IRC logs for #openttd on OFTC at 2008-08-08
⏴ go to previous day
00:33:20 *** Eddi|zuHause has joined #openttd
00:40:49 *** Ridayah has joined #openttd
01:03:55 *** ben_goodger has joined #openttd
01:32:47 *** welshdragon is now known as welshdra-gone
01:53:09 *** lobster has joined #openttd
01:57:22 *** fmauNeko has joined #openttd
01:57:50 *** fmauNeko is now known as fmauNekAway
02:28:41 <CIA-5> OpenTTD: belugas * r14016 /trunk/src/ (blitter/32bpp_anim.cpp gfx.cpp gfx_type.h): -Codechange: Remove some magical numbers
02:50:42 *** michi_cc has joined #openttd
02:50:42 *** ChanServ sets mode: +v michi_cc
03:19:33 *** Brianetta has joined #openttd
03:37:13 <CIA-5> OpenTTD: belugas * r14017 /trunk/src/ (gui.h settings_gui.cpp): -Codechange: DrawArrowButtons now uses Colours enum to specify the colour of the button
03:49:11 <Belugas> it'snmore like good night to me :)
03:49:30 <Belugas> i wish you a good day :)
04:02:48 *** kjetil_ has joined #openttd
04:02:58 *** kjetil_ is now known as Forked
05:13:33 *** Singaporekid has joined #openttd
06:02:28 <CIA-5> OpenTTD: peter1138 * r14018 /trunk/src/settings_gui.cpp: -Fix (r14017): Typos prevented compilation.
06:02:56 *** Sir-Bob has joined #openttd
06:03:56 *** Alberth has joined #openttd
06:21:25 <Brianetta> Early morning coding?
06:22:04 <Brianetta> Fixing Belugas' typos?
06:23:06 * SpComb gives peter1138 the rubber chicken
06:24:14 <Alberth> that's why he gets the lart
06:40:07 *** Vikthor has joined #openttd
06:40:42 * Brianetta launches some clay pigeons for peter1138
06:41:16 <Vikthor> I see I have come in right time to see what's new with destinations :D
06:52:37 *** Marduuhin has joined #openttd
08:05:01 <peter1138> I guess Celestar is out :)
08:53:14 <Wolf01> fonso, just a question, why diagonal leveling and not drag&paint? the last one seem way more useful for me (I like to draw maps)
08:53:32 <fonso> I want to lay diagonal rails
08:53:50 <fonso> and drag&draw requires more precision hand work for that
08:54:38 <Wolf01> I think not, 'cause both use the CTRL key to work, or we need to make a new widget
08:55:13 <fonso> yes, or a second modifier key or a configuration setting. But all of that is no real problem
08:55:50 <fonso> and you can have drag and draw raising and lowering even without that
08:58:41 <fonso> Wolf01: any more discussion about that? Otherwise I'm off for breakfast ...
09:00:03 <blathijs> fonso: The iterator stuff looks cool!
09:00:25 <fonso> I'll be back in an hour or so
09:00:29 <blathijs> fonso: I would only consider making the iterator itself invalid, instead of having it dereference to an invalid tile
09:01:01 <fonso> Like provide an operator(bool)?
09:01:46 <fonso> but I'm hungry now, bye
09:01:59 *** Digitalfox has joined #openttd
09:02:13 <Rubidium> eating and IRC don't go together?
09:02:21 <Rubidium> must've done something wrong then
09:03:34 <blathijs> Hmm, even leaving your IRC on while you're eating somewhere else would help
09:03:49 <blathijs> Then I could have said stuff to him while he was gone :-)
09:04:19 <sid3windr> maybe his laptop doubles as table
09:05:02 <Rubidium> even then you can leave irc open
09:06:23 <Rubidium> last time I closed my IRC was when an exploit in the client was fixed
09:06:40 <Rubidium> yup, they really need to make it possible to patch running binaries ;)
09:06:42 <sid3windr> yeah but if he's running windows the laptop will go into standby if you close the lid to eat on top ;)
09:07:03 <Rubidium> then he wouldn't be running kopete
09:20:33 <peter1138> I don't think 'drag and paint' has any place outside the scenario editor.
09:21:05 <peter1138> Rubidium, irssi has an upgrade command...
09:21:44 <sid3windr> yup, and mostly it doesn't crash :]
09:32:03 <CIA-5> OpenTTD: rubidium * r14019 /trunk/src/yapf/ (follow_track.hpp yapf_costrail.hpp yapf_node_rail.hpp): -Fix [FS#2205]: game crash after order skip while waiting for free path (michi_cc).
09:33:14 <SpComb> peter1138: but /upgrade zeros your /uptime
09:33:39 <Rubidium> yes, that's a major flaw ;)
09:33:41 * SpComb Irssi uptime: 468d 0h 32m 13s, as a fact relevant to this discussion
09:34:00 <Rubidium> better shut down the binary so it frees possibly leaked memory too
09:34:46 <Rubidium> then you must've some ancient kernel
09:35:42 *** Dred_furst has joined #openttd
09:53:51 *** bleepy_ has joined #openttd
09:53:57 *** bleepy is now known as Guest792
09:53:57 *** bleepy_ is now known as bleepy
09:56:38 <Brianetta> My server's desyncing like 0.4 used to
09:56:43 <Brianetta> ie., all the bloody time
09:56:59 *** Alberth has joined #openttd
09:57:03 <fonso> so, where should I declare LevelMode? command_helper.h perhaps?
09:57:34 <peter1138> fonso, no, it's related to map levelling, not the command system.
09:57:54 <Brianetta> It must be grvts.grf
09:57:58 <Brianetta> It's the only change
09:58:05 <fonso> It's also used for clearing
09:58:13 <fonso> I could rename it somehow
09:58:31 <peter1138> Either way it's map operations, not commands, heh
10:01:15 <fonso> I still don't know where to put it. map_func.h?
10:01:43 <CIA-5> OpenTTD: truebrain * r14020 /branches/noai/src/ai/api/ai_airport.cpp: [NoAI] -Fix: AIAirport::GetAirportType didn't check for valid tile (Yexo)
10:01:54 <blathijs> fonso: I meant having an OrthIterator::end or something like that
10:02:21 <blathijs> fonso: like the STL containers .end()
10:02:37 <fonso> ok ... I'll have a look at how they implement that.
10:03:53 <blathijs> fonso: Probably more complicated :-) I think for our case it should be enough to have a bool in the iterator that says "finished" or something
10:04:05 <blathijs> fonso: Or perhaps not explicit, but implicit from the current tile index
10:04:34 <fonso> Providing operator(bool) would be very elegant, I think
10:05:01 <blathijs> fonso: Anyway, by overwriting operator== (and/or operator!=, does that one exist?) you can compare with this special end iterator
10:05:41 <blathijs> fonso: Yeah, but that's a different approach (more like the java approach with hasNext())
10:05:52 <Yorick> I think automatic path reservation with leaving stations is a bad idea
10:06:12 <fonso> So, I'll do it with operator==
10:07:19 <Yorick> what determines if the path gets reserved or not?
10:07:21 <fonso> Can we decide on where LevelMode should go? I'd say either map_func.h because it's related to map functions or map_type.h because it's no function and related to map operations.
10:07:47 <fonso> or it's own header, but that'd be silly.
10:07:50 <Noldo> Yorick: what on earth are you talking about?
10:08:22 <Yorick> Noldo: if train leaves station, its path gets reserved and it sometimes waits on a platform
10:08:53 <fonso> the only terraform_*.h is terraform_gui.h
10:08:53 <peter1138> You're still not making sense.
