letter from my ISP yesterday
22:34:03 <TrueBrain> headline: "WE ARE IMPROVING AND MAKING YOUR INTERNET FASTER"
22:34:12 <TrueBrain> 2 mbps connections become 10 mbps
22:34:25 <TrueBrain> 20 mbps connections will be 25 mbps, for N euro extra
22:34:31 <TrueBrain> my "university" connection?
22:34:35 <TrueBrain> will become a 10 mbps
22:34:41 <TrueBrain> HOW IS THAT FASTER?!
22:34:50 <TrueBrain> and they dropped that in 2 lines in the middle of a letter
22:35:30 <TrueBrain> I am still upset because of it :(
22:36:02 <andythenorth> apparently lack of sleep causes excessive emotional response to events :)
22:36:17 <Rubidium> TrueBrain: maybe it improves the ping times due to less traffic on the line?
22:36:17 <TrueBrain> then I always have a lack of sleep :P
22:36:25 <TrueBrain> Rubidium: hahaha :D
22:36:37 <Rubidium> and that makes the internet *faster*
22:36:39 <TrueBrain> I am pretty sure they can no longer find a way to forward the EAPOL packets tbh :P
22:37:00 <TrueBrain> (which is the authentication method for all university networks in the netherlands these days I believe
22:37:22 <Rubidium> what they told you is that the throughput will be lowered ;)
22:37:28 *** andythenorth is now known as Guest22245
22:37:28 *** andythenorth has joined #openttd
22:37:41 <andythenorth> BT broadband "the most reliable wireless in the uk"
22:37:43 <Rubidium> or that's how they'll spin it if it ever gets to court
22:38:00 <andythenorth> and you can't even swap the stupid crappy wifi router out for a non-BT one :/
22:38:02 <Rubidium> andythenorth: makes me wonder how unreliable the other internet is in the UK ;)
22:38:03 <TrueBrain> Rubidium: I just wonder why they didn't offer me free-for-live upgrade to 25mbps, to compensate
22:38:05 <TrueBrain> that would have been fair
22:38:29 <andythenorth> they claim 'most reliable wireless'
22:38:38 <andythenorth> which is surely lies, as each house varies
22:38:47 <Rubidium> TrueBrain: maybe the university stopped subsidizing?
22:38:49 <TrueBrain> "most reliable" is very vague andythenorth ;)
22:38:59 <TrueBrain> Rubidium: doubtful; but something to call about in the new year :P
22:40:26 <Rubidium> other news: in Enschede on the campus they're working on upgrading the room connections to 1 Gbps ;)
22:40:37 <TrueBrain> nice ... impressive even
22:40:58 <Rubidium> too bad the connection on my work just sucks massively
22:44:19 <TrueBrain> well, clearly Leiden no longer cares enough about its students to supply them with real internet connectivity to university resources :P
22:44:38 <TrueBrain> guess I need to setup a bouncer, to access those sites restricted to university IPs :P
22:45:01 <Rubidium> pff... how many of those are there? ;)
22:45:18 <TrueBrain> papers, journals, ...
22:45:19 <Rubidium> I've maybe needed it, on average, one a year
22:45:37 <TrueBrain> I use it a bit more :)
22:45:49 <TrueBrain> just annoyed; not un-overcomable :P (is that english?)
22:46:12 <Rubidium> though most of the time I was at the uni doing that stuff so I didn't notice
22:46:32 <Rubidium> in the other cases I just opened the VPN connection and it worked
22:46:48 <TrueBrain> we don't have a 'VPN' as such
22:47:03 <TrueBrain> every department has its own method ...
22:47:38 <Rubidium> maybe it's that the student network in E'de is run by students that makes it more useably for students?
22:47:48 <Elukka> Eddi|zuHause: did the bavarians have different colors for different coach classes too?
22:47:56 <TrueBrain> Rubidium: for sure that makes a huge difference :)
22:48:30 <Eddi|zuHause> Elukka: haven't researched that yet
22:49:05 <Elukka> well, i'll get the green one done first
22:49:51 <Elukka> i noticed the prussians at some point moved to a unified green color scheme
22:50:14 <Jogio> I had crashed openttd just before when I wanted to read the readme of a newGRF
22:50:28 <Jogio> is this bug known already?
22:51:07 <Rubidium> not to me; which NewGRF did you want to read the readme from? What operating system are you using?
22:51:17 <Jogio> wait I will copy a error message
22:51:18 <TrueBrain> which OpenTTD version?
22:51:20 <Jogio> Error: Assertion failed at line 1306 of /bamboo/src/widget.cpp: given_width >= this->smallest_x && given_height >= this->smallest_y
22:51:38 <Jogio> I have ubuntu lucid lynx
22:51:58 <Jogio> and it happens in 1.2.0 RC 1
22:52:28 <Jogio> but only when you see the newgrfs of a multiplayer game
22:53:30 *** nudelundbrot has joined #openttd
22:54:57 <Rubidium> Jogio: what is the name of the NewGRF you opened the readme for (which server)?