10:09:27 <peter1138> map_type.h is most suitable without making terraform_type.h, I think.
10:09:45 *** ChanServ sets mode: +o Bjarni
10:09:46 <peter1138> I meant Yorick is not making sense :p
10:09:54 <peter1138> But then I'm not surprised.
10:10:05 <fonso> ok, so it shall be map_type.h
10:10:08 <Bjarni> Yorick: when did you last make any sense?
10:10:19 <peter1138> Bjarni, 1952 I think.
10:10:41 <Yorick> I didnt exist back then
10:27:23 <blathijs> fonso: One could wonder in what cases two iterators are equal
10:27:41 <blathijs> fonso: If both of them have current == INVALID_TILE, I think they should be equal
10:27:59 <fonso> well, I can do it that way
10:28:08 <blathijs> fonso: But what if they point to the same tile, but have different parameters?
10:28:10 <fonso> but it won't work for diagonal
10:28:25 <fonso> At the moment I compare all parameters
10:28:36 <fonso> There is no current in DiagIterator
10:28:56 * blathijs looks closer at DiagIterator
10:29:16 <fonso> Quite frankly, I don't consider the end() thing the best solution, now that I've implemented it
10:29:39 <blathijs> fonso: Then, for DiagIterator, two iterators that both have b_cur == b_max should be considered equal, I think?
10:30:02 <fonso> But it still looks ugly
10:30:41 <blathijs> I do actually agree that the java-style method is prettyer, but this is more consistent with C++ I guess
10:39:28 *** Ridayah has joined #openttd
10:46:55 *** KillaloT has joined #openttd
10:53:13 <Celestar> peter1138: status? ;)
10:56:47 <blathijs> fonso: Sorry, missed your link
10:57:29 <fonso> Actually usage of the iterators looks better now, so perhaps you're right
10:57:32 *** welshdra-gone is now known as welshdragon
10:58:14 <fonso> And are initializers allowed, like in line 56 of the paste?
10:58:50 <blathijs> fonso: Yeah, that looks right
10:59:08 <blathijs> fonso: For reference variables that's even the only way to initialize them
10:59:49 <fonso> I know, but you seem to avoid them where possible, so I thought there may be some coding style rule about them.
11:00:33 <blathijs> Dunno, I don't find them very clear generally, especially when there is also a constructor body
11:00:41 <blathijs> In this case, it's quite elegant
11:00:59 <blathijs> but I haven't been working with the code a lot lately, so can't really say for sure
11:02:41 <blathijs> fonso: Is the overhead you talked about still here?
11:03:06 <fonso> It is reduced to two comparisons and a function call
11:03:07 <blathijs> fonso: I think there is some, you now need two compares on each iteration, instead of one
11:03:55 <fonso> in the previous version I compared all the members on operator==
11:04:01 <fonso> which is even more overhead
11:04:44 <fonso> on the other hand two identical Iterators won't be "==" now
11:05:07 <fonso> That's somewhat counterintuitive, but I hope no one will use them like that.
11:05:13 <blathijs> fonso: You could optimize it a bit and still have that behaviour I guess?
11:07:49 <fonso> ok, I'll do it like that
11:08:07 <blathijs> It does introduce an extra conditional, but for the if (it == DiagIterator::end()) case, a smart compiler should even be able to optimize that away
11:08:56 <blathijs> yeah, so the will be true
11:09:23 <blathijs> Perhaps if you make end a static member var instead of a static method, it is a bit easier for the compiler
11:09:46 <blathijs> Though that should work out fine when it's inlined
11:09:48 <fonso> Then I get into static initialization fiasko
11:10:10 <fonso> It will try to initialize end before the class is there.
11:10:44 <fonso> But perhaps I can circumvent that here ...
11:11:40 <blathijs> Huh? Static member vars should just be possible, right?
11:11:52 <blathijs> but this is a (hypothetical) optimization, so don't bother just yet
11:12:03 <blathijs> I'll see what gcc makes of it with a method
11:12:13 <blathijs> fonso: Do you have the implementation of the constructor for me?
11:14:04 <fonso> actually we don't get static initialization fiasko here as there is no other class involved and I can guarantee that the iterator class is initialized first.
11:14:20 <fonso> so, ok end as const static variable
11:21:51 <blathijs> Hmm, I'm getting undefined reference to `__cxa_guard_acquire' errors :-S
11:23:09 <blathijs> Ah, using g++ instead gcc works
11:24:15 <Celestar> note to self: using debug_level 7 on a large map with many vehicles is ... stupid
11:30:52 <peter1138> Celestar, I fixed a problem with the new cargo list thing that I thought I'd already handled :o
11:31:40 <peter1138> Celestar, but alas my home PC is off now :p
11:32:34 <Celestar> peter1138: I have a sized-based biaser for the PRNG
11:32:48 <Celestar> it's not perfect, but it causes a little insane amount of passengers wishing to go to bus stops
11:34:07 <Celestar> little less insane I mean
11:38:08 <Celestar> int i = 0; for(;;) {}
11:38:13 <Celestar> i is out of scope after the for
11:38:25 <Celestar> int i = 0; SomeStatement; for(;;) {}
11:38:32 <Celestar> i is (correctly) in scope after the for
11:38:54 <Roujin> say, i'm playing a game with the wip cargodests right now - would you mind if I posted screens of it in the forums?
11:39:07 <Roujin> or would it be okay with you?
11:40:53 <fonso> Is the generic name in english for diagonality and orthogonality orientation (as in german) or something else?
11:42:52 <Celestar> I'd say orientation is good enough
11:42:55 <TinoDidriksen> Orientation will work.
11:43:36 <peter1138> Celestar, too many }s ?
11:44:30 *** LilDood has joined #openttd
11:44:32 <TinoDidriksen> There is a flag for enforcing correct for() scope. May change the behavior of that "bug".
11:44:42 <blathijs> fonso: Yipes. Did you know that static variables in a function are initialized on the first function call?
11:45:02 <Celestar> peter1138: other the code wouldn't really compile without access i after the loop, right?
11:45:32 <fonso> blathijs: that wouldn't have been my first static initialization fiasko
11:46:38 <Celestar> peter1138: TIC/TOC is the problem
11:47:00 *** Wezz6400 has joined #openttd
11:47:36 <blathijs> Anyway, that does mean that gcc is quite unable to optimize the operator== there
11:47:53 <fonso> I use a member for end now
11:48:00 <fonso> there is no second class involved
11:48:08 <fonso> which is a precondition for sif
11:49:07 <Wolf01> mmm train crash in czech republic
11:49:50 <Wolf01> and naval crash in Italy
11:50:18 <Celestar> peter1138: TIC opens a block, TOC closes it
11:50:28 <Roujin> since nobody is protesting i guess i may do it..
11:50:59 <Celestar> peter1138: hence the stuff going out of scope
11:51:41 <Celestar> peter1138: GetString is slow as hell (=
11:55:55 <blathijs> fonso: It seems gcc is smart enough to simplify the operator==, but only when it is comparing with local variables
11:56:31 <Celestar> peter1138: factor 10 faster in one function with removing GetString() from a loop
11:57:02 <fonso> blathijs: Does that really matter?
11:57:03 <blathijs> fonso: If I compare with a static const member that is initialized to some fixed value, it doesn't pick it up. But when I compare with a local variable "end" that uses the same constructor, it does pick it up..