22:55:57 <Jogio> the openttdcoopserver maybe
22:56:13 <Jogio> I think it happens to all newgrfs
22:56:41 <Jogio> and I just tested that it crashed to when i click on license
22:57:46 <kais58> warning: attempt to free a non-heap object 'ByteBlob::hdrEmpty' <- that normal on building 1.1.3?
22:58:41 <Rubidium> kais58: if you're interested why, read the comments near the line that causes the warning
22:59:08 <kais58> I just saw it when building it, just checking, still the same on nightlies?
22:59:56 *** AD is now known as Guest22251
22:59:57 <Rubidium> kais58: yes-ish; you will only see it when assertions are disabled
23:00:22 <Rubidium> for releases the assertions are disabled, for nightlies, betas and release candidates they are enabled
23:00:47 <kais58> so it's still an issue then is what I was asking ;)
23:01:23 <Rubidium> kais58: yes, it is an issue... but not in OpenTTD
23:01:37 <TrueBrain> Jogio: any specific grf? Or just a random one?
23:01:42 <kais58> I'll go read the source then
23:02:20 <Rubidium> Jogio: which base graphics set are you using; if OpenGFX, which version?
23:03:39 <Jogio> but the strange is that all works when i click on newGRF settings in main menu
23:03:57 <TrueBrain> is the random one always the same? :)
23:04:02 <TrueBrain> just name one which you know crashes :)
23:05:00 <Rubidium> I've randomly tried half a dozen of #openttdcoop's server but none triggered the error
23:05:40 <Jogio> hmm maybe its only under ubuntu
23:06:11 <Jogio> chips on openttdcoop welcome server
23:06:23 <TrueBrain> can't reproduce it either
23:06:47 <Rubidium> Jogio: do you use 64 bits or 32 bits?
23:08:04 <Jogio> i now use openttd 1.2.0-beta1-linux generic-i686
23:08:28 <kais58> oh, ok, it's just a false warning by GCC
23:09:16 <Jogio> i try to reboot but I don't think it changes something
23:10:54 * Rubidium is off for dreams ;)
23:10:59 <TrueBrain> sleep well Rubidium
23:11:07 <appe_> how did i change my nick in multiplayer?
23:11:26 <TrueBrain> appe_: console, "name <newname>"
23:11:30 <TrueBrain> or quit, and change it in the GUI :P
23:11:35 <TrueBrain> or was it "nick <newname>"
23:14:28 <Jogio> any ideas about my bug?
23:15:19 <TrueBrain> and I guess the only way to find out what goes wrong, is if you can attach a debugger (gdb)
23:15:23 <TrueBrain> not sure if you know how that works
23:15:26 <Yexo> heh, I love people making suggestions about things we already implemented :p
23:15:28 <TrueBrain> otherwise, make a bug report :)
23:15:39 <TrueBrain> Yexo: and make it sound like NOBODY EVER thought about it :P
23:15:55 <Yexo> well, nobody did post it as suggestion on the forum :p
23:16:42 <Jogio> i can compile but I never debugged something
23:17:06 <TrueBrain> well, as long we can't reproduce it, you are out of luck sadly
23:17:23 <Yexo> Jogio: can you reproduce the issue yourself?
23:18:13 <TrueBrain> which language btw Jogio?
23:18:25 <TrueBrain> meh; a crash.log would be useful :D Has all those details
23:19:04 <Yexo> Jogio: if you can compile yourself, configure with "./configure --enable-debug=3", after that "make" and then start the resulting binary in gdb
23:19:08 <TrueBrain> ah, I can reproduce it :)
23:19:18 <TrueBrain> happens in German, not in english
23:21:43 <Jogio> in english it works on my computer :-)
23:22:09 <TrueBrain> CheckForMissingGlyphs causes a recalculation of the windows width, which fails somehow ;)
23:22:53 <Jogio> maybe the german text is longer then the english or something
23:23:15 <TrueBrain> given the error, I expect the rerverse ;) But lets find out ..
23:24:50 <Yexo> this->smallest_x = 4294967281; (=-15 if it was int)
23:29:05 <Jogio> good night and thank you for your quick help :-)
23:29:33 <TrueBrain> I wonder if he understands he helped us finding a bug :P
23:30:04 <Yexo> but for me it's time to sleep, good night
23:33:15 <TrueBrain> FS#4926, to make sure we don't forget :)
23:36:19 *** Brianetta has joined #openttd
23:39:24 <kais58> are the source file md5's listed somewhere?
23:42:17 <kais58> for older releases that is, I can see the current ones
23:56:19 *** Brianetta has joined #openttd
continue to next day ⏵