11:57:28 <blathijs> fonso: It saves a compare :-)
11:57:43 <fonso> I'm somewhat reluctant to optimize against g++ bugs
11:58:06 <blathijs> But, with modern speculative executing processors, the branch shouln't have any real impact I guess
11:58:08 <Celestar> are we optimizing code that is called once per user-command? (=
11:58:39 <fonso> no, once per tile when looping over a rectangle
11:58:43 <blathijs> Celestar: Once per iteration in a tile loop (and if we replace all BEGIN/END_TILE_LOOP with these iterators, this might become more relevant
11:59:16 * blathijs is off cooking lunch, bbl
11:59:20 <Celestar> blathijs: that sounds different
11:59:37 <Celestar> blathijs: are you telling you found a C++ way to do things faster than the macros? (=
11:59:41 <blathijs> fonso: Just leave it like this for now, we'll see about optimizing it later on I guess
12:00:03 <blathijs> Celestar: No, the C++ way is slower because it doesn't get optimized properly :-p
12:01:24 <fonso> ok, I'll make a new version of the patch ready with also LevelMode divided into LevelMode and Orientation and the (atm useless) option to level or demolish in multiple orientations at once
12:02:00 <Celestar> blathijs: you checked the assembly?
12:02:34 *** Sanitarium has joined #openttd
12:02:39 <peter1138> Celestar, er, yeah, try to avoid GetString if you don't need it ;)
12:02:52 <blathijs> pretty funky stuff that g++ generates. I did learn something: Never use static variables in a performance critical function...
12:04:02 <Celestar> peter1138: well, if it's only once per function it's ok. if it's once per loop iteration it can get bad :P
12:07:36 <fonso> I forgot the { at start of functino
12:07:41 * Celestar looks at the profiling output and blinks
12:07:58 <Celestar> 1.36 35.96 0.81 13711130 0.00 0.00 MakeEnglishOriginalTownName(char*, unsigned int, char const*)
12:08:17 <Celestar> why do we call the townname generator 14 million times?
12:09:04 <Celestar> and I _loaded_ the game
12:09:06 <peter1138> Oh... don't know then :p
12:09:50 <Celestar> 1.31 36.74 0.78 123400170 0.00 0.00 ReplaceWords(char const*, char const*, char*)
12:09:54 <Celestar> this is also interesting ..
12:12:05 <SmatZ> Celestar: does this happen with clean trunk?
12:13:02 <Celestar> SmatZ: I'm trying to determine the reason first
12:13:08 <glx> Celestar: this profile looks like you just generated a game
12:13:18 <Celestar> glx: I loaded a game from command line
12:13:24 <Celestar> peter1138: doing so in a minute sec
12:13:31 <glx> then it failed to load it
12:14:36 <Celestar> glx: no. I can see the correct game
12:15:08 <Celestar> glx: check GetStringWithArgs, table 14
12:15:14 <glx> strange because these 2 functions are called only when generating a town
12:15:37 <glx> hmm maybe when displaying town name too
12:16:01 <Celestar> glx: I guess when calling a player company or something
12:16:08 <Celestar> glx: or maybe when calling a station name
12:16:55 <TinoDidriksen> gprof shows the whole call stack, so should know precisely who called it.
12:17:12 <Celestar> TinoDidriksen: haven't scrolled enough (=
12:17:28 <glx> Celestar: GetSpecialTownNameString() calls the generator
12:18:03 <glx> GetSpecialPlayerNameString() can do it too
12:18:06 <SmatZ> so it is called when displaying town name?
12:18:45 <peter1138> What's the point in ...Iterator::end?
12:19:07 <fonso> it's more elegant to use
12:19:14 <fonso> see in terraform_cmd.cpp
12:19:22 <fonso> and blathijs requested it
12:19:24 <glx> SmatZ: of course, the town name is just an int32
12:19:47 <fonso> having the iterator dereferred to INVALID_TILE
12:19:59 <Celestar> peter1138: ::end != :last
12:20:25 <SmatZ> glx: well the function name was looking as it is used for name generating, and Celestar though so too, so I was confused :-P
12:20:26 <peter1138> ::end looks to be some random iterator with no values setup
12:20:38 <Celestar> TinoDidriksen: know of any decent tool that can make a calltree from a profile?
12:20:54 <fonso> it has INVALID_TILE and b_max == b_cur respectively
12:21:01 <fonso> so it's always an invalid iterator
12:21:03 <Celestar> glx: is it prudent to generate a town name over and over and over?
12:21:17 <peter1138> I don't think that's particularly elegant. It's confusing.
12:21:50 <glx> Celestar: doesn't matter, each int32 always generate the same townname
12:21:52 <blathijs> Celestar: Would be cool if the iterators are compatible with STL enough to do std::vector<TileIndex>(DiagIterator(a, b), DiagIterator::end) to build a list of those tiles
12:22:38 <fonso> peter1138: so how would you like it?
12:22:52 <fonso> and please discuss with blathijs
12:23:07 <blathijs> THe alternative would be to have DiagIterator(a,b).end() or something, ie an end that is specific to the range you are iterating
12:23:08 <Celestar> peter1138: updated windows? \o/
12:23:38 <fonso> but I want you to agree first
12:23:38 <blathijs> But I think that would become less efficient
12:23:47 <TinoDidriksen> Celestar, using gprof? 'cause then it should be part of standard output.
12:23:51 * blathijs continues cooking in the meanwhile
12:26:46 <fonso> Actually I can't see what is so confusing about an all-zeroed iterator as end marker.
12:27:14 <TinoDidriksen> Celestar, or gprof --graph rather...seems I had that as default globally.
12:28:02 <Celestar> TinoDidriksen: I see (=
12:30:26 <Roujin> because I got no answer before, I ask once again - do you have something against me posting screenshots from a game with the WIP cargodests in the tt-forums?
12:30:36 <Celestar> peter1138: how often is the cargo list built?
12:30:39 <Celestar> peter1138: in the SVW?
12:31:12 <peter1138> Whenever it changes, and if the item is expanded in the view.
12:33:02 <peter1138> If you change the sort order it doesn't need to rebuild it either.
12:33:23 <Celestar> but anytime the amount changes the cache in invalidet
12:34:04 <peter1138> Previously any time the amount was changed it refreshed the window... which then redid it.
12:34:30 <peter1138> It's not much different, except it is aware of only refreshing the cargo type that's changed.
12:52:09 *** SquireJames5 has joined #openttd
12:53:03 <Celestar> peter1138: any bright idea about FIFO loading?
12:53:48 <fonso> so, peter1138, you're ultimately against *Iterator::end or only against the current implementation?
12:54:23 <hylje> i'm in the belief the C++ implementation of iteration is braindead
12:55:00 <fonso> I can do it any other way you like, but I'd like you to agree on something
12:55:59 <Gekz> there are never enough tiny teddies
12:57:54 <Celestar> heh how can I convert a "dot" file into a graph?
12:58:24 <Alberth> Celestar: dot -Tpng myfile.dot > myfile.png
12:59:30 <Celestar> Alberth: tried that :P
12:59:42 <Alberth> You can also convert to jpg or ps
12:59:44 <Celestar> there result is not what I imagined :P
12:59:53 <Celestar> I get a DNA profile or something (=
13:01:03 <TinoDidriksen> Looks like a scale problem.
13:01:31 <Alberth> Convert to postscript
13:02:35 <Celestar> peter1138: so .. when do we go public on the forums? (=
13:03:30 <Alberth> Roujin is working on that problem :P
13:04:03 <Celestar> Alberth: on what problem?
13:04:49 <Alberth> going public, at least judging by the question he asked here
13:05:38 <Roujin> oh, hey if you don't like it i go and delete it again
13:05:47 <Celestar> Roujin: heh no problem (=
13:05:52 <Celestar> I just hate writing stuff :P
13:05:52 <fonso> bah. I'll come back some other day.
13:06:18 <Celestar> Roujin: can I see it?
13:07:02 <Belugas> beautyfying after compile ad before commit should be avoided like plague
13:09:06 <Roujin> I try to make some screens showing informative stuff like the connections map view..
13:10:25 <Celestar> Alberth: ok .. 3.9MB worth of svg :P
13:11:51 <Alberth> Usual problem is too many details :)
13:14:38 <Celestar> WAAAY to many apparently
13:14:53 <Celestar> peter1138: trying to solve the improvedload problem
13:14:55 <Alberth> What are you viewing?
13:15:22 <blathijs> peter1138: What alternative do you suggest for the iterator thing? Java style hasNext() (or some other way to ask the iterator if it's still valid), or more than one end iterator?
13:15:46 <blathijs> peter1138: The latter could probably be achieved by calculating what the one-past-the-end tile would be I guess
13:15:47 <Celestar> Alberth: openttd call tree (=
13:18:21 *** Yorick_ has joined #openttd
13:18:43 <rortom> oh, nice tools, thanks :)
13:18:56 <rortom> is there something qual for svn?
13:18:57 *** Yorick_ is now known as Yorick
13:21:29 <Roujin> Celestar and others: quick question about the cargodest patch before i write wrong information in my thread:
13:21:53 <Roujin> what exactly does influence the size of the dots in the map view?
13:22:15 <peter1138> It's related to the number of passengers waiting at the moment.
13:22:23 <peter1138> Obviously that's not much use for cargo-only stations.
13:22:41 <peter1138> It is in fact exactly the same as that other paxdest patch ;)
13:23:02 <Celestar> peter1138: we're going to change that are we not? (=
13:23:17 <peter1138> Hence "at the moment"
13:23:43 <Roujin> I will include that information
13:23:55 <peter1138> Probably just add up all the cargo and use that total.
13:24:04 <peter1138> Or rather, all the cargo that's selected.
13:26:02 <Celestar> peter1138: the second possibly, dunno how difficult it is
13:26:05 *** Doorslammer has joined #openttd
13:27:01 <Celestar> that sounds great (=
13:29:03 <peter1138> Adds yet another loop though, heh...
13:30:05 *** ChanServ sets mode: +v tokai
13:30:49 <Celestar> if it's too slow, have the people close the minimap :P
13:30:52 *** Sanitarium has left #openttd
13:31:02 <Celestar> oh man this improvedload is giving me headaches
13:31:43 <Celestar> v->cur_order_index always points to some order in v->orders right?
13:34:36 * blathijs pokes peter1138 again
13:34:40 <Rubidium> and maye when orders get removed it gets invalid
13:37:49 <Celestar> Rubidium: you've giving me a headache with your improvedloading (=
13:44:17 <Roujin> uhm.. there seems to be something wrong with the transfer/income scrolling text
13:44:48 <Roujin> there's no way I've just earned 100.000 DM on this screenshot
13:45:39 <Celestar> Rubidium: could you test my FIFO fix?
13:45:48 <Roujin> I rather get the feeling that the amount that's displayed as "Transfer" is awarded to my account the second the train arrives at the station..
13:47:42 <Celestar> Roujin: er .. there's some shit I did apparently with the display
13:47:45 <Celestar> Roujin: will fix soon
13:48:18 <Wolf01> hailstorm.. the front glass of my car is broken :(
13:50:37 <Gekz> Wolf01: I doubt thats the only damage
13:50:46 <Gekz> I'd say your car just got written off, aesthetically
13:50:56 <Rubidium> Celestar: fix where? in the hg? I'll test it later today
13:51:52 <Wolf01> my mother didn't stay under the storm to look at the damages, we'll look when she'll return home
13:53:38 <Celestar> Rubidium: I'll test it for a minute, then I'll commit it to hg (in half an hour or so)
14:02:33 <CIA-5> OpenTTD: smatz * r14022 /trunk/src/station_cmd.cpp: -Fix (r13957) [YAPP]: Removing station tiles could produce stale reservations, too. (michi_cc)
14:02:51 <Celestar> people you're going to MURDER me :P
14:04:21 <Celestar> std::vector<std::map<StationID, int> > cargo_left;
14:04:49 <CIA-5> OpenTTD: smatz * r14021 /trunk/src/ (station_cmd.cpp train_cmd.cpp): -Fix (r13957) [YAPP]: Overbuilding station tiles with non-track tiles could produce stale reservations. (michi_cc)
14:05:25 <Celestar> glx: at least the C-guys are going to murder me for that
14:05:41 <Belugas> would not make it easier if access is granted to michi_cc ? hehe
14:06:02 <SmatZ> I need commits to have good commit statistics :-P
14:07:06 <glx> Celestar: it's just a nice c++ construct ;)
14:10:49 <Belugas> SmatZ, just change some magic numbers to enums :D LOTS to do ;)
14:10:51 <Celestar> glx: I just notice, whenever I use this std containers, I hardly ever produce segfaults, buffer overflows and other unwanted effects
14:11:07 <Celestar> yEAH! let's make it a numerical-free code
14:11:24 <glx> Celestar: then you don't play with iterators ;)
14:11:39 <Celestar> I've seen things like for (int i == NC_ZERO; i < sizeofsomething, i++) in real life
14:11:48 <Celestar> glx: well, what do you mean by "play"
14:12:12 <glx> delete objects while using an iterator
14:13:16 <Celestar> glx: er .. for (someit it = container.begin; it != container.end; ++it) { do_some_stuff; delete *it; }
14:14:11 <glx> things like that yes, but if not done correctly the iterator becomes invalid
14:14:30 <Celestar> glx: it basically always becomes invalid (=
14:14:37 <Celestar> glx: at least that's my rule of thumb
14:14:41 <SmatZ> doesn't vector / everything call destructor, when you .remove() it?
14:14:44 <Celestar> just use a next pointer or something
14:15:03 *** Singaporekid has joined #openttd
14:15:04 <Celestar> SmatZ: it does, but the iterator is crapped as soon as you remove it's contents
14:15:58 <Celestar> the only place where I do it is when I break; the loop right after the removal
14:16:07 <SmatZ> but they you get segfault on vector destructor
14:16:35 <Celestar> SmatZ: deleting stuff from a vector is pretty evil anyway
14:17:03 <SmatZ> or simply vector.remove(*it++) ?
14:17:23 * SmatZ mumbles something about STL and evilish things...
14:17:49 <Celestar> SmatZ: normally, doing vector.erase does the right thing
14:18:05 <Celestar> and sets the iterator to the right thing (i.e. the next element, or .end())
14:18:39 <Celestar> all the standard containers do this afaik
14:19:09 <glx> vector.erase(it) does that IIRC, but vector.erase(value) doesn't
14:19:27 <Celestar> glx: afaik you are correct
14:20:04 <Celestar> glx: then again, how should vector.erase(value) have an idea that you have some iterator pointing to it?
14:20:19 <Celestar> peter1138: you can pull the fifo fix
14:20:37 <SmatZ> Celestar: maybe vector::erase(iterator &it) ?
14:21:02 <glx> it can't (we had some fun with noai for that)
14:21:10 <Celestar> SmatZ: vector.erase(it) is fine
14:21:22 <Celestar> SmatZ: just all the OTHER iterators (besides it) will be crapped
14:21:34 <SmatZ> Celestar: I fully agree
14:21:50 <Celestar> I think for all other containers, just other iterators pointing the same element will be crapped
14:22:40 <Celestar> oh man. yapp is awesome
14:22:50 <Celestar> it's exactly the signalling system that TTO should have had
14:23:19 <SmatZ> Celestar found YAPP, and that's the end of cargopacket development ...
14:23:24 *** grumbel has joined #openttd
14:24:13 <Celestar> SmatZ: er .. I found YAPP before peter1138 and I started on cargodest
14:24:36 <Celestar> ok who's in for 90 minutes of cargodest playtest :P
14:25:43 <Celestar> ok first I needa compile a new server
14:25:51 <Celestar> glx: should we use newgrfs or not? (=
14:26:25 <Celestar> Roujin: where is that tt forum entry about cargopackets of yours?
14:26:45 <Celestar> glx: *sigh* maybe it will be 60 minuets only. we'll see
14:26:50 <Celestar> glx: but the server will be running anyway
14:28:28 <Celestar> glx: in some std::container anyway (=
14:28:47 <Celestar> glx: can you reproduce?
14:30:55 <Celestar> glx: the playtest will have to wait ^^
14:30:57 <glx> (if I understand the drmingw stack trace)
14:32:14 <glx> >,...StationID next_station = GetVehicleOrder(v, v->cur_order_index)->GetNextUnloadingOrder()->GetDestination();
14:32:40 <glx> when doing that, the segfault happens in GetDestination()
14:33:05 <Celestar> I have to rebuild without inlines
14:33:14 <Celestar> glx: you guys will need to playtest over the weekend
14:33:35 <Roujin> using the 5th-newest entry from your repository atm
14:34:09 <Roujin> will it break my savegame or anything if I upgrade?
14:36:40 <Celestar> make backups nevertheless
14:36:45 <Roujin> lol, i just got a "Cost: .." instead of "Income:.." on a bus dropping off passengers at a stop
14:37:36 <Roujin> uhm.. dunno.. i think it was not 0, but it was away so fast...
14:37:42 <Roujin> i'll see if it happens again
14:38:07 <glx> btw you can lose money (same as with transfer)
14:38:42 <Celestar> Roujin: yeah I know :(
14:40:37 <Roujin> didn't know it can happen
14:40:46 <Celestar> Roujin: it can with transfers
14:41:05 <glx> Celestar: GetNextUnloadOrder() returns NULL
14:41:12 <Celestar> glx: already fixed, commited and pushed (=
14:41:27 <Celestar> I'll be offline for 48 hours, so you need to drop me a mail to celestar@openttd.org (=
14:43:35 <Belugas> and don't danse to much :) it will get you away from coding ^_^
15:00:45 <CIA-5> OpenTTD: smatz * r14023 /trunk/src/yapf/yapf_costrail.hpp: -Fix (r13944) [YAPP]: If YAPF encountered a dead-end that had a signal on it, is could be falsely seen as a valid destination. (michi_cc)
15:02:33 <CIA-5> OpenTTD: smatz * r14024 /trunk/src/yapf/yapf_costrail.hpp: -Fix (r14019) [YAPP]: YAPF also skips tunnels and bridges. Honor that in penalty calculations. (michi_cc)
15:16:13 *** frosch123 has joined #openttd
15:40:36 *** thingwath has joined #openttd
15:42:47 <CIA-5> OpenTTD: glx * r14025 /branches/noai/ (49 files in 11 dirs): [NoAI] -Sync with trunk r13968:14024
15:49:30 <CIA-5> OpenTTD: glx * r14026 /trunk/src/ (network/core/host.cpp player_gui.cpp): -Cleanup (r13995, r14004): trailing whitespaces
15:54:13 <Eddi|zuHause> try to remove any lines like: size="10,10!"
16:14:22 <ben_goodger> good lord, other people use graphviz?
16:16:23 <TinoDidriksen> Of course...it's a great tool.
16:17:37 <ben_goodger> well, I know that, but I've never heard of anyone else using it besides for programmatically generated graphs of debian package dependencies
16:17:49 <ben_goodger> admittedly, this graph of celestar's is probably programmatically generated also, but it's too small to see..
16:19:58 <Eddi|zuHause> my university uses a graphviz plugin for trac to draw ticket dependencies
16:20:36 <Eddi|zuHause> and i use graphviz for my abstraction-visualization in my diploma thesis
16:28:50 <Eddi|zuHause> www.informatik.uni-halle.de/~krause/swt-beamer-vorlage.pdf <- look at slide 20
16:48:28 *** Progman has joined #openttd
16:53:30 <Dred_furst> I just had an idea to extend yapp/yapf to make rail networks run slightly smoother
16:54:26 <Dred_furst> get trains to deaccellerate/accellerate in accordance to traffic in front of them
16:55:06 <Eddi|zuHause> that's exactly the solution...
16:55:28 <Dred_furst> my idea was that if you have a really busy junction you should slow the train down a bit so it hits the junction in a gap rather than having to stop
16:56:19 <Dred_furst> if you do this over a variable number of squares things could run really smoothly
16:56:52 <Dred_furst> like signal 4 squares ahead is red, cut speed to 75%
17:03:21 <Yorick> why did it get removed from openttd?
17:03:39 <Dred_furst> that way when your train hits a busy junction it will slow down and try and hit the junction in a gap
17:03:51 <Dred_furst> rather than hitting the signal just before and going STOOOOOOOP
17:04:41 <Dred_furst> which means everything keeps flowing
17:10:56 <Eddi|zuHause> when do you think it was IN openttd?
17:10:57 <peter1138> Oh, well to be removed it would have had to have been added.
17:11:20 <ben_goodger> hence the source of confusion
17:14:51 <peter1138> TortoiseHg has a nice changeset view showing branches and merges. Is there anything like that for non-Windows?
17:15:04 <peter1138> Or will hg view work fine :D
17:15:47 <peter1138> Yes, it was a patch.
17:15:56 <Rubidium> yup, we removed passenger destinations also 4 times now?
17:16:07 <Roujin> that's not _in_ openttd...
17:16:48 *** [com]buster has joined #openttd
17:17:03 <Yorick> I remember something like that with the openttdcoop 0.4.5-game
17:18:11 <Dred_furst> someone get yellow signal state working with YAPF and YAPP effectively
17:18:43 <Belugas> Yorick, you're just so wrong...
17:18:55 <Rubidium> oh... so paxdest was really in OpenTTD because there was a paxdest openttdcoop game (attempt)
17:25:48 <peter1138> Bah, hg view doesn't automatically update :(
17:26:33 <Progman> Celestar, peter1138: is there some cool updates of your paxdest I can pull? got h:a9505f atm
17:27:41 * Yorick loves celestar too, don't worry :)
17:27:41 <peter1138> a9505f? How ancient!
17:31:22 * peter1138 ponders setting up a server
17:31:54 <peter1138> hg: unknown command 'branch'
17:34:10 <peter1138> I shall comment that line out, as we've not used branches anyway...
17:35:25 <Yorick> why not using the official openttd svn?
17:36:31 <Belugas> because it's not required
17:37:47 <peter1138> Indeed. Anyone can do this.
17:47:16 <peter1138> Whoops, misspelled cargo... never mind.
17:47:41 *** Brianetta has joined #openttd
17:47:57 <peter1138> fonso, will look in moment.
17:52:50 <peter1138> Eddi|zuHause, coded it yet?
17:53:08 <Eddi|zuHause> then i would have posted in development, not suggestions :p
17:54:33 <Belugas> hylje thinks it's brain dead ... as in... cant' understand it? don't want to understand it, way too easy for the task?
17:55:01 <fonso> whatever, he didn't elaborate
17:55:28 <Belugas> you should not have reacted on such a comment ;)
17:55:45 <Belugas> if you do , you would keep on working until the end of times ;)
17:55:55 <fonso> well, I also like the java style iterators more
17:56:07 <fonso> compare the code and you'll see why
17:56:33 <Belugas> to be hmonest, i've got some other code to check, right now...
17:57:38 <fonso> ok, I'm patient. At least there is a new version of the patch to look at when you got time.
18:00:40 <peter1138> Nah, michi_cc was patient.
18:01:01 <fonso> OK, I'm a little impatient at times
18:01:05 *** fonso is now known as fonsinchen
18:01:41 <Eddi|zuHause> i have never seen michi_cc say anything here except "new win64 binary uploaded"
18:02:12 * fonsinchen wraps herself in silence
18:03:49 <XeryusTC> where can the new cargo destinations be found?
18:04:39 <Eddi|zuHause> the new new cargo destinations?
18:07:08 <Roujin> questions at cargodest guys: what do blue lines in the cargo map mean?
18:07:24 <Eddi|zuHause> lines are coloured by cargo type
18:07:35 <Eddi|zuHause> blue==passengers afaik
18:08:06 <XeryusTC> Eddi|zuHause: thank you :)
18:10:20 <Roujin> before, I used trains with passenger and mail carriages
18:10:32 <Roujin> then the lines were white
18:11:05 <Roujin> now I changed part of the network with EMUs that carry only passengers, and the lines there became blue..
18:11:42 <Eddi|zuHause> maybe you should bug peter1138 to colour the lines by the majority of cargo transported ;)
18:14:51 *** Yorick_ has joined #openttd
18:14:54 *** Yorick_ is now known as Yorick
18:16:46 *** SquireJames5 has left #openttd
18:17:55 <peter1138> Bah, now continuing the game that was running on Celestar's server :D
18:18:32 <peter1138> 23859 petern 15 0 40616 14m 3128 R 79.5 2.9 0:52.00 openttd
18:25:09 *** LaSeandre has joined #openttd
18:25:30 *** LaSeandre has left #openttd
18:39:55 <frosch123> goergia? shouldn't they be sleeping?
18:41:02 <fonsinchen> depends on which georgia he means
18:43:06 <Eddi|zuHause> the one south of russia i assume...
18:43:10 <Eddi|zuHause> the one that is in the news
18:44:36 <frosch123> okay I put it way too far east, there it is currently about 21:45
18:44:41 <Belugas> barbarism strikes again the world :(
18:45:13 <Eddi|zuHause> well... while the world is distracted with china and the olympics... nobody will notice ;)
18:50:58 <Wolf01> is there an hotkey for the convert signals tool on the signals gui?
18:51:53 <Eddi|zuHause> i wondered that myself... it is really needed
18:52:20 <Eddi|zuHause> it is as significant as the remove button (R)
18:52:41 <Progman> what about ctrl-click on the signal itself?
18:53:25 <Wolf01> ctrl+click already convert from semaphores to light signals
18:53:50 <Yorick> just click on the signal itself?
18:54:08 <Yorick> I think that is the most intuitive
18:54:35 <Wolf01> I need to toggle quickly the tool, like R for the remove button, as Eddi says
18:54:46 <Yorick> argh, the paxdest doesn't build on windows
18:55:32 <Yorick> I dislike installing and compiling half of the world just to use a paxdest
18:55:48 <glx> you just need some headers
18:55:53 <Progman> just ctrl-click on a signal and it switches between presignals and yapp-signals
18:56:06 <Eddi|zuHause> you only need the boost graph library
18:56:40 <Eddi|zuHause> Progman: not when you set to cycle only between one type
18:57:25 <Eddi|zuHause> Progman: the convert button is there. it is good where it is. the behaviour is good how it is. it just needs a hotkey.
18:57:50 <Wolf01> I think that "C" would be the best one, but we need to steal the event from the rail toolbar
18:58:27 <Progman> but there are 6 different types, what kind of hotkey do you want?
18:58:49 <Eddi|zuHause> Progman: for the CONVERT button.
18:58:55 <Eddi|zuHause> there is only one such button
18:59:02 <Eddi|zuHause> not a hotkey for signal type
18:59:48 <Yorick> what about having that convert button always enabled, and if it's the correct signal type you're clicking, perform the standard operations, else convert
18:59:50 <Wolf01> for signals we might use shift + QWERTY/ASDFGH
19:00:05 <Prof_Frink> Eddi|zuHause: Building a different type of signal over an existing signal should convert it to that type
19:00:19 <Progman> how often do you need the convert button o_O
19:01:00 <Eddi|zuHause> often enough to get annoyed by it not having a hotkey
19:01:51 <Yorick> also I don't like the fact it remembers you've selected a pre- or combo-signal
19:02:05 <Yorick> those are often one-time use
19:02:09 <Progman> hmm, I dont understand this, maybe because I dont play with semaphores ;)
19:02:24 *** XeryusTC is now known as PublicServer
19:03:17 <Wolf01> that's not question about semaphores or not... we need it to replace signals quickly without cycling
19:03:18 *** PublicServer is now known as XeryusTC
19:03:40 *** LilDood has joined #openttd
19:03:43 *** XeryusTC is now known as PublicServer
19:04:56 *** PublicServer is now known as XeryusTC
19:07:52 *** XeryusTC is now known as PublicServer
19:08:19 <Wolf01> what's the best penalty to avoid 90° turns with yapf?
19:09:00 <peter1138> When is euromillions drawn? heh
19:09:19 *** PublicServer is now known as XeryusTC
19:10:07 <Belugas> [15:03] <Wolf01> that's not question about semaphores or not... we need it to replace signals quickly without cycling <--- hehehe "we" :D I don't, personnaly, YOU do ;)
19:14:18 *** Yorick was kicked by Belugas (LIAR!!)
19:15:36 <Yorick> "copying 7.089 items...5 minutes remaining"
19:16:08 * Yorick is currently installing half the world to get paxdest
19:17:09 * peter1138 ponders trying to solve his toolbar issue, or just giving up.
19:18:50 <Belugas> put it on the shelves for a little while, you might find a solution later :)
19:19:16 <peter1138> Belugas, but I need it for my other WIP...
19:19:21 <Yorick> toolbars don't give issues, toolbars are perfect!
19:21:47 <Belugas> will you please mind your businessm Yorick?
19:24:25 <Yorick> hi, what's your question?
19:24:26 <bastl> i installed dedicated server on debian
19:24:40 <bastl> but now i cant find the openttd.cfg
19:24:59 <Yorick> create one at the location of the server executable
19:25:10 <Yorick> otherwise it creates one at the documents
19:25:45 <Yorick> /user/openttd, I think
19:25:55 <Belugas> bastl, have you run the game at least once?
19:26:11 <bastl> i run the dedicated server
19:26:46 <Yorick> have you ran it at least once?
19:27:13 <bastl> but its still running atm
19:27:26 <Yorick> try doing saveconfig in the console
19:29:03 <Yorick> grep: invalid option -- m
19:31:50 <bastl> if i create a new cfg do i need to set all values or just the ones i want to change? and which value do i need to change to start the server with random 1024x1024 maps?
19:32:30 <glx> bastl: check in ~/.openttd
19:32:53 <peter1138> map_x / map_y to change the map size.
19:36:15 *** Belugas_Gone has joined #openttd
19:36:15 *** ChanServ sets mode: +o Belugas_Gone
19:38:14 *** Belugas_Gone is now known as Belugas
19:38:57 <Belugas> i lost my mouse... needed to reboot
19:41:52 <XeryusTC> the sign list has a bug :o
19:42:02 *** LaSeandre has joined #openttd
19:42:11 *** LaSeandre has left #openttd
19:42:19 <XeryusTC> or well, it does in r14013, dunno if it is fixed in the meanwhile
19:56:04 *** Reemo is now known as try_to_get_drunk
19:56:14 <Progman> Celestar, peter1138: a route is somehow dropped all the time :(
19:57:14 <Progman> do you need a savegame?
19:59:30 <Progman> found this bug after a train goes to a depot for maintance
20:27:02 <Progman> Celestar, peter1138: autoreplace totally failes with paxdest, doesn't it?
20:27:33 <Eddi|zuHause> autoreplace totally fails.
20:29:41 <Progman> and you must reload the network by save&load all the time :(
20:32:47 <Progman> patch they working on
20:33:36 <Bjarni> didn't notice any autoreplace problems when I tested it
20:33:59 <Bjarni> but I guess I never tried to replace anything with cargo in it
20:36:53 <Eddi|zuHause> hm... the fan just got silent, this can't be good...
20:37:22 <Eddi|zuHause> i was thinking the north bridge fan
20:38:22 <Eddi|zuHause> i can't see anything...
20:40:08 <Eddi|zuHause> all 4 fans appear to be running...
20:43:41 <Eddi|zuHause> PSU, CPU, GPU, north bridge
20:43:56 <Bjarni> what is "north bridge"?
20:44:42 <Eddi|zuHause> north bridge is the chip that controls communication between the CPU and the important components on the mainboard (RAM, AGP, south bridge)
20:46:00 <Eddi|zuHause> originally i had another fan on the case, but that got de-centered and was very loud
20:48:50 <Belugas> "hello av jus signed onto this site coz big transport tycoon fan but i need someones help with ttdalter but dont no wer to go or who to say it to could someone help me please"fuck...
20:49:47 <Bjarni> you have a fan on the north bridge?
20:51:15 <Eddi|zuHause> yeah, north bridge fans are prone to be noisy because they are much smaller than CPU fans
20:51:45 <glx> and usually you can't put a bigger one because the graphic card is in the way
20:52:17 <Bjarni> we should decrease power used in computers and now you tell me that you guys have active cooling on the bridge???
20:53:28 <Eddi|zuHause> i have the suspicion that, despite the active cooling, the north bridge might be responsible for my heat problems
20:54:19 *** ben_goodger has joined #openttd
20:54:32 <Bjarni> there is a fundamental flaw in computer cooling
20:54:53 <Bjarni> all the heatsinks and fans moves the heat into the air in the case
20:55:17 <Bjarni> cooling elsewhere (like machinery) moves the heat out of the case
20:55:59 <glx> cooling is worse if I open the case
20:56:27 <ben_goodger> Bjarni: the case fans then remove the hot air from the case
20:57:25 <ben_goodger> where is the flaw exactly?
21:01:02 <Bjarni> the flaw is that the case temperature raises way more than needed
21:01:35 <Progman> wth, why is the hotkey F used for "delete all orders" if G is the hotkey for "Goto"?
21:01:59 <glx> because it needs to be changed
21:02:14 * Belugas is afraid to imagine waht a computer designed by Bjarni would look like, to correct the "design-flaw" :)
21:02:19 <Bjarni> if the room temperature is 20°C and the case temperature is 30°C, then (in theory) you would be able to lower your CPU temperature by 10°C if it were able to export the heat directly to the external air
21:02:35 <glx> and the keys followed the buttons order IIRC
21:03:36 <Bjarni> Belugas: I didn't tell about a solution but the problem
21:03:46 <Eddi|zuHause> i don't understand the hotkeys on the order list anyway
21:04:24 <Bjarni> the real issue at hand is that we need to get rid of a whole lot of heat
21:04:29 <Eddi|zuHause> especially when they override keys from the toolbar, even though the order lists are nowhere near the top (e.g. hidden behind timetables)
21:04:50 <Bjarni> we should lower the heat production instead since it would also lower the power usage
21:05:24 <Progman> its okay to have G for Goto but its not okay to have F delete the whole order list
21:05:26 <Eddi|zuHause> where is that configurable hotkeys patch coming anyway?
21:05:42 <Progman> who implemented this crap?
21:06:17 <Prof_Frink> Progman: Well, it's OK to have F delete the whole order list... if you have a confirmation dialogue.
21:06:17 <Bjarni> F = Fault (you made a mistake and want to get rid of all of it before anybody sees what crap you just wrote)
21:06:57 <Eddi|zuHause> Bjarni always mistypes that and hits the "commit" button instead :p
21:07:47 <Bjarni> but I don't have a commit button
21:08:17 <Bjarni> CLI is very useful as you can do all the stuff you like
21:08:35 <Bjarni> in a quick and scriptable way
21:09:02 <Bjarni> also I have yet to find a working SVN GUI frontend. Everything I tried was flawed and useless
21:20:59 <glx> Zuu: there's a little typo for the check box in edit window
21:21:47 <Zuu> Still doing release-tasks.. takes time.
21:22:17 <Zuu> Nope, users.tt-forums.net
21:22:33 <glx> should be better than sf ;)
21:24:24 <Zuu> Have uploaded a new website for the program, and have to change so clients get to that website and not to the tt-forums when they click on the website-button in the about-dialog in 1.1.
21:24:48 <Zuu> And then decide wheither I should spam people with a popup about the alpha or not. :D
21:26:16 <Roujin> you guys were talking about autoreplace some time ago...
21:26:49 <Roujin> ... it also didn't work for engines in my game, but it worked for carriages
21:27:16 <Roujin> - just wanted to inform you. Going to bed now, gnite
21:31:31 *** Yorick_ has joined #openttd
21:31:36 *** Yorick_ is now known as Yorick
21:42:34 *** Yorick_ has joined #openttd
21:42:45 *** Yorick is now known as Guest858
21:42:46 *** Yorick_ is now known as Yorick
22:03:43 *** welshdragon is now known as welshdra-gone
22:23:18 *** tamentis has joined #openttd
22:50:54 <Eddi|zuHause> i have a hornet in my room...
22:51:22 <Eddi|zuHause> it's actually louder than a north bridge fan :p
22:51:52 <ln> certainly, especially with the afterburner on.
22:53:02 <Eddi|zuHause> no kidding, i'm sure it has :p
23:12:02 * Bjarni notes that rejecting Eddi|zuHause's patches might be a good idea if his room is so full of bugs that they even appear in mid air
23:13:21 *** pete is now known as Guest871
23:13:42 <Guest871> hi. i'm having trouble running openttd
23:14:03 <Guest871> i've downloaded the 0.62 source and compiled
23:14:11 <ln> you need the data files from the CD
23:14:17 <Guest871> but when i run it goes through map generation and stops at
23:14:24 <Guest871> dbg: [net] Map generated, starting game
23:14:29 <Guest871> i have the data files
23:14:44 <Bjarni> otherwise you couldn't even get to main menu
23:14:50 <Guest871> i can't get the menu
23:15:09 <Guest871> "dbg: [net] Map generated, starting game" is the last message shown on the console, but no menu appears
23:15:16 <Zuu> Guest871: so you start with -g?
23:15:41 <Guest871> i'm not using any command line switches
23:15:55 <Eddi|zuHause> you have forgotten to install sdl-devel [or similar] and ignored the big warning from ./configure
23:15:56 <Bjarni> it skips the menu and starts a game right away
23:17:15 <Eddi|zuHause> the warning that says "no video driver found, building dedicated only".
23:17:43 <Bjarni> we could ask for config.log
23:17:48 <Guest871> i saw that, it came immediately after a warning about windows only, so i thought it was a windows only warning
23:17:51 <Zuu> Hopefully you don't have to make a system update of hundreds of packages with Yum as I had to do to get sdl-devel on a CentoOS computer that hadn't been updated for ages. :)
23:18:12 <Guest871> htis is a pretty up to date ubuntu, so i shouldn't have any problems
23:19:05 <Guest871> recompiling now, should have results in a few minutes
23:19:47 <Eddi|zuHause> imho, that warning should be changed into an error, the few people that really want a dedicated build can configure that manually
23:20:07 <Zuu> Guest871: Check in aptitude (or synaptics) that you have sdl-devel installed? Or you already have done that and installed sdl-devel?
23:20:26 <Guest871> i've installed sdl-devel and recompiled
23:20:29 <Guest871> everything works now
23:20:34 <Guest871> thanks for your help guys
23:20:47 <Eddi|zuHause> i told you so ;)
23:20:58 <Zuu> Now you only have to learn how to become a Tycoon :)
23:21:17 <Guest871> i've been playing 0.60 for the last few days
23:21:26 <Guest871> because that's the version in the ubuntu repo
23:21:40 <Guest871> but i needed to update to 0.62 to connect to a server some friends of mine are running
23:21:46 <Zuu> Okay, just a kind notice. It's 0.6.0 not 0.60.
23:22:07 <Bjarni> 0.6.2 is the same as 0.6.0 except for some bug fixes
23:22:27 <Guest871> but different enough that i can't connect to a 0.6.2 server with the 0.6.0 client
23:22:54 <Bjarni> because if the same command is executed on both computers then they should behave the same
23:23:15 <Zuu> The client and server must be of exactly the same version. (except for some minor diferences that experts can handle)
23:23:17 <Bjarni> but if one of them is buggy and the other one handles it as intended then you desync
23:26:52 <guru3> quick question, does anything bad happen if a thermometer shoots all the way up to the top?
23:27:53 <Bjarni> you mean a mercury one?
23:28:10 <Bjarni> in fact I know it can break the glass
23:28:53 <Bjarni> hopefully you aren't measuring the temperature of a person ;)
23:28:59 <guru3> well i think it's an alchohol one
23:29:06 <guru3> i just put it on the heatsink of this thin client
23:29:16 <guru3> and some few mm above the markings
23:29:27 <guru3> and when i pulled it off that's where the alchohol was
23:29:28 <Bjarni> get it away from there
23:30:09 <guru3> right i'll file this one away under "not clever"
23:30:14 <Bjarni> it can also cause bobbles to appear
23:31:01 <guru3> i haven't seen any of those yet
23:31:04 <Bjarni> listen to what I'm intend to write :P
23:31:26 <glx> you played too much bobbles bubbles
23:31:31 <Bjarni> <guru3> i haven't seen any of those yet <- I think they appear while it cools down
23:32:01 <guru3> the question to follow this all up is... do you suppose a passive heatsink is too hot at 55C for a 300mhz amd geode?
23:32:03 <Bjarni> if it happens then the solution is to heat it up again and hope they will go away
23:32:28 <glx> guru3: what is the cpu temp?
23:32:46 <guru3> hot enough that thermometer on top goes to the top of the glass? (past 50C)
23:32:49 <glx> and what is the max temp it can handle?
23:32:52 <guru3> i can't measure cpu temp
23:32:55 <guru3> and i have no clue what the max is
23:32:57 <Eddi|zuHause> like that thing they told me when i went to university: "The Professor says A, he actually means B, writes C on the blackboard, but D is correct."
23:33:17 <Bjarni> I would prefer a cooler heatsink
23:33:30 <guru3> yeah it's the bit that worries me
23:33:42 <guru3> but i had this running for a two or three weeks continuously before
23:33:46 <guru3> and it seems to still work fine
23:33:55 * glx won't put fingers on its GPU heatsink (GPU temp is 56C)
23:34:11 <glx> and it's passively cooled
23:34:24 <Bjarni> but the amount of heat that escapes from the heatsink is k(a-b) where a is the heatsink temperature and b is ambient temperature
23:34:41 <guru3> and i'm planning to leave one of these one while i go away on holiday with a webcam as a security box
23:34:54 <Bjarni> so naturally it has to be somewhat warmer than the air
23:36:07 <Bjarni> <guru3> and i'm planning to leave one of these one while i go away on holiday with a webcam as a security box <-- somebody did that in Denmark (he went to work) and when he got back he had a video of the burglars who had visited him more than once
23:36:25 <guru3> a friend that lives about 10 mins away from me had his laptop stolen
23:36:32 <guru3> i'll have my laptop with me
23:36:38 <Bjarni> they took more or less all electronic stuff except the computer that recorded them
23:36:41 <guru3> but i'd prefer my desktop remained firmly seated in my room
23:36:56 <guru3> i was thinking of leaving it hooked up with a cellphone to send me smses occaisonally
23:37:13 <Bjarni> do you have a GPS in the cellphone?
23:37:28 <guru3> the battery doesn't even hold a charge
23:37:44 <Bjarni> you can charge it from USB
23:38:19 <Bjarni> you know you can buy cables that charges cellphones with power from USB ports
23:38:35 <guru3> i was thinking of hacking a power cable to do it though ><
23:38:49 <guru3> some wire strippers and electrical tape work wonders
23:38:55 <Bjarni> getting a real USB cable might do a better job
23:39:02 *** Wezz6400 has joined #openttd
23:39:14 <Bjarni> you can actually buy PCI USB cards with USB inside the case
23:39:27 <guru3> this is a thin client though, there's not that much space inside the case
23:39:27 <Bjarni> I never figured out what they would be useful for before now
23:39:39 <guru3> (there IS a pci slot, but that's beside the point)
23:40:35 <ln> there's no guarantee a cellphone would work inside the case.
23:40:38 <Bjarni> some woman got her camera stolen in Florida. After it was stolen, then camera detected an open wifi access point, woke up and transmitted all the pictures to the owner's computer
23:41:19 <guru3> my idea is to try and get it to send a mms before it gets taken away
23:41:20 <Bjarni> not only did she get all her pictures back, she also got pictures of the thieves, who had taken pictures of themselves when testing the camera
23:41:28 <guru3> not sure if it can send mmses though
23:42:20 <Bjarni> the police got the pictures and went "ohh... those guys. We should pay them a visit" and they still had the camera and some other stolen stuff when the police arrived
23:42:32 <Bjarni> I want a camera like that
23:42:57 <Bjarni> if it's stolen, then it sends pictures of the thieves back home so the police will know who did it
23:43:06 <Bjarni> now all it needs to get is GPS
23:43:58 <ln> Bjarni's stolen camera must be #1 on the cops' priority list.
23:45:46 <Bjarni> luckily nobody ever stole my camera
continue to next day ⏵
